8474 F~ ~~~~~~~~ F 2fLE4 CO?Y China's Rural Industry Structure, Development, i nd Reforrm editcd by William A. Byrd and Liii Qingsong I \N'o -I J d Ba n It vwa r(} I i hibiic.li t ioi China's Rural Industry China's Rural Industry Structure, Development, and Reform Edited by William A. Byrd and Lin Qingsong PUBLISHED FOR THE WORLD BANK Oxford University Press Oxford University Press OXFORD INEW YORK TORONTO DELI-Il BOMBAY CALCUTTA MADRAS KARACHI PETALING JAYA SINGAPORE HONG KONG TOKYO NAIROBI DAR ES SALAAM CAPE TOWN MELBOURNE AUCKLAND and associated companies in BERLIN IBADAN C) 1990 The International Bank for Reconstruction and Development / THE WORLD BANK 1818 H Street, N.W., Washington, D.C. 20433, U.S.A. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior permission of Oxford University Press. Manufactured in the United States of America First printing February 1990 The findings, interpretations, and conclusions expressed in this study are entirely those of the authors and should not be attributed in any manner to the World Bank, to its affiliated organizations, or to members of its Board of Executive Directors or the countries they represent. The maps that accompany the text have been prepared solely for the convenience of the reader; the designations and presentation of material in them do not imply the expression of any opinion whatsoever on the part of the World Bank, its affiliates, or its Board or member countries concerning the legal status of any country, territory, city, or area, or of the authorities thereof, or conceming the delimitation of its boundaries or its national affiliation. The Chinese characters used on the cover and throughout this book stand for "rural industry." Library of Congress Cataloging-in-Publication Data China's rural industry: structure, development, and reform / edited by William A. Byrd, Qingsong Lin. p. cm. Includes bibliographical references. ISBN 0-19-520822-6 1. China-Industries, Rural. 2. Industry and state-China. 3. China-Economic policy--1976- 1. Byrd, William. 11. Lin, Qingsong. HC427.92.C46553 1989 338.951 '009173'4-dc20 89-26665 CIP Contents Preface vii Abbreviations and Terminology xi Contributors xiii I China's Rural Industry: An Introduction 3 William A. Byrd and Lin Qingsong 2 Research Design, Methodology, and Data 19 William A. Byrd and Lin Qingsong PART I. OVERVIEW 41 3 Causes of Rapid Rural Industrial Development 47 Du Haiyan 4 Development Patterns in Four Counties 63 Jan Svejnar and Josephine Woo 5 Market Interactions and Industrial Structure 85 William A. Byrd and N. Zhu 6 Development Issues and Policy Choices 112 He Jiacheng PART II. OWNERSHIP AND RELATED ISSUES 129 7 Ownership and Status Stratification 134 Luo Xiaopeng 8 Private Enterprises: Their Emergence, Rapid Growth, and Problems 172 Lin Qingsong 9 Entrepreneurship, Capital, and Ownership 189 William A. Byrd v vi CONTENTS PART III. PERFORMANCE 219 10 Capital Formation and Utilization 222 Wang Xiaolu 11 Productive Efficiency and Employment 243 Jan Svejnar 12 Regional Imbalances 255 Wang Tuoyu PART IV. LABOR 275 13 TVP Workers' Incomes, Incentives, and Attitudes 280 Alan Gelb 14 The Rural Labor Market. 299 Meng Xin 15 Noneconomic Determinants of Workers' Incomes 323 Wu Quhui, Wang Hansheng, and Xu Xinxin PART V. THE ROLE OF COMMUNITY GOVERNMENTS 339 16 The Role of Township Govemments in Rural Industrialization 342 Song Lina and Du He 17 Why Industrialize? The Incentives for Rural Community Govemments 358 William A. Byrd and Alan Gelb PART VI. COMPARAT'IVE PERSPECTIVES 389 18 Convergence: A Comparison of Township Firms and Local State Enterprises 392 Song Lina 19 Chinese 7VPs in an International Perspective 413 Alan Gelb and Jan Svejnar References 427 Index 433 Preface The emergence and recent rapid growth of rural nonagricultural enterprises (township, village, and private enterprises-TVPS) in China is a striking and in many ways unique phenomenon. Almost ovemight, industrial activity has expanded tremendously and enterprises have proliferated. These changes have brought about large increases in rural personal incomes and significant shifts in the structure of the rural labor force. China's TVP sector has several noteworthy features. * Dynamism. Output has been doubling every three to four years, and hundreds of thousands of new firms have emerged each year on average. * Competitiveness. The sector has gained a large share of the domestic market for many industrial goods and has penetrated some export mar- kets as well. a Small scale. Although enterprises of substantial size exist, especially in the more developed areas, most firms are small concerns with only a few score employees and about Y100,000 in assets. a Diversity. There is great geographic variety in ownership, institutional arrangements, level of development, and degree of industrialization. * Outward orientation. Most output, especially of manufactured goods, is sold outside the community or locality in which it is produced. * Community orientation. Nearly all Tvps are tied in complex ways to their rural communities, and they would not consider relocating. * Factor immobility. Like the enterprises themselves, factors of production and human talent are largely immobile across localities and communi- ties. These characteristics will be analyzed in detail in this volume. Despite considerable scholarly work on China's rural industry in the 1970s and later (see chapter 2), little research of a comprehensive nature has been done. In particular, the rapid development of the TVP sector since 1984 has not been adequately documented and analyzed. In late 1985 the Institute of vii viii PREFACE Economics of the Chinese Academy of Social Sciences (CASS) and the World Bank agreed to undertake a collaborative study of China's industrial TVP sector that would involve extensive data collection, joint fieldwork, shared data processing and analysis, international comparative research, and sponsorship of an international conference. This research followed earlier collaborative work on Chinese state-owned industrial enterprises (see Tidrick and Chen 1987). The main objectives of the project were (a) to gain a better understanding of how China's industrial TVP sector actually functions, (b) to leam about the rural communal institutional and regulatory framework within which it operates, (c) to ascertain the reasons behind the rapid growth of the sector, particularly since 1984, (d) to identify and analyze the problems and obsta- cles faced by the TVP sector as it continues to evolve and develop, (e) to evaluate on an informed basis different institutional and policy options for the future, and (f to shed light on the TVP sector from an international perspective. The main topics covered by the research included the perfor- mance of the TVP sector; the -rvp ownership system; roles and motivations of rural community authorities; the market environment and business prac- tices of TVPS; the allocation, circulation, and utilization of key factors of production; labor and wage systems; and regional variations in TVP develop- ment and rural industrialization. Details on project objectives and research methods are provided in chapter 2. This volume contains the initial research output of the project. The papers were written by Chinese and World Bank researchers on the basis of an agreed set of topics and were presented at an international conference held in Beijing, November 4--7, 1987. All of the papers were revised and edited in light of conference discussions. In addition to the two introductory chap- ters there are ten chapters by Chinese authors and seven by World Bank staff and consultants. All of the former and all but one of the latter focus on vari- ous aspects of China's industrial TVP sector. Chapter 19 provides an interna- tional perspective by looking at firms elsewhere in the world that are similar in some ways to Chinese TVPS. Other fruits of the project have been pub- lished elsewhere or have been circulated internally within China and the World Bank. Given the abundant data collected as part of the project, it is hoped that more published papers and longer works will be forth,coming as the materials and data are analyzed in greater depth. Through this collaborative research project the World Bank has gained a deeper understanding of China's industrial TVP sector. The Institute of Eco- nomics of CASS and other participating Chinese research institutions have benefited from organizing and conducting the first comprehensive, systematic research on China's TVP sector-a project that has applied a wide range of analytical techniques to a large, multifaceted, and relatively complete data set. An international perspective on Tvps has also been gained from analysis of existing literature and through field visits to other countries. Finally, this research has provided a basis for analyzing the policy issues that face the TVP Preface ix sector and for evaluating different policy options. The project represents a further advance in collaboration between Chinese research institutions and international agencies, and we hope that it will lead to similar endeavors in the future. This collaborative research project would not have been possible without the help and active involvement of many people and organizations. We would especially like to thank the leaders and government officials of Wuxi, Jieshou, Nanhai, and Shangrao counties and the many TVP managers and workers who gave interviews and filled out questionnaires during the field investigations. Their assistance was invaluable and was willingly provided de- spite the substantial time and effort required. The research team invariably received the warmest hospitality when visiting the four counties. In addition, the help of government officials and of TVP managers and workers in Yuanping County (Shanxi Province) and Taihe County (Anhui Province) during related fieldwork is gratefully acknowledged. Officials of the TVP Bureau of the Ministry of Agriculture and the State Statistical Bureau also met with members of the research team and provided useful information on the national situation of the TVP sector. On the Chinese side, CAss and its Institute of Economics covered the bulk of financial expenditures. On the World Bank side, a grant from the World Bank research budget and operational funds from the China Division were the primary sources of financial support. A grant from the Ford Founda- tion financed international fieldwork and related costs. The financial assis- tance of all these institutions was essential for the success of this project and is acknowledged with thanks. Just as important as financing was the time of the members of the research team. In addition to CASS'S Institute of Economics and the China Division of the World Bank, the Development Research Institute of China's State Council Rural Development Research Center, the China Economic System Reform Research Institute, and the Sociology Department of Beijing Univer- sity contributed staff time. The project was organized under the overall supervision of, on the Chinese side, Liu Guoguang, vice president, CASS, and Dong Fureng, director, Insti- tute of Economics, cASs, and, on the World Bank side, first Fred Levy, se- nior economist, China Division, and subsequently Gene Tidrick, lead econo- mist, China Department. The respective team leaders for the project were Lin Qingsong, deputy director, Institute of Economics, CASS, and William Byrd, economist, Country Operations Division, China Department, World Bank. Others with major responsibilities for managing the research included He Jiacheng, chief, Development Division, Institute of Economics; Luo Xiaopeng, chief, Macroeconomics Division, Development Research Insti- tute, State Council Rural Development Research Center; and Alan Gelb, senior economist, Financial Policy Division, Country Economics Depart- ment, World Bank. In addition to the contributors to this volume, the following researchers X PREFACE participated in field investigations and other parts of the collaborative re- search project: Gu Xiuliri, Hontg Wei, Hou Xueshan, Li Hanlin, Shu Yan, Christine Wong, Yang Xiaodong, and Yang Xiaodong. Their help with re- search design, data gathering, and other aspects of the work was invaluable. Li Lanting, Li Xiaomei, Lu Xiaoheng, Ma Changshan, Edna Monaghan, Qu Ayang, Wang Shuwen, and Zhang Yi lent administrative and related support to the project, and their help is gratefully acknowledged. The translators of the conference papers were Zhang Ciyun and others from the China Transla- tion Corporation; Zhang Xiaogang and others from the China Daily; Dai Rui, He Baoyu, Wang Lina, aond Zhang Zhixian. In addition to several contrib- utors to this volume, a number of scholars helped organize the international field visits and related background work for the project: Baburao Baviskar, David Ellerman, Saul Estrin, Derek Jones, Y. Y. Kueh, Nicholas Lardy, Lin Yifu, Virginie Perotin, and Janez Prasnikar. An international panel consist- ing of Baburao Baviskar, Ronald Dore, Benjamin King, Dwight Perkins, and Martin Weitzman participated in the international conference and contrib- uted valuable comments. Chinese participants not involved with the project who attended the conference included Du Rensheng, Wang Guichen, Bao Youdi, Chen Xihu, Hua Jianchun, Ren Long, Zhang Yi, and Zhou Qiren; their participation and comments are also greatly appreciated. Abbreviations and Terminology Chinese terms, transliterated according to the Pinyin system, are in brackets. Abbreviations ABC Agricultural Bank of China [nongye yinhang] CASS Chinese Academy of Social Sciences GVAIO Gross value of agricultural and industrial output GVAO Gross value of agricultural output GvIO Gross value of industrial output LMF Labor-managed firm (Yugoslavia) LSE Local state enterprise [difang guoying qiyel PRS Production responsibility system (in agriculture) [baochandaohu or baogandaohu] RCC Rural credit cooperative [xinyong hezuoshe] TEC Township economic commission [xiang jingwei] TIC Township industrial corporation [xiang gongye gongsil. The township govemment organ that is directly responsible for supervising township enterprises. It has different names in different parts of the country and is sometimes combined with the township economic commission. If the two are not combined, the township industrial corporation is sub- ordinate to the township economic commission. TVCE Township and village community enterprise [xiangcun qiyel. Excludes enterprises run by production teams [shengchan duiban qiye]. TVP Rural nonstate enterprise (township, village, and private enterprise) [xiangzhen qiyel. Includes production team enterprises. y Yuan (currency unit). Recent average official exchange rates have been 2.30 yuan to the U.S. dollar in 1984, 2.94 in 1985, 3.45 in 1986, 3.72 in 1987, and 3.72 in 1988. (International Monetary Fund, International Financial Statistics.) xi Xii ABBREVIATIONS AND TERMINOLOGY Terminology Bonds [jizil. Short-term (one to two years) financial instruments issued to workers or local residents by a-vPs or community governments; usually non- transferable. Interest on bonds may be fixed, variable with enterprise prof- its, or a combination of the two. Community government [shequ zhengfuJ. Primarily refers to township [xiang] and village [cun] community govemments but can also include production teams, sometimes termed villagers' small groups [cunmin xiaozu]. Enterprises. The types discussed are: state-owned enterprises [guoying qiye], community enterprises [zhengban qiyel, and private enterprises [minban qiye]. Under community enterprises are township enterprises [xiangban qiye], village enterprises [cunban qiye], and production team enterprises [shengchan duiban qiyel. Township and village enterprises fxiangcun qiyel are referred to as TVCES. Private enterprises include all kinds of nongovern- mental rural enterprises: cooperative firms, partnerships [lianhu qiyel, and individual proprietorships [getihu, geti qiye, or siying qiyel, which may have 100 or more employees. Extrabudgetary revenues and expenditures [yusuan wai shouzhil. Community gov- ernment incomes and expenditures that occur outside the state budget. Includes profit remittances from community enterprises, management fees paid by TVPS, ad hoc levies on the rural population, and sometimes shared above-quota tax collections that are retained by the community government. Management fees [guanli feil. Levies on TVPS by TVP bureaus, industrial- commercial administrative mrlanagement bureaus, or sometimes community governments, usually calculated as a straight percentage of sales revenue, generally in the 0.5-2.0 percent range. Units of measure. 1 mu = 0.1647 acres = 0.0667 hectares. 1 jin = 0.5 kilo- grams. Counties Surveyed Jieshou County, Anhui ]rovince Nanhai County, Guangdong P'rovince Shangrao County, Jiangxi Province Wuxi County, Jiangsu Province Contributors The affiliations shown are those pertaining at the time of writing. William A. Byrd Country Operations Division, China Department, The World Bank Du Haiyan Institute of Economics, Chinese Academy of Social Sciences Du He Institute of Economics, Chinese Academy of Social Sciences Alan Gelb Financial Policy and Systems Division, Country Economics Department, The World Bank He Jiacheng Institute of Economics, Chinese Academy of Social Sciences Lin Qingsong Institute of Economics, Chinese Academy of Social Sciences Luo Xiaopeng Development Research Institute, State Council Rural Development Research Center Meng Xin Institute of Economics, Chinese Academy of Social Sciences Song Lina Institute of Economics, Chinese Academy of Social Sciences Jan Svejnar Department of Economics, University of Pittsburgh Wang Hansheng Department of Sociology, Beijing University Wang Tuoyu Institute of Economics, Chinese Academy of Social Sciences Wang Xiaolu China Economic System Reform Research Institute, System Reform Commission Josephine Woo Country Operations Division, China Department, The World Bank xiii xiv CONTRIBUTORS Wu Quhui Institute of Economics, Chinese Academy of Social Sciences Xu Xinxin Department of Sociology, Beijing University N. Zhu Department of Economics, Georgetown University China's Rural Industry 1 China's Rural Industry: An Introduction William A. Byrd and Lin Qingsong This chapter surveys the institutional framework, historical background, and present situation and importance of China's township, village, and private enterprise (TVP) sector. It is directed particularly at readers for whom the subject is new or whose knowledge about the sector is not up to date. We begin with some definitions. There are three levels of rural communi- ties in China. The township (formerly the commune), now the lowest level in China's government hierarchy, has an articulated government structure. The typical township has a population of 15,000-30,000. The village (for- merly the brigade) is not a separate level of govemment but has governmen- tal functions and a community structure. Villages generally have a population of 1,000-2,000. The production team ("villagers' small group") is purely a community structure, having lost most of its administrative functions as a result of the implementation of the production responsibility system (PRts) in agriculture in the early 1980s. The average team consists of about 30 households with a total of 150 people. The general term community govenm- ment is used to refer to authorities at all three levels. Communities in different parts of the country vary substantially in size and economic power. For example, in Nanhai County, Guangdong Province, the average township has more than 50,000 people and a gross value of industrial output (GVIO) of more than Y35 million. In Shangrao County, Jiangxi Prov- ince, by contrast, the average township has only 20,000 people and a GVIO of only Y393,000. There are some ambiguities in the delineation of the TVP sector, but the broadest definition, which is increasingly popular, encompasses all nonagri- cultural activities in rural areas and small towns other than those on state farms.' Hence the sector includes individual nonagricultural proprietorships and household firms as well as larger concems. Enterprises owned by community governments are referred to as community enterprises, and firms owned by individuals or partnerships are termed private enterprises. Community enterprises owned by townships and villages are re- 3 4 WILLIAM A. BYRD AND LIN QINGSONG ferred to as township and village community enterprises (TVCES); this category excludes production team firms, many of which have been privatized in one way or another. Occasionally township enterprises and village enterprises are discussed separately. Government organs at the township level responsible for supervising township enterprises are referred to as township industrial corpo- rations (TICS).2 The Institutional Framework Since the late 1950s China's rural economy has been organized under rural community structures that are linked to higher levels of govemment primar- ily through the Communist pairty hierarchy. Recent reforms have changed this pattem somewhat by restoring the township as the lowest level of gov- emment in rural areas and by allowing much greater scope for econornic ac- tivities and for the accumulation of assets by private individuals. Neverthe- less, community govemments retain their cohesiveness and have a significant role in the local economy. The most important characteristic of China's rural communities has been their stable membership. The household registration system has prevented large-scale rural-urban or rural-rural migration. Most people bom into a rural community stay there for their entire lives, and their children belong ro the same community. Immobility of population, of course, means immobility of human resources, although "ternporary" migration to cities and to other rural areas to seek jobs is increasingly common. Another key feature is the combination of income sharing within the com- munity with a considerable degree of financial self-reliance in relations with other communities and higher levels of government. Income sharin,g has taken various forms over time and in different areas. Before the reforms, the work-point system for production teams ensured relatively equal distribution of individual incomes within the team. Allocation of jobs in TVCES has been used to balance family incomes. Even coinmunity govemment decisions on the establishment and location of new firms have taken income-sharing con- siderations into account. Finally, the relatively equal distribution of agricul- tural land when the PRS was implemented contributed to maintaining a rela- tively equal distribution of inciDme within most rural communities. Income sharing has been weakened to some extent by the emergence and rapid growth of private economic activities and the abolition of the commune system. The degree to which commu[nities, particularly at the township level, are self-reliant has, if anything, increased since the late 1970s. Fiscal revenue- sharing systems have gradiually been extended down to the township level, particularly after the township became a separate level of administration and public finance in 1984-86. Most townships depend heavily on ex- trabudgetary revenues to meet: public expenditure needs. These revenues China's Rural Industry: An Introduction 5 come primarily from TVPS in the form of profit remittances and management fees. (Management fees do not represent payment for services provided; rather, they are extrabudgetary levies by community governments, usually at the rate of 1-2 percent of sales revenue.) TVPs also generate the lion's share of budgetary tax revenues in the more industrially developed rural communities. Even for poorer, more backward communities, which receive fiscal subsidies, incremental financial resources depend largely on the efforts of the community government and its subordi- nate economic entities. Reforms in the local banking system have made lend- ing ceilings in a township dependent to a large extent on the generation of deposits within the township.3 Similarly, because of the immobility of human resources, rural communities must by and large rely on the human capital of their own members, although some better-off areas have been able to at- tract technical and other skilled persons by offering high wages and other benefits. Community governments have complex and often conflicting roles and re- sponsibilities. They are essentially ministates that provide a whole range of public and social services and (particularly at the township level) set policies and regulations within their jurisdictions. The instruments available to com- munity govemments include the establishment and- ownership of TVCES, deci- sions on community govemment expenditures. a considerable degree of influ- ence over the lending decisions of the local banking system in many areas, and encouragement and support of different types ot private enterprise. Com- munity govemments cannot, however, use certain policy instruments nor- mally available to states. Perhaps most important, they cannot engage di- rectly in deficit financing; that is, their expenditures may not exceed available funds (which include transfers from higher levels of govemment and from Tvps as well as the portion of locally collected taxes allotted to them). Hence they face a relatively hard budget constraint. Since they can- not effectively impose barriers against trade in goods, they cannot meaning- fully protect their enterprises from outside competition. In any case, the mar- kets represented by rural communities, even at the township level, are too tiny to be of much use to local TVPS in most industries. In addition to being local govemments, community governments are profit-oriented economic entities that might be compared with holding com- panies, investment corporattons, or headquarters of loosely controlled mul- tidivisional corporations. Although day-to-day management and business decisions are usually left to enterprise directors, community governments tend to be intimately involved in decisions on important investments, the establishment or dismantling of firms, significant changes in product lines, the appointment ot enterprise management, managerial compensation (and sometimes wages as well), and bonuses. Community govemments can also absorb risk for subordinate community enterprises and can finance invest- ments, both directly and through arranging other sources of funding such as 6 WILLtAM A. BYRD AND LtN QINGSONG bank loans. More generally, they represent their community enterprises in an often hostile administrative environment. Subsequent chapters explore the serious conflicts inherent in this dual role of community governments. But strong fiscal and employment incentives, community govemments' role in business development, and the market envi- ronment for outputs and material inputs in which TVPS function have com- bined to generate an extremely powerful push for the development of TVPS. TVPS fall into several imain categories. Traditional TVCES are under the ad- ministrative control of township or village authorities. Firms at the produc- tion team level often were not independent accounting units; instead they put their proceeds into the pot for collective distribution to their members. Many production team enterprises were privatized during or in the aftermath of the switch to the PRS. Somae township and especially village enterprises have been contracted or leased to individual management. (The contracts typically specify amounts or shares of enterprise profits to be remitted to the community government.) Private enterprises include both individual propri- etorships and partnerships owned and managed by several people; many are based on family or kinship ties. Close personal relationships with community government officials are almost invariably necessary for sizable private enter- prises if they are to navigate in the complicated administrative environment. Only a few TVPs are owned cooperatively by all or most of their workers, but in small communities where the employees of the main enterprise make up a significant proportion of the population, community ownership has some of the characteristics of cooperative ownership. Finally, there is an increas- ing number of joint ventures between different forms of ownership, within the TVP sector as well as with urban state and collective enterprises. Joint ventures involving foreign partners are still rare, but especially in Guang- dong Province, many l'vPs have cooperative relationships with foreign firms. Recent Changes in Rural Institutions China's rural communities have been profoundly affected by the reforms that have taken place since the late 1970s. The most important of these was the PRS, which gave individual households partial rights to land that had previ- ously been vested in the production teams. The PRs also greatly reduced the direct supervision of agricultural production by community govemments and effectively abolished the system of community work-points and collective dis- tribution of income in favor of a return to primarily household-based genera- tion of personal income. A closely related reform was the abolition of the communes and their replacement by townships, now the lowest level of gov- ernment in rural areas. An exirremely important policy change for the TVP sector was the relaxation of restrictions on the large-scale expansion of non- agricultural activities by rural communities. Although these restrictions were China's Rural Industry: An Introduction 7 unevenly implemented and were never entirely effective in some parts of the country, their removal in 1978 created a political and administrative climate conducive to the rapid development of TVPS. Similarly, the removal of re- strictions against household nonagricultural activities and, more recently, against sizable private firms in rural areas stimulated a great spurt of TVP de- velopment. Finally, the PRS brought about a crucial change in TVP wage systems. In the past many TVCES paid salaries directly to the production teams of which workers were members-workers received work-points for their labor in TVPS and then participated in collective income distribution by the teams at year's end. This greatly diluted incentives, since the fruits of hard work were shared not only with other workers in the enterprise but also with all members of the production team. Even in TVPS that paid workers directly, fixed time wages were the norm. All this changed with the PRS. TVPS shifted to direct payment of workers, and performance-based pay-in the form of enterprise- based work-points, year-end dividends, profit-related bonuses, and, above all, piece-rate wages-became the norm. The impact on incentives and effi- ciency in the TVP sector should not be underrated. Regional Variations in Institutional Structure Despite its basically uniform organizational framework, China's TVP sector shows a remarkable degree of local and regional institutional diversity. This diversity is partly attributable to differences in exogenous factors-the tradi- tional resource base (especially human capital and agricultural land), proxim- ity to markets and to urban industry, pre-1949 industrial and commercial tra- ditions, delivery requirements for grain and other crops under the agricultural planning system, and prereform rural industrialization. But local authorities and community leaders have sometimes had a great influence on rural indus- trialization and local institutional structure. In the early 1970s Wuxi County authorities supported the continued development of commune and brigade industries counter to the prevailing national trend. Jiangsu Province (where Wuxi is located) several times successfully ignored national campaigns to shut down and "readjust" commune and brigade enterprises and to concentrate on agriculture. This resistance helped give it a commanding lead in rural in- dustrialization on the eve of reforms in the late 1970s. Nanhai County au- thorities allowed and protected "underground" private firms even in the prereform period. But in many counties local authorities did not effectively promote development of TVPS until very late, which handicapped the rural industrialization effort. Areas in which diversity is evident include the ownership of TVPS, the de- gree to which TVPs are under hierarchical control by townships, the policies of local and community authorities regarding TVPS, and the level of develop- ment and industrialization. 8 WILLIAM A. BYRD AND LIN QINGSONG OWNERSHIP. In the four counties in which intensive fieldwork was conducted-Wuxi, Jieshou, Nanhai, and Shangrao-striking differences were encountered (see chapters 2, 4, and 7). The 1985 share of private enter- prises in the GVIO of TVPs ranged from 3 percent in Wuxi to 51 percent in Jieshou, as against a national average of 17 percent. The relative weight of township and village enterprises also varied considerably; the share of town- ship enterprises was nearly three times that of village enterprises in Jieshou and close to twice as high in Shangrao, but in Wuxi the shares were nearly equal. Only in Nanhai did production team firms account for a substantial part of GVIO in 1985. CONTROL BY TOWNSHIPS. Villages and teams in Nanhai have been rela- tively independent and in particular have had freedom to develop their own firms. Wuxi has a highly integrated hierarchical structure of rural conmmuni- ties, and the township rules with a strong hand. Hence, even though village enterprises there are quantitatively important, they are effectively parl: of the township industrial sector. The degree of administrative control over wages, the prevalence of profit- related bonuses, the methods of appointment of TVP managers, the allocation of labor, and a host of othaer practices can differ greatly among regions, locali- ties, and even communities wiithin a locality. These differences can have an important effect on the behavior and performance of TvPs even though in most of them wages and managerial pay are significantly related to enterprise performance. LOCAL POLICIES. Counties in southern Guangdong Province variously promote large township enterprises, village and production team firms, pri- vate firms, and cooperative ventures with outside and foreign entrepreneurs. Wuxi's administrative climate makes it virtually impossible for large indus- trial private enterprises to emerge. Jieshou County authorities personally sponsor some of the more successful private entrepreneurs, supporting them with bank loans and administrative help. LEVEL OF DEVELOPMENT. The richest and most industrialized counties, such as Wuxi and Nanhai, have already largely achieved the equivalent of middle-income status. The most backward areas-an order of magnitude poorer and often hundreds or thousands of times less industrialized-are no- where near the takeoff sl:age. The industrialized counties face issues of how to organize, finance, and develop the TVP sector most efficiently; the less de- veloped counties are striving merely to get nonagricultural development started. These differences are discussed in more detail in chapter 12. The diversity of the TVP sector in different parts of China means that re- forms and policy measures that: work well in some areas may be unsuccessful China's Rural Industry: An Introduction 9 in others. For example, many poor areas have tumed to industrialization based on private enterprise because they recognize that for a variety of reasons the traditional community enterprise system (epitomized by Wuxi) cannot succeed in backward areas. From another perspective, Wuxi's restrictions against private enterprises may not carry terribly high economic costs there because the local TVCE sector is so efficient, but such policies would be disas- trous in areas where TVCES perform poorly. Historical Background Before 1949 China was a poor and backward agricultural country, and most of its rural areas had only a few small workshops and some handicraft indus- tries. In 1949 the output value of such household sideline production totaled about Y1. 16 billion in fixed 1957 prices. After the founding of the People's Republic of China, sideline production developed rapidly. By 1954 more than 10 million Chinese farmers had taken part-time jobs in the commercial- ized handicraft industry, which yielded a total output value of Y2.2 billion that year. During 1955-57 China launched a large-scale agricultural cooperative movement. Farmers who were working part-time in handicraft industry, as well as individual handicraft workers, were organized into specialized sideline production teams under agricultural producers' cooperatives. The excessive emphasis that some cooperatives placed on agricultural production caused their sideline production to stagnate. Remedial measures by the authorities later helped to restore and develop sideline production in cooperatives. In 1957 the total output value of sideline production in China's rural areas reached Y2.29 billion, about 4.3 percent of the total value of agricultural output. Commune and brigade enterprises, the predecessors of the TVPS, sub- sequently began to emerge and develop on the basis of collective sideline pro- duction. In 1958 the people's commune movement swept across the vast rural areas of China. Dizzy with the "victory of transformation" (nationalization and so- cialization of the economy), China adopted the Great Leap Forward strategy and attempted to reach the "paradise" of communism by building the "golden bridge" of people's communes. Under the slogan of "mobilizing the whole nation to make great efforts to develop industry," rural people's communes established a large number of commune and brigade enterprises by siphoning manpower, funds, and materials from brigades and teams originally engaged in sideline production and by annexing more than 30,000 handicraft cooper- atives that had previously operated in small towns. By the end of 1958 com- mune and brigade enterprises in China employed 18 million people and yielded a total output value of Y6 billion. In the following year that figure catapulted to Y10 billion. 10 WILLIAM A. BYRD AN4D LIN QINGSONG The communization campaign not only dampened farmers' enthusiasm for production but also precipitated a crisis for the national economy. In 1960 the Central Committee of the Communist party and the central govemnment called for readjustment of the national economy and began to check the com- munization campaign. As a result, many commune and brigade enterprises discontinued operations. Some were later returned to production teams, some were changed into handicraft cooperatives, and the rest were closed down. In 1961 the total output value of commune and brigade enterprises nosedived to Y1.98 billion, and the figure dropped further to Y790 million in 1962 and Y410 million in 1963. During the following six years these enter- prises saw a small rebound in their output value, but they remained basically stagnant. Development of comnmune and brigade enterprises resumed during the lat- ter part of the Cultural Revolution period (1966-76). At the North China Agricultural Conference (August-October 1970) the State Council called for efforts to speed up agricultural mechanization. Acting in the spirit of th e con- ference, rural areas began to launch factories to produce agricultural machin- ery and farm tools. Meartwhile, many urban plants, absorbed in the Cultural Revolution, had stopped work. Market shortages became worse with each passing day, and some rural communities took advantage of this opportunity to set up firms. Although such enterprises were then widely regarded as the "tail of capitalism" and faced all kinds of restrictions and attacks, they still witnessed remarkable developrnent. Their output value climbed from Y9.25 billion in 1970 to Y27.2 billion in 1976, an average annual growth rate of 25.7 percent. After 1976, when the Gang of Four was overthrown, commune and brigade enterprises developed further. In 1978 there were more than 1.52 million such enterprises, and their output value reached Y49.3 billion. Gov- ernments at all levels set up aodministrative bureaus to strengthen manage- ment of these enterprises. In 1978 the Third Plenum oif the Eleventh Central Committee of the Chi- nese Communist party declared that "commune and brigade enterprises should strive for great development" and that As long as it is in conformity with the principle of rational economic development, commune and brigade enterprises should gradually engage in the processing of all farm and sideline products that are suitable for rural processing. Urban factories should shift part of their processing of products and parts and components that are suitable for rural processing to com- mune and brigade enterprises and help equip the latter with necessary equipment and technology . . . In addition, the state should adopt a pol- icy of allowing tax breaks or tax exemptions for commune and brigade enterprises in the light of their situation.4 In 1981, to reduce discrimination against commune and brigade enter, prises, the State Council issued regulations which reaffirmed that "commune China's Rural Industry: An Introduction 11 and brigade enterprises already constitute an important part of the rural econ- omy, and their growth is in conformity with the comprehensive development of the rural economy."5 In 1983 the Central Committee further confirmed that "commune and brigade enterprises also belong to the cooperative econ- omy, and great efforts must be made to continue to consolidate and develop such enterprises."6 A series of such policies and regulations brought commune and brigade enterprises into a new stage of healthy, stable, and sustained high-speed growth. During 1980-83 their total output value jumped from Y65.7 billion to Y101.7 billion; the annual growth rate was 13.5-19.2 percent. The Central Committee Circular on Agricultural Work of January 1, 1984, stated that, in addition to supporting actively the development of commune and brigade enterprises, govemments at all levels should "encourage peasants to invest in or buy shares of all types of enterprises and encourage collectives and peasants to pool their funds and jointly set up various kinds of enterprises by following the principle of voluntary participation and mutual benefit." In a March 1984 circular the Central Committee and the State Council agreed to rename commune and brigade enterprises TVPS, since the people's com- munes had been abolished and a large number of partnerships and individual enterprises had come into being.7 The circular also declared, Vigorous efforts should be made to guide TVPS in terms of orientation of development and to exercise administration over them in accordance with the relevant state policies to ensure their healthy development. TVPS should receive the same treatment as state enterprises and are entitled to all necessary state aid. Since then China's TvPs have flourished more than ever. In 1984 the out- put value of TVPS (including TVCES, partnerships, and individual enterprises) totaled Y171 billion, and the figure rose to Y273 billion in 1985 and Y354 billion in 1986. Private enterprises grew faster than TVCES, service activities grew more rapidly than industry, and TVPS in economically backward areas made quicker progress than those in economically advanced areas (see chap- ter 12). Present Situation and Importance The growth of the TVP sector since the late 1970s and particularly since 1983 has been spectacular, whether measured by output, employment, assets, or profits (table 1-1). Nominal gross output value increased at an average annual rate of 24 percent in 1978-86, and GVIO grew by more than 23 percent a year in real terms. (Reliable price deflators for TVP services are not available.) Employment outside agriculture in rural areas grew by nearly 13 percent a year in 1978-86, for a total increase of 124 percent, or more than 30 million people.8 Data on TVCES (see table 1-1) indicate that their wage bills, assets, Table 1-1. Growth of the TVP Sector, 1980-86 (average annual percentage change) Item 1978-80 1980-83 1983-86 1984 1985 1986 1978-86 Value of ouitput Gross (nominal), 16.5 16.5 37.1 34.9 46.8 30.2 23.8 Industrial (nominal) n.a. 15.0 42.3 40.5 50.7 36.0 27.9' Industrial (real)' 16.8 14.9 37.6 38.5 43.0 31.6 23.4 Employment Nonagricultural 7.3 5.4 24.2 38.5 20.1 15.1 12.7 Rural industrial 5.8 1.3 15.9 20.0 13.2 14.5 7.7 Number of firms All TVCEs -3.3 -1 9 4.1 22.5 -4.9 -3.3 -0.1 Industrial TVCEs -2.3 -0.6 5.7 21.1 -5.1 2.7 1.3 Financial indicators (TVCEs only) r..> Assets8 19.2 13.4 25.8 20.9 30.5 26.2 19.4 Profits, 15.0 2.7 13.5 13.4 31.6 -2.1 9.7 Taxes' 7.9 31.3 32.7 34.4 37.2 26.9 25.7 Bank loans n.a. 20.4 61.1 102.6 40.3 47.0 39.2b Gross revenues 19.4 15.9 33.8 36.6 44.1 21.7 23.2 Employment 3.0 2.5 10.7 18.9 7.9 5.8 5.7 Wage bill 17.4 13.8 26.5 36.1 26.0 17.9 19.3 Average wage 14.0 10.9 14.2 14.5 16.7 11.4 12.9 Average wage, state enterprises 11.7 2.5 17.8 19.5 17.3 16.6 10.3 n.a. Not available. a. Gross output value of rural nonagricultural materials-producing sectors (industry, construction, transport, and commerce). b. 1980-86 average, c. Gross output value in current prices was deflated by the implicit price deflator for Chinese industry as a whole (excluding village and subvi!lage midustrial firmns). This was done because sector-specific price deflators are unreliable and result in even higher "real" growth rates. d. Value of fixed assets net of depreciation plus physical working capital. e. Gross profits before payment of income tax but net of indirect taxes. f. Includes both direct and indirect taxes. Sources: State Statistical Bureau (1982, 1985a, 1986a, 1986b, 1987a, 1987c, 1987d). China's Rural Industry: An Introduction 13 profits, and tax payments all increased rapidly in line with the growth of out- put and employment. Profits, however, rose much more slowly than other indicators, suggesting that profitability declined in the TVCE sector. Average wages in TVCES rose 13 percent a year in 1978-86, considerably faster than the rate of inflation. The stagnant number of TVCE firms probably conceals significant activity in establishments and exits of firms. Moreover, a tremen- dous number of new enterprises has been created in the private sector in re- cent years; the number of private enterprises reportedly reached 10.7 million at the end of 1985. Growth in China's TVP sector has been highly uneven over time, as can be seen from table 1-1. There was a sharp acceleration in 1983-86 compared with 1980-83; real GVIO rose by 15 percent a year in 1980-83 and by 38 percent a year in 1983-86. Growth in 1984 and especially in 1985 was sub- stantially higher than in 1986, and in 1987 the TVP sector again boomed. TVP growth performance has also been unbalanced geographically, with some areas, particularly coastal regions, experiencing more rapid growth than backward, less industrialized interior provinces. Between 1980 and 1985 the nominal value of TVP industrial output rose by 415 percent in Zhejiang Prov- ince but by only 37 percent in Nei Monggol (Inner Mongolia). In eight prov- inces the increase was greater than 250 percent; in seven it was below 120 percent. Many, although not all, of the fast growers were more industrialized from the beginning, and a number of provinces that had lagged in rural indus- trialization slid further into relative backwardness. Moreover, provincial growth rates mask great differences within provinces. Table 1-2 provides a more detailed picture of the growth of industrial out- put for TVPS under different forms of ownership. The GVIO of TVCES has in- creased at an average annual rate of more than 20 percent in real terms; it grew especially rapidly in 1984-85. The GVIO of private enterprises (includ- ing production team firms) has grown spectacularly from a tiny base in 1980. This may in part reflect the incorporation of economic activities that had previously escaped the statistical network, but it mostly represents real growth. Differential growth rates have led to substantial changes in the com- position of TVP industrial output by form of ownership (table 1-2). The shares of township and village enterprises have declined moderately, whereas the share of private enterprises nearly quadrupled between 1980 and 1986. As a result of the rapid growth documented above, which implies a dou- bling in size every three years or less, the overall dimensions of the TVP sector have been changing substantially. For example, measured by real GVIO, Chi- na's industrial TVP sector was roughly four times as large in 1986 as in 1980. Thus, any attempt to describe its size and characteristics can be only a snap- shot of a rapidly changing situation. Table 1-3 shows the dramatic changes in the composition of gross social product in China's rural areas. The share of crop cultivation dropped from almost half in 1980 to less than one-third in 1986; the share of other agricul- 14 WILLIAM A. BYRD AND LIN QINGSONG Table 1-2. Growth and Ownership Structure of TVP Industry, 1980-86 Item 1980-83 1983-86 1984 1985 1986 1980-86 Average annual growth of GVIO (percent) TVCEs Nominal 13.8 35.3 38.2 38.8 29.1 24.1 Real 11.4 33.5 38.0 34.3 28.4 21.9 Private enterprises' Nominal 32.8 93.9 65.4 159.3 69.9 60.5 Real (estimated) 32.7 87.7 63.1 146.1 64.7 57.8 All TVPs Nominal 15.0 42.3 40.5 50.7 36.0 27.9 Real 12.8 40.2 40.2 45.8 34.7 25.7 1980 1983 1984 1985 1986 Nominal GVIO, by ownership (percentage of total) Township enterprises 52.8 54.5 50.5 45.1 43.5 Village enterprises 41.8 37.1 39.6 37.9 35.2 Production team enterprises -b _b Partnerships 5.4 8.4 4.6 8.8 21.3 Individual enterprises 1J 5.3 8.2 J Note: Gvio, gross value of industrial output. Figures are available for nominal total rural GVIo, village enterprises' GVIO, and village-and-below GVIO. From these the GVIOs of township enter- prises and private enterprises (including production team enterprises) can be calculated. Figures for GVIO of township enterprises in 1980 constant prices are also available. Real values for other components of rural GVIO were calculated by using the implicit price deflator for China's total GVIO, excluding village-and-below GVIO, for which no deflators are available. a. Includes production teamn enterprises. b. Many production team enterprises were privatized in the early 1980s. Any remaining ones were probably included in the category of partnerships (cooperative firms). c. Derived from separate statistics compiled by the Ministry of Agriculture, which show higher absolute levels of industrial output value of private enterprises than do the figures of the State Statistical Bureau. Sources: State Statistical Bureau (1.982, 1984, 1985a, 1986a, 1986b, 1987a, 1987c, 1987d); China Economic Yearbook Editorial Board (1981, 1985). tural activities remained more or less constant; and industry rose from less than one-fifth to more than 30 percent, nearly equal to the share of crop cultivation. In 1987 rural GVIO exceeded the value of agricultural output for the first time. Construction, transport, and commerce rose modestly as a share of the total. Shifts in the structure of the rural labor force have been less dramatic but nevertheless significant, as is shown in table 1-4. Cultivation declined sharply in 1984 and 1985, but much of the slack was taken up by agricultural sidelines, which more than doubled their share. The proportion of the rural labor force engaged in industrial activities rose only modestly, from 6 percent China's Rural Industry: An Introduction 15 Table 1-3. Composition of Rural Gross Social Product, 1980-86 Item 1980 1983 1984 1985 1986 Total rural gross social product (billions of yuan)a 279.2 412.4 506.8 634.0 755.4 Share of total (percent) Agriculture (cultivation) 49.3 47.1 43.3 36.0 33.1 Forestry, livestock, fishery, and sidelines 19.5 19.6 20.1 21.1 20.0 Industry 19.5 20.0 22.9 27.6 31.5 Construction, transport, and commerce 11.7 13.3 13.7 15.3 15.4 a. In current prices. Includes gross output value of the five materials-producing sectors (agri- culture, industry, construction, transport, and commerce), excluding nonmaterial services. Sources: State Statistical Bureau (1986b, 1987a, 1987c). in 1980 to 8 percent in 1986. Construction, transport, and commerce more than tripled their share, however, rising from less than 2 percent in 1980 to more than 6 percent in 1986. Overall, the movement of labor out of agri- culture in rural areas made a strong beginning in 1984-86, but four-fifths of the rural labor force is still engaged in agriculture and sideline activities. The TVP sector has become increasingly significant in generating fiscal rev- enues in rural areas and providing a growing proportion of the personal in- comes of rural inhabitants. Total tax payments by TVPS in 1985 (including profit tax and indirect taxes) were nearly Y14 billion. These taxes accounted for only 7.7 percent of total national budgetary revenue but for most tax col- lected in rural areas. (The stagnant agriculture tax generated only Y4 billion in 1985.) Tax payments by TVPS have been rising sharply as earlier exemp- tions are gradually phased out, although special treatment remains common and private enterprises, in particular, can evade much of their tax liability. The TVP sector remains more lightly taxed than the state sector (the differ- ence is considerably reduced if extrabudgetary remittances of TVps are taken into account), but its tax burden is much heavier than that of agriculture. Extrabudgetary remittances by TVPS to community governments are impor- tant, especially in the more developed areas. In 1985 TVCES turned over Y6.8 billion in profits to their supervisory community governments, not including management fees, which may have totaled Y2 billion or more. In compari- son, total tax payments by TVCES were Y10.9 billion. More than half of TVCE remittances to community governments were plowed back into reinvestment, but substantial amounts were used for support to agricultural production, for social services, and for community government payrolls, and a considerable amount (Y620 million) was distributed to community members. In 1985 the total of wages and distributions to community members by TVCES was Y31 billion. This represented Y40-Y50 per member of the rural population, or about 15 percent of average per capita rural income. Although 16 WILLIAM A. BYRD AND LIN QINGSONG Table 1-4. The Structure of the Rural Labor Force, 1980-86 Item 1980 1983 1984 1985 1986 Total rural labor force (millions of people) 318.4 346.9 359.7 370.7 379.9 Share of total (percent) Agriculture (cultivation) 82.8 81.5 70.9 67.3 Forestry, livestock, fishery, 80.2 and sidelines' 6.4 6.4 13.3 14.6 Industry 6.1 5.8 6.7 7.4 8.3 Construction, transport, and commerce' 1.6 2.5 4.4 5.5 6.1 Other' 3.1 3.8 4.7 5.2 5.4 a. At year-end. Does not include labor force of state farms. b. The industrial portion of the sidelines labor force has been included under industry. c. Commerce includes storage, catering, and other related personal services. d. Includes real estate admiinistration, public utilities, residential services, consulting, public health, education, sports, social services, scientific research, technical services, banking and insurance, government, Cormmunist party, and the like. Sources: State Statistical Bureau (1986b, 1987a, 1987d). private enterprises would add significantly to the figures, the role of the TVP sector in the generation of rural personal income is less important than its role in production. This is partly because the ratios of net to gross output value are much smaller in industry and, especially, in commerce than in agri- culture. Furthermore, m,ost TVCE profits go into investment and thus affect personal incomes only indirectly. There is great potential, however, for the TVP sector to become more important in the direct generation of income as well as employment. The TVP sector has become the most dynamic part of Chinese industry. It is an important actor in most industries and is dominant in a few. Its share in total national GVIO has risen sharply, from only 3 percent in 1971 (when the movement to promote commune and brigade enterprises was getting under way) to 9 percent on the eve of the reforms in 1978, to 12 percent in 1983, and then, rising very rapidly, to more than 21 percent in 1986. The TVP sector has roughly doubled its share of China's industrial output every five years, starting from a tiny base in the early 1970s but continuing to do so even as the base grew large in the 1980s. The share of the state sector declined from more than four-fifths in 1975 to three-quarters in 1980 to only about five-eighthis in 1986, while that of urban nonstate firms rose moderately. All segments of the TVP sector have seen their shares in Chinese industry as a whole increase, but private enterprises have shown the most spectacular growth in recent years; they rose from negligible levels in the late 1970s to account for nearly 5 percent of China's GVIO in 1986. China's Rural Industry: An Introduction 17 TVPS are active in all but one of the fourteen principal subsectors of Chi- nese industry; only in petroleum is their share truly negligible. Otherwise the subsectoral coverage is surprisingly broad, and the pattern holds even in a more disaggregated industrial breakdown. The most important TVCE indus- trial subsectors are machinery, construction materials, textiles, chemicals, and food processing (in that order); together they account for nearly three- quarters of TVCE GVIO. Only for construction materials is the share far higher than for Chinese industry as a whole; shares of other subsectors tend to be no higher or moderately lower than for industry as a whole. TVP shares in the output of particular industrial products vary greatly, but again the coverage is broad. In 1986 TVPS produced 75 percent of all bricks made in China, 22 percent of the cement, 26 percent of the coal, 26 percent of the silk, 47 percent of the silk textiles, 27 percent of the edible vegetable oil, and 27 percent of the machine-made paper and cardboard. In addition, TVps accounted for about 30 percent of the national output of phosphate fer- tilizer and for high proportions of other industrial inputs into agriculture, such as farm tools and implements. Their share of production of metal- cutting machine tools is only about 4 percent, but they are increasingly im- portant in electronics (where they accounted for 6 percent of total national output value in 1986), in many consumer durables, and even in some more sophisticated industries. Only certain basic process industries with substantial economies of scale (for example, petroleum refining) are impervious to entry by TVPS. In sum, China's Tvps have already become a pillar of the country's econ- omy and are playing a decisive role in rural industrialization and national eco- nomic modernization. The TVP sector represents the only hope for moving rural labor into productive employment outside agriculture, particulaTly as long as restrictions against permanent migration into large cities are kept in place. Without TvP development, much of the gain in efficiency from agri- cultural reforms would not have been realized. At the same time, the TVP sector is increasingly important in Chinese industry as a whole, both in meet- ing rising demand for industrial goods and as a competitor and partner of state industry. Reforms and experiments with varied forms of ownership and management in TvPs have already had a great demonstration effect on urban economic reforms and will continue to do so. From a longer-term point of view, without the transfer of large numbers of farmers to TVPS, China's rural industrialization and national economic modemization will be impossible. The development of TVps has brought mil- lions of farmers into the competitive commodity economy, and this new en- vironment will instill new values and moral concepts. All these factors will have a far-reaching impact on China's modernization. But the development of TVPS will also inevitably bring a series of new problems for overall strategy and for the management of national economic development. Obviously, 18 WILLIAM A. BYRD AND LIN QINGSONG thorough and systematic study of problems in the development and opera- tions of TVPS and their influence on the economy as a whole is needed. This book represents a significant effort toward this end, on the basis of extensive new research. Notes 1. A few TVPS are involved in agricultural production. The definition excludes agricultural production by individual rural households. 2. TICS go under a variety of names, including township industrial corporations, township industry and trade corporations, and township industrial associations. Their functions often overlap those of township economic commissions, which have as one of their main tasks supervi- sion of local industrial development. In some areas the economic commission is directly responsi- ble for township enterprises, and there is no separate TIC. 3. Although there are large (positive or negative) imbalances between the two in many locali- ties, incremental loans to a large extent depend on incremental generation of deposits. 4. "Resolution on Several Questions about Speeding up Agricultural Development," (draft), December 1978. 5. "Several Regulations on Commune and Brigade Enterprises in Implementing the Principle of National Economic Readjustment." 6. "Several Questions on Current Rural Economic Policies." 7. Circular transmitting the "Report on Exploring New Areas for Development of Commune and Brigade Enterprises," submitted by the Ministry of Agriculture and the Party group of the ministry. 8. These figures refer to the nonagricultural, materials-producing sectors in rural areas (indus- try, construction, transport, and commerce) and exclude nonmaterial services. They may con- siderably underestimate TVP ernployment because they appear to leave out large numbers of part- time and seasonal TVP employees and some workers in small private enterprises. Statistics compiled by the Ministry of Agriculture indicate that the total number of TvP employees rose from 22 million in 1978 to 77 million in 1980, an increase of 55 million. 2 Research Design, Methodology, and Data William A. Byrd and Lin Qingsong The recent rapid growth, present significance, and great potential of China's TVP sector as outlined in chapter 1 make it an important subject for research. Accordingly, in 1986-87 the Institute of Economics of the Chinese Academy of Social Sciences (CASS) and the World Bank engaged in a major collabora- tive research project on TVPS. This chapter briefly reviews previous research on China's TVP sector and provides background on how the collaborative re- search project was initiated and carried out. It then looks at the project's objectives and design, including the approach to data gathering, sample se- lection, and methodologies. It concludes with a description and analysis of the data set. Background China's TVPS (formerly termed commune and brigade enterprises) were the subject of considerable research by foreign scholars in the 1970s. Most of this work, however, dealt with the broader subject of rural small-scale industry, including county-run collective and state enterprises. The main full-length works that emerged from this period were Perkins and others (1977), Sigurdson (1977), and Wong (1979). Of other contributions to the literature on Chinese rural industrialization, Riskin (1978a, 1978b) might be singled out. Research on rural small-scale industry by foreign economists was hin- dered in the 1970s by lack of access to data, but some scholars were able to conduct careful, detailed fieldwork. Foreign scholarly interest in the TVP sector continued into the 1980s, al- though paradoxically the volume of published work has been smaller. Much of the research has appeared in compendiums and journal articles or as part of broader works on rural development in China, including Enos (1984), Griffin and Griffin (1984), and Ho (1986). The World Bank has also done some work on China's TVP sector, most notably as part of an economic study of Gansu Province (World Bank 1988). The astonishing growth of the TVP 19 20 WILLIAM A. BYRD AND LIN QINGSONG sector in 1984-85 attracted a great deal of foreign interest, this time more explicitly focused on TVPS, but it has not yet yielded much published output. Chinese economists have been interested in the TVP sector and in small- scale rural industry for an even longer time. How to view these enterprises, and what were the appropriatre policies toward them, were important topics of discussion as early as the Great Leap Forward in 1958. The second wave of TVP development in the early 1970s also engendered considerable research and discussion. But most ChLinese research on TVPs before the early 1980s centered on such political and ideological issues as whether rapid develop- ment of TVPS was compatible with socialism and whether it would harm the administrative division of labor between agriculture and industry. Starting in the early ]i980s Chinese research on the TVP sector turned from the question of whether the sector should be allowed to develop to how and in what ways it should develop. This shift focused work on more concrete economic questions. Some Chinese research on the TVP sector in the 1980s has been careful, detailed, and high in quality; an example is the work of the Development Research Institute of the State Council Rural Development Research Center.' Zhou Qiren and associates have engaged in detailed re- search on a number of aspects of TVP development (for example, Zhoua Qiren and Huang Zhuangjun 1987). Another recent publication, based in part on preliminary testing and research done for this project, is Institute of Econom- ics, CASS (1987). Considerablle research has also been conducted on private enterprises in rural areas. Finally, there has been an extraordinary increase in the available statistical and other source materials on the TVP sector. The Township and Village En- terprise Bureau of the Ministry of Agriculture publishes useful information, as do the State Statistical Bureau (1986b, 1987c, and 1987d) and the Minis- try of Agriculture (1986a and 1986b). Despite the considerable volume of past and ongoing research on the TVP sector, until recently little if any comprehensive work had concentrated spe- cifically on TVPS, as distinguished from rural small-scale industry in general, county-run state enterprises, aind so on. Moreover, most of the work, particu- larly by foreign scholars, was not based on systematic data gathering. Most important, the rapid development of the TVP sector since 1984 has simply outrun past and even much current research. Thus, both the Chinese and the World Bank considered that there was room for a wide-ranging, compre- hensive, policy-orientect collaborative project on TVPS. In September 1985 discussions began on the possibility of a collaborative project on the TvP sector. Detailed work on project and questionnaire design was done in late 1985 and early 1986. A research team was assembled that included representatives of such Chinese institutions as the Development Re- search Institute of the State Council Rural Development Research Center, the Economic System Reform Research Institute, and the Sociology Depart- ment of Beijing University. In early 1986 two counties, Wuxi and Jieshou, Research Design, Methodology, and Data 21 were selected as sites for in-depth fieldwork and enterprise samples. Chinese researchers visited each county to discuss logistical and administrative ar- rangements and to select sample firms. In March 1987 Chinese researchers conducted a pilot investigation, including a test of the draft questionnaires, in Yuanping County, Shanxi Province. The questionnaires were subse- quently revised. In June and July 1986 Chinese and World Bank researchers conducted in- depth fieldwork in Wuxi and Jieshou counties, and some members of the re- search team also spent several days doing fieldwork in Taihe County. The second main joint fieldwork effort took place in February and March 1987 in Nanhai and Shangrao counties. In October and November 1986 some of the Chinese researchers visited cooperative, labor-managed, and other small- scale firms in India, Yugoslavia, Great Britain, the United States, and Hong Kong and held discussions with scholars during the visits. This trip and sup- porting work were financed by a grant from the Ford Foundation. In the course of the trip discussions on data processing and analysis for the fieldwork in China were held at the World Bank in Washington, D.C. The work of processing and analyzing data and preparing research papers began in earnest in April 1987. The International Conference on China's Township and Village Industry, at which the papers in this volume were pre- sented, was held in Beijing, November 4-7, 1987. It was attended by the research team, other representatives of CASs and the World Bank, invited Chinese participants from government agencies and research institutions, and a panel of prominent foreign experts familiar with topics relevant to Chi- na's TVP sector. Project Objectives and Design The first main objective of the research project was to gain a better, more complete, and more solidly grounded understanding of how China's dynamic TvP sector actually functions. Its institutional and organizational structure needed to be thoroughly explored and analyzed. The second objective was to ascertain and assess the reasons behind the rapid growth of the TVP sector, particularly since the early 1980s. The third objective was to identify and analyze the problems and obstacles faced by the TVP sector as it continues to develop, especially institutional features of the TVP sector and problems in the administrative and regulatory environment that TVPS face. The fourth objective was to evaluate future institutional and policy op- tions. Although the research was not intended to generate concrete policy recommendations, it was hoped that the project would improve understand- ing of policy issues in the TVP sector so that future policy choices can be firmly grounded. The fifth objective was to gain an intemational perspective on China's TVP 22 WILLIAM A. BYRD AND LIN QINGSONG sector through review of relevant literature as well as through visits to foreign enterprises by Chinese researchers. To the extent appropriate, the research was to incorporate methodologies used in similar studies in other parts of the world. In line with its policy orientation, it was hoped that the project could be completed relatively quickly and that at least preliminary findings would be forthcoming in a short time. These objectives, as well as logistical and budgetary considerations, influ- enced the design of the research work. The project required considerable in- depth fieldwork that had to be accomplished in a relatively short tirme and with limited resources. C(areful, methodical gathering of both quantitative and qualitative data that coulcl be used in formal statistical analysis was re- quired. With these considerations in mind, it was decided to limit the in- depth fieldwork to four counties, each representative of a different pattem of TVP development (see chapters 4 and 7). The four counties were paired to allow relatively convenient visits on two separate trips, with the joint re- search team spending ten to fourteen days in each. Logistical considerations hence dictated a focus on China proper; the northeast, northwest, and south- west were not included, and even the northern part of China proper received relatively less coverage (see map 2-I).2 The limited number of localities selected made it possible to conduct thor- ough, comprehensive fieldwork in each, on a relatively large sample of enter- prises (an average of thirty in each county), with multiple questionnaires of substantial length. Extensive interviews were held with enterprises ancl with related local institutions. To ensure maximum reliability of the inforrrmation gathered, the bulk of the questionnaires had to be administered in person, and responses had to be double-checked in the field to rectify errors and omis- sions. A relatively large team olf researchers (more than twenty) was required to accomplish this work in a short time. Tradeoffs among the various objectives and constraints affected the sample of enterprises chosen. The bulk of sample firms had to be substantial corncems that kept reasonably good accounts and had at least several years of history to allow a time series. But since a crucial objective was to study the full range of ownership and management patterns in the TVP sector, oversampling of nontraditional forms of ownership and management was considered highly desirable. A difficulty here was that most private enterprises were founded only recently (1984 or later), are relatively small, and tend to keep poor ac- counts. The sampling of different forms of ownership was further complicated by the great differences in ownership patterns among counties. The samnples in two of the four counties contained virtually no private enterprises. The samples in the other two counties included a significant number of private enterprises, but at some sacrifice of the longevity of the firms: 31 percent of sample firms and 53 percent of private enterprises in the sample were re- portedly established in 1984-86. The sample of TVPS was limited to industrial firms, mainly in manufactur- Research Design, Methodology, and Data 23 Map 2.1 The Location of the Sample Counties 1100 / 1200 LIAONING .- D.P.R. OF NEI M NGOL iI JNSH KRA ji HEBEUI * o SHANIUAJINX Gz ANGXI GA OSHANDONGG -3 4 ~ ANA Gogzo *NU Township Pingzhou HEA JINS gXZEIAG 7..)¢; X E PoiSHnOc Copitol ANHLUIAN f~~~~~~~ fA Wunzhuuxi. .zz--s PingzhoWuho Hangzho wshu Nonc + fi /i!3 Notionol Copitols < <--;w-______________'___ -. - Intemolionol Boundorien C'X'CXLa~~~~~~~~S R 24 WILLIAM A. BYRD AND LIN QINGSONG ing. This was done to keep the scope manageable and to maximize compara- bility with previous collaborative research on state-owned industrial enter- prises. Moreover, it was felt that manufacturing is the cutting edge of the TVP sector and holds the key to its future development. This is not to deny that nonindustrial TVPS in commerce, construction, and other services are important, particularly in generating employment, but many of the results of the analysis of manufacturing are applicable to other rural nonagricultural enterprises. Within manufacturing an attempt was made to ensure that each county sample contained some enterprises in certain key TVP industries such as build- ing materials, textiles, machinery, and food processing. The sample was by no means confined to these subsectors, but it did not attempt to cover all industries in the localities concemed. The intent was to represent as wide a range as possible of such characteristics as size, technology, profitability, and scope of markets (local, regional, or national). These considerations, however, could be subordinated to the need to maximize variation in owner- ship and management forms. TVP workers were sampled by taking stratified random samples of employees in a smaller subsample of firms, about twelve in each county. A larger propor- tion of employees in smaller enterprises was sampled to ensure a meaningful minimum number of observations for each firm-roughly twelve employees in most cases, although in a few instances workers did not cooperate. Also, management personnel and technical workers were deliberately oversampled so that they would be adequately represented in the data set. In the survey of township government leaders, an attempt was made to obtain a response from every township where fieldwork was done and also (by mail or in a large group meeting) from every other township in the four target counties. This effort was moderately successful, although in contrast to the other questionnaires, which were administered personally, the re- sponse rate was substantially less than 100 percent. The information gathered in this research project lends itself to a wide va- riety of analytical methods, as is evident in the chapters that follow. At one extreme, the data permit formal statistical and econometric analysis at almost any level of sophistication. Chapters 11 and 13 make the most systematic use of such techniques. At the opposite extreme are the voluminous inter- view records and related materials, which are often detailed enough to serve as case studies and which can be utilized to paint reasonably clear pictures of the operations, institutional structure, and environment of the TVP sector. Chapter 7 is an outstanding example of this approach. Most of the chapters fall between these extremes; they utilize both the questionnaire-based data sets and the interview information, as well as avail- able national and regional statistics on the TVP sector. Although hypotheses are generated and supported oni the basis of these various kinds of evidence, most of the papers do not contain mathematical models or formal economet- Research Design, Methodology, and Data 25 ric tests. This intermediate approach is appropriate for much of the analysis, which to a large extent is still concerned with building up a coherent under- standing of the structure, operations, and environment of the TVP sector. Subsequently, as hypotheses are refined and models developed, it should be possible to utilize the information in the data sets more fully in formal statisti- cal and econometric analysis. The Data Set This section first looks briefly at the four counties chosen for in-depth field- work and at the townships where extra time was spent on interviews and information gathering. It then describes the five questionnaires and the inter- views conducted during the fieldwork. The Four Counties and the Townships The TVP sectors in these counties are discussed in some detail in chapters 4 and 7. Here we will briefly describe their main features. Wuxi County in J iangsu Province was selected because of its extremely high level of industrialization, both in the prereform period and since the late 1970s. For a long time it has been the most industrialized rural county in China, even though it is by no means the most urbanized. Wuxi is also the quintessential example of the traditional TVP system based on TVCES and tightly controlled under the supervision of township governments. This con- trol is especially evident in wage determination and labor allocation. Industry in Wuxi is highly efficient by Chinese standards. The TVCES are relatively large, many of them use relatively advanced technology, and they compete effectively with state industry. Private enterprises are severely ham- pered by administrative restrictions, and sizable ones have not emerged. Mar- kets for labor and land have not developed, but large amounts of financial capital have been brought in by county banks from other parts of the country to support local TVP development, and many enterprises or township indus- trial corporations have issued bonds to employees and local residents. Nanhai County in Guangdong Province is also highly industrialized and is more urbanized than Wuxi. It was selected because of the involvement of its TVP sector in foreign trade and in business relationships with the outside world-most notably, through extensive processing arrangements with Hong Kong businesses. Nanhai also has a diversified structure of rural communities, with much greater independence at the lower levels (villages and teams) than in Wuxi (see chapter 7). This diversified community structure has spawned a variegated pattern of TVP ownership. TVcEs are important but by no means dominant, and production team enterprises, private enterprises, firms con- tracted to private management, joint ventures, and even cooperative entities have emerged. The relatively open labor market has stimulated a large influx 26 WILLIAM A. BYRD AND LIN QINGSONG of migrant workers from poorer areas outside the county, and wage controls are less effective than in Wuxi. Capital flows from outside appear less impor- tant than in Wuxi, mainly because the county is well endowed with savings deposits that can be lent to TVPS. Jieshou County in Anhui Province represents a more or less average level of TVP development. It is much less industrially developed than either Wuxi or Nanhai, and its average rural per capita income in 1985 was somewhat below the national average. The most distinctive features of Jieshou are a lively market in land and the emergence of a substantial number of large pri- vate enterprises with the support and even the personal sponsorship of county and community government authorities. Jieshou is also the only county in the sample that has its own county seat (also called Jieshou).3 This allows county authorities to encourage relocation of successful private enterprises to the town, where infrastructure and support services are more readily avail- able. Private enterprises in Jieshou town are vulnerable to "creeping collecti- vization" in various forms, however. In the rural areas outside the county seat the level of industrialization remains extremely low, and most commu- nity enterprises have gon.e out of business or have been privatized. Shangrao County in Jiangxi Province was chosen to represent a level of rural industrialization and TVP development significantly below the national aver- age. Even though the cotnty is endowed with substantial mineral resources, rural incomes depend primarily on agriculture, industrial firms are small and mostly unprofitable, and industrial development is not yet self-sustaining, de- spite an infusion of capital from the central govemment. The traditional style of ownership and managemeni: of TVCEs has not been successful. TVPS are plagued by such problems as excessive levies by community governments to meet public expenditures, widespread overmanning, poor management, and lack of contact with markets outside the locality. Within each county one or two townships were selected as places where interviews and questionnaires would be concentrated. In Wuxi, Dongjiang township was chosen because, as measured by gross value of industrial output (GVIO) per capita, it is the most industrialized township in the county and, as of 1985, in China. Otherwise, interview time and data gathering were more or less evenly distributed among a dozen or more other townships scat- tered throughout the county. In Jieshou two townships-Hebei and Jinzhai-were chosen. Hebei, which borders on Jieshou town, has become increasingly urbanized and has steadily lost territory to the expanding town. Jinzhai is farther frorn the county seat and is much less industrialized. But the dominance of Jieshou town in the county's industrial TvP development was reflected in the selec- tion of sample firms, particularly since many relatively large private enter- prises are located in the town. Hence in Jieshou County there were really three focal points for information gathering. Research Design, Methodology, and Data 27 In Nanhai the research team spent extra time in two townships,4 Pingzhou and Xiqiao. The former had long been one of the most industrialized town- ships in the county, and many of its enterprises have long histories, some under different names and incarnations. Xiqiao is a well-known center for textile production. Both Pingzhou and Xiqiao are somewhat more developed than the average township in Nanhai. In Shangrao, Shaxi town served as a center for interviewing and data col- lection.5 Shaxi probably has the highest per capita income of all the town- ships in the county and is one of only a handful of townships that have achieved a respectable and stable degree of industrial development. As in Wuxi interview time and data collection were otherwise scattered rather evenly across a number of townships, some of them very poor and backward. The Questionnaires Five questionnaires were distributed: a Worker Survey Questionnaire for TVP employees; an Enterprise Survey Questionnaire for management teams of sample firms; an Enterprise Director Questionnaire for directors of sample TvPs; an Enterprise Quantitative Questionnaire, usually filled out by account- ants of sample enterprises; and a Township Leader Questionnaire. For com- parative purposes some of these questionnaires were also administered to a small number of county-run state enterprises and their employees. As can be seen from table 2-1, each questionnaire contained many items and required considerable effort to fill out and process. This was especially true of the questionnaires that requested detailed statistical information-the Enterprise Quantitative Questionnaire and the Township Leader Question- naire. A mix of quantitative and qualitative-factual, perceptual, and attitudinal-information was requested. Table 2-2 shows the samples generated by the questionnaires. It was in- tended that the three enterprise questionnaires (the Enterprise Survey Ques- tionnaire, the Enterprise Quantitative Questionnaire, and the Enterprise Di- rector Questionnaire) would be administered to the same firms so that data could be combined. This was largely, although not perfectly, accomplished, as can be seen from the table. The distribution of completed questionnaires among counties was fairly even, although the samples for Wuxi and especially Jieshou tend to be larger than those for Nanhai and Shangrao. A considerable amount of information was gathered through the question- naires. Rough calculations indicate that more than 370,000 data items were generated, although some of these are missing values.' Only about half of the data contained in the Enterprise Quantitative Questionnaire has been entered in the computer for statistical and econometric analysis.' Even so, nearly 250,000 pieces of data have been computerized, and this information serves as the main data base for research. Table 2-1. Features of the Questionnaire (number) Observations Observa- entered in Main types Questionnaire Target Respondent Questions tions computer of questions Worker Survey Er.terprise Employees 64 112 112' Personal information; employees attitudes and perceptions Enterprise Survey Enterprise Management 40 170 170 Qualitative behavior team questions, factual cs and attitudinal Enterprise Director Enterprise Director 42 175 175 Personal information; managers perception of role in firm Enterprise Enterprise Accountant 81 1,477b 609" Basic data on firm, Quantitative data 1970-86 Township Leader Township data, Township leader 22 316 315 Basic data on leader's attitude and staff township; leader's perceptions a. A condensed version of this data set, consisting of thirty-six variables, was also generated. It is extensively used in chapter 12. b. The amount of actual data for most sample enterprises is much less than this because few firms were established early and many of them only in 1984 or 1985. Research Design, Methodology, and Data 29 Table 2-2. Number of Questionnaires Completed in the Four Counties Questionnaire Total Wuxi Jieshou Nanhai Shangrao Worker Survey 1,174a 473 297 260' 144 Number of firms covered 49 14 15 11 9 Enterprise Survey 115 34 37 24 20 Enterprise Director 117 35 38 24 20 Enterprise Quantitative 122 32 37 23 30 Township Leader 67 19 27 12 9 a. Two completed questionnaires from Nanhai cannot be used in much of the analysis because the firms could not be identified. WORKER SURVEY QUESTIONNAIRE. The Worker Survey Questionnaire was designed to obtain basic factual information on each employee that could be used to test relationships between various dependent variables and workers' objective characteristics-for example, in wage equations (see chapter 13). It also asked for workers' attitudinal responses on various issues, which could then be systematically related to worker and enterprise characteristics. This questionnaire was drawn up by researchers with extensive experience in questionnaire-based research on Chinese state enterprise employees, and it benefited from ample time for revisions. The questionnaire was administered to stratified random samples of work- ers in, on average, a dozen firms in each county. Larger proportions of work- ers were sampled in small enterprises, and technical workers and managerial personnel were oversampled. Generally, workers were called together to have the questionnaire explained to them and were asked to fill it out on the spot and turn it in before leaving. Chinese researchers administered the question- naires directly, choosing samples from enterprise payroll lists. They gave the questionnaire directly to workers, explained it, waited while workers filled it out, answered any questions, and collected the questionnaires as they were completed. Workers were assured that their responses would be kept confi- dential, and the method of administering the questionnaire ensured confi- dentiality. ENTERPRISE SURVEY QUESTIONNAIRE. The Enterprise Survey Question- naire, designed to complement the Enterprise Quantitative Questionnaire, obtained a broad range of primarily qualitative information about the busi- ness situation and administrative environment of enterprises. The types of responses included multiple choices, prioritization among alternatives, and, in a few cases, rough percentage shares for enterprise factors, inputs, and out- puts. Despite some problems with questionnaire design, most of the data proved useful. The questionnaire was filled out by members of the enterprise's leading group, or management team. This was done collectively or, more 30 WILLIAM A. BYRD AND LIN QINGSONG often, individually by a deputy director of the firm. Questionnaires were ad- ministered in person during enterprise visits and were usually collected on the spot and then checked. Thus there are no serious problems with missing data. ENTERPRISE QUANTITATIVE QUESTIONNAIRE. The Enterprise Quantitative Questionnaire requested detailed statistical information on the sample enter- prises and included a few questions on ownership, location, management sys- tems, profit-sharing methods, and the like. Most items requested data for all years from 1980 and for earlier benchmark years (1970, 1975, and 1978). Only 28 percent of sample firms, however, had been established in 1978 or earlier. Problems included failure to complete certain items and incorrect or inap- propriate information. Data on capital structure turned out to be virtually unusable for these reasons. Even capital stock information is sometimes suspect-for some firms, the reported total value of investment for 1985 or 1986 exceeds the reported figure for the capital stock in the same year. Nevertheless, use of capital stock information in quantitative analysis leads to sensible results for the most part (see chapters 5 and 11). There were inter- nal inconsistencies in some of- the data-for example, revenues minus costs minus indirect taxes might not equal reported gross profits. Finally, in some cases numbers were filled in sc lightly or were so illegible that data entry was difficult. Many TVPS, particularly private enterprises and those located in the more backward areas, keep poor accounts, do not have skilled accountants, and hence simply could not provide all the information requested. Private enter- prises may also have had stronger incentives than TVCES to underreport or overreport various statistics, and, unlike the situation with the Worker Sur- vey Questionnaire, it was difficult or impossible to ensure confidentiality. De- spite all these problems, a large amount of usable quantitative data was gener- ated. ENTERPRISE DIRECTOR QUESTIONNAIR.E. The Enterprise Director Question- naire attempted to probe the attitudes and perceptions of sample 1TVP di- rectors, in particular their views on internal management, on managerial, business, and noneconomic objectives, and on relations with supervisory organizations. Some personal information on the director's background and characteristics was also requested. All of the questions were purely qualita- tive, although some asked for a ranking of alternatives. The questionnaire is the twin of the Enteiprise Survey Questionnaire and shares many of its advantages and drawbacks. Although design problems surfaced in some of the questions, a substantial amount of usable information was generated. In most cases the questionnaire was filled out personally by enterprise directors, Research Design, Methodology, and Data 31 but sometimes it was completed by a deputy director-most commonly be- cause the director happened to be out of town at the time. Unfortunately it is not possible to ascertain exactly which forms were not filled out by the director, but the percentage was probably small. TOWNSHIP LEADER QUESTIONNAIRE. The Township Leader Questionnaire asked for basic quantitative information on the township and its TVP sector, for qualitative perceptions, and for the leader's views and objectives. Unlike the other questionnaires, it was often administered indirectly (by mail), but sometimes research team members met with township accountants or other officials as a group to explain the questionnaire and answer questions. In the townships visited by the research team as part of fieldwork, completed ques- tionnaires were obtained directly and checked. The basic sampling objective was to reach as large a proportion as pos- sible of the total number of townships in each county. This goal was best at- tained in Jieshou, where completed questionnaires were retrieved for 82 per- cent of the townships and towns, and in Nanhai, where the figure was 75 percent. The proportion of townships responding was lowest for Shangrao (29 percent) and in between for Wuxi (53 percent). But in all four counties the share of townships sampled was large enough to be reasonably representative. The Interview Data In-depth interviews complemented the questionnaires and constitute a second main source of research data. As can be seen from table 2-3, more than sixty separate meetings were held in each of the four counties.8 An enor- mous amount of useful case-study, qualitative, and quantitative information was generated in this way. The 101 TVPS where interviews were done repre- sent a large majority of the sample enterprises. The nearly fifty meetings with township and village leaders also provided useful information, as did the in- terviews with county government agencies, financial institutions, and gov- ernment finance entities. Interviews were conducted with an agenda of basic questions in mind, but the agenda was not rigid, and meetings were allowed to meander in directions that were of interest to interviewers or were suggested by initial briefings or written material provided by interviewees. There was some evolution over time in both the questions asked and the organizations interviewed. More public finance units and financial institutions were interviewed during the second set of field visits (in Nanhai and Shangrao) than in the first (in Wuxi and Jieshou). Enterprise visits declined somewhat after the high reached in Jieshou, in part because of the decline in the size of the samples. Most inter- views involved more than one member of the research team. Usually several 32 WILLIAM A. BYRD AND LIN QINGSONG Table 2-3. Types of Organizations Interviewed (number) Category Total Wuxi Jieshou Nanhai Shangrao Enterprises 101 22 33 26 20 Township or village leaders 48 16 8 11 13 Village leaders 8 2 0 4 2 County leaders 8 2 3 1 2 County agencies 52 16 17 8 11 Banks, 12 2 2 5 3 Rural credit cooperatives' 8 1 1 2 4 Tax bureaus' 10 2 2 1 5 Finance bureaus' 9 2 1 1 5 Township industrial corporations 8 6 0 1 1 Township economic commissions 15 5 4 4 2 Other township government agencies 11 3 0 7 1 Note: These figures are incomplete because they include only meetings for which formal inter- view notes were compiled in English. a. Includes organizations at both county and township levels. people from the enterprise, community, or county were present, often repre- senting different organizations or levels of govemment. Generally discussions were open, and interviewees were responsive and forthcoming. Evaluation of the Data Set In this section our objective is to point out and assess some of the flaws in the various types of data as well as in sample characteristics. But despite the problems identified here, the data set is very useful and basically reliable, and the sample is in most respects al good one. Problems in Questionnaire Design Some flaws in questionnaire design came to light when the data were being processed and analyzed. One example is the inadequate space provided for answers on the Enterprise Quantitative Questionnaire, which made responses more difficult to read and increased the likelihood of mistakes in data entry. Sometimes questions were not worded in a way that permitted unambigu- ous classification of responses. More commonly, the possibilities given in multiple-choice questions were not conducive to clear assessment of re- sponses. In a few cases permissible responses did not include some altema- tives that hindsight revealed to be critical. Research Design, Methodology, and Data 33 The Worker Survey Questionnaire was the best designed of the group; the Enterprise Survey Questionnaire and the Enterprise Director Question- naire probably had the most severe design problems. A lacuna in the Town- ship Leader Questionnaire was that it did not request any personal infor- mation on township leaders. The Enterprise Quantitative Questionnaire perhaps asked for an excessive amount of data, which may have had an ad- verse effect on the quality of the information provided, at least for certain items. Despite their flaws, the questionnaires all generated much useful informa- tion. Often, suitable manipulation of response categories can ameliorate de- sign problems. Where this cannot be done, especially careful interpretation of responses and cautious use of the data are called for. Responses to Questionnaires Owing to the way the questionnaires were administered, the response rate was virtually 100 percent for all except the Township Leader Questionnaire, and even for that one it was respectable. Hence problems concern either pos- sible bias in samples (discussed later in this chapter) or incomplete, incorrect, inappropriate, or inconsistent responses in the completed questionnaires. Overall, response patterns show a considerable degree of logic and intemal consistency. This is especially true of the Worker Survey Questionnaire, where the number of nonsensical or inconsistent responses is minimal and compares favorably with similar data in other countries. Broadly speaking, responses to the Enterprise Survey Questionnaire and the Enterprise Director Questionnaire also exhibit logical and consistent patterns. Nevertheless, there are some problems with responses. For example, many sample firms left certain parts of the Enterprise Quantitative Questionnaire blank or filled them out so poorly that the information is virtually unusable. Another problem is that, according to the Worker Survey Questionnaire, workers in Jieshou County are almost unbelievably content with their jobs, firms, and managers; see chapter 13. (The high degree of satisfaction of Jieshou TVP workers may, however, be related to the fact that virtually all enterprises where the Worker Survey Questionnaire was administered are lo- cated in a more urban area, Jieshou town.) Also in Jieshou, private entrepre- neurs reported that they feel more responsible toward "the state" than toward themselves, their enterprises, or their communities. Such obvious biases ap- pear in only a handful of questions, but they show the need for caution in interpreting results when it is plausible that a sizable portion of respondents may have felt pressure (explicit or implicit) to answer in a certain way. In general, responses for all questionnaires except certain parts of the Enterprise Quantitative Questionnaire have been more complete and of higher quality than was originally expected. 34 WILLIAM A. BYRD AND LIN QINGSONG The Suitability of the Four Cournties Since the counties chosen for in-depth fieldwork in effect constitute a sample of four, their characteristics and their representativeness of the more than 2,000 counties of China are an important determinant of the value of the data set. Do the four counties cover the full range of institutional setups for the TVP sector, or are there mainy counties with TVP ownership, supervision, and management systems radically different from those sampled? The same question arises for basic features of the economic environment-goods mar- kets, factor markets, relations with the state sector, and the like. Finally, it would be a matter for concenn if the four counties were not really represen- tative of the main feature they were selected to typify-if, for example, it turned out that Tvp development in Jieshou was considerably above the aver- age for China as a whole or that Shangrao was not significantly below the national average. The four counties do represent an astonishing range of variation in TVP organization and development, not only in terms of the specific criteria for selecting the counties in the first place but also in several unexpected areas. One example is the dispersion of political authority among administrative levels in Nanhai, in contrast to its concentration at the township level in Wuxi (see chapter 7). Another is the personal sponsorship of large private enterprises by county officials in Jieshou. The lively land market in Jieshou is even more striking. The contrasts in labor market development in the four counties are also instructive and probably cover most of the variation that occurs in the country as a whole (see chapter 14). One important model of TVP development is not represented by any of the four counties. The famous "Wenzhou model" is characterized by free develop- ment of private enterprises (mostly household undertakings), a thriving fi- nancial market based to Et large extent on private financial institutions, and extensive commercial relationships with distant parts of China.9 Wenzhou itself was not included in the sample, partly for logistical reasons (travel there is not very convenient) and partly because so many delegations have already visited the area that much inforimation is already available and the receptivity of managers and officials to further in-depth investigations may be strained. Certain elements of the Wenzhou model are, however, present in the four sample counties-the land market in Jieshou, the labor market in Nanhai, ownership patterns in both counties, the "specialized" villages based on household industries in Jieshou's hinterlands, and the movement of tempo- rary migrant labor out of the poorer areas of Shangrao. (In Wenzhou people from the poorer mountainous areas move into the towns in search of better employment; in Shangrao thesc people move out of the county altogether.) But none of the four counties has a highly developed financial market or pri- vate financial institutions. More generally, the four counties may not be able Research Design, Methodology, and Data 35 to provide much insight into what happens when all the elements of the Wenzhou model are brought together and allowed to flourish. It was in the choice of this small sample of four counties that tradeoffs between different objectives were most severe. Inclusion of Wenzhou in the sample would have exacerbated logistical problems and loaded the sample with unrepresentative models, since Wuxi is already included. Substituting another county representative of the Wenzhou model might have skewed the sample too much toward "unusual" counties, since that would have meant dropping Nanhai, Shangrao, or Jieshou. Except for the exclusion of the Wenzhou model, the four counties do seem to provide an adequate range of variation in most dimensions of the TVP sector and cover most of the patterns that are present in substantial numbers of counties. The variety of possible ownership patterns is well represented, and there is a wide range of differ- ences in industrial structure, orientation toward markets, size of enterprises, and so on. As for the degree to which the four counties are representative of China as a whole, Wuxi and Nanhai are far more industrialized than most other counties, but Shangrao and the rural areas of Jieshou are, if anything, some- what below the national average (see chapters 4 and 7). Moreover, in the latter two counties some virtually nonindustrialized townships were visited. Since the sample could include only counties with at least a certain degree of Tvp industrial development (otherwise there would be nothing to study) and the areas of advanced TVP development also had to be covered, the choice of Jieshou and Shangrao seems appropriate. Characteristics of Firms Tradeoffs among research objectives are also evident in the sample of enter- prises. There was no attempt to obtain a random sample. Instead, enterprises were chosen according to certain criteria: (a) reasonable size and enough "history" for meaningful research and time-series analysis; (b) the widest pos- sible range of ownership and management forms, with oversampling of new and unusual variants; (c) coverage in every county of such key TVP industries as machinery, textiles, food processing, and building materials; (d) variation in performance (profitability), including some loss-makers; (e) the maximum possible range in firm size, subject to the constraint that firms must be sub- stantial enough to investigate; (f a range of technologies; and (g) different locations within the county. Sample characteristics were constrained both by the prior choice of coun- ties and by internal conflicts among the criteria, which could not all be ful- filled completely. Criteria c through g were achieved more or less satisfacto- rily. Criteria a and b, however, conflicted, since most private enterprises in the TVP sector were established only recently and are much smaller on aver- 36 WILLIAM A. 13YRD AND LIN QINGSONG age thian TVCES. Criterion b was also severely constrained by the choice of counties, since there are virtually no substantial private enterprises in Wuxi or Shangrao. These problems are evident in table 2-4, which shows the ownership of sample firms. Private firnns make up only 14 percent of the total, and. there are no private enterprises in the sample for Wuxi, only one in Shangrao, and only four in Nanhai. Jieshiou is thus the only county with a sufficient number of private enterprises (twelve) for many types of statistical analysis. The share of private enterprises in the sample is lower than their share in total numbers Of Tvps by counLty but may be reasonably close to the share of large private enterprises in the different counties. Village enterprises are poorly represented in jieshou, and there are hardly any enterprises run by production teams in any county but Jieshou. Thus, although the sample con- tains a sufficiently wide range of ownership and management forms, private enterprises are by no mean-s oversampled-on the contrary, there are too few of them. The close relationship between location and ownership in the sam- ple makes it difficult to disentangle county- and ownership-specific factors. Another question concerns the degree to which the sample firms are representative Of TVPS in their respective counties. Table 2-5, which looks at township enterprises only, since they constitute more than two-thirds of the total number of sample firms, compares sample means with county aver- ages for several statistical indicators. The most striking feature of the sa-mple is that firms tend to be much larger than average for their counties. In gross Table 2-4. Ownership of Sample Enterp-rises (percent, except as specified) Oumersliip - ~Entire sample Share of county subsample category Numtber of firms Share Wuxi Jieshou Nanhai Shangrao Township and town 83 68 78 51 57 87 Township 66 54 62 38 48 70 Town 1 7 14 16 13 9 17 Village 15 12 19 3 22 10 Production team 7 6 3 14 4 0 Private 17 14 0 32 17 3 Partnership' 10 8 0 24 4 0 Individual 7 6 0 8 13 3 Total 122 100 100 100 100 100 Note: All sample enterprises originally classified as "other" have been put in one of the regular ownership categories on the basis of a review of interview notes and a judgment as to which form of ownership is the dominant partner in the case of joint ventures. Some firms were reclassified if it was thought that the original questionrnaire response was inaccurate; this problem arose primarily in jieshou. a. Includes partnerships and family eniterprises. Research Design, Methodology, and Data 37 output value the sample mean for Shangrao is 2.5 times the county average, for Wuxi 5.3 times, for Nanhai 7.5 times, and for Jieshou 8.2 times. Sample enterprises are also somewhat more profitable than county averages; the ratio of profits to fixed capital for sample township enterprises is 22 percent higher than the county average in Jieshou, 55 percent higher in Nanhai, 64 percent higher in Wuxi, and 141 percent higher in Shangrao. Profitability varies greatly, however, and in all counties the ratio of loss-making township enter- prises in the sample is as high as or higher than the ratio in the county as a whole. Thus, although the sample includes a disproportionate number of firms in the upper tail of the distributions of both size and profitability, the lower tail of profitability is also adequately represented. Workers and Directors Despite some problems, stratified random sampling of workers seemed to work reasonably well, and no obvious biases are apparent. In few cases were workers reluctant to fill out forms. Thus the main question concerns the subsample of forty-nine enterprises where the Worker Survey was conducted. In general, this sample seems to mirror the larger sample of firms. For ex- ample, the lack of private enterprises in Wuxi and Shangrao is evident, whereas workers were surveyed in an adequate number of private enterprises in Jieshou and a smaller, but still respectable, number in Nanhai. There was some tendency to select larger, more stable firms for administration of the Worker Survey Questionnaire because these had larger labor forces and were less affected when a given number of employees was pulled off the job to fill out questionnaires. Other than this, there is no evidence of systematic bias in the selection of firms within the sample for employee surveys. The main problem with the Enterprise Director Questionnaire has already been mentioned: in some cases the form was filled out not by the director but by someone else, usually a deputy director. This probably had only a min- imal influence on the results: not many forms were affected; even where the deputy director filled out the form, personal information on the director may have been supplied; and in any case, the personal backgrounds and attitudes of directors and deputy directors in the same enterprise may not differ greatly. Overall Assessment Although some of the problems mentioned in this section are significant and reduce the usefulness of the data set, none raises doubts about its basic valid- ity and value. The painstaking, careful administration of questionnaires en- sured both a high response rate and at least a reasonably high level of quality, as measured by the effort made by respondents. Lack of experience in filling out questionnaires of this kind on the part of enterprise managers, workers, and community government officials created some problems in data gather- Table 2-5. Comparison of Sample Means with County Averages, Township Enterprises (millions of yuan, unless otherwise specified) Wuxi Jieshou Nan hai Shangrao Average township Sample Mean for Sample Mean for Sample Mean for Sample Mean for enterprise mean county mean county mean county mean county GVIO (current prices) 10.081 1.885 2.562 0.314 13.610 1.810 0.246 0.098 Employment (persons)- 451 131 131 55 448 123 82 52 Fixed assets" 2.245 0.488 0.527 0.068 2.249 0.695 0.161 0.114 Total assets' 4.274 0.959 1.022 0.094 5.315 1.194 0.231 0.153 Gross profits 1.780 0.220 0.183 0.018 0.787 0.157 0.022 0.007 Profit tax 0.368 0.040 0.029 0d 0.259 0.053 0.004 0.002 Net profits 1.431 0.180 0.137 0.018 0.494 0.104 0.017 0.006 Profit turned over to community government 0.502 n.a. 0.003 0.004 0.123 0.056 0.006 0.005 Ratio of profit to sales' 19.0 13.5 8.7 9.2 7.5 8.0 8.9 5.7 Ratio of profit to fixed assets' 73.8 45.1 33.0 27.1 35.0 22.6 14.2 5.9 Percentage of loss-making enterprises 12.51 11.7 2.7' 0 17.4' 18.0 23.3' 17.8 n.a. Not available. Note: Includes enterprises run by towns as well as by townships. oVIO, gross value of industrial output. a. At year-end. b. Original value, at year-end. c. Original value of fixed assets plus quota circulating assets, at year-end. d. Exact value Y414. e. Weighted averages were taken for the sample for comparability with county aggregates. f. Ratio of loss-rmaking enterpises to total number or firms filling out the Enterprise Quantitative Questionnaire (including those who left this item blank). g. Reported by county authorities, but obviously not accurate. Source: Information from fieldwork. Research Design, Methodology, and Data 39 ing, but it also meant that many of the respondents found the task fresh and interesting and put considerable effort into it. Shortcomings and biases in the sample should be kept in mind when inter- preting results. In particular, it must be remembered that some important segments of the TVP sector are not well represented-small firms, private firms, the most backward areas with few and small TVPS, and nonindustrial enterprises. Some of these gaps can be filled with aggregated local, provin- cial, and national statistics (see, for example, chapter 12). An important strong point of the sample is that it includes a significant number of poor performers-backward counties (Shangrao and rural parts of J ieshou), even more backward townships (in Shangrao and Jieshou, and, rel- atively speaking, several in Wuxi); and poorly performing enterprises. (In all counties a significant number of sample firms had losses in 1985, and some enterprises had obvious management, financial, technical, or other prob- lems.) This inclusiveness was a significant accomplishment, since provincial and lower-level government authorities naturally would prefer to show re- searchers their "better" localities and firms. Notes 1. Most of this work has appeared in Nongcun, Jirsgi, Shehui (The Rural Areas, Their Econ- omy, Their Society), published by the China Rural Development Research Group, the predeces- sor of the Development Research Institute. Four issues of this serial were published. 2. China proper is the term historically used for an area that contains about four-fifths of China's population. Jieshou County, the northernmost of the four target counties, lies on the southem edge of the North China Plain. Yuanping County, Shanxi Province, where an initial test investigation was conducted, is in the far northwest of China proper. Taihe County in Anhui, where some extra fieldwork was conducted, also lies on the North China Plain. 3. A county seat is a town of reasonable size that belongs to the county and serves as its administrative center. As some county towns grew, they were transformed into independent administrative units at an equal or even a higher level in the government hierarchy. In Wuxi and Nanhai the central city is actually in charge of the surrounding counties. In Shangrao the municipality is administratively separate from Shangrao County. 4. In Nanhai County, as in Guangdong Province as a whole, the terms used for rural commu- nities differ markedly from those used in other parts of the country. To avoid confusion and preserve uniformity, we have used the terms that apply to the closest equivalent unit in the other counties. Thus the community entities in Nanhai that were previously referred to as "dis- tricts" and were recently renamed "towns" are called "townships" here, since they roughly corre- spond to townships in the other counties, although they are somewhat larger. (When specific township-level communities in Nanhai are referred to by name, however, we use "town.") Simi- larly, we use the term "village" for what are called "townships" in Nanhai and the term "produc- tion team" for what are called "villages" in Nanhai. 5. In China the term "town" is used to refer to several different kinds of community govern- ment entity. Shaxi town and most of the other towns visited during fieldwork are equivalent to townships in their administrative level and population, but they are typically somewhat more urbanized. Jieshou town is a county seat and is hence of somewhat higher administrative rank than a township, although it is still well below county rank. The use of the term "town" in Nanhai was discussed in note 4. 40 WILLIAM A. BYRD AND UN QINGSONG 6. Since most firms were not established before the late 1970s, the items for 1970, 1975, and 1978 on the Enterprise Quantitative Questionnaire are usually missing values. Moreover, particularly in the Enterprise Quantitative Questionnaire, some parts were not completed or were completed so poorly that the data turned out to be unusable. 7. This is partly because so many enterprises left some questions blank or responded inappro- priately that the information is virtually unusable and partly because some of the questions (for example, in the income statement) asked for redundant or detailed information that was not deemed necessary for analysis. 8. The number of organizations visited exceeds the number of meetings held because many meetings involved more than one organization. This more than counterbalanced the fact that sometimes representatives of the same entity were interviewed more than once. 9. See Wang Xiaoqiang and others (1983) for the first detailed discussion of the Wenzhou model of TvP development. Also see lInstitute of Economics, CASS (1987), part 1. Part I Overview The four chapters in this part deal with a disparate set of topics, but they all provide general descriptions and analyses of various aspects of the TVP sec- tor. Hence they should be read before the later chapters that delve more deeply into narrower topics. In chapter 3 Du Haiyan analyzes the causes behind the rapid development of China's industrial TVP sector in recent years. He divides the past decade into two distinct periods. 1978-83, when agricultural reforms and the trans- fer of resources from agriculture provided the motivating force for rural indus- trialization, and 1984-85, when expanding market demand, policy reforms, and development of the TVP sector itself, particularly private enterprises, gave the impetus for even more rapid growth of TVP industry. Du notes that both in 1978-83 and earlier, population pressure on limited land resources stimulated the movement of rural agricultural labor into indus- trial activities. He goes on to show how the resources needed for rural industrialization-capital, technology and equipment, land, materials, labor, and markets-were marshaled in 1978-83. Many resources were supplied di- rectly or indirectly from agriculture, but by 1984-85 the ability of agriculture to sustain this contribution was exhausted. Financial flows began to move from the TVP sector to agriculture rather than in the other direction, and agriculture and industry competed for labor in some rural areas. But, as Du points out, rather than subsiding, TVP industrial growth acceler- ated in 1984-85. TVP industrial development had moved onto a self-sustain- ing path, aided by institutional and policy reforms, the most important of which was the legitimization of private enterprises. As a result, the private sector boomed, with even more momentum in many of the underdeveloped areas than in industrially advanced areas (some of which, like Wuxi County, effectively restricted private enterprises from developing rapidly or becoming large scale). The chapter describes both the measures that localities such as Jieshou implemented to promote development of private enterprises and con- tinuing problems and ambiguities, such as the dependence of private enter- 41 42 OVERVIEW prises on local government officials and the conversion of some private enter- prises into collective enterprises. An important element in the 1984-85 TVP boom was the greatly increased supply of bank loans to the sector. When bank lending was subsequently cut back, TVPS turned to other channels, including intemal accumulation, for funds. Perhaps even more important for TVP development was the sharp growth of demand for TVP industrial products in 1984-85, which was fueled by credit expansion and by large increases in both urban and rural personal incomes. The strong growth of intermediate demand from the burgeoning industrial TVP sector itself was another key factor. In the last part of the chapter Du looks at prospects for the future. He sees several grounds for pessimism. Even under conservative predictions of trends in capital-labor ratios, available capital will fall far short of the require- ments for generating employment opportunities for the rural people who will need to be shifted from agriculture to industry over the next several decades. Increasingly severe market competition, both within the TVP sector itself and from a reformed, more efficient state-owned industrial sector, will sap the ability of the TVP sector to expand as rapidly in the future. The government's ability to provide resources for the TVP sector will be increasingly limited, and the impact of policy liberalization will be weaker now that the one-time benefits of earlier reforms have been largely realized. The tight linkages be- tween TvPs and rural community governments-which involve administra- tive intervention by the latter, distortions in investment and employrment decisions, and siphoning of TVP profits to meet community needs-will be- come an increasing liability for TVPS in the future. In chapter 4 Jan Svejnar and Josephine Woo survey patterns of rural indus- trial development in the four counties where intensive fieldwork was con- ducted. After briefly presenting basic data for the four counties and describing salient given, or exogenous, conditions, the authors review the main charac- teristics of the counties' TVP sectors. Wuxi and Nanhai counties are far more developed than Jieshou and Shangrao. Industry is the most important part of the TVP sector in all four counties as well as in China as a whole, but the subsectoral structure within industry varies, as does ownership. The growth and performance of the TVP sector also vary greatly in the four counties, although in all cases the past several years have seen rapid increases in industrial output. TVCE labor pro- ductivity has been much higher in Wuxi and Nanhai than in Jieshou and Shangrao. But indicators of financial performance (profitability) suggest that the TVCE sector in Wuxi is somewhat more efficient than in Nanhai and Jieshou, whereas Shangrao lags far behind, considerably below the national average. Svejnar and Woo find dramatic variation in the role of foreign trade and foreign capital in TVP development, although in all four counties the export market is quantitatively much less important than domestic sales. In Wuxi, Overview 43 although exports are substantial, direct contacts between TVPs and foreign customers are limited, and there is little foreign investment. In Nanhai, by contrast, the TVP economy is very open, and firms rely heavily on foreign funds, equipment, expertise, and markets, most commonly through various kinds of compensation trade and processing agreements with businesses in Hong Kong and other places outside China. Jieshou has only just begun to develop foreign economic ties, and in Shangrao there is not one piece of foreign equipment in any TVP or in any other type of firm. The second part of the chapter analyzes the causes behind these regional patterns. Geographic and historical factors are obviously important, most notably the agricultural resource base and proximity to cities and markets. Each county's endowment of human capital also seems to be significant in explaining the performance of its TVP sector. In both respects Wuxi and Nan- hai are much better endowed than Jieshou and Shangrao, and Jieshou is somewhat better off than Shangrao despite the latter's abundant mineral de- posits. Access to financial capital is considered another important explanatory factor in regional variation. Wuxi and Nanhai are able to attract outside funds to support TVP development, and Nanhai has also made considerable use of foreign capital. But in Jieshou and Shangrao, given the relatively low degree of TVP development, it is probably not the lack of local funds but the poor performance of the TVCE sector, especially in Shangrao, that hampers the flow of financial resources. Still another key factor is county government policy. Svejnar and Woo argue that the county government can play an important, even determining, role in guiding TVP development in areas such as ownership. The timing of the implementation of the production responsibility system (PRs) and related agricultural reforms has varied greatly among the four counties. County gov- ernment policies can also affect labor allocation and mobility. The last part of the chapter generates several models of TVP development on the basis of the patterns observed in the four counties. Although the mod- els differ sharply, it is impossible to assert unequivocally that one is better than the others; all have demonstrated at least the potential for supporting successful TVP development under different circumstances. TVps are able to thrive in a variety of economic and administrative environments. The expe- rience of the four counties suggests that the principal obstacles to TvP devel- opment are inappropriate local government policies, a weak economic base, and inability to compete for material resources and for physical and human capital. In chapter 5 William Byrd and N. Zhu analyze the market environment and industrial structure of the TVP sector. They start by noting some stylized facts about the TVP sector: the primarily market-based allocation of TVP out- puts and material inputs; rigidities and administrative intervention in the al- location of factors of production-land, capital, and often labor as well; the 44 OVERVIEW dominance of TVCES in TVP-sector industrial output; close ties between all kinds of TVPS and their local rLral communities; and the relatively small scale of most TVPS. The ownership and financial structure of TVPS are then exam- ined, and the main characteristics of the product markets that they face are outlined. These characteristics include the tremendous profit and growth pos- sibilities in many markets at the outset; constriction of opportunities in mar- kets for these products over time; and relatively hard market constraints and competitive market conditions for most TVPS. The second part of the chapter focuses on pattems of competition and mar- ket adjustment in the TVP sector. One characteristic pattern is termed "imita- tive competition with neoclassical adjustment." When entry and exit are easy, an initial profit opportunity leads to widespread imitation by new en- trants, and thus to declining prices and profits, until some sort of equilibrium is reached. The "investment competition" pattem is similar but involves ac- tivities with higher investment requirements and often a high degree of "asset specificity" as well. As a result, investment competition can lead to large eco- nomic losses and may be highly risky. A third pattern, "innovative competi- tion," is emerging in some parts of the TVP sector. The chapter also reviews competitive and other interactions between TVPS and state enterprises. In industries where controlled prices are artificially low and a two-tier planning and market system has emerged, there may be a niche for TVPS that would not survive in an uncontrolled market environment. But in many industries characterized by buyers' markets, TVPS can hold their own and sometimes even achieve dominance. The state sector serves as an impor- tant source of business, technology, and personnel for TVPS. Althouglh coop- erative interactions may involve short-term financial and other benefits for state enterprises, in the long run these arrangements may, like interrLational outsourcing, transfer manufacturing capabilities to the TVP sector and threaten the position of state industry. Moreover, the most advanced TVP areas have already moved away from heavy reliance on subcontracting for state enterprises. TVPS' competitive strategies and methods of adjusting to adverse changes in market conditions are then discussed. Among the former are the "rush to be first" and rapid changes in product lines to take advantage of changing market opportunities. Adjustment to adversity occurs in a variety of ways, some more or less automatic, others discretionary. The authors conclude that TVP markets are for the most part reasonably competitive and that at both the enterprise and the aggregate level the TVP sector shows considerable resil- ience and flexibility. The authors next look at industrial structure in the TVP sector, focusing on sectoral composition, firm size, community industrial structure, and econ- omies of scale. The TVP industrial sector is diversified and covers nearly as broad a range of industrial subsectors as does Chinese industry as a whole. Average firm size is small, alithough firms of substantial scale have emerged Overview 45 in Wuxi and other places. The community ties of larger TVPS, among other factors, probably limit the size they can attain. A preliminary test involving sample firms indicates that there are some economies of scale, but this result needs to be interpreted cautiously. An important finding is that in its com- munity industrial structure Shangrao in most respects resembles the national situation much more than does Wuxi or Nanhai. The chapter ends with some conclusions and implications. Potential prob- lem areas are the limitations of imitative competition and investment compe- tition, which may further exacerbate gaps between more and less industrial- ized rural localities; the impact of greater asset specificity on TVP adjustment capabilities and strategies; and limitations on the size of firms. The authors argue in favor of policies to increase the mobility of financial and other re- sources among communities and localities and against administrative restric- tions against entry by TVPS into certain types of industrial activities. In chapter 6 He Jiacheng takes a broad, long-term look at future develop- ment issues and policy choices related to the TVP sector. After describing some overall govemment objectives for the Chinese economy and for the rural sector during the next several decades, the author discusses the complex and critically important interrelationships between the TVP sector and agri- culture. He argues that contradictions between rural industrial and agricul- tural development are a major obstacle. Of particular concem has been the apparent diversion of resources from agriculture to rural industry, which has led some to argue that the latter sector has been developing too rapidly. The limited resources of agriculture and the inability to get around them through foreign trade call for extreme caution in Chinese agricultural policy. In discussing policy solutions the author argues that whatever their merits and disadvantages, financial transfers from rural industry to agriculture, price increases for agricultural products, and other macroeconomic solutions are inadequate for addressing the fundamental problems of Chinese agriculture and their inhibiting effects on rural industrialization. A thoroughgoing land reform is needed to rationalize farm sizes, increase agricultural productivity, and promote regional specialization and interregional trade. The second main area covered is urban-rural imbalances. In addition to the effect of administrative barriers and the antirural bias of China's tradi- tional industrialization strategy, the similar structures of urban and rural in- dustry and the constraints on agricultural development tend to preserve and even exacerbate urban-rural gaps in development and living standards. Still another critical problem is the tension between rural industrialization and urbanization. The small-town urbanization that is occurring as part of TVP development carries benefits but also brings problems, the most important of which are scattered location patterns and the inability to take advantage of economies of scale in urban infrastructure and services. The author asserts that it is essential to improve human capital, particu- larly in the more backward rural areas, and he suggests more emphasis on 46 OVERVIEW education and training, the introduction of entrepreneurs from more ad- vanced parts of the country, and more appropriate income incentives. In ad- dition, industrial policies should ensure that rural localities focus on activities in which they have a comparative advantage, particularly labor-intensive in- dustries. Finally, a consistent strategy of urbanization for towns and smaller cities should be developed. In the last part of the chapter the author argues that although in many respects the ownership and institutional structure of TVCEs resemble those of state enterprises, the differences are important enough to have substantial effects on enterprise behavior and performance. But problems in organization and management hamper TVP development. These include the dual func- tions of township governments; the lack of strong incentives and direct re- wards for the ultimate owners of TVCES-the rural population; insufficient decisionmaking authority for rVCE managers; and shortcomings in labor allo- cation systems, which lead to excessively high wages and labor shortages alongside overstaffing in many parts of the TVP sector. Private enterprises suf- fer from external constraints and discrimination, which distort their behav- ior. To deal with these problems, contract responsibility systems have been instituted for TVCES in, for example, Jiangsu Province. Although these sys- tems are beneficial, they cannot fully solve institutional and organizational problems, which are inextricably linked with ownership issues and prop- erty rights. Reforms in the ownership system of TVCES, including partial privatization (through sales of assets to workers and other individuals) and better treatment of the private sector, will be necessary. 3 Causes of Rapid Rural Industrial Development Du Haiyan The rapid development of China's TVP industry has attracted wide attention both there and abroad. In 1986 there were 6.35 million rural industrial enter- prises, and they employed about 47.6 million people. The value of rural in- dustrial output in that year topped Y241.3 billion and accounted for 32.5 percent of the total value of rural social output and 23 percent of the total national value of industrial output. Between 1978 and 1986 the value of ru- ral industrial output increased about tenfold in Wuxi County and nearly twentyfold in Jieshou. Although growth rates differ among regions, TVP de- velopment was generally rapid. There were, however, sharp ups and downs (see figures 3-1 and 3-2). The output of TVP industry as a whole-whether in highly industrialized Wuxi County or in more backward Jieshou County- rose and declined twice during 1978-86. The fluctuations were so large and the cycles so short that the stability of the entire urban and rural economy was affected. These swings were obviously related to changes in government policies toward rural industries, to changes in the situation of the urban and rural economies, and to economic reforms. The development of rural industry in recent years can be divided into two phases: 1978-83 and the period since 1984. There were important changes in government policies, economic reforms, and the urban and rural econo- mies between the two phases. * The restrictions on farmers' nonagricultural activities were greatly re- laxed after the Third Plenum of the Eleventh Central Committee in December 1978, but for some time there were arguments as to whether the unfettered growth of TVCES would be efficient and compatible with socialism. Government institutions, from the highest authorities down to grass-roots departments, began fully supporting TVPS only in 1984. * The household production responsibility system (PRs) in agriculture was universally implemented by 1983, and by 1984 new reforms were needed in rural areas. 47 48 DU HAIYAN Figure 3-1. Annual Growth Rate of National and TVP Gross Industial Output Value, 1978-86 Percent 80 70 _ 60 _ Industrial TvP sector | \ 50 - 1 '46.0 40 _I 30 20.3 20 - '' 17.211 16.9 10 _ 9.9/ "--% 8.7 / 10.3 8.5 8 .137 - Ntoa industry 1979 1980 1981 1982 1983 1984 1985 1986 * After 1978 the output of agricultural products increased greatly, and by 1984 there were even small trade surpluses in grain and cotton. Since 1984 crop production has stagnated. * Urban economic reformrs accelerated in 1984, with some impact on rural areas and the TVP sector. 1978-83: The Initial Stimulus from Agriculture The growth of rural industry has all along been closely related to conditions in agriculture. For example, in Jiangsu Province, where Wuxi County is lo- cated, the growth of TVP industry has been in inverse proportion to the amount of rural arable land per person and the number of surplus agricultural laborers (table 3-1). It is clear that shrinking agricultural resources were the initial force behind the development of TVPS. In the twenty-five years 1952-7 7 China's rural pop- ulation grew by 55.6 percent, whereas the total area of arable land fell by Causes of Rapid Rural Industrial Development 49 Figure 3-2. Annual Growth Rate of TVP Gross Industrial Output Value, Wuxi and Jieshou, 1978-86 Percent 80 172.8 70 Jieshou y 60.1/ \ 60 - 50 .49.1 o~~~~~~~~~~~~~~~~~~ WXuxi 40 /98 1 J/ /~~~~~ 30 / m 27.5 29. 12.9 26.3' 26.1 20 .18.3/ 16.0% / 10 13.4 4.9 0 -2.7 1979 1980 1981 1982 1983 1984 1985 1986 130 million mu. Arable land per member of the rural population dropped from 3.29 mu to 1.85 mu. In Wuxi arable land per person declined from 2.52 mu to 1.27 mu and in Jieshou from 2.76 mu to 1.36 mu. During the same period agriculture's share of national income fell from 57.7 to 37.1 percent. Yet the proportion of agricultural labor in the national total increased from 82.6 to 84.2 percent, and in 1978 this proportion was more than 70 percent in Wuxi and more than 90 percent in Jieshou. The net output value created by each Chinese rural worker fell from Y323 to Y317. The rural economic situation, as measured by the relation between agricultural inputs and out- puts, thus worsened instead of improving. Between 1952 and 1977 rural people entered nonagricultural activities in three great waves. The first came during the First Five-Year Plan, when large- scale urban industrial construction absorbed a large amount of rural labor. As rural industrialization did not materialize in this period, the mass transfer from farming to nonagricultural activities was expressed in cross-regional population movements and urbanization. 50 DU HAIYAN Table 3-1. TVP Industrial Development and Agricultural Resources in Three Regions of Jiangsu Prcvince Per capita output value Farmland Farmland of TVPs per capita per laborer Counties Year (yuan) (mu) (mu) Suzhou-Wuxi-Changzhou 1980 585.0 1.10 1.94 1984 895.9 1.41 2.03 Nantong-Yangzhou-Zhenjiang 1980 258.5 1.20 2.42 1984 450.1 1.24 2.09 Xuzhou-Huaiyin-Yangcheng 1980 70.0 1.61 4.22 1984 110.9 1.58 3.47 Source: Information from fieldwork. The second wave occurred during the Great Leap Forward movement in 1958, when many communes and brigades set up small industries. The shift to nonagricultural activities during this period mainly involved transfer within localities, although there was also some movement into urban indus- tries. Blind developmeni- of rural industry created many problems and led to low microeconomic benefits and to macroeconomic imbalances. The read- justment of the national economy that followed not only forced some farmers who had entered urban industries to return to their homes in rural areas but also compelled many commune and brigade enterprises to stop operations. The third wave occurred in the early 1970s. In 1970 the State Council urged that the mechanization of agriculture be speeded up. Factories produc- ing farm machinery mushroomed, and the total output value of commune and brigade enterprises rose from Y1 billion in 1971 to Y27.2 billion in 1976. But their development was always restrained owing to policies that gave pri- ority to agriculture and especially to development of grain production. The most conspicuous change in China's rural areas since 1978 has been the adoption of the PRS, a reform that greatly increased farmers' enthusiasm and stimulated agricultural production. Rural productivity increased dramati- cally without great advances in agricultural production technology. In Wuxi, for example, agricultural productivity increased 1.7 times from 1980 to 1985. This, however, aggravated the situation of surplus agricultural labor. New rural economic policies that encouraged a diversified economy in the country- side were as significant as the PRS. Policies that discouraged rural labor from shifting to nonagricultural activities were basically abolished, and the move- ment of farmers into nonagricultural activities was accelerated. The transfer of rural surplus labor to nonagricultural activities involved several important factors: capital, land, labor, equipment and materials, and marketing activity by enterprises. Causes of Rapid Rural Industrial Development 51 Capital Initially, the accumulation of funds for investment in nonagricultural activi- ties was related to the unusual growth of agricultural incomes, which stemmed partly from the growth of agricultural production and partly from procurement price increases that returned industrial profits to agriculture. Be- fore the late 1970s, because of strict price controls on industrial and agricul- tural products, nearly one-third of the real national income created by agri- cultural production was transferred to nonagricultural activities, and this seriously affected farmers' incomes. Since 1978 state procurement prices for agricultural products have increased sharply; the purchase prices for the prin- cipal agricultural products rose by 50-100 percent during 1978-82 (see table 3-2). In 1978-80 alone the rise in procurement prices enabled farmers to reap Y46 billion in additional net income. From 1978 to 1983 farmers' average per capita income in China increased from Y133.6 to Y309.8, or 132 percent; in Wuxi during the same period it rose from Y124.4 to Y412.7, an increase of 232 percent. An increase in farm- ers' bank savings accompanied the rise in their incomes. The saving rate of China's rural residents, as measured by bank deposits, climbed by 7 percent- age points during this period. In Wuxi, a county with high incomes, the in- crease in rural bank savings was all the more striking; between 1978 and 1983 the total amount of rural bank deposits increased by 486 percent, and indi- vidual deposits rose by 600 percent. The growth of rural deposits helped ex- pand local banks' supply of funds and has been a positive force in the develop- ment of TVPS. For the country as a whole the year-end outstanding balance of loans extended to TVCES increased from Y1.5 billion in 1977 to Y7.35 bil- lion in 1982, a rise of 390 percent. The growth of farmers' incomes also added directly to the supply of funds for TVP industry. In Jieshou and Shangrao capi- Table 3-2. Weighted Average Index of Prices Obtained by Agricultural Producers, 1978-82 (1978 = 100) Annual growth rate Index (percent) Product 1979 1980 1981 1982 1978-79 1979-82 Grain 126 137 145 149 25.6 5.9 Edible oils 141 151 161 159 40.3 4.1 Pork 137 155 162 166 37.0 6.6 Beef 120 153 174 226 19.7 23.3 Eggs 122 124 134 135 21.9 3.5 Aquatic products 149 176 199 211 48.6 12.3 Cotton 118 139 137 142 17.6 6.5 Source: State Statistical Bureau (1983). 52 DU HAIYAN tal was often raised by requiring workers to finance the enterprises they joined-in some enterprises each worker might bring in as much as Y10,000 in capital. Capital from agriculture was a forceful stimulant for TVP industry in devel- oped counties such as Wuxi but was less significant in, for example, Jieshou. In 1982 the per capita income of rural residents in Jieshou was merely Y99, and per capita bank deposits were about Y4. 1-19.4 percent and 2.2 percent of the respective figures for Wuxi. Even so, the growth of agricultural income played a positive role in prornoting the development of TVP industry. The transformation of money capital into real capital also helped rural labor shift into nonagricultural activities. At present, the industrial TVP sector is still a "talent-type" economy; the success of a firm depends heavily on the quality of its manager. This affects the sources of technology for TVPS. But different regions have different characteristics. Well-located Wuxi County received its technology directly from urban industries; workers who returned to rural areas during the early readjustment of the national economy in the 1960s became founders and technical personnel of local TVP industry. In addition, TVPS in Wuxi have established cooperative technological rela- tions with their urban counterparts. By the end of 1985 about 856 enterprises in Wuxi had links with 1,097 firms in large cities for the production of 1,159 products and had established cooperative technological ties with more than 350 institutions of higher leaning. At the same time, the county's own tech- nological potential has grown rapidly. TIhe situation is different in Jieshou, which is far from big cities and has been plagued by disadvantages in developing products and technology, largely because of poor transport facilities. Under these circumstances the county found its own way by purchasing technology and talent at high prices. Production of a sprinkling machine, at one time the county's most important product, began in this way. Another channel for transfer of technology is through relatives and friends. A handful of managers has played a vital role in obtaining technology for the startup of TVPS in Jieshou. Land The implementation, beginning in 1978, of the policy of encouraging diver- sified activities in rural areas brought into the open the existing gap in in- comes from land. As the figures below show, industrial income from a plot of land is far higher than the value of its agricultural output. (The data, which were obtained from fieldwork, are in yuan per mu; n.a. = not avail- able.) Wuxi Jieshou Nanhai Shangrao Agriculture 286.7 149.7 195 194.2 Industry 63,359.5 42,711.4 n.a. n.a. Causes of Rapid Rural Industrial Development 53 Hence there is a strong incentive for shifting land to industrial uses. From 1979 to 1984 about 55 million rural residents entered TVPS, which occupied a total of 30 million mu of arable land. Arable land was reduced by another 15 million mu in 1985; 50-60 percent of that was taken by TVPS. Primitive land markets began to emerge in many places. In Jieshou sample firms paid, on average, Y7,930 per mu to purchase the land they occupied and Y1,962 per mu per year to lease it. Land markets formed in this re- gion because of the concentration of infrastructure and the inflow of TVPS. Many TVPS have moved to Jieshou town to obtain electricity and better transport, causing a continuous rise in land prices. Land markets not only favor the rational distribution of land resources but also give firms not run by community governments an opportunity to survive. If the firms have money, they can get land through land development companies run by local govemments. Labor According to Todaro (1969), the transfer of labor from rural to urban areas is affected by the urban-rural wage differential and by the opportunities for nonagricultural employment. In rural China the basic unit of economic ac- counting is the household. When land was contracted according to head count, it did not matter whether a household member was engaged in agricul- tural activities, since the land could be farmed by other members of the fam- ily. As a result, the opportunity cost of a farmer's tuming to industrial activi- ties was practically zero. Besides, there were large discrepancies between personal incomes from agriculture and from work in TVPS. In 1980 rural per capita income from agriculture was Y166, whereas per capita income in TVPS was Y398. In 1985 these figures rose to Y351 for agriculture and Y726 for TvPs. Such wide differences made farmers even more eager to get into the nonagricultural sector. TVP industry at this stage thus faced an abundant sup- ply of labor, which promoted its development. Equipment and Materials After 1978 planned procurement of agricultural products by the state shrank even as farm output grew, which stimulated the growth of TVPS that used agricultural raw materials; between 1978 and 1983 the output value of such TVPS increased by 254 percent. Meanwhile, TVPS that produced energy and raw materials also grew rapidly and helped form a new chain for the flow of materials in rural areas. For example, in Jiangsu Province the share of township enterprises in the total provincial output of certain key industrial producer goods was already high in 1983, as the figures below, obtained from fieldwork, indicate. 54 DU HAIYAN Percent Rolled steel 11.3 Metallurgical materials 18.2 Iron ore 8.0 Fireproof materials 28.6 Nonferrous mnetals 30.0 Copper products 27.4 Aluminum products 21.0 The adjustment of the country's national industrialization strategy also en- abled TVPS to receive some materials subject to monopoly distribution under the state plan. The proportion of state-distributed coal in total coal con- sumed by TVPS in Wuxi in 1980 was 32.9 percent; the figure was 41.3 percent for electricity, 62.7 percent for petroleum products, and 5.3 percent for steel. Some localities, such as Shangrao, have even formulated rules to ensure that a certain proportion of their allocations of materials from the state is distrib- uted to TVPS. TVPS get their equipment from many sources. In highly developed Wuxi, where equipment is often advanced and up to date, there has been a trend since the early 1980s toward transferring equipment out of the county to eco- nomically underdeveloped regions. In backward counties such as Jieshou ob- solete equipment makes up a large share of the total. About 38 percent of the county's equipment was made before 1970, and about 79 percent: of TVP equipment consists of items handed down from state enterprises. The propor- tion of locally made equipment is also high in backward regions. Marketing All along, TVPS have been sensitive to market trends because they have had no access to state commercial channels, and TVP industry has placed strategic emphasis on sales. Most TVPS have followed the pattern of having one person in charge of production, two in charge of purchases, and three in charge of sales. This has led to a widely held belief that the success of TVPS owes much to their marketing. There are more than 10,000 sales personnel in Wuxi TVPS, and Jieshou has thousands of people who market nylon ropes around the country. Generally, the more irregular the production, the higher the ratio of salespeople. TVPS use several methods to improve supply and sales work. * Widespread participation by enterprise staff in obtaining supplies and promot- ing sales. Enterprises in 'Wuxi use three main methods: they designate permanent supply and sales personnel whose salary is fixed but who re- ceive bonuses and whose hotel, food, and transport costs are covered by the firm; they hire part-time supply and sales personnel who receive a bonus for the amount of work they do but who are not recompensed Causes of Rapid Rural Industrial Development 55 for travel and lodging; and they hire temporary supply and sales person- nel who negotiate their terms with the enterprises. * Large bonuses. Incomes of supply and sales personnel are generally sev- eral times higher than the average for all employees and may even ex- ceed a factory director's salary. Some salespeople earn more than Y100,000 a year. Statistics from sample firms show that incomes of sup- ply and sales personnel are still rising. * Ample funds. In TVPS in Jieshou it is generally accepted that 3-5 percent of the value of purchased inputs or of products sold can be drawn out as operational fees and funds to gain access to supply and sales links. State enterprises, with their relatively stricter financial accounting in- spection systems, cannot do this. * Exchange of agricultural and sideline products. Shortages of and high prices for agricultural and sideline products have led TVPS to use them as quasi- currency. This practice enables TvPs to buy raw materials that even state enterprises find hard to get and to sell their substandard but high- priced goods to state commercial departments. In counties where TVPS are highly developed, agricultural and sideline products are often used as a medium for establishing cooperative ties with urban industrial bases. i Nonprice competition. TVPS face an environment in which structural re- forms in state enterprises are just starting. The problem of the soft bud- get constraint in traditional public ownership has yet to be solved. The shrinking of administrative controls and the strengthening of the profit motive for state enterprises under the reforms have led to rampant growth of activities such as barter trade, unplanned sales of products listed in the state plan, and speculation on commodity quotas at state prices. This environment provides an opportunity for TVPS to shift part of state enterprises' profits into their own hands by such means as giving sales commissions, paying sales agents, and making out blank invoices. Hence TVPS, whose backward technology is often coupled with high input costs, can sometimes outcompete state enterprises that have ad- vanced technology and low input prices. Although TVPS have to pay much more for inputs, they boast higher economic benefits (see chapter 18). These factors have enabled TVPS to occupy a substantial share of the market for many industrial goods. 1984-Present: Market Expansion and Self-Supported Progress TVP development in the early stages depended heavily on financial support from agriculture and on the abnormal growth in the supply of production in- puts from agriculture. But by 1984 agriculture could no longer adequately support TVP growth. On the contrary, part of the profits from TVP industry 56 DU HAIYAN came to be transferred to agriculture, especially in industrially developed re- gions like Wuxi. With the rapid development of TVP industry, agriculture fell into "sunset industry" status, dropping from its dominant role to second- ary importance. In 1979-85 TVP industrial profits used to subsidize agriculture in all of Jiangsu Province totaled more than Y4 billion, accounting for about 30 percent of TVPs' after-tax profits. In the two years 1984 and 1985 in Wuxi, Dongjiang township alone allocated about Y6 million of industrial TVP profits to the construction of farm infrastructure and the mechanization of agricul- ture. Every mu of the more than 1,000 mu of arable land of Hujia village in Yanqiao township absorbed an investment of Y600 from the profits of TVP industry. In addition, community governments took part of TVP industrial profits to subsidize the prices of agricultural products and so stimulate agricul- tural production. The subsidy rate in southem Jiangsu was as high as Y0.5 per jin of rice. Competition for labor also has become a serious problem in Wuxi, where 80 percent of the agricultural labor force is already engaged in TVP industry. In Dongjiang township local enterprises are forced to give ten days of leave to all their workers during the two harvest seasons every year, at a cost of Y7 million of output value and Y700,000 of profits. The impetus given to TVP industry by agriculture has obviously fallen short of needs. Despite these problems, in 1984 and 1985 TVP industry not only (lid not shrink but achieved even higher growth rates. The main factors pushing growth during this period were the self-development of TVP industry and changes in microeconomic policies. Local policies in support of TVP industry were instituted. The county government in Jieshou, for instance, used pro- motion of Tvps as an important criterion for evaluating community govern- ment officials and required all townships to have two or three enterprises with an annual output value of Y30,000 by 1984. To supplement this strategy, financial aid, credits, and tax breaks were extended. The burgeoning private economy became an important element in the rapid development of 1rvP in- dustry, especially in economically underdeveloped regions. The Central Committee's Document No. 1 of 1984 allowed private firms to hire workers and relaxed restrictions on them. This brought about a rapid development of private enterprises (see table 3-3). New cooperative economic entities-farmers' joint cooperative enterprises ("partnerships")-have emerged. These enterprises are often initiated and fi- nanced by several households and hire varying numbers of workers. Farmers who hold stocks are both managers and producers. In addition to their nor- mal wages they are entitled to dividends according to the number of shares owned. Shareholders tend to be relatives or fellow village members. These new firms are actually private in nature. In underdeveloped regions the momentum for the development of the pri- vate economy has been all the stronger. Some local authorities, in drawing up their development scenarios, have pinned their hopes on relaxed political Causes of Rapid Rural Industrial Development 57 Table 3-3. Development of Private Enterprises, 1984-86 (percentage of total for TVP sector) 1984 1985 1986 Characteristic of private enterprise Number Share Number Share Number Share Number (thousands) 4,202.2 69.3 10,374.6 84.9 13,425.4 88.6 Partnerships 906.3 15.0 1,121.1 9.2 1,093.4 7.2 Individual firms 3,295.9 54.3 9,253.5 75.7 12,332.0 81.4 Employment (millions) 12.26 23.5 26.52 38.0 33.96 42.8 Partnerships 5.24 10.0 7.71 11.0 8.34 10.5 Individual firms 7.02 13.5 18.81 27.0 25.62 32.3 Output value (billions of yuan) 24.40 14.3 67.89 24.9 102.45 29.0 Partnerships 12.65 7.4 24.50 9.0 31.01 8.8 Individual firms 11.75 6.9 43.39 15.9 71.44 20.2 Sources: Information from fieldwork, and Ministry of Agriculture data. controls and on encouragement of the private economy. The number of pri- vate enterprises in Shangrao, for example, more than doubled within one year as a result of such policies (see table 3-4). In Jieshou, of the 6,330 TVPS in 1985, private enterprises numbered 5,604, or 88.5 percent of the total. Underdeveloped regions like Jieshou have adopted four main supportive mea- sures to encourage the private economy. * Relaxation of controls on registration. Usually a person who applies to open a business gets approval within a week. * Changes in the residence system. In Jieshou rural residents can buy urban resident status for Y20,000. Farmers are allowed to establish enterprises in urban areas on condition that they provide a certain amount of em- ployment to urban residents. * Financial support. A private enterprise can obtain bank loans if two well- managed enterprises act as its guarantors. * Assistance in making connections and signing contracts. According to the rules set by the Shangrao County government, a private enterprise with less than Y200,000 annual output value can call itself a village firm, and a firm with more than Y200,000 annual output value can be called a county collective. These measures have increased the credibility of private enterprises. Policies toward the private economy differ among localities. Wuxi relaxed controls on private enterprises for a short period in 1984, only to discover that private firms posed a threat to TVCES. It soon reversed the policy. The county industrial and commercial bureau does not allow private enterprises to compete with state enterprises and TVCES for raw materials, energy, tech- nology, and human resources. Private transport firms are not allowed to hire 58 DU HAIYAN Table 3-4. Ownership Structu,re of TVPs, Shangrao County, 1984-86 (percent) Type of Shlare of wotl number of TVPs Increase over previous year TVP 1984 1985 1986 1985 1986 Township enterprises 8.0 3.6 3.7 -2.4 4.6 Village enterprises 12.0 5.3 5.4 -3. 8 3.6 Partnerships 1 80.0 91F1 11. } 150.1 4 1.4 Individual firms 79.8 Total 100.0 100.0 100.0 119.3 1.6 Source: Information from fieldwork. workers, private construction firms are not allowed to take on public con- struction projects, and private industrial and commercial households are al- lowed to hire only one or two helpers and three to five apprentices. These restrictive policies and the predominance of the highly developed local TVCE sector have limited the development of private industry, and as a result it is unstable. In Wuxi in the first half of 1986, 2,120 business licenses were issued to private industrial and commercial households, but during the same period some 1,485 households stopped business operations. In the first six months of 1987 the number of licenses issued was 2,601, and 2,072 busi- nesses stopped operations. TVCES have always held the dominant position in Wuxi's TVP industry, accounting for more than 90 percent of total output value in 1985. The second element that spurred the rapid development of Tvp industry after 1984 was changes in the supply of funds. The 1984 policy of supporting the growth of TVPS led to a considerable relaxation of restrictions on bank loans and a consequent credit inflation that directly augmented the supply of funds to TVPS. In 1984 and 1985 national rural loans increased by 60 per- cent. Loans extended by local banks to TVPS in Jiangsu Province reached Y3.39 billion in 1984, an increase of 130 percent over the previous year. Similar increases occurred in Anhui Province. The proportion of loans pro- vided by banks and credit cooperatives in total funds of sample firms in Wuxi was only about 1 percent in 1978 but 23 percent in 1984. For Jieshou's sample firms the 1985 figure was 37 percent, up about 10 percentage points from the previous year. Controls on bank loan quotas in rural areas in 1985 seriously affected the volume of credit extended to T'VPS. In May and June of 1985 a total of about Y300 million in bank loans was withdrawn throughout Jiangsu Province. In this situation TvPs began to search for new sources of funds, including local residents and cooperative investments. In 1985 Tvps all over China mobilized more than Y1O billion from these sources. According to statistics made avail- able by Wuxi's Agricultural Bank, 65 percent of the money pooled by local Causes of Rapid Rural Industrial Development 59 TVPS was obtained by squeezing bank savings, meaning that bank funds were still reaching TVPS through indirect channels. The self-accumulation capacity of TVPS was also expanding. Between 1980 and 1985 total profits used to expand production in TVCES rose by 68.5 per- cent, and the share of this investment in total profits rose from 49.7 to 51.4 percent. The share of profits reinvested in expansion by sample firms in Wuxi rose from 47.9 percent in 1980 to 57.5 percent in 1985. Overall, despite the reduction of agricultural funds in the capital of TVPS in 1984 and 1985 and the return of part of the profits of TVP industry to agriculture, the total supply of funds was enlarged owing to improvements in the functioning of financial markets and greater self-accumulation by TVPS. Changes in product markets were perhaps an even more important factor in the speedy development of TVP industry during this period. As TVPS lack fixed channels for sales, they are susceptible to fluctuations in commodity markets. The loss of macroeconomic control in 1984, which caused an exces- sive growth of demand for both investment and consumer goods, provided an important stimulus to TVPS. Meanwhile, reforms in the system of manda- tory purchases and sales by the state in urban areas allowed TVP products to enter urban commercial channels on a more stable basis. With the development of TVP industry and the expansion of nonagricul- tural activities in rural areas, the share of agriculture has shrunk. In Jieshou in 1978 agriculture accounted for 57 percent of the total value of industrial and agricultural output, industry for only 43 percent. In 1984 the value of industrial output exceeded that of agriculture, and by 1985 the proportion of the value of industrial output in the total reached 55.2 percent. In Wuxi, a developed county, the production structure is much more industrialized, and the value of industrial output is more than eleven times as large as the value of agricultural output. With these changes in the rural production structure, a new process of in- come generation has emerged in rural areas. TVP industry is increasingly im- portant for the growth of farmers' incomes. During 1981-85, Y7.1 billion was transferred by TVPS to local production teams for distribution, and wages of workers in TVPS were Y123.3 billion. The two figures totaled more than Y130 billion, an average of Y326 per member of the rural labor force. Aver- age per capita rural income, which was Y191 in 1980, jumped to Y397 in 1985. In 1985 wages of TVP workers reached Y47.3 billion, an average of Y675 per worker, accounting for 15 percent of average per capita rural in- come. The rise in rural incomes brought about an increase in consumption and changes in the consumption structure. Between 1978 and 1985 the total vol- ume of retail sales in rural areas of China grew by 238 percent. The annual growth rate of the consumption level of rural residents was 4 percent higher than that for urban residents. Meanwhile, there were three significant 60 DU HAIYAN changes in rural consumption structure. (a) The value of purchased commod- ities (as opposed to self-produced goods) in the consumption of the rural pop- ulation in Wuxi increased at about 20 percent a year. (b) The Engel coeffi- cient in the consumption structure dropped by 10 percent. (c) Investment in housing rose; a total of 4.1 billion square meters in housing floor space was constructed in rural areas throughout the country between 1978 and 1985. At Y40 per square meter, this represented an estimated total invest- ment of Y164 billion. The growth of farmers' incomes since 1978 has enabled them to enter the consumption goods market. Rural demand stimulated the growth of TVPS that manufacture consumer goods, thus creating a new source of income for rural residents and further spurring their demand for consumer goods. Rapid pro- duction increases in TVPS thatL produce consumer goods stimulated the devel- opment of firms that manufacture investment goods and further increased rural incomes. While creating income for the rural population, TVP industry was building up an ever wider market for itself. These chain effects have be- come more and more conspicuous since 1984. Prospects for the Future Three basic elements stimulated the rapid growth of China's TVP industry between 1978 and 1985. Policy changes permitted key rural factors of produc- tion to shift to nonagricultural activities. Changes in the macroeconomic en- vironment and in market conditions created space for the survival and expan- sion of TVP industry. And structural changes in TVP industry itself and the chain effects caused by the initial rise in agricultural incomes further spurred growth. At first, policy played the main role, but market conditions and structures assumed greater importance in the second stage of development. In developed regions market conditions have been dominant factors in the growth of TVP industry; in backward areas policy changes have been more important. This analysis provides a starting point for a preliminary assessment of the prospects for the developmen: of China's TVP industry. But it should first be noted that specific conditions--isolated urban and rural economies and sepa- ration of factor markets-gave birth to China's mode of rural industrializa- tion: "leaving the land but not the village, entering the factory but rnot the city." TvPs have not yet completed the split from their mother-farming- and as a result they keep in close contact with the traditional rural social framework, the community. MIost TVP workers still maintain the right to own and control land. Many are still engaged part-time in agricultural production. This incomplete shift of farmers to nonagricultural activities is characteristic of China's rural industrialization. Industrialization has moved ahead while ur- banization has lagged behind. The newly created industrial organizations are still clinging to traditional agricultural structures. TVP development, as a real- Causes of Rapid Rural Industrial Development 61 istic path to restructuring the rural production structure and transferring sur- plus agricultural labor to nonagricultural activities, is only one phase in the process of rural industrialization and is linked to particular historical condi- tions. With the deepening of economic reforms in urban and rural areas, a generalized market for factors of production will emerge, and TVPS will be included in the modem economic system on the basis of an organized, rapid pace of urbanization. From this long-term perspective, the current rapid growth of TVP industry is not a stable phenomenon. TVP industry will face the following difficulties in the future: * Shortages of funds (the biggest difficulty). During the eight years 1978-85 TVPs employed nearly 70 million laborers. The plan is for them to hire another 185 million workers in the fifteen years 1986-2000-about 50 percent of the total rural labor force. But the strengthening of the technologi- cal structure of TVP industry has brought about capital intensification much too early, and TvPs have paid an increasingly high price for employment gen- eration. In 1979-84 each new employee cost TVCES Y3,361 of fixed assets; in 1985 TVCEs had to provide Y5,347 of fixed assets for each new employee. On the basis of the 1985 figure, the transfer of 185 million agricultural labor- ers into nonagricultural activities would require an investment of Y989.2 bil- lion, but the total net profits of TVPS in 1985 were only Y28.74 billion. Even if 40 percent were invested (which is quite generous), that would be only Y11.49 billion. If profits grew at 10 percent a year, the cumulative total would barely reach Y417.0 billion by 2000. Self-accumulation by TVPS there- fore cannot realize the goal, and sufficient state financial aid and a favorable credit policy may not be forthcoming. A wide gap exists between the supply of and demand for funds for TVPS. T The progressive decline in the impact of resource mobilization and policy read- justment. With the reform of state mandatory planning, economic resources in the hands of the central govemment have been reduced, and the govern- ment's ability to support the growth of TvPs by allocating additional resources is disappearing. The effect of policy readjustment will also diminish. The main reason for the unprecedented effect of the government's rural economic policies on TVP industry after 1978 is that the economy had amassed huge amounts of untapped potential during the long period before the reforms. Once policy restrictions were relaxed, this energy was released in a short time. But in the future the effect of policy changes will be progressively re- duced. * Rapid increases in labor costs. Much of China's economic takeoff is de- pendent on cheap labor in the countryside, but this labor cost advantage is dwindling at an alarming rate. Between 1978 and 1986 the total TVP wage bill increased 5.76 times, from Y8.7 billion to Y58.6 billion. Workers' per capita income increased 1.64 times, from Y306 to Y809. In 1978 workers' per capita income in Wuxi sample firms was Y456; by 1985 the figure had risen to Y1,261, higher than the income level of workers in state enterprises 62 DU HAIYAN in the same area. Wage increases in the relatively developed areas seem ex- cessive. If the national average income of TVP workers could be fixed at the 1985 level, Y172.38 billion itn pay-2.65 times the present level-would be needed to enable the nonagricultural sector to absorb 185 million members of the rural labor force. To bear this tremendous labor cost and maintain the present profit rate, TVP industry would have to increase its output value by Y723 billion. Under such circumstances, resource bottlenecks are likely to become even more severe. * Worsened market conditions as a result of sharp competition. The rapid growth of TVP industry in 1984-85 depended partly on a market bloated by the loss of macroeconomic control and conceded by state enterprises owing to their lack of vitality. But reforms in state enterprises will change their or- ganizational form and property relations so as to hold down their demand and build up their competitive strength. TVPS will then no longer be the sole beneficiaries of the expansion of the market mechanism, and their favorable market opportunities will gradually diminish. At present the industrial and product structure of TVPS is highly similar to that of state enterprises. Equal competition in the same market space will be disadvantageous to TVPS, which are plagued by low technological levels and poor product quality. * Obstructions in the organization of TVPs themselves. The most basic orga- nizational characteristic of TVI'S is their dependence on the rural community, as reflected in the deep involvement of community govemments in enterprise management. With the rapid development of TVP industry, this kind of ad- ministrative interference affects firms adversely. Although TVP investments are based partly on the profit motive, the motive of regional development is much stronger. A dispersed regional development pattem will lead to a shortage of working capital and irrational distribution of industries, and TVPS in general will suffer the consequences of being uncompetitive and uneco- nomically small in scale. Local govemments' objective of solving employ- ment problems tends to prevent the rational movemnent of human resources and leads to the coexistence of overstaffing and shortages of workers inr TVPS, especially in developed regions. Furthermore, the system of ownership by community governments gives birth to unreasonable distribution mecha- nisms. In Wuxi, for example, a township draws, on average, Y1.2 million-Y1.6 million a year in administrative and social welfare expenditures from TVPS' profits-equal to roughly 45 percent of their after-tax profits. With the heavy burden of comamunity govemment finances on their shoul- ders, TVPS will be plagued by obstructions to their development. 44 Development Patterns in Four Counties Jan Svejnar and Josephine Woo The TVP sector has become increasingly important in China's economic de- velopment. But there is great geographic diversity in TVP performance. Our research project looked at four counties: Wuxi in the Yangtze River Delta, Nanhai in the Pearl River Delta, Jieshou, about 500 kilometers inland on the North China Plain, and Shangrao, in a valley to the north of the Wuyi mountain ranges. There is much disparity in their economic conditions, with Jieshou and Shangrao lagging far behind Wuxi and Nanhai (table 4-1). Table 4-2 shows some of the exogenous factors that influence the develop- ment and the relative strength of TVPS in each county. Wuxi and Nanhai are more favorably endowed in many respects than are Jieshou and Shangrao. Nanhai does not have as strong an industrial base as Wuxi, but it has greater access to capital and to foreign technology. Shangrao has some advantage over the other counties in mineral resources, but it is poorly endowed other- wise. The availability of skilled labor and of technical and management skills is linked to the level of local industrial development, and in this respect Wuxi and Nanhai are clearly better off than Jieshou and Shangrao. Although these exogenous conditions were important for the initial devel- opment of TVPS, county authorities have been extremely influential in foster- ing (or preventing) particular forms of TVP development. Some have been ardent supporters of TVCES, while others have promoted private enterprise or pursued a more balanced approach toward enterprise ownership. In some cases, mainly because of factors beyond their control, local authorities have been unable to provide any support to TVP development. In this chapter we undertake a descriptive analysis of the structure and functioning of the TVP sector in the four counties. Our goal is to identify the principal features that have influenced Tvp development and to provide a conceptual framework for interpreting the main findings. We describe the characteristics of the TVP sector in the four counties, identify the chief deter- minants of TVP development patterns, and present some implications of TVP experience. 63 64 JAN SVEJNAR AND JOSEPHINE WOO Table 4-1. County Data, 198.5 Item Wuxi Jieshou Nanhai Shangrao Population (millions) 1.02 0.57 0.85 0.70 Area (square kilometers) 960 667 1,153 2,490 Administrative divisions Townshipsa 36 33 16 31 Villages 586 433 236 343 Teams 9,518 3,981 3,328 3,782 County OVAIO (millions of yuan, 1980 prices) 4,680 257 2,541 211 County GVIO" 4,331 148 2,182 50 GVIO of TVPs' 3,650 104 1,518 27 Note: GVAIO, gross value of agricultural and industrial output; GVIO, gross value of industrial output. a. Includes township-level but not vitlage-level towns. b. Including the output of township, village, production team, and private firms. c. 1986 data for Nanhai; estimate for Jieshou. Source: Information from fieldwork. Table 4-2. Influences on TVP Development Favorable - > Less favorable Item 5 4 3 2 1 Local manufacturing industries WX NH JS SR Proximity to large cities NVX, NH JS, SR Natural resources Agricultural WX, NH SR JS Mineral SR WX, NH, JS Human resources Technical and management skills NVX NH JS SR Skilled labor WVX, NH JS SR Local labor supply (1 = shortage) VX, NH JS, SR Access to capital NH WX JS SR Access to foreign technology NIH WX JS SR Note: WX = Wuxi; NH = Nanhai; JS = Jieshou; SR = Shangrao. Characteristics of the TVJP Sector in the Sample Counties This section describes the main characteristics of TVPS that are relevant for our analysis: size, sectoral composition, industrial structure, ownership, growth, performance, and experience with foreign trade and investment. Development Patterns in Four Counties 65 Figure 4-1. The Relative Sizes of the TVP Sectors in the Four Sample Counties, 1985-86 Index (Wuxi = 100) 120 110 100 80 _ , ,-,',' 70 - ,'|,, 6@o0 -XWu Number Employment Gross income __ Wuxi CNanhai []Jlieshou B Shangrao Soure: Information from fieldwork. Size Figure 4-1 shows the relative size of the TVP sector in each county in 1985-86. In number of enterprises, employment, and gross income, the TVP sector in Jieshou and Shangrao is much smaller than in Wuxi and Nan- hai. In terms of industrial TVCES (figure 4-2) Wuxi is clearly superior to the other counties. Its TVPS contributed 84 percent of the total gross value of in- dustrial output (GVIO) of the county in 1985; in Nanhai and Jieshou the corresponding shares were about 70 percent, and in Shangrao the share was only 54 percent. The average size of TVPS, especially TVCES, is large in Wuxi and Nanhai compared with Jieshou and Shangrao (see chapter 5 for details). Sectoral Composition Industrial activities, as measured by TVP gross output value, are predominant in all four counties, but the service sector has a large number of firms, espe- cially in Nanhai and Shangrao. In Wuxi industrial TVPS accounted for 74 66 JAN SVEJNAR AND JOSEPHINE WOO Figure 4-2. The Relative Size of the Industral TVCE Sector in the Four Sample Counties, 1985-86 Index (Wuxi = 100) 100 80 -ui S5ahi0JehuBSaga 70gr:BMormanon Fom fieldwork. percent of 10 ~ ~ ~ ~ ~ ~ /1 value in 1984, whereas the service sector accounted for 22 percent of the number of enterprises and the same share of gross output value. In Nanhai industrial firms constituted on]y 40 percent of the total number of T'VPS in 1986, but they contributed 83 percent of total gross output value. By con- trast, services have a much larger share in the total number of TVPS (50 per- cent) than in gross output valuie (10 percent). In Jieshou in 1985 industrial TVPS accounted for a large share of both the number of firms (66 percent) and the value of gross output (83 percent). In Shangrao in 1986 the share of industry in the number of TVPS (50 percent) was highe-r than in Nanhai, but its share of the value of gross output of TVPS was only 57 percent. 'Frans- port and construction enterprises had a relatively large share in both the number of firms (21 percent) and the value of gross output (32 percent), and other services also had a significant share-30 percent of the number of firms and I11 percent of the value of gross output. Development Patterns in Four Counties 67 Industrial Structure The share of heavy industry in the GVIO of industrial TVCES is much higher in Wuxi (75 percent) than in Nanhai (44 percent). The machinery subsector is the largest in Wuxi, contributing 49 percent of total GviO, followed by textiles, which accounts for about 17 percent, and chemicals and construc- tion materials. In Nanhai the largest subsector is textiles (29 percent), fol- lowed by machinery (21 percent) and construction materials (11 percent)' In Shangrao the share of heavy industry is high (60 percent), but in contrast to Wuxi it is concentrated in the mining and construction materials subsectors (18 percent and 31 percent respectively), and the machinery subsector is insignificant (2 percent). There is hardly any manufacturing in- dustry in Shangrao, and most light industrial activities are in food processing (11 percent), leather processing (11 percent), and traditional handicrafts. In Jieshou the most important industrial subsectors are construction materials (52 percent, mainly bricks) and food processing (20 percent). Jieshou's ma- chinery and textile subsectors have higher shares than those in Shangrao. Products of TVPs range from handicrafts to rather sophisticated industrial equipment. In Jieshou, where the industrial base is weak, the main TVP activ- ities are brick manufacturing, traditional handicrafts such as making pottery and weaving bamboo and reeds, simple food processing such as manufactur- ing beef jerky and imitation protein food, and textile operations such as hand knitting. The skills for such operations are easy to acquire, and in some in- stances residents of an entire village or district specialize in making a single product. These types of operation have started to appear in Shangrao (for example, comb production in Huangshi township), but much of the nonagri- cultural labor force there is engaged in coal and other mining and in hydro- power generation. There is hardly any local manufacturing activity, and sur- plus labor goes outside the county to seek work in the service sectors, such as construction and repair. By contrast, many TVPS in Nanhai and Wuxi are engaged in production that requires a higher level of skill and technology. Some produce household appliances (such as electric fans) and electronic equipment (recorders and calculators) for both domestic and foreign markets. Many have contracts with state enterprises or, in the case of Nanhai, with foreign partners. In some areas, such as Wuxi, TVPS have become keen com- petitors with state enterprises. Some TVP machinery products in Wuxi (for instance, printing and dyeing equipment in Qianzhou) already have won a large share of the domestic market. Ownership Structure The ownership structure of the TVP sector in the four counties varies greatly (figure 4-3). In Wuxi TVcEs are predominant. they constituted 36 percent of the total number of industrial TVPs and contributed almost 96 percent of Figure 4-3. The Ownernip Structure of TVPs, 1985 (percent) By nunber of enterprises By gross output value 6.7 a 27.7 4. 4 4747 49.0 ~ 25 . 3 . 6 > 66.0 SHANGRAO / 18.5 m s 3 .8x88 3.3 15.4 NANHAI 31. 4.7 9.1 31.23. 10.0 JIESHOU 70 > 20 ll 20.7 SHANGRAO 1. O Township enterprises 3 Production team enterprises U Village enterprises 0 Individual enterprises E Partnerships a. Partnerships, 1.0 percent; production tearn enterprises, 1.6 percent; individual enter- prises, 1.8 percent. Source: Information from fieldwork. Development Patterns in Four Counties 69 GVIO in 1985. Development in Nanhai is more evenly spread among the vari- ous forms of TVPs, and private enterprises have a much larger share in the Tvp sector than do their counterparts in Wuxi. In both Jieshou and Shangrao the share of TVCES in the total number of TVPS is relatively small, although their share in gross income is large, especially in Shangrao. Individual propri- etorships in Jieshou have a larger share in the TVP sector than in the other three counties. No production team firms are recorded in the available data for Jieshou and Shangrao, which suggests that privatization of these enter- prises has occurred in these counties. Many production team enterprises in Nanhai are actually run like private enterprises, and some are contracted to individuals. (In addition, many private enterprises are registered as commu- nity enterprises in all four counties, especially in Jieshou; see chapter 7.) Growth In 1980-85 the gross income of TVCES in China grew at an annual rate of 25.1 percent. With the exception of Shangrao, the counties in our sample registered growth rates above the national average during the same period (see table 4-3). TVCES in both Wuxi and Nanhai showed continuous increases in employment and gross income, with gross income growing much faster than employment. In Jieshou development has been somewhat erratic: a de- cline in 1980-81, a sharp increase in 1984, and a drop in 1985. This may be partly attributable to the inconsistency of data over time, especially be- tween 1983 and 1984. In Shangrao data for earlier years are not available, but in 1984-86 both employment and gross income of TVCES grew steadily, the former faster than the latter. Private enterprises have also grown rapidly over the past few years. In Wuxi the number of individual enterprises increased sharply in 1983 and con- tinued to rise in 1984, but growth slowed in 1985-86. In Nanhai the number of individual enterprises more than doubled in 1982, almost quadrupled in 1983, and continued to grow at high rates through 1984-86. In Wuxi devel- opment of TVCES is emphasized at the expense of the private sector, whereas in Nanhai development of all forms of TVPS is encouraged. In Jieshou the number of individual proprietorships surged in 1981 and continued to grow at high rates in 1982-85. In Shangrao the number of individual enterprises increased by 15 percent in 1986 over 1985. In Wuxi the first surge in the growth of TVCES occurred in 1979-80; a slow- down followed in 1981-82. Another surge, in 1983, was followed by acceler- ated growth in 1984-85 and somewhat slower growth in 1986. In Nanhai the pattern was roughly the same. Development of the TVP sector in Jieshou and Shangrao did not take off until 1984. The first surge in the growth of the Tvp sector in Wuxi and Nanhai coincided with the initial changes that occurred after the Third Plenum of the Eleventh Central Committee en- dorsed, in December 1978, a program for accelerating growth in rural areas.' 70 JAN SVEJNAR AND JOSEPHINE WOO Table 4-3. Growth of TVCEs, 1979-86 (annual average percentage growth) All Item W'uxi Jieshou Nanhai Shangrao China By number of enterprises 1979 4.6 n.a. 59.0 n.a. -2.9 1980 11.2 -8.0 -9.6 n.a. -3.7 1981 -3.4 -22.4 16.6 n.a. -6.1 1982 -1.0 34.1 11.2 n.a. 1.8 1983 16.1 4.4 -4.9 n.a. -1.1 1984 35,9 240.3 n.a. n.a. 22.4 1985 22.2 -6.0 n.a. -3.2 -4.9 1986 11.4 33.9 -2.2 4.0 -3.8 1980-85 13.0 28.3 13.1 n.a. 1.9 By employment 1979 17.7 n.a. 4.2 n.a. 2.9 1980 11.1 3.8 6.4 n.a. 3.1 1981 7.6 -25.8 1.3 n.a. -1.0 1982 7.6 28.9 8.6 n.a. 4.8 1983 18.2 0.8 3.1 n.a. 3.9 1984 22.2 187.6 12.6 n.a. 19.0 1985 23.5 -17.3 12.2 37.0 7.9 1986 31.6 4.9 6.9 2.5 5.8 1980-85 15.6 18.1 7.5 n.a. 6.7 By gross income 1979 28.9 0.0 16.9 n.a. 13.8 1980 48.t6 0.0 22.9 n.a. 21.4 1981 14.t6 16.7 43.7 n.a. 12.5 1982 5.9 14.3 11.1 n.a. 15.1 1983 27.9 25.0 16.9 n.a. 20.3 1984 59.4 413.0 n.a. n.a. 36.6 1985 70.4 -12.1 n.a. 13.2 44.1 1986 17.2 n.a. 21.7 9.3 21.7 1980-85 33.3 49.7 36.4 n.a. 25.1 Sources: Estimates based on information from fieldwork; State Statistical Bureau (1982), pp. 194-95; (1983), pp. 206-07; (1984), pp. 184-85; (1985a), pp. 297-98; (1986a), p. 177; (1987a), pp. 226-27. Two aspects were particularly significant for the development of the TVP sec- tor: promotion of nonagricultural sideline activities to augment farm house- holds' incomes and the decision to raise procurement prices for agricultural products, which also helped increase rural incomes. These factors seem to have had an immediate positive impact on TvP development in Wuxi and Nanhai, but they did not have any such effect in Jieshou and Shangrao in 1979-81. In 1984, however, there was a surge in TVP development in all four counties after central government directives reiterated support for rural Development Patterns in Four Counties 71 reforms, stressing the need for diversified rural activities and specifically for the development of TVPS as a means of providing employment and raising the living standards of the rural population.2 Performance The overall performance of the TVP sector in Wuxi and Nanhai has been far better than in Jieshou and Shangrao (table 4-4). Labor productivity is similar in TVCES in Wuxi and Nanhai, but profit per worker is much higher in Wuxi. One reason may be that Nanhai firms are more capital-intensive, as indicated by higher gross fixed assets per worker. Data on enterprises in our sample (table 4-5), as well as aggregate county data, suggest that TVPS in Nanhai rely more on bank loans to finance their operations than do enter- prises in Wuxi, and interest payments may eat up some of the profits. More- over, wages are higher in Nanhai, which adds to production costs. Average production costs for township enterprises in Nanhai were 84 percent of sales revenue in 1985, compared with 65 percent in Wuxi. Both Nanhai and Wuxi are advanced counties in their respective provinces, with profit rates much higher than provincial averages. Their TVCEs also perform much better than the average for China. Gross income per worker for TVCES in Jieshou is only about 32 percent of, that in Nanhai, and profit per worker is about 35 percent. Jieshou enterprises are below the national average in labor productivity, but in return to capital they seem to do relatively well, even in comparison with Nanhai. In Shangrao gross income per worker is only 19 percent that in Nanhai and profit per worker a mere 10 percent; both are below national averages. In their respective provinces, enterprises in Jieshou perform better than the av- erage TVCE, and enterprises in Shangrao do worse. Data on sample enterprises suggest that TVPS in Jieshou are more capital-intensive than in Shangrao, and profitability is also much higher in Jieshou. Production cost as a percentage of sales revenue is similar in Jieshou and Shangrao and close to that in Nan- hai for the enterprises in our samples. Average wages are about the same in Jieshou and Shangrao. The Role of Foreign Trade and Investment Foreign trade and investment in the TVP sector are highly significant in Nan- hai (table 4-6). In Wuxi they are also increasingly important, but in Jieshou and Shangrao their role is small. In all four counties the contribution of TVPS to total county exports is large-about half of total export procurement vol- ume in Wuxi and Nanhai and even more in Jieshou and Shangrao. From the perspective of TVPS, however, the export market is far less impor- tant than the domestic market. In Wuxi total export procurement from TVPS in 1985 was probably about Y60 million,3 or less than 2 percent of their GVIO. 72 JAN SVEJNAR AND JOSEPHINE WOO Table 4-4. Performance of TVCEs, 1985 All Item Wuxi Jieshou Nanhai Shangrao China Gross income per worker (yuan)' 10,996 3,439 10,996 2,049 4,401 Gross fixed assets per worker (yuan) 2,496 968 3,396 1,001 1,807 Profits per worker (yuan)b 1,337 335 944 95 413 Ratio of gross income to fixed assets 4.4 3.6 3.2 2.0 2.4 Profits as a percentage of fixed assets 53.6 34.8 27.8 9.5 22.8 (Profits + taxes) - fixed assets (percent)' 80.1 48.2 49.0 15.4 37.3 a. Wuxi dara represents gross output for industrial rvcEs onLy. b. Total net profit after deduction of income tax. c. Total net profit plus sales and income taxes divided by undepreciated fixed assets valued at original purchase prices. Sources: State Statistical Bureau (1986a), p. 177, and information from fieldwork. Table 4-5. Performance of Sarnple Township Enterprises Item Wuxi Jieshou Nanha.i Shangrao Gross output per worker (yuan) 30,340 17,739 19,015 2,507 Gross fixed assets per worker (yuan) 5,633 3,811 5,526 2,147 Profit per worker (yuan) 4,783 1,450 1,934 293 Ratio of gross output to fixed assets 5.4 4.7 3.4 1.2 Profits as a percentage of fixed assets 84.9 38.1 35.0 13.7 (Profit + taxes) . fixed assets (percent) 108.6 66.2 66.1 23.0 Production cost sales revenue (percent) 65.2 81.7 86.8 84.0 Bank loans - total funds used (percent) 18.3 25.5 47.3 17.9 Average monthly income of workers (yuan) 102 59 174 62 Note: Includes town-run enterprises. There are twenty-three observations for Wuxi, twelve for Nanhai, seventeen for Jieshou, and twenty-six for Shangrao. Source: Enterprise Quantitative Questionnaire. The percentage was the same in Jieshou. In Shangrao the share of TVP ex- ports in the total gross income of TVPS was at most 3-4 percent. In Nanhai the export market is relatively more important; 1986 export procurement from TVPS totaled Y247 million, or about 11 percent of their total gross in- come. Export procurement as a share of total TVP output in Nanhai actually declined until 1986, when siow sales on the domestic market and tight do- Development Patterns in Four Counties 73 Table 4-6. Foreign Trade arnd Investment Wuxi, Jieshou, Nanhai, Shangrao, Item 1985 1985 1986 1986 Foreign trade procurement (millions of yuan)' 100 4.8 477 3.6 From TVPs 29 3.3 247 n.a.b Processing income (millions of U. S dollars)' n. a. n. a. 42 n. a. a. Value of procurement by foreign trade departments. The estimate for Wuxi includes pro- curement through channels other than the county foreign trade bureau (for example, Shanghai and Wuxi municipality); the figure for Wuxi TvPs represents only procurement from Tvps by the county foreign trade bureau. The figure for Jieshou TvPs represents output value rather than procurement value. For Shangrao the amount includes exports through Shanghai as well as through the foreign trade departments of Shangrao prefecture. b. Although the exact amount of procurement from TvPs is not available, information ob- tained in fieldwork indicates that the exports of Shangrao County depend largely on TvPs. c. Income received for processing material supplied by foreign partners. Source: Information from fieldwork. mestic credit forced firms to look outward for market channels as well as fi- nancial resources. Because of Nanhai's proximity to Hong Kong, the foreign sector has a much more important role than in the other three counties. In 1980-86 for- eign investment in the county totaled US$92 million, of which US$27 mil- lion went into joint ventures and US$64 million was for processing arrange- ments. When the central government imposed a tight credit policy in 1985-86, the county government encouraged village and team firms to seek foreign capital through compensation trade and through processing materials for foreign companies. As a result, total processing income increased from US$10.6 million in 1985 to US$42 million in 1986. Many TVPS, includ- ing some production team firms, partnerships, and individual enterprises, established direct foreign contacts through relatives or other social con- nections. In Wuxi there is some inflow of capital from other parts of the country, but the level of foreign investment has been much lower than in Nanhai. This is partly attributable to administrative restrictions against direct foreign contacts by TVPS. In Jieshou there are a few joint ventures that involve for- eign capital; in Shangrao there are none. The Factors That Determine Patterns of Development The size, structure, and performance of the TVP sector in each county depend greatly on geographic and historical factors, the availability and types of local resources, the quality and size of the local labor force, access to technology 74 JAN SVEJNAR AND JOSEPHINE WOO and capital, and the policies pursued by the central and especially the local govemment. Geography and History The size of the TVP sector is closely linked to the level of development of the local economy. Per capita indicators illustrate the disparity among the four counties (table 4-7). In Wuxi and Nanhai there are more market chan- nels, greater access to different types of resources (including physical inputs, financial resources, and technology), and larger demand for manufactured products and for a wide range of services. These counties can therefore make better use of opportunities that arise. This partly explains why the TVP sector in Wuxi was able to take off as early as 1979-80, soon after the central gov- ernment's policy shift toward diversified economic development in rural areas. To capture the conditions of the TVP sector in each county, we need a historical overview. Wuxi County is located in southern Jiangsu Province in the hinterlands of Shanghai. The area has been highly developed since the Song Dynasty (960-1279) and has many small cities and urban clusters. It is rich in agricul- ture and has a strong industrial base. In 1985 Shanghai city and Jiangsu Prov- ince together contributed 18 percent of China's gross value of agricultural and industrial output (GVAIO) and 21 percent of its GVIO. In 1985 Jiangsu led all other provinces in China in the GVIO of construction materials, food and beverages, textiles, leather products, fumiture, chemicals, plastics, and electronics. It also has an advanced machinery manufacturing industry and many art and handicraft products.4 Wuxi municipality, located in the middle Table 4-7. Per Capita Indicators, 1985 (yuan, except as specified) All Item Wuxi Jieshou Nanhai Shangrao China Cultivated land per capita (mu) 0.88 1.2 0.94 0.67 1.39 Average per capita income of rural populationa 754 285 1,029 322 398 GVAIO per capitab 4,588 452 2,989 300 1,164 GVIO per capitab 4,246 260 2,567 71 885 GVAO per capita' 342 192 422 229 279 Note: GVAO, gross value of agricultural output. a. Includes income from all sources, including nonagricultural activities and employment in TVPS. b. Includes the output of TVPs. c. Excludes the output of TVPs. Sources: Information from fieldwork; State Statistical Bureau (1986a), pp. 4, 132, 233, 583. Development Patterns in Four Counties 75 of Wuxi County and since 1983 the supervisory level of govemment for the county's administration, is an important city in the area. The county's strong agricultural and industrial base has helped foster devel- opment of the TVP sector. Rich agricultural produce supplied the raw materi- als for many TVP food-processing and textile industries. Advanced industrial development in the surrounding areas provided opportunities for TVPS to en- gage in processing work for large state or urban collective factories or to pick up products that these factories no longer considered profitable to produce. Moreover, the management and technical skills that had been built up in the area over the years could be drawn on to support the development of TVP industry. An important impetus for TVP development in Wuxi came from the cadres, workers, and educated youths who returned there during the Cultural Revo- lution, many bringing skills learned in their previous jobs in Shanghai and other cities. Prompted partly by the need to find work and partly by the short- age of supplies as many urban factories stopped production at the height of the Cultural Revolution, they established TVPS in rural areas. In the early 1970s, after an intense debate, county authorities decided not to abandon TVPS, even though in accordance with the spirit of the times they stressed the development of agriculture and grain production. By the late 1970s TVP industrial development in Wuxi was far ahead of that in the other counties. Nanhai County is located in a highly developed and urbanized area about 20 kilometers from Guangzhou. Like Wuxi, Nanhai stretches around a mu- nicipality (Foshan) that was given administrative leadership over the county in 1983. The county has several well-developed towns. The area's agriculture produces enough to support its own population and to export a substantial quantity of products to Hong Kong. There were many small factories in the area before 1949, including the first modern silk mill in China, but the indus- trial base is much smaller than in Wuxi. The GVIO (including output of vil- lage industry) of Guangdong Province was only 50 percent of Jiangsu's in 1985, and the output of its machinery subsector was only 42 percent. Light industry, especially food processing and textiles, is predominant. (The pro- portion of light and heavy industry in total ovio was 67:33 in Guangdong in 1985 and 56:44 in Jiangsu.) As in Wuxi, the rural areas benefited from the return during the Cultural Revolution of skilled workers who began to set up their own factories in the countryside. Although Nanhai authorities permitted these factories, they apparently took a less active role than their counterparts in Wuxi in fostering TVCES. Hence TVPS in Nanhai tend to be on a smaller scale. The county enterprise sector is weaker and the private sector is stronger than in Wuxi. Jieshou County is located on the North China Plain at the western border of Anhui Province, in a less favored area than Wuxi and Nanhai. The region is subject to periodic floods and droughts, with serious consequences for agri- 76 JAN SVEJNAR AND JOSEPH:INE WOO cultural output. In good years harvests are adequate for export to other re- gions; in bad years, as in 1985, the area has to be subsidized by grain allocated by the state, and there are severe reductions in personal incomes. (The aver- age per capita income of the rural population dropped from Y345 in 1984 to Y285 in 1985, mainly because of the bad harvest.) The area is much fur- ther inland than Wuxi and Nanhai and has hardly any industrial base. In 1984 Jieshou's GVIO (excluding village industry) was only 6 percent of Wuxi's and 10 percent of Nanhai's. Of the four counties, only Jieshou has its own county seat. During the war with Japan in the 1930s and 1940s Jieshou town, which was not under Japa- nese occupation, prospered as a commercial center. Although its prosperity diminished after the war, sorme commercial traditions have been retained. The area has one of the four important livestock markets in China. Local people engage extensively in marketing activities. They are willing to travel long distances to get supplies and sell products, going as far as Nepal, for instance, to sell locally produced nylon ropes. The development pattem of the TVP sector in Jieshou is somewhat similar to that in Wenzhou, a coastal prefecture in Fujian Province; there too, local resources are lacking, but mar- keting activities have brought prosperity to residents. The TVP sector in iieshou seems to have performed better than that of Shangrao County. Compared with Jieshou, Shangrao was historically more isolated, being hidden away iin a valley among mountain ridges. Although it is now easily accessible by railroad, it is not a commercial center as Jieshou once was, and it is not usually a place where people stop to do business. In 1964 Shangrao County was separated from Shangrao city under the principle, prevalent at the time, that the city should concentrate on industrial develop- ment and the county on agriculture. Unlike Wuxi or Foshan cities, however, Shangrao city had poorly developed industries that failed to exert leadership in the industrial development of the vicinity. The share of agricultural output (excluding village industry) in the total GVAIO of Shangrao County exceeded 75 percent in 1984 and 1985. Most of the county's industries are resource- based, tapping rich coal deposits and construction materials. Almost every township has a few hydropower stations, but many are running losses because of poor management. Some of the county's so-called TvPs are administrative establishments set up to manage labor teams that go outside the county to work on construction and repair jobs. Few commercial activities and services are carried out within the county, although a market for ramie fiber thrived in Shaxi town in 1984-86. Human Resources and Technology More than 60 percent of the rural labor force in both Wuxi and Nanhai now works in industry. To supplement local labor, both counties have hired work- ers from outside-about 40,000 in Nanhai and 10,000 in Wuxi. This repre- Development Patterns in Four Counties 77 sents a great achievement in both counties, where until the late 1970s and early 1980s an important incentive for TVP development had been to find jobs for local residents. By contrast, in neither J ieshou nor Shangrao are there enough industrial and other activities to absorb surplus labor from agricul- ture, and providing employment for the local population is an important con- sideration for county authorities. The employment problem seems greater for Shangrao's TVPS than for Jieshou's. Many of the former are overmanned, which adversely affects their performance and profitability. In Wuxi, county authorities prefer that TvPs become more capital-intensive rather than in- crease the number of workers hired from outside, since migrants from outside the county bring with them a number of social problems. Nanhai has a freer labor market and a larger number of workers from outside the county. As discussed above, TVPS in both Wuxi and Nanhai derived much of their initial technical strength from workers who returned to rural areas in the 1960s and 1970s. These people brought with them not only their technical and management skills but also their networks of social connections, which were essential in obtaining market information and other assistance as the TVP sector expanded in later years. The proximity of these two counties to the large cities of Shanghai and Guangzhou provided TVPS with opportunities to get technical assistance in various ways-for example, by offering short- or long-term employment to workers retiring from city factories or part-time work to technicians in the city and researchers in nearby academic institu- tions. Nanhai also has easier access to foreign technology, through joint ven- tures and processing arrangements, than does Wuxi. Neither Jieshou nor Shangrao has such advantages in access to tech- nical skills and production technology. Many of Jieshou's TVPS are house- holds and small firms that specialize in traditional handicrafts and simple food processing-activities requiring little technical skill. Local residents in Jieshou are rather entrepreneurial, however, and many have taken initiatives to develop new lines of products. They are willing to invest in learning new technology or in sending people to more advanced areas (such as southern J iangsu and Guangzhou) to be trained. They are also responsive to the mar- ket, switching quickly to other products as soon as the market becomes satu- rated with their old products. Through their connections they obtain infor- mation on new technologies and products. By contrast, people in Shangrao seem more risk averse. Their main indus- trial activities are tied to natural resources, and only recently have they started to develop handicraft industries. They are less attuned to the market than in J ieshou. Those who work outside the county engage mostly in manual work rather than in marketing. The county also suffers from a severe brain drain; many of its secondary school graduates who go on to college never return to work in Shangrao. In recent years TVPS in Shangrao began trying to obtain technology by striking up joint ventures or other arrangements with entrepreneurs or institutions in nearby Zhejiang Province and in southem 78 JAN SVEJNAR AND JOSEP1INE WOO Jiangsu. Unlike Jieshou, the county has no joint ventures involving foreign investment and not a single piece of imported equipment. Access to Capital The relatively advanced development and good performance of TVPS in Wuxi and Nanhai offer opportunities for higher returns to capital, and as a result outside financial institutions are willing to invest in them. For instance, in 1984 Wuxi was able to bring in Y40 million in loans from branches of the Agricultural Bank and the Industrial and Commercial Bank in other regions, and a Beijing-based development company came to the county to seek invest- ment opportunities. Moreover, higher per capita incomes lead to an increase in bank deposits, providing a larger base for loans. Some TVCES in Wuxi have issued bonds to finance their investments. Nanhai has the additional advantage of easier access to foreign capital. The sources of foreign exchange include compensation trade, processing ar- rangements, and joint ventures, as well as remittances from friends and rela- tives in Hong Kong and overseas. During 1980-86 the Bank of China lent US$7.8 million, mostly short: term, to TVPS in Nanhai. Personal incomes are higher in Nanhai than in Wuxi, partly because govemment policies in Wuxi have sought to minimize the difference between the incomes of agricultural and industrial labor. Overseas remittances and proximity to Hong Kong also raise incomes in Nanhai. Financial institutions are more flexible, and the banks may be more active in mobilizing funds than in Wuxi. In 1986 Nanhai obtained Y100 million in short-term interbank loans from financial institu- tions outside the county. Enterprises as well as individuals seem to make more use of banking services than in Wuxi. Personal savings per capita in 1984 were Y588 in Nanhai and Y154 in Wuxi. TvPs borrowed about Y900 million of bank loans in 1986, compared with Y574 million in Wuxi. In J ieshou and Shangrao the amount of loans available for TVPS is not small in relation to the size of their TVP sectors. In Jieshou many govemment de- partments, including the finance bureau, tax bureau, supply and marketing cooperative, and foreign trade department, assist TVPS in meeting their capi- tal requirements. In Shangrao in the past local banks were forced to lend to many unviable TVP projects, especially small hydropower stations. As a result, almost half of the value of outstanding loans in Shangrao in 1986 con- sisted of bad debts and overdue payments. Banks have become conservative in their lending policies and do not make maximum use of available funds. TVPS in all four counties complained about the shortage of funds. New loans for TVPS were reduced in 1985 as a result of central govemment efforts to control investment and consumption in the country as a whole. This somewhat dampened the growth of the TVP sector, but growth continued at a remarkably high rate in all four counties. In Nanhai the tightening of do- Development Patterns in Four Counties 79 mestic credit induced increased use of foreign capital by TVPS, which seems to have been beneficial to their development. Government Policies and Strategies As mentioned above, central government policies have had a significant im- pact on the growth of the TVP sector. With the Great Leap Forward in 1958 some beginnings were made in rural industrialization, but many of the invest- ments turned out to be wasteful. During the Cultural Revolution grain pro- duction was emphasized at the expense of other rural activities; some enter- prises, however, were set up in rural areas at the commune and brigade levels, mainly to support agricultural development by providing machinery and tools. Rapid development of the TVP sector occurred after 1978, at first some- what sporadically as a sequel of rural reforms but later with the encourage- ment of the central government. Rural reforms affected TVP development in several ways. The PRs allowed farmers to engage in other activities as long as they fulfilled the terms of their contracts; the reduction of grain procurement quotas relaxed the need for labor-intensive triple cropping and so freed labor for other activities; the in- crease in agricultural output, especially of cash crops, provided raw materials for TVPS; and the increase in rural incomes helped finance the growth of TvPs. For these reasons the timing of rural reforms in each locality influenced local TVP development. The first of the four counties to adopt the PRS was Jieshou, in 1981. (Pro- vincial leaders in Anhui Province had endorsed the PRS in the late 1970s and supported its implementation through the initial debates on these issues; see Wu Xiang 1986.) In 1982 individual firms and private partnerships started to develop rapidly. In Wuxi and Nanhai widespread implementation of the pRs began only in 1982, after the system was officially endorsed by the central government.5 In Wuxi, however, the gradual reduction of triple cropping after 1978 had provided an impetus for the rapid development of the TVCE industrial sector. Partly because of the success of industrial TVCES, local au- thorities were reluctant to encourage development of private enterprises that might give community enterprises too much competition. In Nanhai the PRS was initially implemented only in the poorer townships and villages, but starting in 1979 the local government encouraged large-scale development of brigade and production team enterprises. In Shangrao local authorities seem to have been more sluggish than their counterparts in J ieshou in imple- menting rural reforms. The PRS was not adopted until 1983, and not much effort was made to develop the TVP sector at any level until 1984, when the central government unequivocally promoted TVP development. In the other three counties the central government's clear enunciation of policy in 1984 further boosted the TVP development that had been initiated in earlier years. 80 JAN SVEJNAR AND JOSEPHIINE WOO The ownership structure of the TVP sector in each county is to a large ex- tent determined by strategies adopted by the county govemment. (See chap- ters 7 and 9 for further discussion of ownership patterns and the factors that have shaped them in the four counties.) In Wuxi the community-run econ- omy has a strong base. The government has assumed firm leadership of the TVP sector and is stressing the development of industrial TVCES. Its strategy is to discourage development of TvPs below the village level (to conserve raw materials and energy) and to protect TVCEs as much as possible from competi- tion for human and financial resources. Private enterprises are tolerated, but their development has been constrained by limits on loans, restricted access to inputs, and environmental and other regulations. Ten percent of the prof- its remitted by TVPS is used to finance agricultural investment and support services. The govemment tries to keep the difference between wages of work- ers in TVPS and in agriculture relatively small, and it imposes wage controls on TVCES. Hence the Wuxi lVP sector is much stronger at the township and village levels than at lower levels. In Nanhai, where the government encourages the development of TVPS at all levels, TVCES are much less predominant. During the past couple of years the government began emphasizing the development of core enterprises at the township level but not at the expense of other types of TVP. Restric- tions on loans to private enterprises have been less stringent than in Wuxi, and when macroeconomic control over domestic credit was imposed in 1985, the government encouraged production team firms and even (to a lesser ex- tent) private enterprises to seek financial resources through foreign channels. These measures boosted the development of lower-level TVPS, while the growth of TVCES was somewhat curtailed by the tightening of credit. In Jieshou the government has actively promoted the development of indi- vidual enterprises by helping them to get loans and inputs and by a[llowing them to use the name of the community (the township, town, or county). Some individual firms were given loans directly from township govemment funds. Regulation of individual enterprises is lax; there is no ceiling on the number of workers hired by private enterprises or on wages paid to workers. Hence this sector is relatively large. In Shangrao the government did not get organized to promote Tv]P devel- opment until 1984-85. There is no obvious indication of government favor- itism toward any one form of ,rvp. Because the township governments in gen- eral depend on TVPS to finance their expenditures, they tend to squeeze TVCES for funds, which harms the enterprises' financial viability and long-term de- velopment potential (see chapter 17). From the foregoing discussion it is obvious that the policies of county and community governments have an immense impact on the performance of the TVP sector. Government requirements for remitted profits reduce the funds available to the enterprises for expansion and harm incentives to workers and managers. In this respect Jieshou is consistently much less demanding than Shangrao, as data on firms suggest (table 4-8). Development Patterns in Four Counties 81 Local government can also exert considerable influence over the perfor- mance of TVPS through labor policy. Wuxi's govemment, for instance, has a stronger role in the allocation of labor within the county than Nanhai's, and Nanhai allows more workers from outside the county to work in local TVPS than does Wuxi. Wuxi also imposes stricter constraints on wage levels than does Nanhai, which has helped to keep wage costs for TVPS low and to maintain a high level of profits. (See chapter 13 for a discussion of wages and incentives and chapter 7 for details on the factors that affect local gov- emment policies on wages and distribution.) In Shangrao the result of gov- emment interference in employment and personnel appointments in TvPs has been enterprises that are overstaffed and, in some cases, financially unviable. Implications of TVP Experience in the Four Counties Many factors account for the diversity of TVP development and performance in the four counties. The initial conditions are clearly important and account Table 4-8. Distribution of Profits in Sample Township Enterprises, 1981 and 1985 Item Wuxi Jieshou Nanhai Shangrao 1981 Total profits (thousands of yuan) 847 35 190 15 Of which (percent): Income tax 10.0 20.0 6.8 26.7 Remitted profits- 55.0 0.0 29.5 40.0 Retained profitsb 35.0 80.0 63.7 33.3 Use of retained profits (percent) Investment and working capital 98.8 81.8 78.6 88.9 Bonuses, collective welfare, and so on 1.2 18.2 21.4 11.1 1985 Total profits (thousands of yuan) 1,942 161 787 22 Of which (percent): Income tax 20.7 18.6 32.9 18.2 Remitted profits, 27.7 1.9 15.6 27.3 Retained profitsb 51.6 79.5 51.5 54.5 Use of retained profits (percent) Investment and working capital 87.2 74.2 70.0 91.7 Bonuses, collective welfare, and so on 12.8 25.8 30.0 8.3 Note: Includes town-run enterprises. There are twenty-three observations for Wuxi, twelve for Nanhai, seventeen for Jieshou, and twenty-six for Shangrao. a. Profits remitted to township govemments. b. Total profit less income tax and remitted profits. Source: Enterprise Quantitative Questionnaire. 82 JAN SVE]NAR AND JOSEPHINE WOO for much of the observed difference between the more and less advanced counties. Although such factors as geographic location and natural resources are beyond anyone's control, initial conditions are also significantly affected by endogenous factors-for example, government policies. Hence initial conditions should not be treated as entirely exogenous factors that act sepa- rately from the other determinants of TVP structure and performance. Keep- ing this in mind, and realizing the danger inherent in attempting to gen- eralize from only four cases, we outline four different patterns of TVP development that are exemplified by the four counties in our sample. General Patterns Wuxi represents one extreme of community government involvement. Local authorities have stressed the clevelopment of TVCES: they have supplied initial funds, shared the risk, and served as a supervisory body, appointing managers and allocating workers. Partly because of the favorable conditions discussed earlier (proximity to Shanghai, the presence of skilled labor, an industrial tradition, and so on), the Wuxi system has worked well. The community governments have ensured thLat managerial incentives maximize the welfare of the local community, but they have in most cases left managers consider- able leeway in making production decisions. Workers enjoy job security but have little say in decisionmaking, and their earnings largely consist of piece rates and profit-related bonuses. Although employment generation was the principal goal of Wuxi TVPS in the late 1970s and early 1980s, the elimination of the labor surplus and the advent of competitive pressures have shifted pri- orities toward greater capital intensity and modernization. Workers from out- side the county are admitted on a temporary basis only. The community gov- ernment's leadership role includes providing assistance in product selection, marketing, and financing. Financial institutions are important but regulated, and Wuxi TvPs retain a signilicant part of their profits for reinvestment. In Nanhai local authorities have been much more even-handed in their treatment of different forms of TVPS. Other characteristics of the Nanhai pat- tern are a developed banking sector that TVPS use as their primary source of capital, reliance on markets (including the labor market), and significant ex- port orientation with concomitant transfer of foreign technology. Although the government has emphasized development of all forms of TVPS, it is not as directly involved in the operation of these enterprises as in Wuxi. TVPs use a smaller fraction of their profits for investment and rely heavily on banks, which appear to be less regulated than in other parts of China. Through their contacts with foreign (mostly Hong Kong) firms, Tvps have been able to obtain advanced technology and to modernize their production methods. High productivity combined with significant remittances from abroad have contributed to Nanhai's high per capita income. Since Nanhai, Development Patterns in Four Counties 83 like Wuxi, has exhausted its surplus labor, income maximization based on more capital-intensive modem technology has become the principal goal of most TVPS. TVP managers are selected primarily by the community govem- ment, but they are given considerable decisionmaking authority. Workers have little say. Layoffs are rare, and most workers are on piece rates or other forms of incentive pay. Workers from other areas are used extensively by Nanhai TVPS. In Jieshou local authorities have systematically supported the development of large private enterprises, which are important not only numerically but also in employment and output. The govemment is more supportive of entre- preneurship than in Shangrao and encourages any form of TVP development, including joint ventures that involve foreign capital. Less than half of all en- terprise directors are appointed by the local government. TVPs are responsive to changes in the market and, because of the simple nature of their opera- tions, can switch easily to different lines of production. In contrast to Wuxi and Nanhai, employment creation continues to be an important goal, since surplus labor still exists. Despite the obstacles posed by scarcity of capital, of infrastructure, and of skilled labor, Tvp development in Jieshou has been successful in recent years. Shangrao suffers from its remote location and limited industrial base, and community governments there have often been forced to use the TvPs as sources of revenue for public and social expenditures. TVCEs have struggled to survive, while private enterprises have assumed a large share of total TVP output and employment. Community governments control the appointment of community enterprise directors and determine the annual plan in many rvPs. Generating employment to absorb surplus labor remains an important objective, and brain drain has been a significant problem. As in other coun- ties, workers' pay contains an element of individual incentives, mostly piece rates. Shangrao's TVP development started much later than in the other three counties, but county authorities subsequently promoted TVPs and growth rates have been higher since 1984. Improvement in TVP performance may follow the enterprise reforms and restructuring that the county government scheduled for implementation in 1987. Rich resources of coal and other min- erals have supported the bulk of Tvp activities in the past, but future develop- ment, especially of manufacturing industries, will require innovative ap- proaches to identify new products with market potential and to develop human resources. Implications TvPs have been able to thrive in a variety of environments. In contrast to state enterprises, TvPS are characterized by great flexibility, harder budget constraints, costs of capital that are more reflective of scarcity, and much 84 JAN SVEJNAR AND JOSEPHINE WOO heavier reliance on worker incentives. TVPS have clearly demonstrated their ability to grow rapidly under different pattems of local government interven- tion. Their ability to maintain growth under the restrictive macroeconomic policies of the mid-1980s attests to their viability and superior performance. Our study also suggests that (except for Shangrao) no particular model is clearly superior to others. The Wuxi case, based on TVCES, is impressive, but the extraordinary growth record of Jieshou's TVPS, which include a significant private sector, is also remarkable, and the Nanhai approach of balanced own- ership combined with market and export orientation has also met with suc- cess. The experience of all three counties has drawn significant national at- tention. Wuxi and Nanhai counties and Fuyang Prefecture, where Jieshou is located, are among the fourteen rural counties and districts selected re- cently by the State Council for further experiments in various aspects of rural development. The experience of the four counties indicates that the principal obstacles to the development of TvPs are negative govemment policies, a weak initial economic base, and inability to compete for material inputs and for physical and human capital. Overall, our study suggests that an important stimulus for the TVP sector will be policies that improve the designed functioning of the capital market and promote the development of infrastructure and the acquisition of human capital. Notes 1. "Resolution on Several Questions about Speeding up Agricultural Development" (draft), December 1978. 2. Central Committee Circular on Agricultural Work in 1984, January 1, 1984 (Xinhua Yuebao, June 1984, pp. 82-86), and circular transmitting the "Report on Exploring New Areas for Development of Commune and Brigade Enterprises," submitted by the Ministry ol Agricul- ture and the Party group of the ministry (Xinhua Yuebao, March 1984, pp. 107-11). These direc- tives are also referred to as Documernts No. 1 and 4. 3. Procurement by the county foreign trade bureau was Y29 million; in addition, about Y20 million of TVP exports went through wuxi municipality and another Y14 million through Shang- hai and other cities. 4. Jiangsu is close behind Shanghai in the output of the machinery industry; together they contributed 25 percent of the GvIo of the subsector in China in 1985. Jiangsu is a close runner-up to Guangdong in the output value of art and handicraft products; the two provinces contributed 28 percent of total national subsectoral GVIO in 1985. 5. "Minutes on the National Conference on Agricultural Work," Document No. I (1982), approved the PRS as an integral part of a socialist economy. 5 Market Interactions and Industrial Structure William A. Byrd and N. Zhu China's TVP sector grew rapidly in the 1970s and 1980s and is now a signifi- cant force in Chinese industry. The sector accounted for 21 percent of na- tional gross value of industrial output (GVIO) in 1986, compared with 9 per- cent in 1978 and only 3 percent in 1971. TVPS produce a wide range of industrial goods and have penetrated most of China's domestic markets. They compete not only with each other but also with state enterprises. TVPS have grown extremely rapidly when the economic and political environment per- mitted, most notably in 1984-85, but they have also shown considerable re- silience in the face of adversity. They form the most dynamic and fastest growing part of Chinese industry. This raises a number of questions about the sector's performance, viability, and future. Is its recent growth an aberra- tion? What will happen if state enterprises become more competitive? What are the characteristics of the markets that TVPS participate in (and to a great extent have shaped since the late 1970s)? Is the scale of most TVPS sub- optimal? Is the financial structure of TVPS sound, or are many of them en- gaging in risky, unsustainable, debt-financed expansion? More fundamen- tally, is the TVP sector efficient? This chapter explores a number of related topics concerning market envi- ronment, enterprise response and adjustment, and industrial structure in the TVP sector. We start by briefly describing some basic features of the institu- tional structure and market environment of TVPS, looking at patterns of com- petition and at how firms, communities, and TVP industries adjust to chang- ing business conditions, increasing competition, and adverse developments. We then move on to a preliminary analysis of industrial structure in the TVP sector and a brief exploration of policy issues. Stylized Facts China's TVPs operate primarily in a market environment for outputs and material inputs. Table 5-1 shows the share of outputs and inputs of sample enterprises 85 86 WILLIAM A. BYRD AND N. ZHt1 Table 5-1. Planned Allocation of Sample Firms' Inputs and Outputs, 1985 (percentage of sample or subsample) Share of planned alocation in total a Item 0 0-10 10-25 25-50 50-100 Total sarnple Input 63 13 4 11 9 Output 5 1 8 5 8 28 Input, by county Wuxi 52 16 3 19 10 Jieshou 61 17 0 14 8 Nanhai 90 0 5 0 5 Shangrao 55 15 10 5 15 Output, by county Wuxi 46 7 7 10 30 Jieshou 57 11 6 6 20 Nanhai 74 5 0 0 21 Shangrao 19 6 6 19 50 Input, by ownership Township and town 58 15 3 12 12 Village and production team 65 5 5 15 10 Private 75 19 0 6 0 Output, by ownership Township and town 40 9 8 9 34 Village and production team 80 0 0 0 20 Private 63 13 0 12 12 a. Shares are percentages of firms' annual output value targets specified in plans handed down by village, township, county, and higher-level governments and of principal raw materials allo- cated by these levels of government. In many cases, the reported plan "target" is not mandatory. Inputs allocated by governments often are priced above the controlled state price. Source: Enterprise Survey Questionnaire. allocated by government plans in 1985. FDr most the share is zero, although there is some variation across counties and forms of ownership. But what sample firms refer to as "planning" more often than not consists merely of reference targets or orders from state commercial units-quite different from government directive planning. In Wuxi the share of mandatory planning by state institutions in the allocation of output of TVCES is very small, and moreover it declined from 11 percent in 1984 to 7 percent in 1985. Unlike many state enterprises, rvps buy and sell goods primarily for money rather than through barter.' Of 113 responding sample firms, only 8 reported getting any raw materials through barter-like exchanges, and only 7 obtained any electric power in this way. Thus the money economy is well established for Market Interactions and Industrial Structure 87 the bulk of TVPS. The only partial exception to this pattern is electricity (table 5-2). Most of the power for firms in our samples comes from county government allocations, but self-generation and purchase on the market at high prices are also important. Hence the market is becoming significant in the allocation of electric power. 7VPs, particularly community enterprises, have generally faced rigidities in ob- taining the factors of production-land, capital, and, in many areas, labor. For 43 percent of the township enterprises in our sample, local govemment grants were the most important source of capital at founding. Similarly, 63 percent of sample township enterprises listed the community government as their primary source of labor, through allocation or recruitment. Sample pri- vate firms relied very little on the community government for either capital or labor at founding. Most TVPS of any size borrow from local banking institu- tions, which are subject to community government influence. Table 5-3 shows the sources of land for sample enterprises (also at the time of found- ing). Except in Jieshou, where the market for land is fairly well developed, the bulk of sample firms relied primarily on the community government. Fac- tor markets are not well developed in the TVP sector, as is true for Chinese industry as a whole (see Byrd and Tidrick 1987), and there is not much movement of factors among rural localities and communities. Most TVPs of substantial size are owned and controlled by community govern- ments at the township and village level. The community governments often play a critical role in strategic decisionmaking on investment, new product devel- opment, breakup and merger, and so on. They also take the lead in establish- ing new community enterprises, and they pool funds and absorb risk across subordinate enterprises. In fact, community governments can be compared with diversified holding companies or, when tightly integrated as in Wuxi, with headquarters of multidivisional corporations. Most TVPs are enmeshed in the rural communities to which they belong. Com- munity enterprises belong to community governments and to a large extent Table 5-2. Sources of Electric Power for Sample Firms, 1985 (percentage of total observations) Share of ttal power supply (percent) Item 0-25 25-50 50-75 75-100 County government allocations 22 4 9 65 Exchange 95 4 1 0 Purchase at high market prices 70 6 4 20 Self-generation 54 21 16 9 Other 93 0 3 4 Source: Enterprise Survey Questionnaire. 88 WILLIAM A. BYRD AND N. ZHU Table 5-3. Sources of Land (points) Community govemment Item interventione Purchaseb Rental Other Entire sample 59 26 12 3 By owsnership Township and town 65 26 5 4 Village and production team 59 19 18 4 Private 21 40 39 0 By county Wuxi 75 13 8 4 Nanhai 56 23 21 0 Shangrao 79 14 4 3 Jieshou 29 52 14 5 Jieshou Township and town 39 53 4 4 Village and production team 17 50 17 16 Private 19 51 30 0 Nanhai Township and town 59 34 7 0 Village and production team 68 12 20 0 Private 84 16 0 0 Note: A point-ranking system based on enterprises' responses was used: 3 points for the most important source of land, 2 points fDr the second most important, and 1 point for the least important. The ratings were normalized to add to 100. a. Land acquisition in which the community govemment served as intermediary; use of land without charge; and land acquisition linked to hiring of local residents. b. Includes conversion of land contributed into shares in the enterprise. Source: Enterprise Survey Questionnaire. further their objectives. But even private enterprises are tied to communities in many ways, and in most cases they would find it difficult to relocate. Virtu- ally all directors of sample firms felt that the advantages of their location outweighed the disadvantages and that the main benefits were adminis- trative protection by the community government and ease of arranging land acquisition. Most TXVPs are relatively small. In 1985 the average industrial township en- terprise had fewer than sixty employees and a Gvio of slightly more than Y300,000 (see table 5-5, below), while industrial village enterprises were only about one-third as large. Untrammeled development of TVPs has been going on for only a short time, and administrative restrictions have hindered their expansion. The dependence of the bulk of TvPs on their own retained funds and on funds raised within the township to meet investment needs has also limited their size. Market Interactions and Industrial Structure 89 Enterprise Structure Elements of enterprise structure include ownership, financial structure, and product orientation and structure. OWNERSHIP. (See chapter 9 for an analysis of TVP ownership patterns and trends.) Traditional community enterprises have a relatively clear ownership structure. They are essentially ministate enterprises, but the "states" to which they belong are community governments, which supervise the enterprises through their industrial corporations. At the opposite extreme, small individ- ual firms also have well-defined ownership. In between are some of the larger "private" enterprises such as those in Jieshou, which may be registered as col- lective firms and have close personal ties with local government officials. Many larger private enterprises in Nanhai also have close govemment ties; a few have cooperative forms of ownership. In many private enterprises the capital contributed by enterprise founders is repaid to them fairly early, leav- ing firms with a growing net worth to which no individual has a clear claim. The many community enterprises that have been contracted to individual management add a further layer of ambiguity. FINANCIAL STRUCTURE. Nearly all TvPs have much higher leverage (debt- equity ratio) than state enterprises. A study of a sample of 200 large indus- trial TVPS (Zhou Qiren and Huang Zhuangjun 1987, table 4, p. 48) found that their ratio of total debt to total capital was 0.57 at the end of 1985; the debt-equity ratio was 1.31. Many firms in our sample also have relatively high leverage. Although this might not be considered unusual in other countries, it is far higher than in Chinese state enterprises, most of which got all of their founding capital as state grants. Reliance on internal accu- mulation by TVPS is also substantial, given the initially high profitability of many TVP activities and tax concessions for TVPS in the past. A certain pro- portion of the after-tax profits of community enterprises is siphoned off for community expenditures or development of other TVPS. During the startup phase, however, all after-tax profits can be left with the enterprise for rein- vestment. At the community level the industrial corporations responsible for su- pervising community enterprises function in many ways like financial con- glomerates, holding companies, or the headquarters of loosely managed multidivisional corporations. The key financial roles of the industrial corpo- rations include pooling enterprises' after-tax profits for investment and direct- ing other resources (from local banks and credit cooperatives) to particular investment projects; cushioning subordinate enterprises from short-term fluc- tuations; serving as a short-term financial intermediary by transferring funds from enterprises with surpluses to those with deficits; and facilitating the issu- ance of short-term bonds to local residents. 90 WILLIAM A. BYRD AND N. ZHU PRODUCT ORIENTATION AND STRUCTURE. Traditional community enter- prises tend to focus on a single main product or a closely related set of prod- ucts. Diversification into new product lines often occurs, but it typically leads to discontinuing the old product line or to splitting into several factories. This multiplication of firms is facilitated by the transparency of ownership; any new firms belong to the community government that owns the parent. Private enterprises are less likely to split in this way, and more diversified product lines may emerge over time. For example, a highly entrepreneurial private enterprise in Jieshou started out processing oilseeds, moved into pro- cessing soybeans, then began producing soybean-processing machinery, and now plans to ship frozen chickens to Shanghai. Among rural communities there are examples of both "conglomerate" and "specialized" product structures, as well as cases in which a single dominant firm accounts for a large share of GVIO in the community. The conglomerate pattern is the more common one, but a weaker form of specialization by broad subsectors seems to occur at the township and county levels-machinery pro- duction in Wuxi, which accounted for 49 percent of the total GVIO of TVCES there in 1984, is a good exarmple. Regardless of the precise mix of products, community governments do not set up enterprises that produce the same products and compete directly with each other. Each community enterprise is thus a monopolist within the jurisdiction of its "owner." This is not a seri- ous hindrance to competition, however, since communities are so small that they do not provide viable captive markets for more than a handful of TVP activities. Enterprises belonging to neighboring community governments as well as those belonging to villages nested within townships may compete vig- orously, as the flood of entrants into the soft drink industry in Wuxi illus- trates. Where private enterprises are encouraged, local competition can be even fiercer. The Market Environment The goods markets that TVPS face have only just emerged, largely since the late 1970s. Hence there has been instability and wide, sudden fluctuations, as well as numerous opportunities to make large profits. The state sector left many gaps where latent unfulfilled demand existed. The sharp rise in rural and urban personal incomes starting in the late 1970s also created new de- mand which state enterprises were, at least initially, not well placed to meet. The progressive liberalization of restrictions on TVPS also helped expand mar- ket opportunities. Finally, TV]?s may have an advantage in competing with state enterprises because they are allowed to incur much larger selling ex- penses and to employ salesmen on a commission or consignment basis.2 As the TVP sector expanded, existing market niches became occupied, and competition became increasingly heated. Many industrial product markets in which TVPs dominate or are important have thus been characterized by falling Market Interactions and Industrial Structure 91 or, at most, stable prices and declining profit margins. TVPS that entered the market early, when profit margins were high, could survive and prosper even when margins became narrower, but more recent entrants face greater diffi- culties. The increasing size and changing structure of the TVP sector itself has created numerous opportunities and market niches for enterprising firms, and continued increases in urban and rural personal incomes and in invest- ments generate additional demand. But the new opportunities are often used to best advantage by the more advanced areas and by well-established firms. The markets in which TVPS in the backward areas can most easily get a foot- hold are the "old" ones, where profit margins are now narrow. Although these activities provide employment, the long-term benefit to local TVP de- velopment is much weaker than for the first group of entrants. TVPS face a relatively "hard" and often highly competitive market environ- ment that puts great pressure on them to perform well and is relatively unfor- giving of mistakes. This environment undoubtedly contributes greatly to the efficiency, dynamism, and flexibility of the TVP sector. Unlike many state enterprises, TVPS have no captive markets for their products. In 1984-85 only 4 percent of the gross sales of all TVCES in Wuxi were within the home town- ship. Eighty-two percent of sample firms in Wuxi sell less than 20 percent of their output within the county, and 73 percent of those in Jieshou sell less than 40 percent of their output within the county. Large majorities of sample enterprises in all four counties sell at least 40 percent of their output outside the home province. Thus TVPs are largely outward-oriented in their product markets. More generally, given the lax control over TVP output pric- ing, they are much less likely to face chronic sellers' markets than are state enterprises. The bulk of the output of sample firms is priced by the market, through negotiations with customers, or by the producer. Sellers' markets are similar to captive markets in their impact on producers' incentives and be- havior, and they negate many of the benefits of the market mechanism. Hence the weakness of price controls and the near absence of chronic sellers' markets sharply differentiates the market environment faced by most TVPS from that of state enterprises. The TVP sector's market environment has been unstable and subject to fluctuations because of economic and administrative shocks. This is only nat- ural, since market adjustments had been weak or nonexistent for several dec- ades. Ease of entry, combined with strong pressures in rural communities to develop TvPs as a means of employing surplus labor and generating revenues, can also be a source of fluctuations (see chapter 17). The huge reserve army of rural labor available for relatively simple, labor-intensive activities means that violent downward movements in prices and profits are possible. The weakness of market information networks and the lack of connections be- tween local markets tend to exacerbate fluctuations. Administrative shocks also can engender severe market swings: the national tight credit policy imposed in early 1986 probably had a disproportionately large impact on 92 WILLIAM A. BYRD AND N. ZHU TVPS, both directly and indirectly as state enterprises, facing a credit crunch of their own, cut back severely on subcontracting work they had been, giving TVPS. Patterns of Competition and Adjustment Price competition is intense in many TVP industries, and various forms of nonprice competition occur as well. Brand names are often important in competition-electric fans are an example-and so are product quality and development of new products. Large majorities of directors of sample firms in all four counties considered that they faced "relatively fierce," "very fierce," or "extremely fierce" competition. There are several distinct patterns of competition. Imitative Competition and Neoclassical Adjustment In Jieshou, where products are for the most part relatively simple and enter- prises are rather small, the dominant pattem can be termed "imitative com- petition." A resourceful entrepreneur finds out about a new product or pro- cess that has been successful elsewhere and eams high profits at first. But since entry is easy, a host of irnitators soon springs up, and prices and profits decline sharply. Commonly, the entry of imitative competitors overshoots the appro- priate number of members oif the industry, and after a certain point wide- spread exits and a shakedown may occur. Exit is relatively painless because capital requirements are low and assets can easily be shifted to other lines of business. This pattern involves price and quantity adjustments and entry and exit of firms but little if any innovation in products or technological improvement after the initial entrepreneurial success that opened up the opportunity. The best label for it is "neoclassical adjustment," since it works very much like the neoclassical model of market equilibrium. Price adjustments are fast and often large. Imitative competition ensures that economic profits will be eaten away over time, and the situation tends toward equilibrium. No product market in Jieshou fits this stylized description exactly, but "soy- bean meat" [danbairoul and the machinery to produce it [penbaojil come close (see below). Similar adjustment patterns may be emerging for meat and fruit canning. In many backward areas the brickmaking industry has already set- tled into equilibrium. A CASE HISTORY. The penbaoji (PBJ) story is one of backward integration into production of the machinery for making soybean meat, a product that was in great demand all over China in 1982. It affords an interesting case study of how limited market and technological information can permit large initial profits that are rapidly whittled away as information spreads through Market Interactions and Industrial Structure 93 the market and numerous imitators arise to take advantage of the profit opportunity. The soybean-processing technique was developed in Changchun, the capi- tal of Jilin Province in northeast China, by a professor at the Jilin Provincial Finance and Trade Institute, who designed the PBj. Several state enterprises in Changchun began to make the PBJ, but the technology was not released to other machinery producers. The PBJ is simple, however, and any good ma- chine shop can copy it if it has a specimen. In Jieshou a local entrepreneur set up a small enterprise with thirty-five workers to produce soybean meat. The firm bought first one and then seven more PBJS from J ilin and made huge profits in 1982. Competition heated up as other enterprises entered the in- dustry in 1983 and 1984. Meanwhile, an enterprising private food-processing firm, which had been one of the new entrants into the soybean meat industry, surreptitiously trans- formed itself into a machinery producer and began to produce the PBj-a case study of technology transfer and adaptation that cannot be told in full here. It did this without any help from the original purchaser of the machines, which would not let other local factories see its installed equipment for fear that it would be copied. (Producers of soybean meat even tried to prevent other local firms from finding out the name and address of the producer of the PBJ in Jilin.) A state-owned farm machinery plant in Jieshou also got into production of PBJS and became the largest local producer. As technical information on the PBJ spread widely in Jieshou, a dozen or so small machinery plants gained the capability to produce it. By 1985 the market was becoming saturated, and the largest producer cut output sharply in 1986. Many small machine shops simply stopped producing the PBj. Price trends clearly show this sharp increase in competition and squeezing of profits. The Jilin producer's original price for a PBJ unit was Y9,500. The central mechanism could be bought for about Y1,400, and the rest of the machine could be made by the user (with some assistance from a machine shop) for a total cost of about Y4,000, less than half the Jilin price. When Jieshou firms began making the PBJ in late 1982 and early 1983, the initial sales price was Y7,500, and profits were huge because the production cost was only about Y2,200. (Machinery plants presumably produced the ancillary parts and covering for the PBJ much more cheaply than most users could.) By mid-to-late 1983 the sales price had fallen to Y5,500. In 1984 the original producer brought out a simplified version that was priced at Y2,800-Y3,500 but could be produced for only Y1,500. In 1985 the price again fell, to about Y2,300 for the simplified version. In 1986 the sales price was Y2,000, but since production costs had risen to Y1,700, profit margins were relatively narrow. SOME PROBLEMS. The shortcomings of neoclassical adjustment are not in the mode of competition itself. Adjustments clearly go in the right di, 94 WILLIAM A. BYRD AND N. ZHU rection-production rises sharply, and consumers or users benefit firom the price declines. But where neoclassical adjustment alone is occurring, there may be a certain weakness and lack of dynamism. Market niches are crowded, and margins are squeezed so much that little if any economic profit can be earned, but no further strides in efficiency or product improvement are being made. Markets are functioning well in a narrow neoclassical sense, but they are not very dynamic. The quaestions about neoclassical adjustment thus re- late not so much to what is going on but rather to what may not be occurring. It is too early to tell whether some markets are settling down to static im- passes with minuscule economic profits or whether innovations and improve- ments will subsequently lead to renewed disequilibrium and further industrial development. In the case of the PBJ the introduction of a simplified version as market competition became fiercer is a sign that some producers are trying to improve their products. Buat in the more backward areas the prevalence of imitative competition may mean that the local TVP sector will continue to lag behind the more advanced areas. Investment Competition Investment competition is similar to imitative competition but involves large investments of community resources to build facilities that are capital- intensive and often use advanced, sometimes even imported, technology. It is a means by which latecomers to an industry or to TVP development try to catch up with or leapfrog the frontrunners. In a number of industries many localities have built similar plants with large investment requirements after being shown the way by a few forerunners. State enterprises may become in- volved as well. General market trends under investment competition are broadly similar to those under imitative competition and neoclassical adjust- ment. But since the investments involved are large and often involve sunk capital that cannot easily be shifted to other uses, the potential losses from overshooting and excessive entry are much greater. A good example of investment competition is the aluminum window frame industry. Demand for such window frames rose sharply with the increase in housing construction and the shift from wood to aluminum. This in turn spurred a boom in construction of aluminum extrusion facilities, often with imported equipment. As a result, the market has become saturated, the num- ber of firms in the industry may be excessive, and profit margins are nar- row. These trends are reflected in the experience of an aluminum extrusion plant in Nanhai. In August 1984 the market for aluminum window frames looked extremely good. The enterprise (which had previously produced furniture) and the township government decided to shift to this product. The change required a large investment in imported equipment, financed by bank loans. But by the time the enterprise reached full capacity, the market had been flooded Market Interactions and Industrial Structure 95 with new entrants, and profit margins were squeezed by falling output prices and the rising market price of aluminum. The enterprise now faces problems in servicing its debts and can do so only with tax concessions. When the output price was Y17,000 a ton in late 1984, the gross profit margin on aluminum window frames was about Y6,000 a ton. By early 1987 output prices had declined by 30 percent, the market price of aluminum in- gots had risen by more than 60 percent, and gross profit margins were only Y1,800 a ton, less than one-third the original level. There are now more than 200 aluminum extrusion plants producing window frames in China, but only about 40 percent of them make profits. Although there is an export mar- ket, the domestic price is so much higher than the international price that substantial domestic currency losses are incurred on exports, and the Nanhai plant does not export any of its output. Investment competition is inherently much riskier than imitative competi- tion, particularly for the community governments that finance the large in- vestments involved or guarantee bank loans. The aluminum plant mentioned is by far the most important enterprise in its township, and its financial performance has a great influence on the financial situation of the township government. Since profit margins are initially high, timing is a crucial deter- minant of success. Competitors who come on line early can make profits to repay their large investments quickly, whereas those who are a little late may never recoup their investments. As in the case of imitative competition, the question arises whether there is only a one-time improvement followed by market saturation or whether there can be continuing dynamic change. The greater losses that can occur as a result of the higher requirements for capital investment must be set against any benefits. Moreover, in many cases the equipment is so specialized that switching products in the future may be difficult. But despite the risks, townships in both Nanhai and Wuxi seem to realize that in these kinds of activities success depends on quick decisions and adequate investments. The willingness of county and township officials in both Wuxi and Nanhai to take risks and in particular to invest in large rather than small plants is impressive. Innovative Competition Some parts of the TVP sector are experiencing something akin to innovation, which continually disrupts market equilibriums and leads to improvements in products and efficiency. In the TVP sector this form of competition primar- ily involves not so much genuinely "new" innovations as adaptation and a continual search for new ideas from elsewhere in China and from abroad. This kind of competition is most evident in Wuxi, where an infrastructure is being created to support it. Educational facilities, technical and other training, and importation of experts from the state sector are all part of this strategy. In Nanhai the need to compete in international markets also re- 96 WILLIAM A. BYRD AND N. ZHU quires continual improvement and a constant search for better methods and ideas, which to a large extent are brought in from Hong Kong through com- pensation trade, processing, and other arrangements. Interactions with State Industry COMPETITION. State enterprises are significant competitors in many TVP product markets. At the national level the bulk of TVP industrial output con- sists of product categories in wvhich state enterprises have a dominant share. Because markets for some goods are geographically localized or administra- tively fragmented, the two sectors may be insulated. But overall, there is con- siderable competition between state enterprises and TVPS. For goods that are chronically in short supply at low government-set prices, TVPS may have an artificial market niche. TVPS usually produce these goods at higher cost and sell at higher market prices, but the shortage situation protects them from the full brunt of competition from state enterprises. Under these circumstances TVPS can be viable and competitive even though they are producing goods of lower quality or with more backward technology. The two-tier planning and market system can thus insulate TVPS somewhat from competition with the state sector unless the state enterprises sell a sub- stantial share of their output at market prices. In processing industries subject to the two-tier system-for example, in cement-there are often significant economies of scale. Thus, if prices were decontrolled and state enterprises responded to the higher prices and expanded supply, many TVP producers might have to exit, and some TVPS might survive only because high transport costs allow them to remain competitive in local markets. In a buyers' market, where demand is no greater than supply at the going price and where there may be some excess capacity, competition with the state sector is different. Typically these industries had relatively high prices, which attracted TVPS and put downward pressure on prices and sales of state enterprises. In industries where TVPs have overwhelming economic advan- tages because their labor costs are lower, economies of scale are unimportant, and so on, they may take over the bulk of the market, leaving state enter- prises in a marginal position. In other industries state enterprises may have a strong advantage because of privileged access to key inputs, better technol- ogy, administrative exclusion Of TVPS, or other reasons, and they may be able to repel competition from TVPS. More commonly, high-cost state enterprises can keep a niche in the market only by producing higher-quality or more "modern" products or through better access to distribution channels. Generally, competition with the state sector is more important for techno- logically more sophisticated, capital-intensive goods, which are produced pri- marily in advanced areas such ats Wuxi and Nanhai. In the case of many low- technology, low-value goods that have localized markets (for example, low-quality cement), direct coimpetition between TVPs and state enterprises Market Interactions and Industrial Structure 97 may be limited. In industries where economies of scale are not critical and capital requirements are not great, TVPS tend at least to hold their own in competition. More often than not they do better, leaving state enterprises to struggle with declining market shares and financial problems-for ex- ample, in low-end textiles and simple consumer durables. COOPERATION. Production of parts, assembly work, and the like are often subcontracted to TVPs by state enterprises. Sometimes the latter simply mar- ket TVP products under their own brand names. Or state enterprises can use cooperation to gain access to raw materials. TVPS can even be a dumping ground for state enterprises' old equipment as it is replaced or retired. Many subcontracting agreements, joint ventures, and other arrangements are motivated by state enterprises' need to gain access to certain factors of production that they cannot obtain otherwise, notably land and labor. A re- lated motive is to take advantage of lower-cost labor in TVPS. Although in static terms these arrangements exploit the labor cost differential between the two sectors and are financially advantageous to state enterprises, like in- ternational outsourcing, they may transfer manufacturing capabilities to the TVP sector and thus threaten the position of state industry over the long run. The most advanced TVP areas, such as Wuxi, have moved away from heavy reliance on subcontracting for urban industrial state enterprises. The share of Wuxi's TVP industrial output that consists of such subcontracting declined sharply from more than 70 percent in 1978 to only 21 percent in 1985 (Jiangsu Academy of Social Sciences 1981, pp. 5 and 18, and information obtained in fieldwork). This decline may have been related to sectoral shifts in the composition of Wuxi's industrial output, to withdrawal by state enter- prises that faced reductions in demand and tight markets, or to development of the Wuxi TVP sector itself to such a point that this cooperation was no longer needed. Nanhai's TVPS, in contrast, have extensive ties with Hong Kong businesses (although their links with local state enterprises are less important). TVPs' Compeitive Strategies Several distinct TVP strategies can be identified. There is almost invariably a large premium for being first. Early entrants into a newly blossoming activ- ity can earn very high, lightly taxed profits and rapidly grow to substantial size. Then, even when profit margins narrow, these firms can more than hold their own. Hence many TVPS tend to rush into newly discovered profit oppor- tunities as quickly as possible. This is consistent with the imitation strategy because earlier imitators can still make high profits but with less risk than is taken by the first, pathbreaking entrants into the new activities. Another prototypical competitive strategy is to make rapid changes in product lines to take advantage of changing market opportunities. Both tradi- 98 WILLIAM A. BYRD AND N. ZHU tional community enterprises and private enterprises have shown great flexi- bility in developing new produicts and changing product lines quickly to avoid being stuck too long in industries that face declining market trends. Obvi- ously a key prerequisite for this strategy is low capital intensity or assets that can easily be switched to new activities-in Williamson's (1985) terminol- ogy, asset specificity must be low. It is much easier to institute a radical change in product lines for an agricultural machinery plant or a furniture fac- tory than for an aluminum extrusion plant. Given the increasing severity of competition in most TVP markets, established producers may need mnassive new investments and increasing capital intensity to make substantial im- provements in product characteristics and quality. A strategy of high invest- ment and capital intensification in, for example, Wuxi appears designed to achieve and maintain an edge in increasingly competitive markets; it is not merely a response to high labor costs or labor shortages.3 At the community level, the strategies followed by community govem- ments parallel those of enterprises but with some additional features. Com- munity governments can choose whether their firms should specialize in a dominant line of business or produce a wide array of products, and they can decide how many community enterprises to establish. These questions are intimately related to the issue of business risk, its impact on the community, and how to deal with it. A "conglomerate" strategy can greatly reduce market risk, but specialization can reap certain economic benefits. Having a large number of enterprises can reduce enterprise-specific risk, but fewer, larger firms may offer economies of scale. In the more developed areas, such as Wuxi and Nanhai, the typical pattern at the township level is ten to thirty township enterprises engaged in a range of activities, with some concentra- tion in one or two industries. Although the biggest firm alone does not ac- count for a large share of a township's GVIO, the top three to five enterprises may contribute a substantial share of both output and profits. Thus most townships appear to follow neither extreme specialization nor extreme diver- sification. Adjustment to Adversity The long-term development of China's TVP sector depends on its ability to respond positively and flexibly to adverse changes in the market. Cornpeti- tion has become increasingly severe in many industries, and there are few if any lines of business in which the TVP sector is protected from competition. How well have the TVP sector and different types of TvPs adjusted? There is now enough experience to give a preliminary assessment. ALthough the TVP sector boomed in 1978-85, some industries and regions have experi- enced significant adverse shifts. Moreover, the TVP sector as a whole was squeezed by the national tight credit policy in 1986, which particularly af- fected advanced areas and machinery production bases such as Wuxi. Market Interactions and Industrial Structure 99 At the enterprise level, a number of automatic adjustment mechanisms reflect adverse changes quickly. Price declines, as we have seen, can be steep. Quantity declines may be forced on firms by lack of demand. Output invento- ries may increase for a time, but after enterprises adjust production, there may be excess capacity. Because of the rapid price adjustments, forced changes in quantity are probably not as important for TVPS as for state en- terprises. These market adjustment mechanisms do not involve active re- sponses by enterprises, but they quickly transmit new signals on supply and demand. A sphere in which TVPS make positive adjustments is labor and wages. Performance-based wages lend themselves to reductions in workers' pay be- cause they fall automatically with declines in output or profits. Some large wage reductions have been observed. For example, in a machinery plant in Nanhai average wages of workers fell from more than Y250 a month in 1985 to Y150 a month in 1986. Such examples are by no means unusual. TVP workers bear part of the burden of adjustment, and the enterprise is to some extent cushioned. The clear linkage between wages and output or profits makes declines more acceptable to the workers. There are limits, however; in Wuxi in 1986 average wages continued to rise despite a sharp decline in profits (see below).4 Wage cuts in an adverse situation sometimes lead to voluntary departures by workers and hence to reductions in the work force. In the more developed areas such as Nanhai, high-wage opportunities are available and there is a functioning labor market. In Shangrao, a poor locality, some firms have vir- tually stopped production owing to management problems as well as market difficulties. Since most of the workers are on piece rates, they would be paid virtually nothing for staying, so they go back to agriculture or engage in small-scale commercial activities. Another common practice is to lay off tem- porary workers from other areas. Sometimes even regular workers can be laid off, often with the right to return when conditions improve. But in general, downward wage adjustments seem more common than sharp, involuntary re- ductions in the work force. More active, dynamic responses to a worsening market situation fall under the general category of product adjustments, which include improving quality, changing specific product varieties, cutting costs (often through large new investments), or producing a cheaper product in greater demand. Another adjustment is to improve the product, which may or may not involve an ef- fective price reduction, depending on whether the price is kept constant or is raised. Finally, completely new products can be developed. Financial adjustments deal with an adverse market situation by permitting different institutions to share the burden of adjustment. A decline in profits automatically cuts profit taxes and, in areas such as Wuxi and Nanhai, profit remittances to the local community government.5 Another financial response is to build up debts and payables, but this is limited by the tolerance of banks 100 WILLIAM A. BYRD AND N. ZHU and suppliers. Finally, many T'VPS can obtain tax exemptions in times of mar- ket difficulties. Drastic adjustments that affect the very existence of the firm also occur. The work force may decline if workers leave because of low pay, but shrinkage of enterprise assets through sales or otherwise is infrequent. Other common responses for community enterprises involve contracting, leasing, or even selling the firm to private individuals. Exit by small private enterprises is often easy because little capiital is tied up in the operation. Shutdowns of township enterprises are rare but not unknown; shutdowns occur more fre- quently among village and team firms. One reason for the rarity of exits by township enterprises is that in many areas the bank debts of defunct firms become the responsibility of the township government and have to be repaid from the eamings of other township enterprises in that community. In gen- eral, exit is much more common in the TVP sector than for state enterprises. In Nanhai, for example, about 4 percent of community enterprises-mostly village firms-and roughly 15 percent of private enterprises went out of busi- ness in 1986. At the community level, the community government can to some extent vary the flow of funds between it and higher levels, but its hard budget con- straint (see chapter 17) limits the extent to which this can occur and the length of time over which revenue shortfalls can persist. The ability of com- munity governments to respond to adversity depends crucially on the discre- tionary funds at their disposal. If these funds are large, as in the better-off townships of Wuxi and Nanhai, community enterprises can be given the lee- way to attempt more positive responses. But if most flows of funds from com- munity enterprises to community governments are used to meet essential public expenditures, the comrnunity governments cannot help finance ad- justment by allowing enterprises to reduce their remittances temporarily. Communities that specialize in certain products face great difficulties in ad- justing to adverse changes. Since the financial situation of the community government itself has usually worsened substantially, it is not able to provide much of a cushion for the enterprises. The shortage of funds also hampers positive responses such as establishment of new firms in more promising lines of business. Drastic remedies--shutdowns or at least substantial shrinkage, as well as changes in ownership-may be needed. Overall, the flexibility and effectiveness of TvPs' adjustments to adverse circumstances are impressive, as some examples illustrate. WUXI COUNTY, 1986. After extremely rapid growth of TVP industrial out- put in 1985 (nearly 80 percent: in real terms, following close to 60 percent growth in 1984), Wuxi faced a difficult situation in 1986. Demand was stag- nant or declining as a result of the national tight credit policy and the in- creasing market saturation brought on by widespread entry of TVPS in other Market Interactions and Industrial Structure 101 parts of the country. At the same time, market prices of most raw materials continued to rise. Wuxi's response was to increase TVP industrial output by an additional 21 percent, allow gross profits to decline sharply (by more than 50 percent), reduce total direct and indirect tax payments, but by a smaller margin than the decline in profits,6 sharply increase bank loans as well as fixed and circulating assets, and, surprisingly, increase the number of TVCES by more than 10 percent and the number of TVCE employees by more than 30 percent. Moreover, average wages increased by nearly 10 percent.7 The Wuxi response can be seen as an attempt to keep pushing forward even in the face of adversity, and its success is an indication of impressive flexibility and ability to adjust. THE XIQIAO TEXTILE INDUSTRY. In the early 1980s Xiqiao's TVCE weaving plants thrived because of processing contracts with local state textile produc- ers, whose own facilities could not meet heavy demand, as well as with com- mercial agencies. The TVCES wove raw materials provided by the customer into cloth for a stipulated processing fee. After 1983 private textile factories developed rapidly, also by processing for state units. Increasing competition in textile markets and the national tight credit policy imposed in early 1986 had a severe impact on this industry. State enterprises no longer had extra circulating capital with which to finance processing activities, and market demand for textiles was falling. Thus demand for processing by TVPs declined precipitously. Many private enterprises simply stopped production-80 per- cent of them did so for several months. But community enterprises did not have this option. Xiqiao's textile industry had to make the difficult and painful transition to a system under which firms produced on their own account and hence were responsible for supply and marketing. Problems were aggravated by huge losses on inventories as a result of sharp price declines and the unwillingness of many state purchasers to honor contracts. Although the sector now ap- pears likely to survive (only one community enterprise has closed down, al- though others still incur losses), there have been important shifts in owner- ship structure, and the private sector has flourished. In the weaving industry private enterprises can outcompete community enterprises because of their lower labor costs (they provide few benefits and rely heavily on young female temporary workers from poorer localities in northern Guangdong) and, in many cases, more modern equipment. (Community enterprises often were given discarded obsolete looms by local state textile mills, whereas private enterprises purchased new equipment from state textile machinery producers, who are now allowed to sell to nonstate firms.) Private enterprises are also better able to adjust to the vagaries of textile production and marketing. Be- cause of their low fixed labor costs, many of them can shut down temporarily when market demand falls off and resume operations when it is profitable 102 WILLIAM A. BYRD AND N. ZHU ADJUSTMENT THROUGH OWNERSHIP CHANGES. Xiqiao is an example of ad- justment through changes in ownership structure. When village and produc- tion team enterprises face adversity, the tendency in many areas is to privat- ize them in one way or another. The reverse-collectivization of private enterprises in trouble-does not seem to occur because community govem- ments do not want the burdern. Some successful private enterprises, however, are explicitly or implicitly transformed into community enterprises of a sort. Cases in which successful TVCES have been privatized are rare. Thus there is an asymmetry in adjustments of ownership structure in adverse circum- stances. The share of private enterprises and of community enterprises con- tracted to private management is likely to rise, perhaps even sharply. This occurs not only because of privatization of community enterprises but also because more private enterprises may be established and growth of existing private firms may be relatively fast. Industrial Structure Market interactions and the institutional and economic features of China's TVP sector have implications for industrial structure. In this section we look at selected topics in this vast area: subsectoral composition, firm size, commu- nity industrial structure, and, in a preliminary analysis, economies of scale. Subsectoral Composition As can be seen from table 5-4 (which covers TVCES only), China's TVP sector is diversified and covers nearly as broad a range of industrial subsectors as Chinese industry as a whole. There are some noticeable differences, how- ever. The shares of petroleum and electric power are minuscule for the TVP sector, and the share of metallurgy is much less than in industry as a whole, whereas the share of construction materials is much higher. Wuxi's TVCES are dominated by the machine-building subsector, which accounted for nearly half of their GVIO in 1984, compared with only about one-quarter for China as a whole. Between 1978 and 1984 the textile subsector more than quintupled its share in t:he GVIO of Wuxi's TVCE sector, while other subsectors declined moderately. Construction materials are much less impor- tant in Wuxi than nationally but are still somewhat more important than in Chinese industry as a whole. The industrial structure of TVCEs in Shangrao is radically different from that in Wuxi. Coal mining and construction mate- rials together account for 54 percent of GVIO, whereas machine building ac- counts for only 7 percent and there is-surprisingly, considering Shangrao's low wages-no production of textiles or garments. Extractive and processing industries account for 71 percent of total Gvio, compared with 35 percent for the national TVP sector ancd only 26 percent for the TVP sector in Wuxi. These differences are indicative of the gap between developed and back- Market Interactions and Industrial Structure 103 Table 5-4. Subsectoral Composition of GVIO (percentage of total) China 1985 Wuxi TVCEs Shangrao Subsector All industry, TVCEs 1978 1984 TVCEs, 1984 Metallurgy 8.0 3.4 0.0 6.7 0.5 Electric power 3.3 0.4 0.0 0.0 5.5 Coal and coke 2.5 3.8 0.0 0.0 21.5 Petroleum 4.5 0.1 0.0 0.0 0.0 Chemicals 11.2 8.4 13.1 11.2 11.1 Machine building 26.9 25.5 56.3 48.6 7.1 Construction materials 4.2 18.9 8.5 7.7 32.2 Forestry 1.6 3.0 0.2 0.6 0.6 Food 11.5 7.9 0.9 1.0 4.5 Textiles 15.3 12.5 3.0 15.5 0.0 Garments 2.4 3.7 1.3 1.6 0.0 Leather 0.9 1.6 0.3 0.5 7.6 Paper 1.3 2.4 0.0 0.0 0.9 Cultural and educational activities 2.6 3.0 0.3 2.5 3.7 Other 3.8k 5.4 16.1 4.1 4.8b Note: GVIO, gross value of industrial output. a. Excludes TVPs at the village level and below. b. Residual. Sources: State Statistical Bureau (1986a), pp. 227-28, and information from fieldwork. ward regions. The former have a high share of manufacturing and, more gen- erally, of subsectors such as machine building that take on increased im- portance during economic development. The latter have a high share of transport-protected and extractive industries and of industries that serve pri- marily local markets. The subsectoral composition of Gansu Province's indus- trial TVP sector also fits this description (World Bank 1988, chapter 3 and annex B). The potential of these industries for stimulating dynamic industrial development in backward areas is limited. Finn Size Table 5-5 shows the average size of TVCES (in output and employment) for China as a whole, for a few industrial subsectors, and for Jiangsu Province and Wuxi County. Average firm size is relatively small, although many if not most industrial TVCEs are of medium size and a few are even large by intema- tional standards.8 Differences among subsectors reflect not so much econo- mies of scale as the ease with which small firms can enter the industries con- cerned. For example, the largest average firm size is in textiles and the smallest is in electric power (primarily hydropower). 104 WILLIAM A. BYRD AND N. ZHU Table 5-5. Average Size of Industral TVCEs Township enterprises Village enterprises GVIO GVIO (miiuions Number of (millions Number of Locality of yuan) employees of yuan) employees All China (1985) 0.334 56 0.107 24 Textiles 1.275 135 0.423 59 Machine building 0.465 54 0.229 34 Food 0.153 22 0.027 6 Power 0.051 12 0.010 3 Jiangsu Province (1984) 0.616 102 0.192 45 Wuxi (1984) 1.341 130 0.367 51 Wuxi (1985) 1.655 127 0.510 54 Sources: State Statistical Bureau (1986a), p. 182, and information from fieldwork. The national statistics hide a great deal of regional and local variation. In Jiangsu Province in 1984 the average TVCE was nearly twice the size of the (1985) national average. In Wuxi in 1985 output per TVCE was roughly five times the national level. In more backward areas firm sizes are even smaller than the average. In the more developed regions a few TVps have reached a considerable size. Per firm averages for a sample of 200 large-scale industrial community enterprises studied by Zhou Qiren and Huang Zhuangjun (1987, pp. 41-42), were GVIO, Y4 million; profits, Y560,000; fixed assets, Y1.2 million; ancl number of employees, 370. In Wuxi in 1985 more than half of the 700-ocdd industrial township enterprises had a GVIO of Y1 million or more, and six had a GVIO of more than Y20 million. But there are few if any TVPS in China with gross output values in the Y50 million-Y100 million range and probably none above that level. The limits on firm size have little to do with restricted local or otherwise captive markets, since TVPs are primarily outward-oriented in both product and input markets. Lack of time to develop has been one limitation. Probably more important, however, large Tvps are tied to rural communities of no more than about 50,000 people with limited financial resources. Risk consid- erations further limit the size that most TVPS can attain (see chapter 9). Economies of Scale One criticism of TVPS is that they are suboptimal in size and are therefore inefficient. If economies of scale are significant, the overall efficiency of the TVP sector could be improved by increasing average firm size. A preliminary test for economies of scale was carried out with the use of 1985 data for sam- ple enterprises. The results reported here are tentative. A Cobb-Douglas production function was estimated, with GVio as the de- Market Interactions and Industrial Structure 105 pendent variable and with three factors of production (labor, land, and capi- tal) and three county dummy variables as the independent variables (see the appendix to this chapter). The overall fit is relatively good (R' = 0.90). Dif- ferences in productivity among sample firms in Wuxi, Nanhai, and Jieshou are not statistically significant, whereas firms in Shangrao are substantially (more than 60 percent) less productive. The coefficient for land is not statis- tically significant; a second regression was run without land and the results were broadly similar. The coefficients for labor and capital in both equations (0.81 and 0.44 respectively in the second regression without land) are sen- sible.9 The sum of the coefficients (1.26 in the second equation) is significantly greater than one, and the hypothesis that it is less than or equal to one is rejected at the 99 percent level of confidence. This suggests increasing re- turns to scale in the sample. The implication that some firms are of suboptimal scale is, however, not necessarily true. It is possible that each enterprise is of optimal scale and that larger firms are engaged in different activities than smaller ones. Efficiency could be improved by shifting capital and labor to larger enterprises that produce different goods, but it might not be possible to improve efficiency by increasing the size of plants that continue to produce the same goods. Community Industrial Structure From a number of perspectives the relevant unit in the TVCE sector is the township or village rather than the enterprise. Table 5-6 gives some basic information on the industrial structure of average townships and villages at national and local levels. In China as a whole there must be a large number of nonindustrialized townships, and most villages probably do not have even one industrial community enterprise. The average number of industrial town- ship enterprises per township is less than three, and the average number of village enterprises per village is less than one. Total industrial profits of town- ship enterprises per township in 1985 were only Y81,000, and for industrial village enterprises the figure was only Y8,000. The figures for Shangrao in many respects resemble the national averages, although GVIO, profits, and taxes are lower. Community industrial structure in many townships and vil- lages resembles the pattern in Shangrao more than that in Wuxi or Nanhai. Townships that have industrialized successfully have a large number of en- terprises (up to several score) that generate substantial output and profits and employ large numbers of community residents. Gross profits of township en- terprises per township in Nanhai were twenty-five times the national average and in Wuxi, sixty-eight times the national average; the gap at the village level is hardly smaller. Differences in industrial assets and in TVP employment per community are much smaller, which suggests that the more developed regions hold a clear edge in the efficiency of factor utilization. Table 5-6. The Industrial Structure of Rural Communities, 1985 (average per township or village; current prices except as specified) Fixed Circulating Gross Total tax GVIO assets assets profits payments Number Average Number Number of (millions (millions (millions (millions (millions Item of units population of firms employees of yuan) of yuan) of yuan) of yuan) of yuan)b All China Townships 83,182' 10,149 2.9 160 0.960 1.078 0.081 0.107 Villages 940,617' 897 0.7 15 0.070 0.023d 0.016 0.008 0.004 Shangrao Townships 34 20,674 3.5 179 0.393 0.283 0.137 0.024 0.015 Villages 343 2,049 0.6 12 0.020 0.007 0.001 0.002 0.001 Nanhai Townships 16 53,137 26.1' 3,355' 35.174 10.182 17.467 2.014 3.133 -. Villages 236 3,603 8.7' 279' 1.890' 0.476 0.315 0.232 0.111 Wuxi Townships 35 29,337 30.3 3,840 50.071g 10.22 11.41 5.47 3.16 Villages 587 1,749 5.8 313 2.70' 0.49 0.55 0.37 0.17 Township A, Wuxi 1 18,779 12 3,912 104.02g 20.87 26.26 18.92 6.92 Villages 13 1,445 6.4 422 8.27' 0.84 1.00 1.45 0.50 a. Net value of fixed assets after subtracting accumulated depreciation, unless otherwise specified. b. Includes both direct and indirect tax payments and hence overlaps gross profits to some extent. c. Includes Guangdong and Yunnan provinces, where "townships" are really small communities more akin to villages (with another layer of community govemment, the district-now called towns-between them and the county government), and villages are really comparable to production teams elsewhere in China. Hence all the unit averages are biased downward, possibly by as much as 50 percent. d. Original value of fixed assets with no subtraction of depreciation. e. Figures for all TVCEs, not just industrial enterprises. f. Gross income of village industrial enterprises rather than GVIO. g. 1980 prices. Sources: State Statistical Bureau (1986b), pp. 71, 109, 182; (1987b), p. 98; (1987c), p. 3; and information from fieldwork. Market Interactions and Industnal Structure 107 Conclusions and Issues China's TVP sector has developed into a flexible, dynamic, highly competi- tive, and relatively efficient part of Chinese industry. Its share in national GVIO doubled between 1975 and 1980 and again between 1980 and 1986. If it again doubles by 2000 (a growth rate much slower in relation to the rest of Chinese industry than in the past), it will account for nearly half of national GVIO. Regardless of whether this in fact occurs, the TVP sector today is the most dynamic and competitive part of Chinese industry. An important factor in the success of TVPS may be that community govern- ments are not able to use certain instruments to promote their development. First, barriers to trade imposed by community governments are not effective in protecting community enterprises. The market represented by a single township or even by a single county is far too small to serve most manufactur- ing firms adequately, and restrictions against "imports" or other trade barriers are difficult to enforce. Second, community governments have limited finan- cial resources and face a hard budget constraint. TVCES may be able to draw on community resources to cushion the short-term impact of adverse changes, but this cannot continue very long. Third, community governments with a significant number of firms are typically in charge of a "conglomerate" of different activities and lines of business, even though two or three of them may account for much of total output. This contrasts with the industrial bureaus responsible for state enterprises, which specialize in a particular subsector. The conglomerate industrial structure of many rural communities may preclude detailed regulation or intervention in operations by the com- munity government. Possible Shortcomings in TVP Competition The limitations of imitative competition and the dangers of investment com- petition have already been discussed. To the extent that the TVP sector relies only on these, parts of it may be caught in a rut from which it will be hard to escape. Much competition in other countries, however, also consists of imitating competitors' production activities and making large catch-up in- vestments. And the benefits of investment competition for successful firms and communities are great. The real question is whether and to what extent more dynamic competi- tive strategies and patterns are beginning to emerge in the TVP sector. There is already evidence that this is happening in Wuxi, where competitive strate- gies involve heavy investments, capital intensification, improvements in quality and products, and development of human resources. Nanhai does much less, particularly in the last area (it relies more on Hong Kong), and neither Jieshou nor Shangrao has the resources to do much toward develop- ing managerial, technical, and other needed human capital. Thus the more 108 WILLIAM A. BYRD AND N. ZHU backward areas may remain caught in patterns of imitative and investment competition. Regional and Local Imbalances Various sources of spatial inequality in TVP development are discussed in chapters 4, 9, 12, and 17. Here still another source of regional and local inequality in TVP development has been identified: backward areas may be engaging primarily in imitative competition, which offers limited benefits, or in risky investment competition, which ties up resources and often results in huge losses. Investment competition carries serious risks, particularly in processing industries in which the degree of asset specificity is high. When poorly managed firms have to close down, enterprises and communities bear heavy losses. Since management is weaker in the poorer, more backward ar- eas, enterprise failures and large losses are more likely to occur there than in the more developed regions. In addition to their much weaker base of human and material resources, backward areas, as latecomers, face a difficult environment with smaller profit opportunities in the activities that are best suited to their skills and resources. The fundamental economic advantage of poorer, more backward areas with little TVP development is their low wage rates. But this is of limited use if capital and entrepreneurs are largely immobile and are not attracted to these areas. Instead, there may be some temporary movement of labor from back- ward to more developed areas, which may not directly stimulate TVP develop- ment in the former. Under these circumstances industrial structure in the poorer, less industrialized areas is skewed toward resource-based industries, agroprocessing, and activities (most notably, production of construction ma- terials) protected from outside competition by high transport costs. Asset Specificity and Adjustment The characteristic method by which TvPs adjust to adverse market trends has been to change product lines. This has given the enterprises great flexibility and allowed them to leave stagnating or declining activities and move rapidly into the most profitable activities. It works best for enterprises and industries in which asset specificity is low-that is, in which capital assets can easily be switched to production of other goods. Such is the case in the machine- building industry, since standard machine tools can be used to produce a wide range of goods. In Wuxi the dominant share of machine building makes switching product lines relatively easy and may to a considerable extent ex- plain the great flexibility and efficiency of the TVP sector there. The degree of asset specificity is low in many other subsectors in which TVPS operate. As it has grown rapidly, the TVP sector has become involved in industries with higher asset specificity. Capital intensification, technological improve- Market Interactions and Industrial Structure 109 ment, and acquisition of imported equipment and technology all can involve an increase in asset specificity. If asset specificity indeed increases, it will be- come more difficult for TVPS to engage in their characteristic method of ad- justment. Other ways of adjusting to adversity will have to be relied on to a greater extent, or the flexibility, efficiency, and profitability of the TVP sec- tor will suffer. Size Limitations, Causes, and Remedies The existence of economies of scale in the TVP industrial sector has been suggested but by no means conclusively demonstrated. Even if it is true that many TVPS are of suboptimal scale and that in this sense the sector is not efficient, the policy remedy depends on the reasons for suboptimal scale. To the extent that the small size of most TVPS is simply a function of their lack of time so far to grow to large proportions, the problem will solve itself even- tually, provided that no administrative restrictions are imposed on the sizes of private or community enterprises. All large TVPs belong to rural communities which themselves are of limited size and have limited financial resources. If the limitations of their communi- ties are now or potentially the main constraint on enterprise size in the TVP sector, policies are called for that will increase the mobility of financial re- sources among communities and create financial instruments which success- ful TVPS or community governments can issue. These instruments would allow communities to diversify their risks and enterprises to grow beyond the limits imposed by the financial resources available in a single rural commu- nity. Administrative restrictions against entry into certain lines of activity by TVPS or against "inefficient" small enterprises are not the answer to sub- optimal firm size. TVPS operate in a market environment, and those that are truly inefficient will be weeded out as the functioning of China's markets improves. Exit by private enterprises is already frequent. Clearly specified bankruptcy regulations for community enterprises might facilitate exit by firms that are unviable in the long run, but bankruptcy should not become a means of allowing community governments to wipe out debts that their enterprises incurred to provide funds to those governments for public expend- itures. ' Appendix. Test for Economies of Scale A frequently heard argument is that TVP industrial firms, because of their small size, are not fully utilizing economies of scale, that there is great dupli- cation among these small enterprises, and that efficiency would improve if plant sizes increased. Simple regression techniques and quantitative data col- lected from sample enterprises can be used to make a preliminary assessment. 110 WILLIAM A. BYRD AND N. ZHU Data and Variables The data are taken frorn the Enterprise Quantitative Questionnaire and are for 1985, which has by far the most observations (122). The variables used in the regressions are DJS Locality dummy for Jieshou DNH Locality dummy for Nanhai DSR Locality dummy for Shangrao Y Logarithm of total industrial output value LB Logarithm of year-end labor force LD Logarithm of land used by plant K Logarithm of total capital stock-the sum of the undepreciated value of fixed asseits and "quota" circulating assets (inventories of inputs, outputs, and goods in process) C Logarithm of annual factory cost Production Function Approach A Cobb-Douglas production function was estimated (t-statistics are reported in parentheses under the coefficients). Y= 3.74 - 0.14DJS + 0.12DNH - 0.97DSR + 0.51K + 0.81LB - 0.13LD (3.65) (-0.69) (0.45) (-3.75) (4.59) (5.40) (-1.46) R2 = 0.90 The overall fit, as indicated by R2, is good. The county dummy for Shangrao is significant (and negative) whereas those for the other counties are not. Land (LD) is not statistically significant in explaining output (Y). A second regression was run without LD. Y = 4.27 - 0.07DJS + 0.34DNH - 0.95DSR + 0.44K + 0.81LB (4.89) (-0.33) (1.40) (-3.70) (4.57) (5.43) R2 = 0.go The hypothesis that there are no retums to scale was tested for both of the above-specified regressions: coeff(K) + coeff(LB) + coeff(LD) ' 1 where coeff( ) stands for regression coefficient. This null hypothesis is deci- sively rejected for both the second equation, without land (probability 99.7 percent), and the first (probability 97.8 percent). This suggests that enter- prises are not operating at efficient scale, or at least that shifting factors of production to larger firms in different lines of activity could improve effi- Market Interactions and Industrial Structure 111 ciency. But the failure to take into account differences in production tech- niques among industries gives grounds for caution in interpreting the results. Cost Function Approach A look at enterprise total cost and its relationship to size gives some idea of how the former varies with the latter. The following translog function was estimated on the basis of Fuss and Gupta (1981). C = 0.45 - 0.02DJS + 0.14DNH - 0.03DSR + 0.94Y + 0.0003Y2 (0.38) (-0.28) (2.08) (-0.46) (5.47) (0.04) R2 = 0.97 The output elasticity of cost is less than one, but not significantly. Notes 1. In China such transactions are not pure barter. Rather, they are "exchanges" that consist of parallel, at least implicitly tied, purchases and sales, often at state-set prices to give the appear- ance of conformity with price controls. 2. Many state enterprises complain that they cannot compete successfully with TVPS because the latter provide gifts, banquets, and even direct payments to customers. 3. The TVCE sector in Wuxi added a net total of 86,000 jobs in 1986, increasing the total labor force by over 30 percent despite a difficult market situation. Most of the new employees were local residents. Thus, reports of a labor shortage in Wuxi are exaggerated. 4. Conflicting tendencies were at work because output rose substantially even though profits fell; to the extent that wages were linked with output rather than profits, this would tend to push wages upward. 5. In Shangrao community enterprises are often forced to remit a certain amount of funds to the community govemment regardless of their actual profits (or losses), so this kind of adjust- ment cannot occur; see chapter 17. 6. Indirect taxes increased in line with the growth of industrial output value and therefore offset much of the sharp decline in profit taxes. Overall, the fall in total tax payments in 1986 was less than 20 percent. 7. In addition to the linkage between wages and output, another factor that may in part explain this paradoxical rise is that the pay of township cadres and factory directors is permitted to exceed that of the average employee by only a certain margin. If average wages had fallen, cadres' incomes also would have declined. 8. Generally, firms are considered very small if they have fewer than ten employees, small if they have ten to fifty, and medium-size if they have fifty to one hundred. See Little, Mazumdar, and Page (1987), p. 8. 9. If the coefficients are normalized to add to one, the capital coefficient is 0.35 and the labor coefficient 0.65. In a sense the results may be almost too good; they may reflect rising capital intensity as enterprise size increases, which would tend to lower the coefficient of capital. 10. This is a real danger in places such as Shangrao (see chapter 17). Given the relationships between community governments, community enterprises, and banks, the issue of whether loans of bankrupt community enterprises are the responsibility of their community govemment "own- ers" should be considered carefully. 6 Developnent Issues and FPolicy Choices He Jiacheng Since the decisions by the Third Plenary Session of the Chinese Communist Party's Eleventh Central Cornmittee in December 1978, China's economic development strategy has consisted of three main steps: to double the country's 1980 GNP and to meet the basic subsistence needs of the population by 1990; to raise China's per capita GNP to at least $800 by the end of the twentieth century and to improve conditions generally; and to bring China's per capita GNP to the level of middle-income countries in the first half of the twenty-first century, so that the life of the Chinese people will be rela- tively prosperous. China has largely completed the first step and now must focus on the second. The key lies in the expansion of industrialization and the improvement of life in the countryside. Since poverty and underdevelop- ment in China are mainly confined to the countryside, the second step will have no chance of success if rural per capita income does not increase dramat- ically. The transformation of a low-income country into a middle-incorme one will be accompanied by a shift of labor from agriculture to nonagricultural sectors, including industry. In other countries this shift of labor is often char- acterized by massive migration from the countryside to the cities-hence the close relationship between industrialization and urbanization. The interna- tional experience has been that when the per capita GNP of a large country (one with a population of 20 million or more) exceeds $800, its agricultural labor force falls below 50 percent of its total labor force, and the rural popula- tion falls below 60 percent of the entire population. When an extra-large country (one with a population of 100 million or more) enters this stage of development, the percentages of its agricultural labor force and its rural popu- lation tend to be even lower (see World Bank 1984). These considerations are relevant to China's development prospects to the year 2000. If the country's population program is successful and the popula- tion can be kept to about 1.2 billion in 2000, China's rural labor force will climb to no more than 450 million. The number of people involved in agri- 112 Development Issues and Policy Choices 113 culture should then be fewer than 300 million. Nonagricultural activities in the countryside will have to absorb 150 million people, and the local urbani- zation process will have to accommodate 180 million. In contrast, the pres- ent rural nonagricultural labor force is only 80 million, and the number of people undergoing local urbanization is still relatively small. During the dozen years between now and 2000 China will have to shift 7 million addi- tional workers to nonagricultural activities each year, which means that every year large numbers of people will migrate to towns and townships in the rural parts of the country. All this boils down to the question of whether nonagricultural activities in rural China will develop as expected during the next twelve years. Judging from past experience, this depends to a large extent on an accurate under- standing of the problems and difficulties that face rural industrialization and the choice of workable policies. Research on China's rural industrialization policies has in the past concen- trated on whether rural industries should be allowed to grow. There was some fear that the development of TVPS would undermine state planning and weaken the state economy and that it would not be in conformity with the country's socialist orientation. This issue was basically settled by the Central Committee's "Resolution on Several Questions about Speeding up Agricul- tural Development" in 1978 and was further clarified by Document No. 4 of 1984. The focus of research on rural industry is shifting to such questions as its further development and, particularly, general rural development policy as the foundation of rural industrialization, the help needed by rural industri- alization, coordination of policies goveming rural and urban economies, and the reform of rural enterprise management as the driving force of rural indus- trialization. This chapter looks into these policy issues. The Foundation of Rural Industrialization: Agricultural Development Policy Further industrialization in the Chinese countryside will run into problems arising from the conflict between industrial and agricultural development. Grain output has fallen in the past couple of years after hitting a record high in 1984; the harvest in 1987 was no better than the great harvest of three years before. This raises the question of whether rural industrialization has diverted resources from agriculture. There is some concern that the growth of rural industries may have been too fast. The unsatisfactory situation of rural industry before 1978 was attributable to the traditional industrialization strategy. Conditions had actually long been ripe for developing rural industry, since the agricultural base was capa- ble of supporting industrialization. The nation's average per capita grain sup- ply was 304 kilograms in 1957, 318 in 1978, and 370 in 1986. Naturally this situation was more conspicuous in the relatively developed areas. Wuxi's 114 HE JIACHENG grain output rose from 214,500 tons in 1949 to 417,000 tons in 1970, grow- ing at 3.2 percent a year. Grain output per farmer, however, increased only from 850 to 880 kilograms. During the fifteen years from 1956 to 1970 output per farmer even fell, from 1,250 to 880 kilograms. The rapid pace of rural industrialization after 1978 was prompted by the shift of emphasis in the entire national industrialization strategy and by re- lated adjustments and structural reforms. Controls over the nonagricultural economy in rural China were loosened, the production responsibility system (PRS) was introduced in farming, and state procurement prices for agricultural products were raised. The change in procurement prices transferred consider- able amounts of industrial profits to the farmers, which gave them a financial foundation for entering nonagricultural businesses. In relatively developed Wuxi, price rises between 1980 and 1984 generated 81.9 percent of the total increase in agricultural incornes. This ratio, of course, was not as high in underdeveloped counties because of their greater potential for increasing agri- cultural output, but the transfer of part of urban industrial profits to the coun- tryside through higher agricultural procurement prices probably had a higher marginal effect there. Moreover, rural economic reforms in the past few years have greatly promoted the development of agricultural production and so strengthened the base for rural industrialization. Nonetheless, Chinese agriculture is still far too backward. Several years of vigorous growth on the part of rural industry have left agriculture too weak to provide a stable foundation for rural industry. Rural industries have occu- pied farmland and absorbed able-bodied farm labor to some degree. For ex- ample, the sown area for grain crops fell by 7.73 million hectares during 1978-84 and decreased another 4 million hectares in 1986 alone. The expan- sion of rural industries has led to a relative and even absolute decrease in investment, particularly long-term investment, in agriculture that has jeopar- dized the growth of the sector. Furthermore, the PRS and the rise in state procurement prices for farm products have already brought about a substantial increase in agricultural production and can no longer offer great incentives for continued rapid development. The areas with more advanced agriculture are usually those where rural industries have experienced rnore rapid growth, but they are also the areas where the conflict between industry and agriculture emerged earlier and now appears more acute. The conflict is especially noticeable in Nanhai, where the local TVP sector relies more heavily on private ownership, individual op- erations, and associations, but it is also found in places such as Wuxi, where the principal investors in rural industry are township and village govern- ments. In Wuxi agriculture not only has failed to accumulate funds to support industry but also has absorbed a large share of local industrial profits. Thus the more advanced is a counry's rural industry, the more its industry is out of balance with agriculture. Under such circumstances the present rural industrialization process in Development Issues and Policy Choices 115 China may reproduce the type of relationship between industry and agricul- ture created by the prereform industrialization strategy. The goal of the ear- lier industrialization process was to lay the foundation for an independent industrial system in the shortest possible time. The state limited and even forbade the shift of rural labor to nonagricultural activities, established a land management system that gave political power to grass-roots communities, made agriculture accumulate as much as possible for industry by means of the scissors differential between the prices of industrial and agricultural goods, and excluded the rural population from the urbanization process by imposing the household registration system.' An independent industrial sys- tem was indeed created in a short time but at the cost of a large gap between industry and agriculture and a worsening urban-rural income gap. As in other countries that followed similar strategies, when such a process reached the limits on exacting from agriculture for industry, the countryside became a severe burden on the entire economy. To permit further industrialization, China had to revise its scenario by making industry help agriculture. The current rural industrialization process is aimed at lifting the country- side out of poverty as soon as possible. The PRS permitted farm labor to under- take nonagricultural activities. The scissors differential in the prices of indus- trial and agricultural commodities was reduced. Large loans and tax breaks were extended to rural industries, which in turn were asked to aid farming. This program did cause rural industry to mushroom and freed some areas from abject poverty. But it has failed to provide a basic solution to the conflict between industry and agriculture. Chinese agriculture cannot at present provide rural industry with a reliable foundation. Successful experiences in this regard resemble either the United States, where agriculture has always been a strong base for industrial develop- ment, or Japan, where industry and foreign trade can compensate for the defi- ciency in domestic farming. The United States has abundant natural re- sources for agricultural development, and Japan boasts the broad competitive edge of its industry in international markets. China's per capita agricultural resources are severely limited-the country has one-third of the world's farm labor but only one-eleventh of the world's cultivated land-and the Ameri- can approach is thus not applicable. At the same time, China's urban indus- try continues to face hardships in competing in intemational markets, and its even more backward rural industries cannot be expected to depend solely on foreign trade to overcome the inadequacies of domestic agriculture, at least in the near future. All this calls for extreme caution in managing the relationship between industry and agriculture in rural regions. The slackening growth of farm output since 1984 and great frictions over the development of the nonagricultural sector in recent years have forced governments at different levels to put a new phase of rural reform on their agenda. Some practical moves have also been made, such as easing controls on the prices of farm products, stepping up adjustment of the industrial struc- 116 HE JIACHENG ture in the countryside, and abolishing the state monopoly on the procure- ment and sale of farm staples. These measures have thus far not led to any- thing comparable to the remarkable achievements of the first phase of rural reform. The absence of a proper stress on the development of agriculture in the process of adjusting the rural industrial structure is a grave defect. The second phase of rural reform has failed to pay enough attention to agriculture and to the fact that one of the goals of the reforms is to provide a solid foundation for agricultural development. Agriculture will long remain the ultimate source of the difficulties confronting the advance of China's rural economy. Like it or not, agriculture must still be regarded as the basis of the national economic development strategy. Some hope that aid to agriculture by rural industry will provide a solution to the agricultural problem. Such aid, as it has been implemented over the past several years, involves the use of part of industrial profits to stabilize and promote agricultural production. It can to a certain degree alleviate the im- balance between industrial and agricultural development and may become more important when it is systematically institutionalized. Industry's aid to agriculture now works on a county or township basis. Some villages that have been champions of development have also included aid to farming among the responsibilities of rural enterprises. There are sometimes rules about how many persons each household can send into rural enterprises; no family may exceed the quota. There are valid reasons for balancing the interests of farm labor and rural industrial labor in this way. Yet for economists this kind of balance represents a sacrifice in many aspects-most important, a sacrifice of specialization and comparative advantage. Thus attempts to make rural industry help local agriculture do not offer a perfect solution to the problem of the long-term development of agriculture. Others count on reforms at the macroeconomic level to coordinate the interests of industry and agriculture by drastically increasing prices of agricul- tural goods and the state's investment in agriculture. Reforms at the macro- economic level, especially improvement of the market system, are significant and deserve support, but increasing the state's financial help to agriculture would be difficult. In any case, these reforms would provide little leverage for stabilizing the agricultural base for the rapid growth of rural industry. What is fundamental for ironing out the agricultural problem is a thorough land reform to heighten farmers' enthusiasm for investing in agriculture and build up the momentum for continued agricultural growth. In addition, land reform can rationalize the scale of agricultural production and bring land pro- ductivity to a new peak. Tracle among regions can thus develop according to the principles of geographic specialization and comparative advantage, and rural industries will naturally move toward the most suitable locations, where management will be better and costs lower. At the same time, state procure- ment quotas for farm produce in the areas where rural industry is more devel- Development Issues and Policy Choices 117 oped should be gradually transformed into monetary forms to enable these areas to take ever quicker steps toward industrialization. Coordination of Policies Goveming the Rural and Urban Economies China's rural industrialization faces another category of problems: the imbal- ance between the cities and the countryside. This is a product of the tradi- tional industrialization process, which led to an excessive dichotomy between urban and rural living conditions. The lack of progress in reducing the differ- ence between the rural class and the urban class is conspicuous. In the late 1970s the ratio between per capita income in the countryside and that in the cities was 1:1.7, close to the Indian level of 1: 1.8. If one also considers the difference between the two economic systems-for example, China's full employment scheme, its welfare arrangements for urban residents, and its pricing and marketing system for consumer goods, which gives urban people higher real purchasing power-the urban-rural dichotomy in China was actu- ally even wider in the 1970s. The urban-rural income gap decreased markedly following the introduction of rural reforms in late 1978 and reached its nar- rowest point in 1982. It then widened again, despite the faster growth of rural industries. In 1982 the average income of a rural laborer was 61.8 per- cent of that for an urban worker, but the figure declined to 58.9 percent in 1984 and to 48.4 percent in 1985. What is disturbing is not the gap itself. In the primary stage of such a pro- found social transformation as the Chinese reforms, many factors (some of them transient) can affect the urban-rural income gap. Raising the procure- ment prices of farm produce, adopting the PRs, and allowing farmers to swarm into nonagricultural trades could significantly boost rural income levels. At the same time, the creation of jobs for millions of unemployed urban youths, pay hikes to compensate for the wage freeze of previous decades, and, espe- cially, urban enterprises' thirst for capital investment could lead to rapid in- creases in city dwellers' incomes. The evolving new urban economic structure may generate rapid growth led by technological upgrading, which may further widen the income gap. Moreover, agricultural growth is subject to con- straints, and farmers' incomes can hardly increase when they depend solely on agriculture. This being the case, it is rural industrialization that has the historical mission of closing the urban-rural gap. Rural industrialization has three key components: reducing the difference between factors of production in the countryside and in the cities, rationalizing the division of labor in the urban-rural industrial structure, and synchronizing the development of rural industrialization and urbanization. One obstacle to balanced urban and rural economic development is the immense difference between factors of production in the countryside and in the cities. The essential factor of production is not any form of material capi- 118 HE JIACHENG tal but human capital. Areas with a higher level of rural industrialization are not usually characterized by affluence in natural resources or finances, but they are located where human capital is more abundant. Compared with Wuxi, Shangrao has more natural resources and, relatively speaking, more funds for launching rural industries. Both as an old revolutionary base and as an impoverished area, Shangrao receives sizable financial aid from. higher authorities. But its rural industries are underdeveloped, mainly because of the lack of human capital and the poor quality of township economic commis- sion (TEC) officials, entrepreneurs, technicians, and supply and marketing personnel. In Shangrao the quality of TEC officials, even in the economically most advanced towns, is conspicuously lower than that of their peers in the devel- oped counties. Some are even unacquainted with the basic financial situation of local enterprises or lack rudimentary knowledge of technology, manage- ment, and business. Entrepreneurs in the county lack the qualities of modem industrialists and enterprise managers. Since Shangrao has virtually no his- tory of modem industry, it has at best only some craftsmen, not skilled indus- trial workers. Both the administration and the public in Shangrao tend to underestimate the role of marketing personnel in developing TVPS. Marketing personnel in the county have failed to become an independent interest group or a strong backbone of local enterprises (see He Jiacheng 1987). Shangrao's shortage of human capital undoubtedly has roots in the county's unfavorable geographic and economic position and its short experi- ence as a market-oriented economy as compared with Wuxi. Yet the funda- mental cause of Shangrao's predicament is its failure to augment the local pool of human talent. Efforts in this area have taken various forms in Wuxi. Qianzhou township has brought in more than a hundred industrial pensioners and administrative personnel from Shanghai and Wuxi cities. The county has been trying vigor- ously to tap local human resources as well. Promotions go first to talented local residents. There have been 6,000 such cases, including 1,600 manage- rial staff, 2,000 marketing personnel, and 2,300 technical workers. Wuxi also cultivates its professional manpower. During the past few years it has sent about 200 local personnel to universities and technical schools. Other forms of education that the county has employed include a television university, a correspondence school, radio lectures, exchanges of trainees and technical advisers with enterprises elsewhere, seminars, and encouragement of self- education. Wuxi maintains a social atmosphere conducive to individual crea- tivity through, for example, local policies concerning pay scales, promotions, awards of technical titles, and public opinion. As a result, urban-rural differ- ences in human resources are no longer salient in Wuxi. Another obstacle to balanced urban-rural economic development is the similarity of urban and rural industrial structures. The industrial structure of Development Issues and Policy Choices 119 China's rural economy resembles to a surprising degree that of the entire na- tional economy. The correlation coefficient for fourteen subsectors in rural industry and in China's industry as a whole is 0.721, that for eight heavy industrial trades is 0.793, and that for six light industrial trades is 0.923. (For more details see Development Research Institute 1986.) Such great similarity in industrial structure worsens the prospects for rural industrialization, for the technological upgrading of rural industry, and for full utilization of the advan- tages of urban industry. Urban-rural competition remains on the same low level, making it difficult to institute effective coordination. A retrospective look is needed to explain the cause of this similarity. In response to Mao Zedong's call for mechanization of farming in China by 1980, the 1970 Northern China Agricultural Conference decided to allow and to support with funds and materials the establishment by communes and brigades of factories for making farm machinery and tools. But production of farm machinery was badly suited to the resources and factors of production in the countryside. Farmers had to rely almost totally on state support, but this support was not adequate for successful development of the farm machin- ery industry. Few communes and brigades could actually afford to set up a farm tool factory, let alone a farm machinery factory. More often than not, the factories that were established lost money because of the severe shortage of production and management expertise and the relatively low prices of farm machinery. Faced with this challenge, farmers either closed their factories or went beyond the boundaries designated for their business activities. The more advanced localities such as Wuxi shifted to making the general machine tools needed for farm machinery production in hopes of subsidizing the losses incurred by farm machinery with profits from comparatively high-priced ma- chine tools. By 1978 the structure of rural industry had become highly similar to that of industry as a whole. A significant change came in 1978 with the Central Committee's resolu- tion on speeding up agricultural development. The document declared that all suitable farm products should gradually come to be processed in the coun- tryside, if doing so was economically rational, and that urban industries should allow part of their finished products or components to be processed by commune and brigade enterprises using equipment and technology sup- plied by the urban industries. Because no timely complementary measures were adopted, and perhaps because the scope of business designated by the state was too wide, farmers entered virtually all industrial fields. This further distorted the rural industrial structure and generated fierce competition for raw materials and markets between commune and brigade enterprises and state enterprises. The 1981 national economic readjustment program affected commune and brigade enterprises. In Jiangsu Province, where rural industry was most ad- vanced, the main results included: 120 HE JIACHENG * The closing of a number of production lines and projects that had been started without considering local economic advantages (as occurred, for example, in calcium carlbide, coke, electric fumaces, asphalt felt, asbes- tos, pharmaceuticals, and activities such as animal bone processing) * Restrictions on the development of cotton textile manufacturing and processing, steel rolling,, flour milling, and firecracker manufacturing * Controls on activities that required high energy consumption but turned out substandard goods, such as some pesticide manufacturing and boilermaking * Restraints on the production of commodities that involved high pollu- tion or were not very marketable * Encouragement of production of parts for state-run industry and of co- operation with state enterprises * More careful examination of projects based on nonagricultural materials * Preferential treatment for the processing of farm produce. The readjustment brought about a substantial change in rural industrial structure and the beginnings of more balanced development patterns. As imbalances in the national economy were being corrected, the Central Committee stated its rural industrial policy in Document No. 4 of 1984. This document called for further balancing and readjustment and for priority for food, animal feed, building imaterials, construction, and energy. Again, owing to the lack of complementary policies, especially macroeconomic guid- ance to the newly mushrooming private enterprises, the similarity between rural and urban industry has become more marked in recent years. The last obstacle to balancecl urban and rural economic development stems from the conflict between rural industrialization and urbanization. The prob- lem with the traditional industrialization process was by and large not one of fast industrialization and belated urbanization but rather of distortions in city size distribution. The share of China's urban population increased from 10.6 percent in 1949 to 17.6 percent in 1978. This was 24 percent lower than the average for low-income countries, but correction for the underesti- mate caused by the Chinese definition of urban population yields a degree of urbanization at least equal to the average of low-income countries if not slightly higher. The urban stnrcture, however, appeared undesirable; large and extra-large cities accounted for a big percentage and cities with popula- tions of 100,000 or less accounted for only 32 percent, in contrast to 47 per- cent in India, 42 percent in the Soviet Union, and 41 percent in Brazil. The acceleration of agricultural development and rural industrialization since 1978 has given rise to urbanization and the growth of towns and has to a certain extent cured the inappropriate structure of urban scale created by traditional industrialization. TvPs are mostly clustered in or close to local Development Issues and Policy Choices 121 commercial hubs and attract a large part of the surplus agricultural labor force. The thirty-five towns of Wuxi have, on average, thirteen factories and 1,600 workers per town, and fourteen of them report that workers from sur- rounding villages constitute 30-40 percent of their populations. In eight other towns such workers account for 40-60 percent of the population. At the same time, urbanization as part of rural industrialization has some noteworthy defects. First, the two do not usually keep pace with each other. In some areas the latter lags, in others the former. The sample counties that do not have their own county seats-Wuxi, Nanhai, and Shangrao-are cur- rently engaged in large-scale construction programs to build county centers. Wuxi and Nanhai enjoy a much higher level of industrialization than Shangrao and have both the finances and the need for county seats. Their industrialization cannot realize scale economies and externalities if it is not correlated with urbanization. In contrast, Shangrao cannot justify building a county seat. Its industrialization process has barely begun, its government still operates with state financial subsidies, and it has neither the need nor the capability to undertake the project. If urbanization is divorced from in- dustrialization, it will inevitably lead to inappropriate and high-cost invest- ments, which will retard industrialization. Second, the construction of small towns often does not achieve economies of scale. Township enterprises are commonly located in the towns that serve as seats of township governments, and village and team workshops tend to be located within the precincts of their villages or teams. In 1984 Jiangsu Province had 1,962 townships and 22,592 industrial township enterprises with a total employment of 2,315,376. If all township enterprises were head- quartered in township seats, each town on average would have no more than 11.5 township enterprises and 1,178 workers. Even though some towns are bigger than others, the broad picture is still one of anarchy in developing small towns. In 1983, 2,100 towns were scattered around Jiangsu Province's 100,000 square kilometers. On average, there was one town for every 51 square kilometers, but the density of towns varied in different areas-one for every 70 square kilometers in northern Jiangsu, one for every 42 square kilometers in southern Jiangsu, and one for every 36 square kilometers in middle Jiangsu. The highest density was found in Yangzhou city (Jiangdu County), where there was one town for every 22.6 square kilometers of land area, while in Huaiyang city (Hongze County), which had the lowest den- sity, there was one town for each 108 square kilometers. The problems that crop up in the development of small towns are caused by erroneous guiding concepts and by defects in the management structure. Traditional industrialization neglected the growth of small towns, and the consequences were harmful for rational urbanization. But now the overem- phasis on the construction of small towns and the existing management struc- ture are preventing small towns from expanding into medium-size cities. 122 HE JIACHENG Under such circumstances it will be hard to avoid new distortions in the pat- tern of urbanization. The organizational reforms of the past few years allowed city governments to oversee the counties in their surrounding areas and township authorities to coordlinate the affairs of subordinate villages. These reforms have alleviated problems to some extent but are not a basic solution to the barriers separating the countryside and cities and to the conflicts be- tween towns and surrounding rural areas. The solution to these difficulties is, first, to emphasize the human factor. Financial policies alone (including direct transfer of funds and greater use of urban savings for rural credit) will not resolve the problems. The adjust- ment of the urban industrial structure in coming years will also bring about a huge demand for finance, and it is both impractical and unreasonable to sacrifice urban economic development. Much should be done to augment rural human capital through expanding the urban technology market and increasing flows of personnel to the countryside. For the poor areas, what is most important seems to be improvement of the quality of intermediate talent. In addition, the principle that rural industry should be based entirely on local resources has to be repudiated, and the policy of letting rural enterprises supplement large industries should be revised according to the realities in the countryside-its abundant labor force and insufficient finance. Urban-rural industrial development policies should promote the conversion of labor into funds. Only policies of this kind, especially policies to govem the founding and technological advance of rural industries, can fundamentally solve the problem that when state industry grows, rural industry inevitably appears to be lagging, and the other way around. By shaking off noneconomic ties with state enterprises, TVPS can achieve more coordinated industrial development. More important, there must be structural reforms-of the investment struc- ture and of commercial management, for example-to guarantee the func- tioning of these industrial policies. There should be a strategic focus on developing existing small cities and towns directly affiliated with county govemments into medium-size cities. A long-term development program for small cities and towns should be drawn up as quickly as possible, and the management structure for small cities and towns should undergo reform. First, cities should act as the economic hubs for their nearby counties so as to combine the urbanization efforts of the cities and the countryside. Second, China must strengthen regulation and control of the construction of county seats so that the expansion of old county seats and the building of new ones are covered by an overall program. Third, the construction of towns governed by townships should be more concentrated geographically, and, in particular, the locations of TVPS should be more cen- tralized. But this demands a thoroughgoing reform of the relationship be- tween TVcEs and their govemments. Development Issues and Policy Choices 123 The Driving Force in Rural Industrialization: The Structural Reform of TVPs The third set of problems involves the conflict between the development of rural industry and the management system of TVPS. China's traditional mode of industrialization was not as centralized as the Soviet model. From the be- ginning China let local governments run enterprises taken over from capital- ists and individuals. Later, local industries enjoyed further development when local governments were encouraged to take some initiative. But a highly centralized system of planned distribution and allocation was still the main feature. This system, although it promoted rapid industrialization, con- flicted sharply with the needs of industrialization. The result was a slowdown of the pace of industrial development, low and declining economic returns, and limited benefits for the people. The current system of rural industrialization was also determined by condi- tions created in the past. Since 1978 most TVCES have maintained their origi- nal property relationships, but some smaller firms and those long plagued by poor management underwent changes. Some enterprise assets were sold to individuals at reduced prices or were leased or contracted to them. This kind of transfer of property rights has developed to such an extent that TVCES enjoy rapid growth both in regions where they originally flourished and in a few localities with backward rural industry. In some other backward areas, pri- vately owned, household, and privately managed firms as well as partnerships have moved ahead. Since 1984, when the privately owned and managed economy was given a green light for development, it has mushroomed and spawned different modes of TVP ownership and management. These include the Wuxi mode, based mainly on TVCES, the Wenzhou mode, which empha- sizes private enterprises, and the Gengche mode, with development of both TVcEs and private enterprises. Although property relationships in China's rural industry are different in different regions, public ownership is generally dominant. But this domi- nance is not the same as during the period of traditional industrialization. Not only is the proportion of public ownership currently much lower, but TVCEs are quite different from the state enterprises that were the main form of traditional industrialization. The proportion of TVCES is still high. The share of TVCES in total employ- ment of TVPS is 97.1 percent in Wuxi, 77.5 percent in Nanhai, 96.7 percent in Shangrao, and 64.0 percent in Jieshou. This will restrict future develop- ment in many ways and will particularly obstruct the formation and growth of factor markets. For example, the high proportion of TVCES is closely related to low mobility of labor. In the sample firms the respective shares of workers from the local township, the local county, and other counties are 94, 97, and 3 percent in Wuxi and 51, 85, and 15 percent in Nanhai. In Shangrao 124 1HE JIACHENG the corresponding figures are 81, 96, and 4 percent and in Jieshou 80, 92, and 8 percent. What is more problematic is that like state enterprises, Chi- na's TVPs are subordinated to government administrations and are plagued by the problem of egalitarianism among the workers. But these similarities do not make TVPS and state enterprises identical. In the first place, the number of final owners is comparatively srnall for TVCES. In Wuxi, for example, each township has an average population of about 20,000, whereas the average population of a medium-size city is 100,000-500,000. Second, the obligations of township govemments toward rural workers as members of their communities are relatively weak and are few in number. Township govemments can directly regulate enterprises' eco- nomic activities through plans, as municipal governments do. But nonagri- cultural activities in a township are outside the state mandatory planning sys- tem, and economic exchanges among all nonagricultural firms in a township account for only a small part of their activities. Hence TVPS are controlled by the market mechanism in a strict sense. Unlike urban governments, which represent an entire city, township gov- emments cannot shoulder extensive social functions or pursue many noneco- nomic targets. Since these govemments are unable to support long-term or huge losses in their enterprises, losses suffered by TVCES tend to be much smaller than those that plague state enterprises. For instance, in 1985, 11 percent of township enterprises in Wuxi were loss-makers, and total losses amounted to 3.2 percent of total gross profits of township enterprises. But only 0.6 percent of village firms ran losses. In Nanhai 18 percent of the town- ship enterprises were loss-makers, and losses were 9. 1 percent of gross profits; for village enterprises the corresponding figures were 1.5 and 0.7 percent. Losses in TVCES last for only a short time; if they cannot be tumed into profits, the enterprises have to shift their production or close down. The TVCE form of enterprise organization has not promoted the best distri- bution and utilization of resources. If the industrial TVCE sector is not re- formed, it could have even bigger conflicts with the further development of rural industry in the future. At present, this problem is most prominently seen in the following features. * The township govemment represents the owners of township enter- prises, but it also functions as a govemment administrative unit. Therefore the township govemment has to take into consideration the interests of both the enterprise and the township. The interest of the enterprise is in maximiz- ing profits; the interests of the township include increasing the incomes of all the people in the township, the percentage of nonagricultural employment throughout the township, and the income of the township government. The two targets may conflict. The principal problem is that the township govem- ment adopts policies concerning distribution of enterprise profits that reduce enterprise reinvestment rates and obstruct the rational flow of the main fac- Development Issues and Policy Choices 125 tors of production among townships. To achieve desired social goals, town- ship governments have placed hurdles in the way of the best distribution and utilization of resources that TVCES need for their development. * Legally speaking, the entire population of a township is the final owner of the assets of township enterprises, but the local people are not actually entitled to any individual benefits. Hence they will not behave like owners of a typical joint stock firm. First, township residents do not have a strong desire or willingness to invest in township enterprises, since the investment cannot be inherited or returned as dividends and cannot be transferred else- where in response to better investment opportunities. Second, since town- ship residents do not hold decisionmaking rights in township enterprises, there are no restraints on the representatives of the owners (the township governments) and no effective inspection of the enterprises' financial status and management. This is the reason behind encroachment on the capital stock of township enterprises by a few people and is also a cause of ineffi- ciency and slow growth of capital stock in the enterprises. It eventually leads to conflicts between community residents employed by township enterprises and those outside the firms. 0 Township enterprise directors do not have adequate decisionmaking power, since factories-even those that have adopted contract responsibility systems-are not entirely independent. As a result, enterprise managers are weak and indecisive and are unable to grasp business opportunities. A factory director's decisionmaking authority is strongest for materials, less so for per- sonnel, and weakest for finances. In Wuxi between 1977 and 1984 the pro- portion of total net profits kept by enterprises was between 14 and 46 percent. The figure was 65 percent in 1985 and 60 percent in 1986, considerably higher than the 40 percent ratio in state enterprises. The decisionmaking power of a TVCE director over the use of profits retained by the enterprise is similar to that of a state enterprise director; the only difference is that TVPS are subject to fewer controls over their financial accounts and cash transac- tions. For many TVCEs depreciation funds have to be submitted to township governments for general distribution, and the use of retained depreciation funds is also subject to control by township governments. Overall, TVCE di- rectors cannot expand the capital of their firms on the basis of changes in market supply and demand. * TVP workers have come to have the same behavioral characteristics as state enterprise workers in many respects (see chapter 18). Availability of labor and low wages should have been the biggest advantages of rural indus- try, but the existing TVP system has led to coexistence of high local wage costs and a surplus of agricultural labor in general, of overstaffing in some enterprises and acute shortages of labor in other places. For instance, in Shangrao rural areas have considerable surplus labor, but the wages of TVP workers are much higher than farmers' incomes. Excessive employment in 126 HE JIACHENG some firms may account for 60 percent of the total work force. In Wuxi, al- though many enterprises have already felt the strain of labor shortages over the past several years, most firms suffer from overstaffing, which may account for 10-50 percent of the total number of employees. The problem goes far beyond this. At present, even private enterprises are not set up for the best distribution and utilization of resources. The reason lies in external conditjions rather than internal financial arrangements or structures. Individual enterprises are officially included in the socialist econ- omy, whereas privately managed firms with more than seven employees are considered a supplementary part of that economy. But most local govern- ments have failed to internalize this new attitude. In general, even in places where the private economy is highly developed, such as Nanhai, private en- terprises are discriminated against politically. This is best revealed by the way the private economy seeks survival and development. To develop further, the private economy must expand all the factors of production, including the number of hired workers. But local informal re- strictions on hiring workers and pressure from public opinion have often forced private enterprises into joint operations with township or village gov- ernments, as in Jieshou. Or, as happened in Nanhai, private enterprises sometimes have to link hands with their own workers to expand. But these options are all the least preferred choices for private enterprises. When they operate jointly with township governments, they often have to, in effect, turn themselves over to the laitter, and when they operate jointly with their own employees, they often can obtain only a small amount of capital. There are more failures than successes in the joint management arrangements be- tween private firms and community governments or employees. A related problem is that China's commodity economy is not yet highly developed, and many people are ignorant of the values needed in a modern commodity economy, such as credibility and contracts. The country's judicial work and legislation have yet to catch up with economic development, and most economic agents do not bear in mind the rules of conduct they have to abide by. Over the past several years contract responsibility systems have been intro- duced in TvPs all over the country. The most famous is that of Yanqiao town- ship in Wuxi. A profit contract system gives small or money-losing enter- prises the right to retain incremental profits or the proceeds from reductions in losses, and a quota contract system for large and medium-size firms links wages with profits and allows additional profits to be shared with the entire enterprise work force. Yanqiao's system also involves election of managers, the transformation of fixed salaries into floating wages, and employment con- tracts for workers. This system has spread widely. In 1984, 98 percent of the TvcEs in Wuxi and 90 percent of those in Jiangsu Province had adopted this form of contract responsibility system. Results have generally been good. Development Issues and Pobcy Choices 127 These reforms, which mainly involve enterprise management, did not and could not eliminate the weaknesses of the system regarding the relationship between owners and producers. This is an area for future reform. Community governments can, on the basis of local conditions, sell part of their enter- prises to individuals and implement an intemal ownership system in a planned fashion so that part or all of enterprise assets are gradually sold to workers (see Lin Qingsong and others 1987). Enterprises can establish boards of directors accordingly. Aside from this, community govemments should create favorable conditions for the birth and growth of private enterprises. In doing this, community govemments should publicize the ideological as- pects and, more important, should formulate laws conceming private enter- prises as soon as possible. Note 1. The term "scissors differential" refers to the practice of setting procurement prices of agri- cultural products artificially low and sales prices of industrial consumer goods artificially high so as to generate additional industrial profits that could be used to expand investment. Part II Ownership and Related Issues Ownership is one of the principal areas in which China's TVP sector differs from the state sector, and forms of ownership vary tremendously within the TVP sector itself. Although the effect of ownership on enterprise behavior, efficiency, and overall development is complex and difficult to assess and is often mediated by other factors, different ownership systems are associated with significant differences in outcome. Many of the chapters in this volume deal with ownership in the context of the topics they cover. The three chap- ters in this part focus on ownership itself and on such closely related issues as entrepreneurship and differential administrative treatment on the basis of ownership. In chapter 7 Luo Xiaopeng undertakes a detailed historical analysis of own- ership in the TVP sectors of the four sample counties. He explains how the hierarchy of government administrative levels and the multilevel public own- ership system gave rise to a rigid ranking, or status stratification, of industrial enterprises, including TVPs as well as urban state and collective firms. His- torical, geographic, economic, and even political factors have combined to create substantially different ownership and status stratification systems in different regions and localities, and the four counties well illustrate the range of variation. Luo documents the causal factors that led to significant differences in own- ership and status stratification between Wuxi and Nanhai counties despite their roughly similar level of industrial development. In Wuxi economic power, ownership, and management authority are concentrated at the town- ship level of government. A much more dispersed pattern has emerged in Nanhai, where lower-level communities (villages and production teams) as well as individuals have a considerable degree of independence and economic power. In Shangrao and Jieshou, by contrast, until the late 1970s the TVP sector was tiny and "privileged," with relatively high wages. Luo next examines the differing impact of rural reform, especially the production responsibility system (PRS) in agriculture, in the four counties. 129 130 OWNERSHIP AND RELATED ISSUES In Nanhai rural reform facilitated the emergence of a diversified mixed- ownership system in which a vibrant private sector and various kinds of TVPS under joint community and private management existed alongside the mul- tilayered community enterprise sector. In Wuxi tight controls over wages and labor movements and prohibitive restrictions against sizable private en- terprises persisted and, after a brief period of struggle, consolidation and fur- ther strengthening of the traditional rigid, hierarchical TVP ownership and management system occurred. In Jieshou the PEts rapidly brought about the virtual collapse of the small existing community enterprise sector. This crisis stimulated a positive response by local authorities: the creation of an environ- ment in which private enterprises could thrive and the larger private firms could attach themselves to aind receive support from local and community governments. In Shangrao too the community enterprise sector was weak- ened, but it has been able to survive at a very inefficient and financially weak level. This paradoxically held down the overall level of TvP development in the county. The private sector threatens to overtake the ailing cormmunity enterprise sector. The last section looks at new situations and problems that have emerged since 1984-85. In Wuxi the industrial TVCE sector had its best year ever in 1985, but its financial performance declined sharply in 1986, and increasing problems with investment failures and market saturation for many goods emerged. The county's response was to rigidify the labor allocation system even further and to intensify the trend of TVPS' becoming more like state en- terprises in other ways as well. In Nanhai numerous mixed-ownership firms with rather clearly defined ownership rights and responsibilities have emerged. In Jieshou the problems stemming from unclear ownership of large private enterprises and personal involvement by community and county offi- cials have become increasingly evident. With regard to policy, Luo's main argument is that drastically different op- portunities and resources for rural industrialization have made it impossible to achieve the goal of rapid industrialization in every locality. Large gaps in rural industrialization and TVE' development have arisen between the most developed areas such as Nanhai and Wuxi and the more backward ones such as Jieshou and Shangrao. A crucial lesson is that urbanization must accom- pany industrialization. In chapter 8 Lin Qingsong analyzes the rapid recent growth and develop- ment of the private sector, focusing on the reasons for the phenomenon and on the problems the burgeoning private sector faces. He reviews prereform TVP development, which laid the foundation for the subsequent emergence of private enterprises. The development of private enterprises since the late 1970s can be divided into two distinct phases: the "nascent" stage, 1978-83, and the "flourishing" stage, since 1984. (This classification is the same as that developed by Du Haiyan ifor the TVP sector as a whole; see chapter 3.) Some benefits of the rapid development of private enterprises are outlined, includ- Ownership and Related Issues 131 ing the transfer of surplus labor out of agriculture, more spatial balance in TVP development, more appropriate subsectoral and industrial structure in the TVP sector, and efficiency-improving competitive pressure on community enterprises. Lin next analyzes the causes of the rapid growth of private enterprises, chief among which were policy changes and systemic reforms. Through the PRs, the opening of urban markets to all kinds of TVPS, and other measures, the government wiped out long-standing obstacles against the free establish- ment of firms by individuals. Lin goes on to demonstrate how private enter- prises obtain resources and factors of production, including financial capital, labor, land, mineral resources (for private mining firms), and technical per- sonnel. The last section deals with the problems private enterprises face during and as a result of their headlong development. Lin points out that the tiny scale of most private enterprises is a fundamental problem that will hamper efforts to improve their efficiency and argues that the political environment still impedes expansion of private enterprises beyond a certain size. Ideologi- cal ambiguity concerning sizable private firms leads to an uncertain policy environment, which blunts the firms' incentives to reinvest and encourages a short time horizon. Private enterprises often respond to their hostile admin- istrative and regulatory environment by devoting excessive resources to sales and procurement and by establishing close relationships with local and com- munity government leaders. Finally, ambiguity and uncertainty in policy lead to weak and unclear property relationships in private enterprises. The chapter closes with some policy suggestions. State policy toward private enterprises should be further relaxed and liberalized; government macroregulation of the private sector needs to be improved through a better legal framework, unified finance and accounting methods, new tax laws, and better credit policies; trade unions should be established for the private sector to protect workers' legitimate rights and improve state supervision; there should be more regular government monitoring of private enterprises' opera- tions and financial transactions, with severe punishment for tax evasion and other illegal activities; the government should fully respect the autonomy of private enterprises in choosing their ownership system and management prac- tices; and the voluntary transformation of private enterprises into joint-stock cooperatives should be encouraged. In chapter 9 William Byrd analyzes entrepreneurship, capital allocation and risk-bearing, and ownership, particularly the interactions among these crucial spheres. He starts by deriving a working definition of entrepreneurship suitable for the Chinese context and singling out two important entrepre- neurial activities, the creation of new enterprises and the development of new products. He then describes, with the aid of case studies, entrepreneur- ship in China's TVP sector and under different forms of enterprise ownership. The most salient characteristics of TVP entrepreneurship include alertness to 132 OWNERSHIP AND RELATED ISSUES unexploited profit opportunities; personal commitment and risk-taking; the importance of personal connections; the provision by entrepreneurs of a wide range of services in the absence of functioning markets; and the rootedness of most Chinese entrepreneurs in their home communities. Available empiri- cal evidence indicates that there is greater entrepreneurial activity in private enterprises than in community enterprises and in Wuxi than in other coun- ties. Byrd argues that, given the high degree of immobility among communi- ties, the traditional economic and commercial base of a locality is critical in determining the local supply of entrepreneurs. In this respect Wuxi and Nanhai are much better off than Jieshou, which in tum is better endowed than Shangrao, the only one of the four counties that may have an inade- quate supply of indigenous entrepreneurs. Byrd next analyzes the sources and allocation of capital. Three distinct sources of funds for TVPs are identified: household capital, community capital (which includes bank loans, since banks are viewed as highly responsive to community govemment authorities in their lending decisions), and enter- prise capital. Large community enterprises rely almost exclusively on the lat- ter two and, within community capital, increasingly on bank loans rather than on grants from community governments. Private enterprises mainly uti- lize household capital at the tirne of their establishment, but as they expand, the successful ones increasingly tap community capital, primarily bank loans. Risk-bearing is closely related to the mechanism for allocation of capital and is an important determinant of TVP performance. The traditional TVCE system in advanced areas such as Wuxi involves risk absorption by village and, espe- cially, township governments. In more backward areas community govern- ments are unable to play this role effectively, and the burden of risk falls by default on the local banking system. The chapter goes on to examine interrelationships among entrepreneur- ship, provision of capital, risk-bearing, ownership, and management. Four main models or systems are outlined. (1) The Wuxi system is characterized by communal ownership, risk absorption by the community government, and enterprise-based entrepreneurship and management. (2) Less successful com- munity enterprise systems also have communal ownership, but community governments are forced to take on a more entrepreneurial role (which often leads to costly failures). De facto risk absorption by the local banking system is the norm, and intervention by the community govemment in enterprise management is widespread. (3) "Small" private enterprise systems are charac- terized by the concentration of ownership, capital provision, risk-bearing, and management in the hands of the proprietors and, at most, a few other households. (4) "Large" private enterprise systems still have nominally pri- vate ownership, but their increasing reliance on community capital (in the form of bank loans) means that banks also bear some risk and that private management may be compromised. Byrd argues that whereas in systerns 1 and 3 the different elements fit together harmoniously and consistently, in Ownership and Related Issues 133 systems 2 and 4 they do not, which leads to various problems and inefficien- cies. The last part of the chapter looks at emerging trends and issues. The share of the TVP sector in Chinese industrial output has been rising steadily while that of the state sector has been falling-a trend that is likely to continue. Within the TVP sector the share of private enterprises is rising at the expense of community enterprises, and within the community enterprise sector, it is argued, the share of the top tier (township and town firms) is increasing or at least stable, whereas the share of the lower tiers (village and production team firms) is declining. Thus a "hollowing out" of the TVP sector seems to be occurring. Byrd argues that ownership trends can be understood in light of the analytical framework developed earlier in the chapter. The top tier of community enterprises enjoys better access to financial and other re- sources, greater ability to spread risks (in the more successful areas), and su- perior administrative connections. Private enterprises, however, may have greater flexibility and stronger incentives for proprietors. Smaller rural com- munities and those with inadequate resources may lack the advantages of ei- ther extreme of the TVP ownership spectrum. The consequences that greater independence for the local rural banking system and decentralization of bank ownership will have for community gov- ernments and for TVP performance are complex and hard to predict. These measures should be embarked on cautiously. As a general principle, banks should not be owned by community governments, and the boundaries of their operations should not coincide with those of rural communities. Byrd also suggests that community governments could issue bond-like financial instru- ments to mobilize capital and partially diversify risk. The chapter closes with a brief look at the problems of backward areas and private firms. The need to import entrepreneurial and other human talent into backward areas such as Shangrao is stressed. Byrd also argues that clarifi- cation and strengthening of the property rights of proprietors with respect to larger private enterprises are urgently needed. 7 Ownershiip and Status Stratification Luo Xiaoperig In 1958 the people's cornmune movement swept over rural China. The econ- omy and administration of a given area were put under a single organ, the commune, within which there were no independent property owners. The average size of a commune was about 4,000 households and 20,000 individu- als, but some were much larger. This utopian experiment did not make eco- nomic sense and led to disastrous consequences. Unlike Poland and Yugo- slavia, however, China did not beat a hasty retreat to family farming but instead launched a prolonged program to readjust property relationships. The first step was to restore order in the economy while retaining the com- mune framework. Between November 1961 and February 1962 the basic eco- nomic unit, or accounting unit, was changed first from the commune to the brigade and then to the next lower level, the production team, which nor- mally contained twenty to thirty households. The production team became relatively independent in its management of land, labor, and other com- monly owned property. ComLmunes and brigades, however, retained eco- nomic as well as administrative power because they still organized irrigation projects and other activities that production teams could not handle and be- cause they had the right to acquire and own property. Moreover, corrmmunes and brigades sometimes appropriated resources from production teams with- out compensation. In the nation as a whole, however, production teams became accepted as relatively independent organizers of production and consumption-"relatively" because they could not sell their land, their pro- duction was subject to the state plan, they could market only part of their output, and there were no clear-cut legal regulations to protect their property ftom infringement by their supervisory brigades and communes. The critical grain shortage after the Great Leap Forward of 1958 led to the en masse banishment of about 20 million people from urban centers to their home villages. From then on, change of domicile from rural areas to cities was severely restricted. Rural people were organized into production teams that were obliged to provide subsistence to their members, who lost 134 Ownership and Status Stratification 135 their independence and their freedom to change occupations. Skilled workers might, with the permission of the production team, seek gainful employment elsewhere, but their earnings had to be handed over and shared by all mem- bers of the team, and the worker was paid according to the work-points re- corded during his absence, at the uniform rate for the team. Failure to abide by this rule meant denial of grain rations or even worse. This policy was in- tended to ensure sufficient manpower for farming and to balance the remu- neration for farm and nonfarm.jobs in rural areas. The rural nonfarm sector, chiefly handicrafts, which had been all but wiped out during the Great Leap Forward, was gradually rehabilitated after 1965. Many workshops became part of the commune system and were owned by communes or brigades. Enterprises run by production teams were not re- garded as independent accounting units. Initially all these enterprises had an unofficial status. Only after 1970, when the national govemment decided to use them to promote the mechanization of farming, were they able to ob- tain bank loans and fiscal support. In 1976, with the founding of the National Rural Industry Administration, the position of rural firms was formally lega- lized. The years 1980-83 saw the transition from collective to family farming in China. The production team as an economic entity was almost phased out, communes became townships, and brigades were renamed villages. Since farmers regained the right to choose occupations and to engage in nonfarm activities, a good number of private business firms emerged, with the permission and support of the national govemment. Thus, in addition to community enterprises run by communes or brigades (TVCES), a new sector of private enterprise came into being. In theory, state enterprises belong to the whole people, township enter- prises belong to the people of the township, and so on down the line to pri- vate enterprises. Since each category of firms is attached to a unit in the ad- ministrative hierarchy which has a clearly defined rank, there is also a hierarchy of firms. Although the law is supposedly blind to status differences, govemment policy and administrative procedures treat firms differently ac- cording to their status. State enterprises and their employees enjoy the most privileges-preference in the allocation of material inputs and bank loans at low interest rates, lifetime employment, and so on. Rural community en- terprises, especially township enterprises, are treated preferentially with re- spect to taxes, interest rates, and fiscal assistance. Thus rank is more than a mere classification of ownership types. This chapter examines the TVP status stratification system, ownership sys- tems in different regions, and possible future trends. Each of the four counties studied (Wuxi, Nanhai, Jieshou, and Shangrao) has its own characteristics and represents a region and a model (table 7-1). Although they operate within the framework of central institutions and policies, TVPS in the four counties differ in status stratification and ownership systems. Describing and Table 7-1. Basic Data on the Sample Counties Fiscal revenue Gross output Gross output Industrial of local of agriculture of industry Per capita Per capita output of TVPs governments County and Population (millions of (milions of grain output savings (millions of (millions of year (thousands) yuan) (yuan) (jin) (yuan) yuan) yuan) Wuxi 1978 1,008 325.91 667.44 1,253.0 16.6 425.70 83.20 1980 1,009 277.30 17160727 986.2 35.1 814.45 91.62 1984 1,025 395.57 2,806.02 1,196.3 153.7 2,163.43 205.54 Jieshou 1978 508 67.52 39.09 963.0 6.0 8.81 5.11 1980 531 74.95 48.41 484.3 15.5 13.94 4.52 1984 561 127.06 150.19 716.7 46.7 74.16 7.06 Nanhai 1978 787 363.81 365.29 832.9 66.2 20.51 74.02 1980 804 526.21 467.23 982.8 134.2 98.89 78.39 1984 842 898.29 981.21 886.1 593.5 516.19 109.01 Shangrao 1978 645 91.31 23.39 587.6 6.7 12.57 6.44 1980 661 103.90 29.65 582.2 9.5 14.81 6.74 1984 695 147.49 70.98 681.7 34.0 23.70 8.56 Source: Information from fieldwork. Ownership and Status Stratification 137 analyzing these variations and how they emerged are the basic tasks of this chapter. Status Stratification before the PRS In 1962, when the production team became the accounting unit, communes and brigades in most areas deteriorated into mere administrative organs. They remained economic entities in name only, since they possessed almost no property of their own and their independent business operations were lim- ited. The emergence of community enterprises, however, created much new wealth for communes and brigades and gave rise to the idea of gradually ex- panding the public ownership sector through the development of community enterprises, primarily TVCEs. In 1979 the State Council affirmed that the de- velopment of TVCES would primarily serve agriculture but that it could also expand the collective economy of the commune and the brigade.' Opportunities for TVP industrial development were extremely unequal, however, owing to differing conditions across the country. For example, of the four counties in our study, Wuxi and Nanhai are in rich coastal areas close to traditional commercial centers and were among the earliest localities in China to import Western technology. In contrast, Jieshou has suffered from great disadvantages in its industrial development: it is overpopulated and poor, it is located in the hinterlands of the North China Plain, far from railway trunk lines and urban centers, and in the past it experienced many wars and natural calamities. Shangrao, in the northwestern part of Jiangxi Province, is slightly better off because of its relatively fertile farmland and long history of self-sufficient small farming. These varied backgrounds meant unequal opportunities for developing TVP industry and different economic structures under the three levels of community government-the commune, brigade, and team. Table 7-2 shows the great variation in the conditions for developing TVP industries, as indicated by the total number of industrial establishments and employees in each county. Although per capita income shows a positive cor- relation with the total number of establishments and especially with the number of employees in industrial community enterprises, the correlation be- tween the composition of the three-level community economy and the per capita income level is not as consistent. TVCES in Wuxi account for a higher proportion of aggregate income and aggregate assets than do those in Nanhai, yet Nanhai's per capita income is far higher (an average Y330 per year in 1979-81 compared with Y148 in Wuxi). Industrial Structure and Status Hierarchy Differences in local industrial structure may partly explain the status hierar- chy of the counties. Wuxi was the home of the first modern machinists in 138 LUO XIAOPENG Table 7-2. Structure of the Rural Economy, by County, 1980 Itm Wuxi Jieshou Nanhai Shangrao Per capita rural income from collective distribution (yuan) 146 48 350 58 Number of industrial enterprises, by level 2,306 355 2,784 350 Township 458 102 246 120 Village 1,848 54 575 230 Production team 0 199 1,963 0 Total rural labor force (thousands of workers) 497.2 220.0 357.0 186.0 Workers in industrial enterprises, by level (thousands) Township 43.2 4.6 22.5 14.3 Village 39.5 1.0 5.5 8.3 Production team 0 1.5 42.0 0 Source: Information from fieldwork. China. Retired workers who returned to their home villages in Wuxi pro- vided a tie with machinery-producing state enterprises in Shanghai, which led to close cooperation between the two places and the early development of capital goods production irn Wuxi. Since machinery firms require better management, larger-scale operations, and more capital than other enter- prises, in Wuxi the industrial TVP sector was dominated by TVCES. Nanhai's geographic advantages are similar to Wuxi's, but its history is dif- ferent. Community enterprises grew on the foundation of traditional handi- craft workshops and small-scale industries that traded mainly in nearby mar- kets. Higher personal income levels, the spread of traditional industrial skills, and greater variations in market demand called for a more diversified indus- trial structure, and commaunity enterprises in Nanhai gravitated toward the lower end-brigades and production teams. Cultivated land is scarce in Shangrao, averaging less than 1 mu per person, but it is fertile. The county, situated south of the Yangtze River, has less than 300 people per square kilometer, 25,410,000 mu of forestland, rich deposits of coal, phosphorus, and sulfur, and abundant hydropower. Industry is based on the exploitation of local resources. In 1985 Shangrao's TvPs employed 3,466 miners and 3,069 kiln wvorkers-more than half of the total number of workers in the county's industrial community enterprises. Jieshou is in an area of flat farmland devoid of industrial resources, and before 1979 the extremely bad performance of agriculture left the county with little surplus produce. In 1979 collectives distributed a mere 105 kilograms of grain per person, and even in 1980 the figure was only 168 kilograms. Any surplus was sold to the state under the unified purchasing system. For this Ownership and Status Stratification 139 reason, Jieshou had practically no industrial community enterprises before 1975 except briefly during the Great Leap Forward in 1958. In 1971 there was only one commune enterprise left, and its gross value of output was less than Y50,000. After 1975 there was a sudden upsurge in the number of TVCES as a result of administrative intervention and financial support from higher levels of government. The catchword of the time was, "Go all out developing three lines of industries" (farm machinery, kilns, and grain and oil process- ing). These industries were not market oriented but instead served local needs, and they were mostly dependent on funds and credit from higher-level authorities. Thus their number and scale were limited. The Origins of Community Enterprises The different ways in which community enterprises emerged in the four counties were more deeply influenced by their status stratification and owner- ship systems than by the geographic factors discussed above. THE WUXI MODEL. Wuxi has a highly integrated economy managed by the three levels of administrative organs. The township (formerly commune) is at the head, and the village (brigade) and the production team, although nominally independent in economic affairs, are actually under its strict con- trol. The historical background of this institutional setup is that for a long time Wuxi was under pressure to fulfill high farm production targets and high government grain purchasing quotas. To ensure that these requirements were met, the county banned production teams from engaging in nonfarm activi- ties. Production teams were compensated for this limitation on their activi- ties by payments from TVCES. The commune, to coordinate all nonfarm operations, controlled the ap- pointment of top managers of enterprises at both the commune and brigade levels. These managers could be shifted around much like middle-level man- agers in Western corporations. Managers of brigade firms (village enterprises) did not have to be members or natives of the brigade; in fact, brigade officials were often concurrently managers of the brigade's enterprises. TVCES located in the same commune or brigade avoided competing with each other. The commune was obliged to organize efforts to assist brigades in difficulties. And in decisions about industrial investment and labor allocation, no pains were spared to ensure relatively equal income levels for all brigades and production teams. Industrial investment and employment were thus an important means of promoting equity. The county, by transferring wage payments of industrial TVCE workers to production teams, was able to maintain the high-cost triple- cropping (wheat-rice-rice) system. This is clearly shown by the distribution of income from collective sources in Wuxi in 1980 (see table 7-3). The unique integration of the three levels of community institutions in the economic life of Wuxi was achieved through a special method of labor 140 LUO XIAOPENG Table 7-3. Average Income of Comnmune Members, Wuxi, 1980 Average income Percentage Source (Yuan) of total Fanring 8.78 6.0 Sidelines 45.39 31.1 Industry 91.89 62.9 Wages transferred 66.89 45.8 Profits transferred 24.11 16.5 Other 0.89 0.6 Total 146.06 100.0 Source: Information from fieldwork. compensation in the TVCE industrial sector. Factory workers were not paid directly by their employers but by the production teams from which they came. The production team pooled payments transferred from factories with income from farming and paid each member according to work-points earned during the year. The commune and brigade had a big say in how much money production teams distributed to mernbers or invested. With the growth of the industrial sector, the size of transferred wage payments and profits also grew. To prevent extreme inequity in personal income levels, the commune needed a sizable industrial sector to assimilate surplus labor, distribute profits, and assist bri- gades in setting up new ventures. All TVCES of the commune were obliged to subsidize farming, provide new jobs, and balance local income levels. Al- though TVCES and brigades were, to a degree, independently managed, the commune had the power to appoint the leading officials of both. These offi- cials therefore had to take the interests of the entire community into account in their operations and defer to the political interests of commune leaders. The dominance of the commune and the unique type of enterprise ownership that developed were characteristic of Wuxi. I have not yet found an appropriate paradigm for this peculiar system of property rights. It is heavily tinged with Chinese tradition and very different from the Western system under which the individual is the ultimate owner of property. The Wuxi model as it existed before the PRS might be compared with the traditional Chinese clan, in which three generations lived together. The "grandchildren" (the production teams) depended on their "parents" (the brigades) for a living. If the parents were unable to provide, the grand- children turned to their "grandparents," the commune. The lower levels were nominally independent in financial affairs, but the "eldest generation" had the final say about disposal of property and distribution of income. The source of this power was not only the social status of the communes but also the property they controlled and its impact on the interests of the lower levels. In 1981 coimmune enterprises in Wuxi turned over to produc- Ownership and Status Stratification 141 tion teams wage eamings of Y23.1 million, and brigade enterprises trans- ferred another Y41.3 million. Together these totaled Y64.4 million, or about Y60 per person-one-fourth of the total annual income per capita of the county's rural population. In some communes the proportion was much higher. From the early 1970s income generated by commune enterprises con- tinually increased in proportion to total income, and there is reason to be- lieve that commune enterprises attracted better-quality workers as well. These developments further consolidated the political base of the communes. NANHAI. Nanhai County's community enterprise economy is characterized by relatively dispersed management with villages and production teams as the basic economic units. The township is to a large extent regarded as an admin- istrative organ, and the three levels of the economy are far less integrated than in Wuxi. There are some historical explanations for this pattern. The Pearl River Delta economy was traditionally more diversified and commer- cialized than that of Wuxi. Farmers involved in cash crops and nonagricul- tural activities had to rely for their grain supply on the market and even on imports. After 1949, and especially after the commune system was set up, Nanhai was required to produce grain for its own consumption and for sales to the state, but the pressure was not as heavy as elsewhere because the county had to supply fruit, fish, and vegetables to Hong Kong and Macao. Moreover, government policy in Guangdong Province was always more flex- ible than in the northern provinces. These factors led to higher local income levels. In addition, the county is adjacent to Guangzhou and Foshan and boasts a few big towns of its own. Its greater urbanization no doubt contrib- uted to the diversification of its economy. Each of the three levels developed its own industrial sector, as the follow- ing figures show. Percentage of total gross sales of community enterprises Level 1973 1982 Commune (township) 12.9 19.3 Brigade (village) 9.7 21.7 Production team 77.4 59.0 The aggregate income and wealth of the county was skewed toward the lowest level. Commune enterprises, although well developed in comparison with the Chinese average, were clustered around the towns and were not as integrated with the production teams as those in Wuxi. For example, in Nan- hai commune enterprises did not have to turn workers' earnings over to pro- duction teams for final distribution (see table 7-4). Three factors account for this. * Many of the workers were from urban centers. In 1980, of 80,000 per- sons employed by TVCES, about 55,000 were from rural communities and Table 7-4. Income Structure of Production Teams, Nanhai, 1978-82 (millions of yuan; percentage of total distributed collective income) 1978 1979 1980 1981 1982 item Amount Share Amount Share Amount S'are Amount Share Amount Share Total production team z income' 250.448 - 206.268 - 338.696 - 396.856 - 685.570 - Total distributed collective income 177.437 100.0 142.359 100.0 229.101 100.0 238.721 100.0 449.830 100.0 From brigades 39.681 22.4 38.473 27.0 31.880 13.9 34.280 14.4 24.630 5.5 From communes 0.0 0.0 0,008 0.0 0.032 0.0 0.694 0.3 0.010 0.0 - Not applicable. a. Includes income from private sideline activities. Source: Information from fieldwork. Ownership and Status Stratification 143 the rest were mostly local urban dwellers. Urban residents' wages were paid to them directly. • The average incomes of production team members were not much lower than the average incomes of commune enterprise workers, thanks to high farm incomes and the wages of team members working in TVCES. In 1980 commune enterprise wages averaged Y640 as against the aver- age rural per capita income of Y559, a difference of Y81. When income from sidelines and the lower living costs on farms are taken into consid- eration, there was probably a rough equilibrium, even with freedom of choice of occupation. In fact, many farmers earned more than average workers in commune enterprises, which relieved pressures on com- munes to transfer workers' earnings to farmers. * The commune imposed a surtax in kind to supply grain rations to its factory employees who were rural residents so that they would not have to go home for grain distribution. Brigade enterprises were similar to those in Wuxi in that their workers were largely rural residents and had to obtain grain supplies from the teams to which they belonged. These firms had to see to it that earnings in the facto- ries and on farms were approximately equal. Thus a considerable portion of wages was transferred to production teams for distribution-86.4 percent in 1980. It was not 100 percent because some teams had such high incomes that they could forgo earnings from factories and because grain supplies in some cash crop areas were provided by the state. For example, brigade enter- prises in Jiujiang township transferred only 21 percent of their payroll to pro- duction teams because the area specialized in fresh-water aquaculture, and fish raisers obtained grain supplies as urban residents did instead of from their production teams. In Wuxi, as people got richer, more wealth became concentrated at the commune level. But in Nanhai the development of community enterprises led to a relatively dispersed pattern of distribution of wealth and power and to a high degree of autonomy and coherence in some brigades and teams. Since natural villages in southern China are mostly populated by members of one clan, a strong clan culture and consciousness survives to this day, and the development of TvPs has no doubt reinforced this traditional feature. SHANGRAO. TVCES in Shangrao County had a large share in aggregate in- come, mainly for two reasons. First, incomes at the production team level were so low as to make incomes from TVCES seem relatively high. In 1980, of the 4,867 teams in the county, 86 percent (4,180 teams) had average per capita incomes from collective distribution of less than Y80. Of the total, 61 percent (2,974 teams) were below Y60, 41 percent (2,006 teams) were below Y50, and 20 percent (995 teams) were below Y40. No team had an 144 LUO XIAOPENG average per capita income level of more than Y400; only 2 percent, or 102 teams, were above Y100. Second, proximity to Shangrao city created demand for services, construc- tion, building materials, and so on. Before the PRS, the TVCE sector in Shangrao County was doing quite well, as the figures for 1978 illustrate. Aggregate gross sales Y14,720,000 After-tax profits Y3,530,000 Taxes on business income Y600,000 Industrial and commercial taxes Y130,000 Rate of profit on sales (percent) 28.9 Only Y1,060,000 of after-tax profits, however, came from industrial TVCES, whereas the service sector contributed more than Y2 million, or Y6587 per employee compared with Y149 in the industrial production sector. It is worth noting that in Shangrao, in contrast to Wuxi, most of the earn- ings of workers were not transferred to production teams for distribution. With the growth of the industrial TVCE sector, the number of workers multi- plied, yet the amount of wage earnings transferred actually dropped, from Y1,570,000 in 1978 to Y540,000 in 1981. The extremely low income levels on farms may account for this. TVCE workers were in a good bargaining posi- tion, as the sector generated the lion's share of commune revenues. A regular wage system was used to give incentives to workers who had a strong impact on their firms' performance. This practice spread as an increasing number of people sought jobs in the industrial sector. The difference in income be- tween TVCE workers ancl farmers was more than threefold, and thus the for- mer constituted a privileged group. In short, the resources and local markets of Shangrao allowed brigades and especially communes, with the aid of monopolistic practices, to develop a profitable and growing industrial sector and a privileged group of industrial workers. This weakened the economic ties among the three levels of rural communities. Although a certain amount of industrial development took place, the elite part of the population was always a small minority. This is the essential difference between the ownership systems of Shangrao and Wuxi. JIESHOU. The industrial TVCE sector in Jieshou was similar to that in Shangrao with regard to status stratification and ownership. First, the sector's proportion of income looked relatively large only because of the extremely low farm income levels. The two counties were approximately the same in population, but the average income on farms in Jieshou was even lower than in Shangrao, and the number of industrial workers and the aggregate value of output were insignificant. Second, although Jieshou's industries were lim- ited in scale, they too received privileged treatment, and wage eamings were not transferred to production teams for distribution. The more these indus- Ownership and Status Stratification 145 tries grew, the more they were alienated from the farming population, as transferred wage earnings shrank. The main difference between the TVCE industrial sectors in these two counties was that in Jieshou profitability was even lower, so the underlying strength of the sector was weaker. Although tax and wage rates were much lower than in Shangrao, most TVCES in Jieshou made little or no money. In 1980, of the 102 commune enterprises, 32 employed fewer than ten workers. The firms' aggregate output value was Y450,000, an average of Y14,000 per establishment and Y1,500 per worker. This meant that average net output value per employee was only about Y300, barely enough to pay wages, and consequently many of these community-run industries went out of business. The mortality rate was especially high at the commune level. There were forty fewer commune enterprises in 1981 than in 1980-a loss of 260 jobs, or 13.2 percent of total industrial commune enterprise employment in the county. Such fragility was not seen in the other three counties. The Impact of the PRS The PRS brought tremendous changes to rural China. The land tiller now by contract has the use-right to a piece of land and thereby the right to dis- pose of his own time. Moreover, he has the right to sell farm produce and to acquire productive assets with the proceeds, and his right to seek nonfarm employment has been restored to varying degrees. Such changes, however, emerged from localities' expediencies for coping with their economic prob- lems rather than from conscious efforts to win legal rights for individuals. The economic rights and the degree of freedom of rural people vary in accor- dance with patterns formed in the past. That is why status stratification and ownership in the sample counties retained their palpable distinctions after the implementation of the PRS. Jieshou. Private Firms under Collective Auspices In 1979, after a bad harvest, average farm income in Jieshou County fell to a record low of Y28 plus about 100 kilograms of grain (excluding incomes from sideline occupations). The persistent agricultural crisis went from bad to worse. In the towns small state and collective enterprises, which alto- gether had some 5,000 employees, had little capacity for absorbing the natu- ral increase in job seekers. The troublesome situation was exacerbated by the return of great numbers of people who had been sent to the countryside dur- ing the Cultural Revolution period (1966-76) but now by government order came home looking for jobs. The tremendous pressure on the local govern- ment forced it to take extraordinary measures, and as a result the county was among the earliest to institute the PRS (in 1981) and to accelerate the devel- opment of rural industry through an unusual program. 146 LUO XIAOPENG In 1980 the county's Communist party committee announced temporary regulations covering farming, commerce, and industry. The measures were considered drastic at the time. The policy statement called for expanding the scope of household responsibilLity in farm areas, facilitating the marketing of local industrial goods, and giving decisionmaking powers in purchasing and marketing to public commercial agencies. Conceming the industrial commu- nity enterprise sector, article III of the regulations reads: To hasten the development of TVCE industries and encourage extensive cooperation between enterprises, barriers between various trades, owner- ship forms, and administrative areas are to be eliminated. All firrns may associate themselves in various ways, and all businesses should be con- ducted in compliance with economic laws. Government workers, urban, and rural residents are free to invest in collective industries run by the county and in TVCES. Termis may be prescribed by contracts with regard to shares and dividends. Jobs may be provided through negotiations. Sav- ings of enterprises at all levels may be invested elsewhere as shares of joint operations, obtaining dividends therefrom. State and collective enter- prises and TVCES may, through compensation trade, attract investment and expand the scope of operations. These changes were intended to make a breakthrough in the statutory re- strictions of the county. Along with related ad hoc decisions, this new policy led to a significant result-the emergence of a number of special types of TVPS. Some typical examples follow. THE JIESHOU COUNTY COLLOIDAL PRODUCTS FACTORY. In 1980 a farmer ap- plied for permission to establish a factory in the county seat. A private person was then not allowed by law to use land in town, nor could he open an ac- count for his firm with the bank. The law was sidestepped by designating the business a TvP at the coimmune level under the direct control of the county. This status made the firm eligible for more liberal bank loans than ordinary private enterprises and gave it exemption from taxes. In addition, it obtained easier access to lowv-cost inputs through the govemment channels that served rural industries, advantages in obtaining fiscal assistance from the govemment (such as a Y5,000 interest-free loan for working capital), better credit standing because of being able to use the county's name, and the privi- lege of selecting a plant site anywhere in the county. At the time of our sur- vey the firm was using this last: privilege to relocate its plant from the county seat to a township that offered a better environment for business. We were amazed to find that the firm had just bought 1 mu of land at the new site for Y10,000, although China's constitution forbids private ownership of land. The purchase would have been impossible had the firm not been designated a township enterprise. In return for all these privileges, the enterprise pays Ownership and Status Stratification 147 1 percent of its gross sales as a management fee to the county business admin- istration every year. THE SHUGUANG FOODSTUFFS COMPANY. The predecessor of this firm was a small-scale partnership that made food products. In 1984 it planned to start producing canned beef with support from the local district administra- tion, which was the nominal legal owner for registration purposes but in- vested nothing whatsoever. The firm was actually owned by four individuals who pooled Y20,000 and borrowed Y320,000 from the bank to start with. The partners had full management power. In a short time they retrieved their original investment of Y20,000. Since then, the ownership status of the firm has been ambiguous, but the founder has consistently claimed the enterprise as his own. The district administration-the nominal owner-has never in- terfered with the disposal of the firm's income. Top management was paid no more than Y150 a month (not much more than ordinary workers), but the salesmen, who were paid on a commission basis, could make about Y10,000 a year. Business was so brisk that in its first year the firm recorded a profit of Y180,000 and in the second a staggering Y400,000, tax exempt. These nominally public but actually private firms are not exceptional in Jieshou. (See chapter 9 for another interesting example, that of an elastic ribbon factory. Numerous other cases were encountered in fieldwork.) A check by a county organ in 1985 revealed that eighty township enterprises, or 73 percent of the total, were carrying the "wrong registration," and these were dropped from the classification. But the real reason that the firms lost township enterprise status was not so much that their claim was false as that they were not regularly organized, were unstable, and had no political back- ing. Large and profitable firms such as the Jieshou County Colloidal Products Factory and the Shuguang Foodstuffs Company, which also used false identi- ties, remained township enterprises. Some of the other firms that retained their status were extremely small. For instance, a furniture producer with only two employees remained on the township enterprise list, although it was much like a "ma and pa" store. In spite of such confusion, Jieshou's bold poli- cies achieved what county authorities had intended-the development of a number of industrial enterprises that could not have emerged under the old institutional rules. Before 1984 the position of the central government on private enterprises, especially on their right to hire workers, was not clear. But after 1984 the central government openly supported the rural private industrial sector, and conditions for its development improved. Individual proprietorships and partnerships mushroomed in the towns and countryside of Jieshou, as in the rest of China. These were not enterprises in the regular sense of the word; most were handicraft producers that did not need even a plant or industrial 148 LUO XIAOPENG equipment. The state of these firms in jieshou as of the end of 1984 was as follows. Partnership Individual Number of firms 1,151 4,651 As percentage of total TVP sector 17.4 70.0 Number of employees 5,034 10,494 As percentage of total TVPr sector 16.0 33.4 The two sectors combined generated a total income of Y35 million, or 40.6 percent of the aggregate income of the industrial TVP sector of the county. These figures are not quite reliable, as private enterprises tend to understate their incomes so as to pay less tax, and tax-free community enterprises tend to inflate their incomes so that managers can get promotions. Yet they do indicate the profound changes in status stratification and ownership composi- tion that followed the introduction of the PRS. Conventional community en- terprises are no longer predominant in the Jieshou industrial TVP sector as a whole, and firms that are basically private in nature have become impor- tant. The various ownership forms are now so mixed that the true identity of establishments cannot be ascertained from their nominal status. Owner- ship of many is actually private or undetermined, although all are attached to some administrative organ to facilitate their operations. Wuxi: The Continued Predominance of TVCEs If Jieshou is typical of the poverty-stricken areas in which community enter- prises failed to withstand the shock wave of the PRS and tumed private, were leased out to individuals, or simply closed, Wuxi illustrates the opposite ten- dency. Its TVCEs did undergo a shock, but the sector recovered and became stronger than ever. Township govemments now possess a greater proportion of assets than all other sectors, and their control over other economic entities is even greater than in l-he palst. The first benefit the industrial TVCE sector gained from the PRS was the abolition of wage transfers from factories to production teams, which elimi- nated a large disincentive for factory workers. Under the old system a work- er's eamings were not determined by performance in the factory but by the income level of the production team to which he or she belonged. The conse- quences could be ludicrous; for example, a skilled master might be paid less than his apprentice if the master's team was poorer. In many other areas as well., wage transfers were gradually phased out after the PRS was implemented. This provided greater incentives for the workers but alienated the townships from the production teams and workers from other members of their teams. In Wuxi the problem was not as severe as else- where. The county continued to subsidize high-cost grain production and took measures to alleviate the tension between factories and production Ownership and Status Stratification 149 teams that arose from inequity of incomes. These measures included minimiz- ing income inequalities among production teams (for example, no production team was allowed to engage in factory production of its own); recruiting fac- tory workers and transferring profit disproportionately so as to subsidize poorer production teams; helping poorer villages set up their own factories; and providing liberal bonuses for technicians and skilled workers. In addition to improving worker incentives, the PRS also resolved some long-standing conflicts between TVCEs and production teams. County offi- cials, recalling life in the old days, stated that before the PRs these conflicts had been rather sharp. When extra compensation was given directly to fac- tory personnel, production teams were displeased because they regarded the labor concemed as their own assets. Some production teams even went to the length of crediting more work-points to members engaged in agricultural work to reduce the average value of each work-point so that workers who were away at factories were rewarded less. This problem vanished with the advent of the PRS. The piece-rate wage system could be introduced in TVCES only after the PRS was instituted. The second great benefit that the PRs gave the county's industrial TVCE sector was soaring farm productivity and a great increase in farm incomes. This was most evident in 1983, when average per capita rural income in the county rose to a new high of Y494, up from Y375 in 1982. Growth temporar- ily alleviated the pressure to equalize personal incomes after the suspension of transfer wage payments. Furthermore, the rise in farm production across the nation helped relieve Wuxi from its obligation to wrest grain from the land through the uneconomical triple-cropping system, and much less labor input was required in land tilling. The brisk market for manufactured prod- ucts helped lead the expansion of TVCE industries, which provided more jobs. In 1983 and 1984 more than 70,000 persons were added to factory payrolls, a 40 percent rise. The rapid expansion of employment in the industrial TVCE sector sustained the continuing rise in per capita incomes in the county and made it easier to narrow income gaps among farm households through alloca- tion of factory jobs. County authorities had not anticipated such a phenomenal and effortless expansion of industry. To prevent income gaps between the farm and nonfarm populations from widening, the county nullified the ban on produc- tion team factories. In 1984, when private enterprises mushroomed across China, the county government declared its support for the private sector and encouraged all production teams to engage in industrial activities. Mean- while, TVCEs experimented with a bonus plan keyed to profits and output value to provide stronger incentives. The county's industrial sector was put on "five wheels," with private and production team firms added to the town- ship, village, and county levels. But it was soon discovered that these "five wheels" did not keep the same pace. Private firms and production team enterprises challenged the suprem- 150 LUO XIAOPENG acy of TVCES. They competed in hiring technical and marketing personnel, and their higher remuneration for workers broke the old balance of incomes in the county and caused wider gaps in personal incomes. The impact was so tremendous that many skilled and competent personnel demanded to leave TVCES to set up their own firms, and labor costs were in danger of getting out of control. But the biggest headache for county authorities was the con- siderable number of bogus production team enterprises that were actually pri- vate. These firms, nominally leased to private individuals, could enjoy tax exemptions and other preferential treatment without contributing anly sav- ings to the collective, and they were free to compete with TVCES for markets, personnel, and raw materials. In 1985 Wuxi authorities took steps to reestablish order and consolidate the position of TVCES. The principal measures included * Penalties for skilled persons who left TVCEs for greener pastures. The most effective penalty was that members of the person's family were perma- nently barred from jobs in TVCES. * A thorough check on production team enterprises, partnerships, and individual enterprises. The status of firms was verified, and bogus community enter- prises were cleared out. The criteria for a community enterprise were that the firm's initial investment must come from the savings of a com- munity body, its site must belong to a community, its production equip- ment must be the property of a community, and its distribution of prof- its must comply with regulations. These rules require that deductions be made for a production development fund, a common insurance fund, and a depreciation allowance at the prescribed rates and that profits be remitted to the proper authorities. About 500 firms were removed from the category of production team enterprises, and village control over production team enterprises-which generally cannot even have their own bank accounts-wais reinforced. Finally, the number of partner- ships and individual firms that compete with TVCES was restricted by requiring them to have permission from the village government to regis- ter, although national law requires only the permission of the produc- tion team. * Restoration of control over TVCE wage levels. A new remuneration scheme for factory managers and workers allowed personal incomes to differ under the piece-rate system but tied the average wage level to profits and output value. Moreover, average wage rates were not al- lowed to diverge too much among firms. Enterprise managers' pay was linked to the average wage in their factories; twice the average wage was normal, and three times was the maximum. After the implementation of these measures, production team enterprises declined to fewer than 1,000 by 1985. Partnerships and individual firms de- Ownership and Status Stratification 151 creased as well. Only in sectors such as commerce, transport, and services, which do not compete with industrial TVCES, have private enterprises been growing rapidly. Under the PRS the industrial TVCE sector in Wuxi initially suffered a set- back, but its underlying economic strength and strong county policies al- lowed it to regain its dominant position. This is borne out by table 7-5 on the sources of household income; in 1983 private sources were most impor- tant, but two years later income from community sources had drawn ahead. The breakdown of fixed assets gives a similar picture. In 1986 the county's rural areas had fixed assets, excluding land, of Y1,208 million, and the share of township organs was 49.5 percent or Y599 million-2.1 times above the pre-PRS value. The share of assets owned by townships and villages in com- mon was 39.9 percent, or Y483 million-2.75 times above the level of 1982. And the share of production teams was 7.6 percent, or Y92 million, only a 9 percent increase. Finally, the share of household-operated firms was a mere 2.9 percent of the total, or Y35 million. Nanhai: Production Team Enterprises and Their Turn to Private Operation Before 1979 Nanhai held the same ideological view as Wuxi with regard to production team enterprises, and to ensure adequate manpower for farming county authorities had even ruled that TVCES could not employ more than 10-15 percent of the county's labor force. But in practice the authorities were flexible. They gave quiet approval to some production team enterprises and tolerated the spontaneous emergence of private enterprises. In fact, this has been a significant characteristic of rural economic policy in Guangdong Prov- ince as a whole. In 1980 the county government noted that in one commune that had al- lowed every production team to set up firms, local farmers' incomes had in- Table 7-5. Sources of Rural Household Income, Wuxi, 1979-86 (percent) Community Household Nonproductive Year economy' Partnerships activities sources 1979 62.4 0 25.6 12.0 1980 60.4 0 25.4 14.2 1981 60.7 0 28.1 11.2 1982 60.7 0 31.4 8.0 1983 26.7 0 66.1 7.2 1984 39.8 0 53.6 6.6 1985 47.3 0.9 46.2 5.6 1986 47.6 1.2 44.8 6.4 a. Includes both distributed collective income and wage income from community enter- prises. Source: Information from fieldwork. 152 LUO XIAOPENG creased rapidly. This approachL was promoted across the county under the slo- gan "let three horses draw the carriage." In 1981, 71 percent of the county's production teams startecl their own industrial enterprises, and the income of the three levels' industrial sector jumped to Y572.8 million, up 1.83 times from 1978. Income generated by industrial production team enterprises was Y166.43 million, and income per person from collective sources rose to twice the 1978 level, or Y374, of which industrial and sideline occupations ac- counted for 57 percent. Such impressive successes caused county authorities to hesitate in making the transition from collective to family farming, and even many of the farmers were afraid that they would lose income. Nevertheless, Nanhai, like Wuxi, went over to the PRs in 1983, and farm production increased considerably. Average personal income rose to Y760, double the figure in Wuxi. In the industrial sector great numbers of individual firms and partnerships were set up, and many production team enterprises were leased or sold to private management. In that one year the aggregate value of fixed assets of production teams dropped from Y158 million to Y134 million, whereas the value of privately owned fixed assets rose from Y16 mil- lion to Y104 million. The shock wave of the new system hit the TVCE sector of Nanhai hard, as it had in Wuxi. Skilled and competent personnel, even those at the top management level, wanted to leave and run businesses of their own. County authorities, in trying to stem the tide, were compelled to make rules like those of their counterparts in Wuxi, but these turnLed out to be ineffective because of the different organizational structure of the indus- trial sector and the more flexible attitude of county leaders. In the end, a relatively open job market caine into being, and the private sector evolved into a significant branch of the TVP industrial sector. There are several explanations for the more flexible attitude on the part of Nanhai authorities. * Community enterprises, especially township enterprises, were not im- portant sources of fiscal revenue for the county government or impor- tant job providers and sources of income for the rural population. When these enterprises were adversely affected, the impact was not as strong as in Wuxi (see table 7-6). * In contrast to the situation in Wuxi, factory workers in Nanhai were mostly urban dwellers whose wages were not affected by the production teams' income levels, and so the PRS did not create as much of a shock for township enterprises to absorb. * The remuneration for administrators and managers differed greatly be- tween the two counties. In Wuxi village administrators and managers of township enterprises belong to the same grade. Their salaries are de- cided by township authorities in accordance with county government regulations requiring that relative equity be observed, with the average income of workers in local TVCES the most important criterion. In Nan- Ownership and Status Stratification 153 hai no such relationship existed between township enterprise managers, who are professionals hired by the township government, and village administrators, who by convention must be people from the locality and whose salaries are determined by the village's income level. The personal interests of the village administrators were not affected, and they had no strong motives for impeding the growth of the private sec- tor, especially the shift of production team enterprises to private man- agement. For the county as a whole, the trend toward private enterprises was beneficial. Private enterprises-most of which were handicraft work- shops or contractors that supplied spare parts to larger factories- complemented rather than competed with county and township enter- prises, and they provided jobs, which the industrial TVCF sector was unable to do. The thriving private sector greatly expanded the tax base.2 As the sector grew, so did bank deposits, and banks in turn had more funds to lend to state and community enterprises. In 1985, the Nanhai industrial TVP sector recorded an aggregate sales in- come of Y1,782 million, which was 7.5 times the total in 1978. The break- down was roughly as follows. Share (percent) Township enterprises 38 Village enterprises 31 Production team enterprises 16 Partnerships 6 Individual households 9 The last two sectors employed more than 46,000 people, as many as the township enterprise sector. Private enterprises paid taxes of Y16.83 million, or 12.6 percent of the total fiscal revenue of the county in 1985. Fixed assets owned by rural individuals rose to Y128.43 million, 6.9 times the figure for 1983, higher than the figure for production teams (Y125.63 million, not in- cluding land), and almost equal to the figure for village enterprises. Consider- Table 7-6. Contribution of Industnial Township Enterprises, Wuxi and Nanhai, 1983 (percent) Item Wuxi Nanhai Taxes as share of county's total revenue 34.9 16.1 Workers as share of rural population 13.6 2.9 Transferred eamings of workers as share of total distributed eamings of production teams 41.3 0 Source: Information from fieldwork. 154 LUO XIAOPENG ing that before 1984 banks were not allowed to make loans to private enter- prises, the rate of growth was indeed extraordinary. Most of the remaining production team enterprises have been leased or contracted out to private individuals. This trend is likely to continue. The private sector, including production team enterprises under private manage- ment, will stand on an equal footing in economic strength with the industrial community enterprise sector. Average per capita personal income in Nanhai was Y1,029 in 1985, of which only Y114 (11.1 percent) came from TVCES and the rest came from private sources and production team enterprises. This structure and trend are in sharp contrast with the pattern in Wuxi. The two counties are both well developed, yet their status stratification and ownership structures have be- come immensely different. This is a subject that should be studied in depth. Shangrao: TVCEs in a Jam Before the PRS, there was no comparison between Shangrao's economy and that of developed areas such as Wuxi and Nanhai, but the county was consid- erably better off than Jieshou. The gap between the two poorer counties was at its widest in 1979, when Jieshou was hit by serious natural calamities. In that year average per capita distributed collective income in rural Jies]hou hit a record low of Y28 plus 105 kilograms of grain. But for Shangrao 1979 was a good harvest year. Average per capita distributed collective income was about Y70 per person, and the grain ration averaged nearly 250 kilograms per person. (The above ligures exclude income from household sideline oper- ations, which was equal to about one-third of the distributed income.) Gov- emrnment revenues and expenditures for the two counties also differed widely in that year, as is shown in table 7-7. The biggest difference was in profit remittances from TVCES. For Shangrao the amount was almost comparable to the fiscal revenue and was 5.2 times the figure for Jieshou. The fiscal re- sources of some townships were more than adequate. After the PRS was implemented, Shangrao's TVCES, especially township en- terprises, suffered from deteriorating financial performance. Net profits of township enterprises dropped sharply, from Y2.08 million in 1983 to Y1.3 million in 1984 and Y622,000 in 1985. The number of industrial TVCES fell from 410 in 1979 to 340 in 1985. Of the total, industrial township enterprises went from 120 to 118-practically no change-whereas village enterprises declined from 290 to 222. Inclustrial employment of TVCES fell from 11,358 to 10,294. But the value of fixed assets (at original cost) of township enter- prises almost tripled, from Y5.78 million in 1979 to Y15.75 million in 1985, while the total payroll increased from Y7,390,000 to Y8,420, 000 and average wages rose from Y590 to Y749. These figures for profits and costs indicate a severe decline in economic efficiency. Rising labor costs may account for part of the drop in profits, but invest- Ownership and Status Stratification 155 Table 7-7. Public Finance, Shangrao and Jieshou, 1979 Ratio, Item Shangrao Jieshou Shangraodieshou Fiscal revenue (millions of yuan) 6.20 4.03 1.54 Per capita (yuan) 9.4 7.8 1.2 Expenditures (millions of yuan) 10.30 10.59 0.97 Per capita (yuan) 15.6 20.4 0.76 Profits from lVCEs (millions of yuan) 4.68 0.90 5.2 Per capita (yuan) 7.1 1.7 4.1 Profits from township enterprises alone (millions of yuan) 2.84 0.57 5.0 Per capita (yuan) 4.3 1.1 3.9 Source: Information from fieldwork. ment failures and management weaknesses were more serious; almost Y10 million of new investment in township enterprises showed no return. TVCES confronted a number of difficulties. First, higher income levels in rural areas following implementation of the PRS forced factory labor costs to increase. Second, since farmers were now free to change occupations and to set up their own businesses, many profitable community enterprises in transport and services went private. Township enterprises in these two sectors employed 1,804 persons in 1978 but only 445 in 1983. Some construction firms, al- though nominally not privatized, were in a very strong bargaining position with respect to township authorities, so that when they concealed their ac- tual profits there was not much the townships could do. Third, townships did invest heavily in industry, but unwise decisions nullified their efforts. For instance, a food additive factory was set up at tremendous cost, but the prod- uct found no market. Fourth, competition from the increasing numbers of private enterprises, especially in some labor-intensive industries such as coal mining, caused the township enterprises to operate at a loss. Fifth, township governments, in trying to solve their own financial problems, forced town- ship enterprises to turn over to them more of their profits and even some of their depreciation funds. This amounted to "drying the pond to get the fish." Finally, considerable embezzlement of public funds and corruption in hiring occurred. The failures in the Shangrao industrial community enterprise sector may be justly attributed to mistakes on the part of county and township authori- ties. But these were inevitable for a first generation of grass-roots managers whose experience had been in a dispersed and closed economy. Perhaps their biggest mistake was their failure to recognize the inevitability of problems. The fatal weakness of the industrial TVP sector in Shangrao was the ab- sence of a sizable urban center as the core. This becomes apparent when 156 LUO XIAOPENG Shangrao is compared with Jieshou, which has such a core. There are thirty- four industrial firms in Jieshou town. If these are ignored, the TVCE industrial sectors in the two counties are similar in number of firms, average value of annual output, and average rate of profit (table 7-8); they differ only in the average capital stock per firm,. In Jieshou town many township enterprises are actually private enterprises or are leased to private management, and they are more efficient in their use of funds. The experience of Shangrao also testi- fies to the importance of towns for developing TVPS. In the areas of Shangrao that we visited, one town, Shaxi, which is served by a railway line, stood out as a profit maker. In 1985 the profits of its industrial township enterprises accounted for one-fifth of the profits of the entire industrial township enter- prise sector of the county. Shaxi historically had been a handicraft and com- mercial center, and it is still the busiest town in the county. One town, how- ever, cannot industrialize an entire county. Summary The economic rights ancl personal incomes of farmers advanced tremendously when collective farming was phased out. But complex variations and prob- lems conceming status and ownership patterns in the rural industrial sector have arisen. The old pattern in Jieshou was completely disrupted, and com- munity enterprises survive only in name. The new de facto private enterprises Table 7-8. Main Financial Indicators for Industrial Township Enterprises, Shangrao and Jieshou, 1985 Shangrao Jieshou Total Average per Total Average per (millions enterprise (millions enterprise Item of yuan) (yuan) of yuan) (yuan) Gross sales revenues 13.35 113,100 14.36 131,700 Gross profits of profit-making firms 1.08 9,100 1.30 11,900 Total losses of loss-making firms 0.28 2,300 0 0 Net total profits 0.80 6,800 1.30 11,900 Original value of fixed assets 13.45 114,000 6.70 61,500 Net value of fixed assets 9.61 81,400 6.24 57,200 Value of circulating assets 2.55 19,100 0.98 9,000 GVIO 1980 prices 11.62 98,500 17.11 157,000 Current prices 13.37 113,300 18.73 172,000 Note: GVIO, gross value of industrial output. The number of enterprises sampled was 118 in Shangrao (21 were loss-makers) and 109 in Jieshou (number of loss-makers not available). Source: Information from fieldwork. Ownership and Status Stratification 157 can now select the administrative organ to which they wish to be attached, a change that assists the firms in organizing productive factors but that has caused confusion about property ownership and unfair treatment of different enterprises. In the other three counties farmers have not gained such a degree of independence in running industrial activities; the three-level institutional structure has not changed, even though a new stratum has been added. There are, however, important differences in the industrial organization of these counties. In both Wuxi and Shangrao the private sector is in conflict with the TVCE sector, but the TVCE sector predominates in Wuxi and the pri- vate sector in Shangrao. TVCES in Nanhai are not adversely affected by the growing private sector because of the cushion provided by production team enterprises. They actually benefited from the development of private enter- prises, as they can now borrow more from banks and pay less tax or win ex- emption from tax. Trends, Problems, and Implications The most fundamental factors affecting TVPS will be the inconsistencies in China's reforms and the future course of reform. The changes since 1979 have been the result of the interaction of efforts by interest groups and orga- nizations at the county, township, village, and household levels to maximize their incomes as collectivized agriculture disintegrated. TVPs are striving to set new objectives and find new forms of existence and development. The four counties have to tackle different manifestations of this problem. Wuxi: The Crisis of the TVCEs and Stabilization of Employment In 1985 the total output value of industrial TVCES in Wuxi County grew by 78.3 percent over 1984 to reach Y3.66 billion (in 1980 constant prices), and their total profits shot up to a record Y425 million, an increase of 84.5 per- cent. This rapid rise in profits was accompanied by a sharp increase in average wages and some expansion of employment. The average annual wage for workers in industrial township enterprises rose from Y918 to Y1,068 (a 16.3 percent gain). The increase in the number of rural residents employed and in their average wages brought the income of the rural population in Wuxi County as a whole up to the level of the urban residents of Wuxi city.3 In 1985 industry in Wuxi County employed 282,000 workers, or 52.2 per- cent of the county's total rural labor force. International experience suggests that this share is high even for a highly industrialized city. In fact, after Wuxi County achieved full employment, it experienced a shortage of labor. Industrial production of TVCES in Wuxi County continued to grow at a high rate, and in 1986 gross value of industrial output stood at Y4.444 billion (in 1980 constant prices), an increase of 32.8 percent over 1985. Despite the central government's policy of tightening the money supply, the value of 158 LUO XIAOPENG fixed assets of industrial TVCES continued to grow at a high rate. But the net value of industrial output (in current prices) grew by a mere 0.4 percent for township enterprises, and that of village enterprises dropped by 9.1 percent. Even more threatening were decreases in net profits of 49.7 percent for town- ship enterprises and 55 percent for village enterprises. Government revenue and workers' incomes, however, were not affected. Local governments col- lected less income tax in 1986, but since indirect taxes grew by a big margin, total local revenue increased somewhat. Average wages continued to in- crease, with wages of workers in industrial township enterprises reaching Y1,283. Excessive investment in fixed assets and excessive borrowing of circulating funds reduced the rate of capital turnover to the lowest point in local history. Meanwhile, interest payments and their share of net output value both climbed to a record high (see table 7-9). According to county govemment reports, the principal problems con-front- ing TVCES were caused by intensified market competition and the increased rate of failure of capital investments. Some enterprises with capital invest- ments of more than Y1 million began suffering from overstocks of products just one year after going into production. County leaders are only too well aware of the trend toward increased risks and sharply decreased marginal re- turns on industrial capital, but they see no alternative except to continue to expand TVCES. The profits yielded by these firms furnish the bulk of local revenue and constitute the only source of funds for subsidizing agricultural production. Local farmers depend mainly on income earned by working in these firms, and township and village government leaders draw power and benefits from them. In short, 1VCES constitute the county's political and eco- nomic foundation. Despite the increasing risks, local banking organizations continue to aid the development of TvcWS. The local agricultural bank strives not for maxi- mum profits but for maximum benefits for the county through cooperation with the local government. Ii: is therefore reducing interest rates on loans (within the limits set by the central government); increasing the supply of funds by such means as lending at least YIOO million borrowed from banks elsewhere to TVCEs at cost; aiding relatively poor enterprises; and helping firms that are in the red to regain profitability. Table 7-9. Interest Payments by Industrial Township Enterprises, Wuxi Item 1981 1982 1983 1984 1985 1986 Interest payments (millions of yuan) 1.136 2.163 3.137 5.67 16.50 28.42 As percentage of net value of output 0.6 1.3 1.5 1.9 3.3 5.3 Source: Information from fieldwork. Ownership and Status Stratification 159 Local banks take this attitude because TVCEs are their largest borrowers and because of the banks' own incentive system (which is not discussed here). At the end of 1986 outstanding loans extended to TVCEs by the Wuxi County Agricultural Bank and the rural credit cooperatives operating under it amounted to Y460 million-64.9 percent of total outstanding loans of these institutions and equal to 178 percent of local rural residents' savings deposits at year's end. In contrast, agricultural loans amounted to less than 3 percent of savings deposits. The current crisis is forcing county authorities to make significant policy readjustments. The primary reason for such changes, however, is probably the growing problems in the existing employment system. Owing to fluctua- tions in production and repeated investment failures, many TVCEs are unable to keep their workers fully employed. The instability of employment and of workers' incomes in TVCEs have caused labor movements that are beginning to threaten the traditional pattem of labor allocation and employment. To ensure stable employment and effectively readjust the distribution of income, county authorities are trying to control the distribution of capital and labor. Measures include restricting the flow of labor into the county; en- suring the stability of sales by increasing the proportion of product orders cov- ered by state planning, especially from foreign trade departments; borrowing large amounts of money to upgrade the technology and equipment of county firms and township enterprises and make their products more competitive; intensifying the vertical integration of county firms and TVCES SO that even the accessories required by an order received by any firm will be produced locally; instituting countywide retirement and unemployment insurance sys- tems; and increasing expenditures on welfare and fringe benefits for workers at TVCEs. These measures suggest that Wuxi TVCES will tend to become more like small state enterprises. Nanhai: Joint Public-Private Firms and Widespread Employment of Outside Labor In 1986 industrial township enterprises in Nanhai had the same problem as TVCES in Wuxi. Despite a sharp increase in value of output, they suffered a sharp decline in profits, and their gross losses shot up to Y10.42 million, which amounted to 25 percent of their after-tax profits and was three times the 1985 figure. The net income of private enterprises and of village enter- prises leased to individual managers, however, continued to grow. Produc- tion team enterprises more and more came to be owned by individuals or were turned into mixed economic entities through mergers with private en- terprises. The total personal eamings of partnerships increased by 20 percent over 1985, and the corresponding rate of increase was 18.6 percent for indi- vidual household firms. Industrial private enterprises alone generated an ad- ditional Y24.84 million of income, a 34.9 percent increase. Rising incomes 160 LUO XIAOPENG for workers in private enterprises compelled state enterprises and TvCES to raise the wages of their own workers despite declining profits. This, plus a rapid increase in interest payrnents, caused the financial situation of TVCES to go from bad to worse. The hardest hit were town enterprises with1 work forces made up primarily of urban residents. In 1986 their average ratio of profits to sales was a mere 0.3 percent. Some successful private industrialists have experienced problem.s that range from inadequate space for expanding production to difficulties in ob- taining loans. Although they have accumulated large amounts of capital- perhaps Y1 million-they will be defeated in competition if they fail to ex- pand their operations in good time. Against this background a host of enterprises jointly run by township or village govemments and private industrialists has emerged. The establish- ment of these joint pubLic-private enterprises was motivated by a common desire for mutual benefit based on contracts of equality rather than on re- sponses to political pressures. This is an event of tremendous importance in China's rural industrialization. An outstanding example is the Keda Electri- cal Apparatus Factory in Guanyao town. Lin Jike and others operated a factory in Yanbu town near Guangzhou that produced electrical apparatus and had fixed assets valued at about Y800,000. Because of its limited space and capital the factory had difficulty competing in the market. As it happened, the Guanyao Industrial and Commercial Company, the town government office in charge of industry, needed manage- rial and technical persoinel. It signed a joint operating contract wirh Lin under which it provided a factory building and Y680,000 in equipment in return for a 55 percent share of all losses or gains. Lin Jike was appointed director of the factory, but the Guanyao town government reserved the power to make decisions. Thus the government seems to have annexed a private factory. But the fact is that joint operation enabled the factory to get large bank loans thirough the influence of the town government, and the total sum of such loans has exceeded the total joint investment in the factory. Moreover, the ownership of assets for each side is clearly specified. The assets each owned before the joint operation were not merged to become common property. The deprecia- tion funds for each side's assets are deducted and the remaining sum is calcu- lated separately to avoid controversies when the value of fixed assets is reas- sessed. The two sides share the rights to newly purchased assets on a 55-45 percent basis. Because there are separate systems of taxation and manage- ment for township and private enterprises, the local tax department and the Industrial and Commercial Administrative Bureau separately assess the in- come of each side and collect taxes and management fees from them as before the arrangement. Although this practice seems ridiculous, it shows that Lin and others did not enter the joint operation to evade taxes. Ownership and Status Stratification 161 Under the current status stratification system for TVPS, there are no rules that specifically cover public-private enterprises of this kind. The current sys- tem allows publicly owned firms to become privately owned, and it permits privately owned enterprises to become publicly owned. It tolerates firms that are publicly owned in name but privately owned in reality, as in Jieshou County. But joint public-private enterprises such as the Keda Electrical Apparatus Factory are a profound challenge to the traditional system and practices. According to Nanhai authorities joint operations of this kind are not acci- dental but represent a new trend in the making. Another example is a port machinery plant that was started by rural industrialist Liang Yuanzhang in cooperation with the Pingzhou town government. The two sides each in- vested Y800,000 in the joint project. The immediate motivation was the town government's inability to fill product orders owing to its lack of exper- tise and funds. This joint enterprise is more formal than Keda, as it operates under a board of directors with Liang as vice chairman and concurrently di- rector of the plant. Private industrialists are often shareholders in village enterprises as well as in township enterprises. In Foshan city some private industrialists have even established joint operations with state enterprises. The appearance of joint public-private enterprises in Nanhai is nothing strange. Already ten TVCEs have entered into joint operations with private businesses in Hong Kong, Macao, and foreign countries under the policy of openness. That being the case, why shouldn't they be allowed to cooperate with local private indus- trialists? Another phenomenon in Nanhai is community enterprises that hire work- ers from other places. The Ceramics Factory in Shanyuan village, Nanzhuang town, has a labor force of 1,500. When the local labor supply fell short of demand, it hired 450 workers from other places. Two separate reward systems are practiced for local and outside workers who do the same job. Outsiders are paid according to market prices for labor, converted into standard piece- work wages, whereas local community members are paid on a work-point basis and their wages float according to the factory's revenue. Outside workers eam an average of Y6 a day, local workers Y1 1. The latter workers also enjoy additional fringe benefits, including pensions, free education, and medical services. Shanyuan's income averaged Y3,500 per capita in 1986, the highest among all villages in Nanhai for the second year in a row. Another example is a factory run by the Sixth Production Team of Dagu village, Jiujiang town. In 1986 the factory, by employing large numbers of laborers from other places, achieved a per capita income of Y5,000. The emergence of these en- terprises is an outcome of the open labor market in Nanhai. The county now has more than 100,000 outside laborers, and there is free labor mobility within the county. 162 LUO XIAOPENG Jieshou: Dealings between Local Govemments and Private Industrialists By the end of 1985, 124 TVCES in Jieshou had closed down or suspended pro- duction because of inabillity to repay bank loans totaling Y489,000. In rela- tively developed Wuxi and Nanhai, township and village authorities are ex- pected to pay such debts in tnost cases, but this amount was beyond the capacity of township govemments in Jieshou, so factory buildings and land were lying idle. The Jieshou County branch of the Agricultural Bank of China (ABC), seeing potential for repayment if these facilities were well man- aged, agreed to grant loans to help private industrialists reinvest in the facto- ries, on the condition that they commit themselves to repay the existing debts. This kind of arrangement is also advantageous to community govern- ments, and it has become an important means of liquidating unpaid bank debts of community enterprises. Tax exemption rights and rents for public land and facilities, as well as their own land and facilities, are often used by Jieshou private industrialists as collateral for bank loans. In Nanhai private industrialists are not only good managers but also have accumnulated capital funds on their own. Conse- quently, the risks are bome jointly by community govemments and private industrialists and the ban,k takes no risk at all. In Jieshou, by contrast, the bank could quite possibly fall victim to deals between community govem- ments and private industrialists. The Jieshou ABC has been more involved in clearing up bad debts than banks in any other county, since it was unable to shift responsibility onto community govemments as its counterparts in other regions did. Three conditions enabled the Jieshou ABC to get directly involved in credi- tors' rights and clearance of liability for delinquent loans to TVPS. First, enter- prises are few in number and small in size. In 1985 Wuxi had 3,622 industrial village enterprises alone, and their average output was triple that of township enterprises in Jieshou. In Jieshou only 273 TVPS had transactions with the ABC in 1986. Altogether 499 loans totaling Y3.11 million were granted in that year-only 1 percent of the figure in Wuxi and Nanhai counties. Sec- ond, land and factory buildings constitute the bulk of TVPS' fixed assets, so that there is less specificity of assets. Third, permission for private industrial- ists to lease TVPS or to run such firms under the name of the county or com- munity govemment has increased the mobility of fixed assets. Nanhai has the best developed labor market among the four counties; in Jieshou the development of such a market is hampered by the unlimitecl sup- ply of labor. The market for real estate, however, is liveliest in Jieshou. For example, in 1987 the Zaolin Township Textile Mill moved to the county seat to get access to an uninterrupted power supply. In response to Jieshou's thriv- ing real estate market, many firms have invested heavily in land. Dealings between Jieshou community govemments and private industrial- ists involve a distorted pattern of property rights that not only raises transac- Ownership and Status Stratification 163 tion costs but also gives rise to corruption. The Zhuanji Paint Factory is a case in point; a top township govemment official, in collusion with some relatives, cheated the bank of large sums in loans. Jieshou County has a rela- tively lively and open economy, but many dealings in violation of established rules have led to widespread corruption. Jieshou also encourages private industrialists to help start or manage firms that are nominally registered as township enterprises. These enterprises, however, have to resolve two problems: the ownership of successful firms and the responsibility for liabilities of money-losing enterprises. Successful con- cems started by private industrialists but registered as community enterprises may be transformed into enterprises that actually are collectively owned and that operate directly under county authorities. The Beef and Mutton Process- ing Factory and the Zhongyuan Canned Food Company have already been taken over by the County Labor Bureau. The Shuguang Foodstuffs Factory and the Chunyan Elastic Band Factory also may be about to become county collective businesses. Pressure on private industrialists and businessmen to abandon their owner- ship rights to their firms comes from both the government, which hopes to increase the employment of urban residents, and the employees, who want job security. Improvements have been made in solving the employment prob- lem in China's large and medium-size cities, but in smaller towns urban work- ers meet competition from rural workers. Private enterprises and even TVCES want to maximize profits and would rather employ rural workers, since their labor is cheaper and they can be easily dismissed. Urban workers in most areas still hope to be employed by state enterprises or quasi state-owned ("large collective") enterprises that can provide job security. Workers in large collective enterprises not only have more job security but also have access to housing allocated by their enterprises, receive pensions on retirement, and enjoy other important welfare benefits. Therefore all TVPS with fairly large numbers of urban workers face pressure to tum into large collectives. Private industrialists sometimes also want their enterprises to be promoted to large collective status because they believe they will not be able to main- tain true ownership of the assets anyway and because the consequent change in status for the founders and their families is advantageous. The founders become govemment cadres, and their families are officially recognized as per- manent urban residents and are able to enjoy many privileges not shared by rural residents. Of course, the private industrialists also lose many rights: their personal pay is controlled, and they no longer have the power to hire and dismiss workers freely. No doubt this seriously affects the initiative and efficiency of their enterprises. The Shuguang Foodstuffs Factory and the Zhongyuan Canning Company both produce the same types of products and use the same technology, but their efficiency differs greatly, basically because the former is privately owned while the latter has been turned into a large collective enterprise. 164 LUO XIAOPENG Li Huade and the other founders of the Shuguang Foodstuffs Factory have not yet agreed to exchange their ownership rights for more secure positions. Their present income is high, and market prospects are good. Even more im- portant, the local insurance company has provided them with an arrange- ment that has reduced their risks. In 1986 the enterprise took out various types of insurance (simple personal safety insurance, accident insurance, and family property insurance) for all workers of the factory. In addition, the en- terprise took out annuity insurance of Y50,000 for its principal founders and promised to do the same for other employees in important posts. Li Huade and the others were in fact trying to avoid giving up their right of ownership to the government. But not all private industrialists are able to stand up to the pressures. In neighboring Taihe County an enterprise chiefly composed of urban workers and successfully managed by entrepreneur Li Shanyu earned annual profits in 1985 and 1986 that averaged Y10,000 for each of the twenty-three work- ers. Yet the firm was tratnsformed into a large collective enterprise in 1986 under pressure from the workers. Shangrao: An Intractable Dilemma Shangrao is the only county of the four to possess fairly rich natural resources. Ironically, it is precisely because of this that the county's rural industry is faced with a dilemma. T he difficulties involved in developing small hydro- power stations are a good example. Shangrao has many such stations, with a total generating capacity of 6,955 kilowatts in 1983. Twenty-one plants, with a total generating capacity of 5,885 kilowatts, were set up by people's communes, and forty-nine, with a total capacity of 1,070 kilowatts, by bri- gades and production teams. Although the typical investment for a station was only several hundrecl thousand yuan-far less than investments in large plants-the stations were still highly risky ventures for a poor area like Shangrao. In general, one-third of the investment came from the state, one- third from banks in the form of low-interest loans for equipment, and the remaining one-third from the czommunities themselves. By 1986 bank loans for hydropower equipment had reached Y4.77 million, and the total invest- ment was estimated at more than Y10 million. Although small hydropower stations have brought many benefits to resi- dents and enterprises in the Shlangrao area, management efficiency has not come up to expectations. Serious problems occurred in design, construction, and management, and many stations run at a loss. More than half of the loans extended for hydropower have not been repaid on time. Many town- ships complain that the supply of electricity is not stable and that this seri- ously affects industrial production. Hydropower projects have to be a certain size and be built under certain conditions to be economically efficient. They also require specialized techniques and managerial knowledge and hence rela- Ownership and Status Stratification 165 tively unified and specialized management. The organizational structure and form of ownership of TVCES can hardly cope with such requirements. The small cement plants of Shangrao County have achieved greater suc- cess than the hydropower projects. The simplicity and small size of Chinese- made cement machinery reduces the gap between the management ability of community governments and modem technology. As a result, the produc- tion of low-strength cement has expanded rapidly. The qualities of cement products are highly standardized, and it is easy to disseminate the methods of testing quality. Moreover, sales techniques are simple. But the county has paid a heavy price for its cement plants in energy wastage and environmental pollution. The small cement plants are adapted to the low investment and management abilities of Shangrao's community governments, and they do not offer a solution to Shangrao's basic problem-its inability to make use of modem technology. Similarly, large numbers of farmers have set up small coal pits and brick kilns; but this use of primitive technology does nothing to stimulate production for markets in other parts of the country, and it can- not help the farmers broaden their mental horizons. Conclusions For some time China followed the Soviet model of development and copied. that economic system. The intention was to achieve nationwide industriali- zation rapidly, but the system also favored the town over the countryside and industry over agriculture. Both countries followed the same model, but China, unlike the Soviet Union, has had to cope with the growing pressure caused by its vast population. This has forced it to elevate rural industrializa- tion from a spontaneous activity to a national strategy for providing jobs for surplus rural labor. China's rural industrialization has occurred against the background of priv- ileged development of town and industry in comparison with countryside and agriculture. This determined the fundamental characteristics of the organiza- tional structure of TvPs and their ownership rights. As far back as the 1950s the government established a state monopoly on the purchase and sale of grain, cotton, edible oil seeds, and other important farm products. The pur- pose was to ensure supplies to urban industry while restricting migration of rural residents into the cities and, as a first priority, to feed and clothe all members of the Chinese population. The establishment of a well-organized system of status stratification in rural society under the principle of combining government administration with economic life inevitably followed. The sys- tem guaranteed the privileges enjoyed by the city and by industry. At the same time, it deprived farmers of their property rights to land and allowed urban industry to establish monopolies under state protection. The development of TVCEs has been the result of a compromise, or "deal," between the city and countryside as the latter faced population pressure. The 166 LUO XIAOPENG motives for this deal and its essence were clearly revealed in the early stages of TVCE development in Wuxi County. To this day, the county still links rural industrialization with the fulfillment of state grain purchasing quotas as a gen- eral principle. In the 1960s there was a tacit agreement between city and countryside: the farmers received limited permission to run industries and in return committed themselves to continuing to supply the cities with low- priced agricultural products. In other words, farmers were allowed to own some industrial assets independent of the state industrial establishm,ent, to compensate for their loss of property rights to land. For TVPS this deal had a dual character. On the one hand, in conformity with the rural commune system of combining government administration and economic management, TVPs had to subordinate themselves to communes or brigades. On the othLer hand, they were not covered by urban industrial planning. These two conditions led TvPs to form institutionalized structures of status stratification but made them relatively independent compared with urban industrial enterprises. Important Features of the Ownership System The main features of the TVP ownership system are as follows. • In general, the property rights of TvPs are defined in a clear-cut and complete manner in cormparison with those of urban state enterprises. With state enterprises it is not clear who is responsible for the income yielded by their assets and for the risks from their operations, especially capital investment risks. Such risks can eventually be shifted onto the entire society, and as far as the enterprises are concerned, property rights and liabilities become meaningless. With TVPS, however, com- munity governments are responsible for such income and risks in most cases. Whether all or only some members of a community are directly responsible for the income and risks of a specific firm depends on cir- cumstances, but on no aoccount will it be possible for the enterprise to shift such responsibility, especially responsibility for capital investment risks, onto other communities. • The ultimate owners of rvPs are individual farmers or groups of farmers. The ultimate owners of rvcEs are all of the farmers in the communities within which the enterprises operate. But as a matter of fact, the forms of ownership are diversified, and in some cases even essentially differ- ent, among regions and localities. The rights of individual mem,bers of rural communities with respect to TVCEs are not clear-cut. For example, the statement that "a township enterprise is an enterprise owned by the people in a given township" probably has more practical meaning in Wuxi than in the other three counties. TVPS with different forms of ownership do not have equal status. Farm- Ownership and Status Stratification 167 ers are of lower status than urban residents, members of villagers' com- mittees are lower than township government leaders, and ordinary farmers are lower than members of their own villagers' committees. Township enterprises are of lower status than state enterprises or urban collective enterprises, village enterprises are lower in status than town- ship enterprises, and so on. The higher the grade of a firm, the more preferential treatment and resources it gets from the government and hence the stronger its sense of security. The lower the status of an en- terprise, the more directly is the firm's income linked to that of its own- ers and the clearer are the relations between them. Transfers and changes in enterprise ownership forms can vary according to the ranks of the enterprises involved. The ownership of a TVCE can- not be transferred to another firm of the same rank because workers of all enterprises are residents of the same township or village and there- fore cannot be transferred. The ownership of an enterprise can be changed, however, by raising its status. For example, a township enter- prise may, through transfer of its assets to the state, have its status raised to that of a state enterprise operating under county authorities. Under such circumstances the status of enterprise workers will be raised corre- spondingly. In general, it is difficult for the government of a given level to prevent an enterprise directly under its jurisdiction from getting up- graded if the upgrading is in the interests of the workers. As a conse- quence of the PRs, community enterprises in difficulties can now be sold or leased to private industrialists. Owners of private enterprises can sell them or have them raised to community enterprise status. At present, standard methods and procedures for the transfer and change of owner- ship of TVPS are lacking. Ordinarily these changes are carried out ac- cording to unofficial agreements or unwritten local rules or by adapting official rules and regulations. The Significance and Adaptability of the TVP System There are three main policy targets for the TVP sector in China: supporting agriculture and augmenting farmers' incomes; consolidating and advancing public ownership by developing TVCES; and achieving local rural industrializa- tion to prevent large-scale rural-urban migration and the social and economic problems that arise from urban congestion. Among the four counties covered by our investigation, only Wuxi has met all these targets. In China as a whole, counties like Wuxi are few and are located mainly in suburban areas under the jurisdiction of large industrial cit- ies or in south Jiangsu and the Pearl River Delta, where there are many well- developed medium-size and small cities. In other regions only a small fraction of the total number of townships and villages has attained these objectives. 168 LUO XIAOPENG Nanhai County seems to have achieved the first and third goals, but its efforts to promote the second have not been as successful as expected. Al- though Shangrao has made some progress toward the first and third targets, it is now laden with too many difficulties to go further. Jieshou has aban- doned all three goals. To assist industrial development in its own urban area, it has altered the traditional clefinitions for the status stratification and own- ership of TVPs by removing restrictions against the development of private enterprises and against rural-urban migration. Cases like that of Jieshou are few nationwide; Nanhai and Shangrao are more representative. Following the introduction of the PRs the Chinese government shifted the focus of its policy for rural industry onto the third target. It is now working to correct price distortions between agricultural and industrial products. At the same time it has begun to back the development of private enterprises while allowing the transformation of existing community enterprises through shareholding and through adoption of cooperative and other modern organi- zational forms. In this situation the organizational and ownership structures of TVPS, taken as a whole, have proved to be highly adaptable, allowing alternative policies for areas with different levels of development and different economic condi- tions. For example, the Wuxi County govemment is able to intervene in the operations of grass-roots units and in this way minimize competition among local enterprises, ensure maximum equality and security for all of them, and improve the competitiveness of the county as a whole. Within the sanme basic institutional framework, Jies,hou County goes to the opposite extreme through a laissez-faire policy i:hat allows the privileges of community enter- prises to be transferred but not abandoned. In Nanhai County official respect for individual property rights and freedom to choose jobs has increased the society's understanding of property rights in general and of individual rights in particular. This new relationship of rights is becoming an important means of coordinating the development of community and private enterprises under freer market conditions. Contradictions and Problems There exist deep-rootecl contradictions in the organizational structure and the ownership and status stratification systems of TVPS. Preventing excessive migration of rural residents into cities and consolidating the public ownership of the rural economy are two important policy targets. The principal incen- tives for achieving such targets should be markets and property rights. But such a policy design works only if there are opportunities and capabilities for industrial development everywhere in the countryside. In fact, the distribu- tion of such opportunities and capabilities is extremely uneven. Hence this policy design caused large investment failures in underdeveloped areas such Ownership and Status Stratification 169 as Jieshou and Shangrao, whereas in Wuxi and Nanhai it led to excessive industrial growth as all villages and townships started factories. Since the system of community ownership inhibits the flow of labor and capital, communities or regions with full employment ignore the opportunity cost for the whole society and turn to capital-intensive investments, high social welfare, and high consumption, thus reducing the efficiency of capital utilization and making a fair distribution of income impossible. The standard of living enjoyed by people in some of the townships and villages in Wuxi and Nanhai has approached or even exceeded that of local urban residents. This contributes to fair distribution of income within these counties, but it means greater inequality between them and rural regions that lack opportuni- ties for industrial development. Large investments in scattered public facili- ties and infrastructure reduce the potential economic benefits from urbaniza- tion and concentration. Under the status stratification system of TVPS, many leaders of grass-roots units work as government officials and simultaneously represent the owners of community enterprises. The combination of these two roles in one person inevitably gives rise to serious political and economic interference. Political pressure from the top often causes encroachment on economic interests, which can lead to distortions of economic information, gross violations of law and discipline, and widespread corruption. The dependence of TVPS on administrative organs and on administrative protection of their interests and property rights gravely hampers the optimal development of their organizational structure and makes it difficult for them to use organizational and technical innovation to benefit from the expanded local economy and specialized production. This pattern also constitutes the source of investment failures and of destructive and excessive competition. Although the pioneering spirit of private industrialists has been important, the institutionalized discrimination against private enterprises inherent in the status stratification system obviously prevents them from doing their best. These weaknesses were hidden by the high-speed economic growth of the TVP sector in 1983-85, when the market expanded rapidly. But since 1986 the overall economic efficiency of TVCES has dropped sharply. This suggests that the weaknesses of TVPS may have already become a significant hindrance to rural industrial development. Seeking New Ways to Advance under Harsh Restrictions The present TVP system bears the imprint of the utopian fantasy of thirty years ago. With the gradual disappearance of that unrealistic dream, the harsh reality behind it has become clear to the people. But the system of TVPS unique to China was born under restrictions that are impossible to dis- mantle in a short time. 170 LUO XIAOPENG It is easy to criticize the system's backwardness and the mark of feudalism it bears by making a simple comparison with the organization and property rights of enterprises in Western societies. But China's unique cultural tradi- tions, as well as its population problem, are the legacy of its history. China has no way out but to take these restrictions into account as it seeks the right path and tactics for its advance toward modernization. Moreover, the forces of inertia in the urban economic structure affect the choice of policy altema- tives. In particular, the urban employment and welfare system promotes "equality" for urban citizens but is unfair to farmers and is a great obstacle to China's urbanization. Rural industrial development will, for a fairly long time to come, continue to be an essential part of China's rural development. Nevertheless, as the experience of many other countries shows, urbanization is the only way to modernize. A development strategy that pushes rural industrialization as a substitute for urbanization might prove costly. To smooth the development of rural industry and to make it a harmonious and active factor in the whole process of modemization, it is essential to revise or even redesign the TVP system and its goals. In view of the difficulties and crises that face rvPs, it is necessary to consider how to speed up urbanization over the mediurrm term. A new urbanization strategy is needed as the precondition for reforming urban-rural relationships, the institutional system for urban and rural enter- prises, and other systems. It is impossible in the near future to pull down the barrier between urban and rural areas and effect a relatively free flow of population. Policy should therefore focus on improving 1the coordination and management of TVPS. A pressing task is to set up new types of investment and monetary institutions to serve as efficient policy instruments and means of coordination. Domestic market conditions have witnessed profound changes, but the craving for in- vestment is still strong in rural areas. Lack of managerial expertise and of ability to make sound investment decisions have created great risks for TVps. The first priority at present is to prevent unnecessary grave losses, not to seek optimal development for rural industry. Designing and realizing the TVP system of the future will require a fairly long period of experimentation. China, a big country with a rich cultural heritage, cannot indiscriminately copy the experiences of foreign countries. It will be possible to implant modem organizational principles into the sys- tems of Chinese enterprises only after significant technical and theoretical breakthroughs are made. So far, long-neglected problems have surfaced, new conceptions are being introduced, and systematic studies have been started. Even more important, some irnovations are being tried out in practice. We have every reason to anticipate the establishment in China of a modern, uni- fied system of urban and rural enterprises that preserves the characteristics of the national culture. Ownership and Status Stratification 171 Notes 1. "Stipulations of the State Council on Some Issues Concerning the Development of Com- mune and Brigade Enterprises." 2. Nanhai's fiscal contribution to the next higher level of government was required to in- crease at an annual rate of 7 percent from the 1983 base, and the country was allowed to keep all of the surplus. In Wuxi more than 80 percent of total yearly fiscal revenues has to be handed over to higher authorities. 3. According to sample surveys, the per capita income of urban residents of Wuxi municipal- ity (which includes three counties-one of them Wuxi County-as well as the medium-size city of the same name) averaged Y957.84, compared with Y974.32 (net income, Y818) for people in its rural areas. It is true that some factors that affect the income and expenditures of rural and urban residents are not comparable, but when the more equal distribution of income in the city than in the countryside is taken into account, we can safely say that by 1985 most rural residents in Wuxi County had the same income level and living standards as residents of Wuxi city. 8 Private Enterprises: Their Ermergence, Rapid Growth, and Problems Lin Qingsong Since 1978, and particularly in the past several years, rural China has seen an unprecedented proliferation and growth of industrial private enterprises, which immediately became fierce competitors with community enterprises. Recognition of the role of the private sector is necessary to an understanding of the historical process of industrialization in China's rural areas. This chap- ter discusses China's private enterprises, in particular their environment and the characteristics of their operations. Since the founding of the People's Republic of China there have been three waves of rural industrialization. The first came in 1958, when the peo- ple's communes were established and a nationwide industrialization campaign was launched. The second waove began in 1970 when the State Council, at a conference on agricultural development in North China, called on rural areas to speed up agricultural mechanization. Farm machinery factories set up by rural community governments spread all over China, further boosting the development of commune and brigade enterprises. The third wave came after the Third Plenum of the Eleventh Central Committee in December 1978 shifted the nation's focus to the reform of the economic system. Com- munity enterprises entered a new era of rapid growth, but the main difference between the third wave and the first two was the unprecedented magnitude and speed of the development of private enterprises. The recent development of private enterprises can be roughly divided into two stages: the nascent stage (1978-83) and the flourishing stage (since 1984). Even before 1978 the soil had been prepared to some extent for the development of private enterprise, and a few rural residents had launched some private firms. But under the influence of the ultraleftist policy of the time these firms were regarded as "illegal underground factories" and were strangled in the cradle. Even household sideline production by farmers was banned as the "tail of capitalism." After 1978 the Chinese government abolished policy restrictions against farmers' entering nonagricultural production. As a result, even as community 172 Private Enterprises: Their Emergence, Rapid Growth, and Problems 173 enterprises continued to grow rapidly, the industrial and sideline production of farm households (a fledgling form of private enterprise) enjoyed speedy de- velopment. Between 1979 and 1981 the value of output of farm household sideline production increased at an average annual rate of more than 14 per- cent. During 1981-83, with the introduction of the production responsibility system (PRs) in agriculture, the value of output of farm household handicraft industry jumped by 105 percent. Meanwhile, a large number of private enter- prises began to emerge, including handicrafts in farm households and simple manufacturing enterprises run by individual farmers. Some regions saw the emergence of cooperatives and industrial partnerships jointly run by groups of farm households. During this nascent stage most of the increase in rural industrial production was still contributed by community enterprises. 1984 saw extremely rapid growth of private enterprises. The gross value of industrial output (GVIO) of individually run rural firms catapulted to Y1 1.7 billion, or 9.4 percent of the GVIO of all TVPS, and the Gvio of partnerships and other jointly run firms grew to Y10. 1 billion, or 8.1 percent of the GVIO of all TVPS. To exclude the effect of different methods of collecting statistics, table 8-1 gives figures on the share of private enterprise in the TVP sector for 1984 and 1985 only. Private enterprises account for a large proportion of the totals for number of firms, employment, value of output, and annual growth rate. Since 1984 private enterprises have given a "second push" to the third wave of rural industrialization, as data on industrial TVPS in Jieshou, Nanhai, and Shangrao counties testify.' The Benefits from Development of Private Enterprise As is true of community enterprises, the development of private enterprises can help speed the movement of rural surplus labor out of agriculture, invig- orate the rural economy, raise the incomes and living standards of farmers, increase central and local govemment revenues, and promote the historical process of rural industrialization and national modernization. In addition, the development of private enterprise has some distinctive economic implica- tions. The development of private enterprises may help improve the efficiency of resource allocation and utilization. China is a large developing country with an extremely backward economy, a serious shortage of capital, and an abundant supply of labor. The development of TVPS can help promote the substitution of labor for capital and the more efficient utilization of both fac- tors. According to some estimates, during 1979-84 TVPS increased their total value of output by Y83.6 billion. In so doing, they used 6.9 million rural surplus laborers, which in effect substituted for Y28.1 billion of investment in fixed assets (see Development Research Institute 1986, p. 11). Since pri- vate enterprises are more likely than community enterprises to employ labor- intensive production techniques, their substitution of labor for capital is even 174 LIN QINGSONG Table 8-1. The Share of Private Enterprise in the 7VP Sector Item 1984 1985 Total number of TVPs (thousands) 4,812 4,930 Of private enterprises 3,911 4,076 Number of private enterprises as percentage of total 81.3 82.7 Increase in private enterprises as percentage of increase in TVPs - 139.2a Employment in TVPs (millions) 36.561 41.367 In private enterprises 11.072 13.546 Employment in private enterprises as percentage of total 30.3 32.8 Increase in private enterprise employment as percentage of increase in TVP employment - 51.5 Value of output of Tvps (billions of yuan) 125.44 182.72 Of private enterprises 21.9 36.79 Value of output of private enterprises as percentage of total 17.5 20.1 Increase in private enterprise value of output as percentage of increase of TVP value of output - 26.0 - Not applicable. a. The growth of private enterprises outstripped the overall increase in the number of TvPs because 46,300 community enterprises exited during this period. Source: Statistical materials on China's TVPs. more evident. In Jieshou County in 1986, for instance, each Y10,000 of out- put value of private enterprises required 2.63 workers and a fixed asset invest- ment of Y1,615.5, which meant that private enterprises needed ]Less in- vestment and offered more jobs than state and community enterprises.2 Moreover, private enterprises use scattered resources that cannot be filly uti- lized by state or commtnity enterprises, such as leftover bits and pieces of industrial materials and small. deposits of low-grade mineral resources, and they economize on costly land to a greater degree. Private enterprises also contribute to geographically more balanced. devel- opment. The huge gaps in development among regions before industrializa- tion and the prospect that inclustrialization itself may widen these gaps have long been a headache for China's economic modernization. During 1949-78 the government achieved some results by constantly increasing state invest- ment in backward areas, but it failed to solve the problem. Such "blood trans- fusions" of investment are usually extremely inefficient and sometimes en- courage backward areas to rely on state aid. In the first few years after 1978 the government depended almost solely Private Enterprises: Their Emergence, Rapid Growth, and Problems 175 on community enterprises to push rural industrialization. But the gaps among regions, instead of narrowing, widened alarmingly as a few relatively devel- oped areas took advantage of their rich financial resources, abundant human talents, and favorable geographic conditions and enjoyed rapid rural industri- alization while the backward areas could hardly get off the ground. Since 1984 the blossoming of private enterprises has helped narrow the gaps among regions and has offered new hope for solving the problem of regional imbal- ances (see chapter 12) as rural industrialization progresses. Another benefit of private enterprise is the more coordinated development of nonagricultural production in rural areas and the correction of deviations in the structure of nonagricultural production. China's traditional industriali- zation strategy led to serious distortions in economic structure and low overall efficiency. Until 1984 the same kinds of problems existed in the development of nonfarm activities in the countryside. The surprising similarity between the structure of the nonagricultural sector in rural areas and in the economy as a whole (see Development Research Institute 1986) arose from the efforts of grass-roots rural community governments to boost rural industrialization. The similarity between urban and rural industrial structures has exacerbated the existing structural contradictions in the national economy and has seri- ously hindered the improvement of the efficiency of the national economy and the further development of rural industries themselves. Since 1984 pri- vate enterprises, with their strong market orientation, have begun to break through the limitations that community governments imposed on commu- nity enterprises to further the governments' noneconomic goals. Private en- terprises are actively seeking favorable investment opportunities offered by the imbalances in the national economy, and to a certain degree they help correct structural distortions (see chapter 12). They thus contribute to the improvement of the efficiency of the national economy. The overdevelop- ment of manufacturing industries in Wuxi County, where community enter- prises dominate, and the more balanced structure of production in Wenzhou city, which sees private enterprise as the principal part of its TVP sector, sup- port this conclusion. The development of private enterprises poses a new challenge to com- munity enterprises. In the past rural industrialization shared some of the drawbacks of the traditional industrialization strategy. Centralized control and administrative interference by village and township governments sap- ped enterprises' decisionmaking power and rendered them unable to react promptly to changes in market supply and demand. The "big pot" system of distributing income among firms within a community and among employees of an enterprise dampened enterprises' and employees' initiative and enthusi- asm. Community enterprises still often suffer from economic inefficiency and lack the capability to advance their own development. Private enterprises, which can be more flexible, often obtain more favorable economic opportun- 176 LIN QINGSONG ities and bigger market shares. Their competition in factor and product mar- kets has put great pressure on community enterprises and has forced them to reform their inefficient systems. In recent years Wuxi, where rural industri- alization based on community enterprises is far advanced, has felt threatened by the Wenzhou model, in which private enterprises dominate. Wuxi has sent several teams to study the Wenzhou experience and is now busy prepar- ing further reforms in its TVP system. These experiences show that the impact of the development of private enterprise on China's rural industriaLlization should not be underestimated. Why Private Enterprise Has Mushroomed The vigorous growth of privalte enterprises in China's current wave of rural industrialization is the result of a series of policy readjustments and system reforms. The government tock steps to wipe out the systemic and policy ob- stacles that prevented farmers from freely entering nonagricultural activities and to provide them with the necessary economic conditions for setting up firms. Private enterprises could not have come into being if government agencies at all levels had continued, as during 1958-78, to control produc- tion by compulsory means and to force people "to offer nothing but their labor and on the other hand to have nothing as their own property except that for personal consumption." Starting in 1978 the restrictions thatt barred farmers from freely entering nonagricultural activities were gradually r elaxed. In particular, the implementation of the PRS hastened the dismantling of the system of the "combination of government and commune" and the restric- tions on farmers' rights to own property and change their status. System re- forms in urban areas have broken down the mandatory planning system and the state-controlled distribution system that had long curbed the develop- ment of producer goods markets. The reforms thus wiped out some ingrained obstacles to the development of private enterprise. The recent phenomenal development of private enterprise was possible be- cause of social and economic preconditions brought about by the economic reforms. Rural Monetary Accumulation Funds for launching private enterprises usually come from three sources: the founders' own savings; collective funds-including fixed assets (factory build- ings, machines, and equipment) and floating capital-that are now at the disposal of individuals who operate commune and brigade enterprises under contracts or leases; and loans from banks and credit cooperatives as well as credit from local residents. A survey of typical cases s]hows that founders' own funds usually account Private Enterprises: Their Emergence, Rapid Growth, and Problems 177 for one-third and sometimes all of the total startup investment for private enterprises. During 1978-84 the introduction of the PRS and the rise in pro- curement prices of farm products increased farmers' incomes remarkably. Av- erage per capita net income jumped from Y133.57 in 1978 to Y355.33 in 1984, an increase of more than 150 percent, and farmers' bank savings rose from Y5.57 billion to Y43.81 billion. (The amount of cash that farmers kept in their own hands might have been three or four times that figure, since some were afraid to make their wealth public by depositing their money.) As rural households specialize in different types of production, the income gap between farmers has widened, and the cash and bank savings of prosper- ous farmer households may be thousands or even tens of thousands of yuan. These savings can be tapped by farmers to launch private enterprises, and they are the foundation for credit provided to infant private enterprises. Another important source of startup funds for private enterprises is the col- lective fixed assets and floating capital to which individuals gain access by becoming managers of TVCES through contracts or leasing. Around 1978 some of the TVCES set up during the previous two waves of rural industrialization closed down and others had low profits or chronic deficits. These firms were a heavy burden on rural community governments. Many communes and bri- gades, particularly those with weak leadership, followed the example of the PRS in agriculture and dispersed their holdings. They sold some or all of their enterprises to individuals at extremely low prices, leased them at a fixed rent, or contracted them out in return for a certain percentage of after-tax profits. Before 1984, according to surveys, such firms accounted for as much as 50 percent of the total number of private enterprises in economically backward provinces. Thus a large proportion of private enterprises were originally TVCES. Most of the "capable persons" or grass-roots cadres who obtained own- ership or management rights to TVCES under preferential conditions have proved to be successful managers. So, in a sense, failure in the management of the original TVCES was a factor in the successful development of today's private enterprises. Loans from banks and credit cooperatives and credit from local residents form a third important source of funds. Since 1978 the Chinese government has gradually relaxed restrictions on loans to individual businesses and other private enterprises, and since 1984, in response to government encourage- ment of private enterprise, banks and credit cooperatives have sharply in- creased their loans to private firms. Loans make up a larger proportion of the total funds of private enterprises in advanced than in backward areas and for large firms than for small individually run enterprises.3 Meanwhile, where the commodity economy is advanced and private enterprise is the main part of the TVP sector, the informal credit business has developed rapidly, and in such places as the rural areas of Wenzhou city informal credit is already the main source of funds for private enterprises. 178 LIN QINGSONG Labor Supply Liberating farmers from their dependence on the land and on community governments and giving them freedom over the use of their own laLbor was the precondition for a dependable supply of labor for private enterprises. After 1958 the people's communes used such coercive means as the house- hold registration and grain ration systems to confine farmers to the limited arable land. The abolition of the commune system to a great extent elimi- nated the restrictions that curbed farmers' freedom to choose occupations and move from one place to another. The great numbers of surplus rural laborers that were buried in oblivion in the past began to leave agriculture. Farmers with money or technical know-how and managerial ability started or invested in private enterprises, and others swarmed into the labor market. By the end of 1985, 41,521,400 rural surplus laborers in China were employed by com- munity enterprises and 28,268,900 by private enterprises. It is estimated that there are still more than 100 million surplus laborers in agriculture, so the abundant and cheap labor supply brought into being by the rural reforms is still available to be absorbed by private enterprises. The main reason that large numbers of farmers are attracted to jobs in pri- vate enterprises is higher incomes. Of 951 employees surveyed in fifty-two private enterprises in thirteen counties of Shaanxi Province, 90.6 percent said they sought jobs in these firms to eam more money, 6.5 percent wanted to learn techniques and skills, and 1.9 percent had other reasons. Of those who wanted higher incomes, 6.6 percent (mostly technicians and skilled workers) were from areas with relatively high living standards, 71.7 percent were from families that could afford basic necessities but wanted a higher standard of living, and 21.7 percent were from poor families seeking a basic livelihood. Freer movement of labor maeans an improved supply of workers for private enterprises. For instance, local surplus labor in Xiqiao town, Nanhai County, has already been almost totally absorbed by TVPS, but private enterprises can still recruit large numbers of workers from backward areas in other regions. In 1986 TVPS in Xiqiao employed 1,469 laborers from other regions, and job- seekers are still coming to the town. Owing to the rapid development of TVPS, Lianxiang village in the town has had no surplus laborers since 1984. Thus when twelve local farmers pooled their funds in 1984 to launch the Lianchang Silk Knitting Mill, they had to recruit half of their 120 emaployees from other areas. (The rest were family members and relatives of the share- holders.) But in general, workers in private enterprises come mainly from nearby, and in new firms the first group of employees typically consists of the founders' families, relatives, and friends. In more developed areas em- ployees hired later may come mainly from other townships, counties, or even provinces, since local labor is in short supply and workers from other areas Private Enterprises: Their Emergence, Rapid Growth, and Problems 179 are cheaper, more manageable, and more willing to do heavy manual labor than local people. Land and Natural Resources In the past most private enterprises were small individual firms or household- run factories that usually occupied founders' homes or private plots or land contracted to them by local communities. But private enterprises that have many employees or engage in mining or in producing building materials face a serious problem of land acquisition. Most private enterprises consider two factors in choosing factory sites: the cost of the land and distance from the market for their products. They take into account administrative restrictions, where applicable, but not local gov- emments' employment and nonagricultural output goals. Most firms prefer to set up their operations in small towns and accordingly try to rent public or private buildings. Although the PRS only defines rights to the use of farmland and its output and does not cover the right to transfer and dispose of land, it represents a significant step toward the commercialization of land. In particular, the pol- icy that implicitly allows farmers to make their contract land available to oth- ers gives private enterprises easier access to land. But since there is no clear definition of property rights to land, and since farm households cannot for- mally transfer the property rights to their contract land, private enterprises, rather than purchasing land for new factory buildings, usually lease it or make joint management arrangements under which the land is counted as shares in the firm. At present, mineral resources are not highly commercialized. In the past few years a state policy that allows all types of enterprise to use mineral re- sources freely has given private enterprises access to mining. In 1986 China promulgated its law on natural resources and began taxing their use. But be- cause differential rents for the exploitation of resources were not taken into consideration in fixing the tax rate and because the regulations are not prop- erly enforced, the effective fees for the use of resources are still very low. The case of the Chaoling Coal Mine in Fangshan County, which has the largest employment of any mining TVP in Beijing municipality, is an example of the low fees and high profits in mining. The mine began operating in No- vember 1984 with capital of Y377,000, all in loans. By the end of March 1986 it employed 195 laborers, had an annual production capacity of 40,000 tons, and had produced a total of 23,850 tons of coal with a total output value of Y632,000 (net of Y93,000 in loan repayments). In 1985 the mine tumed over to the state Y16,000 in resource taxes (Y1.50 for each ton of coal mined), Y38,100 in income tax (calculated at the rate of 1.5 percent), and Y1,500 for the individual business administration fee, which left 180 LIN QINGSONG Y44,000 in after-tax profits. The mine was still enjoying exemption from in- direct taxes in 1985. The mines that the state leaves to private enterprises are usually small and scattered. Nevertheless, high product prices and fat profits still make mining attractive to the private sector. Availability of Technology and of Technical and Managerial Personnel Economic reforms have extended the commodity economy to the science and technology sector. The gradual commercialization of technology has offered private enterprises access to the technology market, and many private enter- prises have paid to obtain technology for development. Experience shows, however, that whether private enterprises prosper depends less on technology than on the availability of managerial and technical talent. Success depends on such "capable persons" mrore than on funds, land, labor, and the other technical factors of production because these people can find supplies of the other factors and pull them together to make profits. Earlier, markets were often less important than the availability of technical personnel in determining private enterprises' choice of activities. In recent years greater job and geographic mobility for technical and managerial per- sonnel has given a strong impetus to private enterprise. The technical and managerial personnel of private enterprises are primarily former commune and brigade leaders, skilled craftsmen and educated youth, and former work- ers in state enterprises or TVC:ES. A sample survey of 103 leaders from fifty-two private firms in Chang'an, Zhouzhi, and eleven other counties in Shaanxi Province showed that 1.5.5 percent were former commune and brigade lead- ers, 53.4 percent were technical and managerial personnel from TV1ES, and 31.1 percent were retired workers of TVCEs and state enterprises or skilled rural craftsmen. A similar survey of sixty-three leaders from fifty-six private enterprises in Tianjin showed that 22.2 percent were former leaders of com- munes and brigades or their enterprises, 42.8 percent were managerial and technical personnel frolm TVCES, 30.2 percent were local craftsmen and ordi- nary farmers, and 4.8 percent were from other backgrounds. Some private enterprises even offer high pay to recruit technical personnel or part-time technical advisers from state enterprises or research institutions. These tech- nical and managerial personnel have played a decisive role in the develop- ment of private enterprises. Problems Facing Private Enterprises In the first part of this chapter I suggested that the rapid development of pri- vate enterprises is helping tc speed China's rural industrialization, raise the efficiency of resource allocation, improve the structure and the regional dis- Private Enterprises: Their Ermergence, Rapid Growth, and Problerns 181 tribution of nonagricultural activities in rural areas, and advance rural eco- nomic reform. I now turn to the problems of organization and efficiency that private enterprises confront as they pursue further development. The impression given by most case studies is that private enterprises have strict control over their property, enjoy autonomy in management, are re- sponsible for their profits and losses, are relatively efficient at deci- sionmaking, and are economically efficient because of their direct links be- tween rewards and performance. Many survey reports have concluded that the microeconomic efficiency of private enterprises is generally higher than that of state and community enterprises in the same line of business. For ex- ample, a comparison of the economic efficiency of three private and seven- teen state manufacturers of plastic products, textiles, and construction ma- chinery in Wenzhou city showed that production costs for the private enterprises were 40-42 percent lower than for the state enterprises, the profit rate was 14-40 percent higher, the value of output per Y1 million of total assets was 67-283 percent higher, the profit rate per Y1 million of total assets was 2-25 percent higher, average taxes turned over to the state per employee were 30-1,100 percent higher, and the average income of individual employ- ees was 29-63 percent higher (Lin Zili, personal communication). All of the indexes of economic efficiency were higher for the three private enterprises than for the state enterprises in the same industries. But figures on private enterprises in other regions and nationwide show a somewhat different picture. In 1985 in nine sample townships of Shangrao County the ratio of profits and taxes to capital, the ratio of value of output to capital, and the ratio of profit and tax to value of output were all higher for private enterprises than for local community enterprises; average value of output per employee and average profit and tax per employee were lower for private enterprises than for community enterprises (see table 8-2). Na- tionwide statistics indicate that in 1985 average output per employee, average profit and tax per employee, and average wages in private enterprises were 51.8, 62.6, and 77.5 percent respectively of the comparable figures for com- munity enterprises. Only the private firms' ratio of profit and tax to value of output was higher than that of community enterprises. Moreover, firms polled in surveys are relatively large, but private enterprises in Shangrao County are small, with an average employment of 2.9 persons (6.8 percent of the figure for community enterprises), and private enterprises nationwide have an average employment of 2.65 persons (10 percent of the figure for community enterprises). I do not intend to go into a detailed analysis of the factors that affect the economic efficiency of private enterprises, but their extremely small scale is obviously the most important reason for their low efficiency in comparison with community enterprises. Figures on private enterprises in Nanhai and Jinjiang counties and in Shangrao County support this conclusion. In 1985 private enterprises in seven sample townships in Nanhai had, on average, 182 LIN QINGSONG Table 8-2. Size and Perfifrrance of Community and Private Enterprises, Nanhai and Shangrao Counties, 1985 Community enterprises Private enterpises Both Both Indicator counties Nanhai Shangrao counties Nanhai Shangrao Ratio of profit and tax to capital 32.8 35.4 12.4 48.2 56.1 10.8 Ratio of value of output to capital 215.9 229.6 109.8 237.9 266.8 99.3 Ratio of profit and tax to value of output 15.2 15.4 11.3 20.3 21 10.9 Average output value per employee (yuan) 7,007 8,160 2,123 4,174 4,317 2,926 Average profit and tax per employee (yuan) 1,065 1,260 239 846 907 319 Average number of employees per enterprise 56 61 43 9 13 3 Average value of enterprise assets (thousands of yuan) 183 217 83 16 20 9 Note: Production team enterprises are excluded. Source: Enterprise Quantitative Questionnaire. 12.5 employees-about 4.3 times the figure in Shangrao-and much higher figures for six indexes of economic efficiency than the Shangrao firms. Three comprehensive indexes of economic efficiency (the ratio of profit and tax to capital, the ratio of value of output to capital, and the ratio of profit and tax to value of output) were higher for private than for community enterprises in Nanhai (table 8-2). In 1985 private enterprises in Jinjiang County (Fujian Province) had, on average, 22 employees-660 percent higher than the fig- ure for Shangrao and 76 percent higher than that for Nanhai. Three indexes of economic efficiency (the ratio of profit and tax to value of output, average value of output per employee, and average profit and tax per employee) were much higher for J inj iang private enterprises than for Shangrao private enter- prises and were higher for private enterprises than for community enterprises injinjiang (tables 8-2 and 8-3). The value of output per employee was signifi- cantly higher for private enterprises in Jinjiang than for those in Nanhai. If average employment per private enterprise nationwide can reach the level of Nanhai or Jinjiang, the economic efficiency of private enterprises can sur- pass that of community enterprises. But at present the scale of most private enterprises in China is much smaller than in Nanhai and Jinjiang, and their low capacity for self-accumu- lation will inevitably hinder their development. In addition, the social, polit- ical, and economic environment in China still impedes private enterprises' development and expansion. Private enterprises are still in their niascent Private Enterprises: Their Emergence, Rapid Growth, and Problems 183 Table 8-3. Size and Performance of Community and Private Enterprises, Jinjiang County, 1985 Community Private Indicator enterprises enterprises Ratio of profit and tax to value of output 14.9 15.9 Average value of output per employee (yuan) 3,607 4,873 Average profit and tax per employee (yuan) 537.7 774.8 Average annual wage per employer (yuan) 823 932 Average number of employees per enterprise 62 22 Note: Community enterprises include township and village enterprises. Private enterprises in- clude enterprises run by groups of households with collected funds and private enterprises solely funded by an individual. Other jointly run cooperative enterprises are excluded. Source: Institute of Economics, CASS. stage. Every year a great number of small new firms joins their ranks, which slows the expansion of average firm size in the private sector. Moreover, the living standards of most Chinese farmers are still very low, and even owners of successful small enterprises use thousands of yuan of eamings to improve their living conditions, which means that the amount left for expansion is limited. This situation, of course, is understandable and can hardly be avoided. The main issue is that there is still some fear and concern that the expan- sion of employment in private enterprises may lead to the unchecked spread of capitalism in China. As a result, policy is ambiguous. Political and social opinion has put great psychological pressure on private entrepreneurs and has created unfair competitive conditions for them. This in turn has caused dis- tortions in their behavior and seriously limited expansion of scale and im- provements in efficiency. In 1983 the government made a rule that each owner of a private enter- prise in rural areas could employ one or two hands, and a person with special skills could take on no more than five apprentices. But the govemment adopted a "wait and see" policy; it gave no encouragement or publicity to violations of the rules but did not immediately stop or punish them either. This wise policy left room for the development of private enterprises. But the resulting uncertainty also had negative effects on the private sector's de- velopment. First, uncertainty about policy blunts private enterprises' desire to reinvest and makes them extremely shortsighted in their operations. In their early stages, most private enterprises have strong motives to make and reinvest profits. But when employment reaches the limit stipulated by state policy, increasing concern about political risks dampens entrepreneurs' interest in growth and shortens their horizons in operations and management. They take every opportunity to profiteer by legal and illegal means and to take ad- vantage of loopholes in the state's regulatory system, and they are ready to 184 LIN QINGSONG quit business at any time. They are unwilling to undertake projects that re- quire long-term investment and especially to expand employment, preferring to spend their eamings on building and repairing houses and purchasing other living facilities. Because of such shortsightedness, most private enterprises continue to operate on an extremely small scale year after year. Second, the uncertainty about policy causes deterioration in private enter- prises' business environment and degradation of their capability for self- accumulation. The surveys found that in the supply of production materials and the sale of products, private enterprises and, particularly, large inclividual enterprises encounter more discrimination than do other types of enterprise. They have the most difficulty in obtaining bank loans, and when the state decides to tighten credit, loans to private enterprises are the first to be cut. Thus they may be forced to get loans at much higher rates from informal credit organizations. They usually have to pay higher salaries than commu- nity enterprises to recruit technical personnel in short supply, and in areas with shortages of labor they have to pay higher wages to attract ordinary workers. When they purchase raw materials from state enterprises, they are subject to exorbitant prices set by the seller. When electric power is in short supply, private enterprises are the first to suffer from power cuts. Sometimes they have to sell their products on credit and receive payment after a delay or not at all. To cope with such a hostile climate, private enterprises usually have to assign a large number of staff to purchasing and marketing. In Jinjiang County expenditures on purchasing agents and sales staff alone account for 10 percent of the total profits of private enterprises. In addition, when private enterprises operate under the guise of community enterprises, they must make large voluntary social donations or "pay tribute" to the community govem- ment. In some areas government departments levy a multiplicity of taxes and fees on private enterprises. All this bloodsucking by force or trickery has dras- tically impaired private enterprises' capability for self-accumulation and has restrained their development. The last and most important problem is that policy uncertainty confuses the property relationships of private enterprises and causes retrogression in their development. In the nascent stage the property relationships of private enterprises are relatively well defined and clear. As totally independent com- modity producers that are responsible for all their profits and losses, private enterprises have strong incentives for accumulation and development that are fundamentally different from those of community enterprises. But most small private enterprises, to keep a clear right of ownership over enterprise property, will reduce the speed of their development and stop seeking further advancement when their employment reaches the limits stipulated by state policy. Even if they choose to expand, some private enterprises spontane- ously blur enterprise property relationships to ward off social pressures, guard against unpredictable risks, and improve their operating climate. . Private Enterprises: Their Emergence, Rapid Growth, and Problems 185 In addition, policy uncertainty and the ensuing disorder and inefficiency in the regulatory environment generate actual confusion in property relation- ships. * Some private enterprises operate under other names. Large enterprises funded by only one individual have the most to fear from social and possible political pressures. These firms may tum in part of their profits and pay administrative fees to township and village governments so that they can call themselves TVCES. This enables them to open a bank ac- count, improves their operating climate, gives them government pro- tection, and reduces the pressure of public opinion. Alternatively, firms may voluntarily offer shares to township and village govemments and operate as jointly run cooperative enterprises, or they may provide funds for public investments. They may even declare that the entire assets of the enterprise belong to the state so that they can operate as cooperative enterprises. * Diversified funding and joint management in some private enterprises confuse their property relationships. By using funds from outside and resorting to joint management, enterprises can spread their operational risks, quickly pool small sums of money from individual farmers and use these monies to expand, and, most important, open a bank account under the name of a cooperative enterprise. Such firms develop rela- tively quickly. But they will eventually have to confront the confusion in their property relationships because they are not protected by clearly defined policies and laws. These enterprises are often unstable and find it difficult to pursue further development because there are no clearly defined regulations on the ownership of enterprise property and the dis- tribution of profits. The larger the scale of an enterprise's operations, the more confused its property relationships become and the more diffi- cult it is to coordinate the distribution of benefits. * Enterprise property relationships also become blurred for TVCES that are leased or contracted to individuals. "Contract enterprises" are firms that were originally TVCEs but are now operated by individuals under con- tracts or leases in return for payment of a fixed rent or a fixed percentage of their profits to the original owners. These firms usually operate on a larger scale and have better equipment than other private enterprises. The contractors are mostly local community government cadres, and the rents are often so low that they do not cover depreciation and the value of the land. The contractors can use collective enterprise status to enjoy preferential treatment from banks and the state. TVCES that were contracted out had previously suffered from chronic ineffi- ciency and losses because of the drawbacks in their management system. Most of them quickly eliminated their deficits and became profitable, but 186 LIN QINGSONG their property relationships are becoming more and more confused. The com- munity governments that have contracted out the firms think that enterprise assets belong to them and should be retumed when the contracts expire. The contractors hold that they owII the additional assets created by reinvestment from after-tax profits and that they should actually own the enterprises out- right when the original assets have become a very small part of the total. As community governments insist on their ownership rights over these firms, contractors' aspirations for reinvestment will wane drastically, and the enter- prises will fail to expand. Moreover, the uncertainty of policy causes confusion in the state's regula- tory functions with respect to private enterprise. Some private enterprises use their status as collective enterprises to seek preferential treatment from the state regarding taxation and bank loans. They pay very low natural resource taxes and commonly practice tax evasion. These advantages and the pressure of public opinion encourage private enterprises to try to blur their property relationships. But some private entrepreneurs express concern that since their firms' establishment and development depended mainly on state loans, the firms can be nationalized at any time. In short, confused enterprise property relationships caused by the uncer- tainty in state policy have led to the degradation of private enterprises' devel- opment potential and have forced them to operate on a small scale. This is the fundamental problem that private enterprises face. Some Policy Suggestions In recent years the rapid development of private enterprises has helped speed the transfer of rural surplus Labor to nonagricultural activities and has fur- thered China's rural industrial[ization. Private enterprises have shown great vitality. This indicates that the government's reform policies for the rural economy are appropriate, and the policies and principles adopted concerning private enterprises in rural areas have already yielded remarkable results. But the development of large-scale private enterprises has outrun govemment policies and regulations. Recently the Communist party and the government formally affirmed that the private economy as an employer of hired labor is an indispensable and beneficial supplement to the public economy. Now is the right time to sum up past experience and introduce new policies and regu- lations to deal with the new situation and the problems that have cropped up. Some brief suggestions for state policies toward private enterprises are offered here. * State policy regarding private enterprise should be further relaxed so that it will not only continue to encourage and support the develop- ment of private enterprise in general but will also allow the existence Private Enterprises: Their Emergence, Rapid Growth, and Problems 187 and growth of private enterprises with large numbers of employees. The state should publicly acknowledge these enterprises' legal status through legislation and permit them to make the most of their potential by sup- porting their legitimate operations and protecting their lawful rights. * At the same time that complementary economic and political reforms are speeded up, efforts should be made to improve macroeconomic regu- lation. The govemment should exercise economic, administrative, and legal supervision over private enterprises in such areas as financial trans- actions, income distribution, and labor protection and welfare. For ex- ample, regulations on individual, cooperative, and private enterprises should be developed along with improved examination, registration, and licensing procedures; unified standards should be set for financial and accounting systems; new tax laws should be promulgated; bank loan policies should be adjusted and interest rates for loans raised; and tax and credit instruments should be used to give enterprises industrial guidance, regulate the polarization of personal incomes, and encourage enterprises to reinvest profits. * Trade unions should be set up in private enterprises to protect workers' legitimate rights and to help govemment departments supervise the firms' operations. * State industrial and commercial administrative departments and audit- ing and tax organizations should regularly examine enterprises' financial transactions and operational activities, and law enforcement institu- tions should mete out severe punishment to firms engaged in tax eva- sion and other illegal activities. * The government should respect the autonomy of private enterprises in choosing property ownership systems and forms of operation and man- agement. The gradual transformation of private enterprises into joint stock cooperative enterprises should be encouraged, with the under- standing that no harm will be done to the basic rights of the original owners and that the principles of voluntary participation and mutual benefit will be followed. Notes 1. Between 1984 and 1986 the total value of output of TVPS in Jieshou County went from Y103.75 million to Y137.57 million, a net increase of Y33.85 million. During the same period the total value of output of private enterprises rose from Y45.53 million to Y71.12 million. The increase of Y25.59 million accounted for 65.9 percent of the increment in total value of output of all TVPS. In 1986 the value of output of private enterprises accounted for 51.7 percent of the total value of output of TVPS in the county. Private enterprises' share in the total value of output of TVPS in Shangrao County went from 22 percent in 1984 to 50 percent in 1986, and these enterprises accounted for most of the increase in total value of output of TVPS. 188 LIN QINGSONG 2. The comparison is calculated from statistics on sample townships in Jieshou County. 3. According to estimates based on sample surveys of fifty-six private enterprises in Tianjin in 1985, loans accounted, on average, for 38.8 percent of total funds for private firms with a total investment of less than Y50,000, 43.6 percent for those with a total investment of Y50,000-Y100,000, and 69.9 percent for those with a total investment of more than Yl 00,000. 9 Entrepreneurship, Capital, and Ownership William A. Byrd The importance of entrepreneurship in modem economic growth and devel- opment is increasingly recognized, although there is much debate about its exact meaning and it does not fit well into the framework of neoclassical eco- nomics. Entrepreneurship, broadly construed, may be even more important in developing countries, where markets often function poorly, than in devel- oped countries. The astonishingly rapid growth and successful economic performance of the industrial TVP sector in China would not have been possible without a host of entrepreneurs and a great burst of entrepreneurial activity. A tremen- dous amount of creation of new enterprises and production of new goods has occurred, and functioning markets have emerged where none or only very weak ones existed in the prereform era. TVPS, other than the smallest firms, handle long-distance trade and depend largely on markets outside their own localities. Some TVPs engage in technological innovation, and many others successfully adapt advanced domestic and foreign technology. Entrepreneur- ial performance in the TVP sector has been especially remarkable in an envi- ronment in which ownership and property rights with respect to industrial assets are not clear and pure private ownership is rare except in the smallest concerns. Thus the most powerful reward for small-scale entrepreneurs in other countries-the ability to reap large gains from the "capitalization" of entrepreneurial success in their firms-is absent or at least sharply circum- scribed. One of the main bottlenecks faced by entrepreneurs and small-scale enter- prises all over the world is getting the financial capital they need to start operations and to expand. The capital allocation mechanism for TVPS is hence of great importance. A related question concerns the absorption of financial risk. Even a relatively dynamic system of entrepreneurship and in- vestment decisionmaking is bound to generate unsuccessful investments and enterprise failures. If failing firms are allowed to continue and to absorb addi- tional resources, as typically occurs in the state sector, the efficiency of in- 189 190 WILLIAM A. BYRD vestment will suffer. China's TVP sector has evolved mechanisms for risk- sharing and absorption of losses that seem to have worked reasonably well, at least in the advanced areas. The ownership system is inr.imately related to entrepreneurship, provision of capital, and risk-bearing. In private-ownership economies small-scale en- trepreneurs reap the profits from their activities primarily through ownership of firms. In sizable enterprises the owners ("capitalists") tend to be separate from the entrepreneurs and managers. How the various forms of private and nonprivate ownership in China's TVP sector structure the relationship be- tween capital provision, risk-bearing, entrepreneurship, and management so as to allow relatively efficient operations and decisionmaking on investment is a crucial question. Entrepreneurship and Owvnership There are almost as many definitions of entrepreneurship as there are schol- arly books on the subject (see Hebert and Link 1982). But since a working concept is a prerequisite for analyzing entrepreneurship in China's TVP sector, this section will derive a definition and delineate the scope of activities to be considered entrepreneurial. A primary theme in the literature on entrepreneurship is that the entrepre- neur is an active organizer of ithe factors of production. Schumpeter (1934), whose work on entrepreneurship and economic development has been ex- tremely influential, stressed th.at entrepreneurship is associated with innova- tion and occurs only in the case of "new combinations of the means of pro- duction" organized by entrepreneurs. Others view the concept more broadly. Kirzner (1973, p. 35) asserts that the essence of entrepreneurship is alertness to unexploited profit opportunities and that hence it is an essential part of the market process. Related to the scope of entrepreneurship is the question of its effect on markets. The Schumpeterian approach sees entrepreneurial activity as a destabilizing force that starts the process of "creative destruc- tion," the essence of economic development. Hayek (1948, ch. 2), ]Kirzner (1973, p. 127), and others view it as a stabilizing force that helps markets move toward equilibrium and makes the market process work more smoothly. An important distinction is that between entrepreneurship and financial risk-bearing. Where there is uncertainty and incomplete information, entre- preneurial activities are inherently risky. Some entity (often the one provid- ing funding for the activities) must bear the financial risk. The entrepreneur himself may take great personal risks and commit his own financial resources, but these resources are generally insufficient except for very small ventures. Thus there is a clear analytical distinction between the entrepreneur (the organizer of entrepreneurial activities) and the "capitalist" (the owner and provider of capital, whether an individual, a group, a public entity, or the Entrepreneurship, Capital, and Ownership 191 state). (This distinction eluded many early thinkers on the subject; see He- bert and Link 1982, pp. 112-13.) The working definition of entrepreneurship used here cannot hope to be comprehensive. Although it is recognized that entrepreneurship, broadly construed, occurs throughout a market economy, in minor as well as in major decisions and actions, the former will be excluded from the analysis. For the purposes of this chapter, entrepreneurship consists of activities that lead to new combinations of productive factors or otherwise transform the industrial landscape, in accordance with Schumpeter's stress on entrepreneurship as in- volving truly "new" and innovative phenomena. But in the case of China's TVP sector (as in developing economies in general), "newness" must be seen in a local, microeconomic context. Entrepreneurship need not involve any- thing new from a global or even national perspective; rather, it may adapt advanced technologies from outside, set up new enterprises to produce goods not previously made in the locality, and so on. Two activities that clearly fall within this working definition are the estab- lishment of new Tvps and the development of new products in existing ones. These have the advantage of being relatively clear-cut events for which some statistical evidence is available. The analysis will focus on enterprise creation and new product development as quintessential examples of entrepreneurial activity that will provide insights into entrepreneurship in general. Entrepreneurship in China's TVP Sector Entrepreneurship is thriving in China's TVP sector, as numerous case studies show. Although the examples presented here are of the more successful and striking instances of entrepreneurial activity, they are not atypical, and many similar cases must be omitted owing to lack of space. CASE STUDY I: WUXI PETROLEUM EQUIPMENT PLANT. In the mid-1970s an engineer in the Shengli ("Victory") Oilfield in Shandong Province on the east coast of China, roughly 500 kilometers from Wuxi, was looking for an enterprise that could produce from his rough preliminary design a drill pump for enhanced oil recovery. He had a friend in the Shenyang Aviation Insti- tute in northeast China, more than 1,000 kilometers from Wuxi, who in turn was acquainted with a person in an agricultural machinery plant in the sub- urbs of Wuxi city. That enterprise trial-produced one drill pump but gave up because it could not engage in batch production, costs were too high, and it did not see a good market for the product. An agricultural machinery plant in Dongjiang township, Wuxi County, found out about the drill pump through its close relationship with the suburban factory. Since it felt that the product had good prospects, it began trial production in 1975. The first pump was completed in October 1977, and in December it was taken to the 192 WILLIAM A. BYRD Shengli Oilfield for a test in the field. It was then certified by the Shengli Research and Development Institute. During the first several years only five or six pumps were made each year because of production problems and low demand. (Enhanced oil recovery was not yet widespread in China.) Batch production began in 1982 with 132 units, but only in 1983 did the enterprise start making a profit on this prod- uct. Losses in the first years were covered by profits from production of other goods, mainly welded steel tubes. The Communist party secretary for the en- terprise pushed development of the drill pump despite the early losses because he thought it had a good future. In 1984 the factory signed a long-tenn con- tract with Shengli to deliver at least 500 units a year in 1984-88. This prompted an administrative decision by the Dongjiang township government to split the existing firm into three enterprises: the original agricultural ma- chinery plant, the welded steel tube factory, and the Wuxi Petroleum Equip- ment plant. Production jumped to 700 units in 1984 (from 450 in 1983), and in 1985 output reac:hed 2,580. The Wuxi plant is the only specialized producer of this kind of equipment in China and accounts for the bulk of total domestic production-perhaps 70 percent.' The Wuxi factory is a good example of successful entrepreneurship in com- munity government enterprises. Entrepreneurial decisionmaking and activity take place largely within existing firms and under their managers. Profits from existing production finance initial efforts to develop new products. When a product has demonstrated its marketability and profitability, large amounts of funds may be needed to bring it into full-scale production. These come from community sources-the township government, the local bank, or both. The decision to establish a separate enterprise is administrative rather than entrepreneurial. Entrepreneurship of this type in community enterprises has been occurring all along, although there were many problems and ineffi- ciencies in the prereform period, as well as a long learning curve. CASE STUDY 2: AN ELASTIC RIBBON FACTORY IN JIESHOU. This firm was es- tablished in 1982 by five people, two of whom provided initial investment capital totaling Y450. Another founder, a farmer-entrepreneur, sold the new venture some used ribbon-weaving machines for Y50 apiece. The firm was established primarily to generate employment for the founders' children and relatives. The initial site was provided rent-free by the farmer-entrepreneur in return for a job for his daughter. Production took off rapidly. In 1982 out- put value reached Y40,000 and net after-tax profits Y6,000, and both dou- bled in 1983. Its head start and its efficiency have enabled the firm to do well in an increasingly competitive environment in which prices for ribbons dropped by about 20 percent between the late 1970s and early 1980s and have fallen another 30 percent since then. In July 1985 the factory director (one of the two initial investors) went to Shanghai to buy some machine parts. In a department store she noticed Entrepreneurship, Capital, and Ownership 193 that people were lined up to buy a certain type of decorated elastic ribbon. In August she went back to Shanghai and found out from the department store where the ribbons were made. She was directed to a township enterprise in Wuxi but was not allowed past the gate. In the local hotel she asked about other producers of the same kind of ribbon and was given the name of a vil- lage firm in Shazhou County, also in Jiangsu. The firm, which had 500-600 workers and (in the eyes of the director) the air of a county enterprise, also refused her permission to enter the factory. But some workers who were leav- ing the plant told her how to find the house of a deputy director, the founder of the enterprise. She went there, struck up a conversation with his wife, and remarked on how beautiful the couple's daughter was. Something like a family relationship was established when the Jieshou firm's director became a kind of "godparent" to the child. The Shazhou plant arranged for the producer of the ribbon-making ma- chinery to sell six units to the Jieshou firm. A technician was sent from the Shazhou plant to install the equipment, and some people from Jieshou were sent to Shazhou for training. In retum, the Jieshou firm provided large amounts of agricultural and sideline products at a low price for the annual spring festival in Shazhou. Production of the new elastic ribbons got under way in early 1986. Initially the Shazhou plant was to provide the necessary raw materials, but the technician who was installing the equipment stayed at the Jieshou director's house for over a month and built up a personal rela- tionship with her. He told her that the Shazhou plant was planning to over- charge her firm for the raw materials and introduced her to the original sup- plier, who agreed to sell the materials at a price slightly lower than that paid by the Shazhou plant. The director. of the Jieshou firm is the wife of a govemment official who had served in the public security bureaus of several counties before he was transferred to Jieshou in 1979. The director had seen that elastic ribbons were in short supply on the market and had wanted to go into that field for a long time. The ownership of the enterprise is ambiguous. The two investing founders took back their principal quickly and now earn no.dividends or in- terest. All after-tax profits are plowed back into the enterprise. Although the firm is officially listed as an individualljointly operated enterprise, it has a relationship with the local neighborhood committee, which occasionally provides administrative help. In practice, the director makes all the key decisions. CASE STUDY 3: A FRUIT-CANNING PLANT IN SHANGRAO. This example is ex- ceptional in that an entrepreneur moves across provinces. The director came from Zhejiang Province, where he had run a private garment business that was initially profitable but later lost money and was shut down. Before that he had been in purchasing and marketing. Orange canning was an important activity in his home community, but profits were declining because of in- 194 WILLIAM A. BYRD creasing production and competition. Shangrao, by contrast, had much cheaper labor, somewhat cheaLper oranges, and, as a "backward" area, tax ad- vantages for new enterprises. In June 1986 the entrepreneur went to Shangrao to look into the possibilities. He first contacted a township where he had a business relationship with a factory, but he found that the water supply, an essential element in fruit canning, was not good. Later, through an acquaintance in a county government bureau, he found the right spot, with abundant supplies of clealn water and large numbers of orange trees that had been planted several years earlier. Negotiations with township authorities went smoothly, and once they were completed the factory was put into operation within two and a half months. The Zhejiang entrepreneur invested Y80,000: Y50,000 in equipment and Y30,000 to renovate the factory building. (He reportedly borrowed Y20,000-Y30,000 from friends and relatives; the rest was his own money.) The building, valued at Y100,000, was provided rent-free by the township. In addition, with the help of county authorities the township arranged for a quota, or line of credit, of up to Y100,000 from the local bank. It also guaranteed the enterprise's supply of water and electricity and provided free housing for the director, his family, and some employees who were brought from Zhejiang. These employees included four master technical personnel, a cashier, two salesmen, and the director's two sons, who became rniddle- level managers. The enterprise now has 21 regular employees and vvill use up to 240 seasonal temporary workers. The primary benefits for the township are the employment generated and the 3.5 percent of gross output value that it receives. Another 1.5 percent goes into a pension fund, and 1.5 percent is for workers' bonuses. Given an anticipated gross profit rate of about 10 percent, this leaves about 3.5 percent of gross output value for the director, the residual claimant. As a new food-processing enterprise the firm can be exempted from both direct and indirect taxes for one to three years. The plant's annual production capacity is Y2 million, but the limited sup- ply of local oranges will hold down output in the first few years. (Some or- anges will have to be imported from Fujian Province.) Once local orange trees have reached full maturity, there will be no supply constraints. The di- rector anticipates recovering his investment costs in two years if the tax ex- emptions last that long. He plans to settle down and retire in Shangrao. The firm is officially registered as a jointly run collective firm contracted to indi- vidual management. The director is fully responsible for supply, production, and marketing, and the output is sold all over north China. These case studies illustrate the main features of entrepreneurship in Chi- na's TVP sector: alertness to unexploited opportunities (a characteristic of en- trepreneurship in general), the ability to bring together and organize the ingredients needed for successful production, personal commitment and willingness to take risks, and the use of personal connections [guanxi]. As in other developing countries, many services that are provided by markets Entrepreneurship, Capital, and Ownership 195 in the industrialized countries have to be arranged by entrepreneurs them- selves. Entrepreneurs' rootedness in their home communities is a common pattern. (Even the third case study sheds light on the community ties of TVP entrepreneurs. The person from Zhejiang made Shangrao his home, and the know-how and key personnel for his business came from his original commu- nity.) Entrepreneurship under Different Forms of TVP Ownership The shares of different forms of ownership in the value of TVP industrial out- put for China as a whole and for the four counties in 1985 are shown in table 9-1. TVCES have a dominant share in Wuxi (95 percent), a considerably lower share in Nanhai (74 percent), and even lower shares in Jieshou (49 percent) and Shangrao (65 percent). Production team enterprises account for a significant share of the total only in Nanhai. Private enterprises are pro- portionately most important in Jieshou. The figure for private firms in Shangrao is also high, but it reflects handicraft production and the lack of development of community enterprises rather than dynamism in the private sector. In absolute terms the private sector is most important in Wuxi (more than twice as large as in Jieshou), although its percentage share is smallest there. The locus of entrepreneurship varies for TVPS with different forms of own- ership. Four main variants can be identified: entrepreneurship by the com- munity government itself; managerial entrepreneurship in traditional com- munity enterprises; community and cooperative entrepreneurship in some villages and production teams (a rare variant); and private entrepreneurship. Community governments may become directly involved in entrepreneurial activities if there is no strong base of existing community enterprises (as, for example, in Shangrao) or if development of an entirely new product is under- taken and the idea did not emanate from and is not taken up by an existing Table 9-1. TVP Ownership Structure in China and in the Sample Counties, 1985 (percentage of GVIO for TVPs) Ownership All China Wuxi Jieshou Nanhaia Shangrao Township and town 45 48 36 43 43 Village 38 47 13 31 22 Production team 9 2 - 16 - Private 8 3 51 10 35 Total value of output (millions of yuan) 175,008 3,705 127 1,421 31 - Not given. Note: GVIO, gross value of industrial output. a. Share of gross revenues of industrial TVPs. Sources: Table 9-5 and 9-6 and information from fieldwork. 196 WILLIAM A. BYRD enterprise. Entrepreneurship in this case involves both setting up a new en- terprise and making a (locally) new product. Much more common in localities with a strong base of traditional commu- nity enterprise development is entrepreneurial activity by enterprise leaders, as illustrated by case study 1. The key entrepreneurial decision is to try out and develop a new product; the decision to establish a separate enterprise is administrative, although it does involve a large inflow of capital and hence financial risk-taking of a venture capital nature. Enterprises, as well as gov- emments, may also decide to develop new products that are not closely re- lated to existing ones. The ability of firms to develop and try out new prod- ucts varies greatly with the industry concerned; innovation is particularly easy in the machinery industry, so important in Wuxi. In some cases involving the establishment of village or production team enterprises, collective entrepreneurship by the members of the community seems to occur. Ideas emanate from the members and are discussed at village meetings, and in some cases the decision to go ahead is taken collectively. Collective entrepreneurship is most common at the production team. level. In the private sector decisions to establish firms or to develop new products are made by individuals or partnerships, who invariably also manage the firms. This is no different, at least in theory, from the pattern of small-scale entrepreneurship observed in market economies. The number of important new products that enterprises in our sample re- ported putting into production between 1980 and 1985 (or between time of establishment and 1985) provides some indication of entrepreneurial activ- ity. This information suffers from a number of deficiencies, the most obvious of which is its subjective nature-how main products are defined was left entirely to the firm. The ease of new product development also differs mark- edly for different industries and for enterprises of different sizes. Nevertheless, there appears to be significant, nonspurious variation among counties and ownership types. Table 9-2 provides summarys information on new product development. The raw data were weighted by the inverse of the number of years each enter- prise was in existence (up to five). Five identical observations were then cre- ated for each enterprise set up in 1980 or earlier, four for firms established in 1981, three for those set up in 1982, and so on. The resulting data corrects for the smaller amount of information provided by enterprises set up after 1980 and the shorter time these firms had to develop new products. The table indicates that private enterprises are much more likely to engage in extensive development of new products than are community enterprises at the township and especially at the village and production team levels. But TVCES in Wuxi are significantly more likely to engage in at least moderate new product development than are sample firms in the other three counties. This strongly suggests that the relationship between ownership and new prod- uct development is not spurious. Indeed, cross-tabulations for Jieshou show Entrepreneurship, Capital, and Ownership 197 Table 9-2. Development of New Products by TVPs, 1980-85 (percentage of total new product development in sample or subsample) Ownership or location None Moderate Extensive Full sample 37 37 26 Township and town 35 40 25 Village and production team 57 24 19 Private 13 35 52 Wuxi 25 51 24 Nanhai 29 41 30 Shangrao 62 24 14 Jieshou 34 32 34 Township and town 34 36 30 Village and production team 67 0 33 Private 14 39 47 Note: None, no new product development; Moderate, one or two new products in a five-year period; Extensive, three or more new products in a five-year period. Source: Enterprise Quantitative Questionnaire. that private enterprises are significantly more likely to engage in moderate and especially extensive new product development than are community en- terprises. Because of the data problems, these results should be interpreted with caution, but they are suggestive. Data on the establishment of new firms are difficult to interpret, particu- larly since the available information invariably is for net new enterprise cre- ation, after exits have been subtracted. Moreover, the ease with which new enterprises are created almost certainly varies systematically and inversely with size-most new firms start out small and grow big only later-and size is in turn related to ownership. Finally, decisions to create new community enterprises, especially township firms, are often merely administrative confir- mations of the successful development of new products by existing firms. Nevertheless, the same pattern holds as for the development of new products; Wuxi among the four counties and private enterprises in general are more active in new enterprise creation. In Wuxi the number of TVCEs nearly dou- bled between 1982 and 1985 and increased by another 11 percent in 1986 despite a difficult market and credit situation. Sources and Motivation of Entrepreneurs In the past many managers of traditional community enterprises came from the ranks of rural cadres at the village or production team level. This often was not conducive to good management or to successful entrepreneurship. In places such as Wuxi managers of township enterprises now increasingly 198 WILLIAM A. BYRD rise through the ranks of their enterprises or are transferred from other local firms. There appears to be little movement of community enterprise managers among townships, although movement among villages within a township does occur in Wuxi (see chapter 7). In other localities (for example, Nanhai) directors of village enterprises often serve concurrently as village leaders. This practice is nearly universal at the production team level, even in Wuxi. Many village and production team enterprises have been contracted to indi- vidual management or sold outright to private individuals. The backgrounds of entrepreneurs of this type clo not appear to differ greatly from those of pri- vate entrepreneurs (see below). The motivations and incentives of community enterprise entrepreneurs are complex. As is true of substarLtial non-owner-operated establishments in any country, direct pecuniary rewards are only part of the story. Promotion possi- bilities, the perquisites that come with managing larger firms, pride in one's accomplishments, recognition, and even the perception of helping the com- munity all may play a role. Formal pay incentives for community enterprise managers are typically linkecl to enterprise performance (profits or output growth), to the average pay of enterprise workers, or to both. Nonpecuniary incentives are all the mtore irnportant because there are still limits on how much the incomes of enterpriise leaders can exceed the average level. in the enterprise and the local comrnunity. There are five main sources of private entrepreneurs: lower-level commu- nity government cadres and former cadres; supply and marketing personnel for TVPS; employees of community enterprises, particularly those with some technical skills; community residents who have worked in state enterprises; and farmers, usually with somne business experience. All of these groups are likely to have useful personal networks. Because TVP entrepreneurs are so immobile, the traditional base of entre- preneurship in a rural locality is crucial in determining their supply. Existing community enterprise development is also important. Wuxi has a machinery industry that dates to before the 1930s, and its supply of potential en,trepre- neurs was augmented when, during the early 1960s and the Cultural Revolu- tion, large numbers of local people who had been working in big-city enter- prises returned home. Similarly, urban workers originally from Nanhai were "sent down" to their home villlages and there established private enterprises under the protection of county authorities. Prereform rural industrialization generated an important base of entrepreneurial talent in Nanhai and espe- cially in Wuxi, and commercialized agriculture was highly developed in both counties before 1949. Jieshou was a commercial center of some importance before 1949, and its human resources from that period were never entirely dissipated. In Shangrao, by contrast, there was almost no traditional base of entrepreneurship before 1949, and TVP development during the prereform period was minimal. The large numbers of university and high school gradu- ates generated by Shangrao's relatively good education system left the county Entrepreneurship, Capital, and Ownership 199 instead of augmenting the entrepreneurial base. Thus Shangrao, alone among the four counties, may lack an adequate supply of indigenous entre- preneurs. Making money is a motivation for private entrepreneurs, but it is never the only goal and is sometimes not the most important one. Nonpecuniary goals include providing employment and a secure future for family members and, to a lesser extent, other relatives; improving one's social and official status (for example, by becoming a second-class urban resident as many Jieshou entrepreneurs did); and even contributing to local community wel- fare and employment. Since the right to private industrial property and to the full financial re- wards of entrepreneurship are not secure, the incentives of private entrepre- neurs are distorted. They may use a large part of their firms' profits for their own consumption and housing, or they may demonstrate that they are good corporate and community citizens by plowing all surplus back into the enter- prise. For many, the possibility of changing their household registration status may make takeover of their firms by the community govemment palatable. The family motive is extremely strong for private entrepreneurs, especially since in many rural areas there is surplus labor and nonagricultural employ- ment opportunities are limited. This motive alone suffices to explain a great deal of small-scale private entrepreneurial activity. Capital Allocation and Financial Risk-Bearing This section surveys the different sources of financial capital that TVps draw on and examines how business risk is distributed in the TVP sector. The Capital Allocation Mechanism The sources of capital for TVPS can be divided into household capital, com- munity capital, and enterprise capital, each with a corresponding mechanism for capital allocation. These sources cut across the traditional debt-equity bi- furcation, which is reasonable since in China's TVP sector debt capital often bears a great deal of risk, bankruptcy is not well-established as a means of dealing with financial losses, the concept and role of equity capital are poorly understood, and the notion of individuals reaping large capital gains is not widely accepted. Household capital consists of private funds invested directly in TvPs by households. It includes investment by partners or proprietors, informal loans to private entrepreneurs (usually by friends or relatives of the lenders), pur- chase of bonds issued by community enterprises, and provision of capital in return for jobs. The community government regulates the terms and condi- tions under which individuals can invest in community enterprises, but this is not as true of investment in private enterprises. Household capital is avail- 200 WILLIAM A. BYRD able to finance small firms in significant aggregate amounts, particularly after the sharp increases in rural personal incomes in recent years. Many workers in our samples indicated on the Worker Survey Questionnaire a willingness to invest in an enterprise or buy stock (see chapter 13). But the amount of household capital that can be tapped by any one private firm is limited. Township enterprises can gain access to larger amounts of household capital by issuing short-term bonds with the support of the community government. Community capital consists of investment funds the allocation ancd use of which are determined or at least strongly influenced by community govern- ments, particularly at the township level. This capital includes local govern- ment budgetary funds for TVP investment (typically minuscule), loans of fiscal revolving funds to TVPS, profitLs from community enterprises pooled by town- ship or village industrial corporations, and bank and rural credit cooperative (RCC) loans. The use of retained profits of community enterprises may also be influenced by the community govermment. The Agricultural Bank of China (ABC) is nominally a monolithic state institution with a well-defined headquarters-branch structure and an exten- sive system of top-down credit planning. The RCCS, although supposedly community-based cooperative institutions, have been subject to adrninistra- tive supervision by the ABC. T hus the classification of ABC/RCC loans as com- munity capital needs to be explained. An important reform of credit planning that began in 1979-80 links the total value of loans each ABC branch is allowed to make to the amount of deposits it generates. The difference be- tween loans and deposits in a base year is calculated, and this gap determines the amount of lending in the following year. If deposits rise by a certain amount, loans outstanding can increase by the same absolute amount. The impact of this scheme has been somewhat diluted by bargaining, by the later imposition of separate redeposit requirements, and by the occasional imposi- tion of tighter credit controls by the central government in the interests of macroeconomic stability. Nevertheless, increments in local community de- posits for the most part translate into increments in loanable funds, and hence the level of lencding to TVPS is to a large extent determined by the level of local deposits. The sharp rise in rural personal incomes and bank deposits has made the banking system an important source of community capital. Township govermments take the view that local bank funds form part of the community capital base, and they participate in important lending deci- sions. In the more advanced areas bank offices and township governments have a largely symbiotic relationship. According to one RCC official in Nan- hai, "The RCC has good relations with township leaders. We go to them peri- odically, show them how much money we have, and ask them to think about how they want it lent out." Banks and RCCS do have the theoretical right to veto proposed loans that they find questionable. In Shangrao and Jieshou, where community govermments frequently misused bank funds, this right is Entrepreneurship, Capital, and Owmership 201 becoming more important in practice. In general, township ABC and RCC of- fices cannot lend money freely outside the township-a clear sign that their funds are still basically considered community capital. Enterprise capital consists of the retained profits, depreciation allowances, and other funds of TVPs as well as funds provided by other enterprises. Capital from other firms is largely trade credit, of indefinite maturity and highly un- stable. The allocation of part of the retained profits and other funds of com- munity enterprises is influenced by the community government, but enter- prises may have considerable autonomy in making smaller investments. The intemally generated investment funds of private enterprises and the portion of community enterprise funds that can be used autonomously are the sources most easily tapped by factory directors to finance entrepreneurship. Table 9-3 shows the liability side of the balance sheet for 200 large and Table 9-3. Sources of Funds for Large-Scale TVPs (percentage of total) At End End End June Source founding 1984 1985 1986 State capital' 4.0 1.8 1.6 1.7 Community capitaP 67.6 35.2 30.1 27.8 Bank loans 38.2 28.9 24.1 23.4 Contributions by community governments' 29.4 6.3 6.0 4.4 Enterprise capitalb 13.0 55.8 59.8 63.0 Own funds' 0.0 41.0 40.7 43.2 Other firmsd 13.0 14.8 19.1 19.8 Household capital 6.5 3.5 4.6 4.0 Enterprise's own workers' 5.4 2.8 3.6 3.1 Outside individuals 1.1 0.7 1.0 0.9 Unidentified' 8.9 3.7 3.9 3.5 Note: Figures for 1985 were adjusted to correct apparent arithmetic errors. a. Includes grants and loans from state agencies. b. The share of community capital for 1984-86 is understated and that of household capital is overstated somewhat because "own funds" includes nonrepayable capital provided by the town- ship govemment as well as intemally accumulated capital. c. Includes fiscal revolving funds lent to TVPs, which may be to some extent controlled at the county rather than the township level. Grant funds from the community govemment for 1984-86 are not included. d. May include some debts owed to govemment entities, since it includes payables as well as loans and capital investment provided by other enterprises. The bulk of this capital comes from outside the locality. e. Wages payable to workers and investment contributed by them. f. For time of founding, other funds; for 1984-86, other long-term debt. Source: Zhou Qiren and Huang Zhuangjun (1987), from a sample of 200 relatively large and successful TVPs in ten provinces. 202 WILLIAM A. BYRD relatively successful community enterprises, primarily township enterprises, at four points in time. 'The state accounts for only a minuscule share of the capital requirements even of i:hese larger TVPS run by higher levels of the ad- ministrative hierarchy. The critical importance of the community in provid- ing seed capital for community enterprises at founding is obvious. The community's share diminishes as enterprises develop and generate investment funds internally. The role of household capital in financing community en- terprises is relatively small. Seed capital for new private enterprises comes in the first place fiom the founders, who themselves mnay borrow money from friends and relatives. They may also borrow on the informal credit market and require new workers to contribute some capital. T he bulk of the relatively small capital require- ments of private enterprises at time of founding is thus met by household capital. Successful private firms can rely heavily on reinvestment of internally accumulated funds, and many have generated sizable productive assets pri- marily through intemal accumulation. But beyond a certain point self- accumulation may not be enough to finance rapid expansion, and the po- tential for tapping household capital is also quickly exhausted. Private enterprises then naturally turn to sources of community capital, particularly ABC or RCC loans. The balance sheets of relatively large and successful private enterprises are thus usually dominated by intemal accumulation and bank loans. There is a potenirial inconsistency between ownership, which is in re- ality private (although this may be disguised and the enterprise may have virtually no equity capital on its books),2 and capital allocation, which to a large extent involves community capital (exclusively debt finance). In tra- ditional community enterprises, by contrast, ownership and the capital allo- cation mechanism are much more in harmony. The role of prices (interesl: rates) in TVP capital allocation mechanisms is limited. In the case of household capital, interest rates on enterprise and community bond issues can be distorted upward or downward by community governments. Moreover, when the main benefit of providing capital is get- ting a job in a labor-surplus environment, the interest rate does not mean much, and in fact often no interest payment is involved. Community enter- prises do not pay an explicit dividend on equity. Capital from other firms is often transferred involuntarily (as trade credits) or in return for rewards other than interest payments (for example, in some compensation trade ar- rangements). For community capital, interest rates on ABC and other bank loans are fixed by the central government at below-market equilibriurm levels, and rates on RCC loans are only slightly more flexible. Capital from commu- nity govemments is in the form of nonrepayable grants or low-interest loans. In the more advanced areas such as Wuxi, capital seems to have been allo- cated relatively efficiently despite the lack of price signals. The main reason is probably that the community govemment has strong incentives for making the best possible use of its capital resources. In the first place, these resources Entrepreneurship, Capital, and Ownership 203 are limited, and the community government faces a relatively hard budget constraint. Most of the risk in loan-financed investments is borne by commu- nity governments rather than by banks or enterprises. The community gov- ernment is probably better acquainted with the enterprises and investment projects under its jurisdiction than a provincial or even municipal industrial bureau is with its subordinate enterprises. In any case, when discrete invest- ment projects are large in relation to the total amount of funds available, the interest rate may not be a meaningful guide for the allocation of capital. Financial Risk-Bearing Closely related to the capital allocation mechanism is the mode of financial risk-bearing. Entrepreneurial activities inherently involve risk, most of which is typically not bome by the entrepreneur personally. In the case of commu- nity capital, both the community government and the financial system can absorb risk. When the govemment provides nonrepayable funds to commu- nity enterprises, this is a form of risk capital. In the traditional system the township government also explicitly or implicitly serves as a guarantor for bank borrowings by township enterprises under its jurisdiction, and the vil- lage government does the same for village firms. If an enterprise shuts down, outstanding debts become the responsibility of its community government owner and are usually transferred to another enterprise under the same gov- ernment. The community government's ability to absorb risk by varying its public expenditures may be limited, but it can spread risk across its enterprises to increase the flexibility and ability to absorb losses of any one firm. This risk- sharing is usually ad hoc and informal and is most effective when the commu- nity industrial structure is relatively diversified. Fluctuations and losses that are not life-threatening to the enterprise are absorbed by varying the flows of profits between the enterprise and the community government and, through delayed loan repayments, between the enterprise and the bank. Other measures are needed to deal with severe losses, but only rarely is the firm completely shut down, since bank debts need to be serviced anyway. Only if there is no cash flow at all is it rational from the viewpoint of the community government to shut down an enterprise. The above description applies mainly to areas in which traditional commu- nity enterprise development has been relatively successful. Elsewhere much of the risk associated with TVP activities may be borne by banks or Rccs. The financial resources of community governments in the more backward areas are too limited to absorb much risk, and the financial performance of commu- nity enterprises may be so poor that they cannot credibly cover each other's losses and take over the burden of loan repayment. In Shangrao, for example, a large loan to a hydropower station was collectively guaranteed by all of the other local township enterprises. By the time the loan fell due, nearly all 204 WILLIAM A. BYRD of these firms were running losses and could not even repay their own bank loans, and so the guarantee was meaningless. In this situation bank loans become highly risky, and banks in effect provide risk capital. Banks and RCcs can absorb risk by writing off loans, possibly after selling off the assets of a defunct enterprise, and by allowing firms to delay repay- ments of principal and interest, sometimes indefinitely. The latter method is far more common, especially for community enterprises. Banks often have an incentive to prevent a community enterprise from failing because once it shuts down, there is no hope of recouping loans. In Shangrao ABC offices never call in an overdue loan of a community enterprise. Instead they allow indefinite delays in repayment and request only that the firm service a small part of its debt. Sometimes they even provide additional credit in the hope that the enterprise will be able to turn the situation around. Risk absorption by the ABC in Shangrao is facilitated by the unusual prac- tice of permitting enterprises to pay interest only on the portion of an over- due loan that is actually being repaid. This allows principal payments to be much larger than otherwise, and the total annual repayment of principal and interest may be less than the interest charges alone on the entire loan. With constant annual repayments, the amount of principal repaid decreases pro- gressively, and in many cases a loan will never be entirely repaid. ABc profits may be severely affected. But stretching out repayments and not charging interest when it is due does at least leave open the possibility that a loan eventually will be repaid if the enterprise returns to profitability. Many loss- making TvPs in Shangrao, however, have such severe problems that they will never recover. The risk-bearing and loss absorption functions taken on by the banking system in Shangrao severely undermine the banks' symbiotic relationship with the community governrnent. Banks become very conservative in their lending, and as they gain independence through banking system reforms they tend not to lend all of the money absorbed locally in the form of household deposits. Thus some important linkages are broken, and the severe problems of these localities may be exacerbated. Forcing banks to lend to questionable projects and poorly managed firms, as was done in the past with disastrous results, obviously is not the answer. Bank conservatism is a symptom, not a cause, of the problems of backward areas. The banking system in Jieshou has adopted a different strategy of more aggressive lending, especially to private enterprises, and strong actions to re- cover unrepaid loans by selling off the property of defunct firms. This is more feasible because the main asset of many smaller enterprises is their real estate (land and buildings) and because of the relatively lively real estate market in Jieshou. Banks' reluctance to lend to traditional community enterprises is evident, however. Table 9-4, which pulls together available information on the share of over- Entrepreneurship, Capital, and Ownership 205 Table 9-4. Share of Overdue Loans in Total Value of Loans Outstanding (percent) Share of Locality Date overdue loans Source of loans Wuxi End 1985 8 ABC system End 1986 17' ABC system Dongjiang township, Wuxi End 1985 6 ABC/RCC Jieshou End 1985 8 ABC, loans to TVPs only End 1985 10 ICBC, total loans Hebei township, Jieshou End 1985 8 RCC only Nanhai April 1986 23 ABC system End 1986 8 ABC system Pingzhou district, Nanhai April 1986 15 ABC/RCC End 1986 7b ABC(RCC Shangrao End 1986 46 ABC, loans to TVPs Note: ABC, Agricultural Bank of China; icBc, Industrial and Commercial Bank of China; RCC, rural credit cooperative. a. Less than 4 percent consisted of bad debts considered unrepayable in the long run. b. The share of late repayment loans would have been 17 percent if an unrepaid loan that had financed imports of fake calculators from Hong Kong had not been put in a special category. Source: Information from fieldwork. due loans in total loans outstanding in the four counties, indicates the impor- tance of risk absorption through delayed repayments. The magnitude of over- due loans in Shangrao is striking. Wuxi and Nanhai also sometimes have relatively high shares of late loans, but this is a temporary phenomenon largely related to fluctuations in market conditions, central government mon- etary policy, and credit controls. The willingness of the banking system to absorb risks of private enterprises is severely limited. After a period during which large loans to private enter- prises were encouraged, ceilings have been sharply reduced, and in many cases the bank requires a guarantee from another enterprise or entity. Al- though some borrowers default, the overall repayment record of private en- terprises is probably much better than that of community enterprises, as is suggested by the experience of Jieshou with its disproportionately high share of private enterprises. For household capital provided to private enterprises, any losses are borne by the provider of the capital. For household capital raised by community enterprises, repayment of principal and even the return may be guaranteed by the enterprise or sometimes the community government. Even so, the risk of late repayment has reportedly made individuals in Wuxi less willing to pro- 206 WILLIAM A. BYRD vide capital directly to enterprises and more willing to deposit their money in banks, even at a lowver rate of interest. TVPS' internally accumulated funds have considerable risk-bearing capacity if they account for a substantial portion of total capital. In the more devel- oped areas the return these funds earn is usually free to decline sharply in the face of market fluctuations. In poorer localities, however, community en- terprises may be forced to turn over funds to commtnity governments regard- less of whether they are earning profits (see chapter 17). Capital-often in- voluntary payables-provided by other firms also absorbs some risk. But trade credit is a highly unstable, short-run device, and in any case the receivables of the sample of firms shown in table 9-3 exceeded their payables (although by only a small margin in 19186). Risk-bearing mechanisms for TvPs have obvious imperfections, but they work reasonably well in the rnore advanced areas and for the most part func- tion far better than risk-bearing mechanisms for state enterprises. The flexi- bility provided by the pooling and absorption of risk by community govem- ments in the traditional community enterprise system may in part explain the success of places such as Wuxi. There are serious problems, however, in more backward areas. Interrelationships and Implications The interrelationships among the different elements discussed above can now be analyzed. Several combinations of entrepreneurship, capital provision, risk-bearing, and enterprise management have evolved in the TVP sector. (The concept of management, which has not yet been discussed, is used in this chapter as a catchall that includes the host of routine decisions an,d activ- ities required to keep a firm running on an even keel. Its scope is broad and includes some "smaller" entrepreneurial activities.) The Wuxi System In the Wuxi model the main locus of entrepreneurship is the enterprise itself, and the leaders of exisliing firms are the entrepreneurs. The early stages of new product development are financed by enterprises' own capital, which is allocated in small amounts in a fairly autonomous way by enterprise leaders. The main source of funds for creation of new firms and for large investment projects is community capital. In some respects the community government's role is similar to that of a provider of venture capital or a diver- sified corporation or holding company in a market economy. The commu- nity government also bears some risk and covers financial shortfalls (gener- ally only temporary) of specific firms. Finally, management is in the hands Entrepreneurship, Capital, and Ownership 207 of the enterprise leadership. Some managerial decisions-for example, hiring labor-may be affected or even determined by the community gov- emment, but the day-to-day conduct of business is left to the director. Many of these elements were already present in the prereform period, but some crucial ones have been added or reinforced since the late 1970s. One is the linkage between community bank deposits and lending. Greater auton- omy for enterprise management and the selection of directors with technical and managerial backgrounds rather than former community government cad- res have also helped to improve the efficiency of the Wuxi model and al- lowed Wuxi to remain in the forefront of TVP development. In addition, the fundamental change in the wage system from community work-points to enterprise-based wages with strong piece-rate and profit-sharing components appears to have brought about substantial improvements in incentives and efficiency. For traditional community enterprises in developed areas, the different ele- ments fit together consistently. Community ownership is consistent with the mechanism for allocating community capital and with risk-bearing by the community government, and relatively autonomous management (supported by appropriate worker incentives) promotes efficiency in operations. The hard budget constraint of community governments gives them a strong incen- tive to allocate community capital resources efficiently. Unsuccessful Community Enterprise Systems A somewhat different combination of these elements can lead to slow devel- opment and costly entrepreneurial failures in the community enterprise sec- tor. In the more backward areas that lack an established base of community enterprises, entrepreneurial decisions are made by community officials them- selves. The inherent difficulties of direct government involvement in entre- preneurship are worsened by the poor quality and lack of relevant experience of many community government cadres in these areas. Although the commu- nity government is forced to take a direct role in entrepreneurship, it is un- able to take responsibility for the financial risks of such activities, which are effectively borne by the banking system. Severe conflicts between the inter- ests of the banking system and the community government have arisen and have been exacerbated by the independence that banks gained in the reform period. Finally, in places such as Shangrao there are severe problems with excessive community government intervention in community enterprise management and with heavy financial burdens on enterprises (see chapter 17). The lack of a traditional entrepreneurial base and of experience with establishing and operating TVPS contributes to the difficulties in generating self-sustaining TVP development. 208 WILLIAM A. BYRD Private Enterprise Systems There are a number of variants among private enterprises, but in all cases the initial entrepreneurial impulse comes from the owners or founders of the firm, who in the early stages have full management autonomy. Household capital is the primary source of funds for new private enterprises, and risk is borne by the founders and providers of capital. The relationship among the different elements is thus consistent, probably reasonably efficient, and not too different from the configuration for small new firms in other coun- tries. This changes when private enterprises reach a certain size and their capital needs exceed what can be supplied from internal accumulation and the lim- ited household capital available. They then typically rely more on commu- nity capital from the banking system, which consequently bears some of the financial risks. This leads to a contradiction between private ownership and the mechanisms for allocating community capital and for risk-bearing. The insecurity of private ownership of substantial concerns exacerbates the prob- lem. There may be strong pressure to convert large and successful private en- terprises into community enterprises of one sort or another. This may not be opposed by the private owners because it offers a sounder political posi- tion, better access to community capital coffers, and improved social status for the owners and their families. Conversions are especially easy where, as in Jieshou, many private enterprises officially registered as collective firms in the early 1980s. In Nanhai conversion takes the form of a joint venture by the private en- trepreneur and the cornmunity government. The firm typically becomes a collective that is "contracted" to the owner-founder. The private partner often seems to retain some lcind of claim on the assets he provides to the new enterprise, as well as management authority and status as residua,l claim- ant. He may be able to solidify his position and perhaps even to reap some of his successful firm's large capital gains. In one case the private partner "contributed" all of his enterprise's large debts as well as its assets to the joint venture. Moreover, the assets consisted largely of inventories of goods that were selling very slowly, such as electric fan motors. In another case there may have been side payments to relatives of the owner who worked in gov- ernment offices. Since the collectivization of private enterprises involves only rather large firms, quantitatively this trend is overwhelmed by the rapid growth of the private sector as a whole. But as more and more private enterprises grow to respectable size, the problem could become more serious. Incentives for large private enterprises are distorted, and there is great scope for corruption and other irregularities involving community government officials. Moreover, in the areas where large private enterprises have been allowed to develop, the traditional community enterprise sector was generally unsuccessful owing to Entrepreneurship, Capital, and Ownership 209 the reasons mentioned above and to the predatory fiscal practices of local community governments (see chapter 17). There is no reason why these same problems will not arise in converted private enterprises. This seems to be occurring in Jieshou, where government agencies pad the payrolls of con- verted firms to relieve local employment problems. The director of a large (about ninety employees) food-processing factory asserted that the efficiency of his firm would decline if it became a collective because he would no longer have sufficient autonomy in relations with the local government. In another large private food-processing plant (150 employees), one-third of the workers were hired at the request of government cadres. If outright conversion of private enterprises into community enterprises is avoided, a somewhat different pattern may emerge. Banking institutions al- ready see well-established private enterprises as solid borrowers. If past expe- rience leads banks to distrust traditional community enterprises, as is true in Shangrao and to a lesser extent in Jieshou, the banks may increasingly provide loans directly to private enterprises, with significant implications for the future role of community govemments. This is only a potential long-term trend, since at present lending to large private enterprises is still heavily me- diated by community governments. Trends and Issues This section first looks at trends in the ownership structure of China's TVP sector. It then briefly discusses the future evolution of the local banking sys- tem and how it might affect TVP capital allocation and risk-bearing mecha- nisms. Finally, the section focuses on how to promote TVP development in the backward areas where it has largely failed so far and how to deal with the increasing number of large and successful private enterprises. Ownership Trends Table 9-5 shows trends in the ownership structure of Chinese industry as a whole over the past fifteen years. The two most striking features are the steady decline in the share of the state sector in gross value of industrial out- put (GVIO) in the 1970s and 1980s and the rise in the share of the TVP sector, which began accelerating in 1984. The latter trend makes the changing structure of ownership within the TVP sector all the more important. Unfortu- nately there is no breakdown below the village level, but the sharp rise in the share of below-village industry in 1985 and 1986 suggests that the share of private enterprises has indeed grown considerably. (Before 1980 virtually all of the village-and-below portion of GVIO must have been produced by vil- lage and production team enterprises.) Hence the share of private enterprises is rising steadily at the expense of community enterprises, and within the community enterprise sector the share of the top tier (township enterprises) Table 9-5. The Ownership Structure of Chinese Industry, 1971-78 (percentage of total GyVo) Ownership 1971 1975 1978 1980 1981 1982 1983 1984, 1985 1986 State 85.9 81.2 77.6 75.1 74.3 73.8 72.6 67.6 64.9 62.2 Urban collective 10.9 13.7 13.7 14.4 14.1 14.3 14.4 15.9 15.6 14.8 Urban individual . . . . . . . . . 0.0 0.0 0.1 0.1 0.2 0.3b 0.2 Urban other . . ... 0.5 0.6 0.7 0.8 1.1 1.2b 1.5 Rural nonstate 3.2 5.1 8.7 10.0 11.0 11.2 12.1 15.2 18.0 21.3 Township 1.6 2.6 4.8 5.4 5.9 6.0 6.3 7.7 8.1 9.3 Village I 6.0 6.8 7.5 B village 1.6 2.5 3.9 4.6 5.1 5.2 5.8 1.5 3.1 4.5 Below village i .5 3.1I' 4.5 ° Share of TVP GVIO Township 50.2 51.2 55.5 53.9 53.8 53.8 52.4 50.5 45.1 43.5 Village 39.6 37.9 35.2 Below village 49.8 48.8 44.5 46.1 46.2 46.2 47.6 9.9 17.0 21.3 ... Zero or negligible. Note: Percentages for 1971-80 were calculated from gross industrial output value figures in 1970 constant accounting prices, for 1981-83 in 1980 constant prices, and for 1984-86 in current prices. a. Since data for current prices are not available for 1984, the shares of ownership forms other than rural nonstate industry (for which data in current prices are available) were calculated from their constant price shares. b. Since nominal figures were not available, shares were calculated on the assumption that the relative shares of each in nominal individual and other industrial output value aree the samre a. the re-ea! shares.. c. Of this, 1.6 percent consisted of output of partnerships and 1.5 percent was output of individual enterprises. Sources: State Statistical Bureau (1983), p. 215; (1985a), pp. 20, 239; (1986a), pp. 130, 182, 224, 227; (1987b), p. 3; (1987c), p. 20; (1987d), p. 21. Entrepreneurship, Capital, and Ownership 211 is increasing or at least stable, whereas that of the lower tiers (village and especially production team enterprises) is declining. These trends mean that in addition to a shift from community to private enterprises, a "hollowing out" of the middle of the TVP sector may be occurring. There is a great deal of diversity among provinces, counties, and even local rural communities, however, and in some areas trends may be different. Information from the four counties, which are at least to some degree representative of the rest of China, illustrates the trends. Table 9-6 shows output shares of different types of industrial TVPS in Nanhai during 1978-86. After an initial sharp decline the share of township enterprises has stabilized and even increased somewhat, reflecting the county's policy of developing large "backbone" enterprises in every township. The share of village firms first rose rapidly and then fluctuated at about 30 percent. The share of pro- duction team firms has fallen sharply since 1984. The trend toward privatization of lower-level community enterprises in Nanhai is even more striking if the large number of such firms contracted to individuals is taken into account. Contracting gives the manager effective control over the enterprise and status as residual claimant and often serves as a prelude to outright sale of the firm to the individual concerned. Some- times the very act of contracting seems to involve an implicit transfer of own- ership rights. In one case a village enterprise was contracted to individuals who were required under the contract to turn over Y20,000 in profits to the village each year. The firm became highly profitable, and two years later the contractors bought it from the village, but they paid only Y10,000, the depre- ciated book value of the enterprise's assets. The price did not reflect the in- creased value of the firm (in income-generating capability) during the period of the contract. In effect, the right to capital gains from increased profitabil- ity was implicitly given to the contractors in the act of contracting. If this example is at all representative, a large part of the village and production team sector is already effectively under private management and control. In Xiqiao town, Nanhai County, nearly half of all industrial village enterprises Table 9-6. The Changing Ownership Structure of Industrial TVPs, Nanhai, 1978-86 (percentage of gross income of industrial TVPs) Ownership 1978 1980 1983 1984 1985 1986 Township and town 68 35 32 35 43 40 Village 11 39 36 37 31 35 Production team l 21 26 32 22 16 13 Private J 6 10 12 Source: Information from fieldwork. 212 WILLIAM A. BYRD and more than 90 percent of production team firms have already been con- tracted to individual management. In Nanzhuang town, Nanhai County, there was a wholesale privatization of production team enterprises and a flow- ering of private enterprises in 1984 after the Central Committee's Document No. 1 clearly permitted the development of private enterprise. In Shangrao, as can be seen from table 9-7, township enterprises, village enterprises, and production team firms have all seen their shares in the GVIO of the TVP sector decline precipitously. Community enterprises in Shangrao are in danger of being left behind if they cannot improve their efficiency. Private enterprises have the great advantage of being better able than com- munity enterprises to escape predatory fiscal practices by community govern- ments. Comparable data are not available for Jieshou, but they would show a similar "tipping" away from community enterprises and toward private en- terprises. In Jieshou's rural areas wholesale privatization of village and pro- duction team enterprises through outright sale or contracting occurred when the PRS was implemented in 1979-80. In Zhuanji township all village and production team enterprises were contracted or sold to individuals at that time. In Jinzhai township about half of the registered industrial township en- terprises are actually private firms that had received financial support from the township government. Production team enterprises are dying out in most if not all of rural China. This is not surprising, since with the implementation of the iRs teams lost their main economic and adiministrative function, the exercise of control over agricultural land. Village enterprises are also generally declining, espe- cially if contracting is taken into account. The private sector continues to boom, and township enterprises appear to be holding their own in many lo- calities and even continue to be the leading sector of TVP growth in places such as Wuxi, where village enterprises are also in a strong position. Table 9-7. The Ownership Structure of Industrial TVPs, Shangrao, 1978-86 (percentage of GVIO of TVPs) Ownership 1983 1984 1985 1986 Township and town 43 47 43 32 Village 41 30 22 18 Production team 16 1 0 0 Private 0 22 35 50 Partnershipsa 0 17 23 27 Individual 0 5 12 23 - Not applicable. a. It is likely that a number of production team enterprises simply switched to the joint- household (cooperative) category in 1984 without greatly changing their organizational or man- agement structure. Source: Information from fieldwork. Entrepreneurship, Capital, and Ownership 213 Analysis of Trends The analytical framework developed earlier can aid understanding of owner- ship trends. The communities to which village enterprises and especially pro- duction team firms belong appear suboptimal in size for their capital alloca- tion and risk-bearing roles: they can neither raise sufficient capital for large-scale expansion nor absorb much risk. Production teams, in particular, do not have enough firms to spread the risk to any great extent. In addition, since Rccs are township-level organizations, village and production team au- thorities do not participate much in determining their lending. And when the PRS was implemented, many teams simply wiped out their collective in- dustrial assets by selling or contracting their enterprises to individuals. Township enterprises have the best access to community capital and are part of a high-level risk-sharing system in large communities. The richer, more industrialized rural areas generate more community funds than are needed for basic local public expenditures, and these funds are available for risky TVP investments. Since townships are now established as the lowest level of govemment, township enterprises also have superior administra- tive connections. Private enterprises, however, have greater flexibility and stronger incentives for proprietors, and the pressure of rural surplus labor stimulates the creation of small enterprises. Thus firms at both extremes of the TVP spectrum have strengths that may give them a competitive edge over the TVPS in the middle. By this reasoning villages with more economic resources, more firms, greater diversification of product structure, and better administrative connec- tions with the banking system should be able to survive and even thrive. Villages in Wuxi for the most part fit this description, which partly explains why the village enterprise sector has done so well there. The integration of village enterprises into industrial administration at the township level also is greater in Wuxi than in other localities (see chapter 7). Hence Wuxi's vil- lage enterprise sector should be viewed to a large extent as an integral part of the township enterprise sector. By contrast, townships that are too small and lack economic resources and risk-bearing capabilities may stagnate be- cause they cannot put together the "critical mass" needed for successful devel- opment of community enterprises. Most of Shangrao's townships are in this category. Townships in Jieshou that are distant from the central town also are having great difficulties in developing community enterprises. Thus three different patterns of local TVP development contribute to the overall trends: maintenance of a dominant township enterprise sector in a relatively few counties like Wuxi; a steady shift toward private development at the expense of all levels of community enterprises, as in Shangrao and Jieshou; and, where a more diversified pattern of ownership persists, as in Nanhai, a tendency for the top and bottom tiers (township enterprises and private enterprises) to expand at the expense of the middle tiers (village and 214 WILLIAM A. BYRD production team enterprises). The tendencies toward polarization and pri- vatization of the TVP sector will have interesting implications if thLey con- tinue over a long period. The Future of the Community Banking System In the traditional community enterprise system local banks funnel commu- nity savings into investment projects and enterprises chosen by the commu- nity government. The banks are an integral part of the mechanism for allo- cating community capital, but at least in the advanced areas they bear almost no risk. Reforms in the banking system, changing ownership pattems for TVPS, and increasing factor mobility may drastically alter the role of local banks and RCCS, with consequences for investment financing and risk-bearing in the TVP sector that are important but hard to predict. As long as banks remain community financial institutions that do not freely lend funds out- side, measures to increase their independence in decisions on specific loans are not likely to generate great changes in localities such as Wuxi where the traditional community enterprise system is working well. The reforms may even improve efficiency by providing a further check on inappropriate capital allocation decisions. The linkage of loans and deposits and the near inability to lend outside the locatlity mean that banks, in effect, will still function to a great extent as commtnity financial institutions. Greater independence for the banking system in t]he more backward areas or where traditional commu- nity enterprise development is not working well has more serious conse- quences, since the banks are likely to become more conservative about lend- ing to community enterprises. To the extent that this results in less total lending within the cornmunity, capital could flow out of the community through the hierarchical banlcing system itself if not as a result of deliberate decisions by branches. Within the community banks may prefer to lend to private enterprises, which would accelerate the decline in the share of the community enterprise sector. If local banks increasingly pull capital out of their communities and invest it elsewhere, deposit mobilization, hitherto aided by community government support and campaigns, might suffer. The community government might even establish its own captive financial insti- tutions, which would siphon deposits away from the banks. Making interest rates more important in capital allocation in the TVP sec- tor would be difficult at present. Community governments strive to support TVP development regardless of interest rates, although higher rates do in- crease financial risks. Moreover, even rather high interest rates would proba- bly not attract a great deal of community capital out of the localities where it is generated. This is certainly true of the more advanced areas, where re- turns to local investment are high. Any intercommunity credit flows elicited by higher interest rates would be primarily from poorer communities to more industrialized ones. Entrepreneurship, Capital, and Ownership 215 Decentralization of ownership of banks and RCCs to the township level or even below might have little effect on relationships between banks and com- munity governments in such places as Wuxi. But it could easily promote con- tinued or even greater misuse of community savings by community govern ments in backward areas. Hence decentralization of bank ownership to this level might be dangerous. Decentralization of ownership in the banking system would be counterpro- ductive unless there is competition. But it may be hard to introduce much competition into the banking system at the township level in the near future. Where RCCS have been made more independent from ABC township credit offices, as in Jieshou and Shangrao, there tends to be a strict division of labor and clients between the two institutions. Any competition in the system usu- ally comes when the urban Industrial and Commercial Bank sets up branches in townships, as has occurred in Nanhai. Competition between community and noncommunity financial institutions would be unequal, since the com- munity govemment would invariably favor the former,3 but over time it might undermine the mechanism for allocating community capital. The banking system as presently structured is not well suited to bear a great deal of risk from TVP activities. First, liabilities (deposits) are risk-free, which makes it difficult for banking institutions to hold risky assets such as TVP loans, and banks have no reserve cushions to absorb losses from bad loans. Second, since banks are not owners or formal residual claimants on the in- come flows from TVPS, it is not appropriate for them to absorb a great deal of risk. For community enterprises the primary risk-bearing institution should be their owners, the community govemments. Over the longer term the question of how community govemments them- selves should diversify their risks will become increasingly important. In de- veloped areas such as Wuxi community govemments diversify risks by setting up numerous enterprises in different industries. Even so, Wuxi is buffeted by the changing fortunes of the machinery industry, and Xiqiao town in Nanhai County is even more seriously affected by market changes in the textile in- dustry. In backward areas community enterprise development is difficult and highly risky, and the limited financial and human resources of community govemments leave little room for diversification. The introduction of finan- cial instruments that could be bought and sold by community governments and that allow flows of capital among localities would help diversify risk. Moreover, such instruments would enhance capital mobility and help ease the capital constraints that limit firm size in the TVP sector. Backward Areas The misalignment of the different elements of the entrepreneurship-capital- ownership nexus contributes significantly to the problems of backward areas such as Shangrao. But perhaps even more fundamental is the problem of fiscal 216 WILLIAM A. BYRD predation by community governments in those areas (see chapter 17).4 If this problem is solved, the limited mobility of entrepreneurs among communities, localities, and regions could become a serious constraint. An area such as Shangrao has most of the ingredients needed for successful developrment of small-scale manufacturing inclustries: cheap, relatively well-educated labor, natural resources, and enough capital, including some aid from the central government, to get started at least. What it seriously lacks are entrepreneurs and their networks, as well as good, experienced managers. If these could be brought in from outside antd if the problem of fiscal predation could be resolved, Shangrao would have good potential for TVP development. Private Enterprises The problems and distortions caused by ambiguous government policy toward sizable private enterprises are becoming more and more evident. The irnsecur- ity felt by successful private entrepreneurs may adversely affect their incen- tives once their firms grow to a certain size. Moreover, it leads to all sorts of distortions and to great leeway for corruption in the relationships between these enterprises and community govemment officials. None of the ways of dealing with large private enterprises practiced in the four counties is particu- larly effective or efficient. The central government has made an important first step toward rectifying the problem by formally legalizing large private enterprises and removing the previous (unenforced) ceilings on the number of employees individual proprietorships could hire. But it will take consider- able time and effort to change ingrained patterns of status stratification and discrimination against private enterprises at the local level (see chapter 7). Clarifying and strengthening the property rights of large private entrepre- neurs is clearly necessary and would undoubtedly improve the situation in many respects, but it would exacerbate the inconsistency with the rnecha- nism for allocating community capital. A possible solution would be to have community capital supplied to private enterprises, particularly risk capital, carry certain limited ownership rights-most important, the right to reap profits and capital gains from subsequent development of the enterprise. Bank loans to large private enterprises could be converted into a form of nonvoting equity held by the bank that provided the funds. Since the private enterprises concerned are large and successful, the increased risk to the banking system should be manageable. Notes 1. The Wuxi plant's main competitor is a state-owned machinery workshop under Shengli Oilfield that produces several hundredi units per year. But the state firm's production costs are much higher, and quality is poor. 2. The common practice of "repaying" the founders' seed capital as soon as the enterprise Entrepreneurship, Capital, and Ownership 217 is financially comfortable leaves it with no equity other than intemal accumulation, and it is commonly thought in China that intemally accumulated assets belong to "the enterprise itself' rather than to the founders or owners. 3. But the ability of community govemments to influence local residents' choices of where to put their savings may be limited. Many rural people deposit their money in outside banks, sometimes because they feel that big-city banks offer greater security but more often because they do not want local bank or RCC personnel to know how much money they have. 4. Fiscal predation is related to the inability of community governments to absorb risk. Their largely fixed public expenditures are so great in relation to their normal revenues that they can- not tolerate much variation in the latter and hence cannot absorb much risk. Part III Performance A careful evaluation of the performance of China's TVP sector and its compo- nents is a prerequisite for an overall assessment of the sector and of the reform and policy options for the future. The three papers in this part contain a preliminary analysis of different aspects of TVP performance. In chapter 10 Wang Xiaolu focuses on capital formation and utilization and its impact on TVP efficiency and performance. He first looks at the exter- nal environment for capital formation. Originally TVPs depended on transfers of agricultural surplus (in several forms) to finance their development. Cur- rently, however, there is considerable variation in sources of funding, as a survey of the four sample counties indicates. In all cases community accumu- lation (including retained profits), bank loans, and payables are the main sources of funds, but their relative importance varies by ownership as well as geographically. As a result of administrative pressures and distortions (rather than purely economic factors), community enterprises rely on bank loans much more than do private enterprises. Wang then turns to intemal mechanisms for capital formation. Commu- nity govemments strongly press community goals on their enterprises, whereas enterprise managers tend to emphasize firm-level objectives. A large proportion of the profits of community enterprises is remitted to community govemments, and the proportion of these profits reinvested in the enterprises is highly unstable. As a result, firms depend increasingly on credit financing. Excessive community govemment levies on TVP profits have interfered with the smooth development of enterprises, especially in Shangrao. The rapid run-up of workers' wages and managers' compensation in many community enterprises also saps firms' reinvestment capacities. Wang next discusses capital productivity and the substantial differences among the four counties and among firms under different forms of ownership. Township enterprises generate fewer profits and tax payments per unit of capi- tal than other types of TVPS in the same locality, a phenomenon for which Wang suggests several possible explanations. Capital productivity has de- 219 220 PERFORMANCE clined in the township enterprises of Wuxi and Nanhai, but it has risen in Jieshou, from a much lower initial level. Wuxi and Nanhai have also seen substitution of capital for labor, a tendency that has not been significant in Jieshou. These different pattems are related to local labor market con- ditions-as soon as local surplus agricultural labor is absorbed, communi- ties tend to increase capital intensity rather than hire additional workers from outside. Wang argues that capital intensification, although rational from the viewpoint of individual comLmunities, is inefficient from a natiorkal per- spective. The chapter closes with a discussion of the continuing inability of TVPS to attain larger firm sizes and of the consequences-high costs, scattered loca- tion patterns, excessive competition, and other problems. The persistence of small firm size is traced to the community orientation of TVPS, which limits their expansion. In chapter 11 Jan Svejnar looks at productive efficiency and employment in Chinese TVPS on the basis of theoretical models and econometric ap- proaches developed for the study of cooperative firms in other countries. His analysis of productive efficiency is based on Cobb-Douglas and translog pro- duction functions, augmented by dummy variables that represent different counties, ownership forms, and worker compensation schemes. The analysis of labor utilization is based on an enterprise objective function equation, which is specified as a geometric average of three possible enterprise goals: maximizing profits, generating employment, and increasing average wages above local market-clearing levels. The next section presents the main empirical results. An important finding is that, other things being equal, ownership appears to be unrelated to pro- ductive efficiency. The coefficients on county dummy variables indicate that the productivity of sample firms in Wuxi, Nanhai, and Jieshou counties is more or less the same but thLat firms in Shangrao lag significantly behind. The substantial technical prcigress over time confirms that intensive rather than merely extensive growth has been responsible for the TVP sectDr's ob- served successful performnance. The coefficients on labor and capital suggest that TVPS operate in the zone of mildly increasing retums to scale. Finally, sample TVPS that use a fixed-wage-plus-bonus system or an intemal work- point system tend to be more productive than those that use pure piece rates, a combination of fixed wages and piece rates, or a fixed wage plus year-end dividend. The analysis of labor utilization practices also yields some significant re- sults. In general, TvPs appear to link employment closely to output levels and to economize on labor as wages rise. Thus they recognize the tradeoff between wages and employment, and featherbedding seems not to be a signif- icant phenomenon overall. (This last result contrasts with qualitative evi- dence of overstaffing in some enterprises reported in chapter 14.) Significant associations between ownership forms and compensation schemes are lack- Performance 221 ing, except that private enterprises appear to have paid more attention to local alternative wage levels in their employment decisions in recent years. The chapter closes with a reiteration of the most important findings: own- ership does not appear to be systematically related to productive efficiency; group incentive schemes may be associated with higher productivity than in- dividual incentive schemes; the behavior of local authorities may have a sig- nificant influence on TVP productivity, as is indicated by the relative ineffi- ciency in Shangrao County; TVPS seem to operate under mildly increasing returns to scale and have on the whole experienced rapid increases in produc- tivity over time; and there is some evidence that the employment practices of TVps have been shifting away from the objective of employment generation as such and toward employment levels that reflect the marginal productivity of labor. In chapter 12 Wang Tuoyu undertakes an empirical analysis of regional pat- terns of TVP development, assesses recent and current trends, and draws some policy implications. He first looks at indicators of TVP development and effi- ciency for three groups of provinces, classified according to whether their TVP sectors are developed, developing, or underdeveloped. Wang shows that the level of development of China's TVP sector declines progressively from east to west. Moreover, there are big differences in the internal structure of the TvP sector among the three regions. The share of industry is highest in the developed region, whereas construction plays an important role in the under- developed areas and commerce is relatively more important in the developing region. In general, the developed areas emphasize capital- and technology- intensive industrial activities, whereas in the underdeveloped region labor- intensive activities and industries based on natural resources are more im- portant. The chapter moves on to review recent growth performance in different regions. One of Wang's most important findings is that the gaps in TVCE de- velopment between more developed and less developed regions have been widening. In the TVP sector as a whole, however, this tendency was offset in 1984-85 by the rapid growth of private enterprises, which showed more regional balance. Wang then looks at the prospects for reducing regional imbalances in future TVP development. After reviewing the sources of investment financing for TVPS, he concludes that the possibilities for redirecting resources to the un- derdeveloped region are limited and that there is a danger that regional im- balances will grow worse. A number of policy measures to ameliorate the situ- ation are put forward: free interregional flows of labor and of other factors of production should be encouraged; private enterprises, which have already played a significant role in reducing regional imbalances, should be allowed to develop further; and the limited amount of state support available for the more backward areas should be targeted at sound investment projects and at areas with low per capita incomes and poorer agricultural resources. 10 Capital Formation and Utilization Wang Xiaolu China's industrial TVP sector has developed rapidly over the past several years-much faster than state industry in number of enterprises, value of out- put, and profits. Community enterprises, especially those run by townships and villages (TVCES), form the mainstay of the industrial TVP sector, whereas individual firms and partnerships comprise the private sector. Enterprises run by production teams share some of the characteristics of community enter- prises but increasingly resemble private enterprises in many ways. This chap- ter analyzes capital formation and capital productivity in China's TVE' sector and especially for TVCES. The External Environment for Capital Formation The sources of funds for TVP industry vary considerably among counties. The capital structure of sample firmis in 1985 is shown in table 10-1. After short- term funds such as payables are deducted, the principal and most stable sources of capital are community financial accumulation (including internal accumulation by the firmIs themselves) and bank loans. This section will look at external sources of capital. A subsequent section discusses internal accu- mulation, which TVPs also rely on heavily for expansion. Agricultural Surplus and i'nitial Rural Industrial Development In its early stages TVP industry depended greatly on capital from agriculture. Since Wuxi and Nanhai counties were agriculturally developed, TVP ilndustry there got an early start and developed rapidly. In Shangrao and Jieshou coun- ties agriculture was backward and TVP industry developed slowly. Agricultural surplus is turned into investment for TVP industry in three ways. * Mandatory agricultural accumulation through communes and brigades. This 222 Capital Fornation and Utilization 223 channel was important during the people's commune period but no longer exists. * Personal savings. Most new private enterprises use this channel to get established. Community enterprises often raise funds by requiring new workers to invest in the firm. * Support from financial institutions. The level of lending by banks and credit cooperatives is linked to local deposits. Earlier, the agricultural surplus had a decisive influence on the supply of credit funds, but in some developed regions bank savings now come mainly from industrial income. The Significance of Credit Funds The proportion of bank loans in total capital varies among counties and has changed over time (table 10-2). Wuxi and Nanhai counties are both eco- nomically developed but represent two different patterns. Wuxi's dependence on bank loans was extremely small in the early stages but has increased stead- ily over the past several years. To satisfy demand from TVPS, local banks are systematically borrowing from banks in other areas that have surplus funds and are lending this money to local enterprises. Nanhai TvPS relied heavily on bank loans from the very beginning, and their dependence has continued to grow. In general, bank loans have occupied a dominant place in enterprise capital. Of the two less developed counties, Jieshou relies heavily on bank loans, while Shangrao is the least dependent on bank loans, mainly because numerous investment failures there destroyed local TVPS' credibility and banks tightened controls to cut down on risk. In general, the scale of credit is closely related to local bank deposits. The Table 10-1. Sources of Funds for Sample Firms, 1985 (percentage of total capital) Source Wuxi Jieshou Nanhai Shangrao' Township and village government fundsb 43.5 33.3 22.8 41.3 Bank loansc 16.4 24.1 49.7 4.2 Payables 19.4 13.3 12.0 17.9 Other' 20.7 29.3 15.5 36.6 a. Based on aggregate statistics for six townships in Shangrao rather than for sample firms. b. Includes both grants from community govemments and intemal accumulation by TVPs from retained profits. In addition, sample firms tended to include in this category capital from "owners," even if they were private individuals. c. Includes loans from rural credit cooperatives. d. Includes wages payable, state support funds, loans from govemment departments, capital from individuals and groups, special funds, depreciation funds, and miscellaneous funds. Sources: Enterprise Quantitative Questionnaire and information from fieldwork. 224 WANG XIAOLU Table 10-2. Bank Loans as a Share of Total Capital of Sample TVPs, 1970-85 (percent) Year Wuxi Jieshou Nanhai Shangrao' 1970 4.4 38.3 n.a. n.a. 1975 0.5 44.0 n.a. II.a. 1978 0.9 34.2 n.a. n.a. 1980 8.8 20.7 32.2 7.2 1981 9.0 27.6 34.4 n.a. 1982 7.3 24.8 42.0 n.a. 1983 10.7 26.8 36.4 n.a. 1984 23.0 30.2 50.5 n.a. 1985 16.4 24.1 49.7 4.2 n.a. Not available. a. Based on aggregate statistics for six townships. Sources: Enterprise Quantitative Questionnaire and information from fieldwork. banking system is highly regionalized. Local deposits constitute the basic source of funds for loans to support the growth of local TVP industry, and cross-regional financial exchanges are limited. This creates a feedback: mech- anism whereby wages, personal bank deposits, and loans to local TVPS ratchet upward. Once bank deposits exceed a certain level, they can have a snowball effect on the development of rvp industry. All of the sample counties except Shangrao have reached this stage. This kind of capital circulation has some weaknesses. Lack of a competi- tive mechanism (the result of limited cross-regional capital mobility) affects the optimization of factor allocation and reduces the efficiency of capital uti- lization. Soft Constraints on Credit Fumls TVPs' dependence on bank loans varies greatly with ownership. Community enterprises depend much more heavily on bank credit than do private enter- prises. As the figures below, from the Enterprise Quantitative Questionnaire, show, the share of bank loans in total capital for the sample private enter- prises in Jieshou was consisterntly lower than for the community enterprises.' Percentage of total capital Type of enterprise 1983 1984 1985 1986 TVCES 25.4 29.1 33.3 25.6 Private enterprises 21.0 13.5 14.5 17.7 The discrepancy is even greater in Nanhai, where township enterprises de- pend more heavily on bank loans than do village enterprises.' (Lending to private enterprises was minuscule during 1980-85.) Capital Formation and Utilization 225 Percentage of total capital Type of enterprise 1981 1982 1983 1984 1985 1986 Township enterprises 33.4 39.6 45.9 40.3 60.4 46.7 Village enterprises 9.6 7.2 8.3 4.3 13.7 16.6 Private and production team firms 0 0 0 0 3.5 0 Bank loans have become an essential source of funds for the normal busi- ness operations of township enterprises. This is not because township enter- prises are less risky investments or have better credibility than other types of firms. Indeed, the profit rate on capital for township enterprises in the four counties is invariably lower than that of other types of TVPS, and the proportion of township enterprises in financial difficulties is higher. In 1985, 11 percent of industrial township enterprises in Wuxi suffered losses, and the loss rate (the ratio of total losses to aggregate profit) for all township enter- prises was 3.2 percent, whereas that for village enterprises was only 0.6 per- cent. In Nanhai 18 percent of industrial township enterprises were operating in the red in 1985, and the loss rate was 9.1 percent. Only 1.5 percent of village enterprises suffered losses, and the loss rate was 0.7 percent. In 1986, 30 percent of industrial township enterprises in Nanhai had losses, and the loss rate rose to 61.3 percent, but only 3.1 percent of Nanhai's village enter- prises had losses, and the loss rate was only 1.9 percent.3 Money-losing TVCES typically stay in business, despite their inability to repay debts, for two reasons. First, the community government can shift the responsibility for repaying the debts of a specific firm to its other enterprises, which reduces the risk for both the community government and the bank. Second, in the absence of a capital market, the bank provides the only access to capital for enterprises, and banks keep investing in debt-ridden TVCES be- cause of pressures from community governments acting from nonprofit mo- tives. Moreover, since the principle of a regional balance of loans and depos- its has severed horizontal links among banks, local banks can hardly find any better lending channels on their own. Bank loans are obviously beneficial for the enterprises. Interest rates are low and rarely fluctuate with changes in supply and demand (although there has been some tendency for nominal rates to rise in recent years); part of the loans can be paid off from before-tax profits, which in fact transfers the burden of repayment to government fi- nance; and if firms run into operational difficulties, they can postpone repay- ment. The Agricultural Bank of China (ABC) branch in Jieshou reports that by the end of 1985 the county's TVPS had Y2.09 million in overdue loans. This figure included Y490,000 in bad debts owed by 124 firms that had closed. Overdue loans accounted for more than 20 percent of the Y9 million in out- standing loans to TVPS in the county, and bad debts exceeded 5 percent of the outstanding balance. In 1986 the bank took steps to press for repayment, 226 WANG XIAOLU deduct guarantors' money, and clear bankrupt firms' assets, but it got back only 26 percent of overclue payments and 18 percent of bad debts. The lack of laws and regulations to protect the legal rights of creditors is another important reason for the worsening situation. In general, TVPS (mainly township enterprises) have little trouble obtaining credit, and this partly explains their heavy dependence on debt in their operations. This method of "debt management" (see Zhou Qiren and Huang Zhuangjun 1987) has enabled TVP industry to develop rapidly, but it has also meant that some firms have a shaky financial structure. Severe fluctuations occur because the growth of TVP industry is only too easily affected by changes in the supply of credit. Furthermore, the availability of abundant credit has masked poor management in some firms. Unprofitable or money-losing firms have been able to coexist with successful enterprises and have wasted large amolnts of scarce capital funds. The Internal Mechanism for Capital Formation Capital formation within firms is influenced by the claims and priorities of community governments and by the enterprises' own capacity to make and reinvest profits. The similarities and differences between the goals of commu- nity governments and community enterprise managers as well as the influ- ence that governments have on enterprises are important issues. Township Goals and EnterPrise Objectives In the four sample counties township govemments exercise supervision over subordinate firms directly or through organizations such as township in- dustrial corporations (TICS). TDwnship governments usually have authority to appoint factory directors, inspect factory performance, set annual targets for profits and value of output, allocate labor and manage wages, guarantee enterprises' bank loans, act as liaison agencies in expanding enterprises' busi- ness ties, and help enterprises sell part of their production. The relations be- tween township governments and their enterprises resemble in some ways the ties between state institutions and state-owned enterprises and in other ways those between a general corporation and its branch companies or factories in a market economy. But unlike state institutions on the one hand, town- ship governments face a competitive market that does not allow them to pro- vide monopoly protection to township enterprises. They cannot implement mandatory production plans or determine prices. On the other hand, they differ from market-economny corporations in that they are not merely superior managers or parent corporations of township enterprises. As government in- stitutions, they have multiple objectives, not just profit maximization or en- terprise development, and they often try to push their noneconomic goals on firms. Capital Fornation and Utilization 227 According to responses to the Township Leader Questionnaire, township governments see TvP development as having three main objectives: providing employment opportunities for surplus rural labor in the township, raising local living standards, and increasing township governments' financial reve- nue. Communities that face a surplus labor problem give priority to increas- ing employment; otherwise, the emphasis is on the latter two objectives. The survey results show that township governments act as strong representatives of their communities. They expand their revenues through profit remittances from the firms, and they seek to increase employment and raise local living standards by requiring township enterprises to absorb more labor, by support- ing unprofitable or money-losing firms, and by transferring part of enterprises' profits to local residents. There is a premise that income within the commu- nity should be relatively balanced, and local TVps are asked to pursue steady increases in wages for their workers. Since the employment and welfare tar- gets obviously run counter to profit maximization, township governments weigh the tradeoffs and often sacrifice the profit motive. The industrial development targets set by village leaders are usually similar to those set by township govemments. The difference is that villages may have less power to manage and control enterprises and have less money for public expenditures at the village level. But village leaders are often prompted by a stronger desire to maximize per capita income in the village. Although TVCEs are subordinate to community governments, the objec- tives of TVcE directors differ sharply from those set by these governments. In response to a question about objectives, almost all directors of sample firms (the bulk of which were TVCES) placed "pursuing long-term and stable devel- opment of the firm" first among nine choices. Other targets singled out by directors included "creating famous brand products and improving the enter- prise's credibility," "pursuing maximum profits," "improving workers' per cap- ita income level," and "improving the intemal management of the enter- prise." These choices are all related to the prosperity of the enterprise and its workers. Factory directors placed at the bottom of the list the objectives that are the main ones for township governments-increasing employment, promoting local prosperity, and increasing financial revenue. The motives of enterprises fall into three main categories: long-term devel- opment, maximum profits, and maximum per capita income within the en- terprise. Thus TVP managers have the same set of motivations and objectives as their counterparts in a typical market economy, although in their pursuit of maximum per capita income they exhibit the characteristics of cooperative firms to a certain degree. The differences between the targets of community governments and those of their subordinate firms have affected enterprise operations and have led to some restraints on firms' behavior by community govemments. For in- stance, community govemments' pursuit of employment opportunities may lead to overstaffing of local community enterprises, which affects profits. Table 10-3. Shares and Uses of Profits Remitted to Township and Village Governments by Wuxi Industrial TVCEs, 1977-86 (percent) Item 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 Share of total profits remitted to community governments 82 86 85 66 47 51 58 54 35 40 Use of profitsa Subsidies to agriculture 7 27 19 25 35 36 29 0 0 12 Agricultural investment 20 12 9 12 23 21 7 19 10 0 Subsidies for agricultural mechanization and water conservancy 21 5 3 9 12 8 6 7 6 7 Village and town construction 5 6 4 5 11 13 13 14 16 19 Culture, education, and public health 3 3 2 3 5 6 7 8 9 10 Industrial investment 38 40 52 44 0 0 16 24 27 21 Cadre salaries 0.1 0 3 3 8 9 9 13 11 12 Other 0 8 8 0.1 7 8 12 15 21 19 a. Includes profits retained by enterprises, as well as remitted profits. Numbers may not add to 100 percent because of rounding. Source: Information from fieldwork. Capital Formation and Utilization 229 Community governments' desire for higher financial revenue may lead them to draw off huge amounts of profits for government expenditures and so ham- per the enterprises' long-term development. The community enterprises' pur- suit of maximum per capita incomes may in its turn create sharp discrepancies between the incomes of enterprise workers and those of other members of the community. Community governments may respond by acting to keep workers' wages from growing too rapidly, transferring part of community en- terprise profits to other members of the community, or forcing firms to in- crease their accumulation. Distribution of Profits and Community Government Investment The difference between the profit motive of township governments and that of the owners of capital in a capitalist system can be seen in the distribution of township enterprises' profits and the expenditures of township govern- ments. The method of distribution, which obviously affects enterprises' rein- vestment rate, is a key element of enterprise capital formation. In the four counties surveyed, a large part of township enterprises' profits are usually turned over to township governments for general distribution, leaving only a small portion for reinvestment. Table 10-3 shows the share of profits remitted to township and village governments by TVCES in Wuxi for 1978-86 and the use of these profits. The proportion of profits handed over has fluctuated between 35 and 86 percent but has been declining over the past several years. The amount of remitted profits that goes for subsidies to agricultural production and distribution to community members has fallen sharply. Meanwhile, the proportion spent on township construction, cultural and health projects, and cadres' salaries and bonuses has increased steadily, from 18 percent in 1977 to 41 percent in 1986. The proportion of other spending has also increased, to 20 percent. The share devoted to reinvest- ment in TVP industry has been unstable. It has fluctuated between 0 and 44 percent and has fallen since the late 1970s. The situation in Wuxi suggests that in relatively developed areas public welfare and administrative expenditures are replacing support for farming as the main way in which TVCES' profit remittances are utilized. The declining proportion of reinvestment in industry is making TVCES increasingly depen- dent on credit financing. Surveys indicate that in the other three counties profits drawn by township governments from industrial TVCES were seldom used for reinvestment. The proportion of profits handed over to township governments in Shangrao seems to be larger than that in the other counties. Jieshou, like Wuxi, can be considered an exception. After several local TVCES had failed, the county began to take a laissez-faire attitude toward the local TVP sector. As a result, private enterprises have developed rapidly, and TVCEs enjoy a more comfortable environment. According to statistics for thirty-six sample firms in the county, profits remitted to community govern- 230 WANG XIAOLU ments amounted to only 2-5 percent of total net profits in the past several years. Even when management fees are taken into account, the figure is only about 20 percent of total net profits, which leaves large amounts for reinvest- ment. Excessive levies on TVPS' profits have hindered their smooth development. The problem has been especially prominent in Shangrao, where enterprises are moving ahead at a snail's pace, since they must submit a large part, if not all, of their profits to comrnunity govemments. (Some firms even handed over their depreciation funds when they failed to fulfill the profit target and had difficulty in obtaining bank credit.) The average annual growth rate of TVP industrial output between 1978 and 1985 in Shangrao was only about one-third the level in the other three counties. In developed regions such as Nanhai and Wuxi enterprises are able to obtain an ample and cheap supply of bank credit, which to a great extent helps make up for the shortfall caused by profit remittances, and community governments are willing to act as firms' guarantors to improve their access to credit funds. Guarantees of this kind, however, are more often than not pledged without any collateral. Enterprises also acquire funds through tax reductions and exemptions. In 1985 industrial TVCES in Wuxi were supposed to have paid Y143.4 million in indirect taxes and Y67.5 million in profit taxes. But in fact, reductions and exemptions of Y57.3 million on the former and Y19.6 million on the latter decreased the total tax paid by 27 percent. This is one of the ways in which community governments uphold local interests under the existing government finance system. Distribution of Profits and Self-Accumulation in Enterprises Statistics from sample firms show that a sizable portion of the profits remain- ing after remittances to the community government is used to invest in fixed assets and replenish floating capital and that only a small part is distributed to individuals. The reinvestment rate in Wuxi and 3ieshou counties is higher than in the other two counties. Nanhai devotes more retained profits to bo- nuses than does Wuxi or Jieshou (see table 10-4). Differences in mechanisms for determining wages caused huge gaps in county distribution patterns. From 1980 to 1985 the average annual growth rate of income per worker in Jieshou's sample enterprises was a mere 8.2 per- cent, as against an average annual growth of gross profits per worker of 41.6 percent. Average income per worker climbed by 21.0 percent in sample firms in Nanhai and by 16.8 percen: in Wuxi, whereas the average annual growth of gross profits per worker was only 7.0 percent in Nanhai and 9.6 percent in Wuxi (see chapter 14). There are many reasons for these differences. In addition to variations in labor demand and supply conditions and per capita incomes in agriculture, Capital Formation and Utilization 231 Table 10-4. Uses of Retained Profits by Sample TVPs, 1985 (percentage of total retained profits) Use of funds Wuxi Nanhai Jieshou Working capital 39.0 12.4 49.1 Modernization and development 53.2 63.3 34.0 Total 92.2 75.7 83.1 Collective welfare 1.7 1.9 8.5 Bonuses 7.7 22.3 5.9 Total 9.4 24.2 14.4 Note: Totals are not 100 percent owing to unclear uses of profits and inconsistencies in the re- ported figures. Source: Enterprise Quantitative Questionnaire. there are differences in operational behavior. In the Wuxi and Nanhai sam- ples, community enterprises, which tend to pursue higher per capita income, greatly outnumber private enterprises. Moreover, township leaders, in striv- ing to maximize average incomes in the entire community, permit commu- nity enterprises to raise wage levels rapidly provided that the growth of their incomes is well coordinated with income growth in the community as a whole. In Nanhai community govemments regulate the salaries of enterprise directors on the basis of profits and thereby indirectly control workers' wage levels through factory leaders' interest in profits. This kind of control is rela- tively weak. In Wuxi community governments set the per capita wage level target for firms in advance, in accordance with the principle of "appropriate progressive growth." Workers' wages are linked to firms' quotas for output and profits. Enterprises that overfulfill their targets are entitled to use a certain amount of profits for extra wages or bonuses, and those that fail to fulfill targets are subject to a reduction in the growth rate of wages. But even firms that are operating in the red can maintain their wages at the previous year's level or enjoy a small increase. Although this kind of control plays some role in restricting wage inflation, the speed of increase is still fairly fast. Wuxi wit- nessed a sharp growth of TVCE wage levels in 1984 when it relaxed controls over wages. Profits kept by local TVCES rose by 129 percent, but bonuses in- creased by 280 percent and accounted for 18 percent of enterprise retained profits as against 12 percent in the previous year. Average wages increased 29 percent over the previous year. This wage expansion was not brought under control until 1985. Rapid wage increases in community enterprises mean that employees in effect participate in the distribution of enterprise profits. This undoubtedly helps stimulate workers to care about their firms and work hard. But the rate of increase of wages has exceeded that of profits, and the consequent rapid rise in wage costs may eliminate the competitive advantage that TVPS have 232 WANG XIAOLU because of their low costs. Wage inflation also has some effect on the rein- vestment rate and could be a factor in enterprises' heavier dependence on bank loans and credit. In J ieshou private enterprises make up a larger proportion of the total num- ber of TVPS throughout the county and in the sample. Of the thirty-six sample firms in the county, seven identify themselves as private enterprises, and an- other seven claim to be owned by township govemments but actually belong to private owners. Most of the remaining sample township enterprises had no investment from townshilp govemments. The goals and motivations of these firms may be close to those of private enterprises. In general, private en,terprises' capital accumulation motive, especially for individually owned firms that hire wage labor, may be stronger than that of township enterprises, whereas their desire to increase workers' wages is rela- tively weak. This may be a reason for the slow growth of their wage levels. Impressions from interviews with some private enterprises indicate, however, that noneconomic consideratiions lead these firms to tum part of their profits into consumption or public welfare expenditures, usually when the firm's scale reaches a certain [evel. The low growth rate of wages in sample firms of Jieshou could therefore be attributable to other factors-for example, the firms' high dependence on bank loans. As there are many newly established firms in the sample, the pressure of loans falling due is heavy. In fact, a large percentage of the profits used for reinvestment actually go into repayment of bank loans for equipment and working capital. Stricter requirements for timely repayment of bank loanis by private enterprises have forced the enter- prises to increase their internal accumulation rate, in part by holding down wages. Capital Productivity Data from the four counties show sharp differences in the capital productivity of different types of enterprise. Table 10-5 and data for sample firms indicate that industrial township enterprises are markedly less efficient than other types of firm, as measured by profit and tax per unit: of capital, despite their advantages in scale and equipment. Table 10-6, which shows average fixed assets and number of employees for different types of TVP in Nanhiai and Wuxi, highlights the relative inefficiency of township enterprises. Differences in the profit and tax rate on capital among different types of TVps are not caused by differences in industrial composition. A comparison of profit and tax per unit of c apital for township and village enterprises in the same industries in Nanhai still shows a great gap (see table 10-7). There are several reasons for the lower capital profitability in township enterprises. Township enterprises absorb large amounts of loans and credit funds be- Table 10-5. The Ratio of Profit and Tax to Total Capital for Industrial TVPs (percent) Township Village Production Individual County enterprises enterpnses team finns Partnerships enterprises Wuxi (1985) 29.8 39.8' n.a. n.a. n.a. Jieshou (1985) 32.4a 54.4D n.a. 57.1 56.4 Nanhai (1986) 11.2 34.6 26.5b 31.2b 41.9b Shangrao (1983) 22.0 55.8 n.a. n.a. n.a. n.a. Not available. Note: Profits and taxes are net of losses of loss-making firms. Unless otherwise indicated, the denominator is year-end net (depreciated) value of fixed assets plus physical circulating assets. a. Net value of fixed assets at year-end, plus average value of physical circulating assets during the year. b. Original (undepreciated) value of fixed assets plus value of physical circulating assets at year-end. Source: Information from fieldwork. Table 10-6. Average Size of Industrial TVPs, Wuxi and Nanhai, 1986 Wuxi Nanhai Average value of Average Average value of Average Type of fixed assets, number of fixed assets, number of enterprise (yuan) workers (yuan) workers Township enterprises 492,000 111 977,000 155 Village enterprises 138,000 57 96,000 40 Production team firms n.a. n.a. 45,000 20 Partnerships n.a. n.a. 39,000 14 Individual enterprises n.a. n.a. 8,000 6 n.a. Not available. a. Original value. Source: Information from fieldwork. Table 10-7. Ratio of Profit and Tax to Total Capital in TVCEs in Nanhai, by Industry, 1986 Township Village Industry enterprises enterprises Textiles -4.5 42.4 Garments 19.6 32.7 Leather and fur 23.0 42.9 Plastic products 4.2 74.1 Building materials and other nonmetallic mineral products 15.0 116.7 Metallurgical products 14.4 60.7 Machinery industry 16.3 48.9 Light industry 10.7 54.9 Heavy industry 16.6 76.3 Source: Information from fieldwork. 233 234 WANG XIAOLU cause of their soft constraints in this sphere, and this decreases the effi- ciency of capital utilization. The textile industry in Nanhai is an ex- ample. In 1986 township enterprises suffered large losses, whereas other types of textile firm enjoyed considerable profits. The main reasons for the losses were that township enterprises had stockpiled raw materials and were faced with unsatisfactory market conditions. Excessive use of capital also leads to a reduction in profitability when business opera- tions are normal. * Since township enterprises tend to avoid hiring labor from other places, they are more likely tham any other type of TVP to substitute capital for labor if there are no employment problems in the community. For instance, in 1986 the average value of fixed assets per employee for in- dustrial township enterprises in Wuxi was Y4,200, but for village enter- prises it was only Y2, 100. In the same year the figure was Y6,390 for industrial township enterprises in Nanhai, Y2,400 for village enter- prises, Y2,280 for partnerships, Y2, 780 for cooperative enterprises, and Y1,450 for individual firms. Township enterprises use more fixed assets than other firms to produce the same level of output, and the cost of shifting their product lines is higher than in more labor-intensive enter- prises. * Township enterprises-plagued by noneconomic targets forced orn them by township governments-often suffer from overstaffing. The propor- tion of surplus labor was as high as 20-30 percent in some firms. Other types of enterprise seldom had this kind of problem. This is an obvious reason for higher production costs in township enterprises. The reverse side is that township governments, bound by their community targets, are willing to support unprofitable or money-losing township enter- prises. I Unlike private enterprises, which have to shoulder their entire business risk, township enterprises sometimes do not have clear responsibili- ties for investment decisions. This adds to the possibility of investment failures. - Requiring township enterprises to hand in their profits and the attempt to balance incomes among enterprises may to a certain degree reduce incentives for managers and workers and lead to poorer efficiency in township enterprises than in other types of firm. The latter three factors affect township enterprises more conspicuously in underdeveloped areas, where poor management quality also restricts the de- velopment of these enterprises. We found many examples of management failures in Jieshou and Shangrao counties. Jieshou has adopted a policy that encourages the development of various kinds of private enterprise and has achieved some success in this regard. Shangrao, where industry is still mainly Capital Formation and Utlization 235 based on township enterprises, remains the least industrialized county of the four. Trends in Capital Productivity Although the proportion of private enterprises in the total number of TVPS has increased rapidly during the past several years, TVCES still hold the domi- nant position, and township enterprises are at present more important than village enterprises. It is therefore important to analyze changes in the capital productivity of industrial township enterprises, with the help of historical materials available from several of the counties. Figure 10-1 shows changes in the capital productivity of industrial town- ship enterprises in Wuxi and Nanhai counties.4 Capital productivity has tended to decline unsteadily in both counties but more sharply in Nanhai. Wuxi is greatly influenced by changes in national industrial structure and by economic fluctuations, and the result is abrupt rises and falls. Figure 10-2 shows changes in the ratio of profit and tax to total capital for collective in- dustry (mainly township enterprises but including some firms run by the county govemment) in Wuxi, for industrial township enterprises in Nanhai, Figure 10-1. Capital Productivity of Industrial Township Enterprises, Wuxi and Nanhai, 1979-86 i 0.8 0~~~~~~~~~~~~ 2 0.7- \Nanhai 0.6 0 0A 0 .o 0.4 1979 1980 1981 1982 1983 1984 1985 1986 Note: Total capital is the sum of the net (depreciated) value of fixed assets and the physical ("quota") working capital. Source: Information from fieldwork. 236 WANG XIAOLU. and for thirty-six sample TVPS in Jieshou. The curves for Wuxi and Nanhai in figure 10-2 are similar to those in figure 10-1, but they decline more sharply, indicating that the proportion of wage and other nonmaterial pro- duction costs in net value of output increased while the capital productivity rate decreased. Although capital profitability in Jieshou had not yet reached the level of that in Wuxi and Nanhai by 1985, it showed a consistent rising trend which indicates a gradual improvement in financial performance. The reasons for the decline in capital productivity and profitability in the more developed counties are complicated. Aside from the causes mentioned in the last section-heavier dependence on bank loans and the higher capital intensity that has accompanied full employment-the restricted circulation of factors of production under the township enterprise ownership pattern can easily lead to relative saturation of investment opportunities at an early stage. The rest of this chapter focuses on the analysis of these topics. Changes in the Degree of Capital Intensity In the two relatively developed counties, Wuxi and Nanhai, there has been a clear trend toward substituting capital for labor in TVP industry. Between Figure 10-2. Ratio of Profit and Tav to Total Capital, Wuxi, Nanhai, andJieshou, 1979-86 Woi collective industry 0.5 Nanhai industrial township 0.4 \ enterprises -~0.3- 2 0.2 - Note: Total capital is the sum, of the! net (depreciated) value of fixed assets and the physical ("quota") working capitaL. Collective industry includes industrial township enterpi-ises and "urban" co01ictive industrial firms ofsvarious kinds. Visage enterprises are exCuded In Wuxi (he bulk of the collective industrial sector consists of township enterprises. Sources: Enterprise Quantitative Questionnaire and informnation from fieldwork. Capital Formation and Utilization 237 1979 and 1986 the capital coefficient of industrial township enterprises in Wuxi climbed from 1.46 to 2.01 while the labor force coefficient dropped from 4.13 to 2.73.5 In Nanhai between 1977 and 1986 the capital coefficient rose from 1.5 to 2.15 while the labor force coefficient declined from 6.60 to 2.03. Comparatively speaking, substitution of capital for labor has been greater in Nanhai. This can be seen clearly from the capital-labor ratios in table 10-8. That the capital and labor force coefficients rise and fall together in Wuxi means that short-term changes in efficiency are strongly influenced by macroeconomic fluctuations. For the thirty-six sample firms in Jieshou the labor force coefficient de- clined sharply, from 9.66 in 1980 to 2.60 in 1985, but the capital coefficient did not increase accordingly. On the contrary, it decreased somewhat, from 1.66 to 1.04.6 This strongly suggests that whereas overall productivity grew rapidly in Jieshou, the tendency to substitute capital for labor was not as strong. By 1985 the labor and capital coefficients in Jieshou were 100 percent and 40 percent higher than the corresponding figures in Wuxi, while capital per employee was only 70 percent of that in Wuxi (see table 10-8). That Jieshou's capital coefficient is still higher than Wuxi's is attributable to the difference in efficiency. The changing difference in the ratios of the two coef- ficients indicates that the degree of capital-labor substitution in Jieshou is lower than in Wuxi. The substitution of capital for labor in Wuxi and Nanhai counties occurred in tandem with the absorption of local surplus agricultural labor, but Jieshou still has a lot of surplus labor. (See chapter 14 for calcula- tions on the magnitude of surplus labor in the four counties.) These facts suggest that substitution of capital for labor is a phenomenon brought about by rural industrialization. But because of the influence of tar- gets set by community governments, different development phases have dif- ferent characteristics. When a community has a large amount of surplus labor, community enterprises are more inclined to use labor-intensive factor input mixes. After the community has reached full employment, the ten- dency not to use outside labor often persists because the admission of a large number of outsiders may reduce local per capita income. This leads to a more capital-intensive factor mix, and a decline in the capital productivity rate is hardly avoidable. The behavior outlined above is rational from the perspective of community interests, but in aggregate terms it makes inefficiency in the allocation of cap- ital hard to avoid. National surplus labor cannot be fully utilized in regions that already hold too much of the scarce factor, capital. It is possible to con- clude that the marginal profitability and productivity of capital in developed areas such as Wuxi and Nanhai are now lower than in backward localities such as Jieshou. Table 10-8. Capitl-LaborRati-os, 1977-86 (yuan per employee) County 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 Wuxi, n.a. n.a. 3,538 3,907 4,825 4,961 4,926 5,442 6,466 7,358 Wuxib n.a. 5,100 n.a. 7,200 9,700 10,900 10,900 13,100 15,500 n.a. cW NanhaiP 1,898 n.a. n.a. 2,924 3,551 3,934 5,108 5,927 8,193 9,326 Jieshoub n.a. 6,800 n.a. 4,000 4,400 4,800 5,900 7,300 10,800 n.a. n.a. Not available. a. Average for all industrial township enterprises. b. Sample firms. These figures may be distorted owing to changes in the size of the sample over time. Sources: Enterprise Quantitative Questionnaire and information from fieldwork. Capital Formation and Utilization 239 Economies of Scale and Factor Mobility With the rise of capital intensity in TVPS (especially in township enterprises), the production scale and location pattern of enterprises have an increasing effect on economic benefits, since scale economies are much greater for capital-intensive than for labor-intensive production. Until now, most rural industrialization in China has followed the principle of "leaving the farm but not the community." Undoubtedly this helped with the absorption of rural surplus labor, but it may encourage small production scales and dispersed lo- cation patterns that will affect economic efficiency when capital intensity de- velops to a certain degree. Since factors of production have difficulty in moving across regions and localities, TVCES may experience problems in expanding their production scales and concentrating their locations geo- graphically. The sample firms in Wuxi provide an example. No firms have set up facto- ries outside their own communities. Three firms are in towns, but all the rest are located in the countryside. In 1985 investment from other regions made up only 0.5 percent of the total capital of Wuxi sample firms, and in- vestment by these enterprises in other units accounted for only 0.1 percent of their total fixed investment. Cross-regional investment in Nanhai is a little higher than in Wuxi owing to Nanhai's diversified ownership structure and the stronger functioning of markets. Although the average size of TVCES is still much larger than that of private enterprises, there is a tendency toward slower expansion of firm size. From 1975 to 1986 total employment in industrial township enterprises in Wuxi rose by 502 percent, but the average number of employees per enterprise in- creased by only 23 percent (table 10-9). The original value of fixed assets increased by 2,160 percent, but the average value per enterprise rose by only 360 percent. The same pattern holds for village enterprises. Obviously this kind of industrial development consists mainly of the birth of large numbers of new firms. Especially since 1984 the average size of enterprises, measured by employment, has begun to shrink. In 1986 there were 1,247 industrial township enterprises in the county, an average of 36 per township or town. The number of industrial village enterprises reached 3,486, about 6 per vil- lage. On average, there were 46 industrial TVCES for every 10,000 members of the population in the county, and there were 445 enterprises per 100 square kilometers. The constant birth of small firms leads to sharp competition. Among the 185 relatively large "backbone" township enterprises in Wuxi there are about 40 machinery and equipment plants, of which 7 make cranes and construc- tion machinery, 4 make dyeing equipment, and 3 are general equipment works. In addition, there are 7 steel-rolling plants, 4 metal pipe works, 6 woolen textile factories, and 4 printing and dyeing plants. It is difficult to calculate how many small plants there are in each industry or how many pro- Table 10-9. Average Size of Industrial 7VPs, 1970-86 County and type of enterprise 1970 1975 1978 1980 1981 1982 1983 1984 1985 1986 Wuxi township enterprises Number of firms 155 254 353 458 478 509 524 821 1,017 1,247 Average employment 59 99 122 130 134 136 153 116 131' 114' Average fixed assets (thousands of yuan) n.a. 107 220 292 360 382 442 342 451 492 Wuxi village enterprises Number of firms 859 1,380 1,629 1,848 1,749 1,695 2,034 2,656 3,232 3,486 Average employment 30 28 35 39 44 49 49 47 44' 43' t'i Average fixed assets 41- (thousands of yuan) n.a. 15 32 46 61 76 79 72 108 138 Nanhai township enterprises Number of firms n.a. n.a. 195 246 276 280 287 291 311 332 Average employment n.a. n.a. 105 89 90 97 98 112 129 155 Average fixed assets (thousands of yuan) n.a. n.a. n.a. 178 214 226 359 454 695 977 Jieshou township enterprises' Number of firms n.a. n.a. n.a. 160 120 160 151 375 261 n.a. Average employment n.a. n.a. n.a. 29 33 27 26 26 35 n.a. n.a. Not available. a. 1985-86 employment figure. were adjusted to offset false reporting by firrrs that vere trying tO evade wage controls. b. Includes all township enterprises, but industrial firms account for the bulk of the total number and employment. Source: Information from fieldwork. Capital Formation and Utilitation 241 duce the same products. Since enterprises in the same industry generally be- long to different township govemments, they can produce and sell indepen- dently, and they compete with each other. It is almost impossible to combine them because of their differing ownership and the influence of targets set by different township govemments. This has blocked the expansion of enter- prise scale and led to much waste of resources. For example, Wuxi has set up sixteen aluminum extrusion and processing plants in the past several years. Most of them cannot operate at full capacity, and some face bank- ruptcy because of deteriorating market conditions. The situation in Nanhai is a little different. The average employment scale of township enterprises was stable at about 100 workers before 1983, whereas the capital scale increased continually. But since 1984 both factors have grown rapidly even as creation of new enterprises has increased. There is rea- son to believe that, as private enterprises developed rapidly over the past sev- eral years, the opening of markets and increased factor mobility had some effect on the actions of community govemments and their enterprises. There was no obvious expansion of the scale of township enterprises in Jieshou between 1979 and 1985. Both township enterprises and the number of employees were unstable and changing, as the sector had not yet entered a steady expansion phase. This is different from the trend in developed areas toward a constant increase in the number of enterprises and stagnation in their average scale. The dispersed distribution of firms and scattered infrastructure facilities create other problems. Each community or enterprise has to deal indepen- dently with such problems as obtaining supplies of electricity, water, and heat; gaining access to roads, transport, and communications; and maintain- ing equipment. The inevitable results are higher investment costs, loss of effi- ciency, and difficulties in gaining access to market information and tech- nology. For example, to cope with shortages of electricity in Wuxi, almost every TVCE has installed diesel generators to meet its own needs. The unit cost of electricity from diesel generators is as high as YO.3-YO.5 per kilowatt-hour, several times more than that of power transmitted through the electric power network. In 1986 industrial self-generation of electricity with diesel genera- tors reached 225.883 million kilowatt-hours, which is equivalent to the ca- pacity of a middle-size power plant. If all township govemment deci- sionmakers regarded economic benefits as the most important objective, it would not be difficult to raise funds to build a relatively large power plant, and more than Y10 million could be saved every year. Moreover, if the distri- bution of enterprises were relatively concentrated, considerable savings might be realized on transmission equipment and from reduced losses in transmission. Responses to the Enterprise Director Questionnaire show an awareness of the difficulties caused by the dispersion of TVPS. But since the many ties that 242 WANG XIAOLU link firms with community governments can hardly be severed, the possibili- ties for moving and combining enterprises are small. The survey found that private enterprises have a stronger desire to concentrate their location in urban areas. At present, owing to the effect of community objectives on firms, commu- nity enterprises are facing problems in factor mix, scale, and location patterns that may pose serious olbstacles to further rural industrialization. It seems that the only way out lies in rationalizing the actions of both owners and managers of community enterprises while taking community targets into consid[eration. Objectives set by enterprises themselves can be achieved through aLctivities conducted in accordance with market principles. A diversified ownership sys- tem might be a prerequisite. Notes 1. There were twenty-two sample TVCES. Sample private enterprises-including firms officially registered as community enterprises but deemed by field researchers and county officials to be actually private in nature-numbered five in 1982, seven in 1983, ten in 1984, and fourteen in 1985. 2. Our sample for Nanhai contained no private enterprises or production team firrms for 1981 and 1982. 3. These figures are from information gathered during fieldwork. Since the introduction of higher tax rates for TVPS, hiding profits and reporting nonexistent losses have reportedly become widespread practices, but there is no way to measure to what extent this distorts the data. 4. The index of capital productivity is the ratio of net industrial output value to total capital (the sum of net depreciated value of fixed assets and physical circulating assets). This index avoids distortions caused by differing net-gross ratios of industrial output value. 5. The capital coefficient equals total capital divided by net value of output, or the sum of the net value of fixed assets and annual average outstanding floating capital divided by the net value of output. The labor coefficient equals the annual average number of workers divided by the net value of output (in millions of yuan). 6. The capital coefficients for sample firms were calculated by dividing the original value of fixed assets by the net value of output. They are not comparable with the corresponding in- dexes reported for Wuxi and Nanhai counties. 11 Productive Efficiency and Employment Jan Svejnar This chapter tests several hypotheses about the economic performance of Chirnese TVPS. It provides econometric evidence on the effects that different enterprise ownership and control systems, worker compensation schemes, and geographic locations (levels of economic development) have on produc- tive efficiency and labor utilization in TVPS. As chapters 4 and 7 indicate, the four counties covered in the survey vary substantially in the ownership makeup of their TVPS. Wuxi, for instance, ex- emplifies the model of township and village firms with emphasis on commu- nity govemment ownership, whereas Jieshou has supported the development of private firms. A question naturally arises about the relationship between TVP ownership and efficiency. As table 11-1 shows, the data in our sample vary significantly as to ownership and hence lend themselves to an analysis of this issue. The TVPS in our sample also display considerable variation in the form of worker compensation (see table 11-1). In particular, although a majority of the 122 sample firms use piece rates as the principal form of compensation, some rely on fixed wages supplemented with bonuses, dividends, or piece rates, and others use an internal work-point system. One firm still has a com- munity work-point system; workers accumulate work-points in the factory, but their income depends on that of the entire community because enterprise income is pooled with other community income. Since different compensa- tion schemes provide different incentives for worker effort, discipline, and motivation in general, it is of interest to explore whether differences in com- pensation schemes affect enterprise efficiency. Finally, the statistics in table 11-1 confirm that our data reflect enterprise behavior in both the more developed (Wuxi and Nanhai) and the less devel- oped (Jieshou and Shangrao) counties of China. The chapter focuses on two measures of enterprise performance: productive efficiency and labor utilization. The former is of primary interest for two rea- sons. First, Chinese state enterprises are generally regarded as inefficient, and 243 244 JAN SVEJNAR Table 11-1. Summary Statistics of Variables Used Standard Variable Mean deviation Enterprise ownership' Township and town 0.64 0.48 Village and production team 0.17 0.38 Partnership, family, or individual 0.11 0.32 Joint venture 0.07 0.26 Enterprise location' Wuxi 0.26 0.44 Jieshou 0.30 0.46 Nanhai 0.19 0.39 Shangrao 0.25 0.43 Principal form of worker compensation' Fixed time wage and auxiliary piece-rate wage 0.03 0.18 Fixed time wage and bonus 0.10 0.30 Fixed time wage and year-end dividend 0.03 0.13 Piece-rate wage 0.56 0.50 Internal work-point system 0.07 0.25 Community work-point systemb 0.01 0.09 In Q 13.35 1.76 In L 4.77 0.98 In K 12.38 1.57 In W 4.12 0.53 In W. 3.35 0.73 Timec 5.50 4.68 Note: The variable Wa is the average agricultural wage in the county. a. These variables are binary (dummy) variables that are coded 1.0 in the presence of the given phenomenon and 0.0 in its absence. The mean of this variable reflects the percentage of observations falling into the particular category. b. In this system workers accumulate work-points in the enterprise, but their income depends on the income of the entire community (enterprise revenues are pooled with other community income). c. Time has values of 1 in 1975, . 11 in 1986. Sources: Enterprise Quantitative Questionnaire and Enterprise Survey Questionnaire. TVPS are frequently mentioned as examples of more (productively) efficient firms. Second, as in other centrally planned and developing economies, eco- nomic growth has beert regarded in China as an important performance indicator-one that is closely linked to productive efficiency at the firm level. Labor utilization is of interest because employment generation was the principal goal of most TVPS at: time of founding. Moreover, although Wuxi Productive Efficiency and Employment 245 and Nanhai have largely exhausted their surplus labor and now emphasize income maximization more, in Jieshou and Shangrao the employment objec- tive remains predominant. Hence it is of interest to see whether TVPS' criteria for labor utilization differ among counties and whether they are linked to the ownership structure of the firm. Methodology In this section a method of testing for differences in productive efficiency with the use of a production function estimation approach is developed. An employment equation is also derived that can be used to test hypotheses about the goals TVPS pursue. Productive Efficiency The analysis of productive efficiency is based on Cobb-Douglas (CD) and translog production functions, augmented by relevant ownership, county, and compensation scheme variables. If Q = output, L = labor, K = capital, Z = a vector of ownership and compensation scheme variables, C = a vector of location (county) dummy variables, and T = a vector of annual dummy variables, the basic specification of the CD function can be written (11-1) In Q = xo + Z + RC + jT + ax In L + ct2 In K + E where e is the error term, ,13, , and , are (row) vectors of parameters corre- sponding to the Z, C, and T vectors of variables respectively, and the as are the parameters corresponding to the other variables. In this framework, Z, C, and T are dummy variables that take on a value of 1 when an observation falls into the particular ownership or compensation system, county, or time period respectively and 0 otherwise. For small values of f3, ,, and a the esti- mated coefficients hence approximate percentage differentials in productive efficiency (total factor productivity) among firms with different ownership forms, compensation systems, locations, or time periods. For large parameter values, however-for example ,B > 0.3- it is preferable to calculate the per- centage differential as 100 * [exp(,B) - 1]. (See Halvorsen and Palmquist 1980.) The translog equation was specified similarly as (11-2) In Q = ao + ,Z + C + ,T + e,l In L + a2 In K + cx3 (In L)2 + a4 (In K)2 + as5 (In L) (In K) + e Employment In addition to examining how much output different TVPS can produce from a given set of inputs (productive efficiency), this chapter analyzes patterns 246 JAN SVE)NAR of labor utilization. Most TVPs have three important goals: maximizing prof- its, generating employment, and increasing wages above the conmpetitive (market-clearing) level, in the economy. The relative weight of these goals in the enterprise objective ftnction varies among counties and among types of Tvps. A convenient way to capture these behavioral features analytically is to specify the objective finction of a TvP as a weighted geometric average of these three goals (see IKalai 1977, Roth 1979, and Svejnar 1986). If W = the actual wage (labor income), Wa = the reservation or best altemative wage of TVP workers,' and rl= profit, the objective function U of the TVP can be written (11-3) U La' (W - Wa)OW fl(l-L-Mw) where aL, aw, and 1 -- aL a- dw are the relative weights given to employ- ment, wages, and profitability respectively. Note, for example, that if a TVP maximized only profit, U = [1 as aL = a = 0, and the firm would behave like a traditional "capitalist" firm. In contrast, if the firm only maximized in- come per worker, then U = (W - Wa) as aL = 0 and aw = 1, and the firm would behave according to the traditional labor-management model (see, for example, Domar 1966, Vanek 1970, and Ward 1958). The advan- tage of equation 11-3 is that it nests various forms of enterprise behavior as special cases, which can be identified if one can estimate relatively precisely the values of aL and dw. The empirical work here follows earlier studies by Brown and Ashenfelter (1986) and Svejnar (1986). It exploits the fact that if equation 11-3 is maximized subject to a constant elasticity of substitution (CES) production function, one can in principle identify enterprise employ- ment behavior from the linear estimating equation (11-4) In L = Yo + ez + OC + eT + In Q v cr(l -a) In W - ad In Wa where a = aLIaW, cr the elasticity of substitution, v = a return to scale parameter, and Yo, 0e, and e are undetermined parameters. From equation 11-4 it is clear that i) = 0 implies aL = 0, and, in terms of equation 11-3, the firm does not: expand employment beyond the profit-maximizing level given by the equality of marginal product of labor and the actual wage W. In contrast, if a = 1, the enterprise gives identical importance to increas- ing wages and generating new employment (L = aw). In this case the coeffi- cient on In W is zero in equation 11-4, and the employment behavior of the TVP depends on the alternative wage Wa but not on own wage W. In fact, it can be shown that in this case the firm equates the marginal product of labor to Wa. An important problem in the empirical implementation of equa- tion 11-4 is then, of course, one's ability to approximate closely the concept of the reservation (alternative) wage Wa. Productive Efficiency and Employment 247 In estimating equation 11-4 it is useful to interact a with the variables in vectors Z and C to establish whether a varies systematically with ownership and the level of economic development (proxied by county dummies). Data and Empirical Results As the summary statistics in table 11-1 indicate, the sample displays consider- able variation in the values of most of the variables used in this study. Town and township enterprises account for 64 percent of the observations, village and production team firms for 17 percent. Private firms (partnerships and household, family, and individually owned enterprises) represent 11 percent of observations, and joint ventures of all types make up 7 percent. It should be noted that these figures refer to observations that are usable for the pur- poses of this study. Since different firms had data panels of different lengths, the figures do not necessarily reflect the number of firms of a given type that were sampled. An important point, which is not discemible from table 11 1, is that the data include no observations of private firms in Wuxi or of joint ventures in Shangrao. Estimates of Productive Efficiency Since the CD and translog estimates are very similar, we report the CD results, which are easier to interpret. Table 11-2 presents the CD estimates obtained from three panels of data. The results in column 1 are based on all available data for the entire 1970-86 period. But since the data were collected retro- spectively and many enterprises were reorganized or established after 1970, the estimates may be biased by data errors, which are likely to be more serious in the early period. To assess the severity of this problem, columns 2 and 3 present results based on data from more recent periods. Cross-sectional esti- mates were also generated for several years. Since they tumed out to be simi- lar to those in column 3, they are not presented here. The two main competing views about the relationship between efficiency and ownership are (a) that TVCEs are more efficient than private enterprises because they have better access to new technologies or can take better advan- tage of economies of scale, and (b) that private firms are likely to be more efficient than community-owned firms because the owner-manager has stronger incentives. The results in table 11-2 indicate that when differences in inputs and other variables are controlled, productive efficiency is not re- lated to ownership. Table 11-2 also suggests that enterprises that offer their workers a fixed time wage with a bonus and those that use the intemal work-point system tend to be more productive than those operating under piece rates, a combination of a fixed wage and piece rates, or a combination of a fixed wage and a year- end dividend. The one enterprise that uses a community work-point system, Table 11-2. The Impact of Ownership, Worker Compensation Schemes, and Location on Prodw:tive Efficiency in a Cobb-Douglas Framework Varianle 1970-86 1981-86 1983-86 Intercept' 5.686- 6.769- 7.468- (0.508) (0.450) (0.547) Ownership dummy variables Village and production tearm -0.020 0.028 0.063 (0.110) (0.120) (0.140) Joint household, family, and 0.089 0.008 --0.232 individual (0.151) (0.158) (0.176) Joint venture 0.087 0.130 0.078 (0.141) (0.159) (0.186) Compensation scheme dummy variables Fixed time wage and bonus system 0.276- 0.308- 0.388- (0.162) (0.167) (0.181) Fixed time wage and year-end 0.018 0.047 0.044 dividend system (0.468) (0.473) (0.473) Piece-rate system -0.030 0.005 0.088 (0.103) (0.109) (0.126) Internal work-point system 0.247- 0.302- 0.304 (0.157) (0.173) (0.208) Community work-point system -1.284- -1.333- --1.426- (0.372) (0.382) (0.443) Location dummy variables Jieshou -0.126 -0.139 --0.111 (0.110) (0.121) (0.147) Nanhai -0.165 -0.068 --0.017 (0.117) (0.128) (0.149) Shangrao -1.260- -1.369- -- 1.507- (0.121) (0.137) (0.171) Inputs In L 0.987- 0.966- 0.923- (0.064) (0.071) (0.092) In K 0.164- 0.158- 0.140- (0.040) (0.045) (0.057) Industries Mechanics and electronics 0.413' 0.365- 0.363- (0.114) (0.126) (0. 148) Wood processing and fumiture -0.108 -0.127 -0.204 (0.166) (0.178) (0.205) Chemicals 0.247' 0.190 0.105 (0.128) (0.140) (0.167) Metallurgy 0.795- 0.754- 0.639- (0.153) (0.166) (0.199) Paper processing 0.040 -0.082 -0.291 (0.150) (0.167) (0.196) Food processing -0.046 -0.090 -0.105 (0.165) (0.173) (0.191) 248 Productive Efficiency and Employrnent 249 Variable 1970-86 1981-86 1983-86 Time 1975 0.184 - - (0.427) 1978 0.489 - - (0.418) 1980 0.759' - (0.385) 1981 0.890- - - (0.385) 1982 0.987- 0.095 - (0.383) (0.124) 1983 1.118 0.231" - (0.384) (0.121) 1984 1.161- 0.276- 0.056 (0.384) (0.118) (0.107) 1985 1.540- 0.659- 0.454- (0.385) (0.118) (0.107) 1986 1.554- 0.645- 0.451- (0.437) (0.244) (0.243) R2 0.88 0.88 0.89 N 415 347 240 - Not applicable. 'Significant at 5 percent statistical test level or better. 'Significant at 10 percent statistical test level or better. Note: The dependent variable is the logarithm of the gross value of output. The values in parentheses are standard errors. a. The intercept represents township enterprises in the textile and clothing industry of Wuxi that use a fixed time wage and auxiliary piece-rate system; it also corresponds to the base year of the particular regression. Source: Enterprise Quantitative Questionnaire. which dilutes worker incentives, is found to be greatly inferior in productive efficiency. Other things being equal, group incentive schemes seem to be as- sociated with higher productivity than individual incentive schemes. Group coordination and team spirit may thus be important for efficiency. The coefficients on the county dummy variables indicate that TVPS in Wuxi, Jieshou, and Nanhai are more or less equally productive but that firms in Shangrao lag significantly behind. Shangrao is, of course, the least devel- oped of the four counties but, perhaps more important, Shangrao TVPS suffer from neglect and even exploitation by local authorities (see chapter 17). Since Jieshou is also a less developed county, it may be that different attitudes and policies of local authorities toward TVP development account for the strikingly different productive performances (see also chapter 4). The time dummies indicate that Tvps have achieved a significant rate of 250 JAN SVEJNAR technical progress over time, with the most rapid gains occurring in 1985. The estimates confirm that the rapid development of the TVP sector has not been brought about merely by extensive growth; intensive growth has con- tributed significantly to the observed success. The coefficients on labor and capital are both significant, and they suggest that TVPS operate in the zone of mildly increasing retums to scale. This result is not improbable in view of the small scale of many TVPS. The two input coefficients also suggest that the output elasticity of labor is much greater than that of capital. Finally, there appear to be significant technological dif- ferences among industries. The results in table 11-2 are based on the ordinary least squares (oLS) esti- mating technique, which is a traditional method that is more robust against specification errors than simultaneous equations methods. But since variables K and L are at least to some extent endogenous (that is, subject to enterprise decisionmaking), the oLs regressions of table 11-2 were also reestimated by instrumental variables. The instruments used were price indexes, interest rates, time trends, alternative wages, and firm-specific intercepts. The main results remain unaltered under this estimating technique. Labor Utilization In estimating equation 11-4, one must select a suitable reservation (altema- tive) wage, Wa. After considerable search, the average monthly income of an agricultural worker in each county in a given year was adopted as rhe best measure of the reservation wage. The motivation for using Wa as a proxy for the true (unobserved.) reservation wage Wa' was that it is close to and very likely proportional to W,:. In particular, if (11-5) W.= Wa, then (11-6) In W l = In q + In Wa and using In Wa will produce an unbiased estimate of the coefficient on In Wa, since In Ir will be captured in the regression intercept. As the results in table 11-3 show, the available data do permit one to draw conclusions about the employment behavior of Chinese TVPS within the framework of equation 1 1-4. The coefficients on In W and In Wa reflect the situation in Wuxi (the base county) and suggest that, although it is impossi- ble to identify a from the 1SI80-86 data set, the value of a appears to be zero in the more recent (198:3-86) period. That is, in recent years rvps in Wuxi seem to place no extra emphasis on employment and to operate along the marginal product curve of labor. This implies that they have reduced their labor input in response to actual wage increases, rather than being guided by the alternative wage. Table 11 3. The Impact of Ownership and Location on Enterprise Employment Policy Vanable 1980-86 1983-86 Intercept' -2.038" -1.930 (1.604) (1.776) In Q 0.523' 0.523- (0.025) (0.035) In W -0.531- -0.745' (0.148) (0.219) In Wa 0.643' 0.779 Ownership (0.321) (0.503) Village and production team (VTO) -0.235 0.144 (0.604) (0.875) Joint household, family, and 1.805' 2.842' individual (HFIO) (0.766) (1.001) Joint venture (jv) 0.555 -0.061 (0.965) (1-444) VTO In W -0.012 0.236 (0.197) (0.377) HFIO In W -0.570- 0.073 (0.225) (0.373) jv In W -0.022 0.483 (0.282) (0.441) VTO In Wa 0.026 -0.368 (0.144) (0.367) HFIO In Wa 0.109 -0.975' (0.153) (0.451) jv In Wa -0.144 -0.566 Location (0.174) (0.370) Jieshou 0.076 1.828 (1.131) (2.526) Nanhai 1.623 -1.266 (0.932) (3.115) Shangrao -0.728 2.363 (1.148) (4.706) Jieshou In W 0.560- 0.108 (0.268) (0.360) Nanhai In W 0.504- 0.703- (0.236) (0.321) Shangrao In W 0.669' 0.844- (0.180) (0.244) Jieshou In Wa -0.736' -0.721 (0.245) (0.739) Nanhai In Wa -0.982- -0.523 (0.293) (0.773) Shangrao In Wa -0.451 -1.628 (0.399) (1.499) (Table continues on the following page) 252 JAN SVEJNAR Table 11-3 (continued) Variable 1980-86 1983-86 Compensation scheme Fixed time wage and bonus system -0.148 -0.174 (0.123) (0.141) Fixed time wage and year-end -0.188 -0.229 dividend system (0.316) (0.326) Piece-rate system 0.029 -0.018 (0.078) (0.095) Internal work-point system 0.077 0.104 (0.117) (0.149) Community work-point system 0. 548' 0.796' (0.253) (0.304) Time 1981 -0.095 (0.119) 1982 -0.070 (0.128) 1983 -0.094 - (0.181) 1984 -0.074 0.078 (0.237) (0.155) 1985 -0.219 -0.063 (0.244) (0.156) 1986 -0.242 -0.135 (0.365) (0.374) Industry intercepts Yes Yes R2 0.84 0.84 N 355 232 - Not applicable. 'Significant at 5 percent statistical test level or better. 'Significant at 10 percent statistical test level or better. Note: The dependent variable is the logarithm of employment. The values in parentheses are standard errors. a. The intercept represents township enterprises in the textile and clothing industry of Wuxi that use a fixed-time wage and auxiliary piece-rate system; it also corresponds to the base year of the particular regression. Source: Enterprise Quantitative Questionnaire. The interaction of In W and In Wa with the three county dummies identi- fies TVP behavior in Jieshou, Nanhai, and Shangrao. The estimated parame- ters imply that Jieshou TVPS placed more emphasis on Wa than on W in 1980-86 but that their behavior approximated that found in Wuxi in 1983-86. This suggests that Jieshou TVPS used to operate to the right of the marginal product curve of labor but switched to a more labor productivity- oriented employment policy in recent years. "Labor hoarding" is also ob- Productive Efficiency and Employment 253 served in Nanhai in the 1980-86 data set, but no significant relationship between employment and either W or Wa is observed in recent years. The data from Shangrao display no significant relationship between employment and own or alternative wages. The results suggest that TVPS in Wuxi and Jieshou have changed their employment policies in the 1980s and place less emphasis on employment generation than they did before. The interaction of ownership variables with In W and In Wa permits an assessment of whether employment behavior varies with ownership. Private ownership is the only variable with a significant interaction. The parameter estimates suggest that in 1983-86 private firms paid more attention to alter- native opportunities of workers than did the other types of firms. Different compensation schemes do not appear to have wage-related interactive effects on employment of TVPS. The regressions with these insignificant coefficients are not reported. Turning to the basic (noninteractive) effects of different variables, one can see from table 11-3 that private firms tend to be more labor-intensive and that after industry, output, wages, and the like are controlled, labor intensity does not vary among counties, among forms of compensation (except for the one community work-point firm), or over time within each sample. The one variable that is most strongly correlated with employment is the level of output. The general picture that emerges from table 11-3 is that TVPS link their employment closely to output and, especially in the case of Wuxi and Jieshou, economize on labor as wages rise. To examine the overall validity of the latter statement, equation 11-4 was also estimated without any interac- tions of In W and In Wa with other variables. The exercise yields a significant negative coefficient of -0.18 (0.07) on In W and an insignificant negative coefficient of -0.15 (0.12) on In Wa. These results suggest that TvPs recog- nize the tradeoff between wages and employment and that, on average, feath- erbedding is not a significant phenomenon in these firms. Finally, as in the production function analysis, we reestimated the employ- ment equation by instrumental variables. The results were similar to those obtained in the OLS regressions. Summary and Concluding Observations Our analysis indicates that, after differences in inputs and other variables are controlled, productive efficiency does not vary systematically with the four types of ownership examined. In particular, private ownership and commu- nity ownership appear to have similar effects on productivity. Another find- ing is that group incentive schemes, such as a fixed wage with bonus or an internal work-point system, are associated with higher efficiency than indi- vidual incentive schemes such as piece rates or piece rates with fixed wages. This could signal the importance of teamwork for productive efficiency. TVPS 254 JAN SVE3NAR have experienced rapid technical progress (productive efficiency has risen rapidly over time), and they seem to operate under mildly increasing returns to scale. A comparison of the four counties suggests that there is no significant dif- ference in the productive efficiency of TVPS in the more developed counties of Wuxi and Nanhai and the less developed county of Jieshou. But in Shangrao, where community authorities have sometimes neglected and even exploited TVPS, the level of efficiency is considerably lower. These findings suggest that the behavior of local authorities may significantly influence TVP performance. The employment policies of TVPS, to the extent that they can be identified under our methodology, appear to have shifted away from employment gen- eration as such toward employment levels that reflect the marginal productiv- ity of labor. The findings suggest that TVPS operate with much less featherbed- ding than their state-owned counterparts. TvPs also display a very tight link (almost fixed proportions) between employment and output. Note 1. The reservation wage concept refers to the minimum wage that a worker would be willing to accept in present employment rather than quit. It can also be thought of as the best wage that a worker can obtain in altemative employment. 12. Regional Imbalances Wang Tuoyu The development of China's rural enterprises, like the development of the national economy as a whole, is characterized by great regional imbalances. This chapter surveys the imbalances in China's TVP sector and their influence on the economic development of the country's inland and remote areas. Be- cause the TVP sector has developed rapidly in recent years and because better statistical data are available after 1978, we concentrate on the years 1979-85.l China is a large country, and its regions vary greatly in geography, natural resources, human resources, and material capital. Moreover, regional differ- ences in economic history have created big gaps in current levels of develop- ment, especially in rural areas. Table 12-1 shows the regional distribution of employment, output value, tax payments, and year-end value of fixed assets for industrial and other TVPS in 1985. (The data on fixed asset values include only TVCES.) The following areas account for only about 10 percent of the national totals of these four items: Qinghai and Gansu provinces and the Ningxia Hui and Xinjiang Uygur autonomous regions in northwest China, Guizhou and Yunnan provinces and the Guangxi Zhuang and Xizang (Tibet) autonomous regions in southwest China, Nei Monggol (Inner Mongolia) in north China, and Heilongjiang and Jilin provinces in northeast China. Jiangsu, Zhejiang, Shandong, Guangdong, Hebei, Sichuan, Henan, Liao- ning, and Hubei provinces make up more than 60 percent of the national totals. China's existing TvPs are mainly located in economically developed provinces and in the heartland area made up of Henan, Hubei, and Sichuan provinces; only a few are in remote areas. Tables 12-2 through 12-4 provide further evidence of the imbalances in the TVP sector. Table 12-2 displays value of output, profits, taxes, and year- end value of fixed assets for TVPS in each province and autonomous region, per thousand rural population. As a basis for classification by level of TVP development, table 12-3 compares these indicators with national averages. Table 12-4 shows the proportion of value of rural nonagricultural output in 255 256 WANG TUOYU Table 12-1. China's TVP Sector, by Province, 1985 (percentage of national total) Gross value Tax Orinal value Province E-mployment of output payments of fixed assetsb Anhui 5.0 3.1 2.0 2.4 Fujian 2.9 2.6 2.2 2.4 Gansu 1.0 0.6 0.4 0.7 Guangdong 6.0 7.0 6.2 7.1 Guangxi 2.0 1.1 1.0 1.0 Guizhou 1.4 0.7 0.7 0.5 Hebei 9.8 10.4 11.6 9.1 Heilongjiang 1.6 1.7 1.2 2.2 Henan 8.0 6.1 3.9 5.0 Hubei 5.1 4.1 3.1 5.0 Hunan 4.7 3.5 3.4 4.1 Jiangsu 12.7 20.1 26.5 19.2 Jiangxi 2.7 1.7 1.6 1.9 Jilin 1.4 1.4 1.3 1.4 Liaoning 3.4 5.0 6.6 4.9 Nei Monggol 1.0 0.6 0.4 0.7 Ningxia 0.3 0.2 0.2 0.2 Qinghai 0.2 0.1 0.1 0.1 Shaanxi 2.6 1.8 1.6 1.9 Shandong 9.4 9.0 8.0 11.5 Shanxi 3.1 3.1 3.1 3.4 Sichuan 6.7 5.3 4.4 5.8 Xinjiang 0.5 0.4 0.4 0.5 Xizang 0.0 0.0 0.0 0.0 Yunnan 1.6 1.0 0.7 1.3 Zhejiang 6.7 9.4 9.4 7.7 Selected municipalities' Beijing 1.2 2.0 3.3 2.7 Shanghai 2.0 3.5 7.8 5.6 Tianjin 1.0 1.7 2.6 1.7 a. For simplicity, throughout this chapter the term "province" covers all province-level politi- cal divisions, including autonomous regions. b. Data are for TVCEs only. c. Beijing and Tianjin are also included in the figures for Hebei, and Shanghai is included in those for Jiangsu. Source: Information from fieldwork. Regional Imbalances 257 Table 12-2. Per Capita Indexes of TVP Development, by Province, 1985 (yuan per thousand rural population) Province or Gross value Tax Original value municipality of output Profits payments of fixed assets' All China 32.3 3.4 1.6 8.9 Anhui 19.1 2.2 0.6 4.0 Beijing' 139.2 19.4 11.6 52.0 Fujian 30.5 2.9 1.1 9.6 Gansu 9.8 0.9 0.3 3.2 Guangdong 37.0 4.0 1.6 10.3 Guangxi 8.5 0.7 0.4 2.1 Guizhou 7.1 0.9 0.3 1.5 Hebei 38.0 6.0 1.6 7.4 Heilongjiang 24.7 2.1 0.9 8.8 Henan 24.2 3.6 0.8 5.6 Hubei 28.7 2.9 1.1 9.6 Hunan 19.6 1.9 0.9 6.4 Jiangsu 86.8 5.9 4.9 19.7 Jiangxi 16.3 1.8 0.8 5.2 Jilin 26.8 2.9 1.2 7.3 Liaoning 62.1 7.9 4.1 16.7 Nei Monggol 12.1 1.8 0.4 3.7 Ningxia 18.9 2.5 0.9 3.1 Qinghai 10.1 1.0 0.7 3.5 Shaanxi 19.9 2.3 0.9 5.8 Shandong 36.9 4.1 1.6 12.7 Shanghai" 220.4 24.7 24.9 98.2 Shanxi 40.9 6.6 2.0 12.1 Sichuan 16.2 0.9 0.7 4.9 Tianjin' 126.8 10.9 9.6 34.3 Xinjiang 14.6 7.6 0.7 5.1 Xizang 1.7 1.2 0.0 n.a. Yunnan 9.0 0.7 0.3 3.3 Zhejiang 75.3 6.6 2.0 12.1 n.a. Not available. Note: Hebei does not include Beijing and Tianjin municipalities, and Jiangsu does not include Shanghai municipality. a. Data are for TVCEs only. b. Municipality. Source: Information from fieldwork. 258 WANG TUOYU Table 12-3. Provincial Per Capita Indicators fr TVPs in Comparison with National Averages, 198'i Index against Value of Tax Original value national averages output Profits payments of fixed assetsa Greater than Shanghai Shanghai Shanghai Shanghai 50 percent Beijing Beijing Beijing Beijing Tianjin Tianjin Tianjin Tianjin Jiangsu Jiangsu Jiangsu Jiangsu Zhejiang Zhejiang Zhejiang Zhejiang Liaoning Liaoning Liaoning Liaoning Shanxi Hebei Xinj iang 0 to 49 percent Slianxi Guangdong Shanxi Shanxi Hebei Shandong Hebei Hubei Shandorig Henan Shandong Fujian Guangdong Liaoning Guangdong Guangdong Shandong -30 to 0 percent Fujian Jilin Fujian Jilin Hubei Fujian Hubei Hebei Jilin Hubei Jilin Heilongjiang Heilongjiang HunAn Henan -50 to -31 percent Shaanxi Shaanxi Heilongiiang Henan Hunan Heilongjiang Hunan Shaanxi Arnhui Anhui Henan Jiangxi Ningxia Hunan Ningxia Sichuan Jiangxi Jiangxi Jiangxi Xinjiang Sichuan Nei Monggol Shaanxi Xinjiang Less than -50 Nei Monggol Sichuan Anhui Ningxia percent Qinghai Qinghai Sichuan Anhui Xinjiang Gansu Xinjiang Nei Monggol Gansu Yunnan Nei Monggol Qinghai Yurnan Guangxi Qinghai Gansu Guangxi Guizhou Gansu Yunnan Guizhou Xizang Yunnan Guangxi Xizang Guangxi Guizhou Guizhou Xizang Xizang a. Data are for TVcEs only. Source: Table 12-2. Regional Imbalances 259 Table 12-4. The Value of Nonagricultural Output as a Share of Total Value of Rural Social Output, 1980-85 Province 1980 1984 1985 All China 31.1 36.6 42.9 Anhui 23.2 28.0 30.7 Fujian 29.0 34.7 37.5 Gansu 17.2 16.4 17.9 Guangdong 34.1 36.1 41.6 Guangxi 20.0 18.8 22.9 Guizhou 16.4 24.2 25.6 Hebei 40.7 48.3 54.8 Heilongjiang 15.5 20.0 30.3 Henan 30.7 30.6 34.2 Hubei 25.3 33.6 37.8 Hunan 21.5 28.8 31.2 Jiangsu 52.1 57.4 65.1 Jiangxi 26.6 27.9 30.5 Jilin 24.6 25.7 31.1 Liaoning 35.0 42.3 52.9 Nei Monggol 23.5 19.2 20.1 Ningxia 20.2 23.0 24.9 Qinghai 22.4 24.1 25.9 Shaanxi 25.0 27.5 36.4 Shandong 32.7 37.5 41.8 Shanxi 41.9 47.5 56.2 Sichuan 19.8 25.2 30.2 Xinjiang 13.6 14.4 15.9 Xizang 0.9 4.2 9.3 Yunnan 23.0 20.3 23.6 Zhejiang 40.1 52.7 60.8 Note: Beijing and Tianjin municipalities are included under Hebei, and Shanghai municipal- ity is included under Jiangsu. Source: State Statistical Bureau (1987c), p. 24. the total value of rural social output for each province and autonomous re- gion. The following table compares this figure for each province with the national average. In relation to national average Higher Hebei, Shanxi, Liaoning, Jiangsu, Zhejiang Same Shandong, Guangdong Lower, by less than Jilin, Heilongjiang, Anhui, Fujian, Jiangxi, 30 percent Henan, Hubei, Hunan, Sichuan, Shaanxi Lower, by more than Nei Monggol, Guangxi, Guizhou, Yunnan, 30 percent Xizang, Gansu, Qinghai, Ningxia, Xinjiang 260 WANG TUOYU Table 12-3 and the above breakdown show that the level of TVP develop- ment ebbs progressively fromn east to west-that is, from the economically developed coastal areas, such as the Liaodong Peninsula, the Beijing-Tianjin- Tangshan area, the Eastern Shandong Peninsula, the Yangtse River Delta, and the Pearl River Delta, to the hinterland and western frontier areas. China can be divided into three regions according to level of development of rural enterprises (see map 12-1).z * Region A, with a developed TVP sector, consists of Liaoning, Hebei (including Beijing and Tianjin), Shandong, Jiangsu (including Shang- hai), Zhejiang, Guangdong, and Shanxi * Region B, with develcping TVPS, consists of Henan, Hubei, Fujian, Jiangxi, Hunan, Anhui, Shaanxi, Heilongjiang, Jilin, and Sichuan * Region C, with an iLnderdeveloped TVP sector, consists of Nei Monggol, Ningxia, Gansu, Qinghai, Xinjiang, Guangxi, Yunnan, Guizhou, and Xizang Table 12-5 shows the distribution of TvPs among the three regions according to various indicators. Great regional differences also exist in the internal structure of the TVP sector. In the country as a whole, industry ranks first among nonagricultural production sectors in rural areas, followed by construction, miscellaneous trades (commerce, catering, and services), and transport. But this is not true of all regions. * In the areas with developed TVP sectors (region A) industrial TvCEs ac- count for a large share of TVCES' total number of employees and total sales revenues. The share of industrial TVCES is comparatively smaller where the TVP sector is underdeveloped (region C). The comprelhensive proportion of industrial TVCES (the simple average of their shares in total TVCE employment and in total sales revenue) was more than 70 percent in 1985 in Tianjin, Shanghai, Jiangsu, Zhejiang, Beijing, Table 12-5. The Regional Distribution of China's TVP Sector (percentage of national totals) Rural TVPs labor Value of Tax Original value Region force Employment output payments Profits of fixed assets& A 37.7 51.3 63.9 71.4 64.2 62.9 B 46.1 40.7 31.4 24.7 29.5 32.1 C 16.2 8.0 4.7 3.9 6.3 5.0 a. Data are for TVCES only. Source: Information from fieldwork. Regional Imbalances 261 Hebei, and Shanxi. The comprehensive proportion was less than 55 percent in Xinjiang, Ningxia, Qinghai, Gansu, Xizang, Yunnan, Guangxi, and Hunan. * In region C the construction industry is more important than else- where. In 1985, for example, the cornprehensive proportion of the con- struction sector in the total for TVCES surpassed the proportion of industry itself in Gansu, Qinghai, Ningxia, Xinjiang, and Yunnan. This proportion exceeded 20 percent in nine other provinces and au- tonomous regions-Nei Monggol, Anhui, Fujian, Henan, Hunan, Guangxi, Sichuan, Xizang, and Shaanxi. * The region with developing rural enterprises (region B) has compara- tively large proportions of commerce, catering, and services. In 1985 the comprehensive proportion of these subsectors in the total for TVCES surpassed 10 percent in eleven provinces and autonomous regions-Nei Monggol, Jilin, Heilongjiang, Jiangxi, Shandong, Hubei, Hunan, Guangdong, Guangxi, Xizang, and Shaanxi. Of the eleven, only Guangdong, Shandong, and Hubei are in Region A. There are large regional differences in industrial production structure (see table 12-6). First, the subsectoral composition of industrial TVCES grows pro- gressively "heavier" from the coastal areas in the southeast of the country to the northwest. Comprehensive proportion of heavy industry in total for industrial TVCEs More than 70 percent Shanxi, Guizhou, Yunnan, Qinghai, Ningxia, Xinjiang 60-70 percent Nei Monggol, Liaoning, Jilin, Heilongjiang, Anhui, Henan Shaanxi, Gansu 50-60 percent Hebei, Jiangxi, Shandong, Hubei, Hunan, Sichuan 40-50 percent Beijing, Tianjin, Shanghai, Jiangsu, Fujian, Guangxi Less than 40 percent Zhejiang, Guangdong Regional differences in the composition of heavy and light industries in China's TVCEs are largely determined by the factors of production. China's natural resources are mainly located in the west and north, whereas its agri- cultural resources and population are mainly in the coastal areas of the south and east. Second, although in the country as a whole the leading subsectors in the industrial TVCE sector are, in order of importance, machine building, build- ing materials, chemicals, textiles, and food processing, there are substantial Map 12-1. The LevSel of Development of the TVP Sector 1:.80 1.1~-i'o 1'10 40 i,, ' Region C (Underdeveloped TVP Sector) ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ US.S. U.S.ic BuS.R.e ITNM | InternaWOnAl 80Udar;OS IAgoE Vz Regional Imbalances 263 Table 12-6. Leading Industrial Subsectors in TVCEs, by Province, 1984 (rank as measured by share of GVIO of TVCEs in province) Province First Second Third Fourth Fifth All China MACH BLDG TEXT CHEM FOOD Anhui BLDG FOOD MACH OTH CHEM Fuj ian BLDG MACH FOR FOOD OTH Gansu BLDG MACH CHEM COAL FOOD Guangdong MACH BLDG OTH TEXT CHEM Guangxi BLDG OTH FOOD MACH CULT Guizhou Zhuang COAL BLDG MET CHEM FOOD Hebei MACH BLDG CHEM OTH TEXT Heilongjiang BLDG FOOD MACH COAL OTH Henan BLDG MACH OTH FOOD CHEM Hubei BLDG MACH OTH FOOD TEXT Hunan BLDG FOOD OTH MACH COAL Jiangsu MACH TEXT BLDG CHEM OTH Jiangxi BLDG CULT COAL MACH FOOD Jilin BLDG MACH FOOD OTH COAL Liaoning MACH BLDG OTH FOOD CHEM Nei Monggol BLDG COAL MACH FOOD OTH Ningxia BLDG MACH FOOD COAL OTH Qinghai BLDG MACH COAL FOOD OTH Shaanxi BLDG MACH COAL FOOD OTH Shandong BLDG MACH FOOD OTH TEXT Shanxi COAL BLDG MACH FOOD CHEM Sichuan FOOD BLDG MACH COAL OTH Xinjiang BLDG COAL FOOD MACH CHEM Xizang n.a. n.a. n.a. n.a. n.a. Yunnan COAL BLDG FOOD CHEM MACH Zhejiang MACH TEXT BLDG CHEM OTH Selected municipalities Beijing MACH BLDG MET GARM CHEM Shanghai MACH TEXT CHEM GARM BLDG Tianjin MACH GARM CHEM BLDG MET n.a. Not available. Note: Based on oVIO in 1980 constant prices. BLDG, building materials; CHEM, chemicals; COAL, coal and coking; CULT, cultural/educational activities; FOOD, food processing; FOR, for- estry; GARM, garments; MACH, machinery; MET, metallurgy; TEXT, textiles; OTH, other industry. Source: Information from fieldwork. regional differences (see table 12-6). Machine building and textiles dominate in Jiangsu, Zhejiang, Shanghai, and Tianjin; machine building and building materials in Hebei, Liaoning, Guangdong, and Beijing; building materials and (in second place) machine building in Jilin, Fujian, Shandong, Henan, 264 WANG TUOYU Hubei, Shaanxi, Gansu, Qinghai, and Ningxia; coal and building materials in Shanxi, Nei Monggol, Guizhou, Yunnan, and Xinjiang; and building ma- terials and food processing in Heilongjiang, Anhui, Hunan, and Sichuan. These pattems demonstrate that rural industrial structure is mainly deter- mined by natural resource endowments and the level of economic develop- ment. In general, given a roughly similar environment, the industrial struc- ture in developed areas tends toward comparatively capital-intensive and technology-intensive undertakings and that in underdeveloped areas toward activities that use natural resources and labor more intensively. This does not contradict the common sense of economics. The Evolution of Regional Imbalances in Recent Years Economic growth is inherently a dynamic process that is necessarily accom- panied by imbalances of various kinds, including regional imbalances. The problem is not that imbalances have emerged in TVP development but whether such imbalances will deteriorate in the future, how long the deterio- ration will continue, and what impact this will have on regional economic development, especially on per capita income gaps among regions. Table 12-7 demonstrates that between 1979 and 1985 TVCES in the three regions developed at different speeds. In region A the share of TVCE employ- ees rose by 6.0 percentage points and the share of sales revenues by 6.7. In region B, however, the share of employees decreased by 4.0 percentage points and that of sales by 4.6, and in region C the share of employees dropped by 2.0 percentage points and that of sales by 2.1. If we looked only at the devel- opment of TVCES, we rmight conclude that the gap between the developed region and the other two has been widening and the gap between the devel- oping and underdeveloped regions has narrowed. But during this period pri- vate enterprise grew rapidly in China's rural areas. If we include private en- terprises, between 1984 and 1985 the share of TVP employment in region A dropped by 6.5 percentage points and that of sales by 3.0. In region B the share of employment rose by 4.3 percentage points and that of sales by 2.6, and in region C the share of employment rose by 1.7 percentage points and that of sales by 0.4 (see table 12-8). Although the proportions of TVCE em- ployment and gross revenues in regions B and C dwindled, those of all TVPS expanded.3 Thus the mushrooming of private enterprises in rural areas since 1984 has had a positive impact on the regional structure of China's TVP sector and has begun to contribute toward reducing regional imbalances in the de- velopment of TVPS. The fundamental purpose of encouraging rural nonagricultural production is to promote rural development and raise per capita incomes. It is thlerefore the evolution of the gap in rural average per capita income among regions that warrants our attention. T'he figures in table 12-9 show that the gap wid- ened between 1981 and 1985. Per capita income in region B was 86.9 percent Regional Imbalances 265 Table 12-7. Changes in the Regional Distribution of TVCEs, 1979-85 (percentage of national total) Item Region A Region B Region C 1979 Employees 53.7 38.6 7.7 Revenue 63.0 31.1 5.9 1980 Employees 56.0 36.4 6.6 Revenue 65.2 29.7 5.1 1981 Employees 58.6 35.6 5.8 Revenue 66.8 28.8 4.4 1982 Employees 59.0 35.1 5.9 Revenue 66.4 29.0 4.6 1983 Employees 59.4 34.7 5.9 Revenue 67.5 28.1 4.4 1984 Employees 60.2 34.3 5.5 Revenue 68.0 27.9 4.1 1985 Employees 59.7 34.6 5.7 Revenue 69.7 26.5 3.8 Source: Information from fieldwork. of that in region A in 1981 but dropped to 76.4 percent by 1985, an average annual decline of 2.0 percentage points. Per capita income in region C in 1981 was 74.5 percent of that in region A; by 1985 it had dropped to 67.0 percent, an annual decline of 1.5 percentage points. But since 1983 the gap has grown more slowly. The relative index in region B dropped by 7.7 percentage points between 1981 and 1983 but by only 3.0 between 1983 and 1985. The relative index in region C dropped by 4.7 percentage points between 1981 and 1983 but by 2.8 between 1983 and 1985. The slower growth of the income gap is consistent with the better balance in TVP development brought about by the growth of private enterprises. The role of private enterprises in ameliorat- ing regional income gaps should be recognized and policies adjusted accord- ingly. Prospects for Promoting Rural Nonagricultural Production in Different Regions This section discusses the prospects for nonagricultural development in rural 266 WANG TUOYU Table 12-8. Regional Distribution of Employees and Gross Revenues in China's TVPs, 1984 and 1985 Item Region A Region B Region C Percentage of 7VP employment in national total 1984 57.8 35.9 6.3 1985 51.3 40.7 8.0 Percentage of TVP revenues in national total 1984 66.3 29.3 4.4 1985 63.3 31.9 4.8 Source: Information from fieldwork. Table 12-9. Changes in Relative Per Capita Incomes in Rural Areas, 1981-85 Item Region A Region B RegLon C National average 1981 Per capita income (yuan) 259.95 225.89 193.72 223.44 Compared with regiorn A (percent) 100.0 86.9 74.5 86.0 1983 Per capita income (yuan) 365.86 289.63 255.54 309.77 Compared with regior, A (percent) 100.0 79.2 69.8 84.7 1985 Per capita income (yuan) 468.08 356.47 313.64 397.60 Compared with region, A (percent) 100.0 76.2 67.0 84.9 Note: The per capita incorme figures are weighted averages of the income per rural resident in all the provinces and autonomous regions in each region. The 1981 weights are calculated according to the number of commune members, and those of 1983 and 1985 are worked out according to the size of the niral work force. Source: State Statistical Buireau (1982, 1986a). areas in light of the influence of the current Chinese system on the interre- gional flow and organization of the factors of production. The discussion fo- cuses on sources for the iFunds needed for nonagricultural production in rural areas, since for some timne to come capital requirements will be more of a problem than supplies of labor and land. Prospects for Resource Mobilization Capital comes from three sources: funds accumulated from local agricultural production, funds accumulated from local nonagricultural production, and loans and outside investments. These are here termed, with consideration Regional Imbalances 267 Table 12-10. Distribution of Rural Net Income (percentage of net income) Collective Individual Item Taxes reserves incomes 1979 4.9 14.5 80.6 1985 5.4 8.1 86.5 Change 0.5 -6.4 5.9 Source: State Statistical Bureau (1982, 1986a). for the kinds of statistics available, private bank savings, funds accumulated by TVCES, and outside investments. PRIVATE SAVINGS. Great changes in the distribution of net incomes have taken place since the introduction of the production responsibility system (PRs) (see table 12-10), and private savings in rural areas now play a much greater role in the development of nonagricultural production. Two factors determine whether private savings can be effectively channeled into nonagri- cultural production. * Government policies on private economic activities and investments. If farmers have doubts about the consistency of government policies on the private economy, they will be reluctant to invest their savings in nonagricultural undertakings with slow economic returns and will pre- fer to invest in nonproductive assets such as houses or in nonagricul- tural activities that offer quick economic returns. The degree of development and functioning of local capital markets. Since China's state financial institutions are not yet providing large- scale loans for private investment, the accumulation of funds for private nonagricultural activities depends to a great degree on the development of nongovernment capital markets. In most underdeveloped and devel- oping areas per capita incomes are low, and people have spent the greater part of the recent increases in income on consumption. Because of the underdevelopment of nongovernment capital markets and the poor industrial environment, there is an apparent tendency toward short-term investments, as reflected in the large proportions of private investment in construction, commerce, catering, services, and other trades. In a word, the conditions for turning private savings into nonag- ricultural investment funds are much poorer in region C than in regions A and B. FUNDS ACCUMULATED BY TVCES. The level of development of TVCES differs from one region to another, as does the role of funds accumulated by these enterprises in the promotion of nonagricultural activities. In 1985 the net profits of TVCES in region C were Y549.9 million, of which Y293.5 million 268 WANG TUOYU were used for industrial investment. On the assumption that the net value of fixed assets of each TVCE plus its circulating funds totaled Y75,200 in 1985,4 the net profits for the year could be used to set up 3,903 such firms, or fewer than seven in each county. If the per capita net value of fixed assets plus circulating funds of TVCES was Y2,842 in 1985, the profits could be used to employ 103,283 workers, or only 169 per county.5 This figure is less than 0.2 percent of region C's current rural labor force. The number of jobs cre- ated in 1985 by reinvestment by TVCES in region B accounted for 0.2 percent of the total rural labor force in the region. In the same year TVCES in region A generated Y12.54 billion in profits, of which Y5.58 billion (44.5 percent) were accumulated as production funds. This amount of money could be used to create 1.062 million jobs, or 1.9 percent of the rural labor force in region A. These calculations show that the funds accumulated by TVCES will have only a limited influence on nonagricultural development in regions B and C. OUTSIDE INVESTMENT AND PREFERENTIAL LOANS FROM THE GOVERINMENT. As China's economic reforms proceed and economic actors become more aware of their own interests, outside investment will increasingly be made only with an eye to profits. W'hether a region can attract outside investment depends on its raw materials and transport facilities, its market prospects, and its supply of qualified labor. The first factor is decisive in region C's ability to absorb outside investiment. As for government support, two factors have to be taken into considera- tion: the level of support and its influence on nonagricultural development prospects. The central government will surely give priority consideration to economic efficiency in the future, and its supply of funds to underdeveloped areas will therefore not be large, since the efficiency of capital utilization in these areas is much lower than in more developed regions. Statistics for 1985 indicate that region A could generate a net output value of Y550 by turning one farmer into a worker, as against output figures of only Y213 in region B and Y383 in region C.6 Moreover, every Y100 of investment can generate Y26. 1 in taxes and profits in TVCES in region A but only Y18.7 in region B and Y19.9 in region C. Under the current system, funds channeled into the developed region generate greater economic retums. Another issue is the influence of the interregional flow of funds on TVP development. At present: China's rural funds flow from one region to another through such channels as direct private investment, investment by TVCES outside their localities, and interbank borrowing. Private funds cannot flow from one region to another on a large scale at present because of the underde- velopment of interregional transport and communications, the lack of transregional nongovemrmentail credit institutions, and the comparative iso- lation of private bank savings. Moreover, the tendency of TVPS to strive for maximum profits within the locality and the obstacles to horizontal invest- Regional Imbalances 269 ments have rendered it difficult for firms to make large transregional invest- ments. Consequently the role of government financial organizations and of the rural credit cooperatives under their control is extremely important. The current management system of Chinese banks emphasizes "unified planning, differentiating funds, extending loans according to absorption of savings deposits, and mutual assistance." This system has restricted the inter- regional flow of funds. In particular, the principle of mutual assistance does not permit a bank to reap any economic returns from financial assistance to another region. This situation will inevitably have cumulative effects on re- gional imbalances, since the more funds a region has, the more bank loans it will receive. Of the total bank loans received by China's TVCES in 1985, for example, 18.1 percent went to Jiangsu Province (including Shanghai), 10.6 percent to Shandong Province, 12.0 percent to Guangdong Province, 9.9 percent to Zhejiang Province, and 7.3 percent to Hebei Province (in- cluding Beijing and Tianjin). These five provinces got 57.8 percent of total bank loans to TVCES in that year. Region A received 65.2 percent of total bank loans, region B 31.0 percent, and region C 3.9 percent. Thus the cur- rent bank management system intensifies imbalances in the distribution of China's rural investment funds. As banks have become increasingly aware of credit risk factors, a new situ- ation has arisen in the underdeveloped region. Despite a great demand for funds to develop rural nonagricultural production, banks in these areas have large overstocks of lendable funds. The Guangxi Zhuang Autonomous Re- gion provides an example. In 1985 the regional agricultural bank promul- gated new regulations governing the extension of loans to TVPS. To borrow money for a project, a TVP had to be able to supply 30 percent of the project's total investment and it had to have an economic entity as guarantor. In 1986 the bank allotted Y13 million for equipment loans to TVPS in Yuling Prefec- ture. Since Y2.44 million in funds for special loans were left over from 1985, the prefectural branch had Y15.44 million in funds available for equipment loans, which would be accompanied by Y100 million in circulating capital loans. By April 1986 the bank's equipment loans to TvPs accounted for only 8 percent of its total lendable funds for this purpose, and only 15 percent of its total lendable funds for circulating capital had been extended. Under pressure from the local govemment the bank lowered its requirements for loans in the second half of the year: a TVP needed to have only 10 percent of the project investment, and its guarantor could be a TvP bureau at the county level. How well banks can support local governments' development strategies de- pends to a considerable extent on the degree of local prosperity. In a more developed area bank employees can reap more personal benefits from the local economy, and the bank as a result tends to be more cooperative toward the local government. In a less developed area bank employees can gain little from the local economy, and bank branches are more willing to stay within 270 WANG TUOYU the confines imposed by the vertical management system of the nation's fi- nancial system. Once private savings are deposited in the agricultural bank in region C, they are less likely to be returned to circulation in the local economy, and the difficulties that stem from the limited amount of private savings in region C are exacerbated. It is apparent that under China's current system there are no signs of im- provement in the environment for capital supply in the country's vast under- developed rural areas. The existing market, technical, and labor force condi- tions for developing nonagricultural production in rural areas give cause for worry about a further intensification of regional imbalances. Although the imbalances may be improving somewhat, progress will not be sustained with- out further efforts to restructure the current system. Measures to Ameliorate Regional Imbalances All measures to solve the problem of regional development must start from two supporting bases: human development and income equalization. The fol- lowing are areas in which actions grounded on these bases may be considered. FACILITATING INTERREGIONAL FLOWS OF LABOR. The prerequisite for rais- ing the overall efficiency of an economic system and achieving interregional equality in per capita incomes is a free flow of the factors of production. An important objective of China's economic reforms in recent years has been to break down the barriers of administrative jurisdictions and expand market mechanisms. Spatial flows of the factors of production have indeed increased in recent years-undoubtedly a good beginning. In light of practical condi- tions in China, priority should be given to interregional flows of labor, since without such flows interregional movements of other factors of production will only intensify regional imbalances in TVP development and, what is worse, widen regional gaps in per capita incomes arid encourage the forma- tion of independent communities on the basis of regional interests. The uni- fied national market for other factors of production will inevitably disinte- grate, bringing flows to a stop. Legal and social restrictions hinder both rural-urban and rural-rural flows of labor. The restrictions on rural-urban movement are the inevitable out- come of an industrialization process based on public ownership of property, state monopoly, and exploitation of agriculture. Lack of inputs reduced agri- culture's productivity and ability to accumulate funds, and industry also found it difficult to expand and to increase its financial surplus. Rapid popula- tion growth and low agricultural productivity left little surplus grain to sup- port the shift of agricultural vvorkers to the nonagricultural sector. Moreover, when all private businesses were banned, the state became the only employer and purveyor for urban residents and had to control the size of the urban population according to how much grain was available. Since the govern- Regional Imbalances 271 ment decided who would have access to urban employment, it in effect recog- nized urban residents as holders of vested interests. A principal reason for the government's slowness in opening up urban and rural areas to each other was its concern about the adverse effects of change on the interests of urban residents. Efficiency and equity were not considered when the current policy of separating the urban and rural populations was formulated. This policy should be gradually abandoned. The development of the private economy in urban areas and the increasing exchanges between urban and rural areas in recent years have already created the preliminary conditions for a reorientation of policy. As the country's grain production increases and more grain markets emerge, urban employers should gradually cease discriminating against rural residents when hiring workers. This will not only help narrow the income gap between urban and rural residents but will also lower the costs of industrial production and improve the efficiency of the national economy. As a first step, the Chinese government should relax controls over household registration in small cities and towns and re- form the grain supply system accordingly. Obstructions to the free flow of labor among rural areas are rooted in farm- ers' dependence on land. Under China's traditional system peasants could not engage in nonagricultural activities and were confined to agricultural pro- duction. Thus there was little reason for interregional labor flows. Under the rural economic reforms interregional flows of rural labor have greatly in- creased in recent years. But because the overall environment for nonagricul- tural activities in rural areas is not yet stable, most farmers cannot yet with- draw from their dependence on the land, and many rural households pursue nonagricultural activities only as a sideline. Rural movement of labor is mainly individual instead of by households (see Lu Mai and Dai Xiaojing 1987). Because the income gap among regions is large, people in low-income areas have a strong desire to move once they have abandoned the traditional con- ception of "never leaving one's native land unless it is absolutely necessary." The main issue is the ability and willingness of people in high-income areas to accept them. Rural areas in economically developed coastal regions, espe- cially near big and medium-size cities, have already handed over many of their traditional economic activities to outsiders in recent years because of the emergence of new opportunities for developing nonagricultural produc- tion (see chapter 14). The ownership structure of TVPs also influences the degree of acceptance of outside workers by people in developed areas. Gener- ally speaking, townships and villages launch TVCES with the interests of the locality in mind, and they would rather pay more to local workers than pay less to outsiders. Private enterprises may be more willing to hire outside labor at lower wages. The increased flow of labor in China's rural areas in recent years has pro- moted the development of nonagricultural activities and helped achieve 272 WANG TUOYU more balanced incomes amonlg regions. Such flows should be encouraged by the Chinese govemment. To tum the interregional flow of labor into a real- ity, efforts should be made to institute land reform (see chapter 6) and retrain rural laborers. REFORMING THE MIC]ROECONOMIC BASIS FOR DEVELOPING RURAL NONAGRI- CULTURAL PRODUCTION. The development of nonagricultural activities in China's rural areas started with community govemments' initiatives in set- ting up enterprises. This phenomenon, an inevitable result of China's tradi- tional system, helped to pronmote the rural economy and increase farrners' in- comes. It has great limitations, however. The ultimate objective of community enterprises is to reap maximum benefits for the community. This objective impedes the ftee spatial flow of factors of production and further splits up domestic markets that were already isolated from each other. Fur- thermore, by reproducing or intensifying China's traditional system it creates a new obstacle to economic reform, which is already beset with difficulties. The introduction of the PRS, the development of nonagricultural production in rural areas, and the elimination of the state monopoly in purchasing and marketing have created excellent opportunities for seeking a new road to rural industrialization through launching private enterprises. Our samLple sur- veys show that the basic purpose of private enterprises is to maximize profits on capital. This goal in itself requires a breaking down of spatial barriers, which will provide a microeconomic basis for the free spatial flow of factors of production. The Chinese govemrnent already has a basis for implementing a policy of promoting private enterprise. The changes in people's ideas and the imnprove- ment in their basic living standards have provided a good opportunity for altering the practice under which grass-roots rural govemments are responsi- ble for everything. Where TVCES are highly developed and local governments have been able to guarantee the basic income levels of their residents, au- thorities should try to solve the contradiction of "big govemments with small treasuries" by encouraging the development of private enterprise. From a macroeconomic point of view, the development of private enterprise can, help to weaken the economic functions of community govemments. Where TVcEs account for a comparatively small share of nonagricultural production, changes in their orientation shiould occur, away from a sole focus on maximiz- ing benefits to the comnmunity. The contribution of private enterprises to ameliorating imbalances in the development of nonagricultural activities in China's rural areas in recent years and equalizing per capita incomes interregionally shows that this kind of policy is beneficial. RATIONALIZING STATE SUPPORT FOR DEVELOPING AND UNDERDEVELOPED REGIONS. Since under the conditions prevailing in China it is difficult to achieve free interregional flows of labor, the government should provide Regional Imbalances 273 some help to developing and underdeveloped areas. Such assistance should be given in proper forms and with due consideration for the advantages and disadvantages in each case. It should go mainly to areas with low per capita incomes, large populations, few agricultural resources, and great market po- tential, such as Yunnan, Guangxi, Guizhou, and Gansu in region C and Shaanxi in region B. Attention should be paid to selecting projects according to conditions in each area to promote a regional division of labor. The cen- tral govemment may also help local govemments with agricultural produc- tion and investment to improve the local environment for nonagricultural development. If the Chinese govemment adopts the policies outlined above, regional imbalances in the development of nonagricultural activities in rural areas will persist in the future, but with differences. Such imbalances, unlike the exist- ing ones, will be accompanied by an intensification of regional specialization and an equalization of per capita incomes and will be widely accepted as part of a rational development process. Notes 1. Some Chinese TVPS engage in agricultural production. Because they are small in number, these firms are included in our use of the term TVPS. In 1985, for example, the number of employ- ees in agricultural enterprises accounted for only 6. 1 percent of the total for all TVCFS, and their total revenues were 3.1 percent of those for all TvcEs. The percentages would be even smaller if private enterprises were included. 2. Because first-grade (provincial-level) administrative areas are used as our basic regional de- marcation, differences within each province cannot be taken into account. 3. Statistical materials for 1986 show that for all types of TVPS (including private enterprises) the share of region A continued to decline while the shares of regions B and C continued to rise. 4. This figure is calculated by dividing the total year-end net value of fixed assets and circulat- ing funds of the country's TVCES in 1985 by the total number of these enterprises. 5. Based on the total of the net (depreciated) value of fixed assets plus quota circulating assets of TVCES and their total employment in 1985. 6. These are arithmetic averages of the indexes of all the provinces in each region. The index in each province is the average net value of output created by each TVP employee minus the average net value of agricultural output created by each member of the agricultural labor force. Part IV Labor Labor and wages in the TVP sector are of great interest and critical impor- tance. Lower overall labor costs and apparently superior labor motivation and effort have been principal sources of TVPS' competitive advantages in relation to state enterprises. Wages of Tvp employees are much more closely tied to individual and enterprise performance than are those of most state workers. Various types of piece rates are extremely common, and many firms pay a year-end bonus related to growth of enterprise profits or output. Most TVP workers probably work considerably harder and for longer hours than their counterparts in state enterprises. Labor relations in TVPS seem surprisingly good, particularly considering the hard work, long hours, difficult working conditions, and relatively low wages. Possibly this is because TVP employees are first-generation industrial workers accustomed to the rigors of farm life. Rural labor markets are highly fragmented territorially and are not func- tioning smoothly in most areas. Labor is still rather immobile among prov- inces, localities, and communities, except for temporary migrant labor, mostly in low-skill activities. In such places as Wuxi labor is still allocated to community enterprises by township authorities, and inflows of workers from outside have been discouraged. In other areas there is much less admin- istrative control, but various frictions impede the free flow of labor. The community-based rural institutional structure, the household registration sys- tem, and continuing restrictions against permanent migration to larger cities by rural people all adversely affect labor mobility among rural areas. The three chapters dealing with labor issues shed light on this vast topic from different angles. In chapter 13 Alan Gelb surveys workers' attitudes, wage systems and in- centives, and labor relations in the TVP sector on the basis of data from the Worker Survey Questionnaire. He first describes four models of labor rela- tions that might apply to the TVP sector-the public enterprise and commu- nal models for community enterprises and the Taylorist and Z-firm archetypes for private firms. Gelb argues that community enterprises fit the communal 275 276 LABOR better than the public enterprise model and that private enterprises resemble Z-firms more than they do the Taylorist stereotype of labor relations in capi- talist enterprises. The implication is that pay in all TVPS generally is related to firm (and community) economic performance. Moreover, differences be- tween "communal" community enterprises and "Z-firm" private enterprises are relatively small. The chapter contains interesting findings concerning the characteristics, attitudes, and pay of the Tvp employees surveyed. Regardless of the pay sys- tem actually in use, workers perceive a close relationship between their in- comes and firm profitability. In general, they seem to be fairly content and to regard labor relations in their firms as relatively good. Dispersion in aver- age pay levels is substantial, but it is mainly among rather than within firms. The striking differences a'mong counties for the most part dominate any effects that different forms of firm ownership may have. Local labor markets appear to be more open in Nanhai and Jieshou, where a larger proportion of sample workers came to their present employment through "voluntary" routes, than in Wuxi and Shangrao. The proportions of female workers and young workers are highest in Nanhai, and Nanhai workers indicate signifi- cantly greater willingness to change firms than do those in the other three counties. Dispersion of pay is greatest in Nanhai, which has greater competi- tion among different types of firms and fewer communal constraints; it is rather narrow in Wuxi. Shangrao workers show signs of dissatisfaction with poor management in their firms and-perhaps as a consequence-less con- cern about their enterprises and about their fellow workers' performance. Gelb analyzes wage determination in some detail, estimating several wage equations. He finds that individual worker characteristics, occupation, and firm ownership fail to explain a large proportion of the overall variation in pay levels for sample workers. Simple firm dummies, however, do quite well; together with occupation, age, sex, and number of days worked in the year they explain 70 percent of the variation in pay. This may be because commu- nity income levels are an important determinant of wage levels. Not only is education insignificant in determining wages, but when the effects are (marginally) significant, they go the "wrong" way; that is, higher education is not necessarily associated with higher pay. This result is highly unusual in comparison with findings from similar analyses in other countries. Inter- firm wage differentials reflect the impact of different growth rates of labor productivity. Gelb concludes by suggesting that the community orientation of TVps, re- gardless of form of ownership, may be the main reason why they follow the communal or Z-firm model rather than the public enterprise or Taylorist ar- chetype. Another important finding is that because of the prevailing commu- nal orientation and pay systems, TVP workers behave as though they have substantial de facto equity stakes in their firms. In chapter 14 Meng Xin describes and analyzes the rural labor markets Labor 277 faced by TVPS. Her theoretical foundation is the model of economic develop- ment with unlimited labor supply developed by W. A. Lewis in the 1950s. In this model the unlimited supply of low-wage surplus laborers from the agri- cultural sector permits high accumulation rates and rapid industrialization over a relatively long period until the surplus labor is exhausted. But Meng points out that an unstated precondition for the model to work is a well- functioning rural labor market, and she demonstrates that this precondition is not met in China. The chapter first looks at interregional, intercommunity, and inter- enterprise mobility of labor in the four counties. Flows of labor among counties and provinces are limited except in Nanhai, and flows among com- munities and enterprises are much more important in Nanhai and Jieshou than in Wuxi and Shangrao. The limited labor mobility in Wuxi is especially striking because of the high level of industrialization and TVP development and the emerging labor shortages in the county. (In Nanhai shortages have stimulated a substantial inflow of labor from outside.) Meng argues that the lack of labor mobility in Wuxi is attributable to the domination of TVCES in the TVP ownership system. This contrasts with Nanhai, where the share of TVCES is lower and a large proportion of TvcE employees are temporary labor- ers not under the supervision of community governments. The author goes on to look at wage determination in the four counties. Not surprisingly, TVP wage levels in the two highly industrialized counties, Wuxi and Nanhai, are much higher than in the less developed ones, Jieshou and Shangrao, but wages in Nanhai exceed those in Wuxi by more than 70 percent. Wages have risen at a much more rapid rate in Nanhai and Wuxi than in Jieshou and Shangrao. Wage differentials at the intercommunity and interenterprise levels are large in Nanhai but small in Wuxi and Jieshou. Meng argues that TVP wages are much more responsive to labor supply and demand conditions in Jieshou and Nanhai than in Shangrao and Wuxi. In Wuxi, despite the growing shortage of labor in the TVP sector, agricultural incomes caught up with and even surpassed those of TVP workers, partly be- cause of strict administrative controls over wages in TVCES. An apparent para- dox is that despite greater labor mobility in Nanhai than in Wuxi, intercom- munity and interenterprise differentials are much greater in the former than in the latter. Meng suggests that this is because of administrative controls in Wuxi and the greater degree of community specialization in particular in- dustries in Nanhai. In Jieshou, where the industrial structure is less diversified and skill requirements are more uniform, differentials are relatively small. The higher differentials in Shangrao are attributed to immobility of labor and uneven distribution of natural resources, some of the rents from which accrue to labor. In Wuxi community governments have succeeded in balancing income levels between agriculture and (TVP) industry. The egalitarian distribution system does not appear to harm efficiency, perhaps because TVP workers' em- 278 LABOR ployment and income prospects are uncertain and their wages are strongly linked with performance through various piece-rate systems. In Nanhai com- munity governments are unable to control wages in TVCEs because of well- functioning labor markets and the ability of private firms to attract employees away from community enterprises by offering high wages. This partly explains the rapid increase in TVP wages there. The overall picture that emerges is that China at present does not have a unitary labor market for rural-urban flows or interregional flows. Even though the national situation may approximate the "unlimited supply of labor" of the Lewis model, labor costs in particular rural localities have sky- rocketed owing to localized labor shortages. As a result, the advantages of being able to draw on an unlimited supply of labor at a subsistence wage are lost; in the rapidly developing areas with local labor shortages, rising wages cut industrial accumulation and may also lead to greater capital intensity than is economically appropriate for the nation as a whole. The impact of local labor markets differs according to the level of TVP de- velopment. In the early stages, when there is still surplus local labor to be absorbed from agricultLre, a local labor market can be helpful in giving enter- prises appropriate signials on factor mix and in holding down wage irncreases. Interference in wage determination by community governments can, lead to excessively high wages and bloated payrolls. Once TvP development is rela- tively advanced and local labor shortages have emerged, a functioning local labor market facilitates excessively rapid increases in wages (as occuarred in Nanhai before it was opened[ up to interregional labor flows in 1986). Para- doxically, community governments' involvement in wage determination can help limit wage increases in this situation. In chapter 15 Wu Quhui, Wang Hansheng, and Xu Xinxin loolc at the influence of the structure of alternative labor opportunities on wage cletermi- nation in three of the four counties-Wuxi, Nanhai, and Jieshou. Their main conclusion is that alternative opportunities play a great role in deter- mining wage levels in different localities and can lead to significant differ- ences in income levels even among localities with similar levels of economic development. The authors assert that the structure of alternative labor opportunities is closely related to the ownership pattem and the degree of labor mobility. When the bulk of the rvp labor force is in TVCES (as in Wuxi), community governments can control wages effectively. If there is a substantial number of private enterprises (as in Nanhai), community governments cannot set TVCE wages below market-clearing levels because their employees would move to the private sector en masse. If, however, there is both surplus labor and a substantial private sector (the situation in Jieshou), TVCE wages will be held down by local labor market conditions. A high degree of iriterfirm labor mobility (associated with the existence of a significant private sector) weakens the ability of community governments to control wages. Labor 279 The authors go on to describe how community governments try to balance the incomes of community members. This is easier if TVCEs account for the bulk of local employment and if alternative employment opportunities are limited. In Wuxi TVP profits are used to support agricultural incomes. The structure of altemative employment opportunities also affects wage differentiation within firms, in particular managerial compensation. In Wuxi increases in managerial compensation are tied to increases in workers' pay, and the gap between the enterprise director's pay and the average wage can- not become too wide. In Nanhai, however, TVCE directors' pay is usually linked to enterprise profits, which puts directors in a more favorable position and gives them incentives to hold down wage increases. One reason for these different methods and the lower differentials in Wuxi is that in Nanhai com- pensation for TVCE managers has to be competitive with pay in private enter- prises, whereas in Wuxi the monolithic ownership system allows community governments to exert effective control over management pay differentials both within and among firms. In Jieshou differentials within firms are small because of labor market conditions and the generally simple skill require- ments. In actuality, many TVCEs are basically private enterprises, and their managers give themselves low nominal wages to show a narrow differential to the outside world and to facilitate holding down workers' wages; they can appropriate the returns from "their" enterprises in other forms. The authors describe different modes of wage determination in the three counties for both workers and managers. In Wuxi wages are to a large extent under the discretionary control of community governments, whereas in Nan- hai and Jieshou there is a strong linkage with profits (for managers) or output (for workers), and once the basic linkage is set there is little if any interfer- ence by community governments. The chapter concludes with some additional implications. One is that TVP employees have more community consciousness in Wuxi, where pay is to a large extent controlled and balanced by community government authorities and differentials within communities are small. In Nanhai and Jieshou com- munity governments do not play an important role in income distribution, and workers tend to demonstrate strong individualism (Nanhai) or to identify with their firms (Jieshou). As a result of these differences, the three counties will face different labor problems in the future. Community governments in Wuxi will have to concentrate more on efficient performance in the TVP sec- tor and less on income distribution, whereas the opposite is true for Nanhai. 13 TVP Workers' Incomes, Incentives, and Attitudes Alan Gelb This chapter provides a perspective on workers in TVPs-on their working relationships within their firms, their income levels, and their occupational characteristics. The term `TVP workers" is used here in a broad sense to in- clude technicians, managers, and other employees, except where it: specifi- cally denotes an occupational category. The analysis is based on the Worker Survey Questionnaire, which contains fifty-nine subjective and objective questions. The subjective questions cover workers' views on such topics as industrial relations, attitudes toward work, and reward systems. The objective questions deal with pay and with demo- graphic and similar variables. Since some questions have multiple parts, there are 107 responses, in all, to each questionnaire. Selection of workers within firms was done on a stratified random basis. Proportionately more workers were sampled for smaller firms, and technical and management personnel were somewhat oversampled to ensure sufficient diversity for statistical analy- sis. Researchers helped workers to complete the questionnaire. In the four sample counties 1,174 usable questionnaires were obtained, but 2 were dropped because their firm classifications were probably incorrect. To facilitate analysis, thirty-six questions were selected to yield a clean, compact data set. In addition, certain response options offered in the survey were consolidated and simplified to render them more mutually exclusive and to weight responses according to frequency or intensity. The worker data were then linked up with selected firm variables from other questiornaires, including firm ownership type, work force, sales, and profitability, all for 1985 when possible. The distribution of sample workers and their firms by county and firm type is shown in table 13-1; the 1,172 workers are distributed in 49 firms, with an average of 24 workers per firm. Although there is a wide spread of firms by type and county, township firms dominate, and there are no private firms in Wuxi and Shangrao. Thirty-two enterprises are owned by commun.al gov- emrnments, nine are private, and eight are other types, mostly mixed owner- 280 TVP Workers' Incomes, Incentives, and Attitudes 281 ship. Some minor adjustments were made to the data, but only where there were clear indications of unreasonable responses.' This chapter reports on a first analysis of the survey data. Because only the TVP sector was sampled and there was no control group, the survey can- not be used directly to assess differences between TVPs and state enterprises. The responses can, however, shed light on the nature of the TVP sector itself in relation to various possible archetypes or models, and they can be used to test hypotheses on differences between the characteristics of the TVP sector in the four sample counties. Since the TVP sector has evolved in a wide vari- ety of situations, different types and stages of development can be distin- guished within the sector. These differences and their systematic association with certain counties pose considerable difficulties for the interpretation of tests of hypotheses. For example, the hypothesis that pay is higher in larger firms might be preferred over the proposition that firm size is not correlated with pay. But the first hypothesis might look better only because firms tend to be larger in the more advanced areas where incomes are generally higher. Similarly, one might conclude that private enterprises pay less than community enterprises or em- ploy fewer women, but it might simply be that private enterprises are concen- trated in poorer, less advanced areas. Multivariate estimation techniques are generally preferable to bivariate ones in this situation, but owing to the pre- liminary nature of this analysis, bivariate relationships are frequently re- ported. Another point, as will be seen below, is that the tails of the distributions of some variables are heavily influenced by the inclusion of a few firms that may not be representative of the TVP sector as a whole, although they are interesting. Care should therefore be taken in extrapolating the survey results to the whole TVP sector. Some Broad Hypotheses and Models for TVPs TVPS constitute an "in-between" sector. They are part public, part communal, and part private. They operate partly in free markets for factors, inputs, and products and partly in planned ones. As a framework for the analysis, we develop four archetypes, each with its distinctive set of explicit or implicit labor contracts. First, we look at two models for community enterprises. * The public enterprise model. Ownership is by a large community such as the nation-state. Pay levels are set administratively with little reference to individual or firm performance. Pay is compressed within firms and is similar among firms. Labor is allocated to firms rather than hired on the market, and workers have great job security. * The communal model. Firms are owned and controlled by a small com- Table 13-1. Distribution of Sample Workers and Firms (number) Wuxi Jieshou Nanhai Shangrao Four counties Type of fir Workers Firns Workers Firms Workers Firms Workers Firms Workers Firms s Township 361 11 57 3 121 4 110 7 649 25 Village 37 2 26 2 60 2 21 1 144 7 Private 0 0 128 7 17 2 0 0 145 9 Other 75 2 86 3 60 2 13 1 234 8 All firms 473 15 297 15 258 10 144 9 1,172 49 TVP Workers' Incomes, Incentives, and Attitudes 283 munity, in the sense that the employees form a significant part of the community. Workers have job security by virtue of being community members. They tend to be hired through a mixture of voluntary appli- cation and informal communal ties. Within a firm relative pay levels may be determined by.a range of systems, such as piece rates, time rates, or work-points, but the average level reflects, to a large extent, the income of the community and is therefore tied to the performance of the firm if the latter is important to the community. In the extreme case the work force itself is the community, and the firm becomes a kind of cooperative. The stronger are the small communities (say, pro- duction teams, villages, or the firms themselves) in relation to the larger ones (say, townships or counties), the wider are the likely pay differentials among firms. If labor mobility is low, substantial pay differ- entials can arise among communities. Two models can also be outlined for private enterprises. * The Taylorist model. This model is usually considered to be the dominant type in Westem economies. Labor is a variable factor of production, and tumover is high. Work pattems are systematically simplified to en- able one worker to be easily replaced by another. Except for managers, pay is not tied to firm performance. • The Japanese, or Z-firm model. In contrast to the Taylorist model, this pattern emphasizes strong identification of workers with their firms, low tumover, and close ties between the firm's profitability and payments to labor. This "patemalist-lifetime commitment" model is more plausi- ble where labor mobility is naturally low and family members and part- ners make up a substantial part of the work force. (The term Z-firm was introduced by Ouchi to denote U.S. firms that were following the Japanese model; see Ouchi 1982. It is used here to avoid exclusive iden- tification of this system of labor relations with Japan; many firms in many countries have larger or smaller degrees of "Z-ness." See also chapter 19.) Most firms would be expected to be some combination of the above four models. Moreover, it may not be easy to distinguish a Z-firm from a commu- nal firm except by ownership (private or public). A number of hypotheses about how TVPS fit into these categories can be proposed. * Hypothesis 1. In the TVP sector as a whole, community enterprises tend to follow a communal rather than a public enterprise model, and private enterprises follow the Z-firm rather than the Taylorist model. This implies that pay is generally related to enterprise performance and that differences between private and community enterprises are minimal. * Hypothesis 2. There are significant differences among the four counties. Small communities and firms are stronger in Nanhai than in Wuxi, so that 284 ALAN GELB the former is more "communal" and varied, whereas the latter displays more of the characteristics of the public enterprise model. Jieshou is more market oriented than Shangrao, which displays signs of a more traditional attitude toward rvPs. * Hypothesis 3. There are also significant vintage effects that stem from the timing of the establishment of the TVP sector in relation to the phase of reform. * Hypothesis 4. Because of the communal nature of both commlnity and private enterprises and low labor mobility in China, pay is still largely de- termined by firm and community. Personal characteristics such as type of work and educational level have only a secondary impact. Inequality gen- erated by the growvth of the TVP sector therefore largely reflects the fortunes and dynamism of different firms and communities, not the emergence of wide differentials within firms. These four hypotheses, together with the characteristics of the models out- lined above, imply somne specific propositions that can be tested against the survey data. For example, hypothesis I suggests that there should be few sig- nificant differences between private and community enterprises once regional effects are allowed for. Hypothesis 2 suggests that income levels in Nanhai should be more dispersed than in Wuxi and that the reason is greater between-firm variability rather than especially wide differentials within firms. Labor markets should operate in significantly different ways in the four coun- ties and for workers of different vintages. Such propositions are tested in the following sections. General Features of the Sample The sample respondents are generally young; 13 percent are under 20 years of age and only 14 percent are over 39 years of age. Males constitute 53 per- cent of the total and females 47 percent. Most workers have gone through junior middle school, but only 2 percent have attended any type of college. Sixty-three percent classify themselves as workers (in the narrow sense), and the rest are more or less equally distributed among shift leaders, operations personnel, technicians, ordinary staff, and middle-level staff, with a few ap- prentices. The factory is their main source of income. Sixty-one percent work in their firms for twelve months a year and for six or (mostly in Wuxi) seven days a week. Few work less than ten months in the firm or less than five days a week. Satisfaction Levels Overall, sample workers seem fairly content-80 percent consider themselves to be basically or very satisfied and only 8.5 percent express a desire to leave TVP Workers' Incomes, Incentives, and Attitudes 285 their firms. Seventy-five percent plan to stay with their firms for at least five years, and only 9 percent consider it possible that they might lose their jobs. Many thought that a decision to leave their firms would be "difficult" to carry out. Only 22 percent agree with the statement that TVP workers' status is lower than that of workers in state enterprises, even though most of the re- spondents are rural residents and state enterprise workers enjoy the advan- tages of urban registration. Labor Relations Only 9 percent of sample workers consider that labor relations in their firms involve serious or frequent conflicts, and only 14 percent rate the manage- ment of their firms as poor. (This proportion was significantly higher in Shangrao, confirming impressions from interviews that many Shangrao firms are poorly managed.) Fully 83 percent of the workers see a substantial need for some type of trade union to protect their interests. Labor Allocation At least in a local sense, much of the TVP sector seems to operate in a fairly free labor market. Fifty-four percent of respondents came to their firms through modes that can be considered voluntary: their own initiative, intro- ductions by friends and relatives, or competitive examinations.2 The remain- ing 46 percent arrived through labor allocation, resource pooling, or other arrangements by the local government. This is one of the areas of large differ- ences among counties. In Wuxi and Shangrao 57 percent of the workers were allocated, whereas in Nanhai and Jieshou only 33 percent were. There is also a strong vintage effect; only 26 percent of the workers under 20 years of age were allocated, compared with almost 60 percent for the oldest cohorts. This reflects the evolution of policies relating to the TVP sector. Pay, Profits, and Attitudes toward Work and Investment Pay systems differ considerably among TVPS, and there is a tendency for those in the more advanced counties to be more complex and have a wider range of benefits than those in the less advanced areas. Few of the workers report that time rates are the main method of payment; piece rates and floating wages are more common. Only 19 percent of the workers indicate that bo- nuses are unimportant or unavailable. Other pay components often men- tioned as important include medical subsidies, pensions, and factory subsi- dies. Guaranteed employment for children in the same firm seems to be a less important benefit. Whatever the specific method of payment, workers clearly perceive a close relationship between their incomes and the profitability of their firms. Only 286 ALAN GELB 10 percent see no relationshlip, whereas 71 percent consider the link to be strong or very strong. The relationship is powerful for all counties and types of firms; regressions to predict it have little explanatory power, although it is somewhat stronger in larger than in smaller firms.3 Correspondingly, it appears that a large majority of respondents are pre- pared to assume some responsibility for the quality of their fellow ernployees' work. Only 14 percent are indifferent to poor performance or consider that it is a matter for managers alone; 32 percent would like to speak to the shirk- ers in their firm but feel inhibited; and 54 percent indicate that they would speak out. Again, this response appears to be widespread across counties, types of ownership, arnd occupations.4 In particular, it is not significantly lower in Wuxi than in the other counties. About half of the sample workers regard pay differentials between workers and managers in their firms or among workers as "substantial." Workers are divided on whether differentials should be allowed to widen (17 percent), should stay the same (41 percent), or should be narrower (42 percent). Most workers would deposit surplus income in a bank, use it to purchase consumer goods, or (especially in Wuxi) invest in housing. A smaller but appreciable number would prefer to use funds to purchase producer goods (including farming equipment) or to buy stock in their own or other firms. Although only 7 percent want to set up their own firms, alone or with others, 15 per- cent would buy a "large number" of stocks in their firms if these were offered, 71 percent would buy a "'small" number, and 12 percent would not buy stocks. There is thus evidence of considerable willingness to help with their firms' financing.5 Communal Variables Only 27 percent of sample workers say that the best use of factory profits is direct distribution to workers; 65 percent favor substantial reinvestment of profits in the firm. This response is probably related to the low mobility of labor and the consequent importance of firms in the future income streams of workers and their families. In response to a hypothetical question, 65 per- cent indicate they would accept a 50 percent pay cut and remain to help their firm through difficulties rather than actively look for another job. More than half of the respondents indicate that if they were faced with personal problems, they would first approach other workers or managers in their firms rather than family members or communal leaders. Sample workers consider their income levels to be average or slightly above the levels in their communities; only 2 percent describe their incomes as very high and only 9 percent as low. Interestingly, there is not a systematic rela- tionship between incorme level and perception of income level in relation to the community, despite the wide range of incomes reported in the sam- ple. This suggests that pay tends to be set largely with reference to the TVP Workers' Incomes, Incentives, and Attitudes 287 income levels prevailing in the community of which the firm is a part.6 The general picture is consistent with the hypothesis that the TVP sector follows a communal model and that in some cases the firm itself is a consider- able part of the community. It also confirms the propositions that workers perceive a strong link between pay and profits and that many view their rela- tionship with their firm as long term rather than transient. This is no doubt because of the limited mobility of labor. In this sense TVP workers have de facto equity in their firms even if they do not formally own shares. Many, it seems, would consider a degree of formal shareholding in their firms as well. It is also clear, however, that TVPs are not worker cooperatives, if the term is taken to mean that workers exert a large degree of control. Responses indi- cate that recruitment and dismissal, pay, work assignments, product develop- ment, investments, and use of retained profits are overwhelmingly decided on by enterprise leaders and that the leaders are appointed by the government or chosen by other leaders rather than elected. On average, only 6 percent of respondents felt that the views of factory workers had any weight in such decisions. Surprisingly, Shangrao workers indicate a somewhat higher degree of participation, including election of leaders by workers.7 There is some sen- timent for more decisionmaking power for workers, but the desire is not over- whelming. Only 14.2 percent of respondents indicated that they would like workers to have total or partial decisionmaking power in the areas listed above. Differences among Counties and Firm Types Because the TVP sector differs among counties in so many ways, many if not most relationships between survey variables and location by county differ sig- nificantly from randomness. This section discusses certain systematic patterns and their implications rather than trying to identify all points of difference. The Labor Market As table 13-2 shows, Nanhai and Jieshou appear to have freer labor markets than Wuxi and Shangrao, and younger workers there are less likely to have been allocated to their firms. Private enterprises also operate in a freer labor market. These differences are an aspect of county patterns of labor use. Nan- hai employs more women workers (57 percent) and the most young workers (22 percent of Nanhai workers are under 20 years of age). Wuxi and Shangrao have the largest proportions of old workers. Many of Wuxi's TVPS are long established, and Shangrao's firms are less dynamic than elsewhere, so the proportion of young workers is lower. Nanhai workers are more mobile than those in Wuxi; 27 percent plan to leave their firms within one to three years.8 Faced with the option of taking a 50 percent pay cut or leaving, 53 percent of Nanhai workers, by far the 288 ALAN GELB Table 13-2. Recruitment of Sample Workers (number of workers) Item AUocation Voluntary Total By county Wuxi 270 t91 461 Jieshou 84 209 293 Nanhai 99 154 253 Shangrao 78 66 144 Total 531 620 1,151 By age Under 20 40 114 153 20-29 233 320 554 30-39 168 117 285 40-49 71 57 128 50+ 19 12 31 Total 531 620 1,151 By ownership Public 398 380 778 Private 33 111 144 Total 431 491 922 Pay (yuan per year) Less than 625 82 180 262 625-1,249 212 263 475 1,250-2,499 173 142 315 2,500-4,999 44 31 75 5,000+ 20 4 24 Total 531 620 1,151 Note: For counties, x2 = 3.4; for age, x2 = 54.9; for ownership, x2 = 38.9; for pay, x2 = 51.5. In all cases in tables 13-2 through 13-5, the probability that the distributions are random is less than 0. 1 percent. highest proportion in the sample, opted for the latter. Probably because many Nanhai workers are not tied to their firms, there are far stronger pressures to distribute profits to the work force. Fifty-seven percent considered this to be the best use of profits, compared with only 27 percent in the entire sample. Young female workers are usually the lowest paid. This contributes to a pattern whereby the lowest-irncome workers are freely recruited whereas the highest-income workers tend to be allocated to their firms (table 13-2). In fact, the few college graduates in the sample are the workers most likely to want to leave their firms; 44 percent indicate a desire to leave compared with only 8 percent for the sample as a whole. The other category of workers who wish to leave are the low-paid or those who consider their pay low in relation to the local average. The desire to leave is also related to perceptions of man- agement (table 13-3), especially in Shangrao, where 33 percent consider the management to be poor and 22 percent desire to leave their firms. TVP Workers' Incomes, Incentives, and Attitudes 289 Table 13-3. Management and Desire to Leave the Firm (number of workers) Workers' opinion Desire to leave? of mnanagement Yes No Total Very good 8 167 175 Relatively good 45 761 806 Not very good 31 108 139 Very poor 12 11 23 Total 96 1,047 1,143 Note: x2 = 103.8. Income Distribution The size distribution of annual pay is shown in table 13-4, by county and by type of firm ownership. In the analysis, ownership variables were repre- sented in three ways: on a 2-point scale (public and private), on a 4-point scale (township and town; village and production team; joint, family, and individual; and joint venture with government or foreign entities), and on an 8-point scale in which all the above segments are separate items. The 4-point scale is used in table 13-4. It is apparent that Nanhai has the widest range of pay dispersion and that incomes in Wuxi are more tightly concen- trated, with smaller lower and upper tails. Incomes in Jieshou and Shangrao cluster at the lower ends of the distributions.' The table suggests that the widest dispersion of income levels is in the firms owned by "small" (village and production team) community governments.'° This is not mainly be- cause of wider differentials within such firms but rather because differences in the fortunes of smaller communities permit a greater range of variation in pay levels among firms. The apparently concentrated income distribution of private enterprises may reflect inaccurate reporting of incomes at the upper end or noninclusion of proprietors in the sample. Workers in private enterprises seem not to be unduly concerned over wide differentials, and the proportion of workers who believe that differentials should be wider is the highest in the sample, at 26 percent. (Most workers in private firms believe that differentials are about right.) Only 8 percent of Shangrao workers favor wider differentials, despite the absence of a marked upper tail in that county. Workers' Attitudes Workers in private enterprises and, among the counties, in Jieshou seem to be content-to an almost suspicious extent. Only 11 percent of Jieshou workers agree with the proposition that TVP workers have lower status than those in state enterprises, as against 41 percent in Shangrao, 21 percent in 290 ALAN GELB Table 13-4. Dispersion of Income in Sample Firms (number of workers) Annua income (yuan) Less than 625- 1,250- 2,500- Item 625 1,249 2,499 4,999 5,000+ Total By county Wuxi 47 178 210 36 2 473 Jieshou 90 176 28 3 0 297 Nanhai 43 77 81 36 21 258 Shangrao 84 55 4 0 1 144 Total 264 486 323 75 24 1,172 By ownership Township and town 126 241 234 47 1 649 Village and production team 33 51 22 17 21 144 Private 42 86 14 3 0 145 Other 63 108 53 8 2 234 Total 264 486 323 75 24 1,172 Note: For counties, X2 = 379.2; fcTr ownership, x2 209.1 Wuxi, and 25 percent in Nanhai. Only 3 percent of Jieshou workers are not satisfied, compared with 20 percent in the whole sample, and only 3 percent would like to leave their firms. If anything, Jieshou workers see a lower proba- bility of losing their jobs than other workers. They are more inclined to buy shares in their firms; 23 percent would buy "a lot," whereas 47 percent of Shangrao workers would not buy any stocks. (Given the problematic situa- tion of many of Shangraoo's TVPS, this may reflect simple prudence.) 'ieshou workers are also far more likely to use surplus funds to buy producer goods and to set up their own firms; since there is a high proportion of private enter- prises in Jieshou, the same finding probably holds for workers in private enter- prises. There is therefore sorne evidence that private entrepreneurship is ready to flourish in areas where it is encouraged. Finally, it is interesting to note the absence of a marked generation gap in workers' responses. Substantial age-related differences in attitudinal vari- ables might be expected in a similar survey carried out in a Western country. The reason for the absence of a gap and also for the favorable labor relations climate may be that virtually all of the TVP sector's work force consists of first-generation industrial workers. Private versus Public Firms Some of the most important differences between private and community en- terprises have already been noted. Not all go in the directions one might TVP Workers' Incomes, Incentives, and Attitudes 291 expect. Although there are many statistically significant differences between the responses of workers in public and in private firms, the degree of differ- ence is usually small, so the explanatory power of the public-private distinc- tion is limited. Private enterprises are smaller and less advanced, as is shown, for example, in their narrower range of occupational groups; 72 percent of respondents in private firms, but only 58 percent in community enterprises, are workers rather than technical and managerial personnel. Private enter- prises employ a higher percentage of women and pay less than community enterprises (although this is a "county," or locational, effect; see below). They also operate in freer labor markets. Private enterprises do not diverge from community enterprises in the ways suggested by the Taylorist model, however. Pay distribution does not seem to be less equal than in community enterprises, and responses by workers in private enterprises suggest at least similar, if not greater, enthusiasm for their firms. There are no significant differences between community and private enterprises regarding the degree to which workers have decisionmaking power in the areas mentioned earlier. There is also no significant difference in average profitability between the community and private enterprises in the workers' survey. The dispersion of profitability is larger for community enterprises; table 13-5 shows that no workers were in private enterprises that were extremely profitable or declared losses. The reason may simply be that there are fewer private firms or that they underreport their profits. But it is also true that private firms, unlike community enterprises, cannot turn to local governments for help when they make losses and that the tighter technical and financial constraints placed on them limit their chances of making especially high profits. Workers in private enterprises feel somewhat less well-off in relation to their communi- ties than do those in community enterprises. Wage Equations The variables that determine the pay of TVP workers may be grouped into several classes. County variables represent the stage of development of the lo- cality; counties are represented by dummies with Wuxi as the base. Ownership Table 13-5. Distribution of Workers by Firm Ownership and Profitability (number of workers) Profitability (profits -sales) (percent) Ownership <0 <10 <20 <30 >30 Total Government 37 259 310 164 90 860 Private 0 20 62 19 0 101 Total 37 279 372 183 90 961 Note: X2 = 54.9 292 ALAN GELB variables are measured, as described above, on a 2-point, 4-point, or 8-point scale. Firm dummies, forty-eight in number, distinguish among workers in dif- ferent enterprises and are intended to pick up systematic effects not captured by the intersection of county and ownership type-for example, a large vil- lage firm in a prosperous community. They are a proxy for communal vari- ables, since most firms are identified with smaller or larger communities. The characteristics of the firm are its size (as measured by number of workers) and its profitability (the profit-sales ratio)." Individual variables are age, sex, number of days worked (from data on days per week and months per year), occupation, and education. Occupation is represented by seven (lummies-shift or group leader, operations person- nel, technical personnel, ordinary staff, middle-level staff, apprentice, and driver-with worker as the base. Education is represented by six dummies- no school, four-year primary school, six-year primary school, senior middle school, college, and technical and vocational school or vocational college- with the most common category, junior middle school, as the base. A.n expe- rience variable (age - ]L5) and its square were substituted for age in some re- gressions to permit the familiar inverted-U pattern to emerge. The dependent variable, annual income, was specified in three ways: in five ranges, in nominal value, and in logarithmic form. The results are broadly similar for all specifications. To help interpretation, results are re- ported for log income so that coefficients of unit dummies may be interpreted as percentage shifts in income. Five statistical models are shown in table 13-6.12 Model 1 uses only county dummies. The correlation coefficient (R2), adjusted for degrees of freedom, is only 0.30. Incomes are higher in Nanhai than in Wuxi by 29 percent, but they are lower in Jieshou by 55 percent and in Shangrao by 70 percent. Model 2 adds occupatLional dummies. These all have sensible signs. They suggest that more skilled staff are paid 20-45 percent more than a worker and that apprentices get 32 percent less. Although these variables are mostly significant, the explanatory power of occupation is small. Adding the simple private ownership-community ownership distinction to the county dummies makes little difference. Although workers in private en- terprises do earn less than those in community enterprises, this is mainly a county effect. Adding the 4-point ownership distinction is not much better. Model 3 combines the county dummies, the most detailed (8-point) owner- ship specification, firm size and profitability, occupation, age, sex, and num- ber of days worked. The explanatory power rises to 0.52, and pay relates posi- tively to firm size, age, days worked, and the profitability of the firm (confirming the link asserted by sample workers). An increase of 10 percent- age points in the ratio of profits to sales is associated with a 5 percent increase in the level of TVP pay. Women's pay is 14 percent lower when other variables are held constant. The sizes of the county and occupational dummy coeffi- TVP Workers' Incomes, Incentives, and Attitudes 293 Table 13-6. Wage Equations Model Dummy 1 2 3 4 5 County Jieshou -0.55 -0.47 -0.23 (11.9) (-12.4) (-4.3) Nanhai 0.29 0.14 0.22 (5.8) (3.2) (5.8) Shangrao -0.70 -0.71 -0.37 (-12.1) (-13.8) (-5.6) Ownership Dl -0.03 (-0.8) D2 0.15 (3.3) D3 0.04 (0.5) D4 0.18 (2.5) D5 0.23 (3.3) D6 -0.19 (-1.4) D7 (0.29) (2.8) Firm Work force 0.06 (2.1) Profitability 0.05 (3.5) Occupation Shift or group leader 0.19 0.08 0.11 (3.5) (1.7) (2.7) Operations personnel 0.25 0.10 0.20 (3.8) (1.7) (4-1) Technical personnel 0.33 0.16 0.18 (5.6) (3.1) (4.0) Ordinary staff 0.26 0.10 0.10 (4-1) (1.7) (2.2) Middle-level staff 0.45 0.28 0.30 (7.') (5.1) (6.3) Apprentice -0.32 -0.10 -0.26 (-2.4) (-0.9) (-2.9) Driver 0.74 0.42 0.49 (1.5) (1.0) (1.3) (Table continues on the follouing page) 294 ALAN GELB Table 13-6 (continued) Model Dummy 1 2 3 4 5 Age 0.01 0.01 (7.0) (4.8) Sex -0.14 -0.14 (-5.0) (-5.7) Days worked 0.02 0.01 (14.2) (13.0) R2 0.30 0.34 0.52 0.57 0.70 Note: Figures in parentheses are t-statistics. Blanks are not applicable. The 8-point scale de- scribed in the text is used fcr the ownership dummies; joint venture is the base. The base for occupation is worker. cients are considerably reduced, which suggests that the earlier estimates were correcting for other variables. Data on incomes in the communities of which these firms are part are not available, but it was argued above that TVPS might be considered broadly representative of their communities and that in some cases the firm itself has some status as a community. Model 4 simply consists of the forty-eight firm dummies, which alone have higher explanatory power than model 3. Model 5 adds to the firm dummnies the individual variables of model 3. By the stan- dards of cross-section regressions, explanatory power is high, with a correlation coefficient, adjusted for degrees of freedom, of 0. 70. The pattem of individ- ual coefficients is fairly stable, with somewhat higher occupational coeffi- cients than in model 3. The substitution of experience and its square for age produced the familiar inverted-U relationship whereby the rate at which income rises with age de- clines as age increases. The effect of adding the education dummy coefficients was more surprising. Most of the dummies are small and not significant, but those that are significant suggest that people with some college receive lower pay than would be expected on the basis of firm, sex, age, and occupation. This pattern is most unusual."3 These results suggest that a large part of the dispersion of earnings in the Tvp sector can be explained on the basis of communal and personal charac- teristics. But they also indicate that communal characteristics are more pow- erful, especially near the tails of the income distribution. The reason seems to be that the community is a powerful reference point for the acceptable level of pay in the firm, and so the association between pay and relative in- come status in the community is weak, as noted earlier. Consider, for example, the highest-paying firm in the sample, a large and successful ceramics factory owned by a small village in Nanhai. Workers in this factory receive about Y5,000 a year; pay of the manager (and of village TVP Workers' Incomes, Incentives, and Attitudes 295 leaders) is believed to be about Y15,000. Such pay levels are acceptable only because the factory, in effect, is the community; it employs most of the work- ing population, and it distributes 40 percent of its net income to the village for communal projects, subsidies, pensions, and education. Moreover, the survey data suggest that this firm has an unusually small dispersion of pay. It might be expected that dispersion within firms would relate systemati- cally to pay levels or would differ among counties. Measures of pay dispersion within firms, however, have no significant association with average pay in firms, and there is no significant relationship among counties. Any differ- ences in income dispersion among counties probably have more to do with the power of small as against large government units and with the degree of homogeneity of the area. Regressions were also run with a dummy that separated community insiders and outsiders. This showed only an insignifi- cant effect on pay when firm and other personal variables were included. Finally, we are left with the question of what determines the degree of dispersion of average pay among firms. The regressions indicate that county, size, and profitability make some contribution. These are static variables, however, whereas sizable pay dispersion plausibly reflects dynamic features, notably slow adjustment of the labor market in relation to different speeds of economic development. In a broad sense, such a relationship is observable in the higher incomes in Wuxi and Nanhai, which have developed more rap- idly in comparison with J ieshou and Shangrao. But does it hold on a firm-by- firm basis? Or do the highest-paying firms seem to be stagnating owing to their need to pay high wages? Investigation of this question is not straightforward because data for many firms are not sufficiently complete to allow dynamic comparisons over a num- ber of years. Therefore, to ensure maximum dispersion, firms were ranked in order of decreasing average pay. Those for which there was enough data to calculate growth rates for labor force and current-price industrial output at least for 1981-85 and preferably for 1980-85 were identified, and the high- est five and lowest five of these were chosen. In the top five firms the average growth rates were 27 percent for labor force and 77 percent for output; for the lowest five the rates were 21 percent for labor force and 44 percent for output. This supports the proposition that the higher-paying firms are more dynamic and are characterized by a higher rate of growth of output per worker than the lower-paying ones. These averages, however, are strongly influenced by two firms, one in the lowest five and one in the top five, which began from small bases and there- fore have very high average growth rates. Regressions were also carried out between the rankings of the ten firms by pay and their rankings by labor growth, output growth, and growth of output per worker. Coefficients and t-values were as follows. 296 ALAN GELB Regression Coefficient t-value P2 Pay rank with: Labor force growth 0.54 1.80 0.29 Industrial output value 0.60 2.14 0.36 Output value per worker 0.50 2.03 0.35 These results provide considerable support for the proposition that large pay differentials among firms arise from the "pull" effect of labor productivity gains (or favorable price trends) in the more dynamic firms that also have increased their labor forces somewhat more rapidly. Conclusions This chapter has set forth a number of hypotheses on the nature of ownership and labor relations in the Chinese TVP sector. Although it is only a first look at the data, it does suggest a number of interesting points. In broaLd terms TVPs appear far closer to the communal or Z-firm model than to the other models. This implies that there are few really outstanding differences be- tween community and private enterprises once account is taken of the dif- ferent levels of development of the counties where the respective types dominate. Why should this be so? The reason is probably the close relationship with local communities needed if private enterprises are to obtain permission to go into business and to secure land, bank loans, and so on, as well as the still fragile nature of private lproperty rights. Interviews confirm that private proprietors do not really act as though they had exclusive claim to their firms' equity. Some seek to recover their investments rapidly, leaving the firm in a fuzzy ownership position. Some may see advantages in then changing the firm's registration from private to community enterprise or may be powerless in the face of a decision by the local government to change its ownership. The work force seems to be relatively content, although workers have little say in running their finns. Workers perceive a strong link between pay and performance, both of the individual (piece rates are more common than time rates) and of the firm. This helps to induce a positive attitude toward the firm and some concern f'or its profitability. The link between profitability and pay has a short-run dimension, through bonuses and other benefits. But the long-run dimension may be more important because the average pay of a firm relates to its dynamism and to growth of output per worker, and where there is little intercommunity mobility workers can expect long-term employment in their firms. TVP workers therefore have substantial de facto equity stakes in their firms. This, together with the limited alternative investment options available, may explain their stated preference that profits be reinvested. An issue for policymakers is how and to what extent such an implicit equity stake should be formalized. Workers' attitudes may be influenced by the newness of the TVP sector and the fact that almost all workers were previously farmers. TVP Workers' Incomes, Incentives, and Attitudes 297 How attitudes and labor relations will evolve as the TVP sector matures is an important question. Wage equations confirm the explanatory power of firm variables as against individual variables, although the pattems of most of the latter are as ex- pected. The relationship of pay to age is progressive and of the inverted-U shape. Females eam 14 percent less than males. Technical and management personnel eam 10-30 percent more than workers, and apprentices eam 26 percent less. Educational level does not seem to be a significant factor in earnings, and when it is significant, its direction is the opposite of that ex- pected. Zero or even negative differentials at higher educational levels may be common in China, but they are unusual elsewhere. This is a subject for further research. These pattems are of course not uniform, and the survey reveals too many cross-county differences to be easily summarized here. One notable difference is the much freer labor market in Nanhai than in Wuxi, which leads to a larger proportion of workers with short time horizons, to more pressures for the distribution of surplus to the work force, and perhaps to a two-tier labor force structure. The survey results should be further analyzed to extract the implications for China of different TvP development pattems. Notes 1. In four cases responses on pay were unrealistic; annual incomes far higher than twelve times the greatest monthly income were declared, or income was stated as zero. These cases were modi- fied in line with information on the highest and lowest monthly incomes and months worked. 2. In Wuxi the examination method is based on work force allocations to enterprises by com- munity govemments; only the choice of which specific workers to take is left to the examina- tions. "Voluntary" choice of firms by workers hence does not necessarily mean that firms can choose the number of workers. 3. The explanatory variables tried include county dummies, firm dummies by eight ownership types, firm size, and occupational dummies. In comparison with Wuxi, the perceived link be- tween profits and pay is stronger in Jieshou and weaker in the other two counties. It is stronger in larger firms and greater for the other occupational categories than for workers, except in the case of apprentices. But although many of these variables are significant, the Rz (adjusted for degrees of freedom) is only 0.10. 4. Regressions suggest a significantly stronger reaction inJieshou and from supervisory person- nel. But the &2 is only 0.12. 5. TVP workers do, in fact, frequently participate in financing their firms, through purchasing bonds with a term of one to two years or through the withholding of part of their pay until year's end for use as working capital. The extent to which these contributions represent equity rather than debt (in the sense that bank debts have a preferred claim on firm assets) is not clear; it probably varies with each case and is often ill-defined. 6. A chi-squared test indicates departure from nonrandomness at the 2 percent level in the relationship between pay and relative income, but this does not seem to reflect a systematic progression with pay. Workers were also divided into two groups: those belonging to the five top-paying firms (as measured by the mean pay of the workers in the firm) and others. Those in the top-paying firms did not consider themselves much better-off, in comparison with their communities, than other workers. 298 ALAN GELB 7. In all but one of the decisionmaking areas, responses in Shangrao indicate greater worker involvement than in the other counties. In some Shangrao firms election of leaders by workers seems to have been introduced to counter the tendency of local govemment to make poor ap- pointments out of political considerations (see chapter 17). Interviews with local officials indi- cated that they considered such elected leaders to have greater legitimacy in their firms than appointed leaders. 8. Thirty percent of Nanhai workers indicate that they were in another township, county, or province before joining their present firms. The prominence of workers from outside the com- munity raises the question of whether there is a two-tier labor market, especially in the richer areas. In such a market, "outsiders" would be paid their reservation wage, and "insiders" would share in the surplus generated within the community. This question will be considered in the regressions below. 9. The coefficient of variation (ratio of standard deviation to mean) of pay is 0.49 in Wuxi and 0.70 in Nanhai. Workers in Wuxi and Nanhai usually work full time, whereas those in Jieshou and Shangrao are more likely to have worked for less than twelve months. T'his could increase the dispersion in the latter two counties at the lower end. 10. For township firms the coefficient of variation of income is 0.55, for village firms it is 0.80, and for private enterprises it is 0.55. 11. The preferred measure of profits to assets was not used because of concem about th,e quality of capital stock information. 12. Regressions are all by ordinary least squares, which is not appropriate unless explanatory variables are exogenous. This could be questioned for some variables. In particular, since pay- ments to labor affect profitability, the coefficient of this variable may be biased and is probably less significant than it should be. Occupation might also be considered endogenous, as is the choice of firm for at least some workers. 13. The reason may be interactioni effects between education and other variables, but the pattem of coefficients seems to be fairly stable. In most developing countries, education relates strongly to pay (see Knight and Sabot forthcoming). Human capital has been estimated to ac- count for half of the difference in income between high- and low-income countries (Krueger 1968). Only four of the sampled workers had no formal education. 14 The Rural Labor Market Meng Xin China is known for its large labor force. According to the theoretical model of Lewis (1954) on economic development under a dual-structure economy, an inexhaustible supply of labor is a resource advantage that can check the rise of the wage level and boost accumulation in the process of national eco- nomic development. The validity of the theory, however, depends on the precondition that a labor market already exists. For a long time China re- strained the development of the commodity economy and the formation of a labor market, which meant that there was virtually no flow of labor between urban and rural areas and between the industrial and agricultural sectors. The Lewis theory naturally cannot be applied under such circumstances. After the introduction of the production responsibility system (PRs) in ag- riculture and of price reforms, and in reaction to the separation between urban and rural areas, rural industrialization got fully under way. As a result, industrial and agricultural sectors have come into being within China's rural economy. Has a labor market emerged in the course of rural industrialization? To what extent has the market developed? And what effects does it have on rural industrialization? This chapter will describe, compare, and analyze the formation and development of the labor market in the process of the de- velopment of the TVP sector, as well as the effects of the newly formed labor market on TvPs. Reynolds (1982) defines the purely competitive labor market as follows. * There is full freedom of occupational choice. This implies, among other things, that a young person choosing an occupation that requires train- ing can always obtain the necessary funds. * There is full freedom of exchange. Any employer may hire any worker, and any worker may work for any employer. * There are many employers and many workers in the market, so that no one actor can influence the market price. * There is no collusion on either side of the market. Employers do not 299 Figure 14-1. Changes in Labor Supply and Demand for TVPs in the Four Sample Counties, 1975-85 WvUXI NANHAI Percent Percent S0- -20 80- -20 60- _<>, 40 60- /////////////////- -40 8 40- , 60 40- iE 60 20- --: 80 20 <-44000X8 1975 1977 1979 1981 1983 1985 1975 1977 1979 1981 1983 1985 m Demand for labor in agriculturea E3 Demand for labor in forestry, animal husbandry, fishery, and other activitiesb l Demand for labor by TVsP El Rural surplus labord JIESHOU SHANGRAO Percent Percent 60- /40 60 40 40- _ 60 40_ -60 20- -80 20- _ -80 0- f W 111111111111111lllllllll-100 0_ Io, 1975 1978 1981 1983 1985 1975 1978 1981 1983 1985 Note: For Wuxi, labor in other activities was estimated at 5.2 percent of the total rural labor force (the national average in 1985). For Nanhai, labor in forestry, animal husbandry, and fishery was estimated at 7 percent of the total rural labor force and labor in other activities at 5.2 percent. For Jieshou, labor in forestry, animal husbandry, and fishery was estimated at 5.0 percent of the total rural labor force and labor in other activities at 5.2 percent. For Shangrao, labor in forestry, animal husbandry, and fishery was estimated at 5.0 percent of the total rural labor force and labor in other activities at 5.2 percent. a. See Appendix for method of calculation. b. Actual use of labor. Other activities include science, education, culture, public health, and government. c. Actual employment. d. Difference between total rural labor force and the three sources of labor demand. Sources: Appendix and information from fieldwork. 302 MENG XIN unite to drive down the wage level, nor do workers unionize to drive up wages. This definition is taken as the "perfect state" of the labor market as we examine the development of labor markets in the four counties. This means not that the purely corapetitive labor market is an ideal state but rather that it is the standard for analysis. Most TVP workers come from rural areas. China's agricultural sector has large potential surplus labor, which determines the supply of labor to TVPS. The actual number of workers absorbed by TVPS indicates the demand in the labor market. Thus the labor market is viewed as being in varying degrees of disequilibrium, with the short side of the market (labor demand) determin- ing the actual amount of labor in the TVP sector. Figure 14-1 shows changes in the labor supply and demand in the TVP sec- tors of the four counties in 1975-85 under the assumption that there is local segmentation in labor markets (see the appendix to this chapter for details on the method of calculationi of labor supply to TVPS). From the figure it is clear that in Nanhai and Wuxi labor supply can hardly meet the demand of TVPS, whereas in ShangraD and Jieshou labor supply still greatly exceeds the demand. This should always be kept in mind during the following dis- cussion. Labor Flows among TVPs There are three types of rural labor flows: interregional, among provinces and counties; intercommunity, among villages and townships in the same county; and intracommunity, among firms in the same community. Actual Labor Flows Actual labor flows can be examined from three points of view: the origins of an enterprise's employees and the relative shares of permanent, casual, and seasonal laborers; a firm's ways of employing and dismissing laborers; and re- dundancy of personnel in an enterprise. Table 14-1 shows the family homes of employees in sample enterprises in the four counties and table 14-2 the proportions of permanent, casual, and seasonal laborers. There is a big differ- ence between sample firms in Nanhai and in the other three counties. In Nanhai 51 percent of employees come from local communities, 85 percent from within the county, and 15 percent from other counties and provinces. In the other three counties the bulk of employees come from local communi- ties, more than 90 percent come from within the county, and only a handful are from other counties or provinces. These figures indicate that at present the bulk of labor in China's 1Vps comes from farmers in local communities. Except in Nanhai, intercommunity and interregional flows of labor are weak. The Rural Labor Market 303 Table 14-1. Original Homes of Employees in Sample Enterprises (percent) Same Other townships Other County and township (villages) counties Other type of worker (viUage) in county in province provinces Wuxi All employees 94.0 3.0 1.6 1.4 Workers 94.2 3.5 1.7 0.7 Technicians 78.6 1.4 5.3 14.6 Managerial staff 93.1 0.6 0.5 5.9 Jieshou All employees 79.7 12.7 4.8 2.8 Workers 80.6 12.1 5.2 2.7 Technicians 65.9 18.3 2.4 13.4 Managerial staff 78.1 16.5 2.3 3.2 Nanhai All employees 50.9 33.8 14.2 1.1 Workers 47.7 35.0 16.0 1.3 Technicians 85.1 3.0 10.9 1.0 Managerial staff 53.5 39.7 5.6 1.2 Shangrao All employees 80.9 14.8 2.1 2.2 Workers 83.6 11.4 1.9 2.7 Technicians 68.5 5.6 5.6 20.3 Managerial staff 73.9 18.5 1.7 5.9 Source: Enterprise Quantitative Questionnaire. The proportions of permanent and casual laborers can shed further light on the volume of labor flows. These figures also show much greater mobility in Nanhai than in the other three counties. The data below, however, show that great shifts have taken place in recent years in the relative shares of permanent and casual laborers in Nanhai, indicating that the employment system in that county has undergone remarkable changes. 1980 1981 1982 1983 1984 1985 Percentage of Permanent workers 91.8 87.9 86.3 76.1 48.3 40.2 Casual workers 8.2 12.1 13.7 23.9 51.2 59.8 The methods of recruiting and dismissing workers in sample firms are illu- minated by responses to the Enterprise Survey Questionnaire and the Worker Survey Questionnaire. The latter asked workers how they came into their present employment (see table 14-3). Job assignment by township or village government was cited by 57 percent of sample employees in Wuxi, 44 percent in Shangrao, 38 percent in Nanhai, and 16 percent in Jieshou. Responses to the Enterprise Survey Questionnaire point to a similar conclusion: com- 304 MENG XIN Table 14-2. Permanent, Casual, and Seasonal Workers in Sample Enterprlies (percent) F7rnanent Casual Seasonal County workers workers workers Wuxi 91.2 8.5 0.3 Jieshou 75.8 18.1 6.2 Nanhai 40.2 59.8 0.0 Shangrao 78.7 19.1 2.2 Source: Enterprise Quantitative Questionnaire. munity government assignment constitutes the most important fonn of re- cruitment in Wuxi, whereas enterprises in the other three counities rely mainly on direct recruitment. The Enterprise Director Questionnaire contains two relevant questions: "As the director of the factory, do you have the right to dismiss your employ- ees?" and "In the past year, how many workers have you fired? For what rea- sons?" Enterprises whose directors have the power to dismiss employees ac- counted for 97 percent: of the sample in Jieshou, 82 percent in Wuxi, 73 percent in Nanhai, and 68 percent in Shangrao. The average number of workers dismissed by each sample firm in the previous year stood at 0.87 in Nanhai, 0.75 in Jieshou, 0. 67 in Wuxi, and 0.15 in Shangrao. Although nearly all directors in Wuxi thought that they had the power to dismiss work- ers, the actual number of workers fired was relatively small. Some community governments in Wuxi have issued regulations that ban enterprises fiom dis- missing employees until they find new jobs and require that each d[ismissal be approved by the communit-y government. The primary reason for dismissal given in all four counties was violation of labor regulations. Overall, sample enterprises in Jieshou and Nanhai seem to have relatively more freedom and those in Wuxi and Shangrao less freedom in employing and dismissing workers. The presence of redundant personnel means that the enterprise has little decisionmaking power over its employment level. In addition, it is evidence from another angle of the absence of a free flow of labor. During fieldwork we saw obvious examples of redundancy in Wuxi and Shangrao. The director of the Xuelang Township Special-Shaped Steel Rolling Mill in Wuxi told us that in 1983 the mill set up a meter production workshop and employed seventy-eight workers (mostly women) under a land requisition contract. Later the workshop was closed because its products did not sell very well on the market, but the mill had to keep the workers who had been recruited for the workshop, since their employment was included in the land contract. The director told us that his mill could sack at least fifty workers, or 10 per- cent of the total, without affecting normal production. The director of the Luqu Township Food Products Factory, also in Wuxi, said that his factory The Rural Labor Market 305 Table 14-3. How Employees Got Their Jobs in Sample Enterprises (percent) Method Wuxi Jieshou Nanhai Shangrao Assigned by township or village government 56.8 15.5 38.0 44.4 Examination 15.8 30.7 3.9 2.8 Application 9.2 12.5 15.7 20.8 Contribution of funds 0.9 12.8 1.2 9.7 Recommendation 16.0 28.0 41.2 22.2 Other 0.6 0.3 0.0 0.0 Source: Worker Survey Questionnaire. could cut employment by fifty to sixty workers (25-30 percent of the total) without affecting total value of output, but township government regulations barred it from dismissing any employees. The director of the Dongjiang Township Oil Production Equipment Factory in Wuxi told us that in 1985 the number of employees in his factory doubled from the previous year, not to meet production needs but because of pressures from the township govern- ment to increase employment and obligations resulting from land acquisition contracts. The Lingxi Township Hydropower Station in Shangrao reported that its redundant personnel accounted for about 60 percent of total employ- ment. And the director of the Lingxi Township Fiberglass Plant believed that about half of the plant's employees were redundant. Personnel redundancy in Wuxi and Shangrao is related to township governments' interference in worker recruitment and dismissal. The problem was almost unheard of in Nanhai and Jieshou. The above discussion has given us a picture of labor flows at three different levels in the four counties. The general impression is that whether labor is in short supply (as in Wuxi and Nanhai) or in surplus (as in Shangrao and Jieshou), interregional flows are weak but that the situation in Nanhai is bet- ter than in the other counties. Nanhai has much greater intercommunity and intracommunity labor flows than does Wuxi, and Jieshou has more than Shangrao. Factors That Affect Labor Flows In a commodity economy the supply and demand situation for labor strongly affects labor flows. In general, when the supply of labor is greater than the demand, the mobility of the labor force is small, but the contrary is true when demand is greater than supply. When labor is plentiful, the availability of cheap local labor makes it unnecessary for enterprises to recruit workers from other regions, and low wages and limited employment opportunities discour- age laborers in other regions from coming to seek jobs. But when labor is 306 MENG XIN in short supply, enterprises vie for workers by paying higher wages, and this attracts labor from elsewhere as well. It should be stressed that when there is a labor surplus, the diminished actual flow of labor does not mean a reduc- tion in workers' freedom to choose their professions or in enterprises' freedom to choose laborers. An analysis of labor flows in the four counties in accordance with the above-mentioned logic revealed various distortions. For instance, in Wuxi, where labor is in short supply, labor flows at the three levels are not only much smaller than in Nanhai, where the supply and demand situation is simi- lar, but also smaller than in the other two counties, which have large num- bers of surplus laborers. In falct, there was little tangible labor flow in Wuxi. Also, although the makeup of the places of origin of employees was basically the same in Jieshou and Shangrao, workers and firms in Shangrao had much less freedom in choosing each other than in Jieshou. How could such irra- tional situations arise? Supply and demand are clearly not the only factors that affect the flow of labor. To provide an explanation, we must introduce systemic variables. Lower labor mobility in WVuxi than in Nanhai is directly related to differ- ences in the ownership structure of TVPS. In Wuxi TvPs are dominated by TVCES, whereas in Nanhai 'rvCEs account for only 30 percent of the total. Moreover, since TVCES employ only about 60 percent of the total number of workers in Nanhai's TVP sector and 50 percent of these are casual workers, only 30 percent of the labor force is directly controlled by community govem- ments. As community governments are inclined to restrain the flow of labor, the difference in ownership structure leads to differences in labor mobility. We will now discuss why community governments would want to restrain the inflow and outflow of labor. RESTRAINING INFLOWS OF LABOR FROM OTHER AREAS. Since a community government is both the owner of local TVCEs and the community administra- tor, its goals are not only to develop its firms but also to increase local em- ployment, local residents' income levels, and community revenues. The de- velopment of TVCES is expected to serve these multiple social and economic goals, and the fundamental reason that communities restrain inflows of labor is that these inflows clash with their employment and income objectives. At different stages of economic development the community government has different purposes in promoting TVPS. When there is a large labor surplus in the community, the goal is to expand community employment to its maxi- mum; when all local surplus labor has been absorbed, the purpose becomes mainly to increase per capita incomes. (The revenue objective is always pres- ent.) Under both circumstances the community government tries to restrain the inflow of labor from other regions: in the former case labor inflows would reduce employment opportunities for local community members, and in the latter they would reduce the community's average per capita income. The Rural Labor Market 307 RESTRAINING OUTFLOWS AND INTRACOMMUNITY FLOWS OF LABOR. Com- munity governments in both Wuxi and Nanhai have issued regulations to restrain outflows and intracommunity flows of labor. For example, the flow of labor between enterprises must be approved by the township government; a fine of Y500-Y2,000 is imposed on anyone who goes to seek a job outside the community; and those who leave the community to find jobs outside face noneconomic punitive measures, such as denial of family members' right to a job assignment from the community government and higher tuitions for schoolchildren. Where there are labor shortages, these restraints help ensure the stable development of local firms. The forces that restrict labor flows are basically the same in both Wuxi and Nanhai, but in Wuxi more than 90 percent of TVPs and their employees are controlled by community governments, whereas the figure for Nanhai is only about 30 percent. Thus in Nanhai community governments are unable to control the inflow of labor from other regions, and they can hardly stop the intracommunity flow because gaps among enterprises' wage levels are large and the benefits from taking another job can easily offset losses from punitive measures, such as fines, imposed by community governments. The greater freedom of choice for both workers and firms in Jieshou than in Shangrao is also related to the difference in the two counties' ownership structures. TvcEs account for 64 percent of total sample enterprises in Jieshou and for 97 percent in Shangrao. As the TVCES' owner, the community gov- ernment has indisputable control over their labor forces, so where there are more TVCES, laborers have less freedom to choose their professions and firms have less freedom to choose their employees. Of course, the mobility of labor is not determined only by the supply and demand situation and the structure of ownership. To a certain extent it is also related to the dissemination of information and to social and cultural traditions. Nanhai's local labor shortage and high wage levels, among other factors, make it highly attractive to laborers from other rural areas, and communities in Nanhai are unable to check the inflow of labor from other regions. Never- theless, workers from other provinces and counties account for only 15.3 per- cent of total sample employees. The only reasonable explanations for this phenomenon are that many farmers in backward areas cannot get informa- tion about the labor market, that farmers are still under the traditional psy- chological influence of strong links to their hometown and land, and that the current household registration system hampers interregional flows of labor. Wage Determination for TVP Employees This section describes wage levels and differences in the four counties, the effect on wages of labor supply and demand in general and of supply and 308 MENG XIN demand conditions for labor of different qualities, the effect oif systemic variables on wages in Wuxi, and the reasons behind wage increases in Nanhai. Wage Levels and Differences Figure 14-2 shows average wage incomes in sample enterprises in the four counties during 1975-85. 'There is an evident gap between wage levels in Nanhai and Wuxi and in Shangrao and Jieshou, and there is a tendency for the gap to become wider. In 1985 the average monthly income in sample firms in Nanhai reached Y182, 73 percent higher than Wuxi's figure of Y105, which in tum was 67 percent more than the average in Jieshou, Y63. The average for Shangrao was Y59, about the same as for Jieshou. During 1980- Figure 14-2. Average Per Capita Monthly Incomes in Sample Firms, 1975-85 Yuan 200 180 - 160 - Nanha 140 - 120 - 100 -Wux 80 - 60 .-< _,_ _- 40 - ___--_-- Jieshou 20 - TvPs, all China 0 I I I I I I I 1975 1980 1985 Sources: Information from fieldwork and statistical information on China's rvps. The Rural Labor Market 309 85 the annual growth rate of average income in sample firms was 21 percent in Nanhai, 17 percent in Wuxi, 8 percent in Jieshou, and 6 percent in Shangrao. Although wages in areas suffering from labor shortages are much higher than in areas with labor surpluses, absolute wages and the growth rate of wages are far higher in Nanhai than in Wuxi despite similarities in the two counties' labor supply and demand and the greater mobility of labor in Nan- hai. We explain this phenomenon in a later section. WAGE DIFFERENCES AMONG COMMUNITIES. Since the samples in each county are too limited to explain differences among townships, we base our analysis on statistics on TVPS provided by each county.' The data indicate that in 1985 the variance of average per capita pay (as a ratio to the mean) was 0.03 in Wuxi, 0.14 in Nanhai, 1.61 in Shangrao, and 0.005 inJieshou. WAGE DIFFERENCES AMONG ENTERPRISES WITHIN COMMUNITIES. We se- lected for analysis TVCES in Wuxi's Dongjiang township, various types of en- terprises in Jieshou town, and township enterprises in Nanhai's Yanbu dis- trict.2 The variance of average pay in different enterprises was 0.12 in Dongjiang township, 0.12 in Jieshou town, and 0.27 in Yanbu district. Thus wage differences both among and within communities are large in Nanhai; wage differences at the intercommunity level are rather large in Shangrao; and wage differences at both levels are small in Wuxi and Jieshou. The Effect of Labor Supply and Demand In a competitive labor market the wage level is determined by supply and demand, and the price of labor equals the marginal product of labor. Accord- ing to Lewis, when there is an inexhaustible supply of labor, the wage level of workers in the industrial sector should be a little higher than that of rural laborers so that the industrial sector can attract labor from the agricultural sector, but the differential does not need to be large. When labor is in short supply, the industrial sector has to pay much higher wages than the agricul- tural sector to gain the upper hand in competition for labor. The history of economic development in developed countries (and particularly in the United States) has demonstrated that with abundant funds and preferential conditions the industrial sector can constantly attract labor from agriculture. Therefore the relationship between the wages of industrial workers and the income of rural laborers can shed some light on the effect of labor supply and demand on wage determination. Figures 14-3 through 14-6 show average monthly wage income in sample enterprises, in agriculture and rural industry, and in state enterprises in each county. These indicate that when there are plenty of surplus laborers the in- come of workers in sample firms remains at a higher level than that of local 310 MENG XIN Figure 14-3. Average Monthly Incomes in Agriculture, Industry, and State Eitterprises, Wuxi Yuan 110 100 / 90 _ 80 _ 70 -S 60- 50 -State enterprises - . 40 30 Industry /' 20 Agriculture 10 0) I I I Il l I I I Il I l 1970 1975 1978 1981 1983 1985 Source: Information from fieldwork. farmers, as in Jieshou and Shangrao. But in Jieshou the income curves for sample firms and for local falrmers climb smoothly the same distance apart, whereas in Shangrao the two curves fluctuate irregularly. The correlation be- tween the two curves :is higher in Jieshou (0.99) than in Shangrao (0.92), which suggests that the impact of labor supply and demand on pay is greater in the former than in the latter. When labor is in relatively short supply, the correlation between the aver- age monthly income of workers in sample enterprises and that of local rural laborers is extremely high, as is the case in Wuxi (1.00) and NanhaiL (0.99). But the income of employees in sample firms in Nanhai remained higlher than that of local farmers during t-he survey period, whereas in Wuxi the income of local farmers caught up with and even surpassed that of employees in sam- ple firms in 1984-85. ODbviously the situation in Wuxi was irrational, since labor was in short supply. Therefore we may conclude that wage levels in Wuxi have little to do with local labor supply and demand. Responses to the question "In your enterprise, who decides the pay of a worker who has just joined?" in the Enterprise Survey Questionnaire support the above conclusion. 'The township government" was the answer in 42 per- The Rural Labor Market 311 Figure 14-4. Average Monthly Incomes in Agriculture, Industry, Sample TVPs, and State Enterprises, Nanhai Yuan 180 - Sample 160 TVPS 140 120 - 100 -Idsr 80 60 State enterprises / 60- 40 ..s~~~ Agriculture 20 -20 _ 1975 1978 1980 1982 1984 1985 Sources: Enterprise Quantitative Questionnaire and information from fieldwork. cent of the responses in Shangrao, 31 percent in Wuxi, 21 percent in Nan- hai, and 17 percent in Jieshou (table 14-4). From interviews we learned that community governments in Jieshou and Nanhai rarely interfere in their firms' decisions on employees' pay. Under such circumstances the high correlation between the average income of TVP employees and that of local farmers naturally indicates the influence of labor supply and demand on wages. But in Shangrao community governments not only interfere in enterprises' decisions on setting wages but have also adopted the principle that the salary level in TVPS should be a little lower than that in state enterprises rather than being set according to labor supply and de- mand. (See the discussion below.) Supply of and Demand for Labor of Different Qualities According to the theoretical model of a totally competitive labor market, if the quality of labor is disregarded and labor moves freely, labor should have 312 MENG XIN Figure 14-5. Average Montbly Incomes in Agriculture, Industry, and State Enterprises, fiesbou Yuan 80 70 - State enterprises / 60 1 _ -o 50 _/ Industry 40 - 30 - 20 - 10 1978 1980 1982 1984 1985 Source: Information from fieldwork. a unitary price in labor markets at the same level (interregional, intercommu- nity, or interenterprise). It seems that the above analysis of the data and the conclusions we have reached contradict the theory. For instance, in Nanhai, which has the greatest mobility of labor at the intercommunity and interenterprise levels, wage differences are large at both levels. But the differ- ences are very small at both levels in Wuxi, where almost no flow of labor can be found. Realities are much richer than theories, and many things that can be neglected in theoretical analysis are of great significance in practice. Here the crucial factor is the quality of labor. The industrial composition of Nanhai's TVPs features strong local speciali- zation. For example, Xicliao tcwn is the county's textile center; hardware and machine-building enterprises are concentrated in Pingzhou town; Danzao is the producer of pottery; HeshiLn has a relatively advanced building materials industry; and labor-intensive industries such as garments, toys, and electron- ics are mostly located in relatively backward areas like Guanyao. Local spe- cialization in industrial structure is a principal factor in the big differences in salaries among communities. The Rural Labor Market 313 Figure 14-6. Average Monthly Incomes in Agriculture and Indust7y, Sbangrao Yuan 80 70 - 60 - ~~~~~~~~~Industry 60_ __ 50- - -- -- -- -- -- - - - - Agrculur 40 , 30- 20- 10 O I I I I 1978 1980 1982 1984 1985 Source: Information from fieldwork. Industries at differing levels of technical development have different re- quirements for labor quality, and hence the price of labor varies. Since learn- ing complicated techniques is more costly and the supply of people who com- mand such techniques is scanty, the price of highly trained workers will naturally be higher than that of workers who know only simple techniques. In addition, some techniques can be applied universally, whereas specialized techniques can be applied only in limited areas. Workers who have special- ized techniques take a bigger risk in looking for jobs than those who know commonly used techniques. The higher price that specialized laborers enjoy is actually a kind of compensation for the employment risks they face. The big differences in pay among enterprises within Nanhai's communities are caused mainly by differences among industries. The survey results show big wage differences among enterprises within communities. This is because the survey results do not take into account enterprise scales and do not in- clude firms other than TVCES. For instance, in Yanbu district wages in electri- cal appliance, hardware, and toy factories were evidently too low (an average 314 MENG XIN Table 14-4. Who Decides the Salary for Newly Recruited Employees in Sample Enterprises (percent) Vilage or Government and Enterprise, township enterprise but approved Enterprise County government jointly by government itself Other Wuxi 31.2 9.4 34.4 25.0 0.0 Jieshou 16.7 8.3 19.4 52.8 2.8 Nanhai 20.8 4.2 20.8 33.3 20.8 Shangrao 42.1 10.5 15.8 31.6 0.0 Source: Enterprise Survey Questionnaire. Y1,223 a year in the six sample factories), whereas in machine-building and food-processing factories they were too high-Y2,222 a year in the eight sam- ple factories. The same factors also explain the small wage differentials in Jieshou. In- dustrial structure in the county is less diversified, and TVPs are involved mainly in agroprocessing, production of construction materials, brickmaking, and food processing. As a result, all communities and enterprises in Jieshou need workers of similar quality. Since wage levels are determined mainly by labor supply and demand, wage differentials are small. The wage level in Wuxi is not determined by labor supply and demand, so the small wage differ- ences there cannot be explained in the same way as in Jieshou. Finally, the main reasons for big wage differences among Shangrao's communities are lack of mobility of labor, backward processing industries, and uneven distribution of natural resources such as minerals and water. The Impact of Systemic Variables on Wages in Wuxi County The wage level in Wuxi is lower than in Nanhai, the income of Tvp emnploy- ees has become lower than that of farmers in recent years, and wage differ- ences at the intercommunity and interenterprise levels are small. These char- acteristics cannot be explained by labor supply and demand. In Wuxi the wage determination mechanism is totally different from that of the market. TVCEs account for the overwhelming majority of all TVPS in the county, and all of their assets belong to the communities. In principle this means that the asseits belong to every member of the community, but in reality they do not belong to any individual; instead, they are owned by the community government and are held and used by the enterprises. Each TVCE employee has the ambiguous status of an enterprise owner. This has made TVCEs a special kind of cooperative without clearly defined property rights. In such firms the only way benefits can be shared fairly is thLrough income distribution. Therefore within the enterprise there is a strong mnotive to maximize average per capita income. But since enterprise assets are actu- The Rural Labor Market 315 ally owned by the community govemment, the management of the enter- prises is not allowed to depart too far from community goals. Thus firms' goal of maximizing employees' average income is partially offset by the community goal of maximizing (and balancing) the average income of community mem- bers. This is one of the main reasons that TVP wages in Wuxi are relatively low even though the county is suffering from a labor shortage. During our survey we leamed that when in 1984 the county introduced a contract system and lifted wage ceilings for enterprises, the average wage soared by 28.8 per- cent. In 1985, after community govemrnments restored controls, annual wage growth was reduced to 10.5 percent. The community's desire to equalize average incomes within the commu- nity can be traced to its ownership of enterprise assets. Every community member is supposed to enjoy the same status as owner of these assets as well as equal rights to employment and income distribution. This requires commu- nity govemments to adopt an egalitarian policy in distribution of income be- tween industrial and agricultural laborers. Before 1983 most Tvps in Wuxi handed wages over to workers' production teams and the workers shared in the distribution of production team income. After the introduction of the PRS, community govemments began to use profits remitted by enterprises to subsidize farmers. In recent years, with the emergence of a labor shortage, fewer and fewer people are willing to engage in painstaking agricultural pro- duction. To stop the dwindling of the local agricultural sector, community govemments have been using profits from industrial TVCES to supplement the incomes of agricultural laborers so that their incomes can catch up with or even surpass those of TVP workers. The relatively small wage differences among enterprises in Wuxi's com- munities also reflect community govemments' policy of egalitarian income distribution among community members. Community govemments have adopted flexible policies for enterprises in different situations. For instance, firms that eam fat profits may be required to hold down workers' pay, whereas those in the red may try to raise employees' incomes by delinking the payroll from profits and linking it only to output value. Moreover, keeping wage dif- ferentials among enterprises low is a way of restraining labor flows and main- taining stable development of enterprises. The small wage differences among Wuxi's communities can also be attributed to this consideration. As labor is in short supply, to keep workers it is necessary for each community to pre- vent its wages from sliding much lower than those in other communities. What should be stressed here is that egalitarian income distribution among community members in Wuxi has not had too adverse an impact on the eco- nomic efficiency of TVPS. This may be attributed to the employment, wage, and profit-sharing systems. The fundamental difference between TVP employ- ees and state enterprise employees is that the former face less certainty in their employment. State enterprise employees can never lose their jobs and the related salary, welfare, and benefits, but a TVP employee will lose his job 316 MENG XIN and will have to go back to the fields if the firm is closed. Thus the basic interest of TVP employees is tied to that of their firms. Although community governments try to reduce income gaps, such gaps persist to a certain extent. TVCES in Wuxi have generally adopted the piece- rate wage system, which has clarified the profits made by each individual within an enterprise andl has offset the ambiguity in the relationship between profits of different enterprises. These tangible incentives have stimulated workers' enthusiasm. The practice of linking part of the payroll to enterprise profits also gives employees a limited share in profits and some incentive. Understanding Wage Increases in Nanhai County In the previous section we analyzed the great difference in wage levels be- tween Wuxi and Nanhai counties from the perspective of Wuxi, where com- munity goals restrain enterprises' motives to increase average income. But does this mean that the present high wage level and high wage growth in Nanhai are related to the cooperative nature of its enterprises and community govemments' inability to restrain them? It should first be noted that TVCEs behave similarly in Nanhai and in Wuxi and that community g6vemments in Nanhai also like to control the ever rising level of wages in their firms. But in Nanhai individually run enterprises and partnerships account for a large proportion of the total number of TVPS, and these firms usually pay their employees higher wages than TVCES. This not only compensates for the greater uncertainty in employment but also sharpens competitiveness in the labor market. In addition, agricultural pro- duction in Nanhai has reached a high level of development. In recent years more and more Nanhai farmers have turned to raising flowers, vegetables, and farm animals, and the county is enjoying a high commercialization rate and sharp growth in incomes. Agriculture there can no longer be regarded as a traditional sector; it has already become part of the commodity economy with profit-making as its management goal. Farmers in Nanhai have much higher incomes than those in Wuxi. In 1985 a person whose main occupation was agriculture, forestry, animal husbandry, or fishing had an average income of Y1,132 in Wuxi but Y1,786-58 percent higher-in Nanhai. With labor in short supply, community governments and TVCES in Nanhai are under pressure from both the agricultural sector and private firms. If they cannot keep their workers by offering high wages, the normal development of TVCES will be threatened. UJnder such circumstances community govern- ments in Nanhai allow TVCES to set wages according to labor supply and de- mand. In the final analysis thLe high level of wages in Nanhai is a normal reflection of market regulation when labor is in short supply. Nanhai data on wages are for 1985, however, whereas the data on labor flows reflect the situatiorL in March 1987. The great influx of outside laborers into Nanhai started only at the end of 1985 and the beginning of 1986. We The Rural Labor Market 317 should not assume that the price of labor continued to rise sharply after that. An article in Nanhai Nongcun [Rural Nanhail of February 1987 points out that in 1986 the number of rural laborers in the county engaged in farm pro- duction increased by 3.0 percent over the previous year, while the number engaged in industrial production dropped by 4.9 percent. The main reason was that some TVCES had been contracted out to individuals, and the contrac- tors had gradually replaced local employees with lower-paid workers from other regions. As a result, some local people retumed to farming. Thus the inflow of laborers from other regions has to some extent held down wages of local workers. Labor Market Development and Its Influence on TVPs The description and analysis of labor mobility and wage determination in the four counties have shown the following. * Although development of TVPS is highly advanced in both Wuxi and Nanhai, the two counties show sharply contrasting pattems. Labor flows at the interregional, intercommunity, and interenterprise levels are very low in Wuxi, and wages are not determined by the labor market but by community govemments that strive to equalize local per capita incomes. Enterprises have little freedom to choose workers, and work- ers basically have no choice of profession. In Wuxi the labor market, as defined earlier in this chapter, does not exist. - Among the four counties Nanhai has the greatest mobility of labor. Wages are determined mainly by labor supply and demand. Although community govemments have various regulations that restrict the free- dom of long-term TVCE workers (20-30 percent of the county's total TVP labor force) in choosing jobs, plentiful opportunities for employ- ment and the great attraction of big salary differentials have countered community govemments' efforts to control this 20-30 percent of the labor force. The labor market in Nanhai is the most developed among the four counties. • If we regard Wuxi and Nanhai as two extremes of labor market devel- opment in the more industrialized areas, Jieshou and Shangrao can be seen as two extremes in the backward areas. Although mobility of labor is only slightly better in Jieshou than in Shangrao, workers and firms in Jieshou have much more freedom in choosing each other than in Shangrao. In areas that have substantial surplus labor, freedom of choice reflects the degree of development of the labor matket more than does labor mobility. TVPS in Shangrao suffer from great interference by community governments in wage determination. Community govem- ments there try to make TVPS imitate the fixed salary systems, retire- ment systems, welfare regulations, and even salary adjustment mecha- 318 MENG XIN nisms of state enterprises. In a word, wage determination has nothing to do with supply and demand in the local labor market. In general, community governments in Jieshou do not interfere in enterprises' deci- sions on wages--even those of TVCES. Hence the small differences in wages at the intercommunity and interenterprise levels indicate that the market price of labor has already been formed. The earlier analysis has also shown that at present China does not have a unitary labor market between urban and rural areas and that interregional rural labor markets remain at a low level of development. Despite the inex- haustible supply of labor nationwide, labor costs in some rural areas have sky- rocketed and even surpassed the salary level in local state enterprises (see figures 14-3 and 14-4) Obviously the low level of development of interre- gional labor markets is harmful to the development of China's TVPS. Al- though most rural areas still have a labor surplus, some advanced localities have already seen a sharp rise in labor costs because of the shortage of labor. The advantages that an inexhaustible supply of labor hold for development are gradually disappearing in some rural areas of China. Because of the restricted interregional flow of labor, some areas have al- ready resorted to changing output mix, updating technology, and adjusting other factors of production. Industries and enterprises with different levels of technology usually have different capital-labor ratios. A firm's production scope and technology determine the level of labor costs that it can afford. If labor costs keep rising, the enterprise will have to choose a new capital- labor ratio or switch its production scope and technology level. In developing countries with relatively well-functioning labor markets, the flow of labor from underdeveloped to advanced areas will curb the rise of labor costs in the latter. As a result, enterprises in such areas can enjoy a long period of stable development after they have chosen the industry they will engage in and the level of technology they will pursue. But under present conditions in China's rural areas, TvPs in the advanced areas have to switch to other industries and technologies to counteract the rapid rise in labor costs caused by immobility of labor. We saw many instarLces of this. In Yanbu town, Nanhai, a factory that produced mosquito repellent incense has switched to producing liquefied gas cylinders, and a furniture factory is now producing aluminum materials. In Pingzhou town a knitting mill had to suspend production because of high labor costs, and its machines now sit idle. A plastic products factory in Shatou town has transferred its technology, equipment, and capital to Sichuan Province because of t:he high cost of labor in Shatou. These frequent adjustments will inevitably lead to some waste of investment and will have an adverse impact on enterprises' stable development. Similar cases can be seen in Wuxi. Qianzhou township, for instance, once signed a contract with the Shanghai Foreign 71rade Bureau to produce and export plush toys, but The Rural Labor Market 319 because of high local labor costs, the contract yielded no profits to the town- ship. The township therefore later transferred the contract to a TVP in Taixing County in northern Jiangsu Province, where labor costs were only 50 percent of those in Qianzhou. Flows of any factors of production inevitably call for some investments and risk taking. By comparison, the flow of labor needs less investment, and the risk involved is shared. The transport, dismantling, and installation of other factors of production can be much costlier than flows of labor, and the enter- prise alone shoulders all the risk. Therefore flows of other factors are often less economic than labor flows. There is no doubt, however, that changes in product structure, improved technologies, and flows of other factors can partly compensate for the immobility of labor. Because of the underdevelopment of interregional labor markets, local labor markets have different impacts on TvPs at different stages of develop- ment. In the initial stage of development of Tvps and when there is an inex- haustible supply of labor, the existence of a labor market helps enterprises choose a suitable amount of labor in accordance with the needs of their own development, and it can also help curb the rise in wages. But community interference in the allocation of labor and in wage determination can cause personnel redundancy in enterprises and lead wages to surpass the appropriate level based on supply and demand. This is because communities strive to ex- pand local employment and increase local per capita incomes, and they re- gard the enterprises as a means of achieving their goals. When the development of local TVPs reaches a relatively advanced stage and labor is in short supply, the existence of a local labor market facilitates rapid wage increases and has an adverse impact on the development of TVPS. Before 1985, when the interregional labor market had not yet opened up, wages in Nanhai soared with the growing shortage of labor. If community governments can interfere in setting wages, they will limit wage growth to further general community goals. For example, in 1985 community govem- ment controls reduced the growth rate of wages in Wuxi from 28.4 to 9.8 percent. (Such controls can slow the speed of wage increases to a certain extent but they do not curb the growth of total payrolls, since one of the principal goals of community governments is to increase local per capita in- comes.) Furthermore, if there is no competition among workers for employ- ment, wages will be highly inflexible. The solution to the problems of wage inflation and inflexibility is the open- ing up of interregional rural labor markets. In 1986 Nanhai was able to coax local laborers to return to farming by bringing in cheap industrial labor from other regions. On the premise of an inexhaustible supply of labor in China, the opening up of rural labor markets will certainly create a more favorable climate for the development of TVPS. The analysis in this chapter has shown that labor mobility is related to the ownership and management systems of TVPS. Hence, system reform is a leading issue in the development of TVPS. Table 14-Al. Cultivated Area and Labor Requirements for Crop Cultivation in the Sample Counties Item 1975 1978 1980 1981 1982 1983 1984 1985 Crop area (mu) Wuxi 919,100 915,800 912,900 912,600 908,400 908,400 900,000 877,800 Jieshou 648,000 644,700 642,300 642,300 642,200 642,200 641,900 641,800 w Nanhai 680,600 654,900 655,700 653,700 649,700 645,000 641,400 630,100 CD Shangrao n.a. n.a. 471,200 471,000 471,100 471,400 470,700 n.a. Labor requirements for crop cultivanon (workers) Wuxi 116,955 116,662 116,293 116,255 116,140 115,720 115,490 111,822 Jieshou 86,397 85,953 85,646 85,628 85,624 85,632 85,585 85,571 Nanhai 106,339 102,328 102,455 102,141 101,513 100,785 100,218 98,596 Shangrao n. a. n. a. n. a. n. a. n. a. n.a. 73,546 n. a. Source: Information from fieldwork. The Rural Labor Market 321 Appendix. A Method of Calculating Labor Supply for TVPs To facilitate the calculations we assume that there is no surplus labor in for- estry, animal husbandry, or fisheries. As the proportion of TVP employment in these activities is small and the potential surplus is even smaller, they can be neglected. The great amount of potential surplus labor in China's agricultural sector determines the supply of labor to TVPS. Surplus labor in agriculture is calcu- lated by deducting the labor needed for crop growing, forestry, animal hus- bandry, sideline production, fisheries, scientific research, education, public health, and other activities from the total labor force in rural areas. For the entire survey and the methods of calculation, see China Rural Development Research Group (1985), nos. 1 and 2. The amount of labor needed for crop growing can be calculated by dividing the actual sown area in a region by the per capita crop-growing capacity of local farmers. Owing to different climatic conditions, types of crops planted, soil conditions, and levels of agricultural mechanization, farmers' crop- growing capacity differs among regions. During 1979-83 the Development Research Institute of the State Council Rural Development Research Center conducted a detailed survey on per capita crop-growing capacity among farm- ers in Ji'an Prefecture in Jiangxi Province, Fengqiu County in Henan Prov- ince, and Chuxian County in Anhui Province. The results of the survey are shown in the following table. Survey results Theoretical figure Average Ji'an 6.28 6.45 6.4 Fengqiu 7.01 7.97 7.5 Chuxian 7.75 7.95 7.85 Ji'an has two crops of rice and one of wheat each year and devotes large areas to rice growing; Chuxian grows one crop of rice and one of wheat every year; and Fengqiu usually grows one crop of cotton and one of wheat or two crops of wheat a year. The crop-growing situation in Jieshou is similar to that in Fengqiu, that in Wuxi is similar to that in Chuxian, and that in Nanhai and Shangrao is similar to that in Ji'an. We can therefore use the averages of theoretical figures and survey results to derive a rough estimate of the crop- growing capacity per farmer in each of the four counties (see table 14-Al). From the total labor force of each county, we subtract, in order, the amount of labor needed for crop cultivation (based on the crop-growing ca- pacity per farmer as calculated above), actual labor used in forestry, animal husbandry, and fisheries, and employment in scientific research, education, culture, and government. We are left with estimates of the potential surplus labor from agriculture available for employment in TVPS. 322 MENG XIN Notes 1. Since statistics classified by township were not available for Jieshou, the analysis of the situation there is based on data from sample enterprises in that county and supplementary quan- titative tables in the three communities of Jieshou town, Hebei township, and Jinzhai township. 2. Data on Dongjiang township are from government statistics on TVCES in Dongjiang. Data on Jieshou town are from the Enterprise Quantitative Questionnaire filled out by twenty-two enterprises in the town. Data on Yanbu district are from statistics provided by the TVP bureau in Nanhai County. No comparable data were available for Shangrao County. 15 Noneconomic Determinants of Workers' Incomes Wu Quhui, Wang Hansheng, and Xu Xinxin Analysis of the distribution of personal incomes can focus on actual income distribution or on the opportunity structure for maintaining and increasing incomes. The latter is more significant because it determines future income distribution. Opportunities for labor are related to the extent of free competi- tion and reflect the social structure of a community. According to the eco- nomic theory of pure competition, the price of labor is decided by the rela- tionship between market supply and demand, which in tum determines the opportunity structure of the selling and purchasing sides. But this theory is based on the premises that the supply and demand sides of the labor market have a simple trading relationship on a legally equal footing and that both sides pursue economic gain with no consideration of noneconomic factors such as social status. The wage-setting principle is marginal productivity, and the price of labor depends solely on its economic returns. Obviously this situ- ation of pure greed competition is a theoretical construct that is rarely found in practice. The social structure and systemic environment confronted by TVPS in China's rural communities are far from the scenario given by this theory. China's TVPS generally face a more complete market environment than do state enterprises. TvPs are much less controlled by state plans, they enjoy greater access to the labor market than state enterprises, and, in general, local labor supply and demand have a stronger influence on their wages. Wages in TVPs are directly linked to the enterprises' economic retums (see chapter 13). Community governments, however, exercise varying degrees of control over their firms, and their recognition of community members' eco- nomic rights and freedom also varies, both in content and in degree (see chapter 7). Even if two communities have similar labor markets, their gov- ernments may exercise different forms of noneconomic control over TVP em- ployees' incomes. The relationship between TVP owners and their employees depends on the 323 324 WU QUI-UI, WANG HANSHiENG, AND XU XINXIN form of ownership. China's TVPS fall largely into five categories of ownership: township enterprises, village enterprises, production team enterprises, part- nerships, and individual firms. The first three are owned by community gov- ernments at various levels, while the latter two are private enterprises. The township govemment is the owner of its subordinate enterprises, and it also exercises administrative functions. As the govemment of a community, it protects the interests of community members, improves their living standards and social welfare, and adjusts income distribution among different social groups to achieve general parity within the community. But the township govemment, as requested by higher authorities, must also ensure a steady growth of agricultural production. Accordingly, it has to perform multiple functions and make a proper allocation of resources, including labor, within the community. In pursuit of such goals, township governments canL utilize administrative means, but they may also use township enterprises to reduce unemployment and regulate income distribution. In contrast, private enter- prises rely more on the market for labor. As a result, the relationship between owners and employees is simpler, and workers' incomes are determined more by economic laws. Given the state of labor supply and demand, the ownership pattern deter- mines the extent of market competition for labor. When township enter- prises enjoy absolute superiority and competition from private enterprises is weak, the township govemment is likely to exercise stricter control over em- ployment and wages of township enterprise workers to fulfill its objectives. For example, to ensure adequate labor for farming to boost food production, the township govemment may impose limits on the incomes of those engaged in industrial activities or simply use part of enterprises' profits to subsidize agriculture. But if there is fierce competition from private enterprises, town- ship governments have to relax noneconomic control over wages in township enterprises. Since the 1970s and particularly in the past few years there has been a big upsurge in the development of TvPs in China's rural areas. This has tumed the originally simple rural society of farm laborers into a pluralistic society with diverse trades and strata-a transition that has profoundly af- fected income distribution between workers and farmers, among enterprises, and among individuals. Exactly what are these new income distribution pat- tems? What relationship do they have to the community opportunity struc- ture? What roles do community govemments play in income distribution? And finally, what have been the consequences of these patterns for incen- tives, equity, and efficiency? In this chapter we focus on government intervention in the determination of personal incomes in TVPS in three counties- Wuxi, Nanhai, and Jieshou. (Shangrao was excluded from the analysis because some key data could not be obtained there.) Noneconomic Deterrminants of Workers' Incomes 325 The Opportunity Structure and Sources of Labor The ratio between total employment in TVPS and surplus agricultural labor in a community reflects labor supply and demand. In 1985 there were 323,800 TVP employees in Wuxi, or 80.5 percent of its surplus agricultural labor force. In Nanhai TVP employees numbered 206,839, or 86.5 percent of the county's surplus farm laborers. But in Jieshou TVP employees numbered only 28,900-15.0 percent of the surplus agricultural labor force-and the labor supply situation for TVPS was thus more favorable than in the other two counties (see chapter 14). As a result, TVP workers' average monthly income in Jieshou was Y66, compared with Y100 in Wuxi and Y1 78 in Nanhai. Despite a similar labor market situation, there is a big difference between average incomes in Wuxi and Nanhai counties. Of course, this gap can be attributed to economic factors such as industrial structure, technical capaci- ties, and product prices. But we have every reason to believe that different degrees of control by community govemments over income levels have been an important factor. As can be seen from table 15-1, workers in private enter- prises in Wuxi accounted for only 3 percent of total TVP employment, whereas the figure was 22 percent in Nanhai and 55 percent in Jieshou. Pri- vate enterprises usually shoulder much less responsibility for improving the living standards and social welfare of all community members, and they do not have administrative control over their employees. Their motive is simply to make profits, and they pay their workers mainly according to labor supply and demand. Township govemments' control over the incomes of workers in township enterprises is weakened when there is fierce market competition for laborers from private enterprises and when supply is short. The result is a stronger role for market supply and demand. Community opportunity structure is determined by the distribution of en- terprises with different forms of ownership and by the mobility of community members. When all TVPS in a community are owned by its government, which strictly forbids mobility, community members' choices in employment and pay are limited and the community govemment is in a favorable posi- tion. But when there are many private enterprises and community members enjoy a high degree of mobility, the opportunity structure favors individuals, Table 15-1. Employment Structure of TVPs (percentage of total in TVP sector) Type of enterprise Wuxi Jieshou Nanhai Township 47.1 32.0 23.9 Village 50.0 13.4 53.6 Partnership 9 24.5 1 225 Individual I l 30.1 a. Includes only industrial TVPs. Source: Information from fieldwork. 326 WU QUHIUl, WANG HANSHENG, AND XU XINXIN Table 15-2. Composition and Origin of the TVP Labor Force (percentage of total sample employees) Item Wuxi Jieshou Nanhai Type of employee Long term 91.2 75.8 40.2 Temporary 8.5 18.1 59.8 Seasonal 0.3 6.2 0.0 Origin of employee Same village 94.0 79.9 50.9 Other villages in county 3.0 12.7 33.8 Other counties in province 1.6 4.8 14.2 Other provinces 1.4 2.8 1.1 Source: Enterprise Quantitative Questionnaire. and their employment and incomes depend more on market supply and de- mand. Table 15-2 demonstrates the differences in mobility of community members in the three counties. Nanhai has the greatest intracommunity and intercommunity mobility, Jieshou is second, and Wuxi is last. Therefore in Wuxi the opportunity structure favors community governments, in Nanhai it favors community members, and in Jieshou the situation is in between. It is particularly worth noting that many so-called township enterprises in Jieshou are really private; only a few firms are truly run by township govern- ments. The labor force in Jieshou enjoys fairly high mobility, but actual op- portunities for choosing employment are far smaller than in Nanhai because labor supply exceeds demand. As a result of varied social opportunity structures, workers join firms in drastically different ways in thLe three counties. The percentage of employees designated by community governments is substantially higher in Wuxi than in the other two counties. In Nanhai more workers are employed through the fairly well-developed labor market, and there are fewer government- designated employees. Unlike the situation in Wuxi, community govern- ments are not able to compensate the unemployed or balance incomes of workers in firms with different: economic returns and working conditions. In Jieshou too, workers are recruited by enterprises directly instead of through community governments, but the labor market is not as developed as in Nan- hai. Many people are recommended by their relatives or friends, and conse- quently there are strong blood relationships within firms. TVP employees in Jieshou have a deep sense of belonging to their enterprises. They would ac- cept wage cuts rather than desert their firms. They also have a strong desire to reinvest profits in their enterprises and are content with their present in- comes. One reason is the difficulty of finding jobs in other firms. Noneconomic Determinants of Workers' Incomes 327 A more unified ownership pattem also makes it possible for township gov- ernments to maintain control over the incomes of directors and managerial personnel. In Wuxi community govemments are in a favorable position, whereas in Nanhai, where the ownership pattern is more diversified, directors and managerial personnel are attracted to private enterprises by higher pay, and they enjoy more employment choices and greater mobility (see table 15-3). Community Balancing Mechanisms in Wage Determination The community opportunity structure normally determines the extent of community government control over the incomes of TVP employees. Such control is designed to redistribute economic benefits among different groups within the community. The chief goals of township governments are to raise the general level of welfare and to balance the incomes of individuals in dif- ferent social groups, and the govemments use every possible means-usually administrative-to achieve these objectives. The higher the share of town- ship enterprises' economic returns in total community income, the easier this is to do. As we have seen, township govemments in Wuxi are in a more favorable position to exercise such control. Balancing Incomes of Workers arnd Farmers The average monthly income of township enterprise workers in Wuxi is Y1 10 and that of agricultural laborers is Y108. In Nanhai the figure for township enterprise workers is Y190. We do not have data on farm laborers, but their incomes could not be lower than those of TVP workers because many of the latter are now returning to farming. In Jieshou the figures are Y60 and Y40, respectively. There is little difference between average incomes of TVP work- ers and of farmers in Wuxi. The latter is probably a little higher than the former in Nanhai, and in Jieshou the average monthly income of workers is 1.5 times that of farmers. In contrast to Nanhai, part of the income of Wuxi farmers comes from redistribution of industrial profits. In 1985 township governments in Wuxi set aside Y12.7 million to compensate farmers, whereas in Nanhai such com- Table 15-3. Origins of TVP Managers (percentage of sample total) Origin Wuxi Nanhai Jieshou Same village 93.1 53.5 78.1 Other villages in county 0.6 39.7 16.5 Other counties in province 0.5 5.6 2.3 Other provinces 5.9 1.2 3.2 Source: Enterprise Quantitative Questionnaire. 328 WU QUHUI, WANIG HANSHENG, AND XU XINXIN pensation was only Y6.6 million. Average compensation per farmer was Y66.40 in Wuxi and Y37.40 in Nanhai. Moreover, actual compensation in Nanhai is probably less than 20 percent of the figures on the books. In Jieshou compensation for farmers accounted for 17 percent of the profits enterprises turned over to the state, or only Y40,000. Enterprises' economic returms are diverted to farmers in different ways in the three counties. In Wuxi the main channel is community governmrrent sub- sidies for pig raising, grain production, and cultivation. The Dongjiang town- ship government, for example, pays farmers Y10 for the first pig and Y20 for the second. For every half kilogram of grain sold to the govemment: farmers receive a subsidy of YO. 06, and those who are willing to cultivate more than 10 mu of land receive Y12 in subsidies. Farmers also get similar subsidies from village govemments. In Nanhai subsidies are seldom issued to individual farmers. Another way to balance incomes is through a community pension system. Township govemments in Wuxi stipulate that people who have worked in township enterprises for more than three years are entitled to pensions of at least Y15 a month. Agricultural workers enjoy preferential pension treat- ment, however. Speciatlized farming households receive the same pensions as TVP workers, but the retirement age for male farmers is five years earlier than for workers. Pension systems in Nanhai are available only to employees of township enterprises. and they differ among enterprises. No comrnunities in Jieshou have established pension systems. The money farmers get ifrom re- distribution of TVP profits and the pension system are in fact entitlements for being community members which allow them to benefit from the eco- nomic returns of TVPS. People not employed in TVPS can be compensated through these "membership incomes." In Wuxi substantial amounts of TVCES' returns have been used to increase farm laborers' incomes and welfare. Township governments there feel a great responsibility for balancing the welfare of all community members. More- over, under the county's agricultural policy, township governments are ex- pected to ensure a steady growth of grain production at the same time as they develop TVPS. In fact, the original reason for establishing TVPS there was to promote agricultural production and ensure reinvestment in agriculture (see chapter 7). From the land contracts signed between township governments and farmers we can see the irnportance Wuxi has attached to grain produc- tion. In Dongjiang township, for example, land contracts contain strict stipu- lations on the crop varieties to be grown on contracted farmland; they en- courage grain production, limit cash crops, and forbid leaving arable land idle. But in Nanhai farmers can grow any crops they wish as long as they fulfill state grain quotas, which they can do by buying grain from other locali- ties. In Wuxi farmers must fulfill quotas with their own grain production. This continuing emphasis on grain production has led to a shrinkage in the Noneconomic Deterrninants of Workers' Incomes 329 acreage of cash crops. Historically Wuxi was an important silk producer, but since the 1960s its mulberry acreage has been declining, from 100,000 mu in 1962 to 80,000 mu in 1975, 70,000 mu in 1980, 60,000 mu in 1984, and only 35,000 mu in 1986. Balancing Incomes among Enterprises in a Community Communities also seek to balance wage incomes among TVPS. Differences in average wage income among firms are smallest in Wuxi, largest in Nanhai, and in between in Jieshou. Different factors are responsible for these gaps in the three counties. In Wuxi township governments have exercised stricter control over wages and bonuses in enterprises. Firms' total wage bills must be approved by town- ship governments, and taxes are levied on enterprises when wages surpass a stipulated standard (Y80 per month in Dongjiang township, for example). Township governments in Nanhai exercise indirect control over workers' in- comes through their control over the personal incomes of those who have contracted to run enterprises. Since the contractors' incomes are linked to enterprise profits, higher wages for workers mean lower profits and reduced bonuses for directors, who thus have an incentive to hold down wages. Town- ship govemments in J ieshou have basically no control over TVP wages, which are set by enterprises themselves and largely reflect labor supply and demand on the market. Township governments in Wuxi can achieve an income balance among community members through financial compensation for farmers, improve- ments in social welfare, and control over wages. Such influence is weaker in Nanhai and weakest in Jieshou. Not only do township governments exer- cise more control over wages in Wuxi; community members have less choice of jobs and rely more on community govemments to improve social welfare. In Nanhai, in contrast, private enterprises are more important, employees of private enterprises are recruited through market competition instead of being designated by community governments, and all job hunters have equal opportunities. Township governments therefore have less influence than in Wuxi. Moreover, reducing the wages of township enterprise employees and transferring part of township enterprise profits for redistribution to other com- munity members would lead to unequal opportunities because employees of private enterprises are not affected by such wage limits. And since commu- nity members have more choices in employment, they are less dependent on community governments for bettering their welfare. Community govemments in Jieshou do not have much influence on TVPS because many township enterprises are actually run by individuals. In this situation the market plays a more significant role in determining wages, and the balancing function of community govemments is limited. 330 WU QUHUI, WANG HANSHENG, AND XU XINXIN Income Differences within Enterprises In Wuxi individuals' living standards are strongly related to the general com- munity living standanr, whiich means that the realization of individlual inter- ests depends mainly on the community. But this does not mean that there are no income differences among classes of employees within an enterprise. On the contrary, differences in the human capital of management personnel and workers make it impossible to attain an income balance among directors, other managerial personnel, and workers and at the same time maintain the motivation of managers. In Nanhai and Jieshou, where not every community member shares equally in thLe increased overall welfare of their communities, individuals' welfare depends on the profits of their firms, but employees may or may not share equally in enterprise profits. Information from the sample shows that income disparities within firms vary in the three counties. T'he ratios of average monthly wages, wit]h workers as 1.00, are 1.74 for sales staff, 1.91 for technicians, and 2.07 for directors in Wuxi. In Jieshou the figtures are 1.40 for sales staff, 1.96 for tectnicians, and 1.47 for directors. Inccome differences within enterprises in Nanhai are larger than in Wuxi, although accurate information is not available. Most TVP workers in Nanhai receive piece rates, the norms for which are set by the enterprises themselves. Usually there are no limits on the income dispar- ity between directors and workers, but directors' incomes depend on enter- prise profits. For example, ithe incomes of enterprise directors in Pingzhou town are made up of a basic monthly salary of Y66; a monthly position salary of Y70; a bonus in proportion to the enterprise's profits, divided into eight grades that range fromn Y60 to Y240 a month; and bonuses for above-quota profits, innovative activities, and above-quota sales.' According to local offi- cials some enterprise directors could earn as much as Y30,000 a year under these regulations, but the highest actual income of a township enterprise di- rector in Pingzhou in 1986 was Y5,000. Many township governments in Nan- hai stipulate that enterprises can retain 30 percent of their profits as bonus. Of this, 70 percent is divided among workers and the rest among managers. For instance, managers of the Luocun Gloves and Hides Factory could earn Y10,000 in 1986, and workers received an average monthly salary of Y120 plus a monthly bonus of Y30. These figures show that income disparities within firms are bigger in Nanhai than in Wuxi and Jieshou. The factors responsible for income differentials vary among the three coun- ties. First, there are different contract management systems. In Wuxi twenty- five of twenty-nine sample township enterprises have adopted the collective contract system. In Jieshou 46 percent were contracted by individluals. In Nanhai there are no ieliable data on the number of contracted firms, but case investigations indicate that most are contracted by individuals. Under the collective contract system, management and workers shoulder risks to- gether and hence have common interests. But under the individual or direc- Noneconomic Determinants of WOrkers' Incomes 331 tor contract system, risks are shouldered by individuals, and their personal interests are divorced from those of the workers to some extent. This explains why there are different community government regulations on income distri- bution in Nanhai and Wuxi. In Wuxi's Dongjiang township the basic and position wages of top managers are fixed, but their floating wages are decided by production quotas and by the average level of workers' floating wages. Moreover, the maximum total incomes of directors are determined by the average income level of workers. (They are generally about twice the latter.) Income increases for directors hence depend on income increases for workers. Township governments in Nanhai usually set the relationship between direc- tors' incomes and enterprise profits, leaving workers' incomes to be decided by directors. Obviously this puts the director in a more favorable position in income distribution than workers and other managers. Moreover, the di- rector's personal income will decrease if he raises workers' incomes. Decisionmaking on income distribution has much to do with opportunity structures. In Nanhai the existence of different ownership forms creates a competitive environment that is intensified by the shortage of labor. At the same time, the mobility of labor and managers is great. These two factors create more opportunities for workers and managers and make it difficult for township governments to exercise control over incomes of township enter- prise employees. Inasmuch as the evaluation of enterprises is based on their profits and factory directors have the power to limit workers' wage increases, there are wide income differences within enterprises. Township govemments in Wuxi occupy a more favorable position. Com- munity ownership of firms and weak labor mobility enable township govern- ments to keep a rough balance of average individual incomes among enter- prises and between TVP workers and farmers. They are also able to exercise direct control over income differences among types of employees within en- terprises by controlling enterprises' total wage bills, linking incomes of direc- tors and other managers with those of workers, and setting a maximum per- centage difference between the two, Township governments thus ensure that directors enjoy no more favorable opportunities and rights than ordinary workers, and the incomes of directors and workers fluctuate synchronously. This certainly strengthens the workers' status as de facto owners of their firms. Although as in Nanhai township governments in Jieshou do not have con- trol over TVP wages but instead let directors make income distribution deci- sions, income differentials within enterprises are small. A possible explana- tion is that many of the county's township enterprises used to be private firms and have maintained many of their old methods. Some remain fundamen- tally private enterprises. It is possible to keep workers' wages low when there is heavy pressure from surplus agricultural labor. Enterprise directors also hold down their own nominal wages to keep the gap between their incomes and workers' wages narrow and thus help restrain the latter. This does not mean 332 WU QtJHUI, WANG HANSHENG, AND XU XINXIN that the actual incomes of directors are low. They cover up their actual in- comes because backward regions lack the concept of private property owner- ship and proprietors dare not claim their own enterprises. During fieldwork we met many people who refused to admit that their enterprises were their own property and regarded them as being owned by township governments. Wage Determination Table 15-4 lists the shares of various forms of worker income. "Other" in- cludes welfare benefits offered by enterprises to all of their employees equally in the form of money, that is, membership incomes of employees. This type of income is nonexistent in Jieshou and minuscule in Nanhai, but it accounts for as much as 14 percent of monthly income of sample TVP employees in Wuxi. The appraisal of income distribution hence emphasizes enterprise membership in Wuxi but not in Nanhai and Jieshou. Income parity among employees in Jieshou is realized through equal distribution of basic wages and bonuses instead of through welfare benefits. In Wuxi the income of managerial personnel is composed of basic wages, position wages, floating wages, and annual bonuses based on a work-point system. In Dongjiang township, for example, position incomes of chief mana- gerial personnel are divided into two grades with three classes in each accord- ing to work load, responsibilities, and enterprise size (see table 15-5). Town- ship governments decide to which grade an enterprise belongs. Floating wages depend on six indexes divided into 100 points, with noneconomic as- pects (such as Commtnist party affairs, military recruitment, security, civil mediation, family planning, and office administration) accounting for 20 per- cent of the total. Annual bonuses are decided by the total annual profits of the community's township enterprises and by the increase in profits. Al- though bonuses are related to enterprise performance, the relationship is not completely fixed but rather is decided through comparison with income levels in other enterprises. As a result, the amount of annual bonuses is completely determined by township governments, and horizontal comparisons generate competition among directors of different firms, which forces them to go after production targets set by the township government. Since the floating wages of managerial personnel usually equal the average wage level of workers, in- Table 15-4. Composition of Mvonthly Income (percentage of total monthly income) Iterrm Wuxi Nanhai Jieshou Basic and piece-rate wages 67.5 78.5 92.7 Bonuses 18.8 20.6 7.3 Other 13.7 1.0 0 Source: Enterprise Quantitative Questionnaire. Noneconomic Determinants of Workers' Incomes 333 Table 15-5. Managerial Pay, Dongjiang Township, Wuxi County (yuan) Occupation Grade one Grade two Party secretaries and directors 50 36 Deputy Party secretaries and deputy directors 36 24 Accountants 24 14 Source: Information from fieldwork. come differences between the two are reflected in managers' basic wages, po- sition wages, and annual bonuses. Incomes of managerial personnel in Nanhai are composed of basic wages, position wages, bonuses for profits and above-quota profits, and sales bonuses. Unlike the situation in Wuxi, the determination of these items is explicitly stipulated at the beginning of the year. Except for basic and position wages, the components of income are directly related to enterprise performance. For instance, in Xiqiao the township government stipulates that incomes of man- agement personnel consist of basic wages, food subsidies, and bonuses, with no position wages or other living subsidies. The bonuses of directors and Communist party secretaries are divided into six classes, which are directly linked to the annual profits of their firms (see table 15-6). Bonuses of deputy directors are set at 90 percent of directors' bonuses, and bonuses for other managers are decided by directors according to certain ratios. In comparison with Dongjiang township, incomes of managers are more directly linked to the economic returns of enterprises, which are independent from each other in income distribution. The larger income gap among enterprises in Nanhai shows that wages indeed are directly decided by economic returns. The determination of managerial incomes is simpler in Jieshou, where they are composed of basic wages and bonuses for overfulfilling profit targets. As in Nanhai, township governments make explicit stipulations beforehand and interfere little afterward. The determination of workers' incomes also varies widely in the three counties (see table 15-7). As in the case of managers' compensation, incomes of workers in Wuxi are decided by township economic commissions. Al- though total wages are linked with enterprise profits, monthly pay of workers, even in village enterprises, has to be approved by the commissions. In fact, 3 percent of firms still determine wages according to a unified work-point sys- tem with farmers, a continuation of the past method. This also reflects the emphasis in income distribution on community membership, owing to lack of employment competition. Another interesting feature is that in Wuxi 18 percent of TVPS use work-point systems whereby workers' incomes are based not only on working time, work load, and product quality but also on atti- tudes, seniority, and other noneconomic factors. 334 WU QUHUI, WANG HANSHENG, AND XU XINXIN Table 15-6. The Relatonship between Directors' Bonuses and Enterprise Profits, Xiqiao Town, Nanhai County Arinuai profits Monthly bonuses (tousands of yuan) (y-n) 150 200 51-100 230 101-1150 260 151-200 290 201-250 320 251-300 350 Source: Information from fieldwork. In Jieshou workers' incomes basically consist of fixed wages, piece-rate wages, and bonuses, which are usually decided by directors. Since wage dif- ferences among enterprises within a community are relatively large and in- come gaps within enterprises are narrow, we can say that in Jieshou wages are based on economic retun-s of enterprises and that enterprises rather than individuals are the basic units. In Nanhai 75 percent of workers' income is from piece-rate wages and the remainder consists of supplementary incomes, chiefly bonuses and extra piece-rates. Since there are wider income gaps among employees within en- terprises, we can conclude that incomes in Nanhai are based on individual performance. Conclusions As discussed earlier, forms of ownership and individual mobility vary in the three counties. When township enterprises account for a large percentage of TVPs and labor mobillity is restricted, individuals' choices are limited. The more opportunities for individuals are restricted, the more they have to rely on community governments to raise their living standards. Our analysis indi- cates that different opportunity structures in the three counties have led to different degrees of noneconomic intervention in the incomes of TVP employ- ees by community governmernts. In Wuxi there is virtually no labor market, and community members are mostly assigned to enterprises by community governments, which take into account not only the quality of laborers but also their economic conditions and family burdens. The result is that individuals depend on community gov- emments for employment. Because of govemment intervention, income dis- parities between TVP workers and farmers are not large. These two factors have contributed to the development of a community consciousness that is much stronger in Wuxi than in Nanhai and Jieshou, as we can see from table 15-8. In Wuxi 10 percent of sample workers would prefer that most of their Noneconomic Determinants of Workers' Incomes 335 Table 15-7. The Composition of TVP Workers' Incomes (percentage of total number of enterprises) Wage system Wuxi Jieshou Nanhai Primary Time rates 9.1 29.7 16.7 Piece rates 66.7 64.9 75.0 Enterprise work-points 18.2 0 8.3 Community work-points 3.0 0 0 Secondary None 12.1 34.3 30.0 Piece rates 72.9 60.0 55.0 Bonuses 79.1 2.9 5.0 Dividends 9.1 2.9 5.0 Other 6.0 2.9 10.0 Source: Enterprise Survey Questionnaire. enterprises' profits be used for rural construction, but in Nanhai only 4 per- cent and in Jieshou 3 percent have such a wish. Wuxi community govern- ments are placed in the delicate position of balancing equality and efficiency. They must, on the one hand, try to develop a sense of equality among differ- ent trades and social groups and, on the other, maintain sufficient incentives to stimulate the work initiative of all community members. Community governments in Nanhai and Jieshou have a lighter burden in this respect. They intervene less in income redistribution among different social groups in the community, and workers' incomes are determined by en- terprise performance. In Nanhai there is a relatively developed labor market through which most individuals seek employment. Wages are determined by enterprise efficiency and by labor market trends. This has led to marked in- come disparities between workers and farmers, among enterprises, and among different types of TVP employees. Individuals obtain economic benefits through labor market competition instead of govemment intervention. As a result, individualism is stronger and the sense of community interests much weaker than in Wuxi. Fifty-three percent of Nanhai sample employees would prefer that more enterprise profits be allocated as individual incomes, com- pared with only 27 percent and in Wuxi and 7 percent in Jieshou. In Jieshou as many as 89 percent of sample workers want their enterprise's profits to be used for reinvestment, but only 57 percent of the workers in Wuxi and 36 percent of those in Nanhai have such a desire (see table 15-8). This strong sense of identification with firms in Jieshou can be attributed to scarce job opportunities, the blood relationships common among employees, and the narrow income gaps within enterprises. Distribution relationships within a community cannot but have an impact on the incentives of TvP employees. In Wuxi the incentive role of wages has been weakened and has been replaced by the pursuit of job reputation, social 336 WU QUHUI, WANG HANSHENG, AND XU XINXIN Table 15-8. Employees' Preferences Concerning Use of Enterprise Profits (percentage of total sample employees) Use Wuxi Jieshou Nanhai Individual incomes 27.0 6.8 53.4 Expansion of production 57.3 88.8 35.6 Rural construction 10.1 2.7 4.0 Other 1.5 1.0 0.8 Source: Worker Survey Questionnaire. status, and higher grading of enterprises. TVP employees in Wuxi prefer to remain workers rather than go back to farming, even if they would earn less. They would also prefer to be employed by state enterprises, even at lower pay than in their current jobs. TVP employees in Nanhai and Jieshou, how- ever, care more about their incomes. In Nanhai and Jieshou there are marked differences among wage levels in different enterprises owing to lack of gov- ernment intervention. As a result, incomes are more of an incentive than in Wuxi. Workers in the two counties would be farmers if they could earn more, and they would stay with TVPS for the sake of higher incomes instead of going to state enterprises. 'This tendency to pay more attention to incomes may have a significant effect on enterprise performance and the develop- ment of rural industry. We also analyzed the relationships among different social groups within enterprises in the three counties. In Wuxi factory directors and workers com- monly share risks. Comtmunity governments impose a ceiling on income dif- ferences between directors and workers and control firms' overall wage bills by levying a bonus tax. As a result, there are only small gaps among wage levels in different firm.r and between incomes of directors and workers. As there is a relatively developed labor market in Nanhai, workers obtain in- comes through the market mnechanism on the basis of their human capital, and relationships among social classes are based on economic division of labor rather than on noneconomic factors. In Jieshou the strong blood ties that often exist among employees are cemented by the scarcity of job oppor- tunities, and there is no visible class polarization within an enterprise. When their enterprises face serious financial deficits, 84 percent of sample employ- ees in Jieshou would prefer to stay even if their wages were cut by half. The ratios in Nanhai and Wuxi are 47 percent and 64 percent respectively. When they see other employees shiriking, 70 percent of Jieshou sample workers will directly remind them that they should improve their behavior. In Nanhai only 52 percent and in Wuxi 51 percent of sample employees are willing to do this. All these observations indicate that employees in Nanhai demonstrate the strongest individualism. They regard firms as arenas for economic activities and sources of their incividual benefits. They also have a strong desire for Noneconomic Determinants of Workers' Incomnes 337 labor mobility and rely on themselves instead of on govemment intervention to obtain incomes through market competition. Jieshou is another extreme example; enterprises are like big families in which individuals equally share economic interests, have a deep feeling toward their firms, and exert a strong influence over each other's behavior. The three counties will face different problems in their future develop- ment. Wuxi will have to deal with the problem of efficiency. Employees should be encouraged to care more about their enterprises' performance through greater economic incentives instead of higher social status. Perhaps only systemic reforms could solve this problem. Nanhai will have to concen- trate on the problem of economic equality. The experience of other develop- ing countries shows that with industrialization the gap between rich and poor will further widen unless govemment intervenes. In a country like China where egalitarianism predominates, the size of income disparities that the masses will accept and the disparities that are most conducive to economic efficiency may differ, and further study of this issue is needed. The priority for Jieshou is to strengthen the organizational aspect of TvPs and weaken the existing blood relationships. It is true that family and other close personal relationships help to develop employees' identification with firms and encour- age participation in the management of their enterprises. But we still have to ascertain whether employees' full identification with enterprises can change the nature of enterprises as economic organizations and whether the equal sharing of economic benefits will weaken the incentives of employees. Note 1. In the Huada Electric Fan Factory, for example, the director is entitled to a bonus of Y240 out of every Y10,000 of above-quota profits, Y480 for every Y10,000 of profits arising from inno- vative activities, and Y360 for every Y10,000 of profits from above-quota sales. Part V The Role of Community Governments Rural community govemments-townships, villages, and to a lesser extent production teams-have had a profound impact on the nature, speed, direc- tion, and accomplishments of rural industrialization in China. They estab- lished and own the community enterprises that are responsible for the bulk of the industrial output of the TVP sector. Moreover, where private firms have developed rapidly, they have done so with the encouragement and often the active support of community governments. Thus an analysis of the roles, mo- tivations, and behavior of community governments and their leaders is essen- tial for an understanding of China's rvp sector. The two chapters in this sec- tion look at this topic from somewhat different perspectives. In chapter 16 Song Lina and Du He analyze the role of township govern- ments in rural industrialization. The authors argue that the most fundamental linkage between township governments and their enterprises is in the realm of public finance: the enterprises provide the revenue needed to run the gov- ernment and improve social welfare in the community. China's state budget- ary system and the extrabudgetary funds that are so important at the local level are briefly described. Song and Du then look more specifically at the public finances of township govemments. Townships rely on revenue from community property to fi- nance their operations, and rural industrialization carried out by township enterprises offers a potentially productive and easily tapped source of funds. The production responsibility system (PRS) further strengthened the public finance motivation for rural industrialization by removing agricultural land and related assets from the control of community governments. The authors next discuss the bifurcated role of township govemments-as property owners and as local governments-and the problems engendered be- cause the two roles tend to weaken and compromise each other. They assess the direct participation of township governments in the operations of town- ship enterprises as manifested both in preferential treatment for township en- terprises and in restrictions imposed on them by township governments. 339 340 THE ROLE OF COMMUNITY GOVERNMENTS Factors that generate differences in the behavior of township govern- ments-including natural resource endowments, the level of development, employment pressures, and the quality of the labor force-are then reviewed, and the impact of township governments on local industrialization is ana- lyzed. In all four counties township governments have considerable control over township enterprises, but the focus and orientation of this control var- ies. Township governments in Jieshou and Nanhai regard the key levers of power to be control over personnel (through appointment of enterprise lead- ers) and investment. In Wuxi investment is first, followed by production planning, and in Shangrao control over production planning is considered the most important, with personnel second. The authors provide plausible explanations for this variatiort. In the concluding section Song and Du argue that the present pattern of direct control by township governments over township enterprises will con- tinue for some time andl that rash attempts to change the system radically will fail. What is needed is a proper balance in the roles of township govem- ments and avoidance of vicious cycles in financial flows between township enterprises and township govemments. In chapter 17 William Byrcl and Alan Gelb analyze the motivations and incentives behind comm!unity governments' involvement in the TVP sector and their support for TVP development. The analysis is based on the assump- tion that Chinese community governments and their leaders can be viewed as self-interested and at least somewhat autonomous entities that pursue their own benefits, as distinct from the social welfare of the community as a whole. This basic postulate of public choice theory, which is usually applied to nation-states, has some interesting implications for rural communities in China. Some key features of the rural economic and institutional environment are identified: population and factor immobility; the resulting spatial inequality in rural incomes and industrialization; the ineffectiveness of the fiscal system in redistributing income among localities; compressed income differentials within communities and TVPS; and the need for community governments to rely largely on the resources of their own communities in developing TVPS. The authors go on to describe the flows of funds between TVPs and the local community and to review the strong linkages between Tvp development and the community economy and government. These consist of fiscal link- ages, income and employment linkages, and rewards to community govern- ment leaders-salaries and formal incentives, informal incomes and con- sumption, and promotions. It is argued that community government leaders in virtually all Chinese rural communities have strong incentives to develop the local TVP sector. The personal incentives for township leaders, however, are somewhat less than those for leaders of villages and production teams be- cause the former are state cadres paid under a unified national system. Byrd and Gelb then look at patterns of community government involve- The Role of Community Governments 341 ment in the TVP sector. These include successful development of community enterprises in certain areas that are better endowed with financial and human resources and have easy access to urban markets; fiscal predation against the community enterprise sector by community govemments in poorer, more backward areas; creation of a general policy environment conducive to devel- opment of private enterprises; and more active and direct financial and per- sonal support of larger, more successful private enterprises by community and county leaders. These pattems are leading to a situation in which ownership in the TVP sector is becoming increasingly correlated with the level of devel- opment: localities with successful community enterprises may or may not allow the private sector to flourish, but the backward, hitherto unsuccessful, areas are forced to tum toward developing private enterprise. Fiscal predation occurs when community govemments' regular revenues are insufficient to cover their large, virtually fixed expenditures for local pub- lic and social services and for their own payrolls. Govemments are forced to milk their enterprises for funds regardless of whether the firms are earning profits. Unprofitable community enterprises respond to fiscal predation by "teating" their capital stock (that is, by using their depreciation funds for re- mittances to community governments) or by borrowing from banks. The lat- ter case, in which community governments obtain indirect deficit financing from the local banking system, is especially detrimental. The concluding section suggests that spatial imbalances in rural industriali- zation and TVP development are likely to persist as long as population and factor mobility continues to be restricted. Fiscal predation exacerbates the problems of backward areas. The dual role of community governments as both local governments and business corporations leads to severe conflicts in the backward areas that suffer from fiscal predation, and the govemment role commonly squeezes out the business role. Although local measures such as election of factory directors and contracts between directors and commu- nity govemments can ameliorate fiscal predation, more fundamental solu- tions may be required. These might involve reducing the size of govemments in backward areas (especially the number of people on community govem- ment payrolls) and financing through the state budget certain essential social and public services that are presently paid for with community govemments' extrabudgetary funds. Another, narrower recommendation is for explicit bonus systems, tied to local economic performance, for township leaders. 16 The Role of Township Governments in Rural Industrialization Song Lina and Du He As the study of nonagricultural activities in China's countryside deepens, an increasingly clear picture of the external environment, business operations, and internal systems of rural industrial enterprises is emerging. China's rural industrialization is based on small rural communities, and the interests of these communities have a direct bearing on the behavior and goals of rural enterprises. Community governments not only have all the administrative functions of local government, but they also have some actual autonomy and represent to a great extent the common interests of their members. In this chapter the behavior of township governments is selected as the key variable for study, and its impact on rural industrialization is analyzed. Public Finance and the Functions of Township Governments China's highly centralized public finance system dates back to 1950. In March of that year the Central Committee issued two resolutions containing detailed regulations for the starte financial management system. Except for some local taxes and a few small income items that could be retained by local govemments to offset certain expenditures, all income was concentrated in the hands of the central government. Similarly, all outlays (except local sup- plementary expenditures) were included in the state budget. The fin.ancial outlays of local governments and state enterprises were all to be included in their budgets and had to be reported to and approved by the central govem- ment. In subsequent years this "uniform income and outlay" system inder- went several changes, but it was not until 1980 that China began to break away from overcentralization and excessively strict control and to offer more financial autonomy to local governments and state enterprises. Even under the centralized system, some govemment funds were mobilized and utilized outside the budgetary controls. At first these extrabudgetary funds included only a few items and were very small. Some townships, vil- lages, and departments of local government tried to use their own production 342 The Role of Township Governments in Rural Irdustrialization 343 and business operations to help finance rural education, cultural activities, administrative expenditures, and the like. In 1953 extrabudgetary funds in China as a whole were equivalent to only 4 percent of the state budget, but the figure climbed to 8.5 percent in 1957 and to 20.6 percent in 1960. Read- justments by the state caused the figure to drop to 14.5 percent by 1966, but it then rose again and reached 35.6 percent in 1977. As economic re- forms and decentralization of authority took hold, the figure jumped to 60 percent in 1981 and to more than 80 percent in 1985. In the public finance structure at the county level there are both budgetary and extrabudgetary revenues. Budgetary revenues are allocated by the next higher level of government and must be used for specified purposes, in ac- cordance with the uniform income and outlay system. Extrabudgetary reve- nues include supplementary industrial and commercial taxes, supplementary agricultural taxes allowed by the central govemment, and some other reve- nues from the county's operations and production activities. The four coun- ties we surveyed had some extrabudgetary funds, but these were generally not high in comparison with total budgetary funds. The figures below, obtained from fieldwork, show the percentage shares of industry and agriculture in the extrabudgetary revenues of the four counties. Wuxi Jieshou Nanhai Shangrao Agriculture 0.0 27.4 0.0 49.0 Industry 100.0 72.6 100.0 51.0 The share of agriculture is zero in the two most industrialized counties, it is relatively low in Jieshou, and it is relatively high (nearly half of the total) only in Shangrao. The uses of extrabudgetary funds in the sample townships of the four coun- ties are shown in table 16-1. Townships in Wuxi and Nanhai spend consider- able shares of their extrabudgetary funds on enterprise development, whereas those in Shangrao spend little for this purpose. Table 16-1. Extrabudgetary Expenditures in Sample Townships in Three Counties (percentage of total) Type of expen3diture Wu Nanhai Shangrao Enterprise development 30.5 22.1 1.4 Aid to agriculture 18.3 6.0 4.6 Community development 35.0 38.5 79.7 Monetary rewards for government employees 4.9 14.5 0.0 Administrative expenses outside the budget 11.3 18.9 14.3 Source: Township Leader Questionnaire. 344 SONG LINA AND DU HIE Township Government Operations Township governments are the grass-roots apparatus of the Chinese govern- ment. Their effectiveness is basically determined by the degree of their con- trol over the public wealth of rural communities. In a sense they have some of the characteristics of small nation-states (see chapter 17). Through the township govemment public wealth is accumulated, multiplied, drawn into the income and outlay system, and distributed to every interest group and individual in the comimunity. Township govemments are supposed to regard the overall interest of their communities as their ultimate goal. In fact, however, they have many, not necessarily compatible, tasks-promoting economic development, expand- ing employment, fostering industrialization, and improving local living standards-and they tend to favor those groups that can best quench their thirst for revenue. As a result, they cannot give fair treatment to every group and individual, and their control over public wealth can never be integrative and uniform. In line with the dramatic changes in rural areas brought about by the pro- duction responsibility system (PRs), the township was made an independent level in the govemment financial system in 1984. Townships are now allowed to retain part of the tax revenues collected in the area under their jurisdic- tion. They have control over disbursements from this retained portion of rev- enue, but other budgetary expenditures are still determined and allocated by the county government. Extrabudgetary funds remain separate accounts under the control of township govemments and are not subject to budgetary supervision. The revenue of township governments thus falls into two categories. Budg- etary revenue belongs to the govemment fiscal system and is generailly used for specified purposes such as salaries of state employees, administrative ex- penditures, and operating expenses. Extrabudgetary revenue, which is gener- ated by the business operations and other activities of township governments, is at the disposal of the township govemment. As the state budget can only provide subsistence to township govemments, extrabudgetary revenue be- comes the guarantee foir their effective functioning. THE INITIAL SITUATION. There is a widespread refusal to acknowledge that township public finances existred before the gradual establishment of the for- mal system in recent years. This may be correct theoretically or legally, but not in reality. Township public finances were brought into being by the inad- equacy of funds allocated by the central government from its budget. The state invested much less in agriculture than in the urban industrial system and put only nominal amournts into rural communities. Therefore, even if the township govemment (called the people's commune at the time) re- The Role of Township Governments in Rural Industrialization 345 frained from ambitious public undertakings, it could hardly make ends meet with the funds allocated by the state. During the people's commune period and even earlier, the community government collected funds locally for the central government. In retum, the central government allocated budgetary funds to the community govem- ment through a top-to-bottom process. In this sense the township govem- ment could not be regarded as having independent finances and budget. But it was, after all, a grass-roots organ of state power, and if it did not have some discretionary resources at its disposal, it could not function effectively. So, during the prereform period the township government always controlled some resources for meeting local needs, and it used them for exchanges with lower-level communities-brigades and production teams. This kind of ex- change brought the township govemment some financial income outside the state budgetary system. Some exchanges were in kind rather than in money. For instance, when the local community wanted to build a road or dig an irrigation canal, the township government might offer its sponsorship, and the lower-level com- munity groups would provide the labor. When the township government convened a meeting, it usually called on production teams to provide food for the participants. The township govemment gave not money but the time and energy for organizing projects, and the lower-level groups contributed inputs to undertakings that would advance their own development. But this primitive financial relationship could hardly meet the needs of ambitious township governments, which strove to organize production themselves or to control some profitable entities. THE TRANSFORMATION TO MONETARY FORM. In their efforts to transcend the primitive financial relationship, township govemments did not have many choices. Most township governments had to rely on income from agri- cultural production for primitive accumulation, but agriculture earned rela- tively low financial returns and could provide only limited financial support for the development of rural communities. Capital accumulation in rural communities was adversely affected by the state monopoly on the purchase and marketing of farm produce, by low prices for agricultural products, and by the segregation of the rural and urban economies. The need to make up for the low income from agricultural production gave rise to repeated impulses for rural industrialization. By 1977 Wuxi, after ex- periencing three failures, was achieving the best results of any of China's 2,100 counties in running commune and brigade industries, and the income from these enterprises became the chief source of funds for the county's rural communities. During 1971-78, Y148 million in profits from these firms went into agricultural machinery, bridges and power stations, field terracing, and other agricultural investments. Rural industrialization was obviously the best 346 SONG LINA AND DU HE way for township governments to make up for the deficiencies in their reve- nue and to change the primitive financial relationship based on the exchange of labor and goods into a mionetary relationship based on the profits of com- mune and brigade enterprises. Township Property After 1949 the property system for rural communities underwent a series of changes-land reformrr, the cooperative movement, and, in 1958, the peo- ple's commune movernent. Under the people's communes, the land of rural communities (including private farm plots and farmers' housing plots), as well as large agricultural equipment, became the property of the three levels of rural community government, with the commune at the top. (Farm tools, household necessities, and farmers' houses were exempted from communal ownership.) Since agricultural accumulation was too slow and could not meet the demands of rural development, even in the earlier days of the people's communes some nonagricultural assets came to be included in the public property of rural communities-that is, in the assets of commune and brigade enterprises. But although communities strove to develop rural industry, mac- roeconomic conditions prevented a wave of rural industrialization from oc- curring. When the PRS was adopted, the property of the people's communes was divided. The right to use land was accorded to farm households, ancl in most areas agricultural fixed assets and large farming implements formerly owned by the collectives were dispersed. Table 16-2. Fixed Assets of Industrial Enterprises, Wuxi, 1975-86 (millions of yuan) County Township Vilage Year enterprises enterprises enterprises 1975 50.18 27.18 21.03 1976 43.46 41.87 29.47 1977 72.41 57.32 40.96 1978 80.87 77.82 52.34 1979 92.78 102.15 64.45 1980 105.22 133.61 85.73 1981 119.95 171.96 106.10 1982 138.05 194.25 129.31 1983 167.72 231.83 161.05 1984 203.07 280.40 191.53 1985 225.36 459.17 347.95 1986 262.62 613.10 479.50 Note: Assets are valued at original purchase prices. Source: Information from fieldwork. The Role of Township Governments in Rural Industrialization 347 The share of commune and brigade enterprises in the assets of rural com- munities, however, did not decrease but increased to varying degrees. The original value of the fixed assets of township enterprises in Wuxi was Y27 million in 1975, but by 1986 the figure had risen to Y613 million, a nearly twenty-three-fold increase. In 1985 the fixed assets of township enterprises in Jieshou were worth Y9.8 million-1.9 times the 1980 figure of Y5.2 mil- lion. After 1979 the value of fixed assets of township enterprises in Wuxi exceeded that for county industrial firms, and by 1985 the fixed assets of in- dustrial village enterprises had overtaken those of county firms (table 16-2). The Impact of Changes in Community Property The growth of agricultural productivity and the rapid transfer of surplus farm laborers to nonagricultural activities as a result of the PRS paved the way for rural industrialization by making available the necessary factors-labor, capi- tal, and land. The changes in the composition of community property had several consequences for township governments. First, most township governments did not control much industrial prop- erty, and when agricultural property was dispersed under the PRS, they experi- enced a short-term "anemia" in their extrabudgetary finances and a weaken- ing of their administrative functions. For example, when we visited Wenzhou Prefecture, Zhejiang Province, in 1983, nonagricultural activities were booming, income per capita was rising rapidly, and capital accumulation in rural communities was high. But since the communities' old property had been dispersed and their new property was not yet properly organized, there was a great deficit in government finances. The result was that township gov- ernments became impotent in performing their administrative functions, and their managerial functions were greatly reduced. At that time, basic facilities and public works in the townships of Wenzhou Prefecture were rather back- ward, considering the rate of capital accumulation. Farmers were building three- and four-story houses with kitchens and bathrooms, but their kitchen slops were running in the streets for lack of sewers. Cultural, public health, and other public undertakings were lagging behind other areas. Township governments could not even organize projects but had to entrust the task to popular businessmen and venerable elders. Second, the PRS altered the composition of property under the township governments' control, and their behavior changed accordingly. Their terms of reference shifted from agriculture to industry. The enlarged proportion of industrial wealth in township property further increased the financial power of rural communities. In 1985 in the sample townships of Wuxi the average value of output of industry was 11.3 times that of agriculture. Even in Jieshou, where industries were launched rather late, the value of output of industry was 2.8 times that of agriculture. The share of industry in township government revenue increased with the rise in the share of industrial prop- 348 SONG LINA AND DU HE erty. Income from industrieaI production is now the chief source of township revenue in most areas, and income from agriculture accounts for only 19.1 percent of sample townshipis' extrabudgetary revenues, on average. These consequences, although totally different, are both related to the general trend toward industrialization. The introduction of the PRS was a nec- essary precondition for rural industrialization and opened up prospects for de- velopment; the increased proportion of industrial property in the assets of local communities aff'ected the form of industrialization and the degree of community govemment involvement in the process. After property relationships in local communities have taken their initial shape and the industrialization process has been launched, it is necessary to protect community public finances. In the absence of a strict tax system, the township govemment can safeguard public finances only by participating di- rectly in the management of firms-that is, by owning a portion of the community's property, independent of any interest group. Otherwise, com- munity public finance becomes an empty word. Rural communities benefit financially from industrialization in two ways. First, township government revenue is increased by the taxes paid by the en- terprises, which are at leastc in part retained by township govenmments for their own use. Second, the township govemment draws income directly from enterprises under its control. Under China's specific conditions it is not feasi- ble for township governments to meet all their financial needs from taxes on enterprises, and most tovvnship govemments own and manage enterprises so that they will have a direct source of income. The Bifurcation of Township Government Functions In the process of rural industrialization the township government took on a twofold role-as community administrator and as enterprise owner. 'Whether the township government can function effectively in the first role depends largely on the level of revenue at its disposal. Township governments own certain assets as a means of' achieving the objectives of community public finance. Hence the dual function of township govemments inevitably affects rural industrialization. The behavior of every township government leader has a twofold motive: to discharge govemment duties and to further personal interests. IThis does not mean that most township govemment leaders promote rural industrializa- tion merely for personal gain. Rather, personal and public motivations are intertwined. One reason for this has been the past overemphasis on personal sacrifice and the neglect of deserved material rewards for grass-roots officials. Another reason is the relative decline in the regular incomes of these officials in recent years. When their pay was based on work-points in their home pro- duction teams or came from extrabudgetary revenue, their incomes wvere gen- erally higher or more stable than those of ordinary commune members. But The Role of Township Governments in Rural Industrialization 349 as rural industrialization and economic development have progressed, offi- cials' incomes have not kept pace with those of most local residents. In gen- eral, the salaries of township govemment officials are lower than those of village cadres, and those of village cadres are lower than those of township enterprise employees. The Focal Point of Township Governments' Work Analysis of sample townships shows that most township governments give primary attention to the enterprises they run. This is because the relationship between the govemments and their enterprises is one of broadly based inter- dependence. In our sample, direct and indirect investment by township gov- emments accounted for more than 30 percent of the startup funds for town- ship enterprises. The profits handed over by township enterprises in the sample counties constitute more than 38 percent of township government revenues. As administrators of rural communities, township govemments also obtain revenue through taxation. In the sample, remittances by town- ship enterprises account for 43.6 percent of township revenues and tax pay- ments by enterprises under other forms of ownership for 14.3 percent. Since the township government owns the assets of its enterprises, it has the right as well as the duty to manage this property. Moreover, govemments must participate directly in management to realize the goals of their invest- ment in the enterprises. The township govemment hence has a status similar to that of a board of directors but without all of its functions. The Difficulties of Balancing Roles The township government's role as not only the organizer and manager of rural industrialization but also a participant in it can distort its discharge of government functions. A good township govemment must play different roles simultaneously. Not only does it have to adjust the relations between its enterprises and other interest groups in the community, but it must also straighten out relations among its enterprises (for usually there is more than one firm directly under the township government) and among the owners, managers, and workers of each enterprise. The administrative functions of the township govemment are greatly weakened by its status as owner and beneficiary of rural community property. Township governments have only indirect control over private enterprises, since they have no share in the firms' assets and cannot expect much revenue from them. As long as these enterprises pay taxes according to regulations and abide by the law, township governments exercise little authority over them. There is thus a difference in the degree of protection and control within the community. For example, 79 percent of the sample township enterprises in Wuxi could 350 SONG LINA ANE) DU HE obtain raw materials at bellow-market prices through the township govem- ment, but only 5 percent of the private enterprises could. In the Jieshou sam- ple, low-priced raw materials were distributed only to township enterprises. Township enterprises in the Wuxi sample received 21 percent of their total capital as allocations by township govemments, whereas other enterprises re- ceived none; in the Jieshou sample, such allocations constituted 26 percent of township enterprises' total capital, but other enterprises received none. Township Governments' Influence on Township Enterprises The township govemrnent protects and strictly controls its own enterprises because it depends on them to a great degree. The dual fumnctions of township govemments are mirrored in the twofold character of local protecztionism. Protection of private enterprises is motivated by the township government's role as govemment administrator. The township govemment adopts protec- tive measures for private enterprises only when these firms come into conflict with entities from outside the community or when their development is hin- dered by unfavorable policies of higher levels of govemment. But its protec- tion of township enterprises is motivated not only by its administrative role but also by its status as owner. Preferential Treatment Since township enterprises are launched with financial assistance from town- ship govemments, these firms enjoy advantages from the beginning. In the samples, township enterprises exceed other types of firm in the TVP sector in capital, fixed assets, and snumber of employees and thus have higher pro- ductive capabilities. Their average value of output per worker is 1.64 times that of village enterprises, 2.93 times that of partnerships, and 2.86 times that of individual firms. Moreover, township govemments give their enter- prises priority in obtaining materials in short supply. Township govemments also provide township enterprises with an intangi- ble resource-the use of the governments' credibility. Because of their close association with township govemments, township enterprises enjoy higher status in the eyes of specialized branches of govemment (such as banks, in- dustrial and commercial administrative bureaus, and tax and supply depart- ments) than enterprises under other forms of ownership. Restrictions Despite these advantages, the efficiency of township enterprises is lower than or at least no higher than that of other kinds of TVPS (see chapter 18). There are several reasons for this, related to control by township govemments. First, township govemments' dependence on township enterprises for income The Role of Township Governments in Rural IndustrialiZation 351 increases the enterprises' management costs. In addition to paying taxes, township enterprises bear part of the govemments' consumption expenses. As the "purse" of the township government, the enterprises have to receive guests, send presents, and pay expenses for meetings, travel, and visits on behalf of the govemment. Although no detailed data on these expenditures are available, the seriousness of this problem was evident from our talks with township enterprise leaders. Second, since township enterprises are the chief means by which the town- ship govemment attempts to achieve the comprehensive goals of the commu- nity, maximization of enterprise profits has rarely been pursued wholeheart- edly. Almost all township govemments seek to use rural industrialization to employ surplus labor from agriculture. Because of the stability of township enterprises, rural residents prefer to work in them, as is shown by responses to the Worker Survey Questionnaire. The township govemment usually de- cides the number of workers to be recruited by township enterprises, and en- terprise directors cannot cut personnel at their own discretion even though they are aware that overstaffing adversely affects efficiency. In interviews in Wuxi many township enterprise directors admitted that there was overstaffing, in some cases as high as 50 percent (see chapter 14 for some examples). Third, since the overall development objectives of rural communities have largely been shifted onto the shoulders of township enterprises, many of these firms cannot accumulate very well. This is especially true in Wuxi, where in 1986 industrial income was nearly ten times that of agriculture. To stimu- late agriculture, township govemments demanded that enterprises subsidize farmers. In Dongjiang township in 1986 more than 20,000 mu of land re- ceived a subsidy of Y12 per mu at the expense of township enterprises. Moreover, township industrial corporations (TICS) under Wuxi's township govemments strictly control the operations of township enterprises, espe- cially their distribution of profits. TICS have the right to make plans for the development of all township enterprises, to adjust the reinvestment rate, and to launch new firms. They plan investments to establish new township enter- prises and renovate old ones. Some enterprises may have to wait a long time to obtain the funds needed to increase production or to switch from products in little demand to other lines. These firms will witness further deterioration in their capabilities for accumulation and production. Fourth, the impersonal character of property ownership in township enter- prises means that management is kept at a low level and financial control is lax. Although govemment leaders are not permitted to derive personal monetary profit from the enterprises, the township govemment's right to par- ticipate in the management of its enterprises may become an opportunity for corruption. Officials may encroach on the property of township enterprises in nonmonetary ways to meet their personal needs. Most directors of township enterprises have a strong desire to develop their 352 SONG LINA AND DU HE firms, but they are under the control of the township govemment and wield no all-round management power. Understandably, they lack the attitude of the master, and the result can be serious waste. For instance, in Wuxi the waste rate for materials is worse for township enterprises than for village en- terprises. The practice of workers' showing up but not doing much work is far more serious in township enterprises than in private enterprises and greatly increases the former's costs. Reasons for Differences in the Behavior of Township Governmernts Even under the same rnacroeconomic and policy circumstances, township governments in differerit localities behave quite differently. Several exoge- nous factors influence the behavior of township governments in different areas. 0 Natural resource endowrnent. Shangrao is rich in natural resources, which exceed those of the other three counties in amount and variety. The county has deposits of coal, barite, lead, zinc, silicon dioxide, sulfur, lime- stone, and other minerals. In Wuxi mineral deposits in sample towvnships amount to only 450 tons, or 0.38 kilograms per capita. Jieshou and Nanhai have scarcely any mineral resources, but Jieshou does have fairly abundant agricultural resources; itn sample townships arable land per capita is much larger than in Wuxi, and. average forest coverage is 2.86 times Wuxi's. Arable land per capita is 0.3 mu in sample townships of Wuxi, 1.1 mu in Jieshou, 0.9 mu in Nanhai, and 1.I rmu in Shangrao. Thus, Shangrao has the best mineral resources of the four, and Jieshou has better agricultural resources than Wuxi and Nanhai. Yet Wuxi and Nanhai are far more developed eco- nomically than Shangrao and Jieshou. * The level of economic development. Regional differences in economic de- velopment are shown by the income per capita of the rural population, which is a relatively comprehensive index. In Wuxi in 1985 the figure was Y754, in Nanhai Y1,029, in Jieshou Y285, and in Shangrao Y322. Our investiga- tions confirmed these differences; the standard of living for typical house- holds was highest in Nanhai, with Wuxi second and Shangrao and Jieshou trailing behind. * Employment pressure. A rough indication of employment pressure is ara- ble land per member of the agricultural labor force: the average is 6.1 mu in Wuxi, 2.9 mu in Nanhai, 2.6 mu in Jieshou, and 4.0 mu in Shangrao. But these numbers mask differences in the stages of development of rural in- dustrialization and consequent differences in the degree to which agricultural workers have moved into industry. For instance, the figure for arable land per farmer in Wuxi is high because large numbers of rural population have already shifted to nonagricultural employment. A more accurate measure of The Role of Township Governments in Rural Industrialization 353 employment pressures and requirements in rural communities is the annual percentage change in the industrial and agricultural labor forces, shown here for 1978-85. (The data are derived from fieldwork.) Wuxi Jieshou Nanhai Shangrao Industrial labor force 18.8 32.4 12.6 12.3 Agricultural labor force -12.9 -10.5 -12.7 2.5 These figures indicate the capacity of rural industry to absorb agricultural workers. All counties except Shangrao have managed to shift labor from agri- culture to industry at a fairly rapid rate. In Shangrao the agricultural labor force actually increased between 1978 and 1985. * Quality of the labor force. Some characteristics of the rural population in the four counties were mentioned in chapter 4. The figures below, from the Enterprise Quantitative Questionnaire, show the average number of technical personnel and supply and marketing staff in sample firms. The much higher numbers of the latter in Wuxi and Nanhai are striking, as is, in particular, the low ratio of supply and marketing to technical personnel in Shangrao.' Type of personnel Wuxi Jieshou Nanhai Shangrao Technical 0.53 0.19 0.21 0.66 Supply and marketing 1.15 0.30 1.0 0.21 The differences in microenvironments and in conditions for industrialization reviewed above inevitably affect the decisions and behavior of township gov- ernments and thus the development of rural industries. The Impact of Township Government Behavior on Local Industrialization The ability of township govemments to realize communal targets-con- ditioned partly by the factors discussed in the last section-affects their behavior and to a certain extent determines regional approaches to rural in- dustrialization. In this section we examine broad differences in township gov- ernments' participation in management and control of township enterprises. Management and Control According to responses to the Township Leader Questionnaire, township governments in all four counties are deeply involved in the management of TVPS, especially township enterprises. The index of participation in manage- ment by sample township governments is 0.95 in Wuxi, 0.80 in Jieshou, 0.90 in Nanhai, and 0.78 in Shangrao. We see such participation in management as a positive concept-as assistance, not as predatory meddling. In Wuxi and Nanhai, known for their advanced economic development, the degree of par- 354 SONG LINA AND DUJ HE ticipation of township governments in local enterprises is higher than in Jieshou and Shangrao. The average degree of control over township enterprises exerted by town- ship governments is 0.39 in Wuxi, 0.44 in Jieshou, 0.38 in Shangrao, and 0.39 in Nanhai. The high figure for Jieshou, which boasts a traditional, full- fledged system of politicatl mobilization, tallies with other evidence that TVP industrial development there relies mainly on govemment mobilization and instructions. The county government in Shangrao has also had an important role in local industrialization, but it has been ineffective because of the poor quality of township officials: the low degree of control over township enter- prises is not a rational choice but a situation the government has found no way to avoid. In Wuxi TVCE industries were launched much earlier, and a set of regulations governing the relationship between township governments and their enterprises has already taken shape. Hence the responsibilities, powers, and interests of the government and the enterprises are clearly de- fined. Of the four counties, Nanhai has the most characteristics of a free economy. There is a balanced ownership structure in the local TVP sector, firms rely more on the market than on the township government, and town- ship govemments have allowed more autonomy to enterprise management. Management Prioties The ranking in the Township Leader Questionnaire of respondents' priorities in supervising enterprises reflects the patterns of township government man- agement (or control) in the four counties. Priority Wuxi Jieshou Nanhai Shangrao First Investment Personnel Personnel Production planning Second Production [nvestment Investment Personnel planning Third Personnel Production Production Investment planning planning In analyzing the above choices, the stages of development in the four counties should be kept in mind. For example, sample township governments in Wuxi put personnel in third place, but this does not mean that they pay little attention to the appointmnent of township enterprise leaders. A com- plete set of procedures and standards for appointing and dismissing entelprise leaders has already evolved there, and staffing is no longer the focal point of the township government's 'work. Now enterprises are vying for power over investment and production, and township governments have had to strengthen their control in these areas. In Jieshou, sample township govern- ments put personnel in first place, which reflects the county's reliance on The Role of Township Governments in Rural Industrialization 355 cadres to push TVP development. This situation in Jieshou can be regarded as rational, since in the early stages of rural industrialization the township government has to control enterprise personnel to make its strength felt. The stress on personnel may be typical for localities that are latecomers in the drive for rural industrialization. Sometimes township governments' priorities for supervision of township enterprises are simply illogical. When an area has just begun to see the need for local industrialization and is still in the experimental stage, the behavior of both the township governments and local enterprises may be confused, and their choices of priorities may be irrational. For instance, in Shangrao, sample township governments put power over production planning in first place. But in reality township governments in Shangrao are more concerned about personnel issues than about production planning, and in any case strict control over the latter is not really a good management method. A common feature of the four counties is that, of fifteen indexes for ap- praising managerial behavior listed in the questionnaire, most township lead- ers focus on three: personnel, investment, and production planning. Hence township governments are concerned with some aspects of enterprises' inter- nal management and decisionmaking but give little attention to marketing or sales. As chief owner of the assets of township enterprises, the township govemment, as part of its function, should oversee the operational behavior of the enterprise in the same way as boards of directors in Western countries do. Township governments and their enterprises do appear to have appropri- ate relations with regard to duty, power, and interest. But close examination reveals that governments' duty to their enterprises is motivated by their own financial needs and by the pursuit of wider community goals. Power over personnel most determines relations between township govern- ments and their enterprises, as responses to the Enterprise Director Question- naire show. Since profit maximization by the enterprise conflicts with pay- ment of taxes to the township government, the township government commonly restricts the actions of enterprise management through its power over personnel, and enterprise managers thus become a part of the govern- ment administration to a certain degree. Control of production plans by the township govemment is equivalent to controlling the value and quantity of output and thus achieving income targets for the enterprise and the township government. Control over investment can restrict enterprise expansion and direct resources toward other targets of the community and the township government. Internal control of township enterprises by the township government may restrain the enterprises' expansion and development because the government frequently sacrifices township enterprises' interests for the sake of other goals. But since the realization of community goals must be based on the develop- ment of local enterprises, township governments find themselves in a di- 356 SONG LINA AND DUt Hl lemma. How township govemments, influenced and limited by the specific conditions of their areas, deal with this dilemma affects the development of rural industries. Development Patterns Sample township governments have been able to secure for their enterprises some raw materials that are in short supply. In Wuxi 55 percent of sample township govemments received allocations of raw materials from supply de- partments at the county level or above. The other 45 percent were able to secure raw materials on their own. Only 20 percent of the raw materials allo- cated to township governments in Jieshou came from rural communities; the rest were allocated by the county govemment. In the sample in Nanhai the county government allocated raw materials to only two township govem- ments, which passed them on to township enterprises. In Shangrao 78 per- cent of the township governments received inputs from the county govem- ment, while the remaining 22 percent had to secure raw materials and other inputs on their own. Township govemments in Nanhai exercise control over the management of their enterprises, bui: local resources are limited, and the governments do not have ample materials to trade with township enterprises. For this reason TVPs, including township enterprises, chiefly depend on the market for survival. TVPS in Wuxi are mostly TVcEs. Township govemments exercise control through rough planning, and. enterprises depend for their survival partly on the market and partly on planning. In Jieshou agriculture predominates, and even now a large proportion of local revenue comes from that sector. Local industrialization bore the stamp not only of political mobilization but also of obvious spontaneity. This can be seen from the ownership of local TVPS. In the Jieshou sample a large num- ber of TvcEs are actually individually run, but because individual enterprises have difficulty in requisitioning land and for other reasons, a number of them display the signboards of township enterprises. They do not have close rela- tions with township governments, but they can be taken over by the local govemment once they grow to a certain scale. (For a detailed analysis see chapter 7.) Shangrao has a weak industrial base and underdeveloped rural cornmuni- ties. Township govemrnents strictly control the amount of profits kept by township enterprises, causing great difficulties for the firms' development. To make up for enterprises' losses, township govemments trade local resources for materials in short supply and allocate to the enterprises raw materials bought at preferential prices from county supply departments. But since the capital accumulation of the communities is at a low level and township gov- ernments cannot cut down their expenses, the profits delivered by township The Role of Township Governments in Rural Industrialization 357 enterprises are used to a large extent for the expenses of township govern- ments or are passed on to other community interest groups. Only two sample township governments in Shangrao reported spending part of the income from township enterprises to expand the firms. In Wuxi, by contrast, sample township govemments returned, to varying degrees, profit remittances from township enterprises so that the latter could keep reinvesting and developing. Conclusions Through descriptions of the behavior of township govemments in different microeconomic environments, this chapter has illustrated the roles of town- ship govemments in rural industrialization. In addition, it has examined how the desire of township govemments to increase their financial revenues has become one of their main motives for promoting rural industrialization. Some observers hold that township govemments should cut down their management and control of local township enterprises and develop a new relationship with them through shareholding. But this view neglects the basic fact that once local enterprises are divorced from township govern- ments, local communities will not be able to achieve their goals. The town- ship government-the main force in achieving local community goals-will not function effectively because it will lack funds; revenue from taxation can never be enough to meet the overall development needs of the community. In addition, the tax system in China has never been perfect, and tax evasion by enterprises will further reduce its effectiveness. Therefore, for quite a long time it will be necessary for township govern- ments to directly control some enterprises and to realize community goals through concerted efforts with enterprises. But township governments must maintain a proper balance between their roles as government administrators and as owners and operators of enterprises, to prevent a vicious cycle in the circulation of community wealth as a result of excessive levies on enterprises to meet community public finance needs. Note 1. The higher average number of technicians in Shangrao sample firms probably reflects a looser definition of this category and the predominance of township enterprises in the sample. Correspondingly, the low number of technicians in Wuxi firms in relation to their size and indus- trial structure may be attributable to relatively narrow criteria for classifying employees as techni- cal personnel. 17 Why Industrialize? The Incientives for Rural Community Governments William A. Byrd and Alan Gelb Local community governments at all levels-townships, towns, villages, and sometimes production teams--have played a crucial role in stimulating and supporting the rapid growth of China's TVP sector. Enterprises owvned by these govemments account for the bulk of TVP industrial output (see chapter 9), and where private firms have flourished, it has been with community gov- emments' support and encouragement. Without the deep involvement of community govemments, China's TVP sector could not have grown nearly as rapidly as it did in the late 1970s and early 1980s.' The incentives and motivations of community government leaders have impelled them to estab- lish and finance community enterprises and in many areas to provide support for private enterprises. Where TVP development has not been successful, this has often been despite the best efforts of many community govemment lead- ers and officials. The short-term and long-term implications of reliance on local communities by the most dynamic part of Chinese industry must be considered when evaluating the future prospects of the TVP sector. In this chapter we analyze Chinese community govemments and their leaders as self-interested entities whose actions are geared toward maximizing their own benefits. This approach is applied to community governments, which can be viewed as interest groups with integrated, consistent objectives, and to individual government decisionmakers. The assumption that the gov- ernment has its own independent interests that it strives to further is one of the key postulates of public choice theory.2 This perspective contrasts sharply with the traditional view of mainstream Western economics that gov- emment is benevolent and that the main issue is designing and implementing economic policies that are optimal for the nation as a whole-assumptions shared by the neoclassicatl and Keynesian schools, among others, despite their different policy prescriptions. Although we borrow one of the main postulates of public choice theory, in many respects our conclusions differ, owing to our subject. We l.ook at rural communities with populations of sometimes only a few thousand people 358 Why Industrialize? The Incentives for Rural Community Governments 359 and always fewer than 100,000. Chinese community governments function in an environment of population and factor immobility and with a certain degree of independence in economic policy. They are thus akin in many ways to nation-states, albeit tiny in size. But they differ in fundamental ways from national governments: they have virtually no discretionary authority to tax, and they cannot print money or borrow. One of their main economic policy instruments is the establishment and development of community enterprises. In contrast, govemment ownership of enterprises in industrialized market economies is typically limited to a few sectors or is the result of historical accidents. The Main Characteristics of the TVP Sector's Environment We start by outlining some key features of the environment in which China's TVP sector functions. These include population and factor immobility, spatial inequality in rural incomes and industrialization, ineffective redistribution of income through the fiscal system, and compressed personal income differen- tials within communities and within TVPS themselves. The lack of mobility of factors and funds forces community governments to rely largely on their own resources in developing the TVP sector. Population and Factor Immobility The bulk of China's rural population is rooted in local communities, with little prospect for permanent migration elsewhere. Tiny numbers of students who are successful in their university examinations leave and usually end up in the cities; more important, increasing numbers of rural people go to the cities for "temporary" jobs in construction and other service activities. In some rural areas such as Nanhai there is an open labor market that attracts temporary migrant workers from poorer rural areas. In Jieshou successful farmer-entrepreneurs are allowed to move their production facilities and resi- dences into town, switching their household registration status to that of "long-term urban resident." But both rural-urban and rural-rural mobility re- main limited, and for many analytical purposes it is reasonable to assume complete population immobility. Labor is somewhat less immobile than population because of opportunities for temporary employment in cities and in other rural areas. Nevertheless, mobility is restricted, and most jobs in the TVPs are still filled within the rural community. Capital is also immobile across rural communities, especially townships. Financial flows through the banking system are limited by the "cgap" system of credit planning and the widespread perception that financial resources mobilized in a rural community should be used for investment within that community (see chapter 9). Flows of funds through the fiscal sys- tem are far from sufficient to offset imbalances in the financial resources of 360 WILLIAM A. BYRD AND ALAN GELB different localities, and liows of capital outside the banking and fiscal systems are still small. Finally, movement of managerial, entrepreneurial, and techni- cal human resources in rural areas is limited. Although technical personnel can be attracted from other localities by high salaries and benefits, entrepre- neurs are even more immiobile than ordinary labor, since it is administratively much easier for them to locate their factories in their home commnunities than elsewhere (see chapter 9). This all adds up to an environment in which flows of goods are relatively free (see chapter 5) but flows of key factors of production and of population are restricted. The policy instruments at the disposal of community govern- ments, particularly at higher levels (townships and towns), create an eco- nomic situation that resembles a group of ministates under a free trade regime with factor immobility an-d fixed exchange rates. (An altemative but not in- consistent perspective is to view each rural community as a fixed-membership cooperative firm, with the added complication that lower-level communities are "nested" within higher-level ones.) Dimensions of Spatial Inequality Under these circumstances and given the tremendous imbalances in material, financial, and human resources among localities and regions, substantial in- equality in factor rewards, particularly in personal incomes, is to be expected. Although it is beyond the scope of this chapter to analyze this phenomenon in detail, summary statistics illustrate the magnitude of spatial inequality in rural incomes. Table 17-1 shows the variation in average income per capita among U.S. Table 17-1. Spatial Income Inequality among Provinces of China and U.S. States Highest Lowest (percentage (percentage Coefficient of Area Year of mean) of mean) variation (percent)' China urbanb 1983 137 75 13 China ruralc 1980 194 70 28 China ruralc 1985 195 62 31 United States' 1981 134 72 14 United States' 1986 140 70 17 a. Standard deviation divided by mean. b. Average per capita income of urlban residents in the capital cities of each of the t wenty- nine provinces, based on household surveys. c. Average per capita income of rural residents in each province, from household surveys. d. Average per capita income by state. Sources: State Statistical Bureau (1985c), p. 62; (1986b), p. 202; (1987c), p. 204; U.S. Gov- ernment (1983); Washington Post, August 21, 1987, p. C2. Why Industrialize? The Incentives for Rural Community Governments 361 states and, separately for urban and rural areas, among provinces in China. China's state wage system holds down income inequality among cities to lev- els no higher than state variations in the United States, but rural income inequality among provinces is considerably greater than variations among U.S. states. Regional income inequality in the United States has declined greatly in the past fifty years as a result of improvements in transport and a great increase in interregional mobility. Table 17-2 presents similar statistics for counties in two Chinese provinces and in several U.S. states. Inequality among counties in the United States is somewhat greater than among states; the amount and quality of public ser- vices that local govemments provide varies, and the "Tiebout hypothesis" (Tiebout 1956, 1961) suggests that richer people congregate in localities that provide valued public services. Moreover, county data pick up differences be- tween the rich suburbs of large cities and poorer rural areas. In Jiangsu and Anhui provinces there is relatively little spatial inequality in urban wages, but rural income inequality is substantial-only slightly less than cross- provincial income inequality and greater than inequality among U.S. states. Income inequality within Chinese counties is usually much less severe than inequality among counties in a province. For example, in Nanhai in 1985 income per capita for the rural population in the richest township was 108 percent of the mean for all townships; in the poorest township it was 84 per- cent of the mean; and the coefficient of variation was only 7 percent. In- equality is lower within counties because resource endowments for communi- ties within a county are generally more equal and because county govemments try to hold down the gap between leading and lagging town- ships. Table 17-2. Income Inequality in Selected Chinese Provinces and U.S. States, by County Highest Lowest (percentage (percentage Coefficient of Area Year of mean) of mean) variation (percent) Anhui urban' 1984 128 87 8 Jiangsu urbana 1985 127 85 10 Anhui ruraP 1984 163 49 23 Jiangsu rural' 1985 164 62 23 Michigan 1981 164 60 17 Mississippi 1981 163 65 16 Oregon 1981 137 81 12 South Dakota 1981 183 39 21 a. Average wage per member of the urban work force. b. Average per capita income for the rural population. c. Average gross value of agricultural output (GVAO) per member of the rural population. Sources: Anhui Jingii Nianjian 1985; Jiangsu Jingii Niargian 1986; U.S. Government (1983). 362 WILLIAM A. BYRD AND ALAN GELB Ineffective Redistribution Mdechanisms The degree of spatial income inequality in China's rural areas is especially surprising given China's strong ideological commitment to equality and the powerful redistributional instruments available to the state. But whereas gov- ernment influence over urban wages effectively limits inequality in urban in- comes, there is at present no comparable instrument for smoothing differ- ences in rural incomes.3 Considerable redistribution of income does occur through the fiscal system. Richer, more developed provinces generate tax revenues in excess of expend- itures and turn over surpluses to the central government, which provides sub- sidies to provinces in which local revenues fall short of expenditures. Similar fiscal redistribution occurs within provinces. Although substantial flows of resources are involved, they are far from sufficient to offset the underlying differences in resource endowments and levels of development. Table 17-3 shows variation in average per capita budgetary revenues and expenditures among provinces in China and among counties in Anhui and Jiangsu prov- inces. Variation in expenditure per capita, although much less than in reve- nue per capita, remains Substantial, and there is significant correlation be- tween the two at both the pTovincial and the county levels. If the three provincial-level cities of Beijing, Shanghai, and Tianjin are excluded, the coefficient of variation of' expenditures per capita becomes even higher than that for revenues, and correlation between the two remains high. Thiis sug- gests that there is little if any fiscal redistribution between rich ancl poor "mainstream" provinces in China. (Data were not available for Xizang, a case that probably also involves considerable fiscal redistribution.) In any case, budgetary statistics do not reflect the large share of local public and social expenditures financed from extrabudgetary sources. Such spending is far greater in the richer, more developed localities. Moreover, a large part of budgetary spending, particularly at the lower levels, merely supports the government apparatus and pays for basic social services rather than directly raising rural incomes or financing TVP development. Finally, there is much anecdotal evidence that money transferred to the poorer areas through the fiscal system is not used with maximum effectiveness. Egalitarian Income Distribution within Rural Communities Somewhat counterbalancing the high degree of spatial inequality of income is the tendency toward compressed pay and income structures within rural communities and TVPS. Other things being equal, the smaller the community involved, the greater is this compression. In localities such as Wuxi the town- ship government uses administrative controls to hold down variation in pay among TVPS within the township. Income variation among communities and compression within them is shown by data from the Worker Survey (see Why Industrialize? The Incentives for Rural Community Governments 363 Table 17-3. Average Per Capita Budgetary Revenues and Expenditures Correlation Highest Lowest Coefficient of coefficient (R) for (percentage (percentage variation revenues and Item of mean) of mean) (percent) expenditures Among provinces China, 1985 Revenues 869 28 169 0.72 Expenditures 251 40 65 China, 19851 Revenues 248 56 44 0.69 Expenditures 222 49 48 Among counties Anhui, 1984 Revenues 1,264 25 152 0.58 Expenditures 239 53 46 Jiangsu, 1985 Revenues 391 17 86 0.69 Expenditures 176 63 24 a. Excludes Beijing, Shanghai, and Tianjin municipalities and Xizang Sources: Anhui Jingji Nianjian 1985; Jiangsu Jingji Nianjian 1986; State Council Economic, Technological, and Social Development Research Center (1986). chapter 13). About 60 percent of the variation in pay among sample workers is explained by firm dummy variables, which may be picking up intercommu- nity differences. This is especially striking given the deliberate oversampling of higher-paid employees (technicians, middle-level managers, and the like). Egalitarianism within communities is the result of a combination of fac- tors. Tightly knit communities with fixed memberships (often cemented by family and clan ties) would probably tend toward compressed personal in- come structures under almost any circumstances, and this tendency is strengthened by socialist ideals and policies. In the past the work-point sys- tem of collective distribution virtually ruled out significant inequalities in personal incomes within the communities where such distribution occurred (production teams and sometimes brigades). Moreover, in many areas wages of TVP workers were paid to their communities and were given to the workers themselves only in the form of community work-points, which prevented the industrial labor force from earning much higher incomes than agricultural workers. But probably the most important factor that limits intracommunity income inequality in rural areas today is the relatively egalitarian initial dis- tribution of the main factor of production, land, under the production re- sponsibility system (PRs). Egalitarian distribution of job opportunities in TVPS has also contributed to compression. An assessment of trends that affect income distribution within Chinese 364 WILLIAM A. BYRD AND ALAN GELB rural communities would require a separate study. Ideological constraints against high personal incomes appear to be gradually eroding-for example, the permissible gap between the pay of directors and workers in comrnunity enterprises is generally riising. Specialized households and private entrepre- neurs can eam even higher in,comes. But in some areas, functioning labor markets may hold down differentials between agricultural and industrial in- comes and reduce spatial inequality. Movements of labor may in part explain the considerable apparent reduction in rural income inequality among coun- ties in Anhui Province between 1981 and 1984.1 Whether relatively equal income distribution within rural communities offsets the effect of spatial inequalities is a question for research. From the perspective of this chapter, the different patterns within and among commu- nities are more important than the net result. As shown below, TVP develop- ment is a crucial determinant of communal development levels and living standards. Compression within communities means that the incomes of lead- ing community members (government leaders and enterprise directors) de- pend at least to some extent on the average income levels in the community. Self-Reliance in TVP Development Factor immobility and ineffective redistribution force communities to rely on their own resources (broadly construed) to develop the TVP sector. Hence TVP development is likely to be highly uneven among regions and localities, owing to their unequal resource bases. Table 17-4 provides some indication of the striking spatial inequality in the development of industrial TVPS. Any analysis of the motivations and incentives behind community government support for TVP development must come to grips with these huge gaps, for understanding the anatomy of failure is just as important as assessing the rea- sons for success. Flows of Funds between TVPs and Community Governments Figure 17-1 shows the financial flows among township firms and township and county institutions. Township enterprises pay direct and indirect taxes to the township tax collector; they pay management fees (in effect, a tax, not a payment for services) to their supervisory agency, the township indus- trial corporation (TIC); and they generate after-tax profits, part of which are reinvested and part of which are tumed over to the TIC. Remitted township enterprise profits flow upward through the system and are an important source of funding for township government salaries and overhead as well as for dis- cretionary spending. Budgetary flows are relatively large, but appropriations are mostly earmarked for fixed costs (wages and overhead) and for specific projects. Management fees pay for the administrative support structure for TVPs at township, county, and higher levels. The main sources of funds for Why Industrialize? The Incentives for Rural Community Governments 365 Table 17-4. Variation in Per Capita Value of Rural Indusal Output Highest Loiwest (percentage (percentage Coeficient of Area of mean) of mean) variation (percent) Among provinces China, 1980' 691 4 142 China, 1985 531 2 133 Among counties Jiangsu, 1984b 554 8 119 Jiangsu, 1985' 414 18 97 Anhui, 1984' 1,100 17 161 Among townships Wuxi, 1984 362 42 61 Nanhai, 1985 235 22 63 Shangrao, 1985 474 30 104 a. Based on 1980 gross value of industrial output (GVIO) and 1981 rural population figures by province. b. TvcEs only. c. GVIO for county as a whole. Sources: Anhui Jingji Nianjian 1985; Jiangsu Jingli Nianjian 1986; State Statistical Bureau (1987c), p. 14; and information from fieldwork. investment in township enterprises are retained profits, remitted profits used by the TIC for reinvestment, and bank loans. Township governments now retain for their own discretionary use a portion of above-quota budgetary revenue collected in their communities. This al- lows them to tap into part of the budgetary revenues generated by TVP devel- opment. Where the public finance system is in place at the township level, this revenue-sharing system has been formalized: townships in Wuxi are al- lowed to retain 8-12 percent of above-quota tax collections, and in Shangrao they keep 50-100 percent. In Nanhai the township public finance system has not been formally established, but townships are allowed to retain 1.5 percent of revenues collected locally within the plan and 20 percent of above-quota revenues. Salaries and benefits of township government officials come from three main sources: budgetary allocations for salaries and for part of the overhead for the township leader and a few other cadres; profits remitted from township enterprises; and the shared portion of above-quota tax revenues. The latter two sources are used for the wage and overhead costs for collective (as op- posed to state) cadres in the township government. Figure 17-1 also high- lights the circular financial flows among enterprises, their employees, and banks: township enterprises pay wages and bonuses to workers, who then de- posit money in banks, which then have more loanable funds, the lion's share of which is lent to local TVPS. 366 WILLIAM A. BYRD AND ALAN GELB Figure 17-1. Township Revenue Flows and Community Enterprises t + |Co0unty finance bureau- --t--bunty taxbra | _fogtltr ur e b au I For government *-----Remitted revenues-*-- I salaries arid so on ---Above-quota retention_* N Township _ x1 Twnship financoffice}J government Township ta oGEcej a s (salesa overhead)J Agricultiral tax wnship econonic Other taxes A- Budgetary spending commission I Discretionary spending [w corporanion Industrial-comnercial Bank loans N. and profit: taxes Bankl loans_l Investment allocations N t Bank deposits Reinvestment \ Profi_ . Township enterprise Township enterprise Township enterprise Fo__ I F F~~~~~~~~~__ Pay and b onuses Budgetary flows ----_._ Nonbudgetary flows - -.- I it ui i AL Management fees --------- n I n I!nIA Table 17-5 shows financial flows between community enterprises and com- munity institutions in two very different townships. Township A in Wuxi is the most industrialized township in China as measured by per capita gross value of industrial output (GVIC). Township B in Shangrao is one of the poor- est and least industrialized townships visited during fieldwork. The cornposi- tion of flows as well as their absolute magnitudes differ sharply. Most of the large amount of tax revenue collected in township A is remit- ted to the county and is passed upward through China's fiscal system. In township B, not only does the township keep virtually all tax revenues col- lected, but it receives a large allocation from the county for budgetary ex- penditures. Profits and managemnent fees from township enterprises constitute an important source of township government funds in township A, where sales revenue of TVCEs alone totaled more than Y200 million, but they are very small in township B. The burden on township B's enterprises, however, is paradoxically greater than in township A. Township B's firms lost nearly Y14,000 in 1986 but stil'l had to remit a considerable amount of money to Why Industrialize? The Incentives for Rural Community Governments 367 Table 17-5. Government Income and Expenditure in Two Townships Township A Township B (Wuxi, 1985) (Shangrao, 1986) Share of Share of Amount total Amount total Item (yuan) (percent) (yuan) (percent) Sources of income Tax revenues 13,428,000 56.4 193,500 46.4 Used for quota budget spending 450,000 1.9 185,000 44.4 Retained by township (discretionary) - - 6,800 1.6 Remitted to county 12,978,000 54.5 1,700 0.4 Additional budget allocation by county - - 209,800 50.4 Remittances by township enterprises 10,390,000 43.6 13,200b 3.2 Profits 8,320,000 34.9 n.a. n.a. Management fees 2,070,000 8.7 n.a. n.a. Retained by township 1,035,000 4.4 n.a. n.a. Remitted to county 1,035,000 4.3 n.a. n.a. Total 23,818,000 100.0 416,500 100.0 Available township resources 9,805,000 41.2 414,800 99.6 Percentage Percentage of resources of resources Income Tax revenues 450,000 4.6 191,800 46.2 Remittances from township enterprises 9,355,000 95.4 13,200 3.2 County budgetary allocations - - 209,800 50.6 Expenditures Investment in township enterprises 6,720,000 68.5 19,300' 4.6 Support to agriculture 900,000 9.2 26,000 6.3 Public and social services 1,050,000 10.7 285,700 68.9 Administrative expenditures 138,000 1.4 70,800 17.1 Unidentified (residual)d 1,097,000 11.2 13,000 3.1 Totald 9,805,000 100.0 414,800 100.0 (Notes are on the foLkowing page.) 368 WILLIAM A. BYRD AND ALAN GELB Notes to Table 17-5 n. a. Not available. - Not applicable. a. In 1985 the township public finance system had not yet been established in township A. In 1986 township A was to retain 3.7 percent of tax revenues collected within the quota and 8 percent of above-quota tax collections. b. Since township B's enterprises had net losses totaling Y13,759 in 1986, these remittances were financed by additional bank borrowing and by drawing from depreciation allowances (in effect "eating" the enterprises' capital stock). c. This is probably an overestimate, since it was assumed that all budgetary expenditure "in support of production" other than items clearly specified as agricultural was for investment in township enterprises. d. Total expenditures are assumed to be exactly equal to total township revenue resources, with the unidentified portion of expenditures then derived as a residual. Source: Information from fieldwork. the township government. They accomplished this by incurring additional bank debt and by using their depreciation funds, thus "eating" enterprise cap- ital. On the expenditure side, too, there are striking differences. Perhaps the most important is that in township A nondiscretionary spencLing for salaries, overhead, social services, and the like constitutes only a small pro- portion of total spending (about 12 percent). In township B, by contrast, education and health alone account for more than 60 percent of total ex- penditure and salaries and overhead for close to 20 percent. The absolute level of township government spending on reinvestment in township enter- prises is nearly 350 times as large in township A as in township B. These differences in flows of funds reflect a huge gap in economic cievelop- ment and industrialization (see table 17-6). Differences in industrialization Table 17-6. Economic Indicators for Two Townships (yuan, except as indicated) Township A Township B (percent) Indicator (Wuxi, 1985) (Shangrao, 1986) A/B Population (number) 18,779 23,396 80 Average per capita income of rural population 907 131 692 GVAO per capita 399 197 203 GVIO per capita 11,268 17 663 Sales income of township enterprises per capita 5,846a 14 418 Local revenue per capitab 1,268 9 14,089 Expenditure per capita 522 18 2,900 a. Industrial township enterprises only. b. Includes both budgetary and nonbudgetary revenues but not additional budgetary funds provided by the county for quota budget expenditures. Sources: Table 17-5 and information from fieldwork. Why Industrialize? The Incentives for Rural Community Governments 369 and in community govemment financial resources are far greater even than the large difference in average per capita income, which is roughly 7 to 1. Fiscal redistribution hardly makes a dent in resource inequality at this level. Although these tremendous differences are extreme by Chinese standards, they graphically portray the high degree of spatial inequality in rural industri- alization and fiscal capacity as well as in incomes. Financial flows are somewhat simpler for private enterprises, although there is great variation among localities and enterprises. Figure 17-2 shows some strong community-firm linkages. Private enterprises pay tax to the township tax office, typically on sales income (often 5 percent in industrial- commercial tax plus a certain amount, usually 2 percent, in lieu of the profit tax, which is generally not collected). Management fees are paid to the town- ship industrial and commercial administrative management office, which is in charge of enterprise registration, and sometimes also to the community government. Finally, part of enterprise surplus goes directly to the commu- Figure 17-2. Flows of Funds between Private Firms and the Local Community Township goverru ent Township industrial l l 0 Township governmenr t |and commercial bureau Township ta office Wages and supplements Community govermnent E ! leaders Community l government -- Profit sharing, informal incomes, and consumption I I ~~~~~~~~I I i Bank loans Private entrepreneur I I . Salary, bonuses, housing, and so on II Bank deposits , , ,l I _ r ~~~~Reinvestmenr I i Private firm Profits i I Loans and capital contributions _ I _ ~~~~~~~~Budgetary flows -. j j i ,Pay and bonuses Nonbudgetary flows . Management fees ----- 370 WILLIAM A. BYRD AND ALAN GELB nity govemment and its leaders, sometimes as explicit profit remittances, profit sharing, or salary payments to community govemment leaders, but also in the form of informal payments, gifts, banquets, and so on. Owners typically do not provide additional capital to a private enterprise once it is established on a sound footing; indeed, the seed capital they put in at founding is often repaid to them quickly. Subsequently, the private en- trepreneur typically does not withdraw a large proportion of after-tax profits but instead reinvests profits in the firm. (Of course, he can appropriate some enterprise surplus in the forrn of wages, bonuses, other benefits, housing, and the like.) Private enterprises may obtain significant amounts of capital di- rectly from employees and local community members. Larger and mnore suc- cessful private enterprises may also be able to obtain loans from banks (see chapter 9). Linkages between the TVP Sector and the Local Economy We are now ready to discuss the main linkages between TVPS and rural com- munities. These consisit of fiscal linkages, income and employment linkages, and various rewards for community govemment officials. Fiscal Linkages The importance of TVP development for community govemment revenues is apparent. TVPS generate tax revenues, and the township gets a share of the taxes collected in excess of its quota. The share of tax revenues retained by townships varies greatly. 'The more advanced localities retain a lower share of taxes than do poor localities, but because of their rapid economic growth it is easier for them to exceed the collection plans, which are set largely on the basis of past collections. Profit remittances and management fees of com- munity enterprises are important in the more advanced areas and are the main source of discretionary community government funds. TVPS, especially community enterprises, are highly visible profit centers that make convenient sources of govemmeni: revenue. Moreover, indirect tax rates on industrial production are much higher than those on agricultural output. The agricul- tural tax (really a form. of taxation on land) is frozen at a low level and even so is hard to collect in the poorer areas. Collection of taxes and levies from individual households is frowned on in Chinese rural communities. It is an embarrassment to local officials if they need to rely on such sources, which in any case would not yield much revenue. The fiscal linkage is all the more important because community govern- ments face a rather hard budget constraint. Deficit financing by any level of govemment below the center is not permitted in China, and lower levels of government do not have the right to establish new tax instruments or set tax rates. Why Industralize? The Incentives for Rural Community Governments 371 Given the increasingly close link between community revenue and ex- penditures and the importance of the TVP sector in increasing revenues, TVP development is the single most important determinant of a local community's spending capacity. This is particularly true of resources for discretionary spending-those not required to meet the community government's payroll and overhead. Since discretionary community govemment funds are mainly used for investment in new or existing community enterprises, TVP develop- ment (or lack thereof) can easily become self-reinforcing. Localities such as Wuxi that have achieved successful TVP development generate large flows of discretionary resources, the lion's share of which is used to promote further TVP development. The more backward areas, such as parts of Shangrao and Jieshou counties, must use almost all of their locally mobilized revenues and their grants from higher levels of government to meet their virtually fixed expenditures on payroll, overhead, social services such as education and health, and so on. Since so few discretionary resources are available for com- munity enterprise development, it never gets off the ground, and this perpet- uates the shortage of discretionary resources for TVP development. The fiscal linkage is important to community govemment leaders for a number of reasons. In the first place, it is reasonable to assume that like gov- ernment bureaucrats anywhere, they are budget maximizers (see Niskanen 1968). Community government leaders interested in maximizing their fiscal resources have a strong incentive to promote TVP development. A related benefit is the enhanced fiscal autonomy for community government leaders who can draw on discretionary resources from TVPS. In addition, the pay of community government leaders and officials comes largely from revenues from TVPS; government personnel below the township level are collective rather than state cadres and hence are paid entirely from nonbudgetary reve- nues. Whether they receive any income at all from their positions depends to a large extent on TVP development. Income and Employment Linkages The level of TVP development is an important, if not the primary, determinant of rural per capita incomes. The correlation (R) between the value of rural industrial output per member of the rural population and average income per capita in China's provinces was 0.91 in 1985. Similar relationships hold among counties in the same province and, to a lesser extent, within counties. TVPS have been critically important in generating employment and raising living standards in rural areas. Both township leaders and factory directors, in their responses to questionnaires, overwhelmingly cited increasing employ- ment and raising local living standards as the primary motivations for estab- lishing new firms. In a situation of factor and population immobility, the sharp increases in agricultural productivity brought about by the PRS could not have been translated into higher incomes without TVP development, as 372 WILLIAM A. BYRD AND ALAN GELB rural surplus labor could not have found more productive employment any- where else. In the more developed localities such as Wuxi and Nanhai there has been a wholesale shift in the labor force from agriculture to industry in the past several years. These areas have been so successful in generating nonagricul- tural employment opportunities that they are experiencing a labor shortage (Wuxi) or significant temporary in-migration (Nanhai). In Shangrao, where the TVP sector has not developed rapidly, there is still a large surplus of rural labor that cannot and should not be employed in agriculture. Rewards to Community Government Officials The incentives created by the more general fiscal and economic linkages noted above are important to community government officials and leaders and in many cases may outweigh any explicit personal financial incentives. Carrying on community government business and dealing with superiors in the county government: must be much easier and pleasanter when revenue constraints are eased, community members are fully employed in high-paying jobs, and the local economy is booming. But the personal rewards and career prospects of community government leaders also depend greatly on income, employment, and revenue mobiliza- tion in their communities. These personal benefits can be divided into three main categories: pay increases and bonuses associated with TVP development and other indicators of community economic progress; informal sources of income and consumption generated by TVP development; and prospects for promotion and the related improvement of social status. The first set of bene- fits can be described in a fairly precise manner, but there is considerable vari- ation among localities and among the township, village, and production team levels. Only anecdotal, largely qualitative evidence is available on the second category. The thLird category is important in poorer parts of the coun- try but is actually a negative incentive in better-off areas. PAY. Community government leaders' pay varies significantly with the de- gree of TVP development in the community, and this variation tencs to be greater for villages and production teams, than for townships and towns. There is an implicit but powerful linkage between average income per capita in communities and the maximum socially or administratively permissible level of pay for community government leaders and for directors of commu- nity enterprises. The margin by which their pay can exceed the average in- come of community members (or of enterprise workers in the case of direc- tors) varies among localities and has been increasing, but everywhere and at all times there appears to be at least some fuzzy limit that cannot be trans- gressed. In Wuxi the pay of a factory director is generally limited to no more than Why Industrialize? The Incentives for Rural Community Governments 373 twice the average pay of employees in his factory. Nominal pay and supple- ments for leaders of a village are also about twice the average pay of employ- ees of the village's enterprises.5 Since workers' wages are linked to enterprise profits, there is an indirect relationship between community government leaders' incomes and the profitability of community enterprises. The current reward system reflects measures by the county to limit differentials between community government leaders and factory directors on the one hand and ordinary workers on the other after community government leaders and espe- cially factory directors received what were deemed excessively high incomes in 1984.6 In Jieshou it is harder to discem a clear pattern, but in the private enter- prises surveyed, owners (who are also directors) often appear to be able to earn high salaries only by paying their workers much higher wages than war- ranted by local labor market conditions. Informal incomes of both township and village officials are closely related to community enterprise profits and to incomes of community enterprise employees. When average incomes are relatively low, the fixed portion of community govemment leaders' compen- sation looms more important, especially at the township level. In Nanhai township leaders' pay is typically two to three times the average income per worker in their communities. A similar relationship holds at the village level; since variation in average income per capita among villages is typically greater than variation among townships, village leaders' pay tends to differ much more than that of township leaders. Leaders of successful vil- lages can earn far more than their superiors in the township government; leaders in one township earn about Y2,500 a year in the formal pay structure, whereas incomes of village leaders range from Y2,000 to Y6,000 a year. In- comes of village officials are linked directly to the profits of "their" commu- nity enterprises, unlike the situation in Wuxi. In Shangrao there is also some linkage between local income levels and community government leaders' pay, but the variation may be considerably less, and hence the incentive effects are weaker. The reason is the low gen- eral level of incomes and the influence of state wage scales, which for town- ship leaders set a lower limit on compensation that is high by local standards. Moreover, since the community enterprise sector in most townships has an abysmal profit performance, there is little if any linkage between community enterprise profits and leaders' pay. Income determination differs as between township leaders and leaders of lower-level communities. The former are government cadres on the state budget payroll and are subject to a national salary scale. The official salary of a township leader is roughly Y1,000 a year and is largely invariant across localities and regions with widely differing per capita incomes. If the state salary scale were strictly adhered to, township leaders in advanced localities such as Wuxi and Nanhai would earn considerably less than ordinary TVP workers. Incentive problems are aggravated because-unlike leaders of lower- 374 WILLIAM A. BYRD AND ALAN GELB level communities-township leaders, as govemment cadres, are forbid- den to engage in business activities on the side, which can be potentially lucrative. The richer areas have added supplements to the compensation package of township leaders to make their total pay reasonably competitive. These sup- plements must be financed from nonbudgetary funds, primarily remittances from community enterprises. 'They do raise township leaders' income levels substantially, but the room for variation according to TVP performance and other economic indicators is still limited. In the poorer areas there is no scope for variation at the lower end, since the base pay of township leaders forms a floor. In Shangrao township leaders' pay does not go below about 'Y 1,000 a year, and the maximum for those whose townships perform well is not much more than Y1,500. This may significantly dampen the extent to which lead- ers suffer for the poor economic performance of their townships, although they would still be adversely affected through the nonpecuniary linkages mentioned above. Village and production team leaders are not under the state wage system, and their pay is free to vary across localities in line with local income levels. The top pay of such people in Nanhai can reach as high as Y15,000 (in a village where workers received about Y5,000), and at bottom it can be Y2,000 or even less. The main determinant of village leaders' pay is the prof- its of village firms. In Shangrao village leaders' pay reportedly ranges from Y1,000 to Y2,000 a year. A good example of the strong incentives for village leaders is Xiqiao town in Nanhai, which in 1986 promulgated a detailed incentive pay system for leaders of subordinate villages. Village leaders get Y55 a month of base posi- tion pay. A large floating wage component that depends on the achievement of economic and social targets is paid annually. Up to Y500 a year can be earned for meeting village targets for state grain procurement, tax revenue, and treasury bond quotas (Y150), family planning (Y100), military recruit- ment and militia work (Y 100), and financial and contract management activ- ities (Y150). Subordinate procluction teams in each village can earn up to Y600 a year for fulfilling various targets. Finally, a gradually falling percent- age of the profits of community enterprises goes to the village leader and his coworkers. The profit-sharing rate ranges from 8 percent for profits of less than Y50, 000 to 1 percent for profits of Y2 million-Y3 million. Conceivably, a village leadership group could earn as much as Y57,000 a year through this reward mechanism.7 With annual base wages at only Y660, the possibility of earning up to Y1,100 for meeting various targets and perhaps Y10,000 or more on the basis of community enterprise profits creates strong incentives for village leaders in Xiqiavo to develop community enterprises. Similar incen- tives are in place elsewhere in Nanhai. The income of production team leaders is, if anything, even more closely linked to the performance of community enterprises. Under the PRs the ad- Why Industrialize? The Incentives for Rural Community Governments 375 ministrative role of production teams has been sharply circumscribed, and those without successful enterprises have become largely vestigial organiza- tions whose leaders have no opportunity to earn high incomes. Production team leaders have thus been driven to develop team enterprises (of which they often are concurrently director) or to establish their own firms. In Jieshou one former team leader bought his team's edible oil-processing plant, which had previously been operating inefficiently under his management, and tumed it into a large and successful diversified enterprise that produced machinery as well as food products. Some team leaders in Nanhai get large explicit payments (sometimes more than Y10,000 a year) from production team firms even though they are not directly involved in management. It is impossible to gauge the magnitude of community government leaders' informal incomes with any precision, but rewards of this type are an impor- tant part of their total compensation. Outright bribery may not be common and in any case involves risks for the recipients. Much more prevalent is con- sumption of community government or enterprise resources, from cigarettes to banquets to housing construction. Since more funds are available for such uses in the more developed areas, community government leaders undoubt- edly have an additional incentive for stimulating TVP development. Whereas formal pay incentives generally are related closely to the performance of com- munity rather than private enterprises, private enterprises are as good a source of informal incomes as community enterprises and are sometimes even better. PROMOTION AND OTHER CAREER INCENTIVES. These work quite differently in different parts of the country. In Wuxi and especially in Nanhai successful community govemment leaders are generally unwilling to be promoted to the next higher level. Production team leaders do not want promotion to village cadre status; village leaders, in particular, do not want to become township leaders; township leaders do not want to be county govemment cadres; and county officials reportedly try to avoid promotion to the provincial level. This phenomenon is the result of the relatively low govemment wage scale (in relation to local incomes in the richer areas) and the prohibition against govemment cadres' engaging in private business on the side. The impact of these negative incentives may not be serious and indeed may have some benefits-leaders without promotion opportunities may well be more con- cerned about the long-term welfare of their communities. In the poor, backward areas promotion incentives are much more potent. Not only is there little if any loss of income for promoted community govern- ment leaders, but the change in household registration status from rural to urban residency for those promoted to the township or county level is of in- calculable value. In Jieshou and especially in Shangrao, lower-level commu- nity government leaders actively strive for promotion to state cadre status. Criteria and decisions concerning promotion may be somewhat subjective, 376 WILLIAM A. BYRD) AND ALAN GELB but undoubtedly the economic performance of the community for which the leader is currently responsible carries great weight. Despite the variations in personal incentives for community government leaders among regions and localities and among community levels, it is clear that everywhere and under almost any circumstances community govemment leaders have strong incentives to develop the TVP sector. This is broadly true even where TVP development has lagged or failed. Thus the main forces that differentiate TVP performance among regions and localities are not related to the personal incentives of community government leaders but rather to the prevailing pattems of constraints and their interactions with these incen- tives. Patterns of Commtnity Government Involvement in the TVP Sector In relatively prosperous areats the relationship between community govem- ments and their enterprises tends to be mutually beneficial. But in poorer areas governments are forced to exploit their enterprises, to the long-term detriment of both firms and community. This section examines the reasons behind these dramatically different outcomes and also discusses the condi- tions that lead to encouragement of private enterprise. Support for Community Enterprises in Well-Endowed Areas In better-off areas such as Wuxi and Nanhai that have a strong base of exist- ing community enterprises, rnost community govemments are in a relatively comfortable financial situation. Budgetary revenues far exceed mninimum budgetary expenditures, and some above-plan revenues can be retained for the discretionary use of township govemments. Large nonbudgetary revenue flows from community enterprises greatly exceed public expenditures not cov- ered by the budget. The bulk of these funds, and the substantial profits re- tained by community enterprises, are therefore available for reinvestment to expand the community enterprise sector. Finally, these areas also have high personal incomes and bank deposits, which can be used for loans to local TVPS (see chapter 9). All this means that mobilizing resources for public ex- penditures is relatively easy: basic expenditures as well as numerous extras can be financed from a small part of the profits remitted by community enterprises, leaving adequate funds to support the continued development of the enterprises. All of the linkages discussed in the previous section reinforce the incentives of community govemment leaders to promote the develop- ment of community enterprises. Important policy changes in recent years have strengthened the incentives and increased the opportunities for community govemments in the better- Why Industrialize? The Incentives for Rural Community Governments 377 endowed areas to promote community enterprise development. The linkages in the fiscal and financial systems between local mobilization and funds avail- able for investment in TVPS were put in place only recently. Variation in the pay of community govemment officials probably has become significant only since the late 1970s. The PRS increased both the pressures and the incentives to generate additional employment by developing community enterprises. The removal of political restrictions on rural industrialization and the weak- ening of ideological opposition to raising rural personal incomes were also important. Fiscal Predation in Backward Areas The same set of linkages can have an adverse impact on community enter- prise development in backward areas with low incomes and a weak or nonex- istent base of TVP development. The incentives are basically similar to those in the more advanced areas, but the opportunities for developing community enterprises in the traditional pattem are much more limited. Moreover, the difficult financial situation of community govemments impels them to draw funds from community enterprises for public expenditures, regardless of the enterprises' profitability and ability to pay. Even when the firms are losing money, they still have to make payments to community governments, which they do by using their depreciation funds (in effect "eating" their capital stocks) and borrowing from banks and rural credit cooperatives. Fiscal preda- tion thus occurs because community governments are forced to appropriate the funds of community enterprises to meet urgent and largely fixed public expenditure needs. The basic problem is that the minimum necessary public expenditure re- quirements of community govemments (as determined by their responsibili- ties and payrolls) are at least somewhat inelastic with respect to local income per capita, development, and industrialization. In the advanced areas mini- mum public expenditures are small in relation to local capacity to mobilize resources. In the poorer, more backward localities mobilization capacity is much lower, since it declines more or less in line with incomes and industrial production. Yet the number of community government employees may not be that much smaller, and the govemments are expected to provide and largely finance a full array of education, health, and other social services that may not be much worse in quality than those in more developed regions.8 The need to meet largely fixed expenditures forces community govem- ments in poorer areas to exploit community enterprises in ways that harm their long-term development potential. Community enterprises are starved for development funds, and because their credibility with banks is severely strained, they find it hard to obtain loans even for genuine business invest- ments. In this way a vicious cycle arises. Since community enterprises lack 378 WILLIAM A. BYRD AND ALAN GELB the investment funds needed for healthy development, the fiscal situation of community govemments remains tight, which in tum means that commu- nity govemment leaders are forced to continue milking community enter- prises, and so on. A related form of fiscal predation occurs when community governments require their enterprises to taike on excessive numbers of workers and manag- ers.9 The latter are often former village and production team officials who were moved out to make room for younger replacements or whose jobs were eliminated as a result of the PRS. Excessive payrolls further squeeze the funds available to community enterprises and make it more difficult for them to achieve self-sustaining development. Local political considerations some- times lead to the appointment of inexperienced and even incompetent offi- cials to decisionmaking positions in community enterprises. This not only reduces efficiency but may also have a severe effect on workers' morale. Pad- ding of payrolls and intervention by community govemments in the day-to- day operations and management of community enterprises appear to be more pervasive in poorer, less industrialized localities such as Shangrao. Fiscal predation harms the long-term prospects for TVP development and hence, through the linkages described earlier, the long-term interests of community governments. Although the practice is shortsighted, it is under- standable, given the shcrt-term pressures on community governments. Fiscal predation can be viewed as an attempt by community govemments in back- ward areas to get around their hard budget constraint in the short run by obtaining indirect deficit financing from the banking system. The linkages and incentives that work so well in generating TvP development in better-off areas turn against it in rmore backward areas that lack an established, viable TVP base. Healthy TvcsIS have no trouble generating employment, remit- ting profits to the government, and obtaining bank financing for their operations-indeed, these things come naturally. But these same activities are deadly to the viability and development prospects of community enter- prises if they are forced on them from outside. As in the success stories represented by Wuxi, institutional and policy changes in recent years have affected community govemment involvement in the more backward areas. Before the PRs, paying for govemment salaries, overhead, and social expenditures must have been a considerable burden on the poorer rural communities, but it was spread over the entire community through the work-point system and collective income distribution. Although this undoubtedly hampered development potential in the more backward ar- eas, consumption may have been affected more than investment. In any case, how the impact was handled was to a large extent the choice of the commu- nity govemment. The PRS made agricultare and the community population as a whole much more immune to fiscal predation and left only rural industry as a source of Why Industrialize? The Incentives for Rural Community Governments 379 financing for nonbudgetary expenditures. This provided a strong incentive for community officials to promote TVP development. But where the de- velopment of community enterprises has not reached the take-off, or self-sustaining, stage, community governments have been tempted-even compelled-to milk their enterprises. In contrast to the prereform situation, predation mainly consists of using investment funds (community enterprise depreciation allowances and bank loans obtained by the enterprises) to fi- nance the consumption expenditures of community governments. The fiscal predation evident in relationships between community govem- ments and their enterprises in backward localities raises doubts about the strategy of self-reliance in the provision and financing of social services in rural areas. Self-reliance was highly successful in the past; it has allowed China to reach unprecedentedly high levels of rural education and health with only modest budgetary expenditures. A comprehensive and all-encom- passing rural government and Communist party network was also financed largely from local resources. But fiscal predation and unsuccessful TVP devel- opment in backward areas show that self-reliance has severe disadvantages and high costs. Insulating Community Enterprises from Fiscal Predation Govemment officials in Shangrao, particularly at the county level, are only too aware of the harmful pattern described above. A number of measures have been taken, particularly in a few of the more successful townships and towns, to insulate and protect community enterprises from fiscal predation. Sometimes the community government itself has taken the initiative in this respect, realizing that its longer-term interests lie in keeping intervention and profit levies to a minimum and allowing community enterprises to develop. The best example of a community in Shangrao that has avoided fiscal pre- dation and seems to be on the way to self-sustaining TVP development is Shaxi town. Shaxi has certain advantages: it was a traditional commercial center; it has a fairly solid base of TVP development; there are long-standing commercial ties between Shaxi firms and enterprises in Shanghai; and the large profits eamed from ramie cultivation and trade in 1984-86 provided a source of funds for TVP development. (This money was tapped by requiring new employees of community enterprises to contribute Y1OO-Y5,000 in capi- tal.) Nevertheless, before strong measures were taken to combat fiscal preda- tion, the TVP sector was not developing rapidly. In 1984 Shaxi instituted measures to strengthen the management of com- munity enterprises and improve incentives. A contracting system was estab- lished between factories and the TIC as well as between the TIC and the town government. Factory directors are now elected by the workers, not appointed by the town government, and they have broad authority in decisions on per- 380 WILLIAM A. BYRD AND ALAN GELB sonnel and employment, the use of funds, payment of bonuses, and business operations. Permanent workers were converted into contract workers with renewable terms of one year. Finally, piece-rate wages were instituted in all factories. Only 40 percent of after-tax profits within the plan and 20 percent of profits above the plan are remitted by factories to the TIC and the town govemment. Election of factory directors seems to be effective in helping to stem gov- ernment intervention and fiscal predation. Elected directors may be in a bet- ter position to resist the tendency toward inflated managerial staffs and exces- sive levies of funds by community govemments. Changes and reductions in management teams have occurred after new factory directors were elected. Given the personal ties between community officials and appointed commu- nity enterprise manageirs, it is easier for elected factory directors under con- tract with the community government to get rid of people who are not needed. An elected director is also in a better position to resist pressures from the community govenmment or from his own friends and relatives to pad the payroll. Under the election system the director derives part of his legitimacy from workers, who are also community members. This gives him a more se- cure position and somewhat greater independence and makes it more likely that an arm's-length contractual arrangement with the community govern- ment will work well. '° Worker election of factory directors and the associated contract system are not primarily an attempt to institute greater worker participation for its own sake or to improve employee incentives and morale, although they may have these effects. Instead they serve as a built-in mechanism, devised by commu- nity govemments in their own long-term interests, to limit fiscal predation and excessive government intervention in enterprise operations, especially in personnel decisions. Election may not be the optimal method of selecting enterprise directors, but at least it prevents community leaders from giving in to the temptation to choose unsuitable or incompetent persons. These methods, which seem to have worked well in Shaxi, may not suc- ceed everywhere in Shangrao or in other backward areas. Townships in worse financial straits may not: be able to exercise the requisite self-disciplirne even when dealing with firms with elected managers and contractual arrange- ments. A principal factor behind Shaxi's success was probably its reasonably good financial situation, which allowed it to meet public expenditure require- ments relatively easily. In other townships a weak base of entrepreneurial and managerial skills makes it difficult for community enterprises to achieve healthy profitability in the first place. Although the measures implemented by Shaxi represent an admirable attempt to come to grips with the problems of fiscal predation and excessive govemment intervention, the situation will remain precarious until TVP development becomes self-sustaining and the revenue crunch has been eased. Why Industrialize? The Incentives for Rural Community Governments 381 General Indirect Support for the Development of Private Enterprises Some community and county governments, both in more developed areas and in backward regions, have chosen to create an administrative and regula- tory environment conducive to the development of private enterprise. Where there is a modicum of entrepreneurial spirit and experience in the population (even from forty years ago), relaxation of restrictions on private enterprises in itself usually stimulates substantial development. This has ben- efits for the community economy and the community government whether or not the community enterprise sector is dynamic. In Nanhai government authorities had a relatively permissive attitude to- ward private enterprises even in the prereform period, and there was a burst of private enterprise development after the Central Committee Document No. 1 of 1984 officially sanctioned such development. At present, private enterprises are allowed to flourish, but most are not permitted to take out bank loans above certain limits, and an exodus of community enterprise em- ployees to go into business on their own has been partly stemmed by adminis- trative measures. There are some large private entrepreneurs in the county who got into business early or took advantage of their access to bank loans when these were not restricted. Many of them seek to protect themselves by establishing joint ventures of various kinds with community governments (see chapter 9). Most private enterprises, however, are very small. In Shangrao private enterprises have expanded their share in total TVP sector industrial output, apparently even in the absence of strong measures to en- courage them. Private enterprises, particularly small ones established without the assistance of community governments, may be in a better position to re- sist fiscal predation than are community enterprises. Moreover, these enter- prises raise local incomes, increase employment, and provide small amounts of revenue for the community government through management fees and tax payments. In the more advanced areas, local authorities can choose whether to let the private enterprise sector flourish, but in any case they would be foolish not to actively promote the continued development of the traditional com- munity enterprise sector. Wuxi has chosen to restrict private enterprises, Nanhai to let different forms of ownership coexist. In areas in which commu- nity enterprise development has not been successful, a permissive attitude toward private enterprises has important benefits for community govenmments and their leaders, although these are less direct than the fruits of successful development of community enterprises. If the latter is difficult or impossible to achieve, development of private enterprises may provide at least a second- best solution and help community government leaders meet some of their plan targets. Without explicit positive actions by community governments, however, private enterprises are likely to remain relatively small. 382 WILLIAM A. BYRD AND ALAN GELB Personal and Financial Sponsorship of Large Private Enterprises Community and county authorities sometimes go a step farther and provide direct personal and financial support to large private enterprises. In Jieshou and the surrounding counties, leading county officials personally sponsor a small number of successful private entrepreneurs. Sponsorship facilitates pri- vate enterprises' dealings with local administration and red tape and provides ideological and administrative protection without which the firms would feel vulnerable, given their size and visibility. This close personal relationship between private enterprises and their community government sponsors is ac- companied by a relaxed attitade toward the enterprises' access to cornmunity resources such as bank loans. Community government officials may also gain personal advantages from sponsorship of private enterprises, but the magni- tude of these benefits is hard to gauge. In some Jieshou towrnships the community govemment provides entrepre- neurs with funds and other support to develop firms that may be private in theory but that are under loose administrative supervision by the community govemment. In one township, Jinzhai, four of the seven "township-run" in- dustrial firms are of this semiprivate nature. This approach may be superior to traditional community enterprise development in that enterprise directors may be somewhat insulated from the community government, while at the same time the firm has access to community resources. But there are numer- ous ambiguities and potential problems because property rights are not well defined. Some large private enterprises in Jieshou town have been collectiv- ized, a step facilitated because many of them registered as collective firms when they were established in the early 1980s. A similar strategy that works from the other side of the TVP spectrum is contracting or leasing community enterprises to private management. Usu- ally what happens is that money-losing community enterprises are contracted out to avoid their beingJ a drain on community resources. At the very least, contracting insulates community enterprise management from petty interfer- ence by community officials and other agencies. A good example is Shaxi town, where enterprise directors and employees may receive somewhat higher rewards, but directors do not have complete independence and are not resid- ual claimants. Stronger forms of contracting can fall just short of outright sale to private ownership; in both Jieshou and Nanhai contracting has often been a prelude to such sales. The benefits to community govemments and their leaders from active pro- motion of private enterprises (particularly large ones) are in many respects similar to those from TVPS in general, but there are some differences. The scope for informal payments and personal benefits is probably greater for private than for community enterprises, but flows of budgetary and non- budgetary funds to the community government itself are smaller. Direct Why Industrialize? The Incentives for Rural Community Governments 383 involvement by community governments in larger private enterprises leaves a legacy of confused property rights that may pose problems for successful pri- vate enterprises in the future. Implications and Policy Issues Uneven Development and Ownership Patterns Concern about the problems of backward areas, some of which have failed to establish a self-sustaining TVP sector, should not distract us from the highly successful development and performance of the TVP sector as a whole. The strong incentives for community governments to promote TVP development undoubtedly have been crucial to the success of the sector. But tremendous inequality in the economic resource base" and limited population and factor mobility have led to highly uneven patterns of TvP development that are likely to continue as long as imbalances in endowments and immobility of factors remain unchanged. Under these circumstances community govem- ments that are striving to establish and develop TVPs are largely left to their own devices and resources. Where these are inadequate, progress is bound to be slower. Imbalances are exacerbated by fiscal predation, which may well be the most severe obstacle to TVP development in backward areas, such as parts of Shangrao, that have poor endowments of material and human resources. Fiscal predation is the result of a confluence of factors, including poor re- source endowments, a precarious fiscal situation, surplus labor and cadres, and the hard budget constraint of community governments. The inelasticity of basic public expenditures with respect to levels of income and develop- ment is a fundamental cause of fiscal problems in poorer areas and thus of fiscal predation. Although local and community authorities may be able to ameliorate fiscal predation somewhat, significant changes in national policies may be required to achieve substantial improvements in the situation (see below). As a result of fiscal predation and other factors, ownership is becoming correlated with the level of development. The backward areas are being forced to turn away from a traditional community enterprise development strategy and actively encourage or at least permit the emergence of a large private enterprise sector that sometimes includes sizable firms with assets ex- ceeding Y1 million and with more than 100 workers. The better-endowed areas have a greater degree of choice with respect to the private enterprise sector, but in any case they will continue to develop the traditional TVCE sector. Whether the turn to private ownership will allow backward areas to catch up with the front-runners is an open question, but resource inequality and continuing factor immobility make such an outcome unlikely except in a handful of exceptional cases. 384 WILLIAM A. BYRD AND ALAN GELB Conflicts between Government and Ownership Roles The dual roles of township and village governments as conventional local governments and as business entities in charge of community enterprises do not conflict severely in advanced areas such as Wuxi, where traditional com- munity enterprise development has been highly successful and is now self- sustaining. In fact, in Wuxi the two roles are institutionally separated to some extent. The Tic directly exercises ownership and other economic roles with respect to township enterprises. Its primary relationship with the township economic commission (TrEc) is to turn over to it part of the profits remitted by township enterprises. The TEC has a dual role, but it is not directly respon- sible for supervising enterprises. Finally, the township government is consid- erably removed institutionally from the community enterprise sector, al- though undoubtedly township leaders become involved in important decisions on investment projects, the creation of new enterprises, and the like. But in the backward areas where fiscal predation occurs, there are seveTe conflicts between the two roles of community governments, and the govem- ment role commonly squeezes out the business role. From this perspective., the issue of how to avoid fiscal predation is akin to the question of how to avoid conflicts in the dual roles of township govem- ments. Institutional separation is one alternative, but it runs the risk of dilut- ing community government incentives to promote TVP development (al- though this does not appear to be a problem with the limited degree of separation in Wuxi). Another possibility is to enforce a greater degree of in- dependence for community enterprises with respect to the community gov- ernment so that, for example, community enterprises are required to remit funds to the community govennment only if they make a profit accorcling to a strict accounting definition. Although this alternative would protect indi- vidual community enterprises from fiscal predation and prevent comrnunity governments from using firms to soften their hard budget constraint, it might also weaken risk bearing and financial intermediation by the community gov- emment. (Since in backward areas the community govemment has not been able to play any risk-bearing or intermediary role anyway, this may not be a severe weakness.) Greater independence for the local community banking system might prevent community governments from using it as a means of "deficit financing." Still other alternatives would require nationwide policy changes that are explored below. The Need for Greater Pay Incentives for Township Leaders A primary issue in the area of personal incentives and rewards to community government leaders is the uniform nationwide pay scale for township inagis- trates, who, unlike leaders of villages and production teams, are considered Why Industrialize? The Incentives for Rural Community Governments 385 state cadres and are under the state pay system financed from the budget. Supplements and subsidies from nonbudgetary revenues are necessary to en- sure an adequate salary for township leaders in areas such as Nanhai. Despite these supplements, township leaders in better-off areas are paid less than many village leaders, and variation with TVP performance and other eco- nomic indicators is generally less than for village leaders. A large part of township leaders' rewards may come from informal income and benefits, but this opens the door to corruption. Considerations of scale, risk pooling, and resource mobilization would sug- gest that in most parts of the country the optimal size of rural community for community enterprise development is the township (see chapter 9). But township leaders' personal income incentives for TVP development are weaker than those for village and production team leaders. It would be desir- able to institute for township leaders an explicit bonus system linked to town- ship employment, output, profitability, and related indicators. To be effec- tive, the bonuses would have to be potentially large in relation to base pay, and both components should vary according to the level of personal incomes in the locality. Bonuses could be linked to future as well as current TVP devel- opment and profits, perhaps through performance-based pension benefits. Similar incentive schemes could be instituted at the county level, where vari- ation in pay is even smaller than at the township level. Methods of Preventing Fiscal Predation Unchecked fiscal predation will lead to stagnation of the community enter- prise sector in some areas and to a tum toward rural development strategies based on private enterprise. This pattern may not be welcome ideologically or politically beyond a certain point, and in any case there are many unre- solved problems and inconsistencies conceming private enterprises, espe- cially large ones (see chapter 9). Thus, avoiding or at least ameliorating fiscal predation is an urgent policy issue, but it is by no means easy, as the experi- ence of Shangrao shows. Some specific measures that are within the scope of authority of township govemments themselves (and were actually instituted by Shaxi town) have been discussed above. But these may not work everywhere and, as in Shangrao, they seem most efficacious where community resource constraints are at least temporarily eased by inflows of funds. The question arises as to whether there are any nationwide policy changes that could help the situa- tion. There are, but they would represent a sharp departure from past prac- tices and may require substantial commitments of resources. The fiscal problems of backward townships are related to two of their struc- tural characteristics. In backward areas the government and its activities are too big (in particular, the number of people on the community government 386 WILLIAM A. BYRD AND ALAN GELB payroll is too large) in relation to the resources of the community. Moreover, grass-roots rural commu;nities have borne, by default, much of the financial burden of providing basic services, regardless of their ability to pay. The problem of fiscal predation therefore needs to be attacked from two different angles. First, the size of community governments and in particular the number of cadres (including those who are "placed" in community enter- prises for lack of anything better to do with them) must be reduced to corre- spond better to local cornmunity resources. Backward areas cannot affDrd and probably do not need the full government administrative structure and staffs that may be necessary in the more advanced areas where economic ;activity is much greater. 12 Although retrenchment may be difficult and gradual, the objective should be clear. Second, the relative importance of the social expenditures now being taken care of outside the budget at the township level should be evaluated, and activities that are truly important to national objectives should be in- cluded in the state budget. If such activities as primary education (now partly financed outside the budget) are neglected in the more backward areas, the existing inequality in resource endowments, living standards, and develop- ment prospects will become even worse. Putting essential public and social services under the state budget would require additional budgetary resources, but it would ease rural communities' financial constraints and hence reduce the need for fiscal predation."3 If community public expenditures can be met from budgetary allocations, levies on community enterprises can be limited and profits remitted by them can be used primarily for reinvestment in the TVP sector. Finally, because of factor immobility, solving the problem of fiscal preda- tion may not by itself greatly ireduce spatial inequality in TVP development. But if fiscal predation ceases, community enterprises and their government supervisors will at least be in a better position to acquire the crucial mnissing factors-managerial skills, marketing ability and contacts, and technical knowledge and capabilities-from outside. Notes 1. Rural development programs and industrialization efforts have often failed in other coun- tries because local governments had no incentives to ensure their efficiency. See chapter 19 for a discussion of Tanzania. 2. There is now a large and growing literature on public choice (see Buchanan 1975 and Mueller 1976). Public choice theory can be used to derive some powerful results conceming the harmfulness of govemment intervention, the need for constitutional restrictions against ex- cessive govemment spending, and so on. 3. In the past, restrictions on high personal incomes may have helped lessen variations in incomes among localities. These restrictions have been weakened as rural reforms have pro- gressed. 4. Migration does not reduce the population of poorer areas but rather increases residents' Why Industrialize? The Incentives for Rural Community Governments 387 incomes by allowing them to take higher-paying temporary or seasonal jobs in richer areas. 5. In the early 1970s a general principle was that the income of a brigade (now village) leader in Wuxi should be about 30 percent above the income of a relatively hard-working agricultural laborer. Later, brigade leaders' incomes were more closely linked to the pay of brigade industrial workers. In addition, the allowable differential has widened. 6. Contract systems for factory directors, introduced in 1984, at first linked rewards to enter- prise performance indicators rather than to workers' incomes, and this allowed great pay differen- tials. 7. Although the earlier rewards mentioned are given to individual village leaders, the pay drawn from enterprise profits is so large that it must be split among the leadership group and perhaps among other village cadres as well. 8. Shangrao's educational system is renowned for its quality and for the number of its high school graduates who pass examinations for admission to universities. 9. One township-run hydropower station in Shangrao, for example, had more than 100 em- ployees when 60 were sufficient for the job. Nearly all of the workers were "placed" through personal or work-unit connections. 10. The question then arises as to what to do with superfluous management personnel, who are often unable to perform other economic tasks adequately. Lingxi township has rediscovered the "golden handshake" as an effective way of getting rid of superfluous and incompetent cadres both in the govemment and in the enterprises. 11. This includes human resources. Because of China's strong basic education system, the quality of ordinary labor does not differ greatly among regions and localities, but for a variety of reasons entrepreneurs, managers, and technical personnel are unevenly distributed. 12. In Shangrao the separation of Shangrao city from the county exacerbated the problem by creating two separate and comprehensive govemment structures where probably only one was necessary. 13. Where budgetary funds are limited or if it is decided to give such a reform a more decid- edly "redistributional" character, inclusion of these expenditure items in the budget could be instituted only for townships below a certain threshold of poverty and backwardness. Part VI Comparative Perspectives Most of the earlier chapters in this volume have involved comparative analy- sis, but primarily within the TVP sector. Chapters 18 and 19 attempt system- atic comparisons between TVPS and relevant examples from outside the sec- tor. Chapter 18 compares township firms and the state enterprises that most closely resemble them-small state enterprises administered by counties or municipalities. Chapter 19 looks at analogies to China's TVP sector elsewhere in the world. In chapter 18 Song Lina examines township enterprises and local state en- terprises (LSES), noting that these are the two most closely comparable types of firm in the Tvp and state sectors. The study is based on data from sample township enterprises and from questionnaires completed by a small number of LSEs and their employees in Nanhai and Shangrao counties. Song's main hypothesis is that convergence between township enterprises and LSES is oc- curring and that the two types of firm are becoming more and more similar in organization and behavior. The chapter documents this phenomenon and proposes explanations. Song first points out the structural similarities between townships, on the one hand, and counties or small cities ("urban communities"), on the other. Despite their vastly different sizes, both types of community are characterized by broadly based economic sectors, increasingly similar labor force composi- tions, a variety of firms, and recruitment of employees mainly from within the respective communities. The analogous relationships between the two categories of firms and their owners (township governments and county or city governments) are seen as a principal factor in convergence. In particular, both types of firm have a similar position within their respective communities and provide significant revenues to their governments. Both have better access to inputs and other forms of preferential treatment within their communities, which partly offsets the burden of high profit remittances. Township enterprises and LSES are both under quasi-planning systems established by their respective community au- 389 390 COMPARATIVE PEIRSPECTIVES thorities. The control meclhanisms exercised by community governments over township enterprises and LSES are broadly similar. Song goes on to look at managers of both kinds of enterprise and finds that both sets of directors had surprisingly similar motives for taking their present posts. The directors tend to be outstanding people in their communi- ties, although LSE directors have considerably higher educational qualifica- tions than township enterprise directors. Criteria for assessment of their performance are broadly similar. Township enterprise directors internalize community goals to a much greater extent than do LSE directors, however. In both types of firm directoirs lack meaningful decisionmaking autonomy. Both township enterprise and LSE workers are primarily recruited from within the local community, and their educational levels are broadly similar. Both sets of workers have shifted their primary occupation from agriculture to industry to varying degrees, but they retain strong ties to the land. One important difference is that compensation packages for township enterprise workers tend to be much more closely tied to performance than those of LSE workers. Welfare benefits for LSE employees are extensive, whereas those mandated for township enterprise workers are minimal, but some township authorities are striving to provide benefits similar to those in LSES. Finally, workers' attitudes towaLrd risk, participation, status, and so on are fairly similar. The author compares the economic performance and efficiency of town- ship enterprises and LSE:S and finds that both tend to be less profitable than other types of firm in their respective communities but that township enter- prises earn substantially higher economic returns than LSES. Some reasons for the latter phenomenon are suggested: weaker accounting and bookkeep- ing systems in township enterprises, with much less external auditing and control; lower production costs in TvPs because of tax exemptions, much lower welfare expenditures, use of low-cost or substandard materials, higher labor productivity (owing to performance-based wage systems), and lower capital requirements; and possible differences in the subsectoral composition of industry. The chapter closes with some comments on the future prospects of LSES and township enterprises. Song asserts that LSES will stagnate for the foresee- able future, whereas township enterprises will continue to grow rapidly. Exist- ing differences between township enterprises and LSES will be reduced as convergence continues. In chapter 19 Alan Gelb and Jan Svejnar compare Chinese TVPS with a wide range of local public enterprises, worker cooperatives, and private firms in other countries. Comparators were selected for maximum similarity to TVPS, at least with respect to key aspects of their institutional structure. This meant that many interesting firm types were not included in the analysis. Although some illuminating similarities emerge, in general there are great Comparative Perspectives 391 differences between TVPS and even this select group of comparators, giving grounds for caution in drawing lessons. The chapter starts by looking at two types of firm in Eastern Europe: collec- tive farms, which are often organized as producer cooperatives and engage in diversified activities outside of agriculture, and Yugoslov labor-managed firms. The former are similar in many respects to Chinese TVPS, but they have not developed nearly as rapidly. Furthermore, their role in the economy is much less important, perhaps because Chinese community governments have actively supported TVP development and because Chinese TVPS can compete more easily with the inefficient state sector. Another public ownership com- parator, small-scale rural manufacturing in Tanzania, is briefly analyzed, and it is suggested that one reason for the poor performance of these firms has been their limited fiscal and other linkages with local govemments. The authors then turn to international comparators under private owner- ship. The Mondragon group in Spain's Basque region has some important similarities with China's TVP sector. Gelb and Svejnar argue that Mondragon represents an alternative path for regional development and rural industriali- zation that may become more relevant for China if the TVP sector moves to- ward more private and autonomous, yet still group, ownership patterns in an environment of continuing labor immobility. Other, in certain respects more remote, examples include U.S. company towns, regulation of firms by Swiss cantons, and Japanese-type Z-firms. The last, as noted in chapter 13, share many characteristics with Chinese TVPs and may be highly relevant for China. Gelb and Svejnar then look at tendencies for firms to change ownership forms. In market economies dominated by capitalist ownership, many suc- cessful worker cooperative firms eventually are sold to private ownership be- cause it is impossible to find new workers with sufficient capital to replace the membership shares of those who retire or leave. The opposite transition, from capitalist to worker ownership, can be triggered in a number of ways, including bankruptcy and sale to employees, benevolent voluntary transfers by owners, and transfer of stock to employees for the sake of tax advantages or to mobilize capital. The last part of the chapter assesses the ability of firms to adjust to adverse changes in economic conditions, looking in particular at whether enterprises respond to a downturn by cutting employment or reducing wages. There is some evidence from international experience that communal and Z-firms tend to cut wages rather than employment. Mondragon is a notable example. The response of Chinese TVPS to adversity cannot yet be fully assessed, but it may well be that community enterprises will prefer wage reductions and that private enterprises will engage in other responses, depending on such factors as the number of persons from outside the local community in the work force. 18 Convergence: A Comparison of Township Firms and Local State Enterprises Song Lina Urban-rural dualism in China's economic structure is disintegrating, and industrial developments in urban and rural areas are complementing and infiltrating each other. The factors that affect urban industrialization also affect rural industrialization, directly or indirectly. As extensive commnercial relations are established between urban and rural industrial systems, sepa- ration or semiseparation between urban and rural economies is breaking down. In a sense, rural industrialization in China has taken the urban industrial system as its model. Since their emergence TVps have found themselves in a relatively open "market" environment. Yet this is not a market mechianism based on the principles of fair competition and equal opportunity. Although the situation has improved, discrimination by government agencies, the mar- ket, and the public still exists, and industrial TVPS therefore seek protection. On the one hand, they have to struggle against an unfair economic structure, the consequence of the traditional planned system; on the other, they long to be admitted into the planned economy and to come under government protection. These characteristics are particularly prominent in firms adminis- tered by township governments. Urban state enterprises, by contrast, have begun to feel the weakening of govemment protection that has followed the relaxation of control through central planning. Disturbances in the newly emerging markets have affected them from time to time, and they want to tum to new norms of behavior, but administrative restraints limit their choices. TVPS, which are freer from administrative restraints and more responsive to the market, offer a model for the reform of state enterprises. Both township and state enterprises react ambivalently to their environ- ments. They are aware of the barriers to their development and wish to over- come them through reform, but they also have a habit of dependence on the old institutional frameworkcs and administrative environments. As they seek new paths, they may look to each other's behavioral pattems for norms 392 Convergence: A Comparison of Township Firms and Local State Enterprises 393 lacking in their own systems but conducive to development. The consequent cross-pollination between the behavior of state and township enterprises is leading to a convergence of systems. We hypothesize here that state and township enterprises are becoming more and more similar in operations and structure. This chapter compares the environments, goals, operations, and efficiency of two types of enterprises-industrial township enterprises and local state en- terprises (LSES) in small cities and in counties-with a view to explaining their recent evolution. For the sake of a more scientific and persuasive com- parison, only LSES in the same localities as our sample township enterprises have been chosen for this study. This helps eliminate structural variations and noncomparable elements of regional development and local traditional culture. The data on LsEs are primarily from the industrial statistics of Wuxi city, Shangrao city, Foshan city, and Jieshou County. In addition, ninety Worker Survey Questionnaires were distributed to workers in county-administered LSES in Nanhai and Shangrao, and eighty-one were retrieved. The Enterprise Survey Questionnaire and the Enterprise Director Questionnaire were also distributed to LSES in the two counties. Case study materials and interviews are another important source of information. Structural Similarities between Urban and Rural Communities Superficially, a city differs greatly from a township. For example, the popula- tion of Shangrao city is 142,500, 6.3 times the average population of sample townships in Shangrao County (22,800). The original value of the total fixed assets of LSES in Shangrao city is Y101 million, 5.7 times the average for sam- ple townships in Shangrao County (Y18 million). In 1985 city-administered LSES in Shangrao city paid Y13 million in taxes, profits, and fees, 317.5 times the average (Y40,000) for the township sample. The gross social output value of Shangrao city was Y153 million, 21.5 times that of the average township (only Y7 million). But despite these differences in size, urban and rural com- munities in China are surprisingly similar. The Composition of Economic Sectors Even the smallest community generally includes the five principal sectors- industry, agriculture, commerce, construction, and transport-but the shares can vary. Table 18-1 covers two pairs of proximate communities. In Wuxi city and Wuxi County industry has an overwhelming weight as measured by gross value of output. In the sample townships of Shangrao County, agricul- ture has a substantial share. All sectors, however, regardless of their relative weights, are generally represented in each community. 394 SONG LINA Table 18-1. Composition of Economic Output, Wuxi and Shangrao, 1985 (percentage of total) Communicadions community Iniustry Agriculture Constructon and transport Commerce Wuxi city 89.3 1.0 3.9 1.4 4.3 Wuxi County, sample townships (average) 82.9 7.4 5.8 0.9 3.1 Shangrao city 64.3 3.4 11.4 6.6 14.2 Shangrao County, sample townships (average) 40.3 31.0 18.7 3.8 7.6 Sources: Township Leader Questionnaire and information from fieldwork. Ownership and Administrative Structures As the outline below shows, there is a parallel structure of enterprise owner- ship in cities and townships. City Township LSE5 Township enterprises Big collective enterprises Village enterprises Small collective enterprises Partnerships Private industrial and commercial Private enterprises enterprises The administration of' enterprises under different forms of ownership also exhibits common characteristics. For instance, small collective and private enterprises in cities, like rural private enterprises and partnerships, are under the supervision of local industry and commerce administrations. Both LSES and township enterprises are directly subordinate to the responsible depart- ments of the respective community government. (Whereas in other chapters, community government refers to rural governments at the township, village, and production team levels, in this chapter it also includes county and small city governments.) Big collective enterprises in cities are on the same level as village enterprises; both are partly under the administration of govemnment departments but maintain relative independence in property relations. They are thus less subject to government influence than are LSEs. Labor Composition Cities and townships still differ greatly in their urban-rural population mix. For example, 79 percent of the population in Shangrao city is urban and only 21 percent is rural, whereas, on average, only 7 percent of the population of sample townships of Shangrao County is urban and 93 percent is rural. This contrast, however, is no more than a matter of household registration. Agricultural labor has been rapidly shifting to nonagricultural activities as rural industrialization proceeds (see table 18-2), and the populations and Convergence: A Comparison of Township Firrs and Local State Enwrprises 395 labor compositions of rural communities are gradually becoming more like those of urban communities. Kinship and Native Ties It is generally believed that traditional ties of kinship, clan, and village are much more important in TVPS than in urban enterprises. But this is not en- tirely true. Table 18-3 shows the original homes of workers in sample TVPS. The survey of LSE workers indicates that 27.1 percent are from other coun- ties, cities, or provinces, while the rest are residents of the city or county in which their firm is located. Most enterprises in China recruit their workers in the vicinity, and LSES and TVps are no exception. We will define local, or native, TVP workers as those who come from within the township and local employees of county LSEs as those who reside within the county. Under these definitions natives account for 80.4 percent of sample TVP workers in Wuxi, 70.8 in Nanhai, and 90 percent in Shangrao; in the sample of county LSES 72.9 percent of the workers are natives, which is not significantly different from the percentages in TvPs. Furthermore, TVps have begun to extend their recruiting spheres. Many TVP leaders noted in interviews that because workers recruited among kin and friends could become difficult to handle, they were considering recruiting workers from outside the community. Although doing so would cost the enterprise more in expenditures on shelter and meal facilities for workers, the TvPs would profit by bringing together productive factors from larger areas. The localist and clannish character of TVPS may be weakened by this trend. Enterprises in urban communities have evolved in a different direction. In the past few years unemployment among urban youth has been serious, and many state enterprise workers have retired to vacate places for their sons and daughters. In addition, many state enterprises have established affiliated enterprises ("labor service companies") to create employment for the grown children of their workers (see Byrd and Tidrick 1987). Consequently it is common in urban communities for several family members to work in the same firm. Table 18-2. Changes in the Labor Force, 1978-85 (annual percentage change) Sample toumships, by county industrial Agricultural Wuxi 18.8 -12.9 Jieshou 32.4 -10.5 Nanhai 12.6 -12.7 Shangrao 12.3 2.5 Source: Township Leader Questionnaire. 396 SONG LINA Table 18-3. Otiginal Homes of TVP Workers (percentage of total sample workers in each county) Odter Other villa, township, Sarne same same Odher Other County village township county county province Wuxi 58.6 21.9 8.8 5.2 5.4 Nanhai 53.0 17.8 16.5 9.7 3.0 Shangrao 67.1 22.9 3.6 2.1 4.3 Note: Percentages may not add up to 100 owing to rounding. Corresponding infonnation for Jieshou is not available. Source: Worker Survey Questionnaire. In sample LSES in Nanihai and Shangrao counties each factory director has an average of 1.25 relatives vvorking in the firm; for township enterprises in the same counties the average is 1.54. Urban unemployment has put pressure on the recruiting system of state enterprises. When the govemment cannot solve the problem of employrnent, enterprises and families have to dlo it. In rural communities reliance on enterprises and family connections to provide employment for surplus labor began earlier, with the adoption of the PRS. Thus, as the employment mechanism shifts from govemment control to local absorption, cities and townships are becoming more alike. The Influence of Ownership Systems on Convergence The patterns of relationships between community governments and enter- prises under various forms of ownership are similar in counties and townships. A sort of father-son relationship exists between the county government and its subordinate LSES, as well as between the township government and its sub- ordinate enterprises, but both governments have relatively remote connec- tions with other types of firm. This is because both have ownership rights over the assets of their subordinate firms and depend on them for revenue. Although community governments do make startup investments irn subor- dinate enterprises, the amount of capital they put up is not necessarily large. They also often provide intangible assistance. For example, they may use their influence to help firms obtain bank loans and business licenses. They also requisition land, recruit workers, secure scarce raw materials, and find technology and markets for their enterprises. Interdependence Even if county and township governments do not invest much in enterprises, they are still dependent on them. Their relations with enterprises center on Convergence: A Comparison of Township Firms and Local State Enterprises 397 revenue sharing (see chapter 16); a large share of their revenues comes from taxes and profits tumed over by subordinate enterprises. For example, reve- nues from subordinate LSES and administrative departments accounted for 71.5 percent of Shangrao city's extrabudgetary revenue in 1985, up from 63.9 percent in 1984. The bulk of the extrabudgetary revenues of sample township govemments also were from enterprises, primarily from taxes, fees, and prof- its remitted by township firms. Communities began industrialization at different times and are at different levels of development. This affects community govemments' orientation to- ward the use of extrabudgetary revenues. In Wuxi sample townships spend, on average, 30.5 percent of these revenues on township enterprises, directly or indirectly, but the figure is only 22.1 percent in Jieshou. Similar spending pattems can be found in city govemments. For instance, Shangrao city spends 35.1 percent of its extrabudgetary revenues on enterprises. Management When enterprises tum over to community govemments more in taxes, fees, and profits than the amount retumed by the govemments, the enterprises may have trouble accumulating capital and may find themselves at a disad- vantage in an increasingly competitive market. But community govern- ments, as owners of these firms, may also protect them. In this sense, city and township govemments have functional similarities in the management of subordinate firms. County and township govemments are at different administrative levels and have different privileges in the hierarchy of the centrally planned econ- omy. This situation existed earlier and is continuing in the reform period. In all activities, from the central financing of lower levels of govemment to the planned distribution of needed inputs among state enterprises, the amounts always decrease progressively down the hierarchy, and very little trickles down to township governments from the provincial, city (or pre- fectural), and county governments. The reforms have, however, substantially reduced the control of state en- terprises through central planning. The number of materials directly under state control has been cut from 256 to 20 categories; the number of state- controlled industrial goods was reduced from 123 categories in 1980 to 66 in 1986; the coverage of the value of industrial output by the state mandatory plan has been reduced to 20 percent; and depreciation allowances are begin- ning to be totally at the disposal of state enterprises. Revenue-sharing sys- tems in public finance have given lower levels of government more discre- tionary funds and hence more flexibility, provided that they meet revenue targets. 398 SONG LINA Quasi-Planning Systems The increased funds at the disposal of community governments and the weak- ening of central controls have created a foundation for the establishment of local protective quasi-planning systems. These systems provide benefits for enterprises, particularly those established by community governments, so that they can be counted on to turn over taxes, fees, and profits. In a sense a quasi-planning system operates like a traditional planning system: it tries to infiltrate every link in the operations of the enterprise, from personnel, finances, and materials to production, supply, and marketing. But commu- nity governments can no longer impose inflexible administrative restraints on enterprises, for the extent to which they can control or support them is limited by their own capabilities. Regulation by the community government's quasi-planning system is effective only where the administrative system is highly developed (as in Shangrao) or where the community govenmment is skillful in organizing the economy (as in Wuxi). The administrative capabili- ties of township govenmments vary greatly, and so can their quasi-planning systems. In a macroenvironment where supplies of and demand for inputs and en- ergy are not balanced, community govemments' desire for goods and materi- als is a clear indication of their concern for the development of their enter- prises. Both county and townLship govemments draw on two sources-the central planning system and local efforts-to build up a reservoir of inputs needed by enterprises. Inputs from the central planning system-including even college graduates assigned through personnel departments-can in- creasingly be obtained by monetary means; only the prices (based on cost) vary. In recent years the materials supply departments of city and county govern- ments have often obtained needed inputs through exchange. The materials bureau of Wuxi County is outstanding in this regard. Of the materials sold through the purchase and sales companies of the bureau in 1985, only 5 per- cent were from the state planning system. The rest were obtained by ex- changing materials with or investing in units that produce goods in short sup- ply, such as steel and electricity. In the first half of 1986 the bureau did a business of Y400 million, more than fifty-seven times that done by its coun- terpart in Jieshou. Although raw materials obtained through exchange or in- vestment are more expensive than those assigned by the state plan, the prices are no higher than on the markcet, and the supply is more regular. Therefore firms under various forms of ownership try their best to be admitted into the quasi-planning systems. Although it is almost impossible for township governments to get inputs from the central planning system, they can obtain some materials through their subordinate departments (industrial companies or administrative offices for TVP administration), through multilateral exchanges or investments, or Convergence: A Comparison of Township Firms and Local State Enterprises 399 even through personal relationships. In addition, township governments can obtain from the county materials department medium-price raw materials- inputs priced between the market price and the state plan price. In the distribution of low-price materials, city and county govemments give priority to subordinate LSES and big collective enterprises. Similarly, township governments give priority to their subordinate enterprises. Re- sponses to the Township Leader Questionnaire show that township govern- ments provide extremely few materials to partnerships and none at all to indi- vidual firms, whereas significant amounts are made available to TVCES. Means of Control over Enterprises The supply of favorably priced inputs by community governments to subordi- nate firms is a means of compensating for overtaxation and creating a more favorable environment for them. But community govemments need addi- tional control measures if they want their enterprises to keep bringing in rev- enues and at the same time to have good prospects for development. At present, inputs to China's state enterprises provided through state plan- ning account for only an estimated one-third of the total value of needed inputs, and two-thirds come from the market. Meanwhile, local governments have gradually reduced their planning control over LSES. Ten LSES in Shangrao and Nanhai reported that they no longer depend on state planning for needed raw materials and that state mandatory plans cover only one- eighth of their total output. Local governments, however, have not relaxed Table 18-4. Decisionmaking in Sample LSEs and Toumship Enterprises (percentage of sample) Made by factory Not made by factory Type of Touwship Township decision LSEs enterprises LSEs enterprises Appointment of factory director 30 13.3 70 83.3 Formation of management team 50 50.0 50 40.0 Worker recruitment 30 28.8 70 71.9 Wages 10 18.8 90 62.5 Large production-related expenditures 0 87.5 100 12.5 Production planning 80 50.0 20 18.8 Product pricing 40 n.a. 60 n.a. Proportion of profits to be retained 0 19.4 100 71.9 n.a. Not available. Note: Since not all enterprises responded to all parts of the question, the percentages do not always add up to 100. Source: Enterprise Director Questionnaire. 400 SONG LINA their supervision of enterprises correspondingly (see table 18-4). E.xcept for production plans, LSEs do not seem to have much decisionmaking power. Community governments impose strict control over LSES through their au- thority over personnel and financial matters to make sure that due amounts of taxes, fees, and profits are remitted. For instance, in LSES under the Nanhai County govemment, cadres ranking above section chief (that is, factory lead- ers) are officially appointed by the county government, whereas cadres from section chief down are appointed by responsible county departments. Al- though production plans are made by enterprises, decisions on products and output are made by the counity economic committee or its subordinate indus- trial companies. The sale cf products is also controlled by superior depart- ments, and only products in excess of the quota are marketed by the enter- prises themselves. When it comes to important investments, govemment departments represent LSES in negotiations with banks and financial depart- ments. Decisions on selecting projects or changing products are made by gov- ernment departments. The core of the relationship between LSEs and local govemments is the proportion of profits remitted. Although govemments own the assets of LSES, they cannot avoid adversely affecting the accumulation of enterprise assets in the exercise of their functions. County governments do not behave like independent asset owners; they are concerned not only with increasing the value of enterprise assets but also with carrying out their government func- tions with the help of revenues from enterprises. The consequent: conflict between enterprise development and government activities is manifested in arguments over the amount of profits to be tumed over by LSES. Most of the LSES we studied bargained with government departments over profit reten- tion. This suggests that the functions of community governments undermine the integrity of management of LSES. Township enterprises face the same problem; the decisionmaking power of sample township enterprises is cur- tailed by township governments and their subordinate departments (see table 18-4), sometimes to an even greater extent than in LSES. County and township govemments have fundamentally similar owvnership relations with their subordinate firms. These two types of firm enjoy high po- sitions in their respective communities, receive considerable govemment protection, and are the main sources of revenue for community governments. Therefore LSES and township enterprises have essentially similar relations with their superior community govemments, and their behavior is similar or is becoming more so. Similarities among Managers There are many parallels between managers of LSEs and township enterprises. This section reviews similarities in perceived responsibility to the govem- ment and to the firm, the characteristics of managers, the criteria by which Convergence: A Comparison of Township Firms and Local State Enterprises 401 managers are evaluated, management objectives, and problems stemming from managers' lack of real authority. Responsibility The ownership of LSES and township enterprises is currently rather vague. It is difficult to distinguish or quantify the assets generated by the initial or backup investment by the community government and those attributable to reinvested profits or self-accumulation. As managers of the enterprises, fac- tory directors cannot represent only the owners or the enterprises; they must look after the interests of both sides. Their actions therefore exhibit clear traces of conflict. Directors' strong sense of identification with the government-owner can be inferred from the way in which they come to their positions. In sample township enterprises of Nanhai and Shangrao 88.3 percent of factory direc- tors were directly engaged by the township government, and very few were elected by the workers or hired under contract. In sample LSES of these two counties 70 percent of factory directors were appointed by the government; few came to their positions in other ways. Almost all directors of sample LSES see themselves as responsible first to government departments and only secondarily to the workers. The majority (85.7 percent) of directors of sample township enterprises took the same view. Managers cited "submission to assignment" as their main motivation for taking the position (table 18-5). Characteristics of Factory Directors The differences between directors of LSES and of township enterprises are not very significant. Directors of sample LSES have more education (see table 18-6), but township enterprise directors are generally experienced people with a rich knowledge of nonagricultural activities and are not so different from their counterparts in LSES. In any case, a middle school education can be enough for daily management work in TVPS. Table 18-5. Motivations for Becoming a Factory Director (percent) Motivaion LSEs Toumship entrprises Submission to assignment 88.9 86.7 Sense of social responsibility 88.9 73.3 Interest in the work 33.4 66.7 Improvement of social position 22.3 66.7 Improvement of economic position 0.0 10.0 Greater freedom 0.0 0.0 Source: Enterprise Director Questionnaire. 402 SONG LINA Table 18-6. Educational Levels of Factory Directors (percent) Educational Township enterprise level LSE directors directors College graduate 55.6 0.0 High school graduate 11.1 23.3 Middle school graduat:e 22.2 56.7 Primary school graduaote 11.1 20.0 Source: Enterprise Director Questionnaire. In our samples the ages of the two groups of factory directors fall basically in a normal distribution between 30 and 50 years. Directors of LS:Es have, on average, 14.9 years of work experience and a cadre career (including as factory director) of 13.3 years. Directors of township enterprises are less expe- rienced; they have, on average, 8.8 years of work experience, and their aver- age cadre career is 7.8 years. But 90 percent of the latter group have served in the military, been purchasing agents, or done business before. These ex- periences undoubtedly influence their present work and make therm no less capable than their counterparts in LSES. They are outstanding people in their own communities. Criteria for Assessing Factory Directors Township governments use several criteria to evaluate township enterprise managers. Managers must be able to expand the firm (an indication of mana- gerial ability); they must be conscientious and honest in their work (personal- ity); they must be popular and supported by workers (ability and personality combined); and they must turn over a set amount of revenues to the township govemment (the core relation between firm and government). The criteria used by county govemments with respect to LSE directors are basically similar. In our investigations we also leamed how the two groups of factory direc- tors assess their own abilities. Their selections of desirable personal attributes overlapped on four items: sense of responsibility and drive, technical ability, management ability, and unclerstanding of the market. But directors of LSES chose decisionmaking ability, innovative ability, and democratic ideas as im- portant characteristics, whereas their counterparts in township enterprises chose sense of principle and competitive spirit. Management Objectives of Factlory Directors Differences emerge in thLe choices of management goals by factory directors (see table 18-7). The first three management targets chosen by directors of LSES center on the self-interest of the enterprises, although they do not over- Convergence: A Comparison of Township Firms and Local State Enterprises 403 Table 18-7. Managerial Objectives Rank Tounship enterprise (I = most important) LSE directors directors 1 Promote stable and long-term Promote prosperity development of enterprise of township 2 Produce brand-name goods Improve intemal to raise credibility management 3 Maximize profits; improve Increase financial internal management income of township 4 Increase financial income Maximize profits of locality Source: Enterprise Director Questionnaire. look contributions to the finances of the locality. The choices made by direc- tors of township enterprises, however, include some community government goals: promoting the prosperity of the township and increasing the financial income of the township were ranked first and third in importance, whereas maximizing profits was only fourth. Township enterprise leaders have internalized the goals of community gov- emments to a greater extent than have LSE directors. The comparatively cen- tralized exercise of governmental functions by city and county govemments is closely related to stable financial income and better methods of taxation. Municipal and county govemments are therefore able to engage in much more orderly and standardized management of subordinate firms. Instead of interfering directly, leaders usually make contacts with factory directors through government departments. In rural communities, although such institutions as township economic commissions and industrial offices do exist, township govemments are in- volved more directly in management of township enterprises, for several reasons. • Township communities own fewer enterprises than do county govem- ments. If the leaders of county govemments directly handled the affairs of every enterprise, their workloads would be unmanageable. * To establish an orderly and regular management system takes time. In rural industries a complete reporting system has not yet been estab- lished. The "farmers' style" of management that has evolved in agricul- ture over several thousand years is a chronic problem that must be over- come. * The smaller the community, the more conspicuous the community in- terest, and thus the closer the agreement between the factory director and the community government. In townships, the contribution made by township enterprises to the community is obvious, and the township 404 SONG LINA enterprise director has an important position locally. The factory direc- tor not only has to manage the operations of his own firm but also shoulders the heavy task of community development. In urban communities relarions between LSE directors and community gov- emments are much simpler. Directors only have to stick to fixed targets and regular financial channels, and they do not usually have to assumre many other burdens. Their contribution to the development of their communities is made through tax payments. Unlike township enterprise directors, they seldom have to be entangled in trivial matters. The Lack of True Management Authority Although the managers of the two types of enterprises face different financial environments, their powers are similar and are basically limited to choosing the management team of the enterprise and deciding on the annual produc- tion plan. They cannot: do mruch actual decisionmaking, but they are often partly responsible for en-terprise risks. Under the traditional socialist mode, managers, in the basic sense cf the word, did not even exist. Enterprise direc- tors were administrative officials who were appointed by govemment depart- ments. As far as LSE and township enterprise managers are concerned, the mixture of enterprise property rights and management rights naturally leads to divided self-interest. They have to conform with the wishes of their superior commu- nity govemments, but they also have to ensure the development of their en- terprises. In the sample, only One L:SE director felt he would not face punishment for enterprise bankruptcy caused by improper management, whereas the fig- ure is as high as 13.3 percent for directors of sample township enterprises. Most directors of both types of firm face penalties from community gov- ernments-loss of bonuses, reduction of salary, or even dismissal--if they cause losses. Despite the harsh restrictions on their personal interests, manag- ers have power in name only, and they are not able to take effective control of firm operations. Workers' Characteristics Township enterprise workers are former (or even current part-time) farmers, whereas LSE workers are urban residents. But in general, differences between the two sets of workers are not as great as might be imagined. This section examines the workers' relationship to the land, their personal characteristics, wage systems and benefits, and attitudes, including the intensity of their de- sire to participate in enterprise decisionmaking. Convergence: A Comparison of Township Firms and Local State Enterprises 405 Ties to the Land Unlike workers in most urban communities, employees of township enter- prises are farmers who have just left the land or have not yet completely bro- ken away from agricultural production. In county LSES located near rural com- munities a sizable portion of employees also are fresh from the farm, and some workers' families still live in rural areas. These employees therefore share some similarities with TVP workers. The average household size for sample LSE workers in Nanhai and Shangrao is 4.1, whereas the figure is 5.8 for households of TVP workers in the two counties.' The number of household members engaged in nonagricul- tural production averages 2.2 in the former and 2.7 in the latter. Factory wages are the main income for 92.6 percent of the households of sample LSE workers. Most TVP workers (77.3 percent) also consider income from indus- trial activities to be the main income of the family. Although the economic position of LSE and TVP workers is closely linked to industrial production, it is still, to different degrees, under the influence of the land. In the sample, families of workers in county LSEs hold, on aver- age, 1.0 mu of contracted land; the corresponding figure for TVP workers is about 3.3 mu. Only 27.2 percent of LSE workers' families till their own land; most families employ hired hands or subcontract the land to other people. The majority (84.1 percent) of TVP workers' families still take care of their own land, and households that have contracted all their land to other people account for less than 5 percent. Despite these differences, the number of workers who personally take part in agricultural production is small in both types of firms: 7.4 percent for LSES and 10.0 percent for TVPS. Table 18-8. Educational Levels of Workers in Nanhai and Shangrao (percentage of sample) Educational level LSE workers TVP workers No schooling 0.0 2.1 Primary school 2.5 23.9 Higher primary school 13.6 17.8 High school (secondary vocational school) 38.3 39.3 Senior high school 25.9 17.5 Higher education and technical and vocational training schools 11.1 8.9 Note: Since not all workers responded to this question, percentages do not add to 100 percent. Source: Worker Survey Questionnaire. 406 SONG LINA Personal Characteristics Among sample LSE workers in Shangrao and Nanhai, 76.5 percent are male and 23.5 percent female; in TVPS 48.8 percent are male and 51.2 percent female. The average age of sample TVP workers is 27.8 years; in ILSES it is 35.7. LSE workers had an average of 1.3 jobs before their present employment; TVP workers had 1.2. The average length of service in the firm is about 11.0 years for LSE workers, and their total experience in industrial activities aver- ages 15 years. The average length of service in the present job for TVF' workers is 5.5 years, and their total experience in nonagricultural activities averages 8 years. The educational level and professional qualifications of sample work- ers in TVPs are only a bit lower than those of LSE workers (table 18-8). The Wage System Sample TVP workers earn Y1,434 a year, whereas sample LSE workers earn Y1,302 a year. (According to Nanhai County officials the wages of USE work- ers are generally about 20 percent lower than those of local TVP workers.) The gap may well reflect the different wage systems adopted by LSEs and TVPS. Among sample LSE workers 55.6 percent are on a fixed salary system and only 6.2 percent have piece-rate wages. In contrast, only about 8 percent of sam- ple TVP workers are paid by the hour, and the rest are subject to some form of piece-rate payment. Generally speaking, piece rates effectively mobilize the enthusiasm of workers, and as a result TVPS can attract workers employed by LSES. Many state enterprise workers, especially technicians, would like to work in TVPS. In the state-owned Nanhai Silk Factory alone, 15 workers, mostly skilled technicians, quit their jobs to work in TvPs during January-February 1987; 143 people left the factory in 1986. One-third of the current employees, in- cluding some managers, have part-time jobs in TVPs. According to the factory director, the wage rate for workers in local TVPS in the textile indlustry is YO. 12-YO.28 per meter of cloth, as against only Y0.06 per meter in LSES. An ordinary worker in a state textile firm could see his pay rise from Y100 to Y300 a month if he became a Tvp director. Although the wage system in LSES is not as flexible as in TvPs, many state enterprise directors are beginning to consider the reform of the wage system. But since not all problems can be solved by an enterprise director, real change depends on the improvement of macropolicies. As the higlh-salary attraction exerted by TVPs has caused a labor shortage for LSES in some profes- sions, the LSEs have hadE to recruit workers from regions where rural industry is underdeveloped. An LSE in Nanhai hired more than 2,000 temporary work- ers from Guangxi Province-equal to the number of local staff. The factory has the right to give these temporary workers piece-rates. Eventually, piece- rate systems may replace fixed wages or payment by the hour in LSES. Convergence: A Comparison of Township Firms and Local State Enterprises 407 The Welfare System Although the wages of workers in LSEs are on the low side, their welfare treat- ment is generous. In the sample 60.4 percent of TVP workers are not entitled to housing and subsidies from the enterprise, 41.0 percent of the workers do not receive financial aid for medical care, 52.3 percent are not covered by insurance on the job, 53.9 percent do not get pensions on retirement, and 58.4 percent cannot have their children replace them. By contrast, workers in sample LSEs are entitled to all of these benefits. Some township govern- ments are trying to provide state enterprise-type welfare treatment for their TVP employees. The government in Lingxi township, Shangrao County, has implemented a series of welfare measures for its eight subordinate enterprises. Workers are classified as permanent employees, appointed by the township government, or temporary employees. Permanent employees are entitled to free medical care and pensions and cannot be freely dismissed from their jobs. Workers' Attitudes Differences between workers in county LSEs and in TvPs are not as large as people imagine. Their similarities are especially evident in their attitudes. WORKERS' SENSE OF STABILITY. Among sample TVP workers in Nanhai and Shangrao, only 13 percent fear that they might lose their present jobs, while the rest have considerable confidence in the stability of their jobs. Among sample LSE workers in the two counties, 6 percent believe it is possible for them to lose their jobs. ATTITUDE TOWARD RISK. About 64 percent of sample TVP workers have never planned to run or jointly run a factory with other people. Among sam- ple LSE workers in the two counties, the figure was 59 percent. Only 6-7 percent of sample workers in both types of enterprises actually plan to run or jointly run a factory. ATTITUDE TOWARD THE STATUS OF TVP WORKERS. Among sample TVP workers 16.5 percent agree with the statement that they are one rank lower than state enterprise workers. Among sample LSE workers 19.8 percent hold that opinion, but most do not agree. PARTICIPATION BY WORKERS. TVP workers show a weak desire for and a low degree of participation in enterprise management. The actual degree of par- ticipation by workers in LSES is also low. The proportion of sample LSE work- ers who take part in decisions on appointment and dismissal of enterprise leaders accounts for only 16 percent, and they have even less power in other respects. LSE workers, however, hope to have power in some spheres such 408 SONG LINA Table 18-9. LSE Worzers' Desires for Participation Item Total frequency Precedence Appointment or dismnissal of firm leaders 43.2 1 Development of new products 12.3 2 Use of profits kept by enterprise 9.9 3 Dismissal of workers 6.2 4 Decisions on wages and boniuses 6.2 4 Distribution of work within the factory 2.5 6 Recruitment of workers 0 - New investment 0 - Not applicable. Source: Worker Survey Questionnaire. as appointment or dismissal of enterprise leaders. Table 18-9 shows how work- ers rank their desire for participation in specific areas. These choices indicate that LSE workers' desire for actual participation is low because they think that a good and reliable enterprise management team will run the firm in such a way as to promote their interests. Comparison of Ecoinomic Benefits Comparisons of the economic efficiency of the two types of firms suggest that they are the enterprises with the lowest efficiency in their respective commu- nities and that township enterprises are more efficient than LSEs. This section probes into possible reasons for the latter phenomenon, with specific illustra- tions from the textile industry. Efficiency National statistics on output-capital and profit-capital ratios in state enter- prises and urban collective industrial firms are shown in table 18-10. The ratios are much lower for state enterprises than for urban collective firms. The labor productivity of inclustrial state enterprises, however, is Y15,198 per person, whereas the figure for urban collective enterprises is only Y8,206 per person. Similarly, the economic benefit index of sample township enter- prises in the four counties is lower than that of firms under other ownership systems. (See chapter 10 for detailed statistics.) Owing to similarities in ownership relations and in external environments, the behavior and systems of LSEs and township enterprises tend to converge. In general, both kinds of firm have relatively low efficiency in their respective communities. This conclusion is supported by statistics on industrial firms in Wuxi (see table 18-11). Convergence: A Comparison of Township Firms and Local State Enterprises 409 Table 18-10. Profitability in State and Urban Collective Enterprises, 1985 State enterprises Urban collective enterprises Output value Profit Output value Profit Type of per YIOO of per YIOO of per Y100 of per YIOO of industry fixed assets fixed assets fixed assets fixed assets Heavy 66.9 10.7 204.2 26.7 Light 209.2 19.4 321.0 21.0 Total 95.0 12.4 269.6 23.5 Source: State Statistical Bureau (1987b). Why Township Enterprises Are More Efficient than LSEs Table 18-12 shows that the economic benefits of township enterprises are higher than those of state enterprises at the national level. Analysis of eco- nomic benefits in industrial enterprises in Wuxi reveals the same phenome- non: output-capital and profit-capital ratios in township enterprises are higher than in county LSES (table 18-11). Why are there substantial differ- ences in economic benefits between LSEs and township enterprises even though their environment, operations, and internal systems are converging? Several possible reasons are discussed here. ACCOUNTING SYSTEMS. The accounting system in TVPS is much weaker than in state enterprises because bookkeeping in TVPS is accountable only to the owners. Leaders of township governments will not severely punish their firms for cheating and giving false accounts to government agencies such as the tax, industrial and commercial, and statistical departments; in- deed, they often support such actions. Moreover, not all of the income and expenditures of TVPs have to go through banks, and the enterprises gain con- siderable extra scope for their activities because they can pay less tax and evade banks' restrictions on the use of cash. The bookkeeping activities of state enterprises are controlled by a uniform nationwide accounting inspection system that covers almost every link in the operations of an enterprise, from the wages of employees to expenses for busi- ness trips and from production expenses to nonproduction expense items. Hence accounting management in state enterprises limits the scope of enter- prise activities. Many state enterprises have begun to set up their own small "banks" in recent years, but they are still far less flexible than TVPS. The purchasing and marketing agents of TVPS have sufficient cash and are allowed to offer some individuals gifts, meals, commissions, and the like in return for materials in short supply or for sales opportunities. TVPS can also easily manipulate their business accounts to lower their taxes or to increase 410 SONG LINA Table 18-11. The Relative Efficiency of Different Types of Industrial Finmns in Wuxi, 1984 (yuan) County-owned County couective Township Village Item LSEs firms enterprises enterprises Total output value created by Y100 of capital 202.1 232.8 228.7 277.4 Total profit created by Y100 of capital 23.4 23.7 25.9 33.0 Total tax and profit created by Y100 of capital 31.7 31.6 37.7 45.1 Source: Information from fieldwork. total wages without incurring, for example, a bonus tax. If a state enterprise were to do these things, it would run great risks because of the restrictions under the existing systern and the accounting controls that have been written into the country's laws. These are a great deterrent to everyone in state enter- prises, and only a few dare to defy the law. COST STRUCTURE. Many TVPs have lower costs than state enterprises, as is illustrated by a comparison of costs in the silk industry in Nanhai County (see table 18-13). There are several reasons for these cost differences. First, TVPS (including township enterprises) enjoy tax exemptions during the first several years of operation. Some firms, mainly private enterprises, announce bankruptcy as the tax exemption period draws to an end and then set up again tnder a new name. According to estimates by tax officials in Rongcheng County, Hebei Province, TVPS in that locality have hidden at least 30 percent of their profits. Second, the welfare system iFor workers in state enterprises is quite compre- hensive. State enterprises at all levels offer such benefits as payment of medi- cal expenses (sometimes including part of the medical costs for family mem- bers), insurance on the job, subsidies for heating in winter and cold drinks in summer, and pensions. One factory director estimated that workers' wel- fare expenses amounted to 11 percent of the factory's total wage bill and the trade union fee to 2 percent. Workers in TVPS can never hope to catch up with state enterprise workers in welfare benefits, but these benefits do in- crease production costs for state enterprises. Third, management costs are higher in state enterprises than in TVPS, by about 50 percent according to table 18-13, mainly because of overstaffing and complicated managerial structures in state enterprises. The director of an LSE in Nanhai noted that the factory has ten sections and offices and that none can be abolished because they correspond to those in the responsible govern- Convergence: A Comparison of Township Firms and Local State Enterprises 411 Table 18-12. Economic Efficiency in State and Township EnterpTises, 1985 State entrprises Township enterprises Output value Profit Output value Profit Type of per Y100 of per YIOO of per YIOO of per YI1O of industry fixed assets fixed assets fixed assets fixed assets Heavy 66.9 10.7 176.5 19.6 Light 209.2 19.4 256.2 18.5 Total 95.0 12.4 207.1 19.2 Source: State Statistical Bureau (1987b). ment departments. An LSE does not have control over which intemal depart- ments to set up. As a result, managerial personnel account for about 10 per- cent of the total number of employees in the enterprise. By contrast, TVPS spend very little in this area and have not built up a managerial hierarchy. Often the owner and manager are the same person, and the entrepreneur takes on many duties. Furthermore, according to the director of an LSE silk factory in Nanhai, management fees for LSEs are higher than those for TVPS, especially private enterprises. Fourth, LSES make higher interest payments to banks than do TVPS, mainly because they take out more loans. In the latter half of 1984 the Agricultural Bank in Nanhai County inaugurated stricter measures to control the huge amount of loans demanded by partnerships and individual enterprises, and the amount of lending to them declined. Perhaps more important, the total amount of capital required by TVPS per unit of output (including working cap- ital and fixed assets) is lower than for state enterprises. Fifth, TVPS often use low-cost or substandard raw materials. For instance, when TVPS use polyester-cotton blended yam instead of pure cotton yam, the price difference can well be more than Y2,000 a ton. For some other such substitutions, the price difference is as much as Y4,000 a ton. Sixth, labor productivity in TVPS that produce silk textiles is about 10 per- cent higher than in LSES because the comparatively higher wages in TVPS stimulate workers' enthusiasm. Pay for TVP workers is about Y0. 16-YO. 18 per meter of cloth, whereas the figure is only YO.06 per meter for LSE workers. Many LSE workers have to moonlight, which affects their performance on their primary job. INDUSTRIAL COMPOSITION. National statistics show that the profit created by every Y100 of fixed assets in township enterprises is Y6.8 higher than in state enterprises (see table 18-12). In light industry profitability in township enterprises is slightly lower than in state enterprises, but in heavy industry it is much higher. Further studies are needed to analyze this matter in more detail. 412 SONG LINA Table 18-13. Textile Production Costs in LSEs and TVPs in Xiqiao Town, Nanhai (yuan per meter of cloth) Item LSEs 'IVPs Difference Tax 0.64 0.28 0.36 Welfare (including pensions) 0.16 0.00 0.16 Management fee 0.10 0.05 0.05 Interest paid to banks 0.20 0.10 0.10 Materials n. a. n. a. 0.20 Labor efficiency n.a. n.a. 0.12 n.a. Not available. Source: Information from fieldwork. Conclusions What are the prospects for cotnty LSEs and rural township enterprises? Owing to the limitations of the sample surveyed, especially the insufficient materials on LSES, it is difficult to provide convincing statistics, much less a reliable conclusion about future trends. But a hypothesis can be put forward for future investigation: for a long time to come, the growth rate of county LSES will not be very fast. Generally speaking, these firms were founded much earlier than township enterprises ancl have made a considerable contribution to the community and the community government. Precisely for these reasons, their accumulation cannot be very high, and their pace of development suf- fers. At present, and for some time to come, county LSES will probably experi- ence a fairly painful period as they evolve from leading firms in their commu- nities to backward firms with poor economic benefits. Although township enterprises also suffer from various restrictions, they are in a burgeoning and rapidly growing stage. But since it will take time for the entire industrial system to become an organic whole in the vast rural areas of the country, township enterprises may have a long period of heavy responsibilities. Note 1. The statistics concerning TVP workers used here include workers in enterprises under all ownership systems and thus cannot completely represent township enterprises. But in general, workers in enterprises of different ownership do not exhibit obvious differences (see chapter 13). 19 Chinese TVPs in an International Perspective Alan Gelb and Jan Svejnar The rapid growth of China's industrial TVP sector in the 1980s naturally raises questions about its economic behavior and its significance for the Chinese economy. Other chapters in this volume discuss the institutional and behav- ioral features of TVPS. This chapter compares TVPS with similar firms in other countries in the hope of improving our understanding of TVP behavior and our ability to formulate meaningful policy conclusions. The emphasis is on the ownership and management structure of TvPs and on the relationship be- tween TVPs and local government. Principal Features of the Chinese TVP Sector China's TVP sector has evolved within a state-owned economy, with virtually no markets for land, capital, and labor, with heavily regulated product mar- kets, and with a very even distribution of income and wealth, at least locally. Township and village authorities have conditioned the development of the sector to a great extent. Commercial and corporate law remains undeveloped in China. Although most TVPS, especially the larger ones, are owned and controlled by community governments, there are significant variations in ownership. There are three broad categories of TVPS: community enterprises, partner- ships, and individual proprietorships. The latter two categories are referred to as private firms. Many TVPS fall into an intermediate category: they are heavily influenced by community governments but may have substantial managerial autonomy and mixed ownership. Because of low labor mobility and the prevalence of bonus systems, it could also be argued that complex "implicit labor contracts" exist between firms and their workers. The most distinctive feature of TVPS, however, is the strong connection between firms and community governments. Even private firms are subject to considerable government influence, and all TvPs need permission from the community 413 414 ALAN GELB AND JAN SVEJNAR government to undertalce such basic activities as acquiring land and buildings and taking out bank loans. Implications for Comparators Because of the close relationship between TvPs and community govenaments, we have to choose as international comparators locally owned public and pri- vate firms the operations of which are considerably influenced by local au- thorities. This particular fortn of ownership and control is not observed in many countries. Government ownership of industry, when significant, is usu- ally more centralized and corresponds more closely to the Chinese state en- terprise system. The most similar cases are found among labor-managed and private Yugoslav firms and collective farms and producer cooperatives in East- ern Europe. A few other countries, such as Tanzania, have made limited moves toward this pattern. Some similarities can be found in capitalist econ- omies where there are close links between firms and communities; the power- ful role played by Swiss cantons and township authorities in regulating the operation of firms is an example. Prototypical company towns in the United States and Canada have demonstrated considerable overlap between the in- terests of local governments and of enterprises, although in these cases, a large firm "owns" the community rather than the reverse. It is also important tc) stress the differences between TVPs and somLe types of firms that at first sight might seem to resemble them closely. TVPS evolved within a state-controlled economy, not out of a capitalist or colonial back- ground. They have been motivated by economic objectives, and the politi- cal orientation of local authorities has usually determined only the domi- nant structure of property rights and management. TVPS should therefore not be compared with politically motivated experiments with communal ownership-for example, the enterprises run by workers and local conmmuni- ties under Peru's 1984 Law of Social Property Firms. (Experience with this initiative has been disappointing; see Scott 1979.) Other examples are found in Jamaica, which converted substantial areas of foreign-controlled sugar land into producer cooperatives (Richards and Williams 1982); Algeria, which in- itiated a policy of autogestion, or workers' self-management, for farms and rural industries vacated by French proprietors; and Great Britain, where the Labour government initiated several cooperative experiments in 1974. The results of these initiatives were not inspiring (see Mallarde 1975 and Bradley and Gelb 1983b), but the political environment must be borne in mind when considering their perfornnance. The sugar and dairy cooperatives of India are interesting, but as agro- processing cooperatives they are not comparable with manufacturing co- operatives (see Attwood and B3aviskar 1987 for a review). Some Israeli kib- butzim have substantial industry, but their idealistic origins render them a distinctive case. Chinese TVPs in an International Perspective 415 There is some merit in comparing TVPS with collectively owned firms, such as producer cooperatives, that are established and operated for economic rea- sons. In small communities with extensive family ties the distinction between TVP members and the rest of the community may be blurred, and these TVPS may resemble a cooperative of the whole community. This is especially true in south China. It is also worthwhile to note the similarities between TVPS and capitalist firms that follow the "Japanese" system of long-run labor con- tracting and worker-firm identification. Comparisons with Socially (Publicly) Owned Enterprises The main comparators in this category are in Eastern Europe and Yugoslavia, but some analogies are also found in Tanzania's local state enterprises. Eastern Europe The activities of collective farms and producer cooperatives in many Eastern European countries resemble those of Chinese TVPS. Nonstate enterprises op- erate alongside traditional state enterprises. Many and often all of their activ- ities are outside the central plan, their product prices are to a large extent market determined, and they concentrate on goods that are not supplied ef- fectively by state enterprises. They thus select market niches in which de- mand and potential monopoly rents are high, and they often supply high- quality products at unusually high prices. These activities are tolerated and even encouraged by the authorities because the firms or farms reduce social discontent by supplying products that are in demand and would not otherwise be available. Economically, their existence depends in part on the pent-up demand for commodities, which is a systemic feature in Eastern Europe. On the supply side the shift in terms of trade in favor of agriculture has provided many collective farms in countries such as Czechoslovakia with the resources needed for investment and for diversification into nonagricultural activities. In this respect, too, there is a parallel with China's experience. Slusovice, a Czech collective farm, is perhaps an extreme example, but one that clearly illustrates these points. Since 1968 this farm has extended its membership to include technicians, engineers, economists, managers, computer specialists, and marketing professionals. It has radically mecha- nized, automated, and reduced its agricultural activities, which it maintains, even though it does not have a comparative advantage in them, because it has to fulfill planned delivery targets. At the same time, it has focused on specialized industrial production, high technology development and applica- tions, and services. Slusovice's success in the relatively uncompetitive Czech system has been so marked that the government has used it as a showpiece for foreign visitors and dignitaries. The farm has also reorganized its manage- ment structure (it now has a vice president for public relations), and although 416 ALAN GELB AND JAN SVEJNAR in principle it is under the local government's jurisdiction, it is, in effect, under the patronage of the central govemment. Other examples abound. ]For instance, Eastern European nonstate enter- prises, particularly collective farms, are very successful in bidding for govern- ment contracts for the renovation of historical buildings in direct competi- tion with the understaffed and inefficient state enterprises. A significant difference between Chinese TVPS and Eastern European nonstate firms is that ivps comprise a substantially larger share of the total market. Although no reliable statistics exist on the Eastern European firms, there is no doubt that so far they are relatively insignificant operations with high visibility and high social value. A related difference is that Chinese TVPS have begun competing with state enterprises, but the comparable Eastem Eu- ropean firms have done so only rarely and have focused instead on unex- ploited and highly profitable market niches. Chinese TVPs have therefore evolved beyond the stage of their Eastern European counterparts. This is at- tributable both to the rnore active role of Chinese community governments in fostering TVP development and to the greater ease with which these firms, even at small scale, can penetrate China's relatively less-developed industrial market. State enterprises in China may also be less forbidding competitors than those in Eastern Europe. Yugoslavia Yugoslav industrial firms fall into two broad categories: labor-managed firms (LMFS), which are part of the social property sector, and private firms, which operate independently alongside them. These two types of firms strongly re- semble Chinese community enterprises and private enterprises (especially in- dividual firms), respectively. LMFS are in principle controlled and run by workers, but in practice manag- ers and local as well as regional authorities determine their policies to a large extent (see Prasnikar and Svejnar 1987). Yugoslav private firms are more au- tonomous than LMFS, but they too have to comply with the rules and regula- tions set forth by authorities. In LMFS the capital is owned socially (rather than by individual workers or by the labor force of the firm as a whole), and in this respect LMFS resemble community enterprises in China. Capital in Yu- goslav private firms may be owned by individuals, families, or groups of individuals-a situation parallel to that in Chinese private firms. Both LMFs and Chinese cormmunity enterprises are required by the authori- ties to reinvest a sizable part of their income. In both cases authorities stress growth, employment generation (if unemployment or underemployment is a serious problem), and social goals such as equality of worker incomes as im- portant objectives for enterprises. The authorities also ensure that some re- sources are transferred from successful firms to firms in need of funds for devel- opment. In China this is usually done by the authorities directly, through Chinese TVPs in an Intemational Perspective 417 reallocation of community enterprise profits. In Yugoslavia it is accomplished by pooling resources among LMFS. Finally, labor markets and utilization of labor by enterprises show impor- tant similarities. In both countries the appointment and promotion of man- agers is strongly influenced by local authorities and the Communist party. The size of the labor force often exceeds the level of maximum profitability in Yugoslavia and is known to do so in China in some cases. Layoffs are rare, as exogenous shocks are absorbed primarily by changes in income rather than through adjustment of the labor force. The labor pool from which firms draw their workers is local or at most regional, since labor mobility is limited in both systems. There is more worker participation in the management of Yu- goslav LMFS than of Chinese TVPS. In sum, in Yugoslavia as well as in China, local authorities act in many respects as a holding company that influences important decisions of firms and acts as a risk-sharing institution for the regionally immobile residents of the area. But Yugoslav firms and Chinese TvPs also exhibit important differences. First, although local authorities are influential in Yugoslavia, the republican (that is, regional) and federal (central) governments decide some issues that are within the purview of local authorities in China. TVPs are on average much smaller than LMFS, and whereas TVPS were often spearheaded and are still controlled by local authorities, LMFS are usually large enough to be of regional and sometimes nationwide importance. The size of LMFS and their strategic role in the economy make them analogous to Chinese state enter- prises in some ways. The social ownership of LMF capital implies ownership by Yugoslav society as a whole, whereas Chinese community enterprises are owned by the town- ship or village. This crucial distinction means that in Yugoslavia laws and regulations relating to the required reinvestment of firm income and pooling of resources with other firms often come from the republican or federal level. In China the transfer of resources to other TVPS takes place within the com- munity; in Yugoslavia sharing of capital often transcends regional boundaries and is in effect a form of subsidization of less developed by more developed regions. The influence of higher-level authorities over LMFS also explains why these firms are much larger than TvPs and why they face softer budget constraints, tend to be less flexible, and avoid bankruptcy in circumstances in which TVPS might not. LMFs are large because the Yugoslav government wanted to create sizable firms able to compete with large foreign corporations on the world market and because of the desire of LMFS themselves to grow to monopolistic proportions in the presence of the peculiar capital market features discussed below. The size of LMFS and their close relationship with higher-level authori- ties make the latter more willing to forestall bankruptcy by providing subsi- dies. This in turn reduces the incentives for LMFS to improve their efficiency 418 ALAN GELB AND JAN SVEJNAR and respond rapidly to a changing economic environment. The resulting "le- thargic" behavior of many LMFS resembles that of Chinese state enterprises more than that of TVE'S, which cannot rely on the resources of higher-level authorities and are closely monitored by local officials. This difference in observed behavior is partly attributable to the weaker incentives offered to w,orkers in LMFS and the greater constraints on their in- comes. In particular, most workers in Yugoslav firms are paid salaries, and income differentials are limited by government controls, but most TVP work- ers are on piece rates or other, often profit-related, incentive schernes. In the financial sphere, banks in China may be heavily influenced by local governments, but not by TVI'S. In Yugoslavia banks are effectively controlled by their firm-clients. Interest rates are set, for ideological reasons, well below the shadow price of capital and are usually negative in real terms. Since each firm's control over the bank depends on the size of its deposits and loans, the Yugoslav system gives firms an extreme incentive to borrow excessively, grow in a capital-intensive way, and strive to capture as large a share of the capital and product market as possible.' Demand for bank loans by TVPS in China is also voracious, especially in periods of high growth, and local gov- emrnments may influence the behavior of local bank branches, but direct ties between firms and banks are much less pronounced. Finally, although Yugoslav LMFS share power with managers and authorities at various levels, their workers clearly influence enterprise policy more than do their counterparts in Chinese community enterprises. This means that Yu- goslav LMFS probably pursue the welfare of their worker-members (as distinct from that of the local community as a whole) to a greater extent than do most Chinese community enterprises. Tanzania State enterprises dominate Tanzania's industry; private firms and joint ven- tures, sometimes with foreign partners, are next in importance. Beginning in the early 1970s district development councils (DDCS) were established with foreign assistance, with the development of small-scale local industry as one of their goals. The Small Inclustries Development Organization (SIDO) was founded in 1973. The available information indicates that the results of these initiatives were discouraging. Employment in small manufacturing rose by less than 2 percent a year up to 1978 and declined during 1978-82 as the economy entered a severe recession. The use of installed capacity and of com- mon facilities that had been set up to promote industry was low (see Havnevik, Skarstein, and Wangwe 1985). One reason for this poor performance seems to have been the limited link- age between these initiatives and local governments. The latter had no fiscal or other incentives for ensuring that the ventures were profitable. This left the way open for other objectives, notably patronage, which sometirnes re- Chinese TVPs in an Intemational Perspective 419 sulted in gross overmanning of the few potentially viable firms. Tanzania has begun to rethink such policies as the prohibition of private industrial ven- tures in its 8,000 ujamaa villages. Comparators under Private Ownership The main private comparators are the cooperatively owned Mondragon en- terprises in Spain, company towns in the United States, the Swiss cantonal system, and the "Japanese-type" or Z-firm model. The Mondragon Group In 1956 a priest in the Basque village of Mondragon, Spain, inspired a group of unemployed technicians to buy out a bankrupt firm and start an experi- ment in worker ownership. The venture grew rapidly, and other cooperatives followed, initially in Mondragon itself and then in neighboring towns and cities. The cooperatives established a bank (the Caja Laboral Popular-CLP), a social security system, and technical research and training cooperatives. By the 1980s the Mondragon group had almost 20,000 worker-members, 85 affiliated industrial cooperatives, and a number of other cooperatives in agri- culture and services. It has encouraged the formation of numerous coopera- tive schools and housing associations and has become a significant force in regional development. The performance of the Mondragon group over the past three decades is considered to have been very good (see Bradley and Gelb 1983a; Thomas and Logan 1982 provide an extensive description and assessment). Mondragon's firms differ from Chinese community enterprises in that they are autonomous and are owned solely by their employees. It is also misleading to equate the CLP with a holding company or a township industrial corpora- tion. Representatives from the industrial firms sit on the CLP's board; the bank may have an equity stake in new firms and in firms that have run into serious difficulties; and it provides management and promotional services that help firms overcome some of the disadvantages of their relatively small scale.2 It also helps to pool the risks involved in new starts, which are usually funded jointly by the founding individuals, the cLP, and interested cooperatives, usu- ally those of the founding members. But under normal circumstances the CLP has no say in the operation of industrial firms or in their use of surplus. Mondragon's firms are owned partly as indivisible "socialized" capital but mostly in the form of individual accounts of their worker-members. The ac- counts receive allocations of surpluses and losses based on a distribution for- mula that emphasizes the share of the member in the "wage" bill (wages are anticipations out of profits). Departing members must be bought out with the savings of new and remaining workers, so there is a strong incentive to grow and to hire more labor. This incentive system is suitable for the objec- 420 ALAN GELB AND JAN SVEJNAR tive, so important at the start of the Mondragon experiment, of creating em- ployment in the repressed, war-damaged Basque provinces, especially as pay levels are constrained within a differential of about 3 to 1. This range appears comparable with that seen in community enterprises in China, but such so- phisticated ownership systems are absent in the Chinese TVP sector. The strength of the links between Mondragon firms and their local com- munities depends on the location of the firms and, in particular, on whether the community is small and isolated (as many Basque villages are) or is a large city. In Mondragon itself, now a town of 30,000, the cooperatives are extremely important to the community in providing employment, incomes, and local facilities-10 percent of the profits of all cooperatives go to a social fund used for local facilities and services. A small cooperative in a large cen- ter would have far less influence. But the relationship between comrmunities and firms is more or less an arm's length one, in that firms are considered commercial entities operating in a market economy rather than extensions of the community. Because firms are owned by their members, Mondragon's arrangements for control differ substantially from those of most Chinese TVPS. Members elect a board that appoints the management. Workers have an alternative channel of communication, the social council-a form of workers' council--to air grievances. There seems to be little demand for trade union representation in addition to this channel. To sum up, Mondragon suggests a different path toward regional develop- ment from that of the Chinese community enterprise sector. Its organi- zational arrangements rnay not be ideal in all circumstances; in particular, the ownership system is difficult to reconcile with a mobile labor force. Mondragon nevertheless offers rich lessons for the possible development of the TVP sector toward private and more autonomous, yet still group, owner- ship patterns. The Company Town and Cantonal Systems In the United States as well as in other capitalist countries, individual firms (or a small group) may create a community or dominate it economically. In these "company towns" employment, the local tax base, and property values all depend on the activities of the firm, which in effect "owns" the town rather than the reverse. The company may even influence the selection of public officials. These cases are therefore the opposite from what happens in community enterprises. Some communities that have suffered from departures of firms or that rec- ognize their vulnerability to important company decisions have attempted to attract and develop industry on a community basis. The Jamestown Labor- Management Committee in upstate New York, for instance, was formed in 1972 to rejuvenate the town in the face of the exodus of private capital from Chinese TVPs in an International Perspective 421 the area. The role of the committee was to build a partnership of interests among investors, labor organizations, and government organizations to re- verse the process of industrial decline. This effort met with considerable suc- cess. New industries were established or attracted to the area, and a long period of industrial decline was arrested (see Jamestown Labor-Management Committee 1977). Successful worker and community mobilizations of this type yield eco- nomic benefits and give the community greater control over local businesses. In unsuccessful cases the community may actually become worse off. Local savings contributed by workers and residents are often lost in the attempt to save failing or abandoned firms. For example, Rath Meatpacking in Water- loo, Iowa, was unable to compete in the depressed meatpacking industry and declared bankruptcy despite wage cuts, layoffs, and local investments. Local communities or larger units such as states often compete to attract industry with the aim of maintaining or creating jobs and maximizing the tax base. The competitive approach requires open capital and entrepreneurial markets, as well as skill in negotiating with firms. New York State has at- tempted to attract high technology industry into the Ithaca and Rochester areas, with only limited success. In contrast, the town of Auburn in upstate New York has, through its own initiative, attracted Japanese capital. Access to good technical facilities appears to have been important in the location of certain industries-for example, in the Research Triangle area of North Carolina and in New England. It may thus make sense for communities to pool resources with the aim of creating a good environment for "footloose" private firms. The decentralized system in Switzerland provides another example of close interaction between local authorities and business. Cantonal, township, and communal authorities approve and regulate many aspects of enterprise behav- ior. The rules and regulations concerning, for example, zoning and taxes are usually clearly defined, but such issues as the national or local composition of the labor force may be subject to negotiation. Unlike American company towns, Swiss political and social units usually predate the firms, and in that respect firm-community relationships are more akin to those of Chinese TVPS. "Japanese-Type" Fin7s (Z-Firms) Firms operating in the "privileged third" of the Japanese economy create an enriched employment relationship, sometimes referred to as the "paternalism-lifetime commitment" model (see, for example, Dore 1973). Key elements of this model include a long employment horizon, a strict se- niority system, benefits that tie workers to the firm, measures to increase work- ers' involvement, and a substantial bonus element in total pay, so that pay, rather than number of workers, absorbs the impact of changing business con- ditions. Although the cultural heritage of Japan may be a factor in the preva- 422 ALAN GELB AND JAN SVEJNAR lence of this model, Ouchi (1982) has pointed out that some of the most successful U.S. firms (Z-firms) practice similar principles. Z-firms often derive from strong identification between firms and communities; many origi- nated in small isolated towns and were started by owners with strong ethical principles. Z-firms, although perhaps not direct comparators, are of considerable in- terest because they present an alternative to the Taylorist "hire and fire" ar- chetype for private firms. The Z-firm model may be of great relevance for China because of the degree of community identification and the low labor mobility there. A very high proportion of respondents to the Worker Survey Questionnaire expected to stay with their firms for a long time, and the anal- ysis in chapter 13 suggests that Chinese private firms are closer to Z-firms than to Taylorist ones. Changes in the Form of Firm Ownership The structure of ownership and control of firms is not static but responds to economic and political imperatives. For example, the transition fiom fam- ily firms to corporations coincided with a need to exploit economies of scale and at the same time to diversify against risk through the introduction of limited liability. Some large and well-known U.S. firms, notably Amana (air conditioners) and Oneida (silverware), began as firms owned and run by small religious communities and gradually evolved into typical shareholder- owned firms. The recent and current privatization waves in countries as di- verse as Great Britain, France, and Chile have important economic and po- litical dimensions. In this section we consider the economic factors that tend to shift enterprises from worker ownership to private ownership and the reverse. Frorn Worker to Private Ownership The transition from worker to private ownership in capitalist countries can involve successful firms or failing firms. The latter case is analogous to the acquisition of bankrupt private firms by a new owner and is therefore not of special interest here. The transition of successful worker-owned firms, however, has an important systemic underpinning that affects the operation of these firms and our evaluation of them. The factors that determine the likelihood of transition in the successful cases are the nature of w.orker ownership and the time horizon of the current worker-members. When eachi member of the collective has a clearly defined share in ownership that can be sold to current or incoming members at mar- ket value, the tendency toward private, noncollective ownership is not very pronounced. This is the case in Mondragon. When, however, a significant number of members would like to sell their shares but cannot find new or Chinese TVPs in an Intemational Perspective 423 current members who are willing to purchase them at the market price, the incentive to sell to outsiders may become strong. If the market value of the firm is particularly high and worker-owners do not expect to stay with the firm for long, the incentive to sell becomes even more ptonounced. A tendency to sell a worker-owned firm to private outsiders also exists when ownership by worker-members is collective rather than individual. In this case workers are not assigned individual shares in enterprise assets and therefore cannot individually sell their capital holdings, but if the entire firm is sold to an outside bidder, the workers will divide the proceeds. Collective ownership hence provides an important incentive for older workers to sell the firm, and it often pits older cohorts of members against younger ones. The above set of factors, often referred to as the generational or worker horizon problem, has been responsible for the transformation of many suc- cessful worker-owned firms into private ones. Examples in the United States include the Plywood Cooperatives in the Pacific Northwest, the Vermont As- bestos Group, and U.S. News and World Report. If the promotion of worker collectives or communal ownership is considered desirable, these factors are seen as negative. If growth of employment and output are the goals, the tran- sition may have a negative or a positive effect. Finally, in the debate about the relative performance of worker-owned and private capitalist firms, it is important to remember that the measured performance of worker-owned firms is biased because some of the best-performing ones are no longer worker owned. From Capitalist to Worker Ownership The transition from capitalist to worker ownership usually follows one of four patterns. 1. A benevolent or enlightened capitalist-owner decides to pass his firm on to the labor force. This is a rather special case, but some well-known worker-owned firms originated this way-for example, the John Lewis Part- nership (which has 30,000 members), Scott Bader, and Baxi Heating, all in Great Britain. 2. Capitalist firms give or sell workers shares in the firm to generate capital, to attract workers with special skills, or to elicit greater identification with the firm and hence better performance. The firm remains traditional in struc- ture, and the share of total capital owned by workers is typically rather small. Widespread employee shareholding of this type is a new development, al- though stock option plans for management have long been commonplace. 3. A capitalist firm closes down and becomes bankrupt, and the commu- nity or the workers take over to save the local economy. Private capital is usually unwilling to step in, and this reluctance may signal serious problems that cannot be overcome even under worker ownership. (An example is the case of Rath Meatpacking, noted above.). But-as in the case of the Library 424 ALAN GELB AND JIAN SVEJNAR Bureau, a subsidiary of Sperry Rand-the decision of the parent company to close the firm may simply be motivated by low profitability, and worker- owners can still eam reasonable incomes. 4. As a result of recent tax legislation, capitalist firms in the United States have found it advantageous to introduce employee stock ownership plans (ESOPS) that enable workers to acquire company shares either directly or indi- rectly, through their pension plans. The General Accounting Office esti- mates that in 1986 there were approximately 5,000 U.S. corporations with active ESOPS. Great Britain too has introduced incentives to encourage trans- fer of shares to employees. As in case 2, in most instances ESOPS have not shifted the structure of power significantly. Often workers have not obtained any voting rights, and sometimes ESOP incentives have been used to favor the interests of owners at the expense of workers. The Debate on Efficiency Although many observers believe that private ownership has an edge over public ownership in inclustry, the relative merits of conventional capitalist ownership and worker ownership are still hotly debated. Recent econometric studies indicate that the effect of individual worker ownership on productive efficiency (total factor productivity) is positive or insignificant and that of collective ownership is zero or negative (see, for instance, Jones and Backus 1977, Cable and Fitzroy 1980, Jones and Svejnar 1985, and Defourney, Es- trin, and Jones 1985). Case studies suggest that the particular form of owner- ship, the way in which it was introduced, and the accompanying decision- making rights may influence the effect of ownership on productivity. Systemic Flexibility and Adjustment The ability of TVPS to adjust to changing conditions and their flexibility in relation to state enterprises clepend in large part on institutional arrange- ments that constrain or stimulate their performance. As recent Chinese and, to a lesser extent, Eastern European experience indicates, nonstate firms in a socialist planned economy tend to expand rapidly on the upswing of the economy. This is primarily attributable to their ability to fill niches left by state enterprises and to expand investment rapidly. Their behavior in the more mature phase (when niches have been fully exploited) is harder to predict. Comparative intemational experience provides some support for the prop- osition that communal firms, as well as those organized on the Japanese model, tend to respond I:o economic downtums by reducing incomes rather than by shedding labor, in contrast to "conventional" firms in a market econ- omy. Japan's success in avoiding unemployment is notable (although it may be partly attributable to the existence of a large "unorganized" sector of firms Chinese TVPs in an International Perspective 425 that subcontract work from large firms and so act as a cushion). Among col- lectives the clearest example is perhaps the reaction of the Mondragon group to the severe downtum in the Spanish economy after 1979 (see Bradley and Gelb 1985). Layoffs were minimized, and incomes were reduced to the extent needed to preserve capital. At the same time, the search for new products was intensified, and resources were pooled to permit promising new starts. The performance of the Mondragon cooperatives was considerably better than that of the rest of the Spanish economy during the recession. In the economy at large, hiring ceased and labor was laid off despite regulations that limited dismissals. The result was a sharp segmentation of the labor force ac- cording to age as unemployment rose for all age groups, but especially for younger workers. Wage increases began to moderate only when unemploy- ment became severe. The behavior of Chinese TvPs during a downswing has not yet been fully observed. Given their flexibility during other parts of the cycle, however, they can be expected to display considerable readiness to adjust. The degree and form of adjustment will depend on the severity of the downturn, the availability of altemative products or markets, and the role of local authori- ties who act as both risk sharers and enforcers of social values. One would expect adjustment to take place primarily through reductions in incomes rather than through layoffs. In this sense TVPS resemble Yugoslav LMFS, Japa- nese and Z-firms, and the Mondragon cooperatives. They differ from Swiss firms, many of which lay off foreign workers during recessions, as well as from firms in capitalist economies that rely on temporary layoffs rather than wage fluctuations to absorb the impact of shifts in demand. Like the Japanese, Yu- goslav, and Mondragon firms, community enterprises would probably empha- size flexible labor use within the firm and an active search for new markets. It is more difficult to assess how Chinese private firms that hire wage labor from outside the community will behave during a downtum. Many may rely on labor adjustment, like their counterparts in other countries. Others may adjust like Z-firms, cutting pay and bonuses and trying to maintain employ- ment. The outcome would depend on the relative power of entrepreneurs and community authorities and on the way in which private property rights, which are still rather fragile in China, are interpreted and exercised. Notes 1. Greater capitalization leads to greater control over the bank, whereas a greater share of the product market guarantees greater control over product prices. Both allow the LMF to raise worker incomes, collective consumption, or both more than it otherwise could. 2. Mondragon decided, after a strike at the largest cooperative in 1974, to try to limit maxi- mum firm size to about 500 members to preserve the cooperative spirit; the largest firm now has about 2,500 members. References Anhui Jingji Nianjian 1985 [Anhui Economic Yearbook 1985]. Hefei: Anhui Jingji Nianjian Bianyuan Weiyuanhui. Attwood, D. M., and B. S. Baviskar. 1987. "Why Do Some Cooperatives Work But Not Others? A Comparative Analysis of Sugar Cooperatives in India." Economic and Political Weekly 22, no. 26:38-56. Bradley, Keith, and Alan Gelb. 1983a. Cooperation at Work: The Mondragon Experi- ence. London: Heinemann. . 1983b. Worker Capitalism: The New Industrial Relations. London: Heinemann. . 1985. "Mixed Economy vs. 'Cooperative' Adjustment: Mondragon's Experi- ence through Spain's Recession." World Bank, Development Research Department Discussion Paper 122, Washington, D.C. Brown, James N., and Orley C. Ashenfelter. 1986. "Testing the Efficiency of Employ- ment Contracts." Journal of Political Economy 94(3), pt. 2 (June):S40-S87. Buchanan, James M. 1975. "Public Finance and Public Choice." National Tax Journal 28, no. 4 (December):383-94. Byrd, William A., and Gene Tidrick. 1987. "Factor Allocation and Enterprise Incen- tives." In Gene Tidrick and Chen Jiyuan, eds., China's Industrial Reform. New York: Oxford University Press. Cable, 3., and R. Fitzroy. 1980. "Cooperation and Productivity." Economic Analysis and Workers' Management 14:163-80. China Economic Yearbook Editorial Board. 1981. Zhongguo Jingji Nianjian 1981 [China Economic Yearbook 19811. Beijing: Zhongguo Jingji Guanli Zazhishe. . 1985. Zhongguo Jingji Nianjian 1985 [China Economic Yearbook 1985]. Bei- jing: Jingji Guanli Chubanshe. China Rural Development Research Group (ed.). 1985, 1986. Nongcun, Jingji, Shehui [The Rural Areas, Their Economy, Their Society]. Irregular serial. Nos. 1-3 (1985) and 4 (1986). Beijing: Zhishi Chubanshe. Defoumey, j., S. Estrin, and D. Jones. 1985. "Effects of Worker Participation on En- terprise Performance." International Journal of Industrial Organization 3:197-217. Development Research Institute. 1986. "Lun Nongcun Feinong Chanye de Fazhan" 427 428 REFERENCES [On the Development of Rural Nonagricultural Activities]. Jingji Yanjiu [Economic Research] 8 (August):9-24. Domar, Evsey. 1966. "The Soviet Collective Farm as a Producer Cooperative." Amer- ican Economic Review 56, no. 4 (September):734-57. Dore, Ronald. 1973. British Factory, Japanese Factory. London: Allen and Unwin. Enos, J. L. 1984. "Commune- and Brigade-Run Industries in Rural China: Some Re- cent Observations." In Keith Griffin, ed., Institutional Reform and Economic Devel- opment in the Chinese Countryside. London: Macmillan. Fuss, Melvyn A., and Vinod K. Gupta. 1981. "A Cost Function Approach to the Estimation of Minimurrm Efficient Scale, Returns to Scale, and Suboptimal Capac- ity." European Economic Review 15:123-35. Griffin, Keith, and Kimberley Griffin. 1984. "Commune- and Brigade-Run Enter- prises in Rural China: An Overview." In Keith Griffin, ed., Institutional Reforrn and Economic Development in the Chinese Countryside. London: Macmillar. Halvorsen, R., and R. Palmquist. 1980. "The Interpretation of Dummy Variables in Semilogarithmic Equations." American Economic Review 70, no. 3 (June):474-75. Havnevik, K., R. Skarstein, and S. Wangwe. 1985. "Small-Scale Industrial Sector Study, Tanzania." Report to the Minister of Finance, October 1985. Dares Salaam. Hayek, Friedrich A. 1948. Individualism and Economic Order. Chicago: University of Chicago Press. Hebert, Robert F., and Albert N. Link. 1982. The Entrepreneur: Mainstream Views and Radical Critiques. New Yolrk: Praeger. He Jiacheng. 1987. "Zhuazhu Tuopin Zhifu de Sange Guanjian Huanjie" [On Three Key Links in Letting Poor Areas Prosper]. Jingjixue Dongtai [Trends in Economics] no. 5 (May). Ho, Samuel P. S. 1986. The Asian Experience in Rural Nonagricultural Development and Its Relevance for Clhina. NVorld Bank Staff Working Paper 757. Washington, D.C. Institute of Economics, cAss. 1987. Zhongguo Xiangzhen Qiye de Jingji Fazhan yu Jingji Tizhi [The Economic Development and Economic System of China's TVPs]. Bei- jing: Zhongguo Jingji Chubanshe. Jamestown Labor-Managernent Committee. 1977. "Commitment at Work: The Five- Year Report of the Jamestown Labor-Management Committee." Jamestown, N.Y. Jiangsu Academy of Social Sciences. 1981. Collected Papers on the Experiences of Jiangsu Commune and Brigade Enterprises (in Chinese). Nanjing: Jiangsu Renmin Chubanshe. Jiangsu Jingji Nianjian 1986 Uiangsu Economic Yearbook 1986]. Nanjing: Jiangsu Renmin Chubanshe. Jones, D. C., and D. Back:us. 1977. "British Producer Cooperatives in the Fbotwear Industry." Economic Journal (September):488-5 10. Jones, D. C., and J. Svejnar. 1982. Participatory Firms: Evaluating Economic Perfor- mance. Lexington, Mass.: Lexington Books, D. C. Heath. . 1985. "Participation, Profit Sharing, Worker Ownership and Efficiency in Italian Producer Cooperatives." Economica 52 (November):449-65. Kalai, Ehud. 1977. "Nonsymmetric Nash Solutions and Replications of Two-Person Bargaining." International Journal of Game Theory 6(3):129-33. References 429 Kirzner, Israel M. 1973. Competition and Entrepreneurship. Chicago: University of Chi- cago Press. Knight, John B., and Richard Sabot. Forthcoming. Education, Productivity, and Ine- quality: The East African Natural Experiment. New York: Oxford University Press. Krueger, Anne 0. 1968. "Factor Endowments and Per Capita Income Differences among Countries." Economic Journal 78 (September):641-51. Lewis, W. Arthur. 1954. "Economic Development with Unlimited Supplies of Labor." Manchester School of Economic and Social Studies 22, no. 2 (May): 139-91. Lin Qingsong and others. 1987. "Gongren Gufen Suoyouzhi de Qiye Zhidu: Guowai Feiquxing Qiye Zhidu de Kaocha he Sikao" [Observations and Reflections on Atyp- ical Enterprise Systems in Other Countries]. Jingjixue Dongtai [Trends in Economics] 6 (June):31-37. Little, Ian M. D., Dipak Mazumdar, and John M. Page, Jr. 1987. Small Manufacturing Enterprises: A Comparative Analysis of India and Other Economies. New York: Oxford University Press. Lu Mai and Dai Xiaojing. 1987. "Xian Jieduan Nonghu Jingji Xingwei Qianxi" [An Elementary Discussion of the Economic Behavior of Rural Households at Present]. Jingji Yanjiu [Economic Research] 7 (July):68-74. Luo Hanxian. 1987. Zhongguo Jingli Tizhi Gaige: Nongcun-Zhongguo Nongcun Fazhan Shehuixue [Economic System Reform in China: Rural Villages- Development So- ciology in China's Rural Villages]. Hong Kong: Guangjiaojing Chubanshe. Mallarde, E. 1975. L'Algerie Depuis. Paris: La Table Ronde. Ministry of Agriculture. Various years. Zhongguo Nongye Jingji Ziliao [Statistical Mate- rials on China's Agricultural Economy]. Beijing: Nongye Chubanshe. . 1986a. Zhongguo Nongxuyuye Tongji Ziliao 1984 [China Agriculture, Animal Husbandry, and Fishery Statistical Materials 1984]. Beijing: Nongye Chubanshe. 1986b. Zhongguo Nongxuyuye Tongji Ziliao 1985 [China Agriculture, Animal Husbandry, and Fishery Statistical Materials 1985]. Beijing: Nongye Chubanshe. Mueller, Dennis C. 1976. "Public Choice: A Survey." Journal of Economic Literature 14, no. 2 (June):395-433. Nanhai Nongcun [Rural Nanhail. Periodical. Niskanen, William A. 1968. "Nonmarket Decision Making: The Peculiar Economics of Bureaucracy." American Economic Review 68, no. 2 (May):293-305. Ouchi, W. 1982. Theory Z: How American Business Can Meet the Japanese Challenge. New York: Avon. Perkins, Dwight H., and others. 1977. Rural Small-Scale Industry in the People's Repub- lic of China. Berkeley: University of Califomia Press. Prasnikar, J., and J. Svejnar. 1987. "Enterprise Behavior in Yugoslavia." Advances in the Economic Analysis of Participatory and Labor-Managed Firms 3. Reynolds, Lloyd G. 1982. Labor Economics and Labor Relations. 8th ed. Englewood Cliffs, N.J.: Prentice-Hall. Richards, V., and A. Williams. 1982. "Institutional and Economic Aspects of the Jamaican Sugar Cooperatives." In D. C. Jones and J. Svejnar, eds., Participatory Firms: Evaluating Economic Performance. Lexington, Mass.: Lexington Books, D. C. Heath. 430 REFERENCES Riskin, Carl. 1978a. "China's Rural Industries: Self-Reliant Systems or Independent Kingdoms?" China Quarterly 73 (March):77-98. . 1978b. "Political Conflict and Rural Industrialization in China." W/orld De- velopment 6, no. 5 (May):681-92. Roth, Alvin E. 1979. Axiomatic Models of Bargaining. Lecture Notes in Economics and Mathematical Systems 170. New York: Springer-Verlag. Schumpeter, Joseph A. :1934. The Theory of Economic Development. Cambridge, Mass.: Harvard University Press. Scott, C. D. 1979. "Agrarian Reform and Agricultural Labor Markets." University of East Anglia Occasional Paper 2. Norwich, U.K. Sigurdson, Jon. 1977. Rural Industrialization in China. Cambridge, Mass.: Harvard University Press. State Council Economic, Technological, and Social Development Research Center. 1986. Zhongguo Jingji Nianjian 1986 [China Economic Yearbook 1986]. Beijing: Jingji Guanli Chubanshe. State Statistical Bureau 1982. Statistical Yearbook of China 1981. Hong Kong: Eco- nomic Information and Agency. 1983. Statistical Yearbookc of China 1983. Hong Kong: Economic Information and Agency. . 1984. Statistical Yearbook of China 1984. Hong Kong: Economic Information and Agency. 1985a. Statistical ',earbook of China 1985. Hong Kong: Economic Information and Agency. 1985b. Zhongguo Gongye Jingji Tongi Ziliao 1949-84 [Statistical Materials on China's Industrial Economy 15149-84]. Beijing: Zhongguo Tongji Chubanshe. 1985c. Zhongguo Shehui Tongji Ziliao [China Social Statistical Materials]. Bei- jing: Zhongguo Tongji Chubanshe. 1986a. Statistical Yearbook of China 1986. Hong Kong: Economic Information and Agency. . 1986b. Zhongguo Nongcun Tongji Nianjian 1985 [China Rural Statistical Yearbook 1985]. Beijing: Zhongguo Tongji Chubanshe. . 1986c. Zhongguo Tongi Nianjian 1986 [China Statistical Yearbook 19861. Beijing: Zhongguo Tongji Chulbanshe. 1987a. Statistical Yearbook of China 1987. Hong Kong: Economic Information and Agency. . 1987b. Zhongguo Gongye Jingji Tongji Ziliao 1986 [Statistical Materials on China's Industrial Economy 19861. Beijing: Zhongguo Tongji Chubanshe. 1987c. Zhongguo Nongcun Tongji Nianjian 1986 [China Rural Statistical Year- book 19861. Beijing: Zhongguo Tongji Chubanshe. 1987d. Zhongguo Nongcun Tongji Nianjian 1987 [China Rural Statistical Year- book 1987]. Beijing: Zhongguo Tongji Chubanshe. Svejnar, Jan. 1986. "Bargaining Power, Fear of Disagreement and Wage Settlements: Theory and Evidence from U.S. Industry." Econometrica 54, no. 5 (Septem- ber): 1055-78. Thomas, Henk, and Chris Logan. 1982. Mondragon: An Economic Analysis. Winches- ter, Mass.: Allen and Unwin. References 431 Tiebout, Charles M. 1956. "A Pure Theory of Local Expenditures." Journal of Political Economy 64, no. 5 (October):416-24. - 1961. "AnEconomicTheory of Fiscal Decentralization." InNational Bureau of Economic Research, Public Finances: Needs, Sources, and Utilization. Princeton, N.J.: Princeton University Press. Tidrick, Gene, and Chen Jiyuan. 1987. China's Industrial Reform. New York: Oxford University Press. Todaro, Michael P. 1969. "A Model of Labor Migration and Urban Unemployment in Less Developed Countries." American Economic Review 59, no. 1:138-48. U.S. Government. U. S. County and City Data Book, 1983. Washington, D.C.: Gov- emment Printing Office. Vanek, Jaroslav. 1970. The General Theory of Labor-Managed Market Economies. Ith- aca, N.Y.: Cornell University Press. Wang Xiaoqiang, and others. 1983. "Nongcun Shangpin Shengchan Fazhan de Xin Dongxiang-Wenzhou Nongcun Jige Zhuanye Shangpin Chanxiao Jidi de Kaocha Baogao" [New Trends in the Development of Rural Commodity Production- Report on an Investigation of Several Rural Production and Marketing Bases in Wenzhou]. Reprinted in China Rural Development Research Group no.. 3:69-93. Ward, Benjamin. 1958. "The Firm in Illyria: Market Syndicalism." American Eco- nomic Review 48, no. 4 (September):566-89. Williamson, Oliver E. 1985. The Economic Institutions of Capitalism. Cambridge, Mass.: Harvard University Press. Wong, Christine. 1979. "Rural Industrialization in China: Development of the 'Five Small Industries."' Ph.D. dissertation. University of California, Berkeley. World Bank. 1984. World Development Report 1984. New York: Oxford University Press. . 1988. China: Growth and Development in Gansu Province. A World Bank Country Study. Washington, D.C. Wu Xiang. 1986. Zhongguo Nongcun Gaige [Rural Reforms in China]. Beijing: Guangming Ribao Chubanshe. Xinhua Yuebao [New China Monthly Digest]. Beijing. Zhou Qiren and Huang Zhuangjun. 1987. "Zhongguo Xiangzhen Gongye Qiye de Zichan Xingcheng, Yingyun Tezheng Jiqi Hongguan Xiaoying-Dui 10 Sheng Daxing Xiangzhen Gongye Qiye Chouyang Diaochade Fenxi" [Asset Formation, Operational Features, and Macroeconomic Impact of China's Township and Village Industrial Enterprises-An Analysis of a Sample Survey of Large Township and Vil- lage Industrial Enterprises in Ten Provinces]. Zhongguo Shehui Kexue [Social Sci- ences in China] no. 6 (November):41-66. Zhu Tonghua. 1985. Xiangzhen Gongye yu Xiaochengzhen [Township and Village En- terprises and Small Towns]. Beijing: Zhongguo Zhanwang Chubanshe. Index ABC. See Agricultural Bank of China hai, 160; private enterprise and, Accounting systems, 409-10 57, 58, 80, 177, 184; Shuguang Adjustment to adversity, 98-102; Foodstuffs Company and, 147; TVPS asset specificity and, 108-09; neo- and, 78, 223-26, 269, 359, 365, classical, 92-94 368, 377, 382; in Wuxi, 158, 159. Agricultural Bank of China (ABC): en- See also Credit trepreneurship and, 162, 200, 201, Bankruptcy, 404, 417, 423 202, 204; in Jieshou, 225; in Wuxi, Beijing University, 20 78. See also Bank loans Bonuses, 5, 55, 332, 333, 336, 380, Agriculture: collective farms and, 385, 413 134-35, 414, 415, 416; income Brazil, 120 from, 51, 53, 114; in Jieshou, Brigade industries (renamed village 138-39; labor surplus and, 302; enterprises): history of, 7, 79, 119, output of, 48, 49, 53, 113-14, 115; 135, 137; in Nanhai, 138, 141, prices and, 114, 116, 117; produc- 143; in Shangrao, 144; township tivity of, 50; PRS and, 114, 115; re- govemments and, 345, 346-47; cent changes in, 6-7; rural industri- TVP development and, 9-11, 16, alization and, 113-17, 222-23; 50, 222-23; in Wuxi, 139, 140 township governments and, 343, Brown, James N., 246 347, 348, 352, 356; TVP develop- ment and, 48-55, 59; use of indus- Canada, 414 trial profits to subsidize, 56, 80, Capital: access to, 78, 79; economies 114, 158; in Wuxi, 158 of scale and, 105; entrepreneurship Algeria, 414 and, 189-90, 196, 199-203, 205, Ashenfelter, Orley C., 246 208; private enterprise and, 173; profits and productivity of, 235; Bank deposits. See Savings rural nonagricultural development Bank loans, 56, 87, 115; community and, 267-70; TVCE fund accumula- banks and, 214-215; distribution of tion and, 267-68; Tvp development profits and, 232; Enterprise Quanti- and, 51-52, 61, 232-42 tative Questionnaire and, 224; en- Cement industry, 165 trepreneurship and, 200, 201, 202, Central Committee of the Chinese 203, 204, 205, 207, 209, 216; in Communist Party, 10-11, 47, 56, Jieshou, 146, 162, 163; market in- 113, 120 teractions and, 87, 94, 95; in Nan- Chile, 422 433 434 INDEX Chinese Academy of Social Sciences Contract systems, 126, 330-31 (CASS), 19, 20 Costs: of private enterprise, 181; of Coal, 54, 102, 179 TVPS and state enterprises, 410-11 Commune industries (renamed town- Counties: construction programs in, ship enterprises): history of, 6, 7, 121, 122; geography and history of 119, 134-35, 137; in Nanhai, 141, sample, 74-76; TVPs and, 63. See 143; in Shangrao, 144; township also Jieshou County; Nanhai govemments and, 345, 346-47; County; Shangrao County; Wuxi TVP development and, 9-11, 16, 50, County 222-23; in Wuxi, 7, 140-41, 143 County enterprises. See Local state Community govemments. See Town- enterprises ship govemments; Villages Credit: capital productivity arid, 232; Company towns (U.S.), 420-21 entrepreneurship and, 200, 202, Comparative analysis: accounting sys- 214; market interactions and, tems and, 409-10; cost structure 91-92, 100; nonagricultural devel- and, 410-11; of Eastern Eurolpean opment and, 269; policy of tight, firms and TVPS, 415-16; economic 73; private enterprise and, 177, sector and, 393-94; factory direc- 184; trade, 206; TVP development tors and, 401-04; kinship ties and, and, 58. See also Bank loans 395-96; labor and, 394-95; manag- Cultural Revolution, 79 ers and, 397, 400-04; ownership Czechoslovakia, 415 and, 396-400, 422-25; planning systems and, 398-400; of private Data for study: counties and town- firms and TVPS, 419-22; rural indus- ships covered and, 25-27, 34-35; trialization model and, 392; rural- firm characteristics and, 35-37; in- urban similarities and, 393-96; of terview data and, 31-32; question- Tanzanian firms and TVI'S, 414, naires and, 27-31, 32-33; reliabil- 418-19; worker characteristics and, ity and gathering of, 22; validity of, 404-08; of Yugoslav firrms and TVPS, 37-38; workers and directors and, 37 414, 416-18 Decisionmaking, 125, 398-400 Competition: capital formation and, Depreciation finds, 125, 155 239; community governments and, Directors: bonuses and, 55; cornpara- 5; imitative, 92-94, 108; innova- tive analysis and, 401-04; data set tive, 95-96; investment., 94-95, and, 37; decisionmaking powers of, 108; for labor, 56, 324, 325, 335; 125, 351; education of, 401; elec- market, 62; nonprice, 55; price, 96; tion of, 380. See also Enterprise Di- private enterprise and, 176; soft rector Questionnaire drink industry and, 90; state enter- prises and TVP, 96-97; strategies Eastem European state enterprises, for, 97-98; TVP sector and, 415-16 90, 92-98, 107-08; urban industry Economic development: private en- and intemational, 115; in Wuxi, 158 terprise and, 174-75; strategy for, Construction firms, 58, 66; regional 112. See also Regional imbalances development and, 261 in economic development Construction materials industry, 67, Economic System Reform Research 102 Institute, 20 Construction of county seats, 121, 122 Economies of scale: efficiency and, Consumption, 59-60 105; industrial structure and, Index 435 104-05, 109-11; TVCEs and, and, 190; financial risk-bearing and, 239-42 190-91, 203-06; importance for Education, 95, 118, 368, 386; of di- TVP sector of, 189; motivation and, rectors, 401; wage equations and, 197-98; ownership and, 190-99, 292, 297; of workers, 405, 406 202, 207, 209-12; private enter- Efficiency: analysis of productive, 243, prise and, 195, 196, 197, 202, 205, 245, 247-50, 254; economies of 208-09, 213, 216; supply and im- scale and, 105; LSES and TVPs and, mobility of, 198; townships and, 408-11; ownership and productive, 382; trends analysis and, 213-14; 247, 248, 251, 253, 424; private TVP case studies and, 191-95; TVP enterprise and, 181-83; township ownership structure and, 195-97, government and, 350 209-12; unsuccessful TVCES and, Electricity, 54, 87, 102; shortage of, 207; Wuxi model of, 206-07 241 Equipment: imported technical, 109; Employee stock ownership, 424 oil production, 305; TVP devel- Employment: growth of, 69; guaran- opment and, 53-54 teed, 285; incentives and, 375-76, Expenditures, township, 342, 343, 377; industrial (in Shangrao), 154, 344, 364-70, 370-71, 377, 383, 155; in Jieshou, 145, 241; job cre- 386. See also Extrabudgetary reve- ation and, 117; labor utilization nues and expenditures and, 243, 244-45, 245-47, Exports, 71-72 250-53, 254; in Nanhai, 241; of Extrabudgetary revenues and expendi- outside labor, 161; private enter- tures, 397; remittances to commu- prise and, 183, 184; productive effi- nity govemment and, 15; township ciency and, 243, 245, 247-50, 251, governments and, 342-43, 347; 253, 254; production teams and, TVP institutional framework and, 135; rural-urban and urban-rural 4-5. See also Expenditures; Revenues flows and, 270-71, 272-73; town- ship governments and, 124, 324, Factors of production: economies of 352, 372; TvcEs and, 4, 157-59; in scale and, 105; entrepreneurship TVP sector, 11-13, 47, 61, 62, 77, and, 190; investment and, 319; 121, 123, 125-26; urban, 163; township governments and immo- wage determinants and, 325-27; in bility of, 359-60, 386; urban-rural Wuxi, 157-159 differences in, 118 Enos, J. L., 19 Farmers: incomes of, 60, 177, 178, Enterprise Director Questionnaire, 327-29; nonagricultural activities 30-31, 33, 241-42, 355 of, 47, 172-73, 176; property rights Enterprise Quantitative Question- and, 166; subsidies and, 328 naire, 30, 32, 33, 353; bank loans Financial adjustment, 99-100, 122 and, 224; dispersion of TVps and, Financial structure: community gov- 241-42 emments and, 5; market interac- Enterprise Survey Questionnaire, tions and, 89; support for develop- 29-30, 33, 303, 304, 310, 393 ment and, 57; township Entrepreneurship: backward areas and, governments and, 342-50 215-216; capital allocation and Firm size, 88; industrial structure and, TVPs and, 189-90, 196, 199-203; 103-04, 109; TVCEs and, 239 community banks and, 214-215; Firm status. See Status hierarchy of defined, 191; factors of production industrial firms 436 INDEX Fiscal predation, 377-80., 383, 384, ity and, 271, 364; land contracts 385-86 and, 328-29; ownership and, 289; Food processing, 67 pension system and, 328; PRS and, Ford Foundation, 21 4; rural areas and, 362-64; town- Foreign investment, 72-73, 78 ship govemments and, 324, 360-61; Foreign trade, 71-72 TVPs and, 329, 331 France, 414, 422 Income redistribution, 329, 334, 361 Incomes: agricultural, 51, 53, 114; Govemment expenditures and reve- balancing (among TVPS), 329; nues. See Expenditures; Revenues; change from collective to Taxes household-based, 6; composition of Government policy: private enterprise monthly, 332-34; differences in and, 80, 184, 187; toward rural in- (within firms), 330-32; farmer, 60, dustry, 56, 113, 119-20; TVP devel- 177, 178; gap in urban-rural, 115, opment and, 47-48, 60, 79-81, 84 117, 271; growth of, 90, 91; from Govemment regulation, 184, 185, 187 land, 52; links of community pay Grain: output of, 113-14, 271; pro- with 372-75; of managers, 332-33; duction in Nanhai, 141; production of marketing personnel, 55; owner- in Wuxi, 328; production teams ship analysis and, 138, 141, 143, and rations of, 135, 143; shortage 144, 149, 150, 151, 152, 158, of, 134 159-60; private enterprise and, Grants by local govemment, 87, 89 181; rural communities and, 4, 59, Great Britain, 414, 422, 423 149, 151, 264-65; township enter- Great Leap Forward, 9, 20, 50, 79, prises and, 124; TvcFs and, 69, 71; 134, 135, 139 worker, 61-62, 286-87 Griffin, Keith, 19 India, 414 Griffin, Kimberley, 19 Industrial and Commercial Bark, 78. Gross value of industrial output See also Bank loans (Gvio): growth of, 11; regional var- Industrial development, 172; agricul- iations in, 3, 366; state sector de- ture and, 48-55, 113-17, 222-23; cline in, 209; TVPS and, 65, 66, 85, capital and, 51-52, 61, 78-79; 107. See also Output conditions for, 137-39; equipment and materials and, 53-54; financial Handicrafts, 67, 135, 147-48, 153, operations and, 344-46; funcd accu- 156, 173, 195 mulation and, 267-68; future direc- Hayek, Friedrich A., 190 tions in, 60-62; govemment policy Ho, Samuel P. S., 19 changes and, 47-48, 60, 79-81, Hong Kong, 25, 73, 78, 107, 1411 84; land and, 52-53; marketing Human resources: 76-77, 118 and, 54-55, 62; model for rural, Hydropower stations, 78, 103; tech- 392; 1984 to present, 55-60; nology and, 164-65 1978-83, 48-55; obstacles to, 84; private enterprise and rural, 1.75; Imports, 109, 141 property relationships and township Incentives. See Motivation govemments and, 346-48; regional Income distribution, 289; analysis of, differences and, 8, 108, 265-73; 323, 324; collective contract system rural labor market and, 299; struc- and, 331; community opportunity tural reform and, 123-27; study im- structure and, 325-26; labor mobil- plications and, 83-84; township Index 437 govemments and, 352, 353-57; 326, 334, 335, 336, 337, 352; en- TVP, 82-83; urban-rural economic terprise income differences in, coordination and, 117-22; 330, 331; equipment and, 54; "Wenzhou" model for, 34-35, 123, extrabudgetary funds in, 397; farm 176, 177, 181 machinery plant in, 93; geography Industrial structure: community, and history of, 75-76; government 105-06; economies of scale and, policies and, 8, 79, 80, 84, 162-64; 104-05, 109-11; firm size and, incomes in, 144, 148, 299, 332, 103-04, 109; status hierarchy and, 333, 334; industrial output in, 47, 137-39; subsectoral composition of, 145, 174, 347; investment and, 102-03; TVP development and, 155; labor costs in, 319; labor mo- 65-66, 67; in Wuxi, 137. See also bility and, 303, 304, 305, 306, Status hierarchy of industrial firms 307, 317; land market in, 34, 53, Inputs for TVPS, 85-86 146, 162; management in, 234, Interest, 158, 202, 204, 214 353, 354, 355; marketing and, 54; International Conference on China's materials supply and, 398; owner- Township and Village Industries, ship in, 8, 36, 69, 89, 146-48; 21 ownership forms of TVPS in, Investment: community govemment 192-93, 195, 196-97, 199, 204, and, 5-6; competition for, 94-95, 205, 208, 209, 210, 212, 213, 215; 108; factors of production and, 319; private enterprise in, 57, 162-64, failures in 168-69; foreign, 72-73, 181-82, 382; private firms under 78; in hydropower, 164; interre- collective auspices in, 145-48; pro- gional, 268-69; private enterprise ductive efficiency and, 243, 245, and, 174, 177; in Shangrao, 155; 249, 252, 253, 254; profits and, township, 349, 351; Tvp develop- 145, 164, 229, 230, 232; promo- ment and, 51-52, 61; workers' atti- tion incentives in, 375; residence tude toward, 286; in Wuxi, 158, system in, 57; resource endowment 159. See ako Reinvestment of, 352; Tvcas and, 144-45, 146, Israel, 414 156, 356; TVP income composition in, 329; TVP model and, 283; TVPS Jamaica, 414 in, 55, 63, 65, 66, 67, 69, 70, 71, Japan, 76, 115; flexibility of firms in, 72, 123, 146-48, 162-64, 173; 424; Z-firms in, 421-22 wage and income balancing in, Jiangsu Province, 56, 58, 74, 103, 327; wages in, 144-45, 154, 308, 104, 119, 121, 126. See also Wuxi 309, 310, 311, 314, 373, 375; work& County ers in, 33, 37, 53, 77, 146, 162 Jieshou County: agriculture in, 49, Jilin Province, 93 59, 138-39, 222, 343, 347, 356; Job assignments, 303 backward townships in, 39; bank Joint public-private firms, 159-61 loans and, 146, 162, 163; capital and, 51-52, 78, 236, 237; competi- Kinship ties, 395-96 tive market and, 91, 107; credit Kirzner, Israel M., 190 and, 223, 224, 225; data and field- work and, 21, 26, 31; development Labor: comparative analysis and, in, 35, 137, 138-39, 356, 371; 394-95; competition for, 56, 324, economies of scale and, 241; em- 325, 335; economies of scale and, ployment in, 145, 162, 285, 325, 105; in Nanhai and Jieshou, 161, 438 INDEX Labor (continued) Leather processing, 67 162; from outside area, 161; private Lewis, W. Arthur, 299 enterprise and, 173, 178-79; re- Licenses (business), 58 search and, 34; rural nonagricul- LMFS. See Labor-managed firms (Yugo- tural development and, 270-72, slavia) 272-73; sectoral shifts of, 14-15, Local state enterprises (LSES): ac- 50, 112-13, 114, 115, 121; state counting systems and, 409-10; enterprises and, 96; as substitute for comparison data and, 393; cost capital, 234, 237; TVP development structure and, 410-11; economic and, 53, 61-62, 76-77, 81, 87, 91, sector comparison and, 393-94; in- 97, 99, 116; in Wuxi, 139-40, 157 dustrial composition comparison Labor costs: increase in, 61-62; rural and, 411; kinship ties and, 395-96; labor market analysis and, 318-19 labor sector comparison and, Labor-managed firms (Yugoslavia) 394-95; managers and, 397, (LMFS), 414, 416-418 400-04; ownership and, 396-400; Labor market: defining rural, worker characteristics and, 404-08 299-302; disequilibrium in, 302; freeness of, 287-88; labor costs Macao, 141 and, 318-19; labor shortage and, Machinery, 67, 90, 102, 103, 108, 319; rural industrialization and, 138; farm, 119; penbaoji (soybean 299; surplus in agricultural, 302; meat), 92-93 TrvPs and, 302-07, 317--19; wages Management: community government and, 324 and, 5; costs, 410; failures and, Labor mobility, 413; income inequal- 234-35; fees, 5, 364, 366, 369, ity and, 364; rural-urban and 411; fiscal predation and, 379-80; urban-rural, 270-71, 272-73; town- township govemments and, 351, ship governments and, 359-6C0; TVPS 353-56; TVP, 353, 354 and, 302-07, 317, 319; wage deter- Managers: collective contract system minants and, 326, 334, 337 and, 330-31; incomes of, 332-33; Labor relations, 285, 287; data source objectives of, 227; origins of TVP, and analysis of, 280-82; ownership 180, 327; pay, 286, 297 and, 296; TVP models arid, 282-84 Manufacturing sector, 24, 103 Labor utilitzation and productive effi- Mao Zedong, 119 ciency, 243, 244-45, 245-47, Market environment for TVPs, 323; in 250-53, 254 sample counties, 91, 95-96, 98, Land: arable, 4, 48-49, 352; China's 107, 158 constitution and ownership of, 146; Marketing: rvP development and, economies of scale and, 105; in- 54-55, 62, 118, 184 come from, 52; Jieshou private in- Market interactions: adjustment re- dustries and, 162; market intetac- sponses and, 92-94, 98-102, tions and, 87, 97; private enterprise 108-09; competition and, 91, and, 179; in Shangrao, :138; rVP 92-98, 107-08; description of, development and, 52-53; workers' 85-92; industrial structure and, ties to, 405 102-06, 108, 109-11; market envi- Land contracts, 328-29 ronment and, 90-92, 107; regional Land management, 115 and local imbalances and, 108; Land reform, 116, 346 township govemments and, 87-88, Land rights, 6 100, 102, 107, 109 Index 439 Migration, 4, 34; rural-urban, 112, 167 252, 253, 254; profits and, 37, 229, Mining, 67, 179-80; coal, 54, 102, 179 230, 231; resource endowment of, Ministry of Agriculture, 20 352; revenue-sharing in, 365; risk- Mondragon (Spain), 419-20, 422, 425 taking and, 95; rural-urban migra- Motivation: community governments tion and, 167; technical personnel and, 376; entrepreneurship and, in, 353; textile industry in, 234; 197-98; of factory director, 401; for township governments in, 8, 401; industrialization, 348; pay and, township size in, 3; TVCES in, 384-85; promotion and, 375-76; 141-43, 151-54, 159-61; TVP TVP development and, 364, 371, model and, 283; TVP sector in, 63, 372, 374 65, 66, 67, 70, 71, 72, 73, 82, 84, 97, 123, 124, 153-54, 160-61, Nanhai County: agriculture in, 222; 173; underground private firms in, bank loans and, 160; capital and, 7; wage and income balancing in, 78, 234, 235, 236-37; commune 327, 328, 329; wages in, 99, and brigade industries in, 138, 141, 141-43, 152, 154, 160, 161, 143; community model in, 141-43; 294-95, 308, 309, 310, 311, 312, competitive market and, 95-96, 313, 314, 316-17, 374, 375; work- 107; credit and, 223, 224, 225; ers in, 37, 76, 161, 290, 405 cross-regional investment and, 239; National economic readjustment pro- data and fieldwork and, 21, 25-26, gram (1981), 119-20 27, 31; development in, 8, 35, National Rural Industry Administra- 137, 138, 356, 357, 381; economies tion, 135 of scale and, 241; employment in, Natural resources, 179-80 161, 325, 326, 334, 335, 336, 337; enterprise control and, 400; extra- Ouchi, W., 422 budgetary funds in, 343; financial Output: agricultural, 48, 49, 53, adjustments and, 99; firm failures 113-14, 115; farm household, 173; in, 100; geography and history of, in Jieshou, 145; market interactions 74, 75; government policies and, and TVP, 85-86, 99, 101; ownership 79, 80; grain production in, 141; and TVP, 195; per township enter- imports and, 141; income differ- prise worker, 350; TVP gross, ences among enterprises in, 330, 65-66; TVP growth and, 13, 16, 331; incomes and pay in, 138, 143, 17, 47; value of private enterprise, 152, 159-60, 289, 361, 373, 406; 174, 181; in Wuxi, 157, 158. See industrial structure in, 105; joint also Gross value of industrial output public-private firms in, 159-61; Ownership: adjustment and, 102; bud- kinship ties and, 395; labor in, 34, get constraints and public, 55; 162, 178, 285, 287-88, 318, 372; changes in form of, 422-24; com- labor mobility in, 302, 303, 304, munity government and TVCE, 5, 6, 305, 306, 307, 317; management in, 346-48, 351; comparative analysis 353, 354; ownership in, 8, 36, 69, and, 396-400; contradictions in 114; ownership forms of rvPs in, system of, 168-69; control of in- 195, 198, 205, 208, 211-12, 213; come and, 327; distribution to work- private enterprise in, 159-61, 181, ers and, 291; entrepreneurship 182, 382; production teams and, and, 190-99, 202, 207, 209-12; 138, 141, 143, 151-54, 159; pro- historical background on ductive efficiency and, 243, 249, (1958-80s), 134-37; income distri- 440 INDEX Ownership (continued) labor supply and, 173, 178--79; bution and, 289, 324; of Jieshou land and, 179; market expansion County Colloidial Factory, 146-47; and, 57; in Nanhai, 159-61; natu- labor flows and, 271; labor relations ral resources and, 179-80; owner- and, 296; land, 140, 146; main fea- ship and, 89, 167, 184, 18'5, 186, tures of TVP, 166-67; rnarket inter- 187; partnerships and, 56, 1.50, 159, actions and, 86, 87, 89; private en- 413; policy recommendations on, terprise and, 89, 167, 1L84, 185, 186-87; production costs in, 181; 186, 187, 209-12; private (in West), product orientation and, 90; prop- 419-22; productive efficiency erty relationships and, 184--86, and, 247, 248, 251, 253; regional 187; reinvestment and, 183-84; variations in, 8; of Shuguang Food- rural monetary accumulation and, stuffs Company, 147-48; TVCES and, 176-77; sponsorship of, 382-83; 124, 125; TVP developrment and, staffing problems of, 184; taxes and, 62, 67-69, 80, 413; wage equa- 369; technology and, 180; textile tions and, 291-92; wages and, 324 industry and, 101; urban sector and, 242; wages in, 373; worker re- Partnerships, 56, 150, 159, 413 cruitment and, 329. See also Town- Pay. See Wages ship, village, and private enterprises Penbaoji (PBj-soybean meat machin- Producer cooperatives, 414, 4:15 ery), case history, 92-93 Production responsibility system (PRs), Pension system, 328 50, 168, 173; agriculture and, 6, Perkins, Dwight H., 19 114, 115; defined, 3; effects of, Peru, 414 344, 346, 347, 348; employment Petroleum, 17, 102, 305 and, 377; fiscal predation arid, Piece-rate wage system, 243 378-79; income distribution and, Planning, 86, 398-400 4, 363; privatization and, 6; status Policy documents and rural industry, hierarchy and, 145-57; TVCEs and 56, 113, 120 production teams and, 149, 151; Population: growth in, 48, 49, 112; TVP development and, 47, 79; wage immobility of, 4 system changes and, 7 Prices: agricultural, 114, 116, 11.7; Production structure, 261-64 competition and, 96; controls on, 51 Production team, 6, 79; defined, 3; Private enterprise, 6; adjustment: and, employment and, 135; finance and, 100; administration and, 187; bank 345; grain rations and, 135, 143; in loans and, 80, 177, 184; benefits Nanhai, 138, 141, 143, 151-54, from, 173-76, 381; competition 159; risk and, 213; status hierarchy and, 176; development and, 126; and, 151-54; TvcEs and PRS and, differences between comimunity en- 149, 151; in Wuxi, 148, 149, 150 terprise and, 290-91; efficiency Productivity: agricultural, 50; rural and, 181-83, 243, 247, 248; emer- nonagricultural (by region), 265-73 gence of, 135; employment by, 183; Profit remittances: township govem- entrepreneurship and, 195, 196, ments and, 364, 366, 370, 378; 197, 202, 205, 208-09, 213, 216; TVP institutional framework and, 5 exit by, 109; financial flows and, Profits, 37, 77, 211, 384; capital pro- 369-70; govemment position on, ductivity and, 235; distribution of, 80, 147, 184, 272, 381; growth of, 229-32, 234; distribution of work- 69, 172-73; in Jieshou, 162-74; ers and, 291; expansion and, 59; Index 441 in Jieshou, 145, 164; labor Regulation, 184, 185, 187 utilization and, 246; market inter- Reinvestment: capital formation and, actions and, 89, 91, 101; mining 229; private enterprise and, and, 179; nonprice competition 183-84; township govemments and and, 55; pay and, 296; private en- TVPs and, 365, 368, 370 terprise and, 181, 186; in Shangrao Research, 20; background on TVP, 144, 154, 155; subsidization of agri- 19-21; data set evaluation and, culture by industrial, 56, 80, 114; 32-38; data set on TVP, 25-32; textile industry and, 234; total in- study objectives and design and, dustrial, 105; township govem- 21-25 ments and, 124, 125, 227, 345, Resource allocation, 173, 352 351, 357, 365, 366, 369, 370, 374; Revenues: sharing of, 4, 365, 366, TvCEs and distribution of, 229-32; 397; township, 342-43, 344, 345, per TVCE worker, 71; Tvps and, 347, 348, 350-51, 364-70, 61, 62, 232; workers' perceptions re- 370-71. See also Extrabudgetary rev- garding, 285-86, 287; in Wuxi, enues and expenditures; Taxes 149, 157, 158 Reynolds, Lloyd G., 299 Property relations: industrial develop- Riskin, Carl, 19 ment and township governments Risk-taking: entrepreneurship and, and, 346-48; private enterprise 190-91, 203-06; TVPs and, 166; and, 184-86, 187 workers' attitude toward, 407 Property rights: Chinese tradition Rural credit cooperatives (Rccs), 200, and, 140, 141; entrepreneurship 201, 203, 204, 215 and, 216; farmers and, 166; TVPS Rural nonstate enterprises. See Town- and, 166, 168. See also Ownership ship, village, and private enter- Public-private firms, 159-61 prises Rural sector: community structure in, Questionnaires: Enterprise Director, 4-6; development model for, 392; 30, 31, 33, 241-42, 355; Enterprise employment flows and, 270-71, Quantitative, 30, 32, 33, 224, 272-73; income distribution and, 241-42, 353; Enterprise Survey, 362-64; industrialization and pri- 29-30, 33, 303, 304, 310, 393; vate enterprise and, 175; private Township Leader, 31, 33, 227, 353, enterprise and monetary accumula- 399; Worker Survey, 29, 33, 200, tion and, 176-77; reforms in, 6-7, 280, 303, 351, 362, 393 79; regional variations in, 7-9; rural industrialization and, 117-22, Raw materials, 101, 349-50, 356 172; similarities between urban sec- Rccs. See Rural credit cooperatives tor and, 393-96 Regional imbalances in economic de- velopment: employment and, 264, Salaries, in govemment, 365, 370, 270-71, 272-73; rural nonagricul- 372-75 tural development and, 265-73; Salesmen, 90 TvcBs and, 260-61, 264; TVCE sub- Savings: farmer, 177; nonagricultural sectors and, 261-64; TVP develop- development and, 267, 270; rural ment and, 108, 255-60; TVP de- sector and, 51 velopment during 1979-85 and, Schumpeter, Joseph A., 190, 191 264-65 Sectoral composition: industrial struc- Regional variations among TVPS, 7-9 ture and, 102-03; TVP, 65-66 442 INDEX Service sector, 65-66 Status hierarchy of industrial firms: Shangrao County: agriculture in, 222; before PRs, 137-45; hiring outside backward townships in, 39; bank workers and, 161; impact of PRs on, loans and, 147; capital and, 51-52, 145-57; joint public-private firms 78; commune and brigade industries and, 159-61; local governments in, 144; competitive market and, and, 162-64; origins of comnmunity 107; credit and, 223, 224; data and enterprise and, 139-45; private fieldwork and, 21, 26, 27, 31; de- firms under collective auspices and, preciation funds and, 155; develop- 145-48; poor TVCE performance ment in, 35, 118, 137, 138, 356; and, 154-56; production team en- employment in, 154, 155, 352, terprises and, 151-54; technology 353; entrepreneurship case stjdy in, requirements and, 164-65; treat- 193-95; extrabudgetary funds in, ment of firms and, 135; TVCES in 343, 397; fiscal predation in, 379, Wuxi and, 148-51, 157-59 380, 385; geography an,d history of, Stock ownership by employees, 424 76; government policies and, 79, Subsidies: farmer, 328; food, 333 80; hydropower in, 164-65; in- Switzerland, 414 comes in, 143-44, 289; industrial structure in, 105; kinship ties in, Tanzania, 414, 418-19 395; labor in, 34, 285, 288, 372, Tax breaks and reductions, 115, 230 394; labor mobility and, 304, 305, Taxes, 153, 160, 181, 359, 364, 365, 306, 307, 317; land in, 138; :man- 366; on bonuses, 336; capital pro- agement in, 234, 353, 354, 355; ductivity and, 235; community gov- ownership in, 8, 36, 69; ownership emient and, 370; differences in, forms of TVPS in, 193-95, 198-99, 232; private enterprise and, 369; 203, 204, 205; private enterprise profits and, 230; TVP growth and, in, 57, 181, 182; productive effi- 13, 15 ciency and, 243, 245, 249, 252, TECS. See Township economic commis- 253, 254; productivity and, 105; sions profits and, 37, 144, 154, 155, 229; Technology, 137, 163; hydropower promotion incentives and, 375; rev- stations and modem, 164-65; pri- enue sharing in, 365; rural-urban vate enterprise and, 180; TVP devel- similarities and, 393; subsectoral opment and, 77 composition in, 102; technical per- Textiles, 67, 101, 102, 103, 120, 234; sonnel in, 353; township leader pay silk, 17, 178, 329 in, 374; township size in, 3; IVCES TICS. See Township industrial corpora- in, 143-44, 154-56; TVP model tions and, 283; TVP sector in, 63, 65, 66, Todaro, Michael P., 53 67, 69, 70, 71, 72, 83, 84, 123, Township economic commissions 371; wages in, 99, 144, 154, 308, (TECS), 384; development and, 118 309, 310, 311, 314; welfare mea- Township govemments, 3, 8; agricul- sures in, 407; workers itn, 37, 77, ture and, 343, 347, 348, 352, 356; 289-90, 405 and brigade industries, 345, Sigurdson, Jon, 19 346-47; centralized nature of, Spain (Mondragon example), 419-20, 342-43; community property and, 422, 425 347-48; data set and, 25-27; devel- State Council Rural Development Re- opment issues and, 124-25, 127; search Center, 20 dual functions of, 348-50; eimploy- Index 443 ment and, 124; expenditures of, gions) and, 261-64; money-losing, 342, 343, 344, 364-70, 370-71, 225; in Nanhai, 141-43, 151-54, 377, 383, 386; factor immobility 159-61; ownership and, 124, 125, and, 359-60, 386; grants by, 87, 195, 196, 197, 324; penalties for 89; income distribution and, 324, leaving, 150; poor performance by, 360-61; industrial structure and, 154-56, 207; private enterprise 105-06; investment and, 349, 351; and, 175, 177, 185; production labor immobility and, 359-60; mar- teams and, 149, 151; profit distribu- ket interactions and, 87-88, 100, tion and, 229-32; regional imbal- 102, 107, 109; officials in, 365, ances and, 260-61, 264, 267-68; 370, 372-76; operation of, 344-46; in Shangrao, 143-44, 154-56; ownership and, 5, 6, 351, 383-84; status hierarchy analysis and, profit remittances and, 15, 364, 148-51; structural reform and, 366, 370, 378; profits and, 124, 123-27; in Wuxi, 138, 139-41, 125, 227, 345, 351, 357, 365, 148-51, 152, 157-59, 166, 356. 366, 369, 374; public finance and, See also Brigade industries; Com- 342-50; revenues of, 342-43, 344, mune industries 347, 348, 350-51, 364-70, Township, village, and private enter- 370-71; status hierarchy analysis prises (TvPS)): accounting systems and, and, 162-64; TICs and, 226; town- 409-10; adaptability and, 167-68; ship property and, 346-47; TVCES adjustment and, 92-94, 98-102, and, 5, 227, 349, 350-52; TVP de- 108-09; agriculture and industrial velopment and, 5, 6, 62, 63, development and, 48-55, 113-17, 226-29, 272-73, 358-59, 365, 222-23; bank loans and, 78, 368, 377, 382; TVP financial flows 223-26; capital and, 51-52, 61; and, 364-70; Tvp relations and, capital productivity and, 233-42; 376-83 communal model of, 281, 283; com- Township industrial corporations munity govemments and, 5, 6, 62, (TICS), 226, 351, 384 63, 226-29, 272-73, 358-59; com- Township Leader Questionnaire, 24, pared with other types of firms, 31, 33, 227, 353, 399 414-15; competition and, 90, Townships: defined, 3; TVP data set 92-98, 107-08; cooperation of and, 25-27 TVCES with, 97; cost structure and, Township and village community en- 410-11; defined, 3; development terprises (TVCES): bank loans in patterns and, 81-83; description of, Wuxi and, 158, 159; capital and, 413-414; Eastem European state 200-01, 208, 213; community gov- enterprises compared with, 415-16; emments and, 5, 227, 349, efficiency and, 243, 245-47, 350-52; comparative analysis and, 250-53, 247-50, 254; employment 393-400; defined, 3-4; economic and, 62, 372; entrepreneurship and, reform and, 272; economies of scale 189-90, 191-95, 195-97, 198, and, 239-42; efficiency and, 247; 199-203, 208, 209-12; equipment farm labor income and, 328; firm and, 54-55, 62; financial flows size and, 239; growth and develop- among, 364-70; firm size and, 88; ment of, 11-13, 69-71, 139-45, future prospects for, 169-70; gross 264-65; income and, 69, 71; in output and, 65-66; historical back- Jieshou, 144-45, 146, 156, 356; ground on, 9-11; industrial compo- leading industrial subsectors (by re- sition comparison and, 411; institu- 444 INDEX Township govemments (continued) towns in, 420-21, 422, 423; em- tional framework of, 4-9; Japanese ployee stock ownership in, 424; model and, 283; labor and, 53, Z-firms in, 422 61-62, 321; labor flows among, Urbanization, 112-13, 115, 120, 121, 302-07, 317, 319; land and, 122, 169, 170 52-53; location and, 122; manage- Urban sector: economic reforms and, ment of, 353, 354; managers of, 48; employment flows and, 270-71, 397, 400-04; market envirornment 272-73; employment problems in, and, 323; marketing and, 54-55, 163; industry in, 115; mass banish- 62; money economy and, 87; orga- ment from, 135; private enterprise nizational characteristics and prob- and, 242; rural industrialization lems of, 62, 165-66, 168-69; own- and, 117-22; similarities between ership and, 62, 67-69, 80, 166-67, rural sector and, 393-96 325, 383-84, 396-400; ownership form changes and, 422-25; pay in- Villages: administrators of, 152, 153; centives and, 384-85; perfonnance defined, 3 of, 71; present situation of, 11-18; private firms compared with, Wage determinants: employment and, 419-22; property rights and, 166, 325-27; income differences within 168; regional imbalances and, 108, firms and, 330-32; monthly income 255-65; regional variations in, 7-9; composition and, 332-34; theory research background on, 19-;21; re- and, 323; Tvps and, 323-24 search data and, 25-32, 32-38; re- Wage equations, 291-96, 297 search project objectives and design Wages: balancing, 327-29; bonuses and, 21-25; rural labor market and, and, 55; case of Shuguang Foodstuffs 317-19; structural reform and, Company and, 147; comparative 123-27; Tanzanian firms comlpared analysis of LSE and TVP, 406; distri- with, 414, 418-19; Taylorist imodel bution of profits and, 230-32; gov- and, 283; township govemment in- emment officials and, 365, 370, volvement with, 376-83; wage de- 372-75; income linkage to commu- terminants and, 323-241; Yugoslav nity and, 372-75; increase in, 62; firms compared with, 414, 416-18. in Jieshou, 144-45, 154; labor utili- See also TVP subentries for Jieshou zation and, 246, 250; managerial County, Nanhai County, Shangrao pay and, 5, 286; motivation and, County, and Wuxi County 384-85; in Nanhai, 141-43, 152, Trade: foreign, 71-72; among regions, 154, 160; pay systems and, 285, 116 286-87, 291, 296; private enter- Trade credit, 206. See also Credit prise and, 184, 373; productive effi- Trade unions, 187 ciency and, 247; PRS and, 7; in Transport, 66 Shangrao, 144, 154; township con- TVCES. See Township and village com- trol and, 8; TVPs and, 80, 81, 99, munity enterprises 101, 102, 331, 384-85; TVP sector TvPs. See Township, village, and pri- growth and, 15, 227; TVP sector vate enterprises and township govemment and, 361, 363, 365, 373; of women, Unemployment, 117, 395, 396, 425 292, 297; in Wuxi, 139, 149, Unions, 187 150, 157 United States, 115, 360-61; cornpany Wage transfers, 145, 148, 149 Index 445 Welfare system, 407, 410 191-93; equipment and, 54; extra- "Wenzhou" development model, budgetary funds in, 343, 397; fiscal 34-35, 123, 176, 177, 181 predation in, 378; funds from in- Williamson, Oliver E., 98 dustry and, 345, 347; geography Women: pay of, 292, 297; as TVP and and history of, 74-75; government LSE workers, 406; worker data sam- policies and, 8, 79, 80, 81, 84; ple and, 284 handicrafts in, 148; human re- Wong, Christine, 19 sources in, 118; incomes and pay Work-point wage system, 6, 243, 247, in, 141, 150, 151, 158, 332, 333, 333 372-73; industry in, 47, 105, Workers: attitudes of, 286, 289-90, 137-38, 157, 158; interest pay- 296-97, 407; comparative analysis ments in, 158; investors in, 114, and, 404-08; data set and, 37; edu- 158, 159; kinship ties and, 395; cation of, 405, 406; hiring and fir- labor in, 56, 139-40, 157, 318-19, ing of, 303, 329; hiring outside, 372; labor mobility and, 303, 304, 161; incomes of, 286-87; labor al- 305, 307, 319; machine production location and, 285; labor relations in, 90, 108; management in, 353, and, 285, 287; pay systems and, 354; marketing and, 54; ownership 285, 286-87, 291, 296; perceptions in, 8, 36, 67, 69, 384; planning regarding profit and, 285-86, 287; and, 86, 398; production teams in, research on, 24; satisfaction level 149, 150; profit and, 37, 125, 157, of, 284-85; wage determinants and, 158, 229, 230, 231; raw materials 327-29 and, 349-50, 356, 398; resource Worker Survey Questionnaire, 29, 33, endowment of, 352; restrictions on 200, 280, 303, 351, 361, 393 private enterprise in, 9; revenue- World Bank, 19, 20, 21 sharing in, 365, 366; risk-taking Wuxi County: adjustment and, 99, and, 95; rural-urban similarities in, 100-01; agriculture in, 49, 50, 56, 393; subsectoral composition in, 113-14, 158, 222; backward town- 102; technical personnel in, 353; ships in, 39; bank loans and, 158, TICS in, 351; township size in, 124; 159; business licenses and, 58; capi- TVCES in, 138, 139-41, 148-51, tal and, 52, 78, 234, 236-37; com- 152, 157-59, 166, 356; TVP owner- mune and brigade industries in, 7, ship in, 191-92, 195, 197, 198, 139, 140-41, 143; community en- 202, 205, 206, 207, 212, 213, 214, terprise model and, 139-41, 143; 215; TVP sector in, 57, 63, 65, 69, competitive market and, 91, 95, 70, 71, 73, 76, 82, 97, 98, 123, 96, 98, 107, 158; consumption in, 124, 126, 157, 371; wage and in- 60; credit and, 223, 225; data and come balancing in, 327, 328, 329; fieldwork and, 21, 25, 31; develop- workers in, 37, 76, 283, 285, 286, ment in, 8, 35, 48, 137-38, 357, 287, 289, 290, 291, 295 381; economies of scale and, Wuxi Petroleum Equipment Plant, 239-41; efficiency and, 25, 408, 191-92 409; employment in, 157-159, 325, 326, 334, 335, 336, 337, 352; en- Yugoslavia, 414, 416-18 terprise directors in, 351; enterprise income differences in, 330, 331; Z-firms, 421-22, 425 entrepreneurship case study in, Zhou Qiren, 20 The backlist of publicatLons by the World Bank is shown in the annual Index of Publications, which is available from Publications Sales Unit, The World Baink, 1818 H Street, N.W., Washington, D.C. 20433, U.S.A., or from Publications, Banque mondiale, 66, avenue d'Ina, 75116 Paris, France. Zg" The World Bank Q A striking feature of recent economic development in China is the emergence of rural industrial enterprises, run in some cases by local governments, in others by private groups and individuals. The proliferation and growth of these firms-many of which show great flexibility and energy- have significantly increased personal incomes among rural Chinese and have begun to change the structure of the rural labor force. Rural industry has become the most dynamic part of the Chinese industrial sector. This book, the product of a pathbreaking collaborative effort by Chinese and World Bank researchers, examines the genesis, institutional structure, operations, problems, and prospects of this grass-roots industrial activity. An unprece- dented quantity and variety of data were gathered, mostly in four representative counties, where local government officials, factory managers, and employees were interviewed in depth. The results of the research provide insight into the economic picture at the local level: how firms are started and operated, the relations between firms and governments at all levels, and the present and potential contributions of nonstate firms to rural development, local finance, living standards, and em- ployment. Both successes and problems are illuminated through a number of case studies. A final chapter compares Chinese rural industry with analogous enterprises in other -S. countries. At the time of writing, William A. Byrd was an econo- mist in the China Program Division, Asia Region, The World Bank. Lin Qingsong was deputy director of the Institute of Economics, Chinese Academy of Social Sciences. Cozier designi by Beatrice Sito Oxford University Press ISBN 0-19-520822-6 Oxford