1 44zll How Industrial Reform Worked in China: The Role of Innovation, Competition, and Property Rights Gary H. Jefferson and Thomas G. Rawski In China early atempts at partial reform unliashed forces that, fifteen years later, have brought China's economy to the brink of a market system The participation of tens of thousands of enterprises and millions of administrators, managers, and wrk- ers over the duraion of the reform euentually built a constituency for market- directed change that was far stronger than any official announcment could have producet Gradual and partial reform shifted the economy toward a market system under a regime of growth, improved productivity, accelerated technial change, and rising aeports. Reactions of firms and governments focused increasingly on innova- tion, cost reduction, and ftirther deregulation, deepening the cmnulative inpact of reforn, ratber than on rent-seeking and subsidies. This process of reform is very dif- ferent from the top-down, centrily planned approaci to reform that is zwidely advo- cated by btenational organitions ana economic researchers, but it has produced a durable reform consttuency that easily rebuffed higb-level efforts to roll back reform in the wake of the inflation scare and political repssion of 1989. hina's partial and gradual reform has combined rapid economic progress with institutions and policies that deviate widely from standard prescrip- oJtions for reform. China's reforms present economists with a puzzle. Why has China "grown so fast when conditions thought to be necessary for growth... were absent?" (Blanchard and Fischer 1993, p. 4). How did China's unorthodox reforms spark an economic surge that has far outpaced results in other ex-socialist economies and in many developing nations? Many economists view the reform of former socialist economies as a process of replacing old institutions with new structures in an organized top-down fashion Gay HF Jcffceson is assocatc professor of economics at Brandeis University. Thomas G. Rawski is profs- sor of economics at the University of Pitsburgh. Thc authors received support for this rcsearch from the Hcnry Lace Foundation, the John S. Guggenheim Foundation, the Univrsity of Pittsburghs University Ccnter for International Studies, Brandeis Univsity's Maser Fund, and the World Bank's Project on Industial Rcform and Productvity within Chinesc Enterprises. The authors are gratcful for the research assstnce of John Zhiqiang Zhao and for data and comments from E.C Hwa, NEicholas R Lady, Wei Lo, PCnDCope Prime, Jcffrey Sachs, Indeijit Singh, Wmg T. Woo, and Shahid YusuL Proceedings of the World Bank Annual Conference on Deelopment Economics 1994 01995 The Intrnational Bank for Reconstruction and Developmcnt/ nmEwoRLD BANK 129 ije~ -130 How I)HwmnralRenn lmked m Chmia directed by refornmers. The self-interested response of agents within the economy is expected to stimulate profit-seeking behavior and market activity. If progress is inad- equate, planners can impose further rounds of reform. This type of centrally directed reform has undoubtedly played a role in China's economic transition. It was China's central leadership that initiated economic reforms in the late 1970s, exparrding the role of prices and market allocation, rolling back long-standing barriers to international trade and investment, transferring authority from central planners to enterprise managers and local governments, cre- ating a system of dual (plan and market) pricing for industrial goods, and so on. But unlike the postcomnunist leaders in countries like Poland and the Czech Republic, China's policymakers embarked on a path of reform with no dear vision of what a restructured economy should look like and no consensus about what policy mix or institutional arrangements would best get the economy there (Hua, Luo, and Zhang 1993; Shirk 1993; Naughton 1994). Not surprisingly, policy announcements from the center were partial and tenta- tive. The center ratified but did not direct the momentous shift from collective to household farmin4 Central initiatives in industrial reform focused on the incre- mental relaxation of controls over state-owned enterprises. Even the revolutionary "open door' strategy, reflected in a sequence of central decsions that shattered long-standing barriers to China's participation in the world economy, concentrated on expanding trade and investment actvity in a small number of provinces and spe- cial zones along China's southeast coast. From the usual top-down perspective China's recent economic gains seem remarkably large in relation to the central government's modest reform initiatives. In exploring this anomaly, we focus on an analysis of China's rufarm dynamics that shows how tecdnical innovation, economizing behavior, market-leaning institutional changes, and a multitude of cumulative and mutually reinforcing choices by admin- istrators, managers, and workers reinforced and eventually overshadowed Beijing's partial reform efforts. The focus is on industry, which is both the largest sector of China's economy and the core of its reform problem. Successive rounds of partial reform have cumulated into significant changes in industrial structure, conduct, and performance affecting every type of firm, indud- ing old-line state firns. In this process partial reform initiatives produce unantici- pated outcomes in an interplay of action and reaction among changes in economic conditions, ad hoc policy measures by various levels of government, and uncoor- dinated strategizing by enterprises and individuals. This interaction occurs in an environment of intense competition involving several types of firms, each with its own distinct technical capabilities and institutional constraints. Partial reform expands entry into product markets, and the ensuing intensification of competition erodes enterprise profits and undermines the revenue base at every level of gov- ernment. These financial strains generate prcssures that promote innovation and cost reduction. Government efforts to ease the revenue constraint and enterprise efforts to innovate and reduce costs lead to fresh rounds of market-directed insti- tutional change. Jefferson and Rauski 131 What Needs to Be Explained about China's Industrial Reform China's recent industrial achievements deviate from the orthodox reform scenarios in the continued dominance of the public sector; its improved performance, despite the absence of any plans for privatization; and the continued presence of important defects in the institutional underpinnings of China's industrial economy. Rapid growth is amply documented in World Bank repor;s, which reveal the broad-based nature of industrial expansion. Joint ventures, 'ioreign-owncd firms, and private industry contributed only one-sixth of incremenal output gains during 1980-92 (table 1). Despite the highly publicized economic boom in China's south- em coastal provinces, the region's 36 percent share in agregate industrial output in 1992 was only modesdy above its 30 percent share for 1978-and below the 36.6 percent share in 1952 andustry 1949484. p. 145; Survey 1993, p. 72). The Dominance of Public Enterprise The output share of state-owned industry (firms "owned by the whole people" and direcdy or indirecdy controlled by agencies of the central government) plunged from three-qiarters to less than half between 1980 and 1992 (table 1). This decline reflects dramatic gains by collectives, especially the rural firms known as township and vil- lage enterprises (or TVEs). The 1992 output share of public sector industry, however, which includes state firms and urban and rural collectives, exceeds 85 percent. The output share of private domestc firms remains small, at less than 10 percent? The explosive growth of TVE output has aroused intense interest TVE firms, although different in many respects from state enterprises, are public enterprises. Calling TVEs "nonstate enterprises" conveys the misleading impre-sion that rural colectives operate independently of officialdom. Some authors have speculated that TVE firms "mimic private enterprises" or operate like "loosely-structured coopera- tives" (Singh, Ratha, and Xiao 1993; Weitzman and Xu 1994). TVE operations are closely monitored and often controlled by "local govemment entrepreneurs" (Zweig 1991, p. 720) who "exhibit characteristics of both de facto owners and senior managers of township corporations" (Whiting 1993, p. 6)32 The TVE sector is built on the foundation of earlier industrialization efforts undertaken by local gov- ernments (Perkins and others 1977). Like their predecessors, the TVE firms of the 1980s and 1990s operate 'under close supervision from the township or village industrial departments" (Wong, Heady, and Woo 1993, chap. 9), which contribute start-up funds, appoint managers, and "are intimately involved in major strategic decisions" (Ody 1991, p. iv).3 Improved Performance in State Industry No one disputes the significance of the contributions that TVEs have made to the growth of production, exports, productivity, employment, incomes, and material welfare. Their success shows the insufficiency of assuming that a full and immediate Table 1. Overuiew of Inddustrial Performance in Chitna, 1980-92 1980-92 (percent) Index of real output (1980-110 S17ares in nominal output (percent) Annual Sbare of Ownership type 1985 1990 1992 1980 1985 1990 1992 average growth incremental output State 148 210 257 76.0 64.9 54.6 48.4 7.8 43.6 Collective 247 554 914 18.4 Urban 13.7 13.3 10.3 11.8 11.5 Township-village 9.9 18.8 25.3 26.2 28.8 Private' 21,752 126,057 241,455 0.0 1.9 5.4 6.8 64.9 7.9 Otherb 492 3,530 8,736 0.5 1.2 4.4 7.2 37.2 8.3 Total 176 328 480 100.0 100.0 100.0 100.0 13.1 100.0 Total output (Y billion) 515.4 971.6 2,392.5 3,706.6 a. Privately owned firms employing fewer than eight workers. b. Includes private firms employing eight or more workers, joint ventures, foreign-owned firms, and other ownership forms I Souirce: Yearbook 1993, pp. 409-13; Rawski forthcoming. it efferson and Rawski 133 end to official influence over enterprise management, preferably through privatiza- tion, is a necessary step in the reform of socialist systems. But the inconsistency between recent Chinese industrial experience and free-mar- ket orthodoxy runs even deeper. China's old-line state enterprises have responded to ongoing partial reform by behaving less like passive bureaucratic followers and more like profit-seeking commercial businesses. Reform has altered the objectives, incentives, and corporate culture of state firms, bringing substantial improvements in performance. With state industry accounting for nearly half of industrial output and absorbing 35 percent of aggregate fixed investment, China's recent economic gains could hardly have occurred had state industry served only as a drag on eco- nomic progress.4 This article is not the place for a detailed review of evidence supporting these assertions.5 The discussion here is limited to the following propositions about state industry, each resting on substantial empirical foundations: - State enterprises, formerly devoted to plan fulfilment, now take profit as their chief objective. Data on state enterprise performance generate increas- ingly robust statistical relationships of the ldnd expected from profit-seeking firms operating in a competitive market setting. * State enterprises have achieved substantial increases in labor productivity and steady, although modest, increases in total factor productivity.6 * State enterprises have sharply increased the pace of research and develop- ment, new product development, and process innovation (Jefferson, Rawski, and Zheng 1992a,b). * Exports of state enterprise manufactures, which increased at an estimated annual rate of 18 percent between 1985 and 1992, reflect the impact of greater attention to quality, variety, customer requirements, and cost control7 * State enterprises constitute an important and often crucial source of technol- ogy, equipment, funds, information, expertse, and marketing opportunities essential for the successful development of TVEs. Persistence of Institutional Weaknesses In looldng at China's recent achievements, the challenge is to explain dramatic gains in industrial performance in the absence of comprehensive efforts at the center to promote liberalization and institutional change. The key queston concerns the process that has moved public sector firms reared under state planing to place unprecedented emphasis on efficiency, quality, and innovation, with no program or even credible threat of privatization, and to explain how China's industrial advances have occurred without the features that many economists regard as core elements of a market system. Chinese industry continues to operate in an environment of incomplete specifi- cation of property rights. Rules of commerce are neither clearly defined nor consis- tently enforced. Competing firms in the same industry or locality face widely differing legal, fiscal, and regulatory regimes. Government intervention in business 134 How lnduial Reform Wored China affairs extends well beyond the boundaries observed in even heavily regulated mar- ket economies (Japan, Republic of Korea), often with the effect of softening budget constraints. These difficulties restrict innovation and productivity growth, particu- larly, but not exxdusively, in the state sector.A They have also enabled insolvent and hopelesdy inefficient state enterprises to continue to operate, wasting large amounts of productive resources and requiring subsidies large enough to affect macroeco- nomic stability (Sachs and Woo 1993; Woo and others 1993). The cost of these insti- tutional shortcomings, although difficult to quantify, appears large. A Model for Analyzing Induced Industrial Reform We propose the following model of China's industrial reform, which we see as a cumulative process that begins when partial relaxation of the institutional con- straints associated with socialist planning initiates competition in the markets for industrial products. Competition reduces profits, creating financial pressures that induce technical innovations, promote economizing behavior, and stimulate fresh rounds of market-leaning institutional change. This model rests on four key institu- tional features of China's industrial economy: decentralized supervision, incipient competition, fiscal dependence on industrial profits, and a hierarchy of heteroge- neous enterprise types.9 Decentralized Supervision Central control of industrial enterprises was never as tight in China as in Eastern Europe and the Soviet Union (Granick 1990). Decentralization increased during the late 1960s and 1970s as the central government transferred supervision of many firms to provincial and municipal goverrnents. This system of decentralized super- vision encouraged provinces and localities to create and pursue their own industrial development strategies. When reform began, decentralized decisionmaking also made it possible to introduce piecemeal reforms and to conduct local policy exper- iments without disrupting the whole economy. Successful local reforms inspired widespread emulation. Incipient Competition The term incipient competition" well describes the circumstances of domestic Chinese markets for industrial products on the eve of reform. Actual competition was sharply limited by policies that had the effect of creating strong barriers to entry. Removing these barriers, however, quicldy revealed multiple competitors in nearly every product line. In China, unlike Russia and other countries of the former Soviet Union and Eastern Europe, deregulation leads to industial competition, not to monopoly. Competitive pressures arose from four sources. Rural industry developed widely in the decades prior to reform but was largely confined to fabricating local materi- Jefrson and Rawsei 135 als into goods for local buyers (Perkins and others 1977). Entrepreneurial leaders in hundreds of counties and thousands of production brigades were poised to take advantage of deregulation by bursting into markets that they had coveted for years. China's southern regions, excluded from large-scale inAdustrial investment during three decades of central planning, took advantage of the new open door policy to promote industrial growth with the aid of capital, skill, and commercial contacts from overseas Chinese, most of whom trace their ancestry to the southern coastal provinces. Defense conversion brought strong new entrants into a number of civil- ian industries; by the early 1990s at least two-thirds of output from defense indus- tries consisted of civilian products (Blasko 1994). Finally, China's long-standing policy of building complete sets of state-owned industries in most provinces pro- vided a ready-made source of competition. Fiscal Dependence on Industrial Profits Industrial profits and tax payments are a key component of fiscal revenue at every level of government. State enterprises contributed 80 percent or more of adjusted budgetary revenues in every year during 1978-87; in 1988 state industry accounted for 73 percent of profits and profit taxes from all state enterprises (Sicular 1992, table s and p. 3). The fiscal reforms of the 1980s created a system in which each level of govem- ment collected taxes from enterprises under its jurisdiction, "turned over a contrac- tually specified amount to the next higher level of government, and could keep the residual." The result was "a shift toward local fiscal power at the expense of the cen- ter, as the center's proportion of total government revenue fell" from 50 percent in the 1970s to less than 30 percent in the 1980s (Walder 1994, pp. 17, 19). Chinese Industry as a Hierarchy of Heterogeneous Enterprise Types China's industrial sector displays extreme heterogeneity. There is a hierarchy of domestic industrial enterprises ranging from foreign-linked firms to state enter- prises, urban and rural (TVE) collectives, and private businesses. These groups of firms exhibit systematic differences in technological capabilities, cost structures, and institutional arrangements. There is an inverse relation between innovative capabil- ity and lalbor costs. Among domestic firms, state enterprises have the greatest tech- nical strength. They also have the highest labor costs and suffer the greatest restriction from institutional constraints. TVEs are least affected by insttutional lim- itations. The interaction of these different enterprise types creates a kind of innova- tion and competition ladder. Vernon (1966) and Grossman and Helpman (1991) have developed models of international product cycles and quality ladders that focus on interactions between innovative firms in the "North" and imitators in the "South." Northern firms rely on product innovations to support their high-cost manufacturng operations. Southern finns, with lower production costs, can capture markets from northern 136 How Indwial Rform Worked in China rivals by replicating northern products. The North retaliates with fresh rounds of innovation. Rivalry among different types of producers leads to an ongoing evolu- tion of product characteristics, while the locus of manufacturing activity may shift back and forth between firms located in the 'industrial" North and those in the "developing" South. This approach fits nicely with China's recent industrial history once we extend these concepts to encompass the existence of product cycles and quality ladders within the domestic economy of nations participating in global trade as well as across high- and low-wage economies. Chinese industrial goods rarely match the quality and characteristics of products manufactured by global leaders. But the most accom- plished Chinese firms, on their own or with the cooperation of foreign partners, can produce reasonable substitutes at a low cost. Rapidly growing exports of textiles, garments, footwear, machinery, and consumer durables illustrate China's participa- tion in international quality ladders. At the top of the international quality ladder are the overseas firms that set the innovative pace in global markets. These firms define technical and quality standards that Chinese firrs seek to attain. Chinese industry has its own hierarchy of firms that generates domestic versions of the rivalries, flows, and pressures associated with global product cycles. On the top rung of China's domestic ladder of technological capabilities are foreign-invested firms; below them are state enterprises, urban col- lectives, TVEs, and privately owned firms. We focus on the three largest categories. * Foreign-invested firms. Close foreign links give these firms better access to foreign capital and technologies than purely domestic competitors. While foreign-linked firms operate under many of the restrictions that apply to state enterprises, they enjoy favorable tax treatnent and spedal autonomy in labor-management relations, wage setting, and foreign trade. China's open door policy has allowed these firms to flourish, with beneficial effects on export growth, technology inflows, and the spread of opportunities for purely domestic firms to explore new approaches to production, manage- ment, and sales. - State-owned enterprises. State enterprises were the traditional centerpiece of economic plnnming. Decades of favorable treatment have endowed them with a legacy of technical capabilities surpassing those of other domestic enter- prise types. At the same time, state enterprises bear heavier burdens of social responsibilities and bureaucratic intrusion than any other type of enterprise. - Collective enteiprises. Urban and rural collectives are owned by local gov- ernments and sometimes by state enterprises. Their operations are generally more labor-intensive than state enterprise production; their products typi- cally cluster at the low end of the price-quality spectrum. Some of these firms have begun to apply modern technologies and sophisticated equipment to produce goods that can compete in national and international markets as well as in local markets. There is a well-defined hierarchy of domestic technological capabilities extend- ing downward from joint ventures to state enterprises and collectives. Among purely Jeffeon and Pawsl 137 domestic firms state enterprises enjoy superior endowments of equipment, labora- tory facilities, and skilled and educated workers, technicians, and managers. This resource differential favoring state firms sLws up in the outcomes of quality inspec- tions. In the fourth quarter of 1993, for example, 73.8 percent of goods from a sam- ple of 2,221 firms passed inspection; the success rate was 90 percent for 'big state enterprises" but only 62 percent for "township and private manufacturers" (Ma 1994).10 Managers of urban collectives and TVEs (as well as state enterprises) over- whelmingly identify state firms as the domestic technological leaders in their own industries (Jefferson, Rawski, and Zheng 1992b). As in the transnational version of the product cycle, technology percolates down China's domestic quality ladder. The example of bar coding illustrates how links with international markets necessitate the mastery of suitable technologies and the achievement of specific quality standards that gradually spread into the domestic economy. Export-oriented firms have learned to use bar codes, which they see as "tickets to foreign supermarkets." Today only a few dozen domestic retail outlets use bar codes. But by the end of 1995 "China aims to have bar codes on all its exported goods and 60 percent of domestic commodities" (Sun 1994, p. 8). One imnportant and widely overlooked aspect of Chinese quality ladders is the dependence of many TVE producers on funds, equipment, product designs, techni- cal information, management skills, and subcontracting opportunities obtained from state enterprises (Jefferson and Rawski 1994). In southern Jiangsu province (near Shanghai), a center of booming rural enterprise development, "more than two- thirds of township and village enterprises...have established various forms of eco- nomic and technical cooperation arrangements with industrial enterprises, research units, and higher educational institutions in larger cities" (Xu, Mao, and Yuan 1993). Officials attempting to develop industry in poor localities are encouraged to pursue "joint operations with scientific research organizations or large- and medium-scale enterprises" (Du, Huang, and Chen 1992). The domestic quality ladder also resembles its international counterpart in that cost pressures move in the opposite direction from technological capabilities. The manager of a TVE garment firm that had begun with an infusion of cash and used equipment from a much larger state enterprise observed that 'in the area of prod- uct quality, household producers can't match collectives and collectives can't match state enterprises; but as for costs, state enterprises can't match collectives and the collectives can't match individual households" (interview, June 1993). Wage costs are highest in foreign-invested firms and lowest in TVEs (table 2). The cost advan- tage of collectives over state firms is a staple topic of discussion in Chinese newspa- pers and economic journals, which emphasize the extra burden of pensions, taxes, redundant workers, fringe benefits, and welfare responsibilities assigned to state firms, especially compared with TVEs. The extra cost burdens are large and, in some cases, growing rapidly. For example, state enterprises are obliged to pay retirement benefits out of current income. National data (probably excluding the farm popu- lace) show that the ratio of retirees to active workers increased from 1:26 in 1978 to 1:6 in 1990 (Du and Shang 1993). A 1989 sample study showed that enterprises 138 How Industnal Reforn Worked m China Table 2. Average Wages for Different Classes of Enterprise in China's Textile Industry, 1989-91 (yuan per person-year) Ownerip type 1989 1990 1991 Joint venture 3,663 4,232 5,674 State sector 2,069 2,252 2,377 Urban collective 1,368 1,688 1,862 Township-village 1,132- - - - No avail2ble. a. Average wage for a! TVE industries. Sowca Jefferso, lwski, and Zheng 1994. paid about 232 yuan per worker for medical costs (Du and Shang 1993); by 1993 a survey of 100 units in Hebei province found average annual medical costs of 1,201 yuan-equivalent to 37 percent of money wages (Li and Qin 1993)." Distinct Technical and Institutional Frontiers Firms occupying different rungs of the quality ladder face different institutional regimes, which implies that domestic competition may take the form of efforts to remake institutions as well as products. Economists think of innovations as changes in technology (a new product) or managerial systems (lust-in-time inventory con- trols) that expand the prcduction frontiers for individual firms or whole industries, while the firm's objectives, behavior, organization, and surrounding institutional environment remain unchanged. Since the task of economic reform consists pre- cisely of altering these basic circumstnces, the assumption of institutional stability within and outside the firm is not tenable in transition economies. Firms in transition economies face separate technical and institutional frontiers: the technical frontier embraces the standard idea (in its neoclassical or evolutionary form) that firms can draw on alternative blueprints or techniques to transform resources (induding knowledge and experience) into products. The institutional frontier delineates the set of resource configurations that is attainable under pre- vailing custom, law, and regulation. In the context of socialist systems and transition economies, institutional restrictions prevent the exploitation of many options that are technically feasible and block choices that would be made in the absence of insti- tutional change.12 Chinese authors routinely comment on the long-standing practice of applying separate laws and regulations to firms operating under different ownership arrange- ments. We have already mentioned differences in taxation and labor costs. Regulatory regimes affecting labor unions, environmental hazards, workplace safety, and the like are applied more vigorously to joint ventures and state firms than to urban or rural collectives. TVEs benefit from short lines of command. Business deci- sions are often reached through a single telephone conversation or a meeting of two or three people. State firms, by contrast, often report to multiple supervisory agen- Jeffecn and Rawski 139 cies whose overlapping jurisdictions and competing agendas complicate even rou- tine business decisions (Byrd 1992). Shorter lin:s of command make it easier for TVEs to reach decisions and form business coalitions with domestic or foreign part- ners. TVEs and joint ventures can dismiss workers more easily than can state enter- prises. Jefferson, Lu, and Zhao (1994), using a composite indicator of management authority, find a high degree of decisionmaking autonomy at the enterprise level for nearly two-thirds of 300 TVEs surveyed in 1991 but for less than half of 900 state enterprises. Bankruptcy, long a reality for TVEs, is only now emerging as a possibil- ity for state firms. The Dynamics of Partial Reform in China's Industry We see the dynamics of partial reform in China's industry as a succession of responses to imbalance by both enterprises and governments. Our framework is in the tradition of Hirschman's (1958) analysis of unbalanced growth. The dynamic that transforms partial reform into improved performance is simple and direct: • The government implements partial reform measures that reduce entry bar- riers and lower the cost of many types of transactions. These initiatives have different impacts on the options available to various groups of firms. Partial reform accelerates the domestic product cycle by facilitating the transmission of cost pressures and technologies up and down the hierarchy of industrial enterprises. * The differential impact of reform efforts destabilizes the existing division of industrial resources and product markets among different types of firms. Competition in industrial product markets intensifies. • Stronger competition diminishes the flow of quasi-rents derived from the enforcement of entry barriers and market segnentation. Reduced profitabil- ity inits the growth of wages and bonuses for some firms and throws others into a position of financial loss. The erosion of profits also limits the growth of revenues accruing to local and provincial authorities and to the central govermnent. * Firms react to financial pressures by choosing strategies involving one or more of the following components: restructuring operations, lobbying for further deregulation to facilitate profit-seeking initiatives, and lobbying for government subsidies or official intervention to restore the initial financial position. * Governments also react to financial pressures that reduce their share of total output and destabilize the distribution of fiscal revenue across regions and administrative levels. Officials face conflicting enterprise lobbying efforts, some demanding furither autonomy and deregulation, others seeking protec- tion from the effects of earlier reforms. * These induced responses of firms and governments further erode entry bar- riers and reduce transaction costs. Beneficial feedback effects accelerate every dimension of the reform process by intensifying competition, further dimin- 140 How Indusral Refonn Wb,ked hi China ishing qunsi-rents, and motivating enterprises and governments to undertake new reform efforts. These changes set in motion further rounds of technical development, economizing efforts, and incremental reform. This entire process affccts the attitudes of enterprise personnel and govern- ment officials to the direction and outcome of reform. Changing attitudes affect the objectives and strategies of all participants. Although this mechanism focuses attention on endogenous or bottom-up aspects of the reform process, not every industrial policy initiative undertaken since the beginning of reform represents an endogenous response to the initial partial reform effort. Many policy changes, such as the partial commerciali-r.tion of bank lending and the reduction of budgetary appropriations for industrial research and develop- ment projects, probably represent a combination of endogenous response and exogenous initiative. Furthermore, there is no guarantee that partial reform will succeed. If govern- ments act to stifle competition, equalize financial outcomes for winners and losers, or alter regulations to restore the prereform status quo in product markets, the endogenous process linking initial reforms with innovation, economizing, and fur- ther institutional change may stall. We see this cumulative process of endogenous response as the key to explaining how China recorded unexpectedly strong achievements in both growth and institu- tional change despite modest reform initiatives on the part of its central govern- ment Our discussion focuses on seven propositions about Chinese industrial reform corresponding to each element of the model. Proposition 1: Partial Reform Erodes Market Segmentation, Thereby Lowering Barriers to Technology and Resource Flows China's reforms of the late 1970s restored household agriculture and reopened rural markets; expanded China's participation in global markets for commodities, capital, and technology, and began to loosen controls in the industrial sector. Each of these policy shifts eroded barriers to competition in industrial factor and product markets. Agricultural reforms provided a big boost to China's rural industries by increasing the supply of labor and raw materials and, following the quick rise of farm incomes, boosting demand for output. As Sachs and Woo (1994) point out, China's compara- tively large rural sector creates possibilities for rapid productivity growth that are not accessible to more urbanized states like the Czech Republic, Poland, and Russia. The open door policy brought a rapid inc-ease in imports of industrial goods, many of them competing directly with domestic products. Partial liberalization of the external sector sharply reduced the transaction costs associated with interna- tional inflows of capital, technology, market information, managerial skills, and equipment. New policies speeded the transfer of capabilities and cost pressures across China's borders and throughout the hierarchy of domestic enterprises. The pace of change was particularly rapid in regions of southern China that benefited from proximity to Hong Kong as well as from accelerated deregulation. * Jeerson and RawS&i 141 The inital industrial reforms were directed mainly at large-scale, urban-based state enterprises. The objective was to enliven large-scale industry by encouraging firms to shed the passive mentality of plan followers in favor of self-motivated efforts to take full advantage of available resources. Large firms were allowed to retain a portion of their profits. And because discretionary funds cannot stimulate production or innova- tion unless they can be used to acquire productive inputs, the reforms included mea- sures that diverted industrial resources from planned allocation into market channels. Although intended to stimulate state enterprises, these measures had their largest impact in the TVE sector. Urban-oriented reform enabled TVEs to obtain inputs for- merly reserved for clients of the plan system and to penetrate markets outside their home areas. The relaxation of restrictions on information-sharing, consuling, and technical ties across urban-rural administrative boundaries made it easier for collec- tives and TVEs to adopt new technologies and to produce substitutes that could compete with state enterprise products. These changes shattered constraints that had previously restricted TVE growth. The result was an unexpected growth explo- sion in China's rural industries. These partial and uneven reforms substantially eroded barriers that had long obstructed flows of resources and products across the boundaries separating differ- ent types of firms and different administrative and bureaucratic jurisdictions. As old dhistnctions gradually blurred, resources, products, funds, and information began to circulate in new directions. The new market channels were soon large enough to leave quantitative traces in the form of shrinldng divergences in factor returns among different enterprise groupr f-ble 3). The bright growth prosects and high rates of return enjoyed by TVEs at the start of reform attracted a large inflow of funds, which pushed down returns to TVE capital Naughton (1992) documents a Tablc 3. Proft/Capital Ratios in D#ffernt Segmens of Chinese Industry, Seleted Years, 1980-92 (percent) Befow taxes After ta State Urban State Urban Year enterprises coleives 7VEs enterpries collcatives Tlll 1980 24.8 26.6 32.5 16.0 18.5 26.7 1982 23.4 22.0 28.0 14.4 13.8 20.2 1984 24.2 22.3 24.6 14.9 13.9 15.2 1985 23.8 243 23.7 13.2 15.3 14.5 1988 20.6 19.7 17.9 10.4 11.3 9.3 1990 12.4 - 13.0 3.2 - 5.9 1992 9.7 - 14.2 2.7 - 7.2 - Noc avaiable Note: Rare of murn is the atio of ie nun of profit figures (positve or negtive) for 2i firms to tie sum of net (of depreiation) vaiue of fixed awrs plus average amount of woddng citd in use. Source Stare entcrprises, Yearbook 1991, p. 416; Ycarbook 1992, p. 437. Urban colectives, Industry 1954, p. 85; Industry 1986, p. 87; and Str Stdsicd Bureau. TVEsYcarbook 1991, pp. 377-79; Ykarbook 1993, pp. 396-97. 142 How Idsbial Reform Worked i China related phenomenon: the convergence of rates of return to capital across different branches of industry. These changes occurred prior to the creation of organized cap- ital markets, which remain embryonic even today. Access to high-level technical personnel offers another example of declining mar- ket segmentation. Comparing data from China's 1985 industi census with 1989 survey results reveals a big increase in the availability of engineers and technicians outside the state sector. Using the 1989 survey data, Jefferson, Rawski, and Zheng (1992c) find a surprisingly dose correspondence between the marginal profitability of upper-level technicians within state enterprises, TVEs, and urban collective firms in three branches of industry. Proposition 2: Reform Intensfi0es Competihton in Markets for Industril Products On the eve of reform China's industry was in a position of incipient competition, with large numbers of potential entrants poised to intesify product-market competition. Partial reform rapidly turned tis potential into reality. Booming imports of manufac- ures, swift expansion of joint ventures and other foreign-linked enterprises, and rapid erosion of the economic and administrative barriers preventing state enterprises from raiding each other's customers all contributed to the upsurge of competition. Ihe great- est unpetus to competiton came from the growth of TVE production, which leaped frorn 10 to 2S percent of total industrial output between 1980 and 1990 (see table 1). Competition expanded most rapidly in markets directly affected by the growth of TVE output, but competitive pressures extended to other markets as welL By the late 1980s more than half of industrial products were being sold through markets, and that share has since risen to more than 80 percent. Less than 10 percent of industrial out- put is decided through mandatory plans. Concentration ratios are low and decining (Jefferson and Rawski 1994). 'With competition from manufaured imports on the rise and barriers to domestic trade increasingly porous, it is dear that partial reform has firmly insaUlled rivalrous product markets as a regular feature of everyday opera- tions for most of China's indusrial enteprises. Few firms remain immune from com- petition. The experience of the Luoyang Tractor Works, China's largest manufacturer of wheeled tractors and bulldozers, is illustrative. According to an article in Cina Daily, Luoyang "is trying to improve the quality of its products as well as its market- ing and publicity techniques in a bid to offset..sluggish domestic sales.... The Luoyang tractor complex had been forced to sacrifice more than half of its profits in trying discounts, lotteries, and free delivery of goods to boost sales (Gao 1990). Proposition 3: Competition Erodes Profits and Curtils the Growth of Fiscal Reveues Reform has brought a large decline in industrial profits. Rates of return for state enterprises and TVEs in 1990-92 are less than half what they were in 1980-82 (see table 3). Rates of return (induding taxes) by industry confirm the impression of lemffesn and Ratski 143 declining profits (table 4). Growing opportunities for tax evasion have widened the error margins for reported profit totals. Wang (199?.) cites studies suggesting that the hidden profits of state industry in 1990 may have surpassed the amounts reported to the statistical authorities. Some observers (Sicular 1992) argue that the decline in profitability is mild, confined mostly to state enterprises, and may reflect cyclical factors (the retrenchment of 1988-90). Despite these qualifications, the data strongly support the hypothesis of secular decline (see tables 3 and 4), especially since the boom years of 1992 and 1993 brought no revival in rates of return. Rates of return shown in table 4 also reflect the powerful impact of TVE compe- tition on industrial profitability. In the early years of reform, profits fell fastest in branches with the greatest TVE activity. During 1980-85 falling profitability was concentrated in industries using agricultural raw materials especially beverages, tobacco, textiles, and apparel. During the second half of the 1980s the downward pressure on profitability extended to branches with little direct TVE competition, including power, chemicals, iron and steel, machinery, and electronics, but prof- itability eroded even more in branches with strong TVE activity. The 1991 rates of return for branches with extensive TVE paricapation are less than half the 1980 base- none of the branches with limited TVE particpation experienced as steep a drop in Table 4. Rates of Retum to Capital in Chin=s Industry, 1980-93 (ptrcen) Seato 1980 198S 1991 1992 1993 Indust 25.2 23.8 11.9 10.1 10.6 lgt industry 49.1 31.8 15.0 Farm materials 54.7 32.0 15.9 Nonfarm maerials 39.1 31.5 13.0 Hcavy industry 18.5 20.3 10.2 Sectors with scite TVE womp o Food processing 20.4 175 9.7 Beverage 485 28.4 18.6 Tobacco 326.9 207.4 113.6 Textilc 69.0 26.5 6.6 Apparcl 46.0 26.0 11.7 Leather, hides 30.3 19.4 5.6 Handicrafts 43.0 27.4 12.0 Plastics 31.6 21.2 9.1 Sectors with imited lVE comspetitson Power 20.6 16.0 13.0 Chemicals 22.0 21.7 14.0 Frrous metllurgy 18.3 2S.8 15S Machinery 13.0 19.3 739 Electronic 14.0 24.9 83 Note Rate of return equals txes phs profits as a percetage of the net (of depreciation) value of fixed ase plus workingcaiL owre: Da for 1992 and 1993, Communique 1994. Other data, Industy 1992, pp. 168-73. 144 How IndatiedRefwm Wiked mi China profitability. Singh, Ratha, and Xiao (1993), using provincial data for 198489, also show that faster growth of nonstate industry (collective, foreign-invested, and private firms) is associated with lower profit rates for state industry. Profit erosion also affected government revenue, which has declined sharply as a share of total output from about 30 percent in the early 1980s (as reform began) to 20 percent at the end of the decade and to 14 percent in 1992 (Wong, Heady, and Woo 1993). Slow growth of revenues from industry, the chief source of government income, was the principal cause. Proposition 4: Enterprises React to Market Pressure by Searching for Fincial Gain Enterprises facing competition and declining profit margins have several options for strengthening their financial position. They can improve their performance within existing insttutional limits, pressure the govremment to extend greater autonomy and incentives to the firm, or pursue rent-seeking alternatives (suidies, soft loans to offset losses, lobbying for restrictions to stifle competition). We examine the firm's opportunities in each of these directions. Improve peformance wihin existing instiutinl lits Chinese state enter- prises, particularly firms facing fierce competiion and dedining profits, have demonstrated a substantial capadty to economize and innovate. As stare-owned enterprises came under increasing competitive pressures during the 1980s, total fac- tor productivity improved steadily (Wu 1993). Singh, Ratha, and Xiao (1994) estab- lish an explicit link between competitve pressure and productivity growth by showing that total fictor productivity in state mdustry rose most rapidly in provinces with the largest shares of nonstate production in total industial output. We tested this association between competition and state enterprise efficiency using the following regression equation with 1990 enterprise data: (1) ln(QjL) = -1.25 + 0.63 In(KIL) + 0.09 COMP + 0.65 PCOMP - 0.11 InU') (3.04) (13.94) (2.55) (4.72) (2.54) R2 = 0.32, obs. = 496, where QIL is labor productivity, KIL is the capital-labor ratio, and NKIK is the share of nonindusial capital held by the enterprise.13 COMP and PCOMP are measures of competition-COMP is an estimate of the elasticity of demand for the firm's major product and PCOMP reflects the firm's asessment of the overall competitive pressure it faces.4 Using either panels of provincial data or cross-sectional enter- prise data, we consistently find that competitive pressures motivate firms to improve overall efficiency. The partial reforms of the 1980s also brought a distinct acceleration of innova- tive activity in state enterprises. One survey of eighty stare enterprises found that the Jcffmson and Rawki 145 value of new products as a share of gross output rose from 13.5 percent in 1980 to 18.7 percent in 1985 and to 24.2 percent in 1989 (Jefferson, Rawski and Zheng 1992b). Similarly, in a 1992 survey of 10 percent of China's large and medium-size industrial enterprises, 91.6 percent of the 954 respondents described efforts at prod- uct and process innovation. More than half reported major modification to their main products, 807 percent had new products in the marketplace, 60.8 percent were engaged in major process innovation, and 65.5 percent were implementing new production technologies (Ma and Zhao 1993). A second set of panel data covering 249 enterprises in the textile, celctronics, and equipment industries provides information on output of new products (but not on the intensity of competition) that allows us to examine the impact of changes in profitability during 1984-86 on new product innovation during 198648. We assume a two-year lag between changes in profitability and shifts in the output share of new products. Product innovation is constrained by two factors: diminishing returns to product innovation, captured by NPS86, the initial 1986 share of new products in total output; and financil capadty, represented by PRO86, the 1986 ratio of profit (including tax) to sales. Our estimation yields the following result: (2) NPS8886 = 0.005 - 0.320 PRO8684 - 0.443 NPS86 + 0.371 PRO86 (0.617) (1.970) (7.908) (5.921) R2 = 0250, obs. = 210. Here NPS8886 is the change in the share of new products in output value during 1986-88 and PRO8684 is the change in the profit-sales ratio during 1984-86."S These esumation results confirm that declining profitability creates an incentive to innovate, conditional upon the enterprise's financial capability and the diminishing returns to innovation. Seek greater autonomy and strengthened icentives Along with efforts to improve performance within existing institutional limits, financially pressed enterprises seek greater autonomy and a larger share of residual earnings. Through the mid-1980s circumstances and opportunities were widely regarded as more favorable for collec- tive or private firns. Toward the late 1980s published materials and interviews with factory managers began to reveal a gradual shift toward the view that the autonomy associated with collective ownership had come to outweigh the privileges available within the state sector, leaving state enterprises at a competitive disadvantage State firms complain of administrative interference and cost-inflating obligations that TVE firms and joint ventures often escape. Managers in the state sector have grad- ually emerged as active agents for refornL Seek rents in the form of direct subsdes and soft loans. A final avenue of response for financially distressed enterprises is to seek rents in the form of direct subsidies, soft loans, and a competition-stfling resumption of regulation. We assume that 146 How Indurl Refom Workdi CiMna whenever such assistance is available, enterprises will pursue it as long as the expected payoff exceeds the cost of lobbying. From this perspective the attitude of governments controls the distribution of enterprise resources between economizing and innovation on the one hand and rent-seeking on the other. The following sec- tion considers whether reform has reduced the availability of various types of direct and hidden subsidies for enterprses experiencing financial distress. Proposition 5: On Balance, Government Policy Increases Industry Autonomy and Market Exposure and Hardens Budget Constraints The central facts of life for Chinese public finance in the 1980s indude a slowdown in revenue growth, a significant hardening of budget constraints for subnational governments (Walder 1994), and repeated episodes of macroeconomic instability attributable to fiscal deficits and excessive monetary expansion. Subnational governments are generally more able but less willing to subsidize weak firms and industries than their counterparts at the center. The reason Is sim- pie: fierce competition among development-conscious subnational jurisdictions- Diverting resources from development spending to subsidies threatens to undercut the ability of provinces, cities, counties, townships, and villages to attract domestic and foreign investment. With local revenues increasingly tied to the growth of prof- its from local industry, slow growth of investment endangers the revenue prospects of the same bureaus and officials faced with requests for subsidies and protection. Under these circumstances, how do officials respond to the pleas of firms whose financial interests are damaged by competition? They have two main options: to grant direct or indirect subsidies or to push enterprises toward the market While subsidies continue, evidence shows that government policy has gradually tilted toward sending enterprises to market. Subsidies for TVEs are rare. Loss-making firms are closed and their workers are dismissed. At the start of reform, state enterprises typically expected full compensa- tion for losses. By 1986 the ratio of subsidies to losses for state industry had dropped to 0.8 (table S). Another sharp dropoff in compensation occurred with the retrench- ment of 1988 and 1989. Despite some confusion about the exact timing and scope, it is dear that a decade of partial reform has established a declining scale of partial compensation as the general rule for loss-making state industial firms. Direct subsidy is not the only avenue of government support for weak enter- prises. Public officials can use tax concessions, regulatory protection, and soft bank credits to sustain loss-malcing firms. Tax concessions are limited by the same con- straint as subsidy payments: the high opportunity cost of committng scarce fiscal resources. Regulatory protection runs counter to the general trend of China's domestc and international economic policy. Govermments rely on both these tools in specific instances, but large increases in tax concessions or protective trade restrictions are widely viewed as undesirable and unfeasible. In the words of Vice Minister of the State Economic and Trade Commission Chen Qingtai: 'In the past enterprises turned to the govermnent when they ran into difficulty because the gov- Jeffersn and Rmski 147 emnment could lower taxes and allowed them to retain more profits. This road has now been basically dosed' (1994, p. 48). This leaves the banking system as the primary vehide for large-scale indirect sup- port of weak firms. China's banks certainly experience strong ratfidal pressure to advance funds to weak borrowers. 'Policy loans" that are viewed as unrepayable from the start are extended to both large and small industries at the behest of pow- erful official interests. Finance officials indicate that such loans account for about 30 percent of new lending, with most of the soft credits destined for investment pro- jects (1992 interview). Policy lending is an important component of Chinese industial policy, but it is subject to restrictions. Increases in bank lending have the same inflationary poten- tial as government deficits. Furthermore, the banks, which have developed systems of credit ratings as part of their own profit-seeking agenda (Whiting 1993), have already cut the "policy" component of current lending to less than 10 percent (1993 interview). Banks can be expected to defend their business autonomy with increas- ing tenacity. Pardy for this reason, China's government has begun to implement financial -reforms that will create three layers of financial institutions: a central bank, poliqc banks to support official priorities, and profit-oriented commercial banks. These reforms will shift the locus of conflict without resolving the probIems facing loss-makers. The vice governor of the central bank, reaffirming his determination to 'exercise stringent control over the money supply," insists that the new policy-lend- ing banks "must be careful not to run in the red." Yao Zhenyan, president of one of the new policy linstuitions, makes the same point, emphasizing "the importance of investment effciency" and insisting that "we must ensure the return of princpal, although we are not aiming at profits" (Policy Bank 1994). With highly placed bank officials attacking soft credits even at institutions designed to serve this very need, and with ordinary banks eager to "firther connercialize their business" (Wu 1994), Table S. Loses and Subsidies for C'ise State Enterprises, 1986-91 (biions of yuan) AU publc entprise State-owned idustries Year Loss Subsy Ratio' Loss Sibsid9 Ratio 1986 41.71 32.S 0.78 4.71 0.81 1987 48.17 37.5 0.78 5.07 0.72 1988 52.06 44.6 0.86 7.13 0.73 1989 7436 S9.9 0.80 12.80 9.50 0.74 [53] 1990 93.26 573 0.62 27.88 11.80 0.42 [.34] 1991 93.11 50.6 054 30.02 1450 0.48 [.40] a. Subsidy/lass. b. Infornidon provided by the World BanR Source AUl publi cntrprises: losses, Hm 1992; subsidies, World Bank 1992, p. 242- Smte indusaies Wang. Heady, and Woo 1993; World Bank estimate 148 How IndustridRfoim Wo,kd in China executives and workers of loss-making firms can expect only limited relief from their financial predicament unless they improve their performance in the marketplace. The growing reluctance of government to support weak enterprises is reflected in microeconomic data. Survey data analyzed by Morris and Liu (1993) show that, despite an increase in the absolute level of subsidies, there was considerable hard- ening of budget constraints for state firms during the late 1980s. Other data support the hypothesis that enterprises fadng strong competiton move (or are pushed) toward the market in the expectation that greater independence will help them resolve their financial problems. The following regression, based on 1990 data for state enterprises surveyed in late 1991 and early 1992, evaluates the impact of market conditions on the govern- ment's grant of decisionmaking autonomy (DMA) to enterprises: (3) DMA = 1.73 + 0.12 COMP + 0.10 PCOMP -0.27 PROFIT + 0.02 IND (19.40) (5.02) (3.69) (2.17) (0.46) R2 = 0.10, obs. = 572. In this regression DMA is a composite measure of enterprise control over production and marketing decsions.'6 COMP and PCOMP are the same measures used in equa- tion 1. PROFIT is the ratio of profit (or loss) to the gross value of output in 1990. To control for differences in autonomy that are specific to the light-heavy industry mix, the equation includes IND, a dummy variable in which 0 represents heavy industry and 1 represents light industry. The regression results show that competition (measured by the firn's estimate of the elasticity of demand for its products and the degree of competitive pressure from rivals) is associated with a relatively high degree of managerial autonomy, which we interpret as synonymous with greater exposure to market forces. Low or negative profits also contribute to a greater transfer of pro- duction control and marketing rights to enterprise management. Recent developments in the woolen texdle industry illustrate the government's propensity to assist troubled firms with offers of deregulaion rather than direct or indirect subsidies. During 1990-91 the requirement that woolen textile exporters sell through foreign trade corporations insulated them from international market changes and led to large inventories and losses. In response to pressures from pro- ducers the government allowed woolen textile companies to export direcdy to over- seas customers. Chinese firms soon began to produce semifinished inventories that could be more quicldy transformed into final goods that conformed to the specifi- cations and just-in-time production requirements of overseas customers. Although government intervention continues to cushion some firms, especially state enterprises, against the consequences of weak performance, limited resources, fear of inflation, and dcanging attitudes have increased the iikelihood that firms and their workers will bear the financial consequences of market outcomes. There is a growing gap between financial outcomes for successful and unsuccessful firms. Loss-nmaking industrial enterprises, formerly eligible for full compensation as part of Jefferson and Rawski 149 official administrative routine, face growing difficulties under China's steadily deep- ening reforms. In the rural sector, losses bring a quick exit for enterprises and dis- missal for workers. Subsidies continue, but even for urban state enterprises, subsidies have dropped from fill coverage of losses to well below half Workers associated with loss-making enterprises face a growing probability of sanctions such as slow wage growth, deterioration of boriuses, erosion of health benefits and other nonwage income, layoffs with only partial wage paymenr, delayed wage payments, compulsory transfers and, most recently, dismissal. Proposition 6: Feedback Mecbanisms Ampliy and Extend the Reform Process The consequences of reform are not limited to a linear progression in which new policies intensify competition, reduce profits and fiscal revenues, and create pres- sures for better industrial performance. At every stage we observe feedback med1a- nisms that reinforce the momentum of beneficial change. The success of some enterprises in reducing costs or developing new products reverberat;:s up and down the domestic quality ladder, escalating the pressure on rival enterprises to follow suit. Every reform that relaxes institutional constraints on market entry, enterprise autonomy, or technological change shortens the distance separating adjacent rungs along the ladders of technology and cost, increases the probability of competi- tion-enhancing innovation, and raises the risks facing enterprises that are slow to reform. Reductions in fiscal resources caused by flling profits or tax evasion (itself an outcome of reform-induced expansion of enterprise autonomy and financial mechanisms) increase pressures on enterprises by reducing the chances of successhfl rent-seeling, further widening the gap between "winners" and "los." Groves and others (1994) provide a quantitative illustration of feedback mecha- nisms that shows how state enterprises use grants of autonomy to strengthen work incentives and raise productivity. Their analysis of sample data indicates that enter- prise autonomy is associated with large shares of discretionary payments in worker compensation and with high shares of untenured contract workers in the labor force. Their statistical analysis confirms the expected positive link between these incentiLve changes and productivity growth. Thus incremental grants of enterprise autonomy appear to feed back into faster productivity growth, which in turn inten- sifies competition, and so on. Proposition 7: China's Decision to Create a Market-Based Economic System Is an Endogenous Outcome of the Partial Reform Process Chinams initial reform efforts sought to improve economic performance; there was no dear picture of what the economy should look like after reform. Partial reform initiated a learning process that expanded the horizons of all participants. Competition among firms organized under heterogeneous institutional arrange- ments opened the door to a dynamic and interactive reform process in which spe- cific policy initiatives have different effects on the opportunity sets of firms facing 150 How Indtal lRf