Sv/P7 a7 Indonesia The Challenge of Urbanization Andrew M. Hamer Andrew D. Steer David G. Williams WORLD BANK STAFF WORKING PAPERS Number 787 WORLD BANK STAFF WORKING PAPERS Number 787 Indonesia The Challenge of Urbanization Andrew M. Hamer Andrew D. Steer David G. Williams The World Bank Washington, D.C., U.S.A. Copyright (© 1986 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 Manufactured in the United States of America First printing January 1986 Second printing April 1988 This is a working document published informally by the World Bank. To present the results of research with the least possible delay, the typescript has not been prepared in accordance with the procedures appropriate to formal printed texts, and the World Bank accepts no responsibility for errors. The publication is supplied at a token charge to defray part of the cost of manufacture and distribution. The World Bank does not accept responsibility for the views expressed herein, which are those of the authors and should not be attributed to the World Bank or to its affiliated organizations. The findings, interpretations, and conclusions are the results of research supported by the Bank; they do not necessarily represent official policy of the Bank. The designations employed, the presentation of material, and any maps used in this document are solely for the convenience of the reader and do not imply the expression of any opinion whatsoever on the part of the World Bank or its affiliates concerning the legal status of any country, territory, city, area, or of its authorities, or concerning the delimitation of its boundaries, or national affiliation. The most recent World Bank publications are described in the annual spring and fall lists; the continuing research program is described in the annual Abstracts of Current Studies. The latest edition of each is available free of charge from the Publications Sales Unit, Department T, The World Bank, 1818 H Street, N.W., Washington, D.C. 20433, U.S.A., or from the European Office of the Bank, 66 avenue d'I6na, 75116 Paris, France. Andrew M. Hamer is a senior economist and David G. Williams a senior urban planner in the Operations Support and Research Unit of the World Bank's Water Supply and Urban Development Department; Andrew D. Steer is a senior economist in South East Asia Division of the East Asia and Pacific Regional Office. Library of Congress Cataloging-in-Publication Data Hamer, Andrew Marshall. Indonesia, the challenge of urbanization. (World Bank staff working papers ; no. 787) 1. Cities and towns--Indonesia--Growth. 2. Rural- Urban migration--Indonesia. 3. Urban policy--Indonesia. 4. Urban economics. I. Steer, Andrew. II. Williams, David G. (David Glyn), 1938- , III. Title. IV. Series. HT371.H27 1986 307.7'6'09598 85-29537 ISBN 0-8213-0688-X - iii - Abstract Indonesia's urban population is growing rapidly. Almost 40 million residents will be added to its cities in the last two decades of the century. This growth is the outcome of millions of individual household and business decisions. It is best managed inclirectly, through improved national sectoral and urban management policies. Suggested policy modifications are included in the text. Direct interventions to contain city size or redirect migration are usually inefficient and counterproductive; most such initiatives should thus be avoided. - iv - Preface Over the last two years, a number of reports have been produced dealing with regional and urban issues in Indonesia. This work, in turn, has been encouraged and closely followed by the Government, which is in the process of reviewing certain policies and institutions dealing with these matters. Because other countries are facing similar questions, it seemed appropriate to repackage and, in part, to expand upon the findings of these reports, making them available to a wider audience. The Key reports on which this study is based include: Indonesia: Policies and Prospects for Economic Growth and Transformation, Report No. 5066-IND, East Asia and Pacific Regional Office, The World Bank, April 26, 1984. Indonesia: Urban Services Sector Report, Report No. 4800-IND, East Asia and Pacific Regional Office, The World Bank, June 25, 1984. Indonesia: Selected Aspects of Spatial Development, Report No. 4776- IND, East Asia and Pacific Regional Office, November, 1984. Indonesia: Policies for Growth and Employment, Report No. 5597-IND, East Asia and Pacific Regional Office, The World Bank, April 23, 1985. Please note that the documents cited as "reports" in this paper, such as the foregoing, are for distribution within the World Bank only. v Table of Contents Page No. Preface I. Summary of Conclusions and Recommendations ........................ 1 General observations...... ........o........... 1 Industrial policy reforms .... .o...... . ..... . ...... .... . .2 Reforming urban management. .. . . . ... .. ....... .......o. .4 II. Urban Spatial Policies and {institutions: An Introduction ...... , *7 III. Human Settlement Patterns in Indonesia....... ...-................ 11 A. Urban Growth: Past Trends and Future Prospects...............11 Recent pattern of urban growth..... ....... o.....................11 The challenge of future urbanization....o.o.o................ 12 B. Rural-Urban Migration and Urban Population Growth............. 17 Data limitations ....... ......................... .... 0.0..... .... 17 Permanent migration o..o...o.o.o.o..o...o ... o......................... o....19 Unrecorded migration..........................................20 Urban quality of life and rural-urban migration...............21 C. Urbanization and the Growth of Large Cities ...................29 Difficulties in defining urban growth by city size ....o .... 29 Trends in urban growth bly city size...........................29 IV. Influencing Urban Development ............o..*...*oo .............. 33 A. Some Basic Principles.................. ...................... 33 Determinants of city size in the course of development........ 33 The impact of natLonal production patterns..oo.........,....34 The relative attractiveness of cities and city size ...........37 Limits to public intervention in shaping city size............38 B. Influencing Industrial Location ................... ......... . 41 Why focus on industrial location? .... . . .*****4l (Continued) - vi - (Table of Contents - Cont.) Page No. The spatial bias of sectoral policies ......................... 43 The trade regime and city size effects ................ 44 Impact of transport sector inefficiencies on city size..*..... 46 The locational impact of pervasive micro regulation........... 48 Industrial estates as tools for regional dispersal ......o..... 54 The role of public industrial investment in regional development ....................... .................. oo.....*.ee..... *57 Regional development and the rationing of industrial finance ..... o.--.... oo.... .o..o..................... ..o.o.o......59 C. Management and Finance of Urban Services ...........-........ 62 The financial implications of REPELITA IV urban goals......... 62 Challenges facing public administration..o..o.. o....ooo...... 67 Alternative sources to finance urban finance ........l..........71 Local borrowing as a tool....o.-oo ..... o-o-oo-ooe ........ 77 Financing urban housing services. ...o.......o....o*.oo..o.... 84 I. Summary of Conclusions And Recomendations General observations 1. Though the urban share of Indonesia's population is still relatively low, its absolute size and annual growth is significant enough to warrant serious concern among policy-makers. During the final two decades of the century, almost 40 million residents are expected to be added to Indonesia's cities. 2. Urban growth is the outccome of millions of individual household and business decisions concerning lociation for new or expanded activity. This process, which includes rural-urban and interprovincial migration, helps to redirect population and employment growth toward areas of emerging economic opportunity. Such a redistribution of growth is a common feature in the history of all nations and one whLch, in the main, yields the most efficient and equitable outcomes when managed indirectly, through improved urban management and national sectoral policies. Direct interventions to force growth to follow different paths is usually costly and counterproductive, and forces already over-extended public administrators into areas where they are at a comparative disadvantage. 3. What is true about efforts to stem rural-urban migration in general is equally true of any attempts the shape the city size distribution of the urban population by crude administrative measures. There is a tendency for the share of urban population of different city size categories to become more evenly distributed in the course of economic development. Unfortunately this results from a complex set of factors; most of the features that determine the changing relative attractiveness of cities are exogeMus in the short-run. -2 - Policy levers work best when they are applied indirectly, and when results are allowed to emerge over an ample period of time. 4. A set of policy recommendations are found throughout this paper, and are summarized below. Their net effect will be to modify the rate at which different size cities will grow. Unfortunately, little can be said about the precise way in which these changes will take place. There is no available theoretical framework of city size distribution, no model that one could estimate with data and use to predict reforms in relative city sizes in the presence of policy changes, and then compared with actual Indonesian data. Nevertheless, what is clear is that even marginal declines in the growth of larger cities will result in the reallocation of millions of urban dwellers over time. The lack of precise estimates of policy impacts is thus no grounds for inaction or for the use of controls whose record is generally poor. Industrial policy reforms 1. Because of its growing importance, its inherent locational flexibility, and its multiplier effects, the industrial sector is a natural focus for policy reforms aimed at managing urban growth. Certain measures are to be avoided in this process. Use of business location licenses or of residence permits should be set aside. Public administrators rarely have the foresight to manage urban growth so directly. This is especially true in countries like Indonesia, where the central bureaucracy is already over- extended, given its size and its quality. Instead, industrial policy initiatives should concentrate on five areas. First. large dividends can be reaped from improving transportation and telecommunications access to domestic and international markets. This is important for a country like Indonesia, where geographic fragmentation has resulted in an economic archipelago. with fewer links between provinces and islands than required to help fulfill the -3- goals of efficient and equitable development. It is crucial that the planned program of improving inter-island shipping and port performance be carried out vigorously, along with an extension and upgrading of interregional road links. Until access is significantly improved, physical proximity to the main consumer markets, especially in and around Jakarta, will dictate the location of all producers except those thaLt are resource-bound or those too small to cater to more than a very localized market area. 2. Second. improved physical access must be accompanied by measures to reduce the incentives to operate businesses near the centers of government power. In the past, day-to-day access to customs officials, tax officials, regulators, and sectoral policy-makers. especially those in Jakarta, was necessary. This type of incentive is counterproductive to spatial goals of decentralized development, and has no other redeeming value. It should be neutralized through the continued simplification of rules and regulations. 3. Third, the explicit price signals governing location behavior require modification, especially those associated with trade regulation. The existing trade regime encourages the premature growth of large cities by promoting the growth of inefficient import-competing activities which are dependent on the agglomeration economies found in economically pre-eminent centers. Once again, such sweeping protectionist incentives are not justified on national development grounds, and they retard decentralized growth as well. Reforms are needed if efficient and equiLtable growth is to occur, nationally and across regions. 4. Fourth, the government should be encouraged to develop and expand the use of industrial zones and estates targeted selectively in areas of proven growth or well-documented potential. At this stage of development, the areas where industrial growth and diversification will tal place are clear. It is in these areas where improved infrastructure is a necessary, and almost sufficient, condition for continued growth. Elsewhere, on the other hand, the obstacles to industrial growth are so many that radical improvements in public services will not be cost-effective. Such areas should share in the gradual upgrading of infrastructure that can be expected as economic development constraints give way over time. 5. Fifth, during this period, when the rules that govern the behavior of financial institutions is changing, considerable attention is required to improve access to investment and working capital for small and medium-size firms. Particularly important is the consideration of proposals that would promote long-term lending by the various regional development banks. For the foreseeable future, other institutions do not have a sufficient incentive to provide long-term funds for smaller firms. The public sector, using own and foreign resources, will be forced to demonstrate the profitability of such lending activity over the next decade. Reforming urban management- 1. Improved access to public services in urban areas plays a crucial role in economic development by modifying the relative attractiveness of particular centers over time. This is true both directly, by reducing the operating costs incurred by business, and indirectly, by helping to attractive skilled workers and professionals to such centers. At the same time, the central government cannot continue to operate in this area as it has in the past. While cities require improvements, the rural areas of the country, where most of the population lives, are seriously deficient in services they have a right to expect on equity grounds alone. This enormous task confronts a central government that faces serious financial constraints and that cannot -5 - effectively make all the many decisions that each city investment strategy requires. 2. The solution most likely to achieve this goal of improved services for the cities is one thal: incretases local autonomy and financial responsi- bility. The government should give serious consideration to granting kotamadya status to all urban areas with a threshold population of 20,000- 50,000 and above. At that point :Ln the development of a city, the demands for local coordination of investments, maintenance, and cost-recovery are generally beyond the means of central, provincial or even district authorities who have responsibilities for large rural populations as well. With increased autonomy there should be consLderable attention paid to upgrading the technical and planned skills for local government workers. Very useful in this regard would be the introduction of multi-year, rolling, consolidated development budgets which would bring together financing and physical planning information for all projects. 3. In an age of central government austerity, plans should be put into place for graduating urban areas from an intergovernmental fiscal regime based heavily on straight central grants to one based on a combination matching grants, loans, and increased local tax effort and user charges. Such a system is well-established in many developing and most developed countries; it appears practical for Indonesia. Once in place, it will allow improvements to be made in urban areas while rural demands for basic needs are addressed. It will permit improvements to be made in large cities, like Jakarta, without subsidizing its residents well-being. It will make possible the extension of services necessary for decentralized economic development across a large number of cities without burying the central government in debt and administrative red tape. From this will emerge a stronger administration at all levels, and the central government will finally be able to fulfill its mission to plan selectively and strategically for the enormous changes that face the nation over the following decades. 4. An integral part of any attempt to better provide for the needs of the cities is a shelter plan which mobilizes the financial resources and the potential sweat equity available in the urban areas. In the past, the government has channeled public resources largely mobilized elsewhere to finance new housing at high levels of subsidization. Though the national public housing authority has developed an innovative strategy to reach households at the lower half of the urban income distribution, most of the benefits of the housing programs have accrued to the upper two deciles. 5. Reforms are now necessary. Housing finance programs must be expanded, as the goals proposed for REPELITA IV have emphasized. Beyond that there is a need to develop pioneering efforts to help the poor half of the households through an imaginative mix of sites and services programs, core unit construction, housing upgrading support and guided land development. The macroeconomic realities inherent in the present climate of austerity, combined with the need of meeting rural demands, both require financing mechanisms that mobilize private resources for housing and allocate them to all income groups. This, in turn, means that housing finance institutions must operate according to increasingly commercial principles, raising and lending funds at market rates. Equity concerns should be addressed by a limited amount of cross-subsidization with higher income borrowers paying more for loans so that low income borrowers can pay less. Beyond that there is a need to structure lending so that the profile of beneficiaries more closely resembles that of the urban households as a whole. This could be achieved by relying more heavily on sites and services programs and core housing financing; by -7- introducing home improvement programs; and by restricting the size of individual loans. In this manner, housing finance institutions can play an active role in ensuring that labor markets within and across Indonesia's cities function more smoothly by removing barriers to labor mobility. II. Urban Spatial Policies and Instruments: An Introduction 1. In Indonesia, and across the developing world, national policy-makers tend to fear the rising tide of urbanization. More specifically there is apprehension that urbanization will concentrate population in one or a very few centers, drawing resources away from other areas and thwarting attempts to achieve political integration. Furthermore, the difficulties associated with managing individual urban areas are often given as reasons for imposing some type of control on urban development. Cities growing "uncontrollably" appear to generate unmanageable problemsi whose symptoms include slums, congestion, and pollution. Given that urbanization patterns are shaped by millions of individual decisions on where to live, work, and invest, the key issue is whether this growth should be managed directly or deaTft with indirectly, by improving national sectoral and urban management policies. 2. Indonesia provides an interesting case, 1 cause it has actually attempted to influence the rate of inmigration into the capital city. An early effort was made in 1970, when the Governor of Jakarta established a system of formal urban residence permits. New migrants were to receive a temporary residence card, upon deposit with the authorities of a sum equivalent to twice their return fare. After six months, a permanent card was to be made available on proof of permanent employment. Failing that, a migrant was to be provided with funds to return to his home district. This mechanism was allowed to lapse and was not rigorously enforced. Linked to this was another tool which is still being utilized. Attempts have been made, over time, to limit new migrant employment opportunities by barring street vendors and becak (bicycle rickshaws) operators from areas of the city. 3. More systematic efforts to force economic activity out of Jakarta or other large centers and into more peripheral locations were never undertaken, except as a side-effect of industrial policies whose objective was a broader one of regulating the degree of competition in each economic subsector by creating licensing barriers-to-entry. For a growing number of these regulated activities, locations in and around Jakarta were deemed "saturated" and further new investment was barred. 4. This example is but one in which the government has pursued ostensibly non-spatial policies that can have spatial effects, whether intended or not. Many, including licensing, are inspired by a belief that Indonesia's low level of economic development, combined with a history of regional and ethnic unrest, requires a high level of central supervision over many facets of economic and political life. These policies are discussed in this paper, but it is useful to review them briefly at this point. 5. Historically, the macroeconomic environment has been characterized by a fairly high degree of regulation which has exacerbated the tendency for the production of tradeable goods and services to concentrate in areas where scarce infrastructure facilities are present, namely the largest centers (especially Jakarta). Thus at the same time attempts were being made to stem migration into Jakarta, the trade regime in place provided a heavily protectionist environment which encouraged the premature emergence of otherwise unprofitable industrial activities whose requirements at this stage of development can only be met in and around Jakarta and, to a lesser extent, -9- Surabaya. Businesses also found it necessary to have access to central bureaucrats not merely to facilitat:e the establishment or expansion of plants, but also to negotiate customized treatment regarding taxation, additional protection, processing of goods through ports, and industrial financing. 6. Another fundamental barrier to decentralized development can be found in the Indonesian system of intergovernmental fiscal relations. Ir.. countries where local governments are responsible for mobilizing a significant portion of public sector resources, one tends to observe a systematic, and coordinated, expansion of local public services with minimal reliance on central government subsidies.,1/ Localities can compete among themselves to attract entrepreneurs and skilled workers by offering different bundles of local services, local taxes, and user charges. In such settings, large centers are likely to be relatively expensive places to do business and to live, because they face diffiicult problems in dealing with congestion, pollution, waste disposal, and crime. Employers who do not require the special advantages of big cities, may find lower cost, alternative sites in smaller centers, especially as interregional communications improve. This in turn, helps redirect migration flows. 7. In Indonesia, city resiLdents have received priority access to centrally funded, and heavily subsidized, public ser-ices, while the much larger, and poorer populatiLon of the countryside, has failed to keep pace. Metropolitan centers, in turn, have benefitted most of all. Thus, at a time when there is much concern about migration into large cities, a highly centralized system of intergovernmental relations discriminates in their favor 1/ J.V. Henderson, A Framework for International Comparisons of Cities, Urban and Regional Report No. 80-3, Wiorld Bank, March 1980; B. Renaud, National Urbanization Policies in Developing Countries, New York, Oxford University Press, 1981. - 10 - by providing relatively high levels of local services without commensurate local cost recovery. In this sense, the observed lag in the improvement of local services elsewhere in the urban system, and even in the countryside, is as much due to the lack of local resource mobilization and planning as to the overall level of development of the economy. 8. One additional instance of implicit spatial policy-making is found in the planning of infrastructure investments of regional or national significance. Though one might argue about the overall level of effort devoted to port development, highways, or industrial zones and estates, the fact remains that current levels of development have forced the public sector inevitably to favor some areas over others. The best port, the best road links, and the most extensive facilities for industrial estates exist in Jakarta. Other areas might have been given greater attention under other scenarios, but few would argue against a spatial concentration of these types of investments at early stages of development. In fact, where aggressive efforts have been made in poor countries to scatter these types of investments, they have been uniformly criticized by outside observers. 1/ That having been said, it remains clear that the totality of policies that impinge on urban development in Indonesia do not provide a consistent set of signals to the millions of economic decision-makers in the country. The argument made in this paper is that a self-conscious review of the facts, and careful attention to a framework for assessing spatial development, can lead to improved policies that support the gradual and consistent increase in the number of locational options for firms and workers. In the final analysis, the spatial configuration of urban settlements is seen to depend less on 1/ A.U. Sekhar, Industrial Location Policy: The Indian Experience, World Bank Staff Working Paper No. 620, 1983. - 11 - direct interventions by social engineers working at the center and more on the indirect outcomes of a mult[plicity of policies, most of which merit adoption on non-spatial grounds alone. III. Human Settlement Patterns in Indonesia A. Urban Growth: Past Trends and Future Prospects Recent pattern of urban growth 1. Population growth has accelerated over the last two decades, as substantial reductions in mortality outpaced fertility declines. Indonesia grew at a rate of 2.3% per annum, in the 1970s, compared to earlier rates of 2.1% (1961-1971) and 1.5% (1930-1960). Added to this has been a noticeable expansion in economic activity, starting with the economic recovery of the late 1960s. These two factors haLve led to increasingly rapid urban growth. According to census data, the percentage of population living in desa I/ classified as urban has risen from roughly 15% in 1961 to 17% in 1971 and, more sharply, to 22% in 1981). Measured urban population Increased at 2.6%/annum during 1961-1971, accelerating to 4%/annum in the next decade. If attempts are made to count the rising tide of semi-permanent Javanese migrants who flow to urban areas for the better part of each year, then the 1/ In Indonesia, the differentiation between urban and rural area is made at the "village" or desa level; a desa is either urban or rural. Desa units are agglomerated into sub-districts (Kecamatan), which are, in turn, grouped into districts (Kabupat:en/Kotamadva) and then provinces. 2/ The average annual growth rate for 1971-1980 has been adjusted to take into account changes in the definition of what constitutes an urban desa. - 12 - recent rate of growth would have to be modified upward, to perhaps 4.5% per annum, or almost twice the national population growth rate. 2. Using measured urban population for comparative purposes, Indonesia is now more urbanized, on average, than all low income countries (17%), though it still lags considerably behind the rate for all middle-income countries, (45%). More important for the future, however, is the growth rate of the urban population, which already equals the total for middle income countries as a whole and exceeds rates among East Asian and Pacific countries like the Philippines (3.7%/annum) or Malaysia (3.3%/annum) (Table 1). 3. Provincial data on measured urbanization are given in Table 2. At about 25%, the proportion of total island population defined as urban is highest on Javal/. This reflects, in part, the presence of the capital city, Jakarta, which alone accounts for over six million of the 33 million urban inhabitants. Excluding Jakarta, the urban proportions do not vary greatly among the most populous islands, and range from 16% on Sulawesi to 21% on Kalimantan. The remaining Eastern islands have a lower average of 13%. Urban growth rates, as captured by the census, exceed the national average in all island groupings outside Java and range from 3.6%/annum on the Eastern islands to 5%-6%/annum on Sumatra, Kalimantan, and Sulawesi. The challenge of future urbanization 1. The difference in urban growth rates, extended into the future, implies a significant redistribution of the increasing urban population over j Since it is likely that Javanese urbanization is understated by the failure to count semi-permanent residents, all data on urban Java must be understood as providing lower bounds for actual totals. -13- Table 1 Indicators of Urbani:ation: International Conparisons Urban Population Percent of Average Annual Growth as % of Total Labor Force Rate of Population 1970-81 Pbpulation, 1981 in Industry Total Urban Indonesia 23 15 2.3 4.0 (4.5) Middle Income Cbontries Total 45 21 2.4 4.1 Asia and Pacific 32 18 2.3 3.9 Lw Income Countries 21 15 1.9 4.4 Comiparator Countries Philippines 37 17 2.7 3.7 Malaysia 30 16 2.5 3.3 Thailand 15 9 2.5 3.4 India 24 13 2.1 3.7 Pakistan 29 20 3.0 4.3 Nigeria 21 19 2.5 4.8 Egypt 44 30 2.5 2.9 a/ Includes estimate of growth oif urneasted seasonal mlgration; see para 1. Source: Wbrld Development Report 1982, 1983; Wbrld Atlas 1982; and Wbrld Hank's "Socdal Indicators". - 14 - Table 2 Total and Masured Urban Ppulation and Adiusted Growth by Province Average Amul Growth 1980 Census 1971-1980 Total Urban % Total Urban Population Population Urban Population Population (est) a Aceh 2,611 233 8.9 2.9 North Sumatra 8,361 2,127 25.4 2.6 West &mtra 3,407 433 12.7 2.2 Riau 2,169 588 27.1 3.1 Jambi 1,446 183 12.0 4.1 South Sumatra 4,630 1,267 27.4 3.3 Benr*ulu 768 72 9.4 4.4 Lanpung 4,625 577 12.5 5.8 TIUAL SIMA3RA 28.016 5.481 19.4 3.3 5.1 DKI Jakarta 6,503 6,072 93.4 3.9 3.9 West Java 27,455 5,771 21.0 2.7 Central Java 25,373 4,756 18.7 1.6 D.I. Yogyakarta 2,751 607 22.0 1.1 East Java 29,189 5,720 19.6 1.5 TOTAL JAVA 91.270 22.926 25.1 2.0 3.0 (JAVA EMG. JAKAREA) (84,767) (16,854) (19.9) (1.9) (2.6) West Kalinantan 2,486 417 16.8 2.3 Central Kalimantan 954 98 10.3 3.4 South KalinEntan 1,218 485 39.8 2.2 East Kalimantan 2,065 441 21.4 5.7 TOTAL KALIMANYN 6,723 1.441 21.4 3.0 6.4 North Sulawesi 2,115 355 16.8 2.3 Central Sulawesi 1,290 115 9.0 3.9 South Sulawesi 6,062 1,096 18.1 1.7 South East Sulawesi 942 88 9.3 3.1 TIOAL SULAkESI 10.410 1.654 15.9 2.2 5.1 Bali 2,470 363 14.7 1.7 West Nusa Tenggara 2,725 383 14.0 2.4 East Nusa Tengra 2,737 205 7.5 2.0 Mbuku 1,411 153 10.8 2.9 Irian Jaya 1,174 237 20.2 2.7 TOTAL EASTERN ISLAS 10.517 1.341 12.8 2.2 3.6 TDTAL INDNESIA 146.935 32.846 22.4 2.3 4.0 aI/ hese are estimates based on Imperfect informtion. Sotce: 1980 and 1971 Population Censuses issued by Central Bureau of Statistics. - 15 - time. Bank staff estimatesL/ suggest that, between 1980 and 2000, Java's urban population will grow from 23 million to 41 million, while the urban totals on the Outer Islands will rise from 10 million to 31 million. Though it may be argued that the Outer IslaLnds will face difficulties in attempting to urbanize at a rapid pace, the underlying policy message of the cited estimates remains valid. The growth of those Outer Island cities may help relieve the pressure on the cities of Java, by diverting rural-urban migration toward Outer Island centers. Many of these migrants may not be interested in the Outer Island rural transmigration sites now being promoted by the central government; they might, however, welcome the opportunity to find urban employment on the Outer Islands. 2. Given the probable rates of decline in fertility and mortality and the likely trends in the pattern of migration among provinces, estimates of total and urban population can be made through to the year 2000 (Table 3). Following such a scenario, the growth of Indonesia's population is projected to decline from an annual rate of 2.3% in the 1970's to 2.0% and 1.8% in the next two decades. This decreased growth rate is unlikely to be matched by a similar deceleration in urban population growth. A conservative projection would assume a continuation of the recent 4%/annum urban growth rate over the remaining two decades of the century. By 2000 the urban population would then total 72 million or 34% of the population. Such a scenario implies that half of the population increase in the 1980's and fully two-thirds of the Population increase in the next decade will take place in urban centers. 1/ These estimates assume inter-lprovincial and rural-urban migratin patterns are unchanged from 1975-1980 rates, and that Indonesia reaches a stationary population by 2025. - 16 - Table 3 Projected Growth of Total and Urban Population in the 1980s and 1990s Mid-Year Population Increase in Population 1971 1980 1990 2000 1980-1990 1990-2000 Total (million) 119.0 146.0 178.7 213.6 32.7 34.9 Urban (million) 22.9 32.6 48.6 71.9 16.0 23.2 Percent Urban 19.2 22.3 27.2 33.7 49 67 Source: Bank Staff estimates. Total population projected to grow at rates of 2.04% and 1.78% per year in the 1980's and 1990's respectively. Urban population projected to grow at 4.0% per year throughout the period. - 17 - 3. These projections, based on assumptions about likely rural-urban flows on Java, may be modified marginally by actions taken to sustain the rate of economic growth in rural areas. Nevertheless, there is greater potential for influencing the pattern of urbanization than for controlling the overall total of urban population. Elements for such a strategy are discussed below. B. Rural - Urban Migration and Urban Population Growth Data limitations 1. Of the estimated 9.6 million increase in urban population between 1971 and 1980 roughly half was accounted for by the natural increase of the existing population and half by migrants and their offspring. By comparison, between 1961-1971 fully two-thirds of the growth was due to natural increase alone (Table 4). Migration is largely an intra-provincial phenomenon; twice as many migrants moved within provincial boundaries as across them. Even this exaggerates the importance of inl:er-provincial migration because Jakarta is classified as a province separate from its hinterland, West Java province. If Jakarta-West Java migration were considered as "local" in nature, then the overwhelming bulk of migration would be intra-provincial. Unfortunately, the data on migration to urban areas are fragmentary. Little information is available on within province migration. For inter-provincial migration, data are available on provinces of origin and of destination but the breakdown into urban and rural desa is available only at the point of destination. Thus, it is not possible to say whether migrants came from urban or rural areas, nor is it possible to identify the number of migrants leaving urban areas, with the exception of Jakarta. Within these constraints, some conclusions about migration behavior can be reached. - 18 - Table 4 Components of Urban Growth: 1961-1971, 1971-1980 1971-1980 1961-1971 million % million % Increase in Urban Population 9.63 100 6.76 100 Natural Increase of Existing Urban Population a/ 4.63 48 4.59 68 Migration and "Urbanization" 5.00 48 2.17 32 Migration from Other Provincesb/ 1.66 17 n.a. n.a. Migration from within the Province LI 2.92 30 n.a. n.a. Natural Increase of Migrants .d 0.42 5 n.a. n.a. ia Based on estimated fertility and mortality in urban areas. S/ Derived from 1980 census data on gross movement to urban areas. Assumed that level of migration from urban areas is half level of in-migration. sc Derived as residual. d/ Derived from age-specific fertility and mortality behavior. Source: Central Bureau of Statistics and Bank staff estimates. - 19 - Permanent migration 1. Permanent inter-provincial migration is relatively rare in Indonesia. In 1980, less than 7% of the population (roughly 10 million) were living outside their province of birth. Of these, approximately half had migrated to urban areas. In Eact, 96% of those living in rural desa and 84% of those in urban desa were born in the province of residence. However, though the amount of inter-provincial migration is low, it does appear to be accelerating. Of the 5 million urban migrants who crossed provincial boundaries, fully half had arrived in the previous decade alone, and one-third had settled in the previous five years.1/ The 1980 census also provides evidence that inter-provincial migration is sturdy enough to withstand administrative controls. Jakarta attracted 48% of the total interprovincial migrants to urban areas over the last decade, a proportion similar to that of the previous decade (47%). This occurred in spite of attempts to stem the flow of migrants to the capital through the introduction of residence permits. 2. Movements from rural to urban centers within provinces are the most important source of migration for all areas excluding Jakarta, which is itself a province. Thus, for example, the majority of migrants to Surabaya, Semarang, or BOTABEK (suburban Jakarta) come from their respective immediate hinterlands. This is probably generalizable to all urban areas, except for some of the transmigration centers in Sumatra and Kalimantan. Furthermore, this source of migration is growing over time.The estimated 3 million rural- jJ A definitive judgment on trends would, of course, require information on the rate of "return" migration and on migrant death rates. The 1980 census does suggest that a small portion (12-15%) of inter-provincial migrants again move across provincial boundaries within five years of the original move. This suggests that the rate of inter-provincial migration has, in fact, accelerated over time. - 20 - to-urban within province migrants recorded between 1971 and 1980 exceed the total number of migrants to urban areas for the preceding decade by 750,000.1/ Unrecorded migration 1. The above discussion severely underestimates the degree of rural- urban mobility on Java. Seasonal migration to urban areas is an increasing phenomenon; it makes conventional measures of urbanization an imperfect proxy for actual city growth. Seasonal migrants include a varied group of indi- viduals; some spend only a month or two in urban areas during slack periods in agriculture, while others spend most of the year in urban areas and return to their rural homes for only a month or two. The census fails to include these migrants for a variety of reasons:, city "residents" must have lived in an urban area six or more months to be counted as migrants and many seasonal migrants fail to qualify under this rule; in addition, unregistered workers in Jakarta tend to exclude themselves from that city's total population count; finally, many seasonal migrants, even those spending most of the year in a city, continue to record themselves as residents of their home villages. 2. Studies of rural and urban labor markets in Java, made during the last few years, emphasize that circular migration is high and growing rapidly. These studies suggest that at least 25% of rural households on Java have at least one family member working part-time in urban areas. This would imply a total of 3.75 million people, equivalent to 16.5% of the measured urban population and just over 50% of the measured 1980 urban employment on Java. Given the circular migration, the average impact on urban areas is less severe, but it is quite likely that approximately 15-20% of the urban work j/ There are no direct data available on within-province migration. However, given estimated rates of urban growth, natural increase and interprovincial migration, intra-provincial migration can be estimated as a residual. - 21 - force consists of temporary migrants. Finally, the fact that this form of migration apparently has grown rapidly over the last decade requires urban population and labor force growthl rates be revised upward. Under plausible assumptionsl/ one can assume, for example, that Java's urban population grew at 4% and not 3% over the last decade, while its urban labor force growth exceeded 6% per year, instead of the 4.7% annual rate implied by the census. 3. To understand why this phenomenon is not likely to fade in significance, one need only lookc at the changing sectoral composition of employment in rural and urban areas (Table 5). Although agriculture provided three-quarters of rural employment in 1971, it accounted for less than one- tenth of the increase in rural employment over the 1971-1980 period., During that decade, over six million children of rural households entered the labor force, but only one in nine found regular employment in agriculture. These new entrants into the labor force are the key to the urbanization process in Indonesia today. Rough estimatesj/ suggest that of the 5.4 million "surplus" workers, about 3.1 million stayed in rural areas, with half entering the service sector and half joining the rural industry ranks. A portion of these became seasonal workers in the cities of Java. The remaining 2.4 million moved permanently to urban areas; of these, less about 400,000 found employment in urban industry, whiLe the remainder joined the service workers. Urban quality of life and rural-urban migration 1. The ability of Indonesia's urban areas successfully to absorb these inmigrants is noteworthy. More imnportant is the fact that across the cities, 1/ The calculations that follow assume that the percentage of rural households with at least one family member working in the cities rose from 15% in 1971 to 25% in 1980. g It is assumed that the new labor force entrants from non-agricultural households joined the same sectors as their parents. - 22 - Table 5 Structure and Growth of FWployment in Urban and Rural Areas 1971-1980 Rural Urban % of Increase X of % of Increas_ X of Total 1971-1980 Total Ibtal 1971-1.980 Total Enployment ('000) Increase Enployment ('000) Increase (1980) ('000) Agriculture, Forestry, Fishing 67.3 458 8.5 9.3 231 6.2 Industry 13.0 2265 41.8 28.8 1191 31.9 Manufacturirig 7.9 1062 19.6 14.1 679 18.2 Construction and Utilities 2.6 670 12.4 5.9 252 6.7 Other IndListry A 2.5 533 9.8 8.8 260 7.0 Services 19.7 2690 49.7 61.9 2312 61.9 Trade, Restaurants, Hotels 10.3 1197 22.1 24.9 859 23.0 Other Services 9.4 1493 27.6 37.0 1453 38.9 Total 100.0 5408 100.0 100.0 3738 100.0 .S Tranuport, Storage and Camuunicaticns. Sources: 1980 data from 1980 Census, Series S, No. 2; seasonally adjusted to mid-year. In addition, ezployment in unspecified sectors is allocated proportionally to all sectors. 1971 data from 1971 census adjusted to make consistent with definiticm in 1980 cnsus. Total enployment is from series C while the sectoral breadcmn is from series D. lhe data are also adjusted to include estimates for E. Tivr, and nral Irian Jaya (exdluded frcm 1971 census). - 23 - incomes and access to publLc services are higher among all groups than ever before. Table 6 summarizes trendEs in rural and urban poverty during the last decade. Through reductions in the! proportion of individuals with consumption below poverty lines is broadly evident across Indonesia, regardless of location, the urban poverty figures are particularly heartening. While half of all urban residents could be considered destitute in 1970, when the economy was just entering a period of recovery after decades of neglect, the proportion had fallen to only one in five by 1980. Across urban Indonesia, incomes of the poorest 40% of the population rose at 4% per annum for the 1970-1980 decade. The migration t:o the cities also facilitated the growth in incomes of the poorest rural dwellers, whose incomes rose at an annual real rate of 2.4%/annum.I These data can be supplemented by other which suggest rising urban expenditures on consumer durables and decreasing proportions of urban household budgets being devoted to food purchases. 2. Complementing these findings, it is clear that urban dwellers have experienced improved public service coverage, in spite of the fact that overall urban population expanded from 23 million to 33 million over the 1971- 1980 period. In urban and, to a lesser extent, rural areas, the opportunities for primary education are becoming universal (Table 7). As a result, j/ It is possible that urban poverty is underestimated by the surveys from which these data were derived. Seasonal migrants and semi-urban residents in the periphery cities are likely to be ignored. In addition to the extent that the very poor are highly clustered in particular neighborhoods, the sampling techniques used may misestimate the number of poor residents. - 24 - Table 6 Trends in Pbverty Incidence in Urban and Rural Areas: 1970-1980 Percent of people with consumption below the desigiated poverty cut-off line) Region 1970 A/ 1976._/ 1978 b/ 1980 -/ JAVA Urban 56.3 33.8 27.5 20.9 d/ Rural 67.0 62.7 65.0 52.9 Total 65.0 57.3 57.9 46.9 OTHER IANDS Urban 40.8 28.0 21.2 17.3 Rural 43.9 39.6 34.3 30.3 Total 43.2 37.3 31.8 28.0 Urban 50.7 31.5 25.2 19.7 i Rural 58.5 54.5 54.0 44.6 Total 57.1 50.1 48.5 39.8 a/ January-April. b/ Average for year. Cl February. d/ In this table the 1978 SUSENAS definition of urban areas used for 1980 data. Scrce: Social EXpenditure Surveys (SUSEMAS). - 25 - Table 7 Perceritage of School Age Population Attending School a/ by Age Group 5-9 Years 10-14 Years 15-19 Years 20-24 Years Urban Rural Urban Rural Urban Rural Urban Rural 1971 Census 43.8 33.1 75.4 58.5 42.3 18.1 16.1 3.2 1980 oensus 69.2 55.3 139.5 77.0 53.4 25.7 15.4 3.4 al Includes school, university, technLcal colleges and full-tdie training courses. Source: 1971 Population Census, Series D (March 1975). 1980 Population Census, Serieis B, No. 2 (February 1983). - 26 - illiteracy among urban dwellers aged 15-29 has fallen to 5%-7%, compared to 28% for older inhabitants. New rural inmigrants are making a particularly impressive contribution to the gradual upgrading of the labor force: fully 42% of those who migrated to urban centers in the four years preceding the 1980 census had more than a primary education. 3. Good progress has also been made in providing safer drinking water to urban areas. Between November 1980 and the end of the Third Five Year Plan, REPELITA III (March 1984), it is estimated that the proportion of urban dwellers with access to reasonably clean water (piped or pumped) 1/ rose from 35% to 47%; coverage was extended to an additional 3.8 million individuals in less than four years. 4. Through the Kampung Improvement Program, initiated in 1969, the government has put into effect a far-reaching national program to upgrade the informal, unplanned, and unserviced "urban villages" which house over half of Indonesia's city population. Initial upgrading in the largest 15 cities brought basic services -- especially improved access, drainage, and water -- to large numbers, including 3 million people in Jakarta and 400,000 in Surabaya. The program has subsequently been extended over the last 5 years to cover nearly 15,000 hectares across more than 100 cities, affecting another 10-15% of the estimated kampung population. 5. Much, of course, remains to be done. The urban road network, for example, is inadequate, with survey results for areas outside Jakarta suggesting that half of the facilities are either seriously deficient in physical quality or unable to carry projected traffic loads without 1/ It is reasonable to assume that piped water and most pumped water is clean enough to drink directly after boiling. Pumped water includes handpumps installed by individual households. - 27 - extensive, repeated congestion. In Jakarta, 60% of the road system is subject to serious congestion. Such problems are being exacerbated by the 15% average annual increase in the number of motor vehicles across Indonesia. In addition, coverage of such services as formal solid wates disposal, safe human waste disposal, and even electric power, is still only one-third to one-half the level required for universal availability in urban areas. This gap must be addressed even as absolute urban population growth over the next two decades far outstrips the 10 million added in the 1970s. 6. Simultaneously, the public sector must face the serious problems of the rural areas. While much public concern is focused on disparities in per capita regional gross domestic product, the evidence suggests that more attention should be paid to the enormous rural-urban gap (Table 8). Urban areas across all regions have per capita consumption levels which are at or considerably above the overall national average. Furthermore, and excluding the Eastern Island and Irian Jaya (which includes only 7.5% of the nation's population), there are few differences in urban per capita consumption across regions. By contrast, per capita consumption in rural areas is uniformly below that found in urban areas in the same region. In Java alone, rural capita consumption levels are half those found in urban areas. As noted, the incidence of poverty in rural areas is much higher than in urban areas. 7. To address these rural problems, and to cope with the macroeconomic constraints created by stagnating oil export revenues, the central government must reassess the whole question of who pays for public servicese There is a prima facie case for shifting resource mobilization for urban public services down to the relatively prosperous urban areas themselves, leaving the center free to target its grants to the rural and urban poor. - 28 - Table 8 Regional Variations in Per Capita Oitput and Coisunption, 1980 (Index: Indonesia = 100) Per Capita Per Capita Per Capita Cbnsuption Regional Product Regional Product Total Rural Urban (incldinng mining) (excluing mining) Java 76 89 97 66 129 Sumatra 167 127 116 102 138 Kalirantan 226 172 129 118 139 Sulawesi 75 100 87 80 120 Eastern Islands and Irian Jaya 83 82 77 72 98 Indonesia 100 100 100 77 129 Source: Central Bureau of Statistics, Bank Staff Estiiates., - 29 - C. Urbanization and the Growth of Large Cities Difficulties in defining urban growth by city size 1. Assessing the importance and growth rate of different size grouping of cities in Indonesia is di:fficult. Apart from the 54 kotamadya (municipalities), and the towns designated to become kotamadva, cities and towns in Indonesia have no administrative identity or boundaries. Some major urban centers, with populations in excess of 150,000 are not kotamadya, while some kotamadya are small urban centers. The roughly six hundred cities that are not kotamadya each consist of one or several desa. For statistical reporting purposes, these cities are non-existent entities, and can only be identified by a laborious aggregation of desa-level data, using arbitrary, unofficial "city" boundaries. Comparisons over time are further complicated by the growth of official cities beyond their recognized boundaries, by periodic changes in the boundaries of some official cities, and by changing definitions of urban desa. The information that follows uses unofficial but realistic boundaries for existLng kotamadva and accepts the boundaries employed by sub-district (kecamatan) heads in responding to a 1978 Ministry of Interior survey of urban conditions in their area. The resulting breakdown of city size distribution, for 1980, and of selected growth rates by size category, for 1961-1980, are reported in Table 9. Trends in urban growth by city size 1. As of 1980, Indonesia had thirty-six medium and large sized cities, with populations in excess of 100,000, and an additional 293 with populations of between 20,000 and 100,000. As a group these 329 centers include all but 13% of the 33 million urban residents of Indonesia. Five urban centers have populations in excess of one million and house 40% of the urban population; - 30 - Table 9 A: Size Distribution of Indonesia's Cities and Towns. 1980 Size of CLty Total Inhabitants Number of Cities in Cities of this Size (million) Less than 20,000 4.3 369 20,000 - 100,000 6.8 293 100,000 - 500,000 6.1 27 500,000 - 1 million 2.5 4 MDre than 1 million 13.1 5 Total 32.8 670 Source: (a) 1980 Census for population of kotamdya roughly adjusted for the major cities of Jakarta, Bandung, Surabaya, and Surakarta to take acomunt of growth of metropolitan areas outside kotamsdya boundaries. (b) 1978 Dalan Nbgeri "Cwiat" survey updated to 1980, assuming 8% growth of all cities between 1978 and 1980. B: Growth Rates of KbtgMaya by Size Size of City Average Population Growth i (Percent per Year 1961-1971 1971-1980 Large Gities (over 500,000) 3.5 4.1 Medium Cities (100,000 - 500,000) 2.4 3.8 Snall Cities and Towns (below 100,000) 3.0 3.2 Total 3.2 3.9 i To the extent that kotmadya boundaries have in general not been eapended in line with peripheral urban growth, these growth figures probably underestimate the growth of the Kotmd a. Instead of the 3.9% overall annual groeth rate presented here, a figure of, say, 4.2% is probably closer to the truth, implying a rate of about 3.7% p.a. for rnnkotanudya. Sotwoe: 1961, 1971, 1980 Population Census documents issued by Central Bureau of Statistics. - 31 - DKI Jakarta alone, with 6.4 million residents, accounts for one-fifth of the urban population. Thirty-one medium-sized centers (100,000 - 1 million) contain roughly one quarter of the urban population, with the rest found largely in the small cities.. 2. In the absence of time series data for all urban areas, growth rates over time can be proxied by the performance of kotamadya in different size categories. Growth rates in the 1970s exceed those in the 1960s for all groups, coincident with the upsurge in economic growth in the recent past. Large and medium cities are growing at or about 4% per annum, while small cities are expanding at a slower rate of 3.2%/annum. Significantly, in light of repeated concerns that Jakarta is expanding too rapidly, the rate of population increase in the capital area over the same period is 3.9%/annum, a rate which would have to be adjusted marginally upward to take account of the above-average performance of the excluded BOTABEK suburban hinterland and to include the impact of the growth in seasonal migration. 3. Judged by the standards of other developing countries, the 1980 share of large cities in the total urban population of Indonesia is quite low (Table 10). While Jakarta included 19% of the urban population, the 1980 proportion for low income countries as a group was 28%, while for middle income countries it was 29%. Jakarta's share is also quite low by the standards of other countries in the region. In addition, these proportions can be supplemented by data which point to the level of population dispersal across city sizes. In Indonesia, for example, the nine largest centers account for less than half of the urban population, a level which is rarely matched in the developing world. - 32 - Table 10 Indicators of Urban Concentration-Selected Large Countries 1/ Share of Largest City in Urban Population, 1980 Indonesia 23% Bangladesh 30% Brazil 15% Mexico 32% Nigeria 17% Pakistan 21% Philippines 30% Thailand 69% Vietnam 21% Low Income Countries 28% (excluding China and India) Middle Income Countries 29% j/ National populations vary from approximately 50 million to 150 million. Source: World Development Report 1983, Annex Table 22, New York, Oxford University Press (for the World Bank), 1983. - 33 - 4. Given these data, there is only limited evidence that Indonesian urban growth to data has been in any sense unbalanced. Certainly mere size is no grounds for present concern about the larger cities, since there are innumerable cases of multi-mill.Lon population centers in the developing world. In addition Jakarta's share of urban population and of national population is not so high as to elicit serious debate on geopolitical grounds. The key issues instead are whether cities in each size category are being efficiently managed, whether local governments can compete for skilled workers and employment by tailoring local taxes and expenditures to suit their demands, and whether the macroeconomic environment is neutral with respect to the spatial distribution of economic activity. This, rather than concerns about optimum geography, should be the guiding principle for policy-makers. IV. Influencing Urban Development A. So=! Basic Principles Determinants of city size in the course of development 1. Such a perspective can b,e derived from an examination of how cities of different sizes typically evolve as the pace of urbanization increases in any given country. While the degree of urbanization appears to be closely associated with the level of economic development, the link between urbanization per se and urban concentration is not obvious. As the proportion of urban population increases, the population of most individual urban areas tends to grow, but the number of urban areas grows faster and thus the share - 34 - of any large urban area also falls.1/ This does not exclude the possibility that a higher proportion of the urban population is housed in an increasing number of larger urban areas. More important than these indirect relationships between urbanization and urban concentration are two other factors: the impact of national production patterns and the policies that affect the relative attractiveness of different cities as centers of production and consumption. The impact of national production patterns 2. Cities and their size are shaped by the tradeable goods and services they produce for export to other areas and, through such base activities, by the demand they generate for purely local goods and services. The development process that increases overall urbanization also reshapes the nature of tradeable goods and services produced in a nation's urban centers. Cities do not grow primarily because of any sharp increase in the proportion of workers employed in industry. Instead the composition of national industry activity changes; in addition, larger fractions of an increasing service sector are devoted to business services. This has implications for what types of urban centers are likely to grow relatively rapidly at each stage in a nation's development. A stylized version of this process can be explored at this point so as to make the validity of recommendations in later sections more apparent. 3. Given the existing resource base and the friction encountered in transportation and communication between regions of a country, early urbanization is built around resource-based tradeable activity dispersed across the economy. These include mining and agroprocessing activities, .j/ Given the relatively low level of urbanization in Indonesia, it is not surprising that Jakarta's share of the urban population has not yet declined. - 35 - construction materials, beverages, textiles and handicrafts, and the repair and production of simple structures and machines. In addition, ports involved in interregional or international trade, and administrative and agrobusiness service centers proliferate. Activities which suffer no impediments to location in and around major ports are attracted by large, accessible centers and begin to define the specialiZed profile of these larger cities; a good example is the concentration of industrial processing of most food products at this stage. Major ports are aLlso ideally located when the process of substitution of imported consumer goods gets underway; this is so because firms engaged in such activities frequently wish to use imported inputs while avoiding the high costs of internaL trasportation. When promoted artificially by heavy tariff barriers in a country with limited infrastructure, import substitution will provide a significant boost to the rate of growth of major ports, while "taxing" the rural sector and the smaller centers dependent on its health. There follows a stage when the most rapidly growing tradeable urban activities are associated with the emergence of engineering and metallurgy subsectors. In response to the relatively underdeveloped nature of transport, communications, skilled labor supplies, and information on markets and technology, the initial expansion of these sectors favor larger port centers where savings associated with economies of agglomeration are abundant and where imported inputs are available. Larger urban areas act as economic supermarkets, providing a wide variety of inputs, services, skills, and clients for firms; and consumer goods, services, and opportunities for workers of varying skills. 4. As consumer and producer goods' production processes become more standardized; as cheaper, more frequent, and reliable interregional transport becomes available; as other national infrastructure is improved with - 36 - development; and as education investments boost the supply of higher skilled workers; a redistribution of growth takes place favoring medium sized centers of 100,000 to one million residents. This occurs both because of the fact that such smaller centers can begin to cash in on their inherently lower costs of living and because standardized production and marketing is less sensitive to the economies of agglomeration provided by very large cities. Now the most efficient locations for the dynamic sectors tend to be own-industry concentrations of particular lines of activity in smaller centers. At a lower cost than possible in big agglomerations, clusters of similar firms can attract most of the special services and inputs they require. The result is a landscape dotted with medium-sized centers that specialize in particular activities and trade extensively with the rest of the urban system. At this point in the development process, large cities come to depend on two sources for continued growth: new, unstandardized high technology and business services, which benefit from the economies of scope provided by big cities; and, more generally, the attraction of resource-independent or "footloose" activities in search of large consumer markets somewhat protected by the nature of final delivery transport costs. 5. Note should be taken of two additional phenomenon associated with the changing profile of national output. Decentralized development emerges less from the physical relocation of activity from one type of urban area to another than through the differential rates of growth of local-based entrepreneurs and of new investors. Except for supporting deconcentration in and around large centers, at distances unlikely to exceed 50-100 kilometers, big city factories and service establishments do not transmit their dynamism to medium-sized centers through relocation or branching. Furthermore, decentralized industrialization is limited in the earlier stages of economic - 37 - development by the inherent locational inflexibility of most industrial activity. As long as infrastructure ;constraints prevail, especially with respect to interregional access, firms will locate in historically preeminent centers unless a) their resource-based inputs are highly perishable or suffer large weight losses in processing; b) their product is a processed resource going directly to export markets; or c) their output is marketed locally at prices which are competitive with production coming from more distant and larger urban centers. In this environment it is quite useless to try to lure or force activity away from large cities with easy access to a major port.- The relative attractiveness of cities and city size 1. While the changing profile of national production in tradeable goods and services provides a basis for the emergence of medium-sized centers, there is a second factor at work det:ermining the speed of such decentralized development: the rate at which city attributes can be reshaped to attract new types of activity. Crucial to this process is the improvement of local public services and of regional and interregional access. On one level this appears to be an intuitively obvious matter: better infrastructure reduces the costs of doing business, under reasonable tax and user charge regimes. Less apparent is the important role these improvements play in attracting and 1/ The applicability of this framework to Indonesia is implicit in the various regional economic development studies performed over the last decade. Examples include the Java Regional Study of East Java (1975) and of Central Java (1977) and the, Southern Coast Development Plan for East Java (1980), all prepared by t:he Japan International Cooperation Agency. Also worth citing are the Sumatra Regional Planning Study (1977) by Robert Nathan Associates, and the Sulawesi Regional Development Study (1979) by the Canadian International Development Agency. - 38 - retaining high skilled workers, managers and entrepreneurs.j/ Work in other developed and developing countries has demonstrated that high and low skilled labor are poor substitutes for one another. This means emerging tradeable goods and services are associated with fairly rigid skill mix requirements. Decentralized growth as described above thus depends on mechanisms which overcome the disamenities of emerging centers. While some amenities critical to urban development in economically advanced countries -- like climate or natural recreational assets -- are beyond the scope of public intervention, many amenities are associated with the quantity and quality of public services. As the level of amenities rises in any area, businesses in that location are able to offer lower compensation packages to skilled workers than would otherwise be the case. At some stage, these reductions may mean not merely greater scope for expansion of individual firms but the differences between the existence or absence of many lines of activity. Limits to public intervention in shaping city size 1. Before contemplating specific courses of action by the public sector, the Indonesian government should consider the large-scale replicability of any potential action. At any one point in time, only a few central initiatives will be possible, and those must husband financial and administrative resources to accomodate all the other competing demands for public intervention. Unfortunately the very conditions of development which call for remedial action limit the funds and the competent personnel available to implement such initiatives. In addition, many of the processes of development 1/ Recent work on the United States and Brazil confirm this assertion. See V. Henderson, Urban Development in Brazil, Urban Development Discussion Paper, Water Supply and Urban Development Department, 1983; and V. Henderson, "Population Composition of Cities: Restructuring the Tiebout Model," Working Paper No. 82-16, Brown University, 1982. - 39 - -- especially those described above -- are so complex and subject to so many interactions that omniscience is impossible. For these reasons the temptation to assign an optimal size to a city in any location as a goal of public policy must be resisted. Government could try to throttle the growth of cities above a certain size by banning a whole range of activities or licensing any attempts at local expansion but the effort would mire the bureaucracy and its subjects in red tape and the consequences of success might be very harmful to national economic growth. Or, public officials could take it upon themselves to declare a city to be too small and then attempt to flood the area with all manner of integrated packages of subsidies and technical assistance so as to create producers of "missing" services, parts, and components "necessary" to reach an optimal size. That approach too would fail because the effort would involve a level of resources that would doom any effort at replicability. Among the most notable of these types of initiatives is that of "creating" regional balance by building greenfield new cities of particular sizes needed to provide for missing city sizes. 2. In the end the spatial configuration of population and employment is the indirect outcome of other policies, which shape the behavior of millions of economic actions across the urban landscape. What is needed is a set of efficient policies across the difEerent sectors where government operates, at all levels. As argued below, tlhese policies should ensure a macroeconomic environment supportive of efficLent and broadbased growth, backed by an orderly expansion of interregional investments in transportation, communi- cations, power generation, and education. Cities, taken as entities, should be free to grow and develop with as little regulation and control as is consistent with a sober assesesment of national goals. The amenities represented by local public serv:Lces should continue to be extended over an - 40 - increasing number of medium and small cities, without neglecting the absolutely large needs of major centers. This will require that public service consumers at all levels pay a greater share of the cost of providing such services. 3. Local communities must take on a larger share of planning, executing, and maintaining such investments, freeing the center to use its scarce resources to intervene in a more selective fashion to shape the future of Indonesia. One necessary ingredient is the extension of municipal status to a larger number of de facto cities. A major effort should be undertaken to designate as kotamadva all urban areas above a certain size, say 20,000.1/ While it is true that in Indonesia the goal of increasing the involvement of local governments in local affairs will require considerable efforts to upgrade the proficiency of local government officials, there is no plausible alternative. The central government bureaucracy is already overextended; the detailed intervention in local affairs practiced at present cannot be continued without incurring increasingly heavy efficiency costs. The time for devolution of authority is at hand.2/ The problems of an urban community of over 20,000 residents are surely more than enough to occupy local authorities without superimposing the additional responsibilities of governing surrounding rural areas. The problem is all the more acute when the functional urban areas cut across county boundaries and involve different local bureaucracies 1/ A Ministry of Home Affairs 1978 survey of subdistrict (kecamaten) administrators identified the de facto boundaries and population of all cities and towns. From this one can estimate that there are roughly 329 urban centers with 20,000 or more residents. By contrast, there are only 54 kotamadva. Some unofficial cities have population in excess of 150,000, including Cilacap, Purwokerto, Banyuwangi, and Sorong. g/ Merely transferring powers to existing kabupaten (county) authority is inadequate. Counties usually contain large numbers of rural and urban dwellers. - 41 - B. Influencing Industrial Location Why focus on industrial location? 1. One way to affect the spatial pattern of economic activity is to guide the location of tradeable commodities toward or away from particular types of urban areas. Tradeable commodities generate export activity to other areas of Indonesia and the world at large. This, in turn, creates new opportunities for non-tradeable commodity production, as new sources of demand are generated for these "local" goods and services. This strategy is sensible even when tradeable commodities constitute a small percentage of economic activity in an area. This is because the production of tradeable commodities have multiplier effects, supporting several non-tradeable jobs for every job devoted to "export" activity. 2. Another reason for having a spatial policy grounded in the support of industrial activity is the alleged flexibility of industrial as opposed to service sector jobs. Industrial jobs are supposed to be more "footloose" and less tied to a particular location precisely because of their "export" orientation. In addition, they are less reliant on proximity to natural resources than agricultural actLvity, with the exception of industrial establishments operating with inputs that are highly perishable or that suffer considerable weight loss during processing. This flexibility should not be exaggerated. With exceptions, most private, domestic industrial activity originates in the area of residence of the founder-entrepreneur. The quality of the market serviced from that location will determine the degree of realized activity. Public and foreign investors do have more flexibility but they too are severely circumscribed by perceived locational constraints at the time of the establishment of the firm. The capital-intensive nature of - 42 - ventures based on the exploitation of natural resource endownments (e.g. fertilizer, natural gas, tin, nickel, copper, bauxite, pulp and paper, basic chemicals and forestry products) tends to attract much public sector investment and a significant portion of foreign investors. These sectors, however, are likely to require locations near the source of raw materials. Other foreign and public investments may feel constrained by the same factors that limit most location decisions, especially where the level of overall development is still low. The few locations offering the best initial opportunities for profit-making will receive the bulk of the investments. Once these investments, made in the early stages of development, are in place, these firms tend to act very conservatively with respect to new investment locations. Great modifications in the environment across space may elicit limited responses from existing entrepreneurs. They tend to favor deconcentration of activity away from the initial centers of activity, to locations rarely more than 100 kilometers from the headquarters location. New opportunities in locations that are not historically preeminent will tend to be exploited by entrepreneurs spawned in those regions. The dynamism of preexisting large centers is thus rarely transferred to new areas of activity, regardless of how their relative attractiveness has been improved. 3. These considerations are useful in assessing the relative merits of different policy initiatives. These fall into three categories: sectoral policies, microspatial initiatives, and urban service strategies. Sectoral policies can be divided into five major areas: trade, regulation, interregional transportation, government productive investment, and industrial finance. Among the microspatial initiatives worth considering are industrial zones and parks. The introduction of improved levels of urban services across all city sizes is discussed in a final section. - 43 - The spatial bias of sectoral policie!s 1. Public policies need not be explicitly spatial to have major impacts on the location of industry. Sectoral policies can favor the industrial development of some locations over others. Where these "implicit" spatial policies go unrecognized, their impacts can create barriers to decentralized industrial growth. These implicit spatial policy effects are difficult to measure with precision. The data problems involved are substantial and the eventual impact of implicit policieas depends on the way in which factors of production (labor, capital, land) adjust to them. Some conclusions can be reached, however. 2. Looking across induistrial subsectors and across cities, it is clear that the input requirements of dlifferent production processes will vary widely, as will the supplies of inputs at different locations. In the extreme, some inputs, such as easy access to an international port, to government officials, or to agglomeration economies, may be available at only a few, specific locations. At the same time, individual subsectors will differ in the emphasis they pLace on such variables. Under these circumstances, promotion policies favoring subsectors which feel location- bound will result in long-term subsidies for particular urban areas. If the individual industrial policies are ill-advised on macroeconomic grounds, the differential urban impact is one more reason to consider reforms. If the promotion policies are justified but the economic environment is such that economic agents, in urban centers other than those benefitted, are denied access to resources on reasonably equal terms with which to remedy locational deficiencies, then the public sector should act quickly to remedy the matter. Only where subsector promotion policies are lustified, and where - 44 - there is no cost-effective way to diminish the number of severely location- bound factors, can the government remain passive to the spatial consequences of particular promotion policies. For, ignoring implicit spatial policy impacts amounts to neglecting the important fact that the selective reinforcement of particular urban centers increases the attractiveness of such places relative to the rest of the urban system. The trade regime and city size effects 1. Ideally, for a country that combines a relative abundance of labor and a bountiful supply of natural resources, trade policies in Indonesia should support the development of a mix of labor-intensive and resource- dependent capital-intensive subsectors. These would include areas where there are comparative advantages over the short and medium-term. In fact, in Indonesia trade policy to date has been inappropriate, and needs reformulating on macroeconomic grounds. Such changes as have been suggested in other Bank documentsL/ would encourage greater efficiency in investment decisions, a better use of resources, and greater competitiveness in international trade. 2. Work performed in other developing countries2/ suggests that heavily protected sectors tend to locate in large urban centers, where they can satisfy their exceptional need for imports, sophisticated factors of production, and for access to government officials. This locational pattern- implies that the industrial value-added of a major metropolis, under such a 21 Especially relevant are the recommendations made in Indonesia: Selected Issues of Industrial Development and Trade Strategy, Report No. 3182-IND, 1981. 2/ A summary of recent research can be found in A. Hamer, Decentralized Urban Development and Industrial Location Behavior, Water Supply and Urban Development Department Report No. 24, the World Bank, May 1983. For evidence in East Asia, see Philippines: Industrial Development Strategy and Policies, The World Bank, May 1980. - 45 - trade regime, will be boosted by artificial inducements. Elsewhere, in smaller centers away from the more developed core of the country, this value- added bonus will be much smaller or non-existent. As noted, this promotion of subsectors in particular locations has long-term developmental effects, which encourage additional growth over and above that recorded at any one point in time. The benefits of a change in policies would take time to become fully effective. As long as access to national and international markets is poor, as long as infrastructure services are available umt..dLlly across cities for business and skilled workers, and as long as central government favoritism is worth seeking, there will be constraints on diversified development outside a handful of centers. 3. Nevertheless, a reorientation of promotional programs, providing roughly equivalent, moderate protection across all sectors, would allow production to restructure iltself spatially over the long-term. Both the Outer Islands and the smaller centers iLn Java would benefit. The subsectors that would be most favored have fewer ties to big city economies of scale and would thus have less reason to locate near Jakarta than the present mix of national output. These include the labor intensive industries, such as wood and cork products, saw milling and wood processing, tanneries and leather finishing, processed tobacco, kretek cigarette production, and certain textile products. One should add certain capital-intensive, natural resource-based industries which are located on Sumatra, Kalimantan, and the less developed islands: fertilizers, petroleum refinery products, natural gas, tin, nickel, copper, and bauxite. In contrast, incentives would be cut back for protected sectors that presently cluster in and around Jakarta. These would include motor vehicle assembly, pharmaceuticals, paints, plastic goods, electronics, - 46 - ref rigerators and washing machines, electrical appliances, and textiles and garments dependent on imported inputs. 4. With the passage of time, economic development will be accompanied by a loosening of general locational constraints and by a gradual expansion of the more sophisticated sectors of production. At that point, more cities will be capable of accommodating the demands of such advanced production, and modernization will no longer coincide with growth in one or two centers. Impact of transport sector inefficiencies on city size 1. Improvements in the trade environment need to be accompanied by other changes, including those involving interisland and international shipping. Under present conditions the largest urban areas benefit disproportionately from the relative isolation of Indonesia's cities. Historically, sea-borne trade has been saddled with high costs. The inter-island general cargo scheduled service operates at high nominal rates and suffered from inadequate schedules and inefficient routing. The composition, size, and age of the ships has not been adequate to meet local traffic needs. Compounding this and accounting for the bulk of the problem, port costs were inflated because of difficulties created by inadequate infrastructure, lack of sufficient equipment, cumbersome stevedoring arrangements, and inefficient management of overall operation. Pilferage, damage, and unofficial payments added yet another layer to shipping expenses; some estimates placed these additional charges at 100-150 percent of normal shipping costs. These problems reduced not only potential domestic trade but exports as well. 2. The central government is well aware of these problems and has prepared an action program to deal with issues in the maritime sector. A hierarchy of ports has been developed, four of which are to be gateway ports (Tanjung Priok in Jakarta, Surabaya, Belawan in Medan, and Ujung Pandang). - 47 - These gateway ports will be supported by six to eight regional trunk ports, serviced in turn by feeder ports. This should generate economies of scale in operation, and reduce investment requirements at each location. In addition, the multiplicity of port charges for piloting, mooring, stowing, and for small boats and harbor-master fees, are to be simplified or eliminated. Other charges, such as those for loading or for storage at port warehouses, are to be reduced. To accelerate cargo-handling, responsibility for round-the-clock loading and unloading operations will be transferred from shipping companies to specialized firms. Port labor wages will be increased and stevedoring fees will be reduced. To facilitate ioreign trade, foreign flag vessels will face a reduction and simplification of procedures. To ease inter-island goods transport, various requirements are being removed, including inter-island freight documentation, fiscal statements and tax-payment receipts. To improve the handling of goods and documentation, such tasks can now be performed by forwarding companies, importers, and exporters, as well as appointed clearing agents. Finally, major ports have been placed under autonomous Port Administrators, who will coordinate and improve operations. 3. According to detailed studies by consultants working on the Maritime Sector Development Program, reductions in direct transport costs of up to 50X can be expected from the introduction of these measures. The reforms will also cut indirect transport costs, especially unofficial payments, by reducing and simplifying customs procedureis and regulations, thus cutting the number of needed contacts between officials and customers; and removing the ambiguity and discretion inherent in the regulations that remain in force. In particular, new measures introduced in 1985 will drastically reduce the required import and export procedures to a bare minimum. Export goods are no longer subject to regular customs inspection. Imports will be inspected at - 48 - the points of origin by Government-appointed surveyors, who will also certify the type, quality, and volume of goods, the prices at points of origin, the tariff category applicable, and the appropriate import tariffs. Under these procedures, Indonesian customs officials will limit themselves to ensuring that this minimal documentation is completed, whereupon the cargo will be promptly released. Assessment of import duties and other taxes will now be made by the importers and exporters themselves, (on the basis of surveyors' certifications) and payments will be made directly to designated banks. 4. One other area could benefit from significant improvements. Most ports have heavily congested access roads that cause delays in and limit the growth of traffic. Many medium-sized ports./ are poorly linked to the surrounding hinterland, severely restricting their growth potential and, with it, the potential for decentralized growth. The locational impact of pervasive micro regulation 1. In the past, Indonesia's industry has suffered from widespread and poorly administered microregulation of production units. The government is now making efforts to simplify and streamline these controls. Numerous licenses and regulations have tended to hamper decentralized development, by making physical proximity to central government officials a factor in the location decisions of large and medium sized firms and by using administrative measures to attract investments away from or towards particular locations. 2. An important characteristic of the regulatory and licensing system has been its explicit lack of uniformity, a feature it shared with the trade regime previously described. Specific licenses may be required for new investments and expansions, which vary by product and by region. Firms may 1/ This issue is raised repeatedly the earlier-cited regional economic development studies. - 49 - also be subject to licenses regulating domestic trading, domestic and sea transportation export-import activities, as well as the more common permits governing wages, personnel nationality, worker safety, and pollution. This regulatory system is divided into two main groups, known commonly as the BKPM and the BRO sectors. 3. The BKPM category of firms covers those which have applied for licenses through the Investment Coordinating Board, the BKPM. All manufacturing firms with any degree of foreign equity and all purely domestic operations applying for tariff exemptions and other investment incentives must work through the BKPM. Other domestic firms whose size exceeds certain capital stock limitsl/ must register under the provisions of the Company Regulations Ordinance (BRO). Remaining firms must simply register their existence with local governments. 4. The BKPM operates in accordance to an Investment Priority List (DSP), which is revised periodically, and includes inputs from various government institutions. The DSP explicitly reflects attempts to balance productive capacity and demand, both nationally and by region, while protecting or promoting market shares of firms owned and operated by economically disadvantaged elements of the community. On this basis, the BKPM negotiates with affected firms through its Jakarta office, and proceeds through pre- application, licensing, and project implementation. As noted, once a project reaches the production phase, a series of licenses are needed which regulate the freedom to operate on a day-to-day basis. 5. Companies regulated under the BRO are subject to licensing for both new investments and capacity expansion. Licenses are issued primarily by the 1/ Total investment of below 10 million rupiah, excluding the cottage sector, with investments of up to 500,000 rupiah. - 50 - Directorate General of Miscellaneous Industries of the Ministry of Industry in Jakarta. Until recently no formal DSP system was used for BRO firms; since 1982 the Department of Industry has introduced investment restrictions; once again, the guiding principle behind such a system is the avoidance of significant excess capacity in industrial subsectors by controlling the degree of competition. Historically, BRO firms have found licenses for new or expanded capacity easier to obtain than BKPM firms.j] Experience with the new DSP list is too recent to allow a judgement to be made as to its liability. In general, however, it is safe to assume that the net effect of the DSP lists is harmful to economic development, in general, and regional development, in particular. A case can be made that very large, private or foreign investments with substantial impacts on their region should be reviewed by the central government. Beyond that, licensing entagles entrepreneurs in the web of an overworked central bureaucracY that can ill- afford the time to make seasoned judgments on each new proposed investment involving, say, 10 or more employees. Furthermore, the record is clear in demonstrating that overcapacity problems have not been avoided by regulation; problems occur inspite of the necessaril- imperfect forecasts of central planners. Given this fact, the government should reconsider the argument that widespread. official barriers to entry in different subsectors and regions does much more than reward firms for mastering hurdles that have nothing to do with performance. 6. From the viewpoint of spatial policy two issues are crucial: regulation is used to control the location of investments; and regulation is .jf As in the case of the more highly regulated BKPM firms, additional permits are needed; the number varies depending on the geographic scope of operations. - 51 - employed in a manner that rewards firms for maintaining close day-to-day contact with government officials at Jakarta and at major ports. The process of getting an investment license from BKPM or BRO forces firms to negotiate with government officials in Jakarta. But recent simplifications of investment procedures will drastically reduce the time and number of steps required to receive approval for new investments. Furthermore, this process does not affect location behavior directly, since, once a license is granted, only limited contact with the offLcials who granted those initial permits is required. The use of locational constraints in allocating investment licenses is, however, a major problem at this stage. 7. BKPM restrictions on location have expanded over time. A reading of the 1982/1983 list of investment priorities suggests that 250 of the 1877 industrial items cited are subject to location restrictions, while another 75 items are closed to all new investment. The most common locational restriction requires that particuLar investments be outside a given province or island, or in a location outside a set of listed provinces or islands.l/ As an example, seventeen items in metal products and metal building materials are restricted to locations outside Jakarta; sixteen categories are barred from the JABOTABEK region, including various types of electrical appliances, metal products. and metal building materials. Seventy-four items are forbidden from locating in Java, and run the gamut from printing for cans, printers ink, and plastic bags, to steel bars for concrete, metal desks, ceramics, roofing nails, and gymnastics equipment. Such restrictions, whatever their intent, distort efficient location behavior by restricting access to plentiful supplies of labor, large nearby potential markets, well 1/ A less frequent restriction specifies a particular location where investment must take place. - 52 - developed infrastructure, and abundant natural or semiprocessed resources. Among the undesirable consequences one can list the following: decreased overall investment by firms which refused to invest in mandated areas; more capital-intensive investment by firms forced to locate in inappropriate locations and faced with costlier non-Java labor; reduced export potential given higher costs; and reduced realization of economies of scale. Furthermore, because the regulatory system covers enterprises that are responsible for much of the manufacturing value added but little of the sector's employment, the primary impact of regulation is likely to be on industrial productivity rather than on the industrial employment the regulators wish to disperse. The appropriate response to a perceived pattern of choices which is viewed as too narrow is to modify the external conditions that bring about that reaction. Microeconomic regulation is more likely to impose dead weight' losses on the society at large than to encourage decentralized development. 8. Location controls aside, there is a second set of bureaucratic barriers which has encouraged large and medium-sized firms to locate in and around Jakarta or near major international ports. Day-to-day operations in the context of a permit-ridden environment encouraged ad hoc compliance and the active search for rulings granting exceptions on an individual basis. In addition, accounting practices followed for tax purposes have historically required the obtaining of official letters of clarification from taxation officials; in the extreme, tax liabilities sometimes were explicitly negotiated with tax assessors.jJ Furthermore, import procedures have been cumbersome; the process of actually moving goods through the ports was helped immeasurably by thorough familiarity with officials at major ports. Firms had jj It is hoped that the new tax code reforms will eliminate the more unfortunate consequences of the system in place. - 53 - to cope with the delays and invisible payments required to obtain a multiplicity of stamps and approvals as imports move through 29 major steps. Like taxes, tariff liabilities on imports were often subject to ad hoc decisions by customs officials. These problems should now be greatly reduced. However, individual industrial enterprises can still petition the Ministry of Industry for reduced tariffs on their imported inputs, and/or an increase in the protection of their output from competing imports. As with any ad hoc, discretionary goverrunent action, this encourages close face-to- face negotiation with respomsible government officials in Jakarta. 9. Casual empiricism suggests that the firms most susceptible to this centralizing effect are medium-sized enterprises that are too large and too dependent on imports to avoid controls, and, simultaneously, too small to be able to afford a subsidiary office: devoted principally to expediting decisions from the central government. FLrms in subsectors which are not resource- dependent processors are likely to use their "footlooseness" to gravitate to Jakarta or the other major port cities to seek relief from the bureaucracy. Highly regulated environment are often criticized on macroeconomic grounds. One can now argue that there are spatial impacts which tend to preserve historic distributions of invest:ments across regions, in spite of rapid increases in "hinterland" interregional and local infrastructure. and in market potential. This bias is not likely to be reversed by administrative fiat at the investment licensing stage. What is needed instead is a dismantling of most controls on competition, restricting the close supervision now accorded BKPM firms, in general, to a high priority category of firms engaged in large-scale, capital-intensive industries which require massive investments and which have vast impacts on the national and/or regional economy. - 54 - Industrial estates as tools for regional dispersal 1. The existing literature on estates is in general agreement about the constraints inherent in using estates to promote regional dispersal.l/ The first is that the opportunity cost of underutilized public investment in industrial infrastructure is so high that countries like Indonesia cannot afford to rely on "leading infrastructure" as an inducement for regional industrial dispersal. Second, estates facilitate development only when the absence of industrial sites and services is the major factor inhibiting industrial growth. Therefore estates need to be focused on major regional centers or in areas where they promote natural resource processing for export in products which the country has a distinct competitive advantage, such as forestry products. It is not useful to use estates as a device to develop "pioneer" locations that rank low in market potential. 2. In Indonesia the best guide to strategic growth centers is found by examining the locations of large and medium establishments2/ whose expansion fueled the rapid growth of industrial production during the first three Five Years Plans. Such establishments account for approximately 80% of industrial value-added and 60% of wage employment. They represent the cutting edge of industrial development, "introducing and adapting new technologies into the Indonesian economy and providing a reliable supply of good quality, relatively cheap and simple producer and consumer goods that are needed by other / Country experience is reviewed in the cited reports on the Philippines and on Malaysia. See also A. Uday Sekhar Industrial Location Policy: The Indian Experience op.cit. g/ Includes all establishments with 20 or more employees. - 55 - sectors"*./ Finally they are, historically, the most likely to be willing to pay for the industrial amenities that an industrial park tends to include.-/ 3. Evidence from the 1980 survey of medium and large establishments suggests that 80X of that employment is concentrated in 15 areas, which are either free standing or continuous clusters of cities and their counties. Included are the: zones in and around Medan, Palembang, Greater Jakarta, Bandung, Cirebon, Semarang, Tegal-Pekalongan, Surakarta, Pati-Kudus, Yogyakarta, Kediri', Malang, Greater Surabaya, Madium, and Pontianak. 4. The existing industrial estates in Indonesia generally follow the suggested guidelines..I Approximately 1300 hectares are available for industrial uses in six industrial estates in Jakarta. Two hundred forty five hectares are available in one estate in Surabaya. Additional estates exist in Semarang, (800 hectares), Cilacap (240 hectares), Cilegon (area around Krakatau Steel Complex), Medan (80 hectares), and Ujung Pandang (86 hectares). Various estate areas are being planned, some in areas earlier defined as industrial complexes (Cirebon, Surakarta) and others in areas where industrialization is in iLts infancy (Lampung and Padang in West Sumatra; Samarinda and Banjarmasin in Kitlimantan). These estate plans should be reviewed carefully to determine if the coordinated availability of infrastructure is the key constraint whose removal will bring accelerated development; if the answer is negative their development should be discouraged. / P. McCawley, "Industrialization in Indonesia: Developments and Prospects," Development Studies Centre, The Australian National University, Occasional Paper No. 13, 1979. 2/ Small and cottage establishments in Indonesia almost never utilize electric power, for example. 3/ Information based on data contained in Indonesia: Executive's Fact Book 1982/1983, The Chase Manhattan Bank, Jakarta, 1983. - 56 - 5. Another, less ambitious approach to easing industrial development is the so-called industrial zone.l/ In these areas infrastructure is provided either jointly by major project owners and the government or by one of these parties alone; no effort, is made, however, to prepare serviced lots as would be the case in an estate. These zones are being developed as part of a build- up of the basic chemical industries, rubber and cellulose, and organic and inorganic chemicals. They take advantage of the fact that these industries attract large-scale investors whose infrastructure requirements can be utilized most efficiently when shared with other users. Zones being developed at this stage include areas in Aceh (Lhok Nga, Lhok Seumawe), West Sumatera (Padang), South Sumatera (Palembang), West Java (Cibinong, Tangerang, Cilegon, Bekasi, Cikampek), Central Java (Cilacap), East Java (Gresik. Madura, Probolinggo), East Kalimantan (Bontang, Sesayap), South Sulawesi (Gowa, Tonasa), and East Nusatenggara (Kupang). Most of these areas can coincide with identified centers of industry. In other cases (including those yet being planned) the locations are not traditional ones and their development should be reviewed carefully before proceeding. 6. New enthusiasm exists for developing export processing zones (EPZ) and export estates.2/ EPZ are industrial estates that operate as extraterritorial entities where inputs and outputs can enter and be exported overseas without the payment of excises and duties, and with minimum customs and regulatory procedures. Often the enterprises operate under rules (such as 100% foreign ownership or special tax treatment) otherwise inadmissable in the 1/ Information on areas drawn from Indonesia: Executive Fact Book, op. cit. 2/ Shannon Free Airport Development Co., Limited, Export Processing Zone and Export Estate Masterplan for the Republic of Indonesia, draft final report, 1982. - 57 - nation's territory. As such, the EPZ aim to attract internationally mobile firms that face many competing offers; their existence is a frank admission that the national industrial promotion and control procedures inhibit initiative. Export estates would be geared at firms exporting half or more of their output while selling the rest domestically. Establishments in these estates would not enjoy duty-free entry of materials. but they would be subject to simplified customs and regulatory procedures, have nearby support services,and participate, wiLth other exporters, in export certificate schemes that provide rebates for duties of imported inputs incorporated into exported goods. Often, such estates contain bonded warehouses where goods originating within or outside the country may be stored without paying any taxes until they are released. To date, Indonesia has an EPZ, in the port area of Jakarta, as well as a bonded waLrehouse in the same area. While export processing zones would be useful as components of an outward-looking industrialization strategy, their locations would tend to reward a handful of preeminent port areas with excellent transport and communications facilities. Only in those cases where an export estate can be used to process bulky and difficult-to-transport natural resources, such as forestry products, would non-traditional locations be considered. While they have other advantages, these estates have only a minor role to play in the future dispersal of industry. The role of public industrial investment in regional development 1. If one combines wholly owned and joint venture public industrial establishments, their total work force includes 143,000 jobs (1980) equivalent to 15% of all medium and large enterprise employment (Table 11) It can be argued that these establishments have a special role to play in pioneering the development of non-traditional locations. To the extent these firms are - 58 - Table 11 Distribution of Employment in Public and Private MediumrLarge Establishments by Province (1980) Publicl/ Private D.I. Aceh 1.1% i Sumatera Utara 0.8% 4.4% Sumatera Barat 1.5% 0.5% Jambi - 0.6% Sumatera Selatan 3.9% 1.7% Bengkulu - - Lampung 0.2% 0.4% DKI Jakarta 8.9% 18.8% Jawa Barat 17.4% 20.1% Jawa Tengah 18.2% 19.1% D.I. Jogyakarta 2.4% 1.5% Jawa Timur 40.7% 26.2% Bali 0.5% 0.6% Nusa Tenggara Barat 0.1% 0.2% Nusa Tenggara Timur 0.2% Kalimantan Barat 0.1% 1.7% K.alimantan Tengah 0.1% 0.8% Kalimantan Selatan 0.1% 0.9% Kalimantan Timur 0.2% 0.6% Sulawesi Utara 0.1% 0.2% Sulawesi Tengah - - Sulawesi Selatan 2.5% 0.4% Sulawesi Tenggara - 0.6% Maluku - - Irian Jaya 0.7% - Total 100.0% 100.0% Total Jobs 143,153 827,375 jj Includes private-public joint ventures, with 20,223 jobs. By comparison fully-owned public enterprises have 122,930 employees. 2/ Percentages rounded off. Source: Central Bureau of Statistics - 59 - resource-oriented, the proposition is acceptable. But if this is taken to mean that public ventures should accept reductions in profits to promote regional development then it is at more debatable proposition and, beyond a limited point. an unacceptable one. Already many public enterprises are neither profitable nor well-managed. To expect a further departure from commercial goals is unrealistic. I'he constraints on the national treasury are such that these ventures should be sited in existing areas of development except where the resource orientat:Lon of the project justifies an alternative location. It does, however, make sense to coordinate the development of priority industrial zones and estates with the emerging locations of major employers, public and private, espiecially where non-traditional locations are concerned. This is likely to help minimize the danger of a dissipation of effort and scale across an excessively large number of sites. Regional development and the rationing of industrial finance 1. Previous studiesl/ have provided a number of stylized facts concerning industrial finance in Indonesia, and these are relevant to any regional dispersal strategy. Undier the system rules which prevailed until June 1983, public enterprises, jOint venture/foreign companies, and firms controlled by non-ethnic Indonesians had relatively easy access to either private or public industrial finance. To meet their working capital and investment fund requirements, they could rely upon at least some of the following major sources: state, private, and foreign commercial banks; foreign-offshore sources; and the domestic informal market. This system operated best in the short-term funds market, though larger customers were 1/ A good summary is found Ln "Financing Policies for Industrial Development", in Annex A of Indonesia: Selected Issues of Industrial Development and Trade StrategS, op. cit. - 60 - granted "rollover" privileges that lengthened these credits. Domestic long- term lending was dependent on government guidelines and backed by infusions of cheap liquidity credits. The system wrapped financial institutions in a maze of restrictions weakening competition for lendable resources and dulling the incentives to lend to the economically most attractive projects. As a result, a key role was played by demands for customer credentials and by burdensome collateral requirements, both to the detriment of loan evaluation. 2. Under the new system, the banking community is supposed to operate in a more competitive environment and rely heavily on deregulated interest rates to mobilize and disburse funds; the role of liquidity credits for priority sectors is to be sharply contracted. One interesting question is how the new system differs from the old in the treatment of establishments outside Jakarta and its West Java environs. Little is known about the regional differences in the supply of short-and long-term credit, but some plausible assumptions can be made. Large scale customers, regardless of location, are equally well treated under either regime, having the collateral and connections needed to obtain loans domestically and overseas. The same should be true for the larger of the medium-sized establishments, whose ranks include plants with 20 to 100 employees. Smaller establishments, especially the small-scale enterprises have always had limited access to the organized financial system, with the result that their level of indebtedness is relatively low and their accumulation of liquid assets is relatively high. This is especially so outside the Jakarta region, where branch banks have limited authority for credit requests above a certain low threshold. Since smaller firms are less - 61 - concentrated across regions than larger establishments,l/ it follows that regional development has been affected. The new, deregulated system, by modifying incentives to mobilize resources, may increase lendable funds across all regions. However, in an environment of centralized banking accustomed to short-term lending and procedures favoring high-visibility firms, the removal of government support for longer-term lending could worsen the situation of smaller firms outside the capital area. Borrowable funds are likely to gravitate to prime customers as determined by the central bank offices. 3. Among the many proposals that might be made to improve the results of the system now in place, there are two which merit attention as encouraging decentralized activity. First, rates for short- and longer-term loans for smaller size firms could be allowed to rise sharply, to provide a commercially acceptable spread over cost-of-funds. This will encourage an expansion in lending and reward financLal institutions that invest in loan evaluation, supervision, and collection. IPart of the annual allocation of central government funds for developmental purposes could then be used to reduced eligible borrower costs. In addition, the central government, through Bank Indonesia, could work aggresively to upgrade the role of regional development banks. These banks should be capable of capturing a higher percent of gross assets of the financial system (now 1%) and of diversifying away from its concentration on working capital and local treasury functions, and into longer-term loans. To the extent that liquidity credits are preserved for smaller-scale establishmenits, the GOI should consider channeling a higher percentage to these entities, which place high priority on local development. 1/ For a review of the evidence see Indonesia: Selected Issues in Spatial Development, O2. cit.; and Indonesia: Selected Issues of Industrial Development and Trade Strategy. _. -62- C. Management and Finance of Urban Services The financial implications of the urban goals of the Fourth Five Year Plan (REPELITA IV) 1. Indonesia has made progress in extending public services, in a systematic way, from large to medium and then small cities. The record of the last decade, reviewed above, makes the extent of this effort very clear. The achievements during the third Five Year Plan, REPELITA III (1979-1984) involved an average annual expenditure by all levels of government of 285 billion rupiahs (1982 prices)e-/ The government may be preparing to make even greater commitments over the remainder of the decade, with potential spending, as proposed by line agencies, averaging 529 billion rupiah per annum (1982 prices). This goal is consistent with decentralization objectives in that it attempts to equip medium and small centers across Indonesia with the level of amenities required to attract and retain higher skilled workers. This, in turn, will help promote the diversificatica of the economic base of such cities, while meeting the basic needs objectives of population as a whole. 2. These goals, though laudable, must share the arena with others. The improvement in service levels should not provide an indirect mechanism to subsidize any one city or set of cities. Local government taxation and cost recovery levels are low by international standards, and further provision of subsidized urban servies would exacerbate urban-rural inequalities. Urban dwellers are, in material terms, much better off than their rural counter- parts, and thus in a position to contribute more to the costs of meeting such services. Finally, increasing the contribution of urban governments would jj The approximate exchange rate was $1=Rp.630 (1979-March, 1983) and $1Rp.970 (after April, 1983). - 63 - help to release central resources for rural development. This requires that investments and recurring operating costs of urban services be paid for by the users, to a degree consistent with equity concerns. The same conclusions are forcibly reached by an examination of macroeconomic conditions and their effects on short and mediumrterm government revenues. The Fourth Five Year Plan, REPELITA IV, will be implemented during a time of scarce foreign exchange resources and modest domestic economic growth. The central government can no longer expect to rely, as before, on large-scale programs of grants to finance activities of local interest. Finally as more programs are extended from a few dozen centers to hundreds of medium and small cities, the administration capacity of the central government and its sectoral line agencies in Jakarta will be fully, occupied with dealing with selected issues that transcend local boundaries; by necessity, local planning, local execution, and local operation and. maintenance of public service projects will have to become the norm. Increased local funding would heighten local interest in the efficient management of its affairs. 3. Local governments, at the provincial level and below, operate in the shadow of the central governmenit. As Table 12 makes clear, even in the financing of urban services, the central government is the dominant force. In the recent past this preeminent role has been built upon oil revenues which, by 1980, accounted for over half of all central resources. The targets outlined in Table 13 suggest that the transition from REPELITA III to REPELITA IV could involve a potential increase in urban service investment expenditures equivalent to an annual rate of real growth of 13%. These rates of growth far exceed the expected growth rates3 for central government revenues. After growing for several years at an average real rate of over 142 per annum, the central government development budget for the period 1981-1989 is estimated to - 64 - Table 12 Annual Pattern of Spending for Selected Urban Services (percent distribution) Central Local Govt Item Percent of Govt and Cost Total Overall Total Recovery a/ Spending Water Supply 95 5 100 29 Sanitation and Drainage 52 48 100 13 Kampung Improv. 66 34 100 7 Urban Roads and Traffic Manag. 33 67 100 23 Public Transport 96 4 100 6 Others 10 90 100 12 Total 64 36 100 100 a/ Includes local borrowing from central government, and local spending of central government, and local spending of central INPRES grants. Source: Average 1979/80 expenditures in current prices, as reported in Indonesia: Urban Services Sector Report. - 65 - Table 13 Pbtential Financing ReciuiremEnts for Investnent in Urban Services in the 1980s i Average Annual Potential Average Equivalent Annual Expenditure a/ Amnal Expenditure i Rate of Real Growth 1979/80-1982/83 1984/85-1988/89 (Z) - Rp. billions, constant 1982 prices - Water Supply 82 220 19 Drainage and Sanitation 38 96 13 Yampung Inproveient 19 38 13 Urban Roads (includirg toll roads) 62 100 12 Traffic Mnagement 5 10 17 Public Trawsport 45 125 15 Others b/ 32 42 4 Total 285 529 13 a/ Development expenditure by all levels of goverment. Therefore, these totals do not correspond to central government budget figures. Routine expenditures uthold be over and above these levels. b/ Assuied to grow in line with urban population growth. Sobrce: Indaiesia: Urban Services Sector Report No.4800-IND, April, 1984. Derived from government and World Bank staff estimates. In some cases, e.g. wter supply, financial requirements are derived from physical targets (coverage) endorsed by the central government. - 66 - increase at an average real rate of only 5.5% per year. So drastic has been the combined effect of reduced oil production and lowered oil prices, that even this low growth rate will require major central government initiatives in new domestic taxation and in reduced subsidization. 4. To a degree, it may be appropriate that a higher proportion of central government development be allocated to urban services. A case can be made, for example, that central investment in some services is very low by international standards; in addition, one can argue that public funds in the difficult years ahead should be redirected toward less import-intensive activities, including public service investments. But clearly this shift can only be limited in scope. Under REPELITA III central government expenditures for urban service development accounted for 3.2% of central development funds. Under the new conditions of fiscal stringency, it would probably be difficult to shift that share upward by more than one or two percentage points. If one assumes the share increases to 4.5%, then the central government's role in funding potential urban services investments could fall from two-thirds to one-half of the proposed average annual total. This would leave an important gap that would either be filled by local government or by loan funds accompanied by increased cost recovery. 5. One waV to get the maximum impact out of these central resources is to make use of matching grant formulas.j] These grant formulas should allow central revenues to be funnelled to localities that show initiative in taxing their local resources. Because this would reward tax effort, it would be 1/ The rapid expansion of INPRES grants for local projects, which often require local donation of land; the creation and strengthening of provincial planning authorities; and the introduction of catalyst funding for kampung improvement and solid waste; are all important building blocks for such an approach which are already in place. - 67 - consistent with equity concerns as. well as developmental ones; disadvantaged locations could be rewarded for the! ability to mobilize resources from a small base, even if the absolute amounts collected were small. To meet legitimate central government goals of ensuring that all localities meet certain minimum service standards over time, the grant formulas could reward poorer areas more heavily than wealthier areas, at any given tax effort level. 6. The government should also consider other reforms in the management of intergovernmental relations. Historically, Indonesia's need to promote development after Independence, while relying on relatively few skilled administrators, led to a relatively centralized approach to decision-making. However, this was accompanied by a proliferation of specialized agencies operating with different administrative and financial procedures, so than centralization did not result in a coherent or coordinated approach to dealing with local governments. Challenges facing public administration 1. Indonesia is divided into 26 provinces for administrative purposes. These are further divided into 301 second level units, which are known as regencies (kabupaten) or municipalities (kotamadva). Broadly speaking the 247 regencies are rural in character, while the 54 municipalities are pre- dominantly urban. There are sepaLrate regional administrations at both the provincial and second level of government. Both have their own legislative and administrative arms. The legislative arms consist mainly of locally elected representatives. The executive arms are headed by nominees of the legislature who are appointed by the central government Ministry of Home Affairs. These regional executive arms are in turn organized into a number of local departments (the Linas) which parallel central departments in a number of areas, but are responsible for the provision of local government - 68 - services. Thus most medium and large size cities have their own legislative and administrative bodies. 2. In keeping with the philosophy of Indonesian government, provincial governors and the chief executive of municipalities and regencies exercise a dual role. They are responsible for the administration of local government and are also the local representatives of the President. In this latter role, they are responsible for generally overseeing the work of the central government and for coordinating the effects of central and regional government agencies. This means that the central government departments operating at the regional level are subject to the authority of the regional heads of government ministries. Although this structure in theory permits a distinction to be made between central government and regional government activities, in practical terms the division of responsibilities between central and regional government is far from clear cut. The arrangement has two drawbacks. Firstly, the overlapping of central and local government means that it is difficult to avoid some duplication of effort, inconsistencies and omissions. Secondly, the process of coordination itself can often prove time consuming and inefficient and it necessarily places considerable demands on the limited administrative resources of both central and regional governments. Thirdly, this duplication in responsibility is compounded by fragmentation of funding from central government. Grants for local services and administration are provided from several central ministries. They are usually not related to each other or to local priorities and do not constitute incentives to local governments to plan for efficient and rational services, or to improve local revenues. In all, the system leads to a great deal of confusion, and supresses potential initiative at the local level. - 69 - 3. Central government departments play a dominant role in the provision of public services in Indonesia. Responsibility for technical aspects of urban services is vested largely in various departments of the Ministry of Public Works (Pekeriaan Umun, P.U.). However, the Ministry of Health (Kesehatan) is involved in monitoring and improving the quality of water supplies and sanitation facilities; it also constructs small-scale schemes for water supply and human waste disposal in rural areas. The Ministry of Communications (Perhubungan) is responsible for traffic control and public transport. And street liglhting is provided through the National Electricity Corporation (PLN). Almost all these central government departments have established a set of branch offices at the provincial level and frequently central government sub-branches have been set up in individual cities and towns. These local office are responsible for implementing the development programs of the central departments, as well as administering ongoing programs. There are strong vertical relationships within specific departments which means that coordination between different departments tends to be weak, especially at the local level. This is a particularly serious problem for urban services, because of the need to pursue an integrated approach to the development of densely-populated settlements. 4. As a country of such great size and diversity, Indonesia faces difficult choices in trying to strike a balance in the degree of the centralization of the publlc administration. It is undoubtedly true that as Indonesia develops, more responsibility for the implementation of local services must be devolved to lower levels. However, it is important that central guidance on policy and resource allocation decisions be broadly maintained. Among the steps ithat are consistent with both goals, the following can be mentioned. Across all sectoral programs, efforts should be - 70 - made to reduce and even eliminate overlapping responsibilities for investments at the local level between central line agencies and local authorities. Furthermore, each sectoral agency should consolidate the multiplicity of existing financing sources for local programs. into a single funding channel to be disbursed through one allocation mechanism. In addition, local authorities should create consolidated development budgets which would bring together financing and physical planning information for all projects, thus providing a means for monitoring, evaluation and considering future initiatives. Finally, long-term planning should include multi-year forward, rolling programs which would be prepared to cover all spending for all agencies. Once approved by the central government, then all annual budgets for local spending, regardless of source, would be prepared within the guidelines. All these measures would benefit immeasurably from the speeded-up creation of autonomous municipalities. as recommended earlier in this report.l/ 5. Given these new goals, it is clear that serious attention needs to be devoted to training needs and manpower dzvelopment at the local level, accompanied by systematic, though selective, delegation of authority. As background to this exercise, three general points need to be kept in mind. First, technical and managerial staff at the central level are less constrained by skill deficiencies than by size. As the REPELITA IV investment programs are put in place, the chief role of the central agencies must increasingly be to advise and guide local implementation rather than to design and approve contracts. Second, the technical staff at the kotamadva/kabupaten level is adequate in size, but deficient in skills. These staff must be the Ij In Jakarta and Surabaya, three-year rolling plans are being introduced. This system will be extended to other large cities over the next two years. - 71 - primary target for skill upgrading programs. Finally, the local planning and coordinating units will face expanded responsibilities over time; though their skill mix and size are adequate at present, their training needs should be periodically revaluated as devolution continues. 6. Recognizing these factors, the central government has already embarked on a major expansion of training and manpower development programs. The Ministries of Public Works and Home Affairs both have large programs, and the Ministries of Finance, Commutnications and Planning also have organized courses. The challenge now is to expand these to cover neglected sectors (eg. sanitation programs), while strengthening the content of such courses. In addition, these initiatives must be accompanied by visible progress in delegation of authority.. With the support from provincial agencies, the second level governments already have a substantial technical capacity to design and supervise smaller projects. There is greater variability in local managerial capacity. Even here, however, it is possible to devolve certain routine duties to local. administrators while assigning even greater responsibilities to stronger provinces and, in turn, to stronger local governments. This approach will permit central and provincial manpower to concentrate their energies on strategic planning and on assisting the weakest local agencies. Alternative sources to finance urban improvements 1. There are three alternative sources of finance that could and should increase in relative importance: local taxation, borrowing by local authorities, and cost recovery. The fact that these alternatives are being considered now, because of the financial constraints being faced by the central government, should not obscure the value of more decentralized financing and planning under any scenario. Local economic development is a - 72 - function, in part, of increased local government initiative and of the ability, at the local level, to facilitate the expansion of the economic base through timely, coordinated, and strategic interventions in the provision of public services. 2. A careful program of devolving of planning and financing authority to the local level would also help the public sector to solve the critical problems of choosing which areas will receive what investments. The central government does not have the knowledge necessary to identify priorities and execute programs in a coordinated fashion across hundreds of urban areas. Instead, with appropriate arrangements dealing with technical assistance and financial aid, localities should be encouraged to rely more heavily on local taxes and user charges, and compete among each other for borrowable funds. This will tend to minimize extravagance and planned investments that fail to conform to local conditions and local economic prospects. 3. At present local government spending is based on central grants, assigned taxes and royalties levied by the center for local use, direct revenues, and loans. While the distribution of revenue sources varies across local governments, the largest contribution, roughly 65 percent, comes from central grants. Local tax revenues and charges account for another 25 percent, while borrowings and assigned revenues (mostly the property tax) each add another 5 percent. Non-grant local revenues have the greatest potential for expansion over the medium-term, especially since they only amount to an average of about 2.5% of personal per capita income across Indonesia. This low average level of effort is associated with remarkable variations in provincial and district level revenue-raising per capita. Excluding obvious cases explained away by the extraordinary high or low level of local economic activity, most variation is hard to understand. It suggests that one way to - 73 - encourage greater local taxation i,s to develop incentive systems that link an increasing proportion of central grants or loans to local tax effort. One can demonstrate quite easily that there is considerable scope for such improved local performance. 4. All the recommendations that follow assume increased local ability to set tax rates and charges for existing. taxes; and to introduce new taxes. In addition it is assumed that incentives will be in place to identify relevant taxpayers, to correctly assess their individual liabilities, and to efficiently collect the assessments. At present the incentives for local effort are limited, as are the options for tax reform. All possible taxes and charges are specified in national legislation. These laws describe the nature of taxes and charges; while rates are specified by local legislation. Local tax legislation is subject to approval by the Ministry of Home Affairs, in consultation with the Ministry of Finance in cases where a tax is being introduced for the first time. Local service charges need to be approved by the provincial governors. In practice, most important tax and charge rates are standardized nationally and are, effectively, national taxes, locally collected and assigned to i:he regions. 5. Using data on the kotarnadva as examples of urban local revenue sources, one finds that 45% of average receipts are composed of taxes, with the remainder being charges. Of the taxes raised, on average half are Ipeda land and property taxes.l/ Th4s Ipeda property tax is the most widely collected direct tax in Indonesia, and its rolls include 3.5 million registered urban payers. Unfortunately this type of tax is poorly j/ Since DKI Jakarta is a province, the proportions there are different; taxes account for the bulk (83%) of per capita revenues. Property tax yields, in turn, are equal to only 9% of local tax receipts. - 74 - exploited. As a result this type of taxation, which provides the most important source of revenue (20-50 percent) for most developed cities in the world, accounts for only 5 or 10 percent of total resources for Indonesia's cities. There are various reasons for the poor performance of the property tax. First, the tax rates are very low, both absolutely and relative to rates prevailing in other countries. The tax is assessed at a modest 0.18% of the market value of commercial land and 0.09% for residential land; buildings are taxed at even lower rates. Second, the tax rolls are incomplete; a severe problem of underregistration appears to exist. Third, there are a large number of low assessments the collection of which is not cost-effective. Fourth, there is evidence that higher-value urban properties are deliberately underassessed. Finally, local authorities have only a limited role to play in generating what are essentially central revenues assigned to localities. As a result, the tax effort varies greatly by city, for reasons that are unclear. In addition, the growth per capita in urban Ipeda yields are below recent GDP growth rates, when the opposite should be the case. 6. There seem to be considerable room for reform. The setting of rates, the determination of assessments, and the collection of taxes should be made local in character. The central government should confine itself to creating incentives for higher tax effort, and to providing technical assistance to improve the quality and coverage of the local property cadasters. Under a reformed system, tax rates should be raised; the scope for increased tax effort is illustrated by the fact that many cities in developing countries have tax levels of 1% of market value. A ten-fold increase in the property tax efforts would be a major undertaking, but one which could be achieved over a ten-year period. During that period, tax rates could be raised, assessments - 75 - could be brought closer in line with market valuations, and collection efforts could be redirected to more affluent taxpayers. By the early 1990s, receipts from the urban property tax could finance about a third of all new investments in urban services, and could, almost singlehandedly, compensate for any declining relative contribution from the central government. 7. Other sources of local revenues include vehicle taxes, particularly the annual license fees and vehicle transfer fees collected by provinces. There are, in addition, poorly exploited provincial gasoline sales surcharges and kotamadya parking charges. The license and transfer fees contribute 80% of provincial tax receipts; they are particularly important to DKI Jakarta where 19% of all registrations and 70% of all vehicle transfers occur.i1 Reliance on this source of revenues serves a dual purpose of diversifying the local tax base, earmarking funds for local road improvements, and discouraging the rate of increase in vehicular traffic. The yield from taxes could be increased by dropping current exemptions applicable to government vehicles and by raising fees for trucks and buses to levels comparable to those applicable to cars. 8. At present the provincial surcharge on gasoline yields no more than 3% of provincial tax receipts and amounts to only Rp. 1.05 per liter. The same rationale presented above could be used to justify higher surcharges; raising the rate to Rp 10 per liter would yield enough revenues to finance the increase in the projected annual urban road investments for REPELITA IV. Together with use of a portion of the license and transfer fees, this would allow all road investment programs to be easily financed. jj The license fee is equivalent to 2-3% of market value. The transfer tax is 10% of assessed value at the time of the first transfer and 5% thereafter. - 76 - 9. Other fiscal measures could also be contemplated which would help to discourage vehicle use, especially in congested city centers. This course of action would be consistent with the philosophy that it is easier to encourage public transportation by penalizing private vehicle use than it is to provide purely positive incentives for public vehicle use. One simple measure would be to raise parking charges in larger cities. Currently these vary from Rp. 50 to Rp. 200 (Jakarta only) and are invariant with length of stay. These rates, levied as they are on the affluent, could surely be raised to conform with charges typical in other international centers, eventually reaching levels of Rp 1000 - Rp 2000. This would provide an attractive substitute for the more difficult scheme of licensing vehicular access to city centers on a daily, monthly, or yearly basis. 10. Less is known about other existing taxes at the local level, especially the reasons for marked variations in receipts per head and in growth rates over time. These taxes are assessed primarily on entertainment, recreation, and economic activity. It would be useful to undertake a set of studies on the procedures involved in assessing, collecting, and accounting for local revenues with a view to improving the efficiency of assessments and collections, and of reducing the costs of collections. In some cases some taxes might prove to yield negative net revenues and be candidates for- elimination (e.g. taxes on alcoholic drinks, radios, dogs, and non-motorized vehicles). In other cases, new taxes and/or improved collection procedures would suggest themselves. While many uncertainties exist concerning these miscellaneous taxes, the associated receipts per capita are low by international standards. Special consideration should be given to unexploited sources of revenues, such as that which could be derived from extending present kotamadya/ kabupaten tax on economic activity from restaurants and - 77 - hotels to retailing and manufacturing. Even at a modest rate, such a tax could yield revenues sufficient to finance a significant fraction of new urban services investments over the nexi: decade. Local borrowing as a tool 1. In global terms, borrowing by local governments is presently unimportant. It accounts for only 1%-1.5% of local government spending, and 2%-3% of kotamadya/kabupaten spending, in particular. Most loans, however, are directed at municipal servicies, and may account for up to 20% of water supply and kampung improvement activities. Such loans do not yet play an appropriate role as a rationing device. Most funds are borrowed at subsidized or zero interest rates and appear to be made without much relation to local resource availability. It would seem desirable that wealthier local communities be required to finance a higher proportion of their development from loans rather than grants, foLlowing the example of Jakarta and Bandung in water supply investments. 2. Lending to local authorities has important advantages over grant finance. First, it encourages imcreased cost recovery for services, as well as greater tax effort, especiaLly when debt service payments are linked directly to charges for the services improved by the proceeds of the loan. Second, the use of loans will help reduce the burden on the central government budget, by generating new sources of funding in the form of debt repayment with interest and in the form of bank loans. 3. Third, lending to local authorities provides an indication of relative needs and priorities at the local level, and thus encourages better resource allocation. Finally, It provides an incentive to control costs, improve accounting standards, and maintain and operate the system efficiently. In fact, consideraltion should be given to creating a municipal - 78 - loan board or of expanding the role of existing regional development banks in this field. While initially the bulk of funds would come from central government loans, grants, and foreign aid allocations, in the long-run attractive yields might draw new sources of funds, including institutional savers (pension funds, insurance companies). With this type of system in place, and realistic cost recovery policies, it will be possible to encourage urban service improvements across the hierarchy of cities, without fearing that funds made available to large cities will consist of subsidies provided at the expense of deserving populations in other centers. 4. Ideally, such a financial institution should be linked to the entire planning and budgeting process. Three-year rolling expenditure plans could be prepared at the local level. These could be appraised at the provincial level. At the central level, a municipal development board consisting of relevant Directors General from the Ministries of Home Affairs, Finance, Public Works, Communications and Planning, could review the proposals and authorize financial packages consisting of grants and foreign or domestic loans. The loan components could be channelled through holding banks (perhaps the regional development banks) to the local authorities. A revolving fund would accumulate at the central level, permitting government grants to be reduced gradually. This approach would take time to implement. In the interim, preparatory work is needed to a) develop policies and procedures for allocating such grants and loans; b) provide guidelines to local governments for developing multi-year investment and revenue programs; c) define the criteria and methods used by provinces in appraisal; and d) provide a framework to coordinate external aid for urban development. 5. With few exceptions there is virtually no direct cost recovery for capital investments in urban services in Indonesia. The general policy has - 79 - been that recurrent expenditures should be recovered through user charges, where possible, but that the initial capital investment should be provided to the users in the form of grants. This policy has been based on arguments of affordability, fairness, and externalities. The poor are deemed incapable of paying their way; charging poorer communities for improvements earlier provided freely to more affluent neighborhoods is considered unfair; finally, some services are assumed to benefit the community at large and, it is argued, should be paid by all. Unfortunately, in the present environment of fiscal stringency, communities face the choice between highly subsidized services for the few and widespread services based on greater levels of cost recovery. This can be illustrated for particular sectors. 6. At present, the policy of the central government with respect to water tariffs is that they should cover operation, maintenance and depreciation. Within those guidelines, local water authorities have widely different tariff structures and installation charges, some of which encourage waste (e.g. when unit charges are constant irrespective of consumption), some of which discourage access to the system (e.g. when house connections exceed the actual costs of installation by a wide margin). The government is reviewing the degree of regressivity existing in such charges, with the intent of improving the progressivity of the tariffs and of decreasing the practice of excessive connection fees. Bwt the principal issue is how to finance the potential increase in urban water supply investments from Rp. 82 billion to Rp. 220 billion per year (1982 prices). The present policy of user charges, when applied to forcasted investment programs, will recover less than 20X of annual investments by 1991. By contrast a policy based on replacing grants with 20 year, 4% per annum loans would generate, in debt repayments alone, enough to finance nearly 40X of annual investments by 1991. The effects would - 80 - not be immediately apparent in the first few years, but it would be necessary to implement the system changes immediately in order to reap large advantages within a decade. 7. Planned carefully, such improved cost recovery should not unduly burden the poor. Evidence from other countries indicates that consumers are willing to pay about 4% for their income for clean water on a regular basis, and up to roughly 7% when they are paying off debts for connection charges.l/ Low income households are already paying much higher rates when they rely on water sold by street vendors. These proportions can be preserved for urban consumers with house connections. An appropriate pricing policy also needs to be developed for the poorest 20% of the population that cannot afford house connection. A judicious combination of grants for the provision of standpipes and tariffs paid by a small group of families responsible for individual units would allow recovery of operating costs and of some capital expenses. 8. The same general principles can be applied to sanitation services. Under an existing pit latrine program, beneficiaries contribute to investment costs through the provision of labor, while maintaining the latrines after construction. This arrangement is not excessively burdensome to the central government and allows mass replication from a financial standpoint. Another program provides for the construction of communal wash-house/toilets. These are presently financed by a grant except for the land, which must be donated. Operational cost recovery is not an issue, but the present method of pricing has drawbacks, being based on per use fees collected by a paid 1/ Survey undertaken in selected East Java cities in preparation for the World Bank's First East Java Water Supply Project suggest these guidelines could be met in Indonesia while covering all operation and maintenance costs plus depreciation. - 81 - collector/cleaner; this limits use and increases expenses. If such facilities were assigned to small groups of families, who would pay a fixed monthly fee for unlimited use, it would be possible to recover most capital costs in addition to operating and maintenance expenses. For example, for roughly Rp 2000 per family per month, 5-7 families can pay for water charges, pay for annual emptying of the pit, and repay an interest free loan in four years. Only the most destitute households would require other solutions, such as larger, multiple-unit communal facilities. 9. Sewerage system coverage is very limited in Indonesia; however, the upcoming Jakarta sewerage project provides an opportunity to consider cost recovery options. The government is committed to recovering operation and maintenance costs plus depreciation or debt service, whichever is greater. The plan envisions substantial cross-subsidization, with non-residential customers paying considerably more than residential ones; this will keep average household charge to about Rp. 800 per month (1982 prices), equivalent to 1.4% of the household income at the 50th percentile. Such a program is not replicable where non-residential. properties are proportionately fewer. Extensive cost recovery, which is indisputably necessary where large-scale extensions are contemplated, may limit sewerage system investment to pockets of larger cities. 10. Responsibility for house-ito-house collection of garbage lies with the neighborhood association. Charges are levied and are designed to cover labor costs of collection and transport to the transfer station or to the local dump site. On the average Rp5O - Rp200 is collected per month from each household benefitted from the system. Capital investments, such as transfer stations, and even hand-carts for house-to-house collections, have usually been provided as a grant from the central and local governments. A major expansion of the - 82 - solid waste program is required and the government is encouraging neighborhood solutions to hold down costs. A gradual increase in charges could make an important contribution to refinancing the expanded program. In addition, there is scope for increased cross-subsidy from high income residential and commercial (markets) areas to poorer areas. 11. An additional area for close scrutiny involves public transportation. Bus fares are controlled for each type of service, and government owned systems in Jakarta and eight other large cities are explicitly subsidized. Publically-owned companies receive Rp. 30 billion a year in capital grants, as well as subsidies to cover annual operating losses and to purchase diesel fuel at below international market prices.jj This is unfortunate from various perspectives. There is little evidence that such subsidies limit the diversion of passengers to private vehicles. Furthermore the controlled fares system has done widespread damage to private bus companies and obstructed the development of higher-quality, higher priced microbus service; this is likely to increase congestion by reducing public transit service. Finally there is little reason to subsidize a "basic need" for middle and lower income individuals who happen to live in a handful of large cities. It is precisely by placing the burdens of congestion on all users of roads -- private and public vehicles -- that a balance is reached- between the attractive features of large centers and its costs. The doubling of fares on large city buses in January 1983 was an important contribution to resolving the issues in this field, and points the way to full cost recovery at some point in the future. 1/ Public transport is also assisted by generous access to subsidized credit through the state banking system. - 83 - 12. There is one final candidate for cost recovery through user charges: neighborhood improvements whose benefits cannot be allocated to each household by charging specific tariffs. 'Included among these are road surfacing, drainage, footpaths, and water mains. These investments do have one feature that makes cost recovery possible: they tend to increase land prices in the adjoining areas. In many parts of the world betterment or "valorization" taxes are able to recoup most or all of the investment costs. 13. Jakarta has the only example of a betterment tax in Indonesia. It is being used in particular areas of the city on an experimental basis; it allows for the recovery up to 60X of tlhe cost of new or improved infrastructure through the use of a special tax. The recoverable cost is divided between beneficiaries according to the length of road frontage, with properties directly fronting onto the road paying more than those only indirectly connected. The tax is calculated and announced before improvements are made; it must be paid within threse years after completion of the work, with financial penalties for late payment and denial of building permits until the tax is paid. The tax has been fairly successful in recovering costs where applied. The special tax could be extended both within Jakarta and other cities. If project standards are kept moderate, construction takes place in stages, and the major burden is placed on middle and upper income beneficiaries, the taxes should be affordable to all income groups even if the proportion of cost recovery is increased substantially above 60%. Before the program is extended it is important to review problems connected with unregistered property owners and lack of data on land values. Aggressive action in this regard would have the welcome side-effect of improving property tax collections, in general. - 84 - 14. Existing legislation in Jakarta also allows for the collection of an "excess" profits tax tied to land value increases in excess of 300% over the first two years following the completion of public works; it would be collected on change of ownership, and would thus be relatively painless. Though this section of the special tax was never put into effect, it would represent a useful way to recover additional costs while preventing disproportionate windfalls gains from accruing to landowners. It deserves closer scrutiny. It would be useful to consider, for example, lowering the threshold to 200% and increasing the period of application. In addition, to avoid double taxation for betterment, landowners should be permitted to deduct taxes already paid from the excess value surcharge. Financing urban housing services 1. Once resources are found to provide lower income groups with basic services, there is a need to consider shelter issues. Residential investment is tied to infrastructure investment and together they play a major role in determining the level of internal efficiency and productivity of cities. Looking across cities, one can argue that an important ingredient of national spatial policy is the existence of a reasonably smooth mechanism that will generate shelter options in the quantity, quality, and locations demanded by a shifting and growing labor force. 2. In Indonesia, 150,000-200,000 new urban households seek housing each year, be it shared, rented, or owned. An unknown fraction of existing households seek to change the characteristics of their shelter through moving or upgrading of existing units. Though information is scanty, it is clear that most urban housing in Indonesia is built by individual households, using incremental or staged construction, and ignoring the formalities associated with building construction permits and orderly land registration. Financial - 85 - institutions play virtually no role in this "informal" market. In fact the major outside catalyst is provide-d by the provision of basic infrastructure through programs like the Kampung Improvement Program. Direct public sector initiatives are largely confined to financing formal or licensed housing. The public housing agency, PERUMNAS is producing fewer than 18,000 housing solutions per year, roughly half' of which are serviced sites and half are units of varying quality and size. In addition, through the housing finance institution, Bank Tabugan Negara (BTN), the government allocates public resources largely obtained through liquidity credits from Bank Indonesia to finance private construction of about 32,000 units yearly. All told, about one-quarter to one-third of new housing construction in urban areas involves government financing. This resem'bles conditions in other East Asian countries such as the Philippines and Thailand.1/ Except for a largely symbolic effort by a new private mortgage company, the only additional institutional financing available in these cities come through the efforts of some employers who help their workers, and of insurance companies who lend to policyholders. There is no formal assistance for home upgrading or for financing the transfer of existing units. 3. Not only is formal assistance limited, but it tends to be inequitably distributed. In fact, the bullk of the private units financed by BTN are affordable to only the top quintile of the household the households in urban areas. The smaller PERUMNAS program, because of its reliance or "sites and services," and "core" units, is far more accessible; eighty percent of the housing solutions financed are valued at less than Rp. 3 million and are accessible to households with incomes at the median or less. The average jJ See B. Renaud, Housing and FLnancial Institutions in Developing Countries. Urban Development Department Report No. 19, The World Bank, December 1982. - 86 - PERUMNAS solution is affordable to households with incomes low enough to place them in the lowest quintile of the urban household income distribution. The equity issue is all the more troublesome because both PERUMNAS and assisted private construction are heavily subsidized. Public housing solutions are provided through mortgages carrying interest rates of 5 to 9 percent, while private housing is financed at 9 percent. By contrast, mobilizing resources from other parties would force BTN to pay rates as high as 16 percent.jJ Since the bulk of the financing goes to private construction activity, it is not surprising that 60 percent of the implicit interest rate subsidy accrues to households in the upper two deciles of the income distribution. 4. The challenge facing policymakers is how to devise a strategy to increase housing production in Indonesia's cities so as to benefit households at all income levels while decreasing the role of the public treasury, given macroeconomic realities and the critical needs of the rural sector. The rewards for developing a successful approach are great. Not only is a basic need met in the process but a dynamic housing construction "industry" generated significant number of low and medium skilled jobs while requiring only limited amounts of foreign exchange. 5. The government has set ambitious targets for itself. This is evident in the goals established for REPELITA IV; with an estimated 350,000 units scheduled for financing, this would more than double the effort expended during REPELITA III. While this would still be modest in terms of overall demand, attempts to service a more significant share would be difficult. This is particularly so because during REPELITA III the bulk of resources used to 1j During the period 1976-1983, the direct budget subsidy for public housing units was Rpl20 billion, and the interest rate subsidy for private units exceeded Rpl50 billion, or equivalent to more than one-third of the amount lent for the latter units. - 87 - finance public and private initiatives were either outright grants (34 percent) or Bank Indonesia funds (60 percent) lent at a nominal rate of 3 percent. The weighted average coEst of funds was therefore below 3 percent, a situation unlikely to persist in the future give the macroeconomic realities. Significantly larger resources must be mobilized for housing with far less reliance on grants or heavily subsidized funds from Bank Indonesia. In fact, World Bank estimates suggest that a fairly optimistic scenario would still require that more than half of the new resources come from institutions (pension funds, insurance companJies) and holdexs of new savings instruments receiving rates of interest in the 15-16 percbnt range. As a result the average weighted cost of funds wrould more than triple to about 10 percent. All borrowers will have to pay more for borrowed funds, with rates varying from levels somewhat above the average weighted cost of funds to the prevailing market lending rates (i.e. 12-18%). Additional resources could be obtained by raising the prese!ntly very low down payment requirements particularly for higher priced tnits. For example, down payments could be increased from the current 5-10% to 10-30% depending on the price of the unit. With this overall approach in place, there would be little need to depend on the Ministry of Finance or Bank Indonesia for subsidies. 6. World Bank estimates suggest that, even with higher interest rates, it will be possible to provide affordable housing for the various segments of the population with incomes above the first two deciles. The poorest of the poor would benefit as well, through the greater availability of rental rooms that characteristically accompany the expansion of housing construction outside the wealthiest neighborhoods. The remaining challenge will be to provide incentives for, and to remove obstacles in the way of, developers who wish to construct units for the lower half of the income distribution. Here, - 88 - efforts are needed to remove regulatory barriers to private construction of core units with less than 36 square meters. In addition private access to mortgage and construction finance should be made conditional on bona fide efforts to build a range of housing, with special emphasis on units affordable to urban households in the two deciles bracketing the median income levels. Ceilings on the size of individual loans may be necessary to guarantee that the wealthiest households do not receive a disproportionate share of the available funds. 7. Beyond that, more imaginative solutions must be sought. As noted, one can assume that in Indonesia's cities more than half of all new housing accommodations (including rental of rooms newly added) is provided for by the so-called informal market, especially in the Kampung areas. Given the continuation of vigorous Kampung Improvement Programs, which will provide various public services to the sites, there is a critical need to mobilize and relend funds for housing upgrading. While there are a few precedents in this area 1/, consideration should be given to schemes permitting financial institutions to provide community lines of credit. This would allow legally constituted community organizations to borrow funds for housing improvements by its members. This system would significantly multiply the number of beneficiaries receiving small loans, primarily for construction materials and some skilled labor help. At the same time some of the paper work associated with loan disbursements and collections could be passed down to the community organizations. While there are risks associated with involving community groups in this way, the idea has enough merit to warrant testing on a pilot basis. 1/ One country that has experimented with various pilot home improvement programs is Colombia. - 89 - 8. Much of the housing problem, especially for lower income groups, does not reside in the lack of housing units, but in the lack of serviced land available for housing. Some of this deficit can be addressed by land purchase and development of serviced sites by a public housing corporation (such as PERUMNAS). However this approach is time consuming, places an additional burden on the government budget, and has sometimes led to the selection of inappropriate sites, which have been difficult to sell. An innovative approach which would complement that of the public housing corporation and is now being tried in Jakarta and Denpasar is "Guided Land Development." This, in essense, is a simplified form of land readjustment whereby the government provides coordinated infrast:ructure to growth corridors at the city perimeter and establishes agreements with groups of landowners for the delineation of public space for minor rights of way and community services. Through these arrangements and by providing assistance to landowner associations on simple subdivision layout, large volumes of minimally-serviced land can be released onto the market through private sale. A betterment tax on the developed land both encourages landowners to se11, and recoups the costs of most of the local services. 9. These various initiatives, taken together, will begin to address the varying demands of households that share rooms, rent units, and own homes. Actions taken to promote informal housing are particularly beneficial because experience elsewhere demonstrates that among households in the lowest half of the income distribution, it is informal sector that provides most affordable housing at many different locations for owners. renters, and roomers while economizing on scarce national administrative and financial resources.A/ Ij B. Renaud, op. cit.; WeP. Strassman, The Transformation of Urban Housing. Baltimore, Johns Hopkins Press, 1982; A. Hamer, Bogota's Unregulated Subdivisions: The Myths and Realities of Incremental Housing Construction. World Bank Staff Working Paper No. 734, May, 1985. The World Bank Headquarters European Office Tokyo Office U 1818 H Street, N.W. 66, avenue d'Iena Kokusai Building Washington, D.C. 20433, U.S.A. 75116 Paris, France 1-1 Marunouchi 3-chome Telephone: (202) 477-1234 Telephone: (1) 47.23.54.21 Chiyoda-ku, Tokyo 100, Japan Telex: WUI 64145 WORLDBANK Telex: 842-620628 Telephone: (03) 214-5001 RCA 248423 WORLDBK Telex: 781-26838 Cable Address: INTBAFRAD WASHINGTONDC ISSN-'-3-2115/ISBN 0-8213-0688-X