China - Long-term issues and options (Vol. 1 of 7) : The main report

Report No. 526-CHa China Long-Term Issues and Options The Main Report May 22, 1985 East Asia and Pacific Regional Office FOR OFFICIAL USE ONLY ,. - . , .= Documeiet of the World Bank This report has a restricted distribution and may be used by recipients onl in the perforrnanceof their official duties. its contents may not ot-herwise Abe dlisclosed without World Bank authorization. ,~~~~~~~~~~~. . ~.,,, < _ g  CURRENCY EQUIVALENTS The Chinese currency is called Renminbi (RMB). It is denominated in Yuan (Y). Each Yuan is 1 Yuan = 10 jiao = 100 fen In early 1984 the official exchange rate of the Yuan to the US dollar was around Y 2 = US$1. The internal settlement rate (ISR) of Y 2.8 = Si, however, was used in most merchandise transactions. The officiaL exchange rate is now about Y 2.8 = $1. On January 1, 1985, the Government abolished the ISR. WEIGHTS AND MEASURES Chinese statistics are usually in metric units; in addition, mu and Jin are often used: 1 mu = 0.1647 acres = 0.0667 hectares 1 jin = 0.5 kg FISCAL YEAR January 1 - December 31 TRANSLITERATION The Pinyin svstem is used in this report. TERMI NOLOGY (1) With the recent reorganization of rural administration, "communes" have been replaced by "townships," and "production brigades and teams' by "villages." This report retains the former terminology in one respect, however, by referring to "commune and brigade enterprises." (2) The term "national income" is used in this report to encompass both the Chinese measure (net material product) and the Western measure (gross national product). Where the context makes the distinction between these two measures important, they are more precisely identified. Note: In tables, individual items may not sum exactly to totals because of rounding errors.  FOR OFFICIAL USE ONLY Preface The World Bank initiated its economic work on China in 1980 with a review of China's achievements over the previous three decades and of the main economic and sectoral deveLopment issues, published as China: Socialist Eco- nomic Development. Since then. the Bank has sought to improve its understand- ing of China through both project work and economic and sector studies. The latter have included China: The Health Sector (1984), and a general review of recent economic trends and poLicy developments (Report No. 4072-CHA), as well as studies on industrial finance, the rural financial system, urban develop- ment in Shanghai, and agricultural pricing. Research focusing on enterprise guidance and on structural change is also being undertaken in collaboration with the Economic Research institute of the Chinese Academy of Social Sciences. The present report, which draws on this earlier work, looks at some of the issues and options that China will face over the next two decades. It was prepared principally by members of an economic mission that visited China twice in 1984, for four weeks in February/March and for five weeks in April/ May. In addition to Beijing, the mission went to three provinces: one coastal and relatively high-income (Jiangsu); one inland and average-income (Hubei); and one interior and low-income (Gansu). It received a lot of infor- mation, as well as numerous valuable comments and suggestions, from officials and others in these provinces, as well as from those in many central agencies and institutions, including: the State Planning and State Economic Commis- sions; the Ministries of Finance, Agriculture, Coal, Communications, Educa- tion, Foreign Economic Relations and Trade, Labor and Personnel, Petroleum, Railways, Urban and Rural Construction, and Water Resources and Electric Power; the State Statistical Bureau; and various universities and research institutes of the Chinese Academy of Social Sciences. A series of seminars was organized by the Technical-Economic Research Center under the State Council. The generous and thoughtful assistance of all these people in China contributed greatly to the preparation of this report. The Bank mission was led by Edwin Lim (mission chief) and Adrian Wood (deputy mission chief), and also consisted of William Byrd (econo- mist), Mats Hultin (senior education adviser), Erh-Cheng Hwa (senior econo- mist), Timothy King (senior economist), Jacques Yenny (senior transport econo- mist), Umnuay Sae-Hau (research assistant), Betty Ting (interpreter), Luc De Wulf (senior economist, International Monetary Fund), Benjamin King (consul- tant on statistics), Wouter Tims (consultant on planning and agriculture); and the following teams: Agriculture: J. Goering (team leader, April/May), Tom Wiens (team leader, February/March), Lang-Seng Tay (irrigation specialist), Lo-Chai Chen (fishery consultant) and Fred Bentley (consultant on arid agricul- ture); Energy: Roberto Bentjerodt (senior economist, coal projects), Weigong Cao (power engineer), Abdel El-Mekkawy (engineer, petroleum proj- ects), Robert Taylor (energy economist), and Darrel Fallen-Bailey (con- sultant); D.C. Rao (Assistant Director, Energy Department) led the team in the field; IThis document has a restricted distribution and may be used by recipients only in the performance of | their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  - ii - Industrial Technology: Gene Tidrick (team leader), Anupam Khanna (indus- trial economist), Reza Amin (industrial specialist), and Josephine Woo (research assistant); Location and Trade: Ian Porter (team leader), Vernon Henderson (consul- tant on urbanization), John Sheahan (consultant on industrial location and trade) and SamueL Ho (consultant on rural nonfarm activities). The following also contributed to the preparation of the report: Wlodzimierz Brus (consultant an socialist economies), Gerhard Pohl (energy and transport), Robert Drysdale (Annex A); Helena Ribe, Nikhil Desai (Annex C); Shujiro Urata (Annexes D and E); and Lily Uy (Annex F). Larry Westphal, Carl Dahlmzan and Bruce Ross-Larson organized background work on technology. Behrouz Cuerawi-N, Tejaswi Raparla, and Kong-Yam Tan helped with the multisec- toral model, the input-output table and data for international comparisons. Ann Orr, Kenneth HiLl, Moshe Syrquin, J.V.S. Sarma, Kenneth Cochran, Chang Hsin, Liu Ying and Cai Jinyong undertook research. Linda Mitchell and Terrice Bassler edited the report; and Helen Kung assisted in its processing. The report also benefited from comments of a review panel consisting of Anne 0. Krueger, Luis de Azcarate, Kemal Dervis, Janos Kornai (consultant) and managers of the East Asia and Pacific Regional Office. The report is in seven volumes, a Main Report and the following annexes: Annex A: Issues and Prospects in Education Annex B: Agriculture to the Year 2000 Annex C: Energy Annex D: Model and Projections Annex E: China's Economic Structure in International Perspective Annex F: Transport In addition, the following background papers have been prepared in connection with the work on the report and are available from the China Division: I. "Rural Nonagricultural Development: The Asian Experience and Prospects for China" 2. "'Urbanization: International Experience and Prospects for Chinag 3. "'Alternative International Economic Strategies and their Relevance for China" 4. "International Experience in Budgetary Trends during Economic Development and their Relevance for China" 5. "p'roductivity Growth and Technological Change in Chinese Industry" 6. "Technological Development in Electronics" 7. "The Environment for Technological Change in Centrally Planned Economies" 8. "managing Technology Development: Lessons from Developing countries" 9. "crowth and Structural Change in Large Low-Income Countries"  Table of Contents Page No SUMMARY AND CONCLUSIONS ......................... i-xxxiv 1. GROWTH AND CHANGE A. Objectives ............... , . .......... 1 B. Macroeconomic Change. 3 China's Present Economic Structure in International Perspective. 3 PossibLe Implications for China's Future Development 8 C. Microeconomic Change .12 D. Inequality .16 InequaLity in China in International Perspective 16 Possible Future Trends in Inequality .19 2. ILLUSTRATIVE PROJECTIONS A. Description of the Model ..22 Growth ..................... ......................---... 23 Structure ..23 B. Three Alternative Projections ..25 QUADRUPLE ..25 MODERATE . .26 BANCE ..26 C. Growth and Demand ..28 D. Production and Investment .. 32 E_ Employment and Income ..36 F. Projections and Predictions ..42 3. AGRICULTURAL PROSPECTS AND POLICIES A. Production Possibilities ..43 Prospects for Major Crops ..45 Prospects for Livestock . .45 B. Potential Demand-Supply Imbalances . . 46 Alternative Outcomes in 2000 . .47 Changing Imbalances over Time . .49 Product Mix within Agriculture . .51 C. Changing Food DeAnd ..53 Alternative Food Consumption Patterns . .55 Managing Food Demand ..58 D. Managing Household Agriculture ..59 Output Pricing and Marketing 59 Input Pricing and Allocation . .61 Areas for Increased State and Collective Activity 62 E. Agricultural Incomes Policy .. .63  -ii - Page No 4. ENERGY DEVELOPMENT A. Demand and Supply: Trends and Balances . . 66 Energy Intensity .................... 67 Demand and Supply Prospects .. 67 B. Coal Production, Transport and Utilization . . 69 Coal Mining .. 70 Coal Transport .. 71 Coal Utilization .. 73 EnvironmentaL Issues .. 76 C. Oil and Gas .............................................. 77 Oil Production .. 77 Natural Gas .................. ...................... 78 Petroleum Refining .................. 79 D. Electric Power .. 81 Demand Projections .. 81 Investment Planning .. 81 E. Rural Energy and Biomass .. 83 F. Investment, Planning and Prices .. 85 Investment Requirements .. 85 Planning and Coordination .. 87 Technology Transfer and Development . . 88 Demand Management and Pricing . . 89 5. SPATIAL ISSUES A. Specialization and Trade .. 91 Agricultural Specialization .. 97 Industrial Specialization and Location . . 99 B. Transport, Commerce and other Services . .......... 106 Transport Needs ..................... 106 Developing Commerce ..1................ 11 C. Urban and Rural Development ................ 113 Urban-Rural Balance ................... 116 Development of Towns and Cities ............. 120 D. Rural Poverty and Regional Inequality . .......... 122 6. INTERNATIONAL ECONOMIC STRATEGY A. China in the World Economy ................. 131 External Environment ................... 133 Benefits and Costs of External Economic Contacts . .... 134 Managing an Open Economy ................ 134 B. Foreign Trade Prospects and PoLicies ............ 139 Primary Products ..................... 141 Manufactures ....................... 142 Services ...... 149 C. External Balance and Finance ................ 150 Structure of Capital Flows .................... . ........ 153 D. Regional Issues .......................................... 154  - iii - Page No 7. MANAGING INDUSTRIAL TECHNOLOGY A. Technology in Perspective . . . 156 Other Determinants of Efficiency and Quality ... 158 Economically Efficient Technology . . . 159 B. Planning for Technological Development . . . 161 C. The Make or Buy Decision . ..................166............. 166 Importing Technology ................................... 166 Research and Development . . . 168 D. Incentives for Innovation and Diffusion . . ................ 170 Diffusion of Technology . . ...................... 171 The Role of Small Enterprises .. . 174 E. Obsolescence and Scrapping . . . 174 8. HUMAN DEVELOPMENT A. Education and Training Issues ............................ 178 Basic Educational Priorities ........................... 180 Specialized versus General Education ................... 182 The Role of Manpower Planning .......................... 184 B. Employment and Productivity .............................. 185 Productivity and Growth ................................ 186 Urban Employment ....................................... 186 Underucilized Rural Labor .............................. 189 C. Allocating and Motivating Workers ........................ 191 Employment Choices of Enterprises and Workers .......... 191 Efficiency, Fairness and Stability in Wage Determination ........................................ 196 D. Population Issues Beyond the Year 200 .201 Changes in the Age Structure .205 Birth Planning Policy .209 Supporting the Elderly .210 9. MOBILIZING FINANCIAL RESOURCES A. Alternative Sources of Saving ............... 216 Government Saving .................... 217 Enterprise Saving .................... 218 Household Saving ..................... 220 B. Subsidies and Transfers .................. 221 China's Subsidies in Perspective ............ 222 Future Options for Price Subsidies . .......... 224 Future Options for Transfers ............... 227 C. Revenue Needs and Sources ................. 229 Revenue Sources .. 231 Implications for Saving ................. 234 D. Tax Reform .. .............. 238 Indirect Taxes ...................... 238 Profit and Asset Taxes .................. 239 Personal Income and Wage Taxation ............ 241 Taxation of Agriculture .. 241  - iv - Page No 10. DEVELOPMENT MANAGEMENT A. The State and the Individual ............................. 243 Issues of Income Disparity ............................. 244 Social Security and Welfare ............................ 246 Social Security and Services in Rural Areas ............ 249 B. The State and Enterprises ................................ 251 Control and Management of State-Owned Enterprises ...... 251 Competition ............................................ 253 Motivation of Workers and Managers ..................... 256 Management of Strategic and Key Enterprises ............ 257 Diversification of the Ownership System ................ 259 C. Investment and Price Reforms ............................. 262 Investment Decisions and Financing ..................... 262 Price Reform ........................................... 268 D. Economic Planning .......................... 273 New Approach to Planning ............................... 273 Tools for Planning ..................................... 278 E. Overview ................................................. 279 Tables in Text 1.1 Size Distribution of Industrial Enterprises ................... 14 1.2 Rural Income Distribution ..................................... 17 2.1 Aggregate Growth Rates, 1981-2000 ............................. 29 2.2 Composition of Final Demand ................................... 29 2.3 Sectoral Gross Output Growth, 1981-2000 ....................... 33 2.4 Composition of Capital Stock .................................. 35 2.5 Sectoral Employment Shares .................................... 37 2.6 Sectoral Shares of National Income ............................ 40 4.1 Energy Supply and Demand, Year 2000 ........................... 69 4.2 Share of Solid Fuels in Commercial Energy Use ................. 73 4.3 Estimated Investment Requirements in the Energy Sector, 1986-2000 ................................................... 86 5.1 Inter- and Intraprovincial Dispersion of Production Activities, 1982 ............................................ 92 5.2 Gross Output and Profits of State-Owned Industry, by Province and Subsector, 1982 ............................. 92 5.3 Industrial Concentration by Region and Key Cities, 1982 ....... 96 5.4 Extent and Composition of Rural Nonagricultural Activities .... 102 5.5 Urbanization and Economic Development ......................... 114 5.6 Personal Income per Capita in Urban and Rural Areas, 1982 ..... 115 5.7 Distribution of Employment and Population: Alternative Projections, 1981-2000 ...................................... 118 5.8 Growth in Urban and Rural Incomes: Alternative Projections, 1981-2000 .................................... 119 5.9 Urban-Rural Income Disparities: Alternative Projections, 1981-2000 ...................................... 120 5.10 Rural and Provincial Poverty, 1982 ............................ 122 5.11 Rural Development in Dingxi County (Gansu) and Wuxi County (Jiangsu), 1983 ................................. 125 5.12 Variations in Wage Levels, by Province and Sector, 1982 ....... 127 5.13 State Budget Revenues and Expenditures, by County, 1983 ....... 128  Page No 6.1 Total Foreign Trade Ratios: Alternative Projections, 1981-2000 ............................................ ...... 139 6.2 Manufactures in Total Merchandise Exports .143 6.3 Sectoral Foreign Trade Ratios: Alternative Projections, 1981-2000 .144 6.4 Growth in Foreign Trade: Alternative Projections, 1981-2000............... ... 145 6.5 Structure of Foreign Trade: Alternative Projections, 1981-2000 .145 6.6 Foreign Balance and Debt: Alternative Projections, 1981-2000 .152 7.1 Total Factor Productivity in State-Owned Industry, 1952-82 157 7.2 Sources of Successful Innovations in European Firms .168 8.1 Educational Attainment, Selected Countries, by Age Group, Level and Sex .179 8.2 Alternative Population Projections, 1980-2100 ................. 204 8.3 Population of Working :e (15-64) under Three Alternative Projections, 1980-2100 .206 8.4 Population Aged 65 and over, under Three Alternative Projections, 1980-2100 .208 9.1 International Comparison of Saving, 1976-80 .216 9.2 Subsidies in China, 1981 .223 9.3 Subsidies and Transfers in Other Countries .223 9.4 Alternative Scenarios for Subsidies and Transfers .228 9.5 Illustrative Projections of Government Expenditures .230 9.6 Illustrative Projections for Figure 9.2 .233 9.7 Illustrative Projections for Figure 9.3. 236 Figures 1.1 Composition of Aggregate Demand. 4 1.2 Production Structure. 5 1.3 Employment Structure. 6 1.4 Enrollment Ratios in Formal Education: China (1965-83 and Targets for 2000); Other Developing Countries (1965-80) 20 2.1 Share of Agriculture in Total Employment 1960-82 .39 3.1 Projected Crop Exports and Imports .50 3.2 Food Intake per Capita in Selected Economies in Relation to Income, 1960-81 .56 3.3 Share of Animal Products in Total Food Intake and Per Capita Income in Selected Economies, 1960-80 .57 5.1 Spatial Variations in Gross Agricultural and Industrial Output per Capita, 1982 .95 5.2 Urban/Rural Distribution of Employment, Capital and Gross Output under Alternative Projections, 1981-2000 117 6.1 Exports of Goods and Nonfactor Services as Z of GDP, 1978-82.. 132 6.2 Trade in Goods and Nonfactor Services, 1978-83 ................ 140 6.3 Direction of External Trade, 1983 ............................. 147 6.4 External Balance, 1978-84 ..................................... 151  - vi - Page No 7.1 Output and Cost of Production of JA 1-1 Sewing Machine by Enterprise . ... 159 8.1 Three Alternative Population Projections, 1980-2100 . ...... 203 9.1 Sources and Uses of Funds in the Economy, 1981 . . .............. 215 9.2 Income Flows and Taxation .... 232 9.3 Saving Rates and Shares ............ ............ . . 235 Boxes 1A Size and Growth of Enterprises in Market Economies . . 13 2A Wages, Profits and Prices .. 41 3A World Grain Trade, Past and Future .. 52 4A Cogeneration for Industry and District Heating . . 75 5A The Spatial Distribution of Dairy Production . . 98 5B Locational Issues in the Iran and Steel Industry . ....... 101 5C Locational Issues in the Bicycle Industry ........... 103 5D Trade, Pricing and Transport of Cement ............. 109 5E Agriculture on the Loess Plateau .............................. 124 5F The Benefits and Costs of Migration to Communities of Origin ............................ 126 6A Contrasting Types of International Economic Strategy . ..... 135 7A Continuous Casting of Steel ................................... 161 7B Strategic Options in Power Station Boiler Technology . ..... 163 7C A Successful Strategy in Electronics .............. 165 8A The Second Health Care Revolution ................ 209 IOA Dispersed Social Ownership in the Federal Republic of Germany ..................................................... 2 54 lOB Monetary Control ........................ 267 Maps 5.1 Net Material Product per Capita by Province, 1982 .94 5.2 Provincial Budget Expenditures and Revenues, 1982 .129 Relief and Administrative Divisions .282  SUMMARY AND CONCLUSIONS 1. China's ultimate economic objective is to catch up with the de-vel- oped countries, while maintaining a socialist system in which the benefits of prosperity are widely shared. Major, though uneven, progress toward this goal was made in the past three decades. During the next two decades, there could be substantial further progress. The foundations for rapid and equitable growth in the twenty-first century could also be Laid. But this will require steering a difficult course, in both development strategy and system reform. 2. This report looks at some of the key issues for China in the next 20 years, and at some of the options for addressing them, especially in light of international experience. It covers many specific topics, including agricul- ture, energy, technology, transport, industrial location, internal and exter- nal trade, population, education, employment, and finance (health was addressed in another recent report). The first section of this summary assem- bles some of its conclusions in the context of a more general discussion of the pattern of economic growth. Common themes regarding the system of econo- mic management and planning are puLled together in the second section. A third section covers some related aspects of sociaL policy. At che end of the summary there is a brief overview. 3. The most careful forecasts are often confounded. The lessons of international experience are often ambiguous and controversial. In any event, they are hard to apply to China, a country that in important respects differs from all others, and is not easy for outsiders to understand. The projections of this report are thus not predictions, and its suggestions are tentative. They are no more than an attempt to contribute to the debate in China about the difficult questions that must be confronted during the country's unique socialist modernization. Pattern of Growth 4. National income per capita in the developed countries is (in terms of purchasing power) approximately ten times higher than in China, and will increase in the future by at least 2% and perhaps 3X per year. To catch up by 2050, China's per capita income would have to increase at an average rate of at least 5.5% and perhaps 6.5Z per year. Such rapid progress has been rare elsewhere. In 1960-82, excluding small and oil-dominated economies, only two deveLoping countries achieved annual per capita income growth above 5%: South Korea (6.6%) and Greece (5.2%). More generally, only one country - Japan - has indisputably caught up with the developed nations from a position of eco- nomic backwardness. 5. China's past record, on the other hand, is quite encouraging. From 1952 to 1982, despite relatively rapid population growth and periods of acute economic mismanagement, per capita national income grew at an average annual rate of 4.0%, with phases of significantly faster growth. Of particular rele- vance is China's performance during the past few years of policy and system reforms: from 1979 to 1984, per capita national income grew at 6.8% per year.  - ii - 6. Although the basic objective is to improve living standards and eli- minate poverty, the Government has set targets of quadrupling the gross value of industrial and agriculturaL output (GVIAO) between 1980 and 2000 and increasing per capita national income from about $300 to $800 (about 5Z per year). If China's future investment efficiency were similar to the average of the past three decades, but allowing for increased investment in economic and social infrastructure, quadrupling GVIAO would require an investment rate of about 30Z of national income. This is comparable to the rates in other fast- growing East Asian economies, as well as East European countries, and in line with China's past investment rate. With reasonable improvements in energy and materials use, and provided that the population in 2000 is fairly close to the official target of 1.2 billion people, quadrupling GVIAO would allow the Gov- ernment's per capita income target to be attained. 7. The target growth rate of per capita income is far above the 2.6-3.6% range projected by the World Bank for middle-income countries in 1985-95. It seems feasible because China's projected population growth is unusually slow (about 1% p.a., as compared with an expected 2.2% in middle- income countries), and because China's investment rate has consistently been unusually high (the 1982 average for middle-income countries was 24Z). But its attainment depends crucially on the efficiency with which resources are used. For example, if the overall efficiency of investment were similar not to China's 1952-82 average but to the 1957-77 average - roughLy comparable to that of the Soviet Union in 1950-75 and India in 1960-82 - the growth rate of per capita income would be about one percentage point Lower. 8. If agricultural or energy production (discussed below) increased unexpectedly slowly, or if domestic or internationaL misfortunes adversely affected China's efficiency or investment rate, growth could be slower still. In any event, experience in China and other socialist countries strongly suggests the desirability of cautious planning. Unrealistically high growth targets cause fluctuations, shortages and inefficiency; while aiming too low has few adverse consequences. Indeed, continuation of China's recent practice of setting modest annual and medium-term targets could greatly facil- itate reform of the system of economic management. Sectoral Balance 9. Experience elsewhere also confirms the importance of striking an appropriate balance among the different broad sectors of the economy. In par- ticular, favoring industry at the expense of other sectors does not seem to generate rapid and efficient growth. In China, incentives for agricultural production have recently been substantially increased, but additional measures to support agriculture (discussed below) will be needed. Moreover, emphasis on quadrupling GVIAO should not imply neglect of infrastructure and services, which are vital for industrial and agricultural efficiency. The adverse con- sequences of past underinvestment in electricity and transport are well recog- nized in China; but there are still difficulties in diverting sufficient funds from directly productive investment into infrastructure. And the importance of commerce and other services (including finance and enterprise support ser- vices) has only recently come to be appreciated.  - Liii - 10. The share of the service sectors - including commerce - in output and employment in China is at present strikingly small by comparison with other developing countries, despite an unusually large education and health sector. The pattern in China, however, is characteristic also of the Soviet Union and Eastern Europe. Continued emphasis on increasing the physical volume of production (for example, the target of quadrupling GVIAO) might cause this pattern to persist. But China could also shift to a different pat- tern of growth, with accelerated development of the service sectors. 11. Rapid service sector expansion and reform of the svstem of economic management are closeLy reiated options. Expansion of commerce would go hand- in-hand with expansion or market regulation and increased specialization of production units and localities. Expansion of enterprise support services, including finance, accounting and law, wouLd likewise be a corollary of increased enterprise autonomy and specialization. And expansion of personal services would reflect a change in the relacive weights attached to planners' priorities and peoples' preferences. Moreover, rapid growth of services pro- bably couLd not be accomplished through administrative directives and central- ized resource allocation, but would have to be pulled by demand. Individual and collective enterprises may also be better suited than state enterprises to providing many types of services. :2. This association with system reform makes it likely that the bene- fits of faster service sector growth would more than offset its costs (in cerms of investment, skilled labor and - to a lesser extent - materials). In addition to the direct benefit to households from more convenient retail com- merce, and from more restaurants, cailors and so on, there would be indirect contributions to the efficiency of industry and agriculture. A bigger, better-equipped and more responsive commerciaL sector could reduce require- ments for both circulating capital and - by permitting greater speciaLization and realization of economies of scaLe - fixed capital. Enterprise support services also make fuller use of specialized equipment and personnel, while tinancial institutions can contribute to better investment decisions. Larger commerce and enterprise support service sectors may also contribute to reduc- ing material input requirements, especially by accelerating improvements in product quality, which make the cost of materials a smaller fraction of the price. 13. An alternative development path, involving both greater emphasis on services and more efficient resource use, is therefore analyzed in the report. With a level of investment efficiency similar to the average in Japan in 1950-80 and in all middle-income countries in 1960-82, it could attain the Government's target national income growth rate with an investment rate of only 26%, and hence faster growth of consumption. (The growth rate of indus- try would be 7% p.a. rather than 8%, however, and hence GVIAO would less than quadruple.) On this path, China would also have the option of maintaining the investment rate at 30Z, which would cause national income tc grow one percent- age point faster than the target rate (and GVIAO to slightly more than quad- ruple).  - iv - Human Resources 14. Faster service seccor expansion could also make it easier for China to absorb growth in the labor force, averaging about 10 million workers per year in 1980-2000. In the usual pattern of development - even in rapidlv industrializing countries - the main intersectoral employment shift is from agriculture into services, with a more moderate increase of employment in industry, which is less labor intensive. In China, however, employment has grown faster in industry th,.n in services. Continued emphasis on increasing the physical volume of production could perpetuate this pattern. This is especially likely because the prospective decline in the primary school age group will cause the number of teachers - now approaching a quarter of all service sector employment in China - to increase only modestly. Service sec- tor employment in 2000 might thus remain as low as 70% of industrial employ- ment (similar to the ratio in the Soviet Union in .959). 15. On the alternative development path, by contrast, service sector employment in 2000 could be as much as 150% of industrial employment (similar to the ratio in Japan in the early 1950s). Such a change in the composition of nonagricultural employment could ease some problems in China's urban areas, which now suffer from a combination of surpluses of some sorts of Labor and shortages of other sorts of labor (including competent, able-bodied workers even at relativeLy low skill levels). A much higher ratio of service seccor to industrial employment wouLd help to reduce this structuraL mismatch becween labor demand and supply, since service sector jobs tend to be disproportion- ately filled by the categories of labor that are now in surplus in urban China - particularly women. 16. Perhaps more important, faster service sector expansion, by raising the overall level of nonagricultural employment, could reduce the proportion of China's labor force that must remain in agriculture, where the earnings and contribution to production of most workers are likely to remain unsatisfactor- ily low for the next two or three decades. On the material-production- oriented quadrupling path, the proportion of the labor force in agriculture - excluding brigade industries - might decline by 2000 to onLy slightly under 60%. On the alternative development path, though, with the same growth rate of nationaL income, it could decline to nearer 50Z, a difference of perhaps 40 milLion workers. 17. Even the latter percentage, which implies an absolute increase in agricultural employment of 13 million, is much higher than most Chinese pro- jections, which envisage a decline in the agriculturaL employment share to under 40%. This would be normal for a country with a per capita income roughly double the Chinese target for 2000. Given the growth rate of nonagri- cultural output implied by this target, and the present very large agricul- tural employment share, such a large decline in agricultural employment could be achieved only if nonagricultural labor productivity grew slowly. This might well be incompatible with efficient industrialization. In the early decades of the twenty-first century, however, agricultural employment is likely to shrink rapidly.  18. Especially because many of the children of today's farmers will move into nonagricuitural employment at some point during their working lives, it will be important for China to continue to emphasize the wide diffusion of basic educacion, and to improve its quaLity. International experience strongly suggests that this couid contribute not only to faster economic growth, but also co a less unequal distribution of its benefits. in recent years, however, there have been problems in maincaining the primary school enrollment ratio (although it remains well above the developing country aver- age); and the secondarv schooi enrollment ratio has dropped below the devel- oping country average. 'Wide variations among localities in enrollment ratios and school quality also persist. These problems are being aggravated by increased locaL financial self-reliance in basic education. as well as by key schools (recently modified in name, but less in substance) at primarv and secondary levels. Policy changes, and greater financial support from central and provincial authorities, will be needed if the important target of making nine years of basic education universaL by 2000 is to be attained. 19. As regards advanced education and training, the need for rapid prog- ress to make up for the damage inflicted by the Cultural Revolution is well recognized in China, but some changes in emphasis seem advisable. University- Level courses are too specialized, partly because many educational institu- Lions are subordinate to sectoral ministries and other agencies. The Govern- ment's present policy of vocationalizing secondary education could suffer from similar problems. The case for sound generai education, supplemented by sub- sequent, briefer training in highly specialized skills, is particularly strong in China over the next few decades. Rapid technical advance and structural change will require conscant alterations in che skill composition of the labor force. The slow prospective growth of the labor force in the twenty-first century will reduce the scope f:r achieving such aLteracions through changes in the pattern of training of new Labor force entrants. It will correspond- ingly increase the required amount of retraining, which is more difficult for people whose original education was highLy specialized. Food and Agriculture 20. Despite its long history of food problems, China's per capita con- sumption of calories and protein is currently similar to the average for middle-income countries. The share of animal products in the diet, however, is that of a low-income country, and a major issue is how far and fast this share should increase. This issue is hard to address because of uncertainty about China's agricultural production potential: depending on whether the recent remarkable surge in production wanes or persists, the future trend growth rate of agricultiral output could lie anywhere between 2-3% and 5-6% per year. 21. In addition to the successful measures already implemented in recent years, especially introduction of the production responsibility system, the Government could take a number of steps to enhance agricultural growth pros- pects. These include improvements in agricultural research, education and extension services; irrigation and drainage projects; better nutrient balance in fertilizer supplies, as well as changes in the pricing and distribution system so that fertilizer is allocated among crops and localities more in  - vi - accordance with its potential contribution to production; increased availabil- ity of agricultural credit: and improvements in ruraL transport, storage, and marketing faciLities. 22. To permit consumption of animal products to rise from 6Z of total caloric intake in the early 1980s to 15Z by the end of the century, gross agricultural output would have to increase at about 4.5Z p.a., with crop pro- duction growing at 3.6% p.a. (less than the 1965-83 average) and animal hus- bandry at 7.5% p.a. (well above past rates). This agricuLtural growth rate could simultaneously satisfy the other demands of economic growth at the gov- ernment's target rate, with agricultural exports and imports more or less in balance in 2000. But it is very high by international standards - well above the 3.0% average for middle-income countries in 1960-80, and surpassed only by Thailand and the Philippines, which started with more uncultivated land and lower yields_ 23. To achieve such rapid growth of animal production would also not be straightforward. China's pastures are now seriously over-grazed, which makes it unlikely that present targets for beef and mucton production growth can be achieved. Much higher priority than at present would need to be given to improving techniques for the production of poultrv, a potentially efficient converter of feed into animal protein. The most important issue, however, appears to be availability of grain and protein-based feeds, o-i which an increasing proportion of animal nusbandry will have to be based. This is a matter partly of the backwardness of China's feed-processing industry, partly of achieving the required level and pattern of crop production. 24. With rapid growth of animal husbandry, the Government's present targets for increasing production of specific crops would tend to create a large shortage of the coarse grains needed for animaL feed, and a large sur- plus of rice, whose consumption is likely to grow only slowly. Feedgrain imports could make up any conceivable shortage, probably without an apprecia- ble increase in world prices; but it would be possible to export only a part of the potential rice surplus at economically attractive prices. Even if overall crop production targets can be attained, it may thus be necessary to switch some rice land either into coarse grains or into other crops that could be more easily exported. 25. The future pattern of agricultural production and foreign trade thus depends quite heavily on the rate of increase in consumption of animal pro- ducts. The possibility of a rapid increase cannot be ruled out, but there are some significant problems and uncertainties, which could be greater in the twenty-first century. (Even with a favorable feed conversion ratio, an increase of 10 percentage points in the share of animal products in the aver- age diet is equivalent to the disappearance of 35-40Z of China's cultivable land, or a 35-40% increase in population.) Particularly undesirable would be an increase in per capita consumption of animal products to a level that could not be sustained - because of a slowdown in crop production growth, poor feed- meat conversion ratios, or unwillingness or inability to sustain large feed- grain imports - and therefore had to be reduced.  - vii - 26. Although there appears to be little danger of a decline in China's present very satisfactory food intake level, caution is accordingly needed in managing food demand. It seems especiaLly important that consumers should pay the full economic cost of animal products, without subsidies or administrative restrictions on prices. Japanese experience suggests that high prices can bring consumer demand for animal products into line with limited supplies in a socially acceptable manner. By contrast, the experience of countries which have subsidized and rationed consumption of animal products has been extremely unfavorable. Energy Production and Use 27. Because China's initiaLly high consumption of energy per unit of output offers great scope for conservation, it would be possible to attain the Government's target of quadrupling GVIAO with a much smaller proportionate increase in energy production - though probably not as small as the originally envisaged doubling. Electricity production probably has to quadruple by 2000 to meet demand, which will require a huge amount of investment (averaging about 2% of national income), although it is technically feasible and in line with current plans. Crude oil production in 2000 is hard to predict, since it depends heavily on success in replacing the outpuc of existing fields with that of new discoveries, but the 200 million ton target (twice the present Level) couLd be more than enough to meet the likely Level of domestic demand. 28. It is coal, however, that will bear most of the burden of bringing 'uture energy production into line with demand. How much coal will be needed could vary widely, depending on the rate and sectoral pattern of growth, as well as on the degree of energy conservation achieved within individual sec- tors. With growth of national income at the Government's target rate, the material-production-oriented quadrupling path would require 1,400 million tons of coal to be produced in 2000 on the most optimistic assumptions about prog- ress in conservation, and more probably 1,600 million tons (as compared with 770 million tons in 1984). At the same national income growth rate, the alternative - more service-oriented and efficient - development path mentioned earlier could reduce coal demand in 2000 to 1,200 million tons on the most optimistic conservation assumptions, and 1,400 million tons on intermediate assumptions. 29. Especially because less optimistic, but still plausible, assumptions about progress in conservation could increase the potential demand for coal in 2000 to over 1,800 million tons, it will be vital to promote economical use of all forms of energy. The mixture of administrative controls and financial incentives applied in recent years has led to impressive energy savings. But a quota-based energy allocation system is fundamentally unsuited to achieving large and economically rational reductions in energy intensity over the longer term. It gives enterprises an incentive to exaggerate their needs, and does not take enough account of the widely varying economic costs of reducing energy consumption in individual enterprises. It is very probable that more energy could be saved, with less waste of other resources, if enterprises were more strongly motivated to reduce all costs and faced with prices that better reflected the economic values of all their inputs and outputs.  - viii - 30. it should be possible to mine the required quantities of coal, given China's ample reserves and the straightforward production technology, although this would require some advance pLanning, substantial investment, and con- tinued support for small collective mines. Much more of a challenge will be Eo transport the coal (as discussed below), and to use it cleanly and effi- ciently. The share of coaL in totaL commercial energy use is likely to remain at 70-75% (compared to 30-35% worldwide), which will require it to be put to an unusually wide range of uses, including many where backward technology and small scale in China today result in wasteful and dirty combustion. 31. To prevent further increases in alreadv unacceptably high levels of urban air pollution, as weli as to economize on fuel, will often require the replacement of decentralized and uncontrolled coal-burning in households and enterprises with centralized, large-scaLe, environmentally controlled combus- tion to produce cleaner forms of energy (gas, electricity, steam, hot water) Eor distribution to final users. This in turn will require China to import and disseminate new technology - which will also be critical in existing Large-scale uses such as steel and cement production - and to invest heavily in suitable production and distribution facilities. Careful studies will be needed to select the least costly of the various alternatives (e.g., coal gasification versus cogeneration and district heating) in each case; and their implementation and operation will entail much cLoser cooperation chan at present among a variety of agencies and organizations. Transport Priorities 32. There is concern in China that economic growth over the nexc decade or two might be held back by transport shortages, especially given the length of time needed to construct new railways and roads. But the voLume of freight transport in China is now so high (per unit of output) by international stan- dards - though Lower than in the Soviet Union - that it is hard to predict how rapidly it will need to increase in the future, given any particular trend growth rate of national income. This wiLl depend on the sectoral pattern of growth (for example, industry is more transport-intensive than services), and on its spatial pattern (which influences average transport distance). It will also be affected by system reform, which will on the one hand tend to increase transport requirements because of increased specialization and exchange, but couLd on the other hand greatly reduce wasteful use of transport facilities, for example through less cross-hauling, more preliminary processing of mater- ials, and more rational location decisions. 33. If national income were to grow at the Government's target rate, with continuing strong emphasis on material production, and with agriculture and industry using transport only moderately more efficiently than in the past, freight volume could reach 3,200 billion ton-kilometers by 2000 (approaching four times its 1980 level). But with the same growth of national income achieved through faster expansion of services and slower expansion of industry, and with somewhat greater efficiency in transport use, freight volume in 2000 would be 2,400 billion ton-kilometers. Even in the latter case, China's projected freight-intensity would remain high by comparison with other countries, which suggests that the possibility of even slower growth of freight volume cannot be ruled out.  - ix - 34. Although the overall volume of transport requirements in the longer term is highly uncertain. there will certainly have to be significant changes in the composition of transport investments, as well as in organization and management. AgriculturaL and industrial specialization will involve much movement in fairly small Lots of a great variety of goods, often perishable or fragile, over short to medium distances, to and from dispersed origins and destinations. This is generally only feasible bv road. The demand for road passenger transport will aLso increase rapidly. However, China now has a rural road network half the size of India's and fewer trucks per ruraL person than the impoverished Sahelian countries of West Africa. Major intercity highways, the vehicle fleet, and vehicle fuel production and distribution are also at present strikingly backward. 35. That road transport should play a larger role in the future is widely recognized in China. Yet there seems to be no strategic plan to bring this about. Underutilized agricultural labor couLd, as in the past, be mobi- lized to build up an adequaEe rural road network. Buc higher levels of gov- ernment (central and provinciaL) will need - as in ocher countries - to play a much more active role in planning, coordinating and financing the road devel- opment schemes implemented at lower levels. In addition, it will be essential to produce a much larger number and wider range of more fuel-efficient trucks, to improve the utilization of vehicle fleets, to upgrade the qualicy of motor fuels, and to distribute and price these fuels more rationaLly. All this will require improved coordination among higher-level government ministries and agencies. 36. Railways will remain the most efficient means of transporting raw materials and heavy industrial products over long discances. There will be a large increase in demand for coal transport, even with more coal washing and minemouth power generation. But there is apparently much scope for economiz- ing on rail transport of all commodities through stronger incentives for cost reduction and less administrative intervention in materials allocation and pricing. It also seems physically and financially possible to accommodate the overall increase in rail transport demand. This couid be accomplished largely by switching from steam to diesel and electric traction, although substantial double-tracking and some new lines will also be necessary. 37. Increased use of water transport could appreciably reduce the load on the railways, expecially if future heavy industrial investments were appro- priately located. And China's entire transport system could benefit from better inter-modal coordination: this is partly a matter of improving tech- nology and infrastructure (e.g. containers and their handling facilities); but it also depends on rationalizing tariffs, increasing competition, and encour- aging new specialized entities to provide an appropriate variety of transport, leasing, storage and transfer services.  Specialization and Urbanization 38. China's size and past emphasis on local self-sufficiency offer opportunities for enormous economic gains through increased specialization and trade among localities. rn recent years, agricultural specialization has pro- ceeded rapidly. In industry, by contrast, specialization does not seem to be increasing, and indeed local planning bureaus are still promoting the develop- ment of a wide range of industrial activities, with continuing emphasis on local self-sufficiency. This appears to result partly from transpo-t prob- Lems, but also from shortcomings in -ommerce and material supply, the nature of the fiscal system, distorted prices and insufficient competition. It could have a serious adverse effect on China's overall investment efficiency. 39. To prevent continued proLiferation of industrial plants of subopti- mal scale and product quality, there is thus an urgent need for measures to reduce the present conflict between what is rational for a particular locality and what is rational for China's whole economy. Of particular importance would be better transport, fiscal reform, and changes in the system of econo- mic management (discussed below), regarding enterprise motivation, competition and prices. In addition, the central government might consider legal measures to back up its existing general prohibition of barriers to interlocality trade, including possibly the establishment of a special regulatory institu- tion, with the power to levy large fines. 40. Sub-optimal plant size is not China's only current industrial loca- tion problem. There is also too much concentration of industry in Large and medium-sized cities. Research and experience in other countries suggest that economic efficiency requires weight-reducing heavy industry (e.g., iron and steel production) and agroprocessing to be located near raw materiaL sources, with other space-intensive or standardized industrial activities in special- ized small cities. Industries such as petrochemicals and bicycles benefit from large-scale operation, but not from proximity to other industries - or at any rate not enough to justify the much higher economic cost of land and labor in large cities. By contrast, industries whose technology is evolving rapidly or which must respond to constantly changing demand, as well as many special- ized service activities, benefit substantially from mutual proximity and clus- tering in large urban areas. Development of both large and small cities, as well as smaller towns providing goods and services to rural areas, could thus contribute to rapid and efficient growth in China; but planners and prices (discussed below) will need to give better guidance on what activities should go where. 41. How much the overall level of urbanization should increase in the decades ahead is a difficult issue, partly because the distinction between urban and rural is itself blurred in many densely populated parts of China. In any eyent, substantially increased rural-urban migration seems probable and desirable. Indeed, without some rural-urban migration, China's overall urba- nization rate (as officially measured) would decline from 20% in 1981 - quite normal for a low-income ccuntry - to 19% by 2000, simply because the natural population growth rate in existing urban areas over the next two decades is likely to be much lower than in rural areas. This could cause the general labor shortages that have already appeared in some Chinese cities to become  - xi - widespread within a few years. Increasing urbanization would probably be a more economically efficient response to such shortages than accelerated auto- mation of urban production. An urbanization rate in 2000 of 30% (implying that more than a third of the urban population would have come from rural areas) would still involve an unusually large amount of rural nonagricultural employment. Lower-middle-income countries today (at a similar income level) have an average 34% of their popuLation in urban areas. 42. Whatever the overall rate of urbanization, efficienc development of towns and cities will require a very large increase in urban infrastructure and services. The administrative responsibilities for their provision, and potential financing and cost recovery mechanisms, need to be carefully reviewed. In many countries, efficient and equitable development of urban services has been seriously constrained either by fragmenced and unclear allo- cation of responsibility among different administrative levels, or by inade- quate growth of urban revenue sources. In China, moreover, special attention should be given to the potential impact of economic and social reforms on the level and composition of local government revenues and expenditures. Foreign Trade and Capital 43. China's recent resumption and expansion of external economic con- tacts have raised exporcs to 9-10% of national income, which is well within the normal range for large courtries. However, inflows of foreign loans and direct investment have remained small, partLy because of foreign trade sur- pluses. For the future, there appear to oe two large and related issues. The first is how much, and in what directions, to further increase external con- tacts. The second is how best to manage such contacts, and in particular whether to continue the present insulation of the domestic economy from the world outside by administrative intermediation and separate price systems. 44. Over the next two decades, China will need to change the composition of its manufactured exports. Exports of textiles and clothing, in particular, are likely to grow fairLy slowly. China will therefore need to make a major effort to expand exports of machinery and metal products. Rapid growth in such exports would permit rapid growth of intermediate and capital goods imports, as well as consumer goods imports (which can stimulate important improvements in domestic production). Exports of manufactured goods could also be used to pay for purchases of licenses and technological assistance, which in many advanced as well as developing countries have proved to be very effective ways of upgrading domestic technology. 45. As regards foreign borrowing, the Government's policy of cautious progress seems well founded. For a low-income country to run foreign trade surpluses, as China has done in recent years, has some obvious disadvantages, as compared with a net inflow of foreign capital to supplement domestic savings. But experience in other countries, including socialist ones, has shown how easy it is to slide into excessive or unproductive borrowing, espe- cially when domestic enterprises and planners are not sufficiently conscious of investment costs or responsible for bad decisions.  - xii - 46. However, China also seems wise to encourage direct foreign invest- ment, less for the foreign capital or advanced technology it brings than for the demonstration effect of modern management techniques. The example of, and competition from, well-run foreign companies can help domestic firms identify weak links in management, product design, material supply, and so on, and spur them to make changes they might otherwise never consider. But such examples of the way modern industry operates in more advanced countries are effective only if contrasted and competing with locally managed firms. This requires foreign and joint ventures to be spread - as is increasingly the case in China - among a wide range of localities and activities, rather than confined to special zones or particular sectors. 47. International experience suggests that decentralization of foreign trade to the enterprise level can also make an important contribution to realizing the potential benefits of increased externaL economic contacts. Direct exposure of exporting firms to foreign buyers and competitors has proved elsewhere to be an extremely effective way not just of learning in the abstract about new and better products and processes, but also of learning how (and being put under pressure) to introduce them in practice. Similarly, allowing enterprises - including commercial enterprises - and consumers to choose directLy between imported and domestically produced goods could greatly increase competitive pressure on Chinese producers to introduce new, better, and cheaper products. 48. Such decentralization of foreign trade could have some potentially serious disadvantages, too, including damage to promising but newly estab- lished industries in China, as well as greater exposure to world price insta- bility and unpredictable fluctuations in foreign markets. But skillful use of policy instruments - including the exchange rate, tariffs, and selective administrative interventions - could minimize the costs while securing many of the benefits. China will also always be a large and diverse economy, with a relatively smaLl foreign trade sector, which greatly reduces the possible damage that unexpectedly unfavorable external developments could inflict, even though particuLar industries and localities might experience quite sharp swings in prosperity. 49. Especially following the recent unification of foreign exchange arrangements - and given the current account surpluses of recent years, large reserves, and the ready availability of external capital - China appears to be well placed to move toward a decentralized and mainly indirectly regulated foreign trade system, without the problems (of external insolvency or internal inflation or deflation) that many other countries attempting a similar move have simuLtaneously had to contend with. To do this immediately, or at one stroke, would probably be unwise, since most domestic enterprises do not yet have appropriate motivation or independence, and since some prior changes of domestic prices would ease the adjustments required. But it could proceed in parallel with internal economic reform.  - xiii - Managing the Economy 50. How far China will have risen into the middle-income range by 2000 will depend crucially on how successfully the economy is managed. This is a matter partly of mobilizing financial and human resources, and of importing and developing modern technology, partly of putting these things to effective use. China's past record of economic management, for all its unevenness, is superior to that of other low-income countries, in terms of both growth and poverty reduction. But maintaining the past pace of economic progress is likely to require greater efficiency. 51. Efficiency in the relevant sense means ceaseless, intense efforts by producers to reduce costs, to increase productivity, to improve product qual- ity, to introduce new products, and to seek out and respond to changing needs and opportunities. It involves bold yet frugal investment decisions, increas- ing specialization, and the constant displacement of more expensive or infer- ior products and processes by cheaper and better ones. 52. These conditions for rapid, sustained growth are at present far from fully met in China, except in agriculture, where outstanding progress has been made in recent years with the introduction of the production responsibility system. Recognizing this, the Central Committee of the Chinese Communist Party in October 1984 announced a program of urban economic reforms, with invigoration of state enterprises as its central theme, to be implemented over a perio'I of about five years. Much of the discussion in the present report is premised on the Central Commictee's decision. 53. To be efficient, enterprises must be motivated to improve their eco- nomic performance; they must have some freedom of maneuver; they must be faced with economically rational prices; and they must be subjected to competi- tion. None of these elements is individually easy to establish, and the absence of any one of them reduces or nullifies the benefits of the others. In addition, the state must retain the ability co direct the overall pace and pattern of development. The essence - and the difficulty - of a successful and comprehensive reform thus lies not only in bringing cogether the several elements of market regulation, but also in combining them with an appropri- ately modified system of planning. Enterprise Motivation 54. The unsatisfactory results of direct administrative control of enterprises - whether by central ministries or local governments - lead irre- sistibly to the principle of greater enterprise independence. Yet establish- ing independent and appropriately motivated enterprises may be the hardest single aspect of reform in a socialist system - and perhaps also the most fun- damental, since an unmotivated enterprise cannot be expected to motivate its workers, or to respond to price and tax signals, or to compete. 55. Appropriate enterprise motivation should include a strong desire to increase profits. Avoidance of losses is an important aspect of this, but is insufficient, since China needs enterprises that not merely pursue a passive strategy of staying out of troubLe, but actively seek to increase production  - xiv - and sales, and to cut costs of all kinds. With rational prices and compeci- tion, the best single measure of enterprise performance in this regard is usu- ally medium-term profits (especially after deduction of the cost - deprecia- tion and interest - of the capital employed). The active desire of peasant households to increase sales and cut costs, because their standard of living depends on it, lies at the heart of China's recent agricuLtural successes. Outside agriculture, the same motivation is a natural feature of individual and family enterprises, and of small enterprises owned and operated collec- tively by their workers. In medium-sized and large enterprises, however, things are not so simple. 56. State Enterprises. Some important enterprises should clearly remain subject to direct government control, and should not primarily pursue profit. At a minimum, this appLies to public utilities (e.g. electricity, railways, and telecommunications), large-scale mineral exploitation, and many defense- related industries. Direct control of other key enterprises might also help to achieve specific development objectives, including the technological upgrading of particular sectors. 57. What is less clear is the best way to organize and motivate the remainder - probably the great majority - of state enterprises, which would be relatively independent. Once a suitable economic environment is created through price reform and increased competition, and provided that indirect levers such as taxes and credits are skillfully applied, pursuit of profit should lead most state enterprises in economically appropriate directions. It would thus no longer be necessary for particular individuals within enter- prises to be charged specifically with representing the interests of the state. But each of the alternative possible internal management arrangements Las strengths and weaknesses. 58. Giving direct control of state-owned enterprises to their workers is a socially attractive possibility. Japanese experience also shows the econo- mic advantages of strong worker commitment to the enterprise. But there are also disadvantages, especially because state enterprises should operate for the benefit of the whole society and not only of those who work in them. Experience in YugosLavia and elsewhere suggests that worker control could result in excessive wages and worker benefits, inadequate labor discipline and effort, restriction of employment, and comparative indifference to profitabil- ity. 59. Experience in developed capitalist economies suggests that giving control of enterprises, instead, to their managers would increase their pro- pensity to expand, improve and innovate. But completely independent managers sometimes choose a quiet life or, more commonly, pursue personal power through expansion with insufficient attention to profitability. In socialist coun- tries, moreover, managerial control tends to differ only slightly from worker control, since managers find it hard to resist worker demands for greater benefits or to insist on the often unwelcome changes in work practices that are needed for innovation and increased efficiency. This is already a problem in China: studies of experimentally reformed enterprises controlled mainly by their managers reveal large increases in worker benefits, but small increases in economic efficiency.  - xv - 60. An alternative approach, common in non-socialist countries (and now being tried in Hungary), is to give strategic decision-making authority in each enterprise to a Board of Directors. Such a board could contain some representatives of society at large, as well as of the workers. But to impart the necessary motivation, the board would mainly have to represent institu- tions with a strong interest in the enterprise's profits. It could insist that the enterprise managers behaved accordingly, partly by appointing and dismissing them, partly by linking their remuneracion to profitability. The managers might in praccice make most decisions, but would be greatly influenced by their uLtimate accountability to the directors. 61. Such boards shouLd be free from direct government intervention. But, precisely because these are state enterprises, this may be difficult to achieve. In China, for inscance, even if the boards were to consist of repre- sentatives of the Ministry of Finance or a new Ministry of State Property, rather than of sectoral ministries or the governments of the localities in which the enterprises are situated, informal connections and pressures for direct administrative control could persist. A possible solution might be to spread the ownership of each state enterprise among several different institu- tions, each in some way representing the whole people, but with an interest mainly in the enterprise's profits rather than directly in its output, pur- chases or employment. (Examples of such institutions, in addition to central and local go:ernments, are banks, pension funds, and insurance companies.) Such a system of socialist joint stock ownership could perhaps be creaced initially by suitable dispersion of the ownership capital of existing state enterprises. Over time, it could be reinforced by a more diversified pattern of investment finance (discussed below), with a variety of state institutions acquiring financial interests in existing and new enterprises. 62. Collective and Individual Enterprises. China's policy is to encourage the coexistence of state and other sorts of enterprises. In small- scale economic activities, whose importance wil' increase with economic devel- opment and reform, individual and collective enterprises are already making a valuable concribution. They could be even more productive if given better access to mace:ials, credit, skilled labor, and technology. Their potential role in medium- and Large-scale activities, however, remains to be determined. 63. Experience elsewhere suggests that genuine collectives flourish only in certain lines of activity, and rarely above a certain size. Many of the most successful examples elsewhere are in commerce, and are marketing or pur- chasing cooperatives. There are far fewer examples of medium-sized genuine worker collectives, or of any sort of collectives in manufacturing. China's numercus township and village enterprises may appear an exception; however, these are not collectives in the ordinary sense, but community enterprises. They have been successful partly because of restrictions on personal mobility among communities, which has given local residents a strong common interest in their establishment and profitability. 64. Collectives and community enterprises could in the future constitute a significant proportion of China's medium-sized enterprises. Many individual enterprises could also grow rapidly to medium size, if allowed to do so, and could be an important dynamic force in the economy. As an alternative to the  - xvi - present restrictions on the size of individual enterprises, these enterprises might be obliged to sell (at a fair price) majority ownership to one or more state institutions once they reached a certain size. The then minority owners could continue as managers, supervised by a board of directors representing all the owners. 65. Accounting and Law. The best way of motivating independent state enterprises, and the best balance among state, collective, individual and mixed enterprises, can only emerge from experience and experiment over a pro- tracted period. But certain necessary conditions for success can already be identified. 66. One is improvement of enterprise accounting and auditing, which at present leave much to be desired. Without accurate accounts, subject to thorough, compulsory and independent audits, with severe penalties for non- compliance, it is hard to see how any system of independent enterprises - state or nonstate - could function properly. 67. Moreover, a decentralized economy cannot function properly without a comprehensive system of commercial and contract law, and institutions and per- sonnel to implement it. Laws are also needed to regulate the economic activi- ties of independent enterprises. Among other things, laws or decrees should prohibit specified types of monopolistic, anti-competitive, or exploitative behavior, with legal institutions to interpret and implement these prohibi- tions. At present in China, the situation is unsatisfactorily vague, with enterprises and local officials free to place their own interpretations on general guidelines from the center. Competition 68. As the Central Committee's 1984 decision clearLy states, competition among enterprises is crucial to efficient growth - without it, customer requirements are neglected, innovations diffuse slowly, and market-regulated prices are distorted for monopoListic gain. One key ingredient of competition is allowing purchasers to buy from the best and cheapest source. China has already taken some significant steps in this direction, but further progress is needed. It could probably be assisted by dismantling the annual production planning and allocation system, either in one step (as in Hungary) or more gradually, by reducing each year in each enterprise the proportion of material requirements coverea by the state allocation and, in parallel, the obligatory state production and procurement quota. 69. It would also be assisted by dismantling administrative barriers to internal trade, which would be easier if fewer enterprises w..-.-e directly con- trolled by government organs anxious to avoid competition within their "fami- lies" of enterprises. Consolidation of enterprises into large sectoral corpo- rations, by contrast, would normally tend to reduce competition, although this could be offset by increased exposure to foreign competition. Equally impor- tant would be measures to enlarge, diversify and strengthen commerce and rela- ted services. Individual and collective enterprizes might be given a greater role in whoLesaling and material supply, and better access to transport facilities.  - xvii - 70. Another key ingredient of competition is free entry of new producers into particular markets. One step in this direction for China could be to allow larger enterprises to diversify more freely out of their existing lines of business. Provided that prices are rational, such diversification can transfer profits earned in one line of activity into other lines where the economic returns to investment are greater. And it can permit enterprises to transfer workers who can no longer be profitably employed in one line to other lines, without having either to discharge them or to keep them idle. 71. Another important step toward free entry could be to allow indivi- dual and collective enterprises to operate in more lines of activity, includ- ing technologically dynamic sectors. Studies in other countries have shown that small enterprises are just as innovative as large ones. (Chinese visi- tors to foreign countries are often given a misleading impression in this regard, because they are usually taken only to weLl-known, large enterprises.) Particularly because of the strong incentive that small enterprises have to innovate, the overall pace of technological advance could very probabLy be accelerated by letting them compete for some of China's skilled manpower, technology imports and research support. 72. Lastly, effective competition has to involve the elimination of obsolete or unwanted products and processes. Subsidies or protection for backward technology and inefficient enterprises - other than on a temporary basis to permit reorganization - hold back the growth of efficient enter- prises, and remove a powerful negative stimulus to innovation and improved efficiency. At the same time, this aspect of competition can obviously cause hardship, especially for displaced workers. Social policies and institutions (discussed below), and gover,unent retraining schemes, can alleviate this hard- ship. But the conflict between the gains of the majority from greater compe- tition (and faster technological change) and the losses of particular individuals and groups cannot be completely eliminated. Prices 13. The Government is well aware of the need to estabLish economically rational prices, without which the decisions of independent, profit-oriented enterprises would often be inefficient for the whole economy. The Government also recognizes that this must be accomplished by changing the price-setting system to give market supply and demand forces a greater roLe, and not only by administered changes in prices set by the state, which tend to lack the flex- ibility, complexity and precision needed in a modern economy. The prices of many minor items, and of some transacticns in more major items, have already been successfully decontrolled. 74. Further progress in this direction is impeded by chronic shortages of many goods. It is feared that these shortages, if prices were no longer subject to direct state control, would cause general price inflation, as well as obstructing the allocation of materials to key projects. For this reason, it is often argued in China that further price decontrol should be postponed until rising production has eliminated most of the shortages.  - xviii - 75. Soviet and East European experience suggests, however, that chronic shortages are not the temporary result of inadequate production capacity. but an enduring feature of administrative economic management, which can be elimi- nated only by systemic reforms, including price decontrol. This view is sup- ported by China's experience in the past few years, especially in agricuiture. where reLaxation of direct controls has turned long-standing shortages into abundance. The same could happen in industry. Increases in specific prices could eliminate specific shortages by stimulating supply and reducing demand. These price increases, moreover, would tend to reduce purchasing power over (and hence the prices of) other goods, provided that the government kept strict control of the budget balance and credit. 76. Nonetheless, there is substance to the fear of immediate and general decontrol of prices in China. The downward inflexibility of some prices, and of wages, could obstruct the smooth adjustment outlined above. The fulL response of demand and supply to price changes may take years. China Lacks experience in indirect fiscaL and monetary regulation of the general price level; and other countries with much greater experience still suffer from rising prices. The elimination of shortages, moreover, would also require increased competition and stronger enterprise motivation to hold down costs, including investment costs, which cannot be brought about overnight. For these reasons, price decontrol probably has to be gradual though steady - per- haps in conjunction with the gradual dismantling of annual production planning and allocation mentioned earLier. 77. Some steps, however, could be taken more quickly. Large increases, either administratively imposed or market-determined, in the prices of most forms of energy - and some other materials - are a case in point. Without them, there wouLd be littLe chance of achieving the essential reductions in energy use discussed earlier. There is concern in China that large energy price increases would have ripple effects, since energy-intensive industries and enterprises could not absorb all the increased cost through conservation and profit reduction, and would therefore have to raise their prices. But most such ripple effects should in fact be weLcomed (and cushioning through reduction of taxes and profit remittances avoided): experience elsewhere suggests that increases in the prices of energy-intensive products, by dis- couraging their use, can make a vital contribution to raising the economy's overall energy efficiency. 78. Agricultural prices are another instance. With the new system of household agriculture, if production is to be efficient and surpluses and shortages of particular commodities avoided, prices must be allowed to respond flexibly to supply and demand trends. At the same time, it would be desirable to avoid the large short-term price fluctuations - often because of weather - that characterize unregulated markets for agricultural comnodities. The Gov- ernment's present strategy is to decontrol completely the prices of minor items, and for major items to have two-tier pricing (flexible prices for amounts in excess of official, fixed-price, procurement contracts). A possi- ble alternative would be to have wholly market-determined prices, but with a price stabilization scheme - Government intervention to keep fluctuations at any given time within a pre-determined range by purchasing for addition to stocks at the lower end and selling from stocks at the upper end.  - xix - 79. A third instance concerns the relationship between producer and retail prices, which needs to be altered not only to reduce the presently high level of subsidies, but also co encourage consumers to buy less of things (such as energy or animal products) in short supply or whose production costs are increasing, and more of things in abundant supply and whose production costs are falling. This could be achieved by establishing relatively rigid margins between producer and retail prices, sufficient to cover the costs and normal profits of commerce as well as indirect taxes (whose rates could vary from commodity to commodity). 80. In the near term, what is mainly needed are some substantial increases in the retaiL prices of staple food and coal, and in rents. These should be as fully and accurateLy compensated as pocsible, by wage increases and income supplements to households with high dependency ratios (discussed later). Special supplementary interest payments on saving deposits which would otherwise lose part of their real vaLue might also be needed (financed perhaps by a corresponding special levy on borrowers, whose loan repayment burden would otherwise decrease in real terms). A large-scale program of selling state and enterprise-owned housing to its present tenants, who would otherwise face increased rents, could soak up a large fraction of existing saving deposits, and hence reduce the possible scale of panic buying in anticipation of price increa!es. 81. There is room for disagreement as to whether such a major realign- ment of retail prices, incomes and assets should be done in one step, or more gradually. The social problems that have sometimes followed large retail price increases in ocher countries seem to have occurred mainly because of the absence of compensation - often deliberate, because of the need to reduce real consumption in the face of economic difficulties. In China, however, there is no need for a cut in household consumption, and full compensation could be provided (although this would mean no net improvement in the state budget bal- ance). A carefully prepared and well-explained one-step adjustment could thus be quite acceptable, and would avoid the protracted uncertainty and delays to other necessary reforms associated with gradual adjustment. But it would also obviously increase the cost of errors in preparation (especially calculating the required changes in prices, wages and other forms of compensation) or implementation. 82. Following this initial phase of adjustment, subsequent rises and falls in producer prices could be reflected in retail prices. Experience in other countries suggests, however, that it would be inadvisable to provide continuing automatic compensation for retail price increases in the form of general wage or income indexation, since this tends to aggravate inflation. Even selective indexation (of state income supplements to poor households, for example, or for pensions) should be approached cautiously. 83. A fourth instance concerns prices that influence the spatial loca- tion of economic activities. One is transport: prices for each mode should more accurately reflect long-run marginal economic costs, without cross- subsidies among different-length hauls, to provide incentives to reduce the present waste of rail facilities, to encourage greater use of roads, and to enable proper assessment of the costs and benefits of locating particular  - xx - activities in particular places. The other is urban Land: for social as weLl as economic reasons, enterprises and planners should be made to feel the dra- matically varying economic usefulness (or opportunity cost) of different sites - higher in coastal cities than in more remote regions, higher in large than in small cities, higher near the center than in the outskirts (in a city the size of Shanghai, experience in other countries suggests that a centraL site is worth approximately 150 times as much as one in the suburbs). 84. A differentiated urban land tax reflecting these variations could be introduced. Alternatively, a competitive rental market might be estabLished - which would in principle be more efficient, and quite consistent with public ownership of all land. (Either or both could provide a useful supplement to municipal revenues.) This wouLd encourage better use of existing sites, and more economically efficient location decisions. The presently huge incentives for rural-urban migration would likewise be reduced to more economically rational proportions, especially if differing land values were reflected in house rents and wages, and hence in the costs of living and employing workers in large urban areas. 85. Finally, further strengthening of the linkages between internal and world prices could be beneficial, though there would need to be some excep- tions - for example, rice, where an increase to the present world price could aggravate incipient overproduction. For most goods, especially manufactures, enterprises should eventually directly feel the relationship between their value (for exports) or cost (for imports) to China in world markets and their domestic prices. Such linkages (which are now standard in Hungary, and have begun to be introduced in China) are essential for the efficient functioning of a decentralized foreign trade system. They would also contribute to a much-needed widening of price differentials between low and high quality pro- ducts, without which enterprise incentives co innovate and improve quality will remain small. Labor and Wages 86. To reduce the risk of shortages of highly skilled workers constrain- ing China's growth in the next two decades, measures to improve their alloca- tion and motivation seem worth considering. Among the more important possi- bilities would be freedom of job choice, for both new graduates and exper- ienced staff, aLlowing employers to recruit and release freely and competi- tively, and higher and more flexible salaries. Provided that they were accom- panied by reform of enterprise motivation, competition and prices, these measures could help move highly skilled labor to where its economic contribu- tion was greatest, speed the diffusion of new technology, improve motivation (with better matching of jobs to personal preferences and skills just as important as more money), and send much clearer signals of needs and priori- ties to education and training institutions. 87. More generally, it will be essential to make better use of all categories of labor. This is because long-term growth of average per capita income will be almost entirely determined by growth of labor productivity, most of which must come not from movement of labor out of agriculture into other sectors, but from higher productivity within each sector. Unskilled  - xxi - Labor is currently in surplus: but this will change in the twenty-first cen- tury; and experience elsewhere shows how hard it is at a late stage to break out of an established pattern of low productivity and indifference to labor costs. 88. It might therefore be desirable to give enterprises progressively more discretion in deciding how many, and which, less-skilled workers to employ. This should in principle encompass not only the right to dismiss idle or negligent workers. but aLso to release or reject workers who are simply not needed for production. Lifetime employment would not necessarily disappear (Japanese experience confirms its potential advantages in training and motiva- tion), but would be likely to be limited to larger enterprises, which would be highly selective recruiters. 89. Some open unemployment would be an inevitable consequence of these changes. In addition to its adverse human consequences (which could be miti- gated by changes in social policies and institutions, discussed below), this would involve waste of human resources. The waste of unemployment, however, has to be weighed against the waste of human resources associated with the present employment system, which transfers unemployment from the streets into the factories and discourages managers from realizing the full productive potential of their workers. It is also important to bear in mind the tremen- dous potential for creation of new jobs in the service and small enterprise sectors. Nonetheless, movement should be gradual, perhaps with an early end to the obligation of enterprises to accept unwanted new workers (usually as part of a package also containing some useful recruits), and subsequent release of unneeded workers each year up to a specified percentage of an enterprise's labor force. 90. Greater flexibility of less-skilled wages might smooth this transi- tion, as might stronger linkages between individual pay and performance. But these changes, too, could be only gradually introduced, to avoid a sharp drop in unskilled wages. It will also be necessary for some time to retain admin- istrative control of total and average wages (and other benefits) in state enterprises, to prevent excessive increases in response to pressure from workers. In this regard, the Government is contemplating formal linkages between the remuneration of workers and the performance of their enter- prises. But experience elsewhere suggests that this could generate inequities and misrepresentations that might undermine aggregate wage control. It might be preferable to have wage guidelines that were more uniform across enter- prises, coupled with greater managerial discretion in the distribution of wages among workers within particular enterprises. 91. Gradual wage adjustments could also help remote and backward locali- ties. At present, nationally uniform state sector wages, and unofficial pressure to tie collective sector wages to state wages, make it harder for enterprises in these localities to compete effectively against enterprises in more advanced regions, or in international trade. Nonagricultural employment in backward areas might therefore increase faster if unskilled wages in these areas declined in relative terms. At the same time, though, these areas pro- bably need to offer skilled workers higher wages than elsewhere.  - xxii - Investment Decisions and Finance 92. In energy, transport, education, health and defense, the Government winl undoubtedly retain direct control over most investment decisions, mobi- lizing the necessary finance through the budget. The same will apply to major land and water development projects. Budgetary saving to finance other sorts of investment may also have to remain substantial, if the high aggregate saving rate necessary for rapid growth is to be attained. But households and enterprises could contribute a much larger share of aggregate saving than in the past, and could make an increasing proportion of investment decisions. This pattern is already established in agriculture and other small-scale activities. In larger-scale industry and services, however, the degree to which investment decisions should be devolved to enterprises remains an important issue. 93. Such devolution could have a number of advantages. It would give greater meaning to enterprise independence, especially since the economic per- formance of an enterprise at any given time depends heavily on previous investment decisions. It would increase the medium and long-term responsive- ness of production to evoLving needs and demands. It would make fuller use of the detailed knowledge of enterprise managers. It might also provide a way of strengthening incentives for sound investment decisions, in whose outcome enterprise managers could be given a substantial direct personal financial interest. Experience suggests that this would encourage bold and innovative thinking combined with thorough analysis, hard-headed calculation, and the avoidance of waste. 94. Delegating more responsibility for investment decisions to indepen- dent enterprises might also help in striking a better overall balance between infrastructural and other investment. China's local governments have proved dedicated and resourceful investors, but are probably unduly biassed in favor of directly productive industrial investment. Making enterprises independent, with more investment responsibility, could thus give local governments a stronger incentive - as in other countries - to improve infrastructure in order to attract industrial investment. 95. China's recent mixed experience with devolution of investment deci- sions to enterprises, however, confirms that without other systemic changes such an arrangement can have undesirable results. Inappropriate managerial incentives, irrational prices, lack of competition, and shortcomings in commerce and material supply can lead enterprises to undertake investments of little economic merit, while neglecting other investments of much greater use to the economy. 96. Reforms in enterprise motivation, competition, and prices couLd eventually overcome many of these problems, making it possible to consider devolving a much greater proportion of investment decisions to enterprises and, relatedly, allowing state enterprises to retain a much larger proportion of their profits. More mobility of investment funds, through more varied channels, could also increase the efficiency of decentralized investment decisions.  - xxiii - 97. Investment flows in China in the past have been largely vertical, with most savings being mobilized upward through the budget and dispensed downward through sectoral ministries and state banks. In addition, there has been some compartmentalized reinvestment of savings generated within particu- lar sectors and localities. In the future, although vertical flows and com- partmentalized reinvestment will remain important, they could advantageously be supplemented and replaced to an increasing extent by horizontaL investment fund flows, of three main types. The first, mentioned earlier, would be allowing existing enterprises to diversify into new lines of activity. The second, which overlaps with the first, is direct investment linkages among economic units: enterprises could increasingly be permitted to invest in other enterprises, establish new enterprises, or participate in various sorts of joint ventures. The third sort of horizontal flow is through socially owned financiaL institutions. acting as intermediaries between the suppliers and users of resources. 98. Such financial institutions, offering interest rates that reflected the scarcity of capital, could mobilize funds from households and institu- tions, including enterprises with limited internal investment opportunities. They could make these funds available to potential investors, especially enterprises of all kinds, perhaps mainly in the form of loans, but also in the form of ownership capital - partly-because many worthwhile projects are too risky to be largely loan-financed, partly because directly sharing in the risks and benefits would motivate financial institutions to provide more assistance in project design and implementation, including information on mar- kets and technology. These institutions could in effect create a socialist market for investment funds. 99. For such a market to function weLl, the financial institutions shouLd be numerous and diverse, acting in most respects as independent state or collective enterprises. Nonetheless, government regulation of their activ- ities would be essential. At a miirimum, as in all other countries, rules to protect depositors would need to be designed and enforced, with other ruLes and central bank intervention in the financial market to regulate the overall suppLy of money and credit. In addition, the government could - as for many years in Japan and South Korea - influence the allocation of funds in accord- ance with strategic economic objectives. This could be done in various ways, including channelling some budgetary savings through tiaese institutions and subsidizing or taxing loans for particular purposes, as well as by selective direct ccnitrols on interest rates and loan allocation. 100. For China, continued experimentation and exploration in the area of investment decision-making and financing will be necessary, especially since the experience of other countries provides no precedent for a socialist finan- cial market. The relative shares and importance of the various possible ele- ments discussed above - vertical and compartmentalized, as well as horizontal - should be allowed to evolve with experience and the development of institu- tions. In general terms, however, it seems likely and desirable that diversi- fied investment arrangements will emerge as both consequence and cause of a diversified pattern of encerprise ownership.  - xxiv - Planning 101. Reform of the system of economic management along the lines of the Central Committee's 1984 decision will reduce the Government's direct involve- ment in production, commerce, prices, employment, and so on, but should not diminish its responsibility or capacity to promote and steer China's develop- ment. The preceding discussion has touched repeatedly on areas in which con- tinued, and often enLarged, direct government involvement is needed, espe- cially in establishing the physical, educational, scientific, institutional and legal infrastructure essential for rapid and efficient growth. The Gov- ernment will also continue to intervene directly in some other key sectors and enterprises, though more in investment decisions than in current operations. 102. Indirect Control. In the remaining areas of the economy, where decisions will increasingly be made by households, farmers, and independent enterprises, there will be a shift from direct to indirect control. This, however, need not diminish - and indeed might well increase - the Government's ability to manage the economy. For up to now in China, what has not been directly controlled has often not been properly planned. The broader reach of indirect controls, which shape the environment in which all economic units operate, could more than compensate for the associated loss of direct control over particular economic units. 103. Instruments of indirect control fall into two main categories - those connected with credit and those connected with taxes or subsidies. A third category is prices. But indirect control through administrative deter- mination of prices has the drawback of being able for any particular commodity to govern either demand or supply, but not in general both, and hence having to be supplemented with administrative quotas either on demand (in the case of shortages caused by a low price) or on supply (in the case of surpluses caused by a high price). Moreover, reduction in the scope of administrative price determination will gradually change prices - apart from their tax or subsidy element - from instruments of government control into independent indicators of value, cost and scarcity. These both influence and are influenced by the actions of producers and consumers, usually in economically appropriate direc- tions. Government intervention, however, may still be needed in some cases to damp fluctuations, as well as to set one very basic price, namely the exchange rate, which strongly affects exports and imports in a decentralized economy. 104. Control of the overall level and growth of credit, which depends not only on caution in financing budget deficits by note issue or borrowing from banks, but also on regulating the whole banking system (according to well- established rules), is a fundamental element of indirect macroeconomic manage- ment. And, as mentioned earlier, preferential interest rates and repayment terms in certain sectors or regions, as well as direct government intervention in the allocation of credit, are feasible and effective ways of influencing the composition of investment. But experience elsewhere suggests that they should be used cautiously, to avoid undesired side-effects. Low interest rates on rural credit, for example, can lead to uneconomic machanization in areas with surplus agricultural labor.  - xxv - 105. Personal income taxation, including progressive rates and selective relief, can influence the proportion of household income saved and the compo- sition of consumer demand. Above-average indirect taxes on particular goods (for example. luxuries) can reduce the amounts purchased and supplied by driv- ing a larger wedge between the price paid by the customer and that received by the producer; and vice-versa for below-average indirect taxes or subsidies (as on food, children's clothing, books and the arts in many countries). Taxes on enterprise profits, including selective concessions bv product or locality, can be designed to regulate the total amount of funds available to enter- prises, the proportion of profits reinvested, and the amount of external Einance used; and can stimulate investment in certain activities or regions while discouraging iC in others. Similarly, import tariffs can be used to promoce increased production in particular sectors through import substitution (although wichout complementary export incentives they produce an inefficient bias toward the domestic market), and in certain cases it is rational to tax exports. 106. This range of instruments can enable planning and markets to coexist reasonably harmoniously, though it involves greater emphasis than hitherto in China on managing demand rather than supply. Direct control of supply will, as mentioned earlier, continue in some important sectors. But in most cases, rather than dictating what should be produced, the Government could use taxes and credits to guide the changing composition of demand and to influence the relative profitability of different sorts of production and investment, allow- ing supply to respond through the decentralized decisions of enterprises and farmers. 107. Experience in other countries suggests that this indirect approach can be as effective as direct controls in shaping the pattern of development, and often more economically efficient. But the relationship between objec- tives and indirect instruments is much less precise and predictable, espe- cialLy in the absence (as in China) of much systematic study of the behavior and responses of economic units. This element of unpredictability can be reduced by research, and by flexible and responsive implementation, with adjustment of policies to achieve desired objectives over the medium term. But it can never be eliminated, and many countries (including Hungary in recent years) have experienced practical difficulties in indirectly regulating their economies. The increased uncertainty associated with greater reliance on indirect controls thus has to be weighed against the inefficiency of direct controls. 108. Scope of Plans. To the extent that direct control of production and allocation was reduced as part of the reforms, annual plans would become less significant in themselves, although they (and the budget, with which they should be linked) would remain important as a means of monitoring and adjust- ing medium-term plans. The medium-term plan, by contrast, would become the core of the entire planning process. Although the instruments for its imple- mentation could be different, the general form of the plan document might be much the same as at present. Its preparation could serve even more than now as a focus for the exchange of views among different government agencies, and between the center and local governments (who would still have their own, com- plementary, medium-term plans, at least at the provincial level). Other sorts of planning could also play an important role.  - xxvi - 109. One is what is called in other countries "policy planning". This is the design of packages of policies to achieve specified medium-term objec- tives - for example, improving road transport, or increasing machinery exports, or raising primary school enrollment. It is of particular value when indirect poLicy instruments are used, or when several different government organizations are involved, or when different objectives and poLicy instru- ments are interrelated. For this reason, policy planning is generally the responsibility of a high-level government agency, with strong analytical capa- bilities, which can evaluate alternative approaches to a given objective, make recommendations to the political leadership, and monitor implementation. 110. Another is location planning. Even with more economically ratienal pricing of transport and urban land, most countries have found that zoning plans and regulations are necessary to preserve urban amenities. Moreover, in industries where a single optimal-sized plant would supply a high proportion of the national market (for example, automobiles or electronic chips), but many alternative sites would be equally attractive in economic terms, the Government needs to arbitrate among competing localities to prevent wasteful duplication. In other industries (cement and fertilizer, for example), gov- ernment intervention is required because the location and scale of all the enterprises need to be simultaneously determined to minimize transport plus production costs. t11. In addition, Japanese experience suggests the potential usefulness of long-term sector planning, especially where major technological zhanges are expected or desired. (In China, such plans wouLd provide a particularly use- ful framework for establishing industry-specific priorities for modernization - a very difficult task because of the comprehensiveness as well as the back- wardness of the whole industrial sector.) However, Japanese experience also suggests that such plans should not be designed and imposed from above, and must have competitiveness in internacional markets as their not-too-distant objective. Through close contact with che enterprises in an industry, sec- toral pLanners need to create a consensual "vision" of future deveLopment, which increases the consistency (and reduces the risks) of investment deci- sions. This vision has to be backed up by, for example, government financial participation in - or loans to - key investments, incLuding pilot plants, or by temporary restrictions on imports and assistance to exports. But enter- prises that wish to pursue a different strategy should often be allowed to do so: though this may seem wasteful, the flexibility it provides has proved valuable, since neither planners nor enterprises can pick technological winners consistently. 112. Long-term (10- to 20-year) plans in individual sectors need not, and probably cannot, be coordinated in detail. But they could benefit from a pro- jected Long-term framework for the whole economy - a broad and fairly aggre- gated "vision" of future growth and structural change, worked out and regu- larly revised by central government planners. Such a plan could also provide guidance to enterprises in industries without specific sectoral plans, as well as for the planning of direct state investment in infrastructure and natural resource development, and as a context for the medium-term plans.  - xxvii - 113. In all these areas, although the common sense and good judgment of planners wilL remain fundamental, China could make some significant technical advances in planning methods, including greater use of quantitative models of various kinds. This could be assisted by improvements in economic training (for both planners and other relevant people). Increased use of computers could permit better and more flexible use of the large amounts of data now collected at lower levels. These, and other sorts of economy-wide and sec- toral information, should increasingly be disseminated to lower-level deci- sion-makers, including enterprises - who at present receive far Less informa- tion from the Government than in many other countries. Social Issues and Policies 114. China has been outstandingly successful in reducing extreme poverty. Although there is substantial income inequality between urban and rural areas and among different rural localities, the hunger, disease, high birth and infant death rates, general illiteracy, and constant fear of desti- tution and starvation that haunt very poor people in other countries have been more or Less banished. It will be of the utmost importance in the future co preserve and build on this achievement. But wilL the policies and institu- tions that have worked so well in the past meet the social objectives of a middle-income China? And are they compatible with changes in the system of economic management? 115. To a large extent, the answer to both questions is "yes." Social ownership of land and other productive assets, widely dispersed access to basic education and health care, and a guaranteed minimum food intake should all contribute to maintaining much less inequality and poverty than in other developing countries. Continued encouragement of agriculture and of other small-scale and labor-intensive activities should also - in light of exper- ience both in China and elsewhere - play a vital role. 116. However, economic reforms and greater technological dynamism could increase income disparities, while economically desirable changes in the price, wage and employment systems could make it more difficult to use these things as instruments of social policy. Demographic trends, and major changes in the pattern of disease, will also pose new problems. Some changes and innovations in social policies and institutions may thus be necessary. Income Disparities 117. Improving the allocation and motivation of skilled labor could involve some substantial increases in relative wages. Even higher incomes could be earned - and in some instances already are being earned - through entrepreneurship. Although perhaps not yet fully accepted in China, entrepre- neurship can make an important economic contribution. Entrepreneurs perceive and supply hitherto unmet needs, initially at a high profit, thus luring others into the same activities, which increases supply and drizes profits down. If allowed to, entrepreneurs in China could also - as in other coun- tries - play a major role in technological innovation, whose risks may require the stimulus of high rewards.  - xxviii - 118. In such cases, there is unavoidably some conflict between consider- ations of economic efficiency and the desire for minimal income inequality. But this conflict can be eased by progressive income taxation (tax rates that rise with income level) and by high indirect taxes on luxury goods. China already has a progressive personal income tax, aimed mainly at resident for- eigners, which could be extended to cover, say, the top quarter of personal incomes. Although this would have administrative costs, it would probably make it easier for the central government to prevent the informal income redistribution that is now reportedly occurring in some localities. 119. International experience also suggests the desirability of not attempting to conceal economically necessary income disparities. In some countries, for example, bribery has been aggravated by keeping the wages of public officials low for political reasons. In other countries, formal incomes are fairly equal, but for important people there is access to shops providing goods not available elsewhere, and sometimes even unrecorded but reguLar supplements to official salaries. Since these and simiLar practices quickly become well-known, their social benefits are temporary, while they seem to have a high long-term cost in terms of corruption and the evolution of a society based on special privileges. Social Security 120. At the other end of the income spectrum, a social security system consistent with a reformed economy is urgently needed. It will be increas- ingly difficult to reconciLe the need for greater enterprise independence and efficiency with the present role of enterprises as providers of social welfare benefits and life-time employment. Prices and wages will be increasingly governed by economic considerations, which will make it much harder to use them to guarantee minimum urban living standards. Faster structuraL change will involve contraction and closure of some enterprises. Some skills will become obsolete, and some workers - often through no fault of their own - may become at Least temporarily unemployed. 121. In devising new institutions and policies to tackle these problems, China might draw on the experience of the developed Western countries, all of which over the past few decades have established sociaL security systems aimed at providing an adequate income for aLl citizens and at eliminating poverty due to old age, illness, disability, and unemployment. These systems have not been free of problems, but have effectively attained their main objectives. All but a few developing countries, however, have been deterred from introduc- ing such comprehensive social security systems by their financial cost and administrative complexity. The challenge for China will thus be to devise a social security system that is effective, affordable and workable for a lower- middle-income country. 122. In developed countries, the social security system consists mainly of a state-run insurance scheme, financed primarily by compulsory wage-related contributions from workers and their employers, which provides old-age pen- sions, unemployment benefits, sickness and maternity benefits and so on, to those who have contributed. In addition, some noncontributory income supple- ments are usually provided to households whose per capita income (including  - xxix - social insurance benefits) would otherwise fall below some minimum level. This would be one possible approach for China, which could be implemented by consoLidation and extension of the present enterprise-based Labor insurance and welfare arrangements. 123. An alternative approach might be for the state in China to concen- trate its resources on noncontributory income supplements to low-income house- holds. These could perhaps advantageously be financed partlv from the taxes on high incomes mentioned above, although some additional contribution from general revenues might well be required. The total fiscal cost of these sup- plements could be substantial, but, because they would be targeted on the poor, would be a fraction of the cost of equivalent price subsidies, which benefit rich and poor alike. Their administrative cost could be minimized, and their effectiveness maximized, by making use of existing institutions. In urban areas, there is the household registration system used to administer grain rations. This might be a particularly useful basis for providing income supplements to households with high dependency ratios that would otherwise on balance lose from the upward adjustment of retail prices and wages mentioned earlier. In rural areas, the scheme could be implemented by the administra- tive successors to communes, brigades and teams. 124. In administering either state-run social insurance or an income supplement scheme. China, Like other countries, would face some difficult choices concerning coverage and benefits. There might be a uniform urban scheme, including the self-empLoyed as well as state and collective empLoyees. It could be hard, however, to extend the same scheme to rural areas, because incomes are much lower and vary widely among localities. A compromise aLso always has to be struck among the desire to alleviate existing poverty, the need to give people an incentive to avoid poverty, and the wish to minimize budgetary and administrative costs. For example, setting unemployment benefits high, though it reduces hardship, also reduces the incentive to seek (or remain in) work, particularly if the alternative to unemployment is a monotonous or arduous job at a Low wage. 125. Especially if noncontributory income supplements were chosen in preference to state-run social insurance, the state should encourage collec- tive and local development of contributory pension and insurance schemes. Some of these pension and insurance schemes might, as in other countries, be organized by enterprises and other employers. But they should be financially independent (or "funded"), so that, for example, a scheme would be able to continue paying pensions even if the enterprise that had originally organized it were closed. Such pension and insurance schemes, which would have substan- tial funds to invest, might play an important role in a socialist financial market. 126. Because many complex issues are involved, and because improved social security arrangements could be essential to a successful reform of the economic system, a commission could perhaps be established soon to formulate proposals. These might be implemented in stages, and should probably initi- ally be of a modest kind. But the long-term objective should be to maintain China's social achievements, as in the past, well above those of other coun- tries at comparable income levels.  - xxx - Housing and Social Services 127. China's Large and medium-sized enterprises are now unique in the degree to which they provide housing and education and health services to workers and their families. As with social security arrangements, it would be much easier for enterprises to function as independent, efficient economic units if these obligations were removed from them. This would enable managers to concentrate on economic activities, facilitate labor mobility, reduce the social problems caused by the release of unneeded workers and closure of inefficient enterprises, and diminish the present reluctance of many workers to consider employmenc in small urban enterprises. 128. The social services now provided by enterprises could be taken over by local governments. So might some of the housing. But experience in China and elsewhere (including Eastern Europe) suggests the advantages of giving households as much responsibility as possible for their own homes, through individual ownership or housing cooperatives. This can provide a powerful stimulus to personal saving, release government resources for other purposes, and produce a higher quality and better maintained housing stock. The Govern- ment could usefully assist this, as mentioned earlier, by selling more exist- ing urban housing to tenants, as well as by "sites and services" projects (planning and providing basic utilities for new individually-constructed hous- ing), coupled with technical assistance and limited subsidies or tax conces- sions to housing cooperatives. At the same time, universal entitlement to a minimum standard of accommodation could be maintained, through both income supplements and direct provision. 129. Increased central and provincial government support of rural social services seems to be needed, especially in poor and backward localities. The introduction of the production responsibility system adversely affected social services in such places, wich a sharp decline in cooperative health programs, as well as in primary and secondary school enrollment rates, especially among girls. This is unfortunate, especially since powerful eviden-> from other countries confirms that improvement of basic education and heaLth is one of the most effective ways of helping poor people, as well as contributing to economic growth. Thus although self-reLiance may remain an appropriate policy in more prosperous rural areas, the state should consider financing a larger proportion of the costs of social services in poor areas, for !xampLe through matching grants. Population 130. A further reason for taking financial and other steps to maintain and improve basic social services is that studies in other countries have shown that both education - especially of females - and health are major con- tributors to voluntary fertility reduction. As noted earlier, the expected slow rate of population increase, due also to successful birth planning poli- cies, is an important ingredient of China's relatively favorable economic growth prospects. Given, in addition, the limited amount of cultivable land in China, the achievement of replacement level fertility (2.1 children per woman) is clearly desirable.  - xxxi - 131. It might not be so desirable, however, to hold fertility well below the replacement Level for a long time (which would be necessary, for example. to attain the ultimate population of 700 million suggested by some Chinese demographers). This is because the gains from having a smaller population in the late twenty-first century could be more than offset by the costs of tran- sitional changes in the population's age structure. Of particular concern would be the long-term impact of very low fertility on the labor force, which could decline rather rapidly in the middle decades of the twenty-first cen- tury. Negative labor force growth might have some economic advantages, but would inevitably cause problems of adjustment. There would also be a rise in the average age of the labor force, with shortages especially of younger, more recently trained, more adaptable workers. In addition, maintaining very low fertilitv could cause the proporcion of elderly (65 and over) dependents to bulge temporarily above its ultimate long-term level. 132. The increase in the proportion of elderly people will be a serious social issue in the twenty-first century even with a more moderate reduction in fertility. It will increase from about 5% today eventuallv to about 20%. Though there will be an offsetting decline in the proportion of young depen- dents, providing financial sunport co elderly people is more costly, and more difficult ro organize. China's existing institutions, including the family, for supporting elderLy people are unlikely to be able to cope with their increased numbers. This strengthens the arguments mentioned above for start- ing now to plan long-term reform of the social security system. The early establishment of wider and better pension schemes in rural areas would, among other things, make it easier to reduce fertility. Health 133. The aging of the population will also contribute to the fundamental changes that are occurring in the nature of China's health problems. These have already moved from the pattern typical of low-income countries to one where the leading causes of death are heart disease, strokes, and cancer, as they are in high-income countries. These chronic diseases constitute a fundamental challenge for China, which other countries have had difficulty in meeting despite vast financial outlays. 134. One essential ingredient of the required "second heaLth care revolu- tion" will be prevention. Some steps - particularly control of salt intake and tobacco consumption - could be taken straight away. But much additional research is needed to identify and field test appropriate preventive strate- gies, which are much less straightforward than with communicable diseases. Another essential ingredient will be strategies for dealing with the large number of cases of chronic disease that will, inevitably, occur. These stra- tegies should include capacity for treatment where treatment holds promise of results and can be afforded; they should pay careful attention to affordable plans for rehabilitation of incapacitated individuals; and they should be concerned with humane care for the terminally ill (an area in which there have been major and quite affordable advances in Western medical practice).  - xxxii - 135. A third ingredient of the second health care revolution will be designing an insurance and financing structure that encourages prevention, and discourages current tendencies toward overuse of faciLities and introduction of high-cost procedures. Assembling these three ingredients will, inevitably, be difficult. But, to the extent that success is achieved, China could become a world leader in the effective and humane handling of the burden of chronic disease without succumbing, as other countries have, to endlessly costly investments in medical technologies of limited efficacy. Rural Poverty 136. Increases in agricultural labor productivity should permit agricul- tural incomes to grow quite rapidly over the next two decades. It is not easy, however, to find a pLausible set of assumptions under which agricultural productivity would grow fast enough to narrow the large agriculture-nonagri- culture earnings gap (even in proportionate terms) until the twenty-first cen- tury. Because members of farm families will increasingly be engaged in non- agricultural activities, their total incomes will rise faster than their earnings from agriculture, but could still remain unsatisfactorily low. Addi- tional measures to raise them are worth considering. 137. Higher agricultural product prices and lower prices for agricuLtural inputs couLd conflict with other objectives of agricultural pricing policy, including avoidance of surplus production and encouragement of efficient input use. The present two-tier pricing system, though, would in principle make it possible to boost the incomes of farmers by increasing offical procurement prices, while allowing market-determined above-contract prices to balance sup- ply and demand for the various agricultural commodities. But this would be quite complicated, especially if unacceptable inequities among farmers in dif- ferent places and situations were to be avoided. Direct income supplements to broad categories of farm households might pose fewer administrative problems. 138. Increased government expenditure on things that mainly benefit the agricultural population, financed by taxation of other sectors, could provide another means of transferring nonagricultural productivity gains to those who must remain farmers. Development of agricultural and other rural infrastruc- ture, improvements in agricultural research and extension services, and sup- port to rural social services are obvious examples. Another possibility is subsidies to rural participants in social security schemes. 139. Though generally low agricultural incomes are an important problem, the most serious rural poverty in China is concentrated in specific locali- ties, where incomes are far below the average. (For example, in Dongye town- ship, in Dingxi county, per capita income in 1983 - the township's best-ever year for agricultural production - was about one-sixth of the national rural average.) Many such localities will benefit absolutely, and some relatively, from reform of the economic system - this appears to be the reason why the production responsibility system has not yet generated the anticipated increase in overall rural income inequality. But others will tend to remain very poor, and could even become worse off, especially if they are remote or have meager land and water resources.  - xxxiii - 140. To solve the problems of poor localities would generally require packages of complementary measures, whose composition would vary according to the particular circumstances of the locality and other government priori- ties. Improved education and heaLth services, in addition to meeting immedi- ate social needs, build up capacity to work and innovate. Focused research and extension services can release untapped agricultural potential. Agricul- tural specialization and industrial development can be made possible by better roads and other infrastructure, and assisted by temporary subsidies for local industry. Allowing part of the population to move elsewhere, thus increasing per capita land and water availability, can also be an effective and economi- -al way of reducing poverty. Overview 141. To sustain rapid growth for several more decades will be a hard and complicated task for China. 3etter access to advanced foreign technology than in the past will make it easier, but wilL not be sufficient: the cumulative economic effect of individually inconspicuous improvements in processes and products has been shown to be larger than that of radical innovations, and the size of the technological gap between the best and the average enterprise in an industry to be as important as the technoLogicaL level of the best enter- prise. Nor will availability of energy, land and other natural resources - despite their importance - be the deciding factor. Cbina's economic prospects will depend, rather, on success in mobilizing and effectively using all avail- able resources - especially peopLe. 142. This in turn will depend largely on success in reforming the system of economic management, including coordinated progress on three fronts. First, greater use of market reguLation to stimulate innovation and effici- ency. Second, stronger planning, combining indirect with direct economic con- trol. Third, modification and extension of social institutions and policies to maintain the fairness in distribucion that is fundamental to socialism, despite the greater inequalicy and instability that market regulation and indirect controls would tend to cause. 143. On each of these fronts, there are promising ways forward, but also problems and hard choices to be faced. In addition, it is hard to overstate the importance and difficulty of striking a correct balance among the three. Very few countries have combined state and market regulation in such a way as to produce rapid and efficient growth, and fewer still have also managed to avoid intolerable poverty among substantial segments of their populations. On the contrary, there are far more countries in which unhappy combinations of plan, market and social institutions have produced neither rapid growth, nor efficiency, nor poverty reduction. 144. There is thus a vital need to guard against losing the strengths of the existing system - its capacity to mobilize resources, as well as to help the poor - in the course of overcoming its weaknesses. This will surely not deter China - a successful pioneer in many areas - from moving ahead. But it argues for a gradual advance, with experimentation and evaluation at each step, even though a one-stroke change would in principLe involve fewer inter- nal inconsistencies. Experience in Eastern Europe also suggests the impor-  - xxxiv - tance of moving steadily, and of trying to avoid ill-judged steps in the direction of market reguLation that subsequently have to be reversed or admin- istratively tampered with, thus creating needless uncertainty. 145. Not all the steps need be small, though. In some cases, despite the greater risk of error, it may be best to introduce substantial packages of simultaneous reforms. Nor need progress be slow. What has been accomplished in China's rural areas in the past few years has provided not only an example, but also an excellent opportunity and indeed a vital need, for complementary and similarly rapid progress in the urban economy. Though in many ways more complicated and troubLesome than rural reform, urban economic reform probably has the advantage of not needing to be so uniform. Other countries have suc- cessfully applied different management methods in different sectors and enter- prises, and China should be able to do the same, while constantly seeking to refine and improve the mixture. 146. In system reform, and in the many other areas covered by this report, both the potential for progress and the problems involved are so large, and there is so much that is without historical precedent, that an even-handed and credible conclusion may be impossible. At a mininum, though, China's long-cerm development objectives seem attainable in principLe, and if recent experience is any guide, there is a good chance that they will be attained in practice.  1. GROWTH AND CHANGE 1.01 Over the next two decades, China - now a Low-income country - will become a middle-income country. This chapter explores possible implications for China's development strategy of international experience during the low- to-middle-income transition. Its three main sections deal with macroeconomic change, microeconomic change, and inequality. But first some objectives - both for China and for this report. A. Objectives 1.02 China's ultimate economic objective is to catch up with the devel- oped countries, and as quickly as possible. The impressive progress of the first three decades of the People's Republic, described in an earlier World Bank report (China: Socialist Economic Development, 1983) represents the first step towards that goal. The next step envisaged by Chinese leaders is to give most Chinese people a relatively comfor1able living standard and eliminate the worst manifestations of poverty.' Towards this objective, quadrupling of agricultural and industrial output between 1980 and 2000 has been set as a target, with per capita GNP to increase from about $300 to $800 (in 1980 dollars). Beyond this, the path is as yet uncharted. But the direction of development during the next two decades will powerfully and perhaps irreversibly influence China's economic prospects in the twenty-first century. 1.03 The magnitude of the task ahead is readily illustrated. In real (purchasing power) terms, GNP per capita 2^n the developed countries is approx- imately ten times greater than in China,- and will increase in the future by at least 2% and perhaps 3% per year. To catch up by 2050, China's GNP per capita would have to increase at an average rate of at least 5.5Z and perhaps 1/ "A reLatively comfortable living standard" is a rough translation of the Chinese term xiaokang, which originated from a Chinese classic Li Ji. In that context the term was contrasted with datong, which refers to a very rich society. Some people in China now classify the various stages of development in terms of improvement in living standards, as follows: jihan (poverty), wenbao (basic needs satisfied, which is considered applicable to the present situation in China), xiaokang (which is to be achieved by the end of this century) and fuyu (rich). See Liu Guoguang (ed.), Zhongguo Jingji Fazhan Zhanlue Wenti Yanjiu (Issues of China's Economic Development Strategy), Shanghai 1983, p. 98. 2/ At actual prices and exchange rates, the gap is much larger - a factor of thirty rather than ten. But the huge nominal income gap between rich and poor countries is apparently due more to differences in prices than to differences in purchasing power, large though the latter are. See 1.B. Kravis and others, World Product and Income (World Bank, 1982), and China: Socialist Economic Development (World Bank, 1983), Main Report, para. 3.32.  6.5Z per year (the Government's 2000 target implies a rate of 5.0%). Such rapid progress has been rare elsewhere. In 1960-82, excluding small countries and economies dominated by oil, only two developing countries achieved annual per capita income growth above 5Z: South Korea (6.6%), and Greece (5.2%). 1.04 More generaLly, only one country - Japan - has indisputably caught up with the developed nations from a position of economic backwardness. Sev- eral Latin American countries for example, including Brazil and Chile, entered the modern growth phase at least as early as Japan, and have subse3ouently experienced periods of rapid growth, but still remain far behind.- Even the Soviet Union, with an exceptionally high rate of investment in physical and human resources over many decades, has raised its per capita GNP in ret4 terms to only about half the US (and two-thirds of the West European) level._ 1.05 China's past record, on the other hand, is quite encouraging. From 1952 to 1982, despite relatively rapid population growth and periods of acute economic mismanagement, per capita national income grew ae ,an average annual rate of 4.0%, with phases of significantly faster growtn.-_ Of particular relevance is China's performance during the period of policy and system reforms that began in 1979 (described in more decail in World Bank report 4072-CHA, March 1983). From 1979 to 1984, largely as a result of much accele- rated agricultural growth, per capita national income grew at 6.8% per year. 1.06 The objectives of the present report are to identify some of the key issues and problems that China may face in the next 20 years, and to review, in light of international experience, some of the options for addressing them, with particular emphasis on decisions that might need to be taken soon. As well as considering issues within specific sectors, such as agriculture and energy, and intersectoral linkages - the ways in which prospects and problems in each sector depend on circumstances and decisions in other sectors - the report gives special attention to cross-sectoral or economy-wide issues, including human and financial resources, spatial location of economic activ- ity, and reform of the system of economic management. 1.07 The future is inherently unpredictab'le, and the most careful fore- casts are often confounded. The lessons of international experience are com- monly either obvious (though frequently ignored) or ambiguous and controver- sial, and in any event are hard to apply to China, a country that in important respects differs from all others, and is not easy for outsiders to under- stand. Thus although this report is about planning in the broadest sense, its projections and calculations are not predictions, and its suggestions are no more than an attempt to contribute to the debate in China about the difficult 3/ See for example L.G. Reynolds, "The Spread of Economic Growth to the Third World: 1850-1980", Journal of Economic Literature, September 1983. 4/ This is the conclusion of several Western studies. Soviet calculations suggest smaller gaps. 5/ Statistical Yearbook of China, 1983, page 6.  questions and problems that must be confronted during the country's unique socialist modernization. B. Macroeconomic Change 1.08 China's future economic growth will require and cause changes in economic structure. At the macroeconomic level, the pattern of structural change associated with rising income has been broadly similar in most coun- tries, and many of its features are likely to recur in China. But because individual countries have deviated significantly from the average pattern, it is also important to consider China's particular starting point and circum- stances. 1.09 6 igures 1.1 to 1.3 summarize some essential results of available research.- Each figure juxtaposes three things: China in 1981; a typical large low-income country; and a typical large middle-income country. The first figure deals with the sectoral composition of domestic demand, the second with the sectoral composition of production, and the third with the sectoral composition of employment. China's Present Economic Structure in International Perspective 1.10 The figures show that China is in essence a low-income country with an unusual pattern of domestic demand and production. Output per agricultural worker is small and agriculture's share of total employment is large - similar to the typical low-income country. Output per worker in manufacturing is higher than in the typical low-income country, because of massive investment in heavy industry; but with nonagricultural employmen only 30% of the total, this raises China's per capita income only modestly.-' 1.11 The most unusual aspect of demand in China (Figure 1.1) is the high share of savings - unique among low-income countries, well above the average for middle-income countries, and matched only by Japan and the East European socialist countries. However, although household consumption comprises only 6/ The discussion in this section is based on Annex E, which compares the present structure of China's economy with those of other countries, and on Background Paper 9, which presents the results of empirical research on the structural changes normally associated with the low-to-middle- income transition in large countries. Figures 1.1-1.3 are derived from Tables 3.1, 3.6, 3.8 and 3.9 of Annex E. In the figures, per capita GNP is taken to be $300 and $1,500 (1981 dollars) for the typical low-income and middle-income country, respectiveLy. In the Annex tables, the predicted values for large lower-middle and upper-middle income countries are presented separately. 7/ As explained in Annex E, China's per capita GNP in 1981 is $300 at Chinese prices and official exchange rates, and is tentatively estimated to be $350 at prices comparable to those in other developing countries.  - 4 - Figure 1.1: Composition of Aggregate Demand (shares of GDP) Fofen Baroa-O Saiwing Ivsmn Irvestrrent S-n ~~~~~~~-biC _________--________ j nn t RJblic to to _ --- (unptionb J ConumCon sumpti Public=PuIi Consmmfion Househo4d Consumption HoLsehold Consumption Household ConRrhton China TYPiCot TVPpi LaOW Lrge Low-InCome Middle-Income Country Country World Bank-27327  -5- Figure 12: Production Structure (sectoral shares ot GDP) ROG MFG ~~~~~~~~~~~~~MFG SER~~~S SER~~~~~~~~~~~S! SE!?~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~~~~~E ____________ _ _ _ _ _ _ _ PAN UN NNN~~~~M AGR ~~~~~~~~AGR? AG!? China TVpical TyPical Large Large Low4rcome Middce-mnCme Cotunfty Coutrv Legenct AGR = Agculture MIN = tnwng MFG = Manukactriing 11,11' - -1n struckure (eleclrcilv. trmort. onsaruction) SEP -- Serices (induding haoing) World Bar*-27326  - 6- Figure 1.3: Employment Structure (sectoral shares of total labor force) MFG hAFG~~~~~~~~~~~~F INF INF SER SER INF MIN- AGR AGR AC.R China TypicCd Typical ILow-income Miciddle-1ncoe tCountrV Counfty Leoend: AGQ --Agriculture MI1N = Mnirs LIFG = Nmoulocturing INF = In*yshutLue (elecb dy onspo,t.constuction) SEPR -Sefvies (irncludirng housing) V* Bcznk-27325  half of China's low per capita income, basic living standards - as measured for example by calorie intake, literacy and life expectancy - are similar to those of middle-income countries. This has been achieved partly through a share of public spending (particularly on education and health) well above - and better focused on basic needs than - that of the typical low-income country, partLy through imposing an austere but cost-effective pattern of household consumption. Although the share of food in household consumption expenditure is quite normal for a Low-income country, staples such as grain constitute an unusually high proportion of food consumption. 1.12 Especially because exports and imports are small in relation to total output (though not unusually so for a very large country) and almost balance in both agriculture and manufacturing, China's unusual pattern of demand contributes to an unusual pattern of production (Figure 1.2). The high saving share has its counterpart, because of the need to produce a large volume of investment goods, in a manufacturing share far Larger than that of the typical low-income country, and similar to that of the typical middle- income country. The share of agriculture in total production, by contrast, resembles that of the typical low-income country, simply because of the need to allocate a substantial proportion of limited resources to feeding the popu- lation. The share of infrastructure (electricity, transport, construction) is also simiLar to that of other low-income countries, despite China's large manufacturing share. 1.13 The share of services in output is thus umuch smaller (about 17% of CDP) than in the typical low-income country (35%) or middle-income country (40%). The share of education and health services is (as already mentioned) significantly above average. The Low overall service share thus mainly reflects the small shares of commerce (including restaurants), banking and finance, and miscellaneous business and personal services. This in turn reflects the small share of household consumption in national income and the high proportion of food consumed directly by its producers, which tend to reduce trade (especially retail), as well as past official neglect of personal services. In addition, China's present system of economic management, and the high degree of enterprise and local self-sufficiency, reduce the provision of (especially wholesale) trade, financial and other business services. 1.14 Although the share of agriculture in total employment in China is similar to that of the typical low-income country, the composition of nonagri- cultural employment is rather different (Figure 1.3). The share of manufac- turing in employment is significantly higher than in the typical low-income country, but by less than its share in output, since output per worker in manufacturing in China is unusually high (which in turn mainly reflects the unusually low share of labor-intensive consumer goods within the manufacturing sector). The share of services in employment in China is correspondingly lower than in the typical low-income country. 1.15 International comparisons of economic efficiency are difficult. But the available evidence (reviewed in Annex E) gives considerable support to the view that China's economy is relatively inefficient. Usage of intermediate material inputs - most conspicuously energy, but also other industrial and agricultural products - seems high. (Indeed, inefficient use of industrial  -8- intermediate inputs appears just as important as the high saving rate in explaining the large share of industry in China's total production.) The data do not permit a firm conclusion regarding efficiency in the use of fixed capi- tal, but the amount of circulating capital used appears unusually high. Possible Implications for China's Future Development 1.16 Partly because China now differs in important respects from any other deveLoping country, low or middle income, the macroeconomic structural changes that occur during the typical low-to-middle-income transition (summa- rized in Figures 1.1-1.3) are of only limited relevance to China's future development. Nonetheless, some implications - explored further in subsequent chapters - can be drawn from individual elements of this transition, the experiences of some other atypical countries, and China's unusual starting point. 1.17 Saving. China's saving rate is already so high - about 30Z of national income - by international standards that the usual low-to-middle- income trend of a rising saving rate (Figure 1.1) is unlikely to apply. In principle, indeed, China could reduce saving to a level - perhaps 20% of national income - similar to other developing countries. But this would be incompatible with China's long-term growth targets. Middle-income country levels of output per worker require about five times as much capital per worker as in a low-income country; to accumulate this capital rapidly requires a high investment rate, as for example in other fast-growing East Asian economies. 1.18 China has, though, the option of supplementing domestic savings with external finance. In recent years, as Figure 1.1 makes clear, China's balance of payments surplus has added to the need for domestic savings - reducing con- sumption below the already low level required to finance domestic invest- ment. By contrast, the typical large developing country (Low or middle income) runs an external deficit, with domestic investment exceeding domestic savings, thus easing the conflict between consumption and growth. Recent international experience, however, has underscored the need - discussed fur- ther in Chapter 6 - for caution and good judgment in external borrowing, and the essential role of matching increases in export earnings. 1.19 Efficiency. A high rate of investment (in education and training, as well as physical capital), though necessary, will not be sufficient to attain China's growth targets. It must be accompanied by increased efficiency and improved technology, which Western economists call growth of total factor productivity (TFP) - the difference between the rate of growth of aggregate output and that of some aggregate of inputs. 1.20 Though different studies have often yielded significantly different estimates for particular countries and time periods, most studies suggest that TFP growth has typically contributed about one-third of aggregate net output growth in middle-income developing countries, and nearly one-half in indus- trialized market economies. In the Soviet Union, TFP growth has apparently been slower than in most industrialized market economies, and has contributed a smaller proportion of output growth. Rough calculations for China suggest  -9- that TFP growth during 1952-82 was on average low by comparison with other developi7g countries, though better in some sectors and subperiods than in others.- 1.21 Research on the causes of variations in TFP growth across countries and sectors and over time has had limited success, partly because of diffi- culty in quantifying potential explanatory variables (such as improved tech- nology and knowledge, increased division of labor, incentives to economize on materials and capital, and to use workers more effectively). However, differ- ences in TFP growth among sectors are apparently less systematic and pro- nounced than differences among countries: in a fast growing country, TFP growth in all sectors tends to be higher than in other countries. This - like China's experience since 1978 - confirms that the general orientation of poli- cies and the system of economic management are crucial determinants of the pace of economic advance. 1.22 The same finding - and much other evidence, including China's own experience - confirms the need for a balanced allocation of resources among sectors. In particular, favoring industry at the expense of other sectors does not seem to generate rapid and efficient growth, even though industry is usually the fastest-growing sector and the shift of employment from agricul- ture to industry contributes significantly to overall productivity growth. Countries that have stimulated agricultural production and incomes have gene- rally experienced faster, rather than slower, industrial growth. Similarly, even in countries where manufacturing production has grown very rapidly, infrastructure and services - which are vital for industrial (and agricul- tural) efficiency - have absorbed a much larger share of resources. For exam- ple, in South Korea in 1965-74, manufacturing accounted for only 27% of total annual fixed investment, excluding housing, and in Japan in 1951-65, 30Z. In China, the corresponding share in the period 1950-80 seems to have been close to 40% (Annex E, Table 3.11). 1.23 Consumption. Like the domestic saving rate, the share of public consumption in national income in China is already above the normal level for a middle-income country, and will thus not necessarily follow the rising trend observed in other countries (Figure 1.1). Indeed, it could even fall, given the large weight of education in government expenditure, because of the large and highly untypical prospective decline in primary school enrollment in China over the next two decades, as a result of low birth rates (see Annex A). Increases in enrollment at higher levels, improvements in educational quality, and pressure for greater public expenditure on health, administration, defense and other public services will, on the other hand, tend to increase public consumption. Clearly, the Government has many options, whose budgetary impli- cations can only be assessed in relation to possible trends in subsidies, transfer payments, and public investment - discussed in Chapter 9. But public consumption (much of which is in fact essential investment in human skills and knowledge) in the narrow sense of current purchases of goods and services seems unlikely to rise faster than national income in the next 20 years. 9/ Annex E, Table 3.16; Chapter 7, Table 7.1.  - 10 - 1.24 Since household consumption in China is probably more concentrated on basic necessities than in the typical Low-income country, it will almost certainly change, as per capita income rises, in the same directions as it has changed in other countries. These include a decline in the share of expendi- ture on food, with an increase in the relative weight of nonstaple foods, especially meat; a slight rise in the share of expenditure on clothing; and a sharp rise in expenditure on other items, most notably those connected with housing, travel and recreation, and with a disproportionate increase in expen- diture on reLevant se-vices (as distinct from goods). 1.25 However, with any given leveL and distribution of income, the com- position of household consumption varies significantLy among countries. This appears to be caused by variations not so much in national tastes (which influence only the details) as in the relative prices and availabilities of different items, which in turn vary partly because of natural resource endow- ments, partly because of government policies. For example, people in Japan, as compared with those in the US, eat less meat and dairy products and are less well housed, mainly because Land is scarcer, but also because imports of agricultural products are restricted to protect Japanese farmers. In the Soviet Union and Eastern Europe, retail trade and other services are a much smaller share of household consumption than in Western countries, mainly because pLanners have allocated few resources to these activities. 1.26 China's future pattern of househoLd consumption, though following certain universaL trends, will thus to some degree accommodate itself to phy- sical constraints, as well as to government policies affecting the supply and prices of particular goods and services. Some of the Government's policy options (especiaLly as regards consumption of food, energy and services) are examined in Later chapters. Naturally, at a given level of aggregate real consumption, if Chinese people are induced - by high prices or restricted quantities - to consume less of one thing, they must also be induced or allowed to consume more of someching else. But planners often find it diffi- cult to identify items (other than basic necessities) whose consumption should be disproportionately increased. In most market economies, by contrast, pro- fit-seeking enterprises seek continuously to produce consumer goods and servi- ces whose consumption can be expected to increase unusually rapidly (sometimes with the assistance of advertising). Changes in China's system of economic management (Chapter 10) may thus be needed, among other reasons, to enable household consumption to grow at the same high rate as national income. 1.27 Foreign Trade. Choices connected with the level and pattern of for- eign trade are less significant for large countries than for small ones, since the former generally have more balanced resource endowments, higher internal transport costs, and domestic markets large enough to realize economies of scale in basic industrial activities. China appears to be an exception to this generalisation only in respect of resource endowment: its 0.1 ha of cul- tivable land per person is one of the lowest ratios in the world. The long- term possibility of net agricultural imports paid for with manufactured exports - the pattern observed in South Korea and Japan - thus requires care- ful investigation.  - 11 - 1.28 But China's foreign trade options (Chapter 6) are significant mainly because of the relationship elsewhere between foreign trade and economic per- formance. Developing countries - including Japan - that have become sucessful exporters, especially of manufactures, have generally also experienced faster growth. The causation runs in both directions; but exposure to trade seems to be a necessary ingredient of rapid and efficient growth. Thus, although the ratio of exports (and imports) to national product is typically no higher in large middle-income countries than in large low-income countries, the possi- bility of raising China's foreign trade ratio seems worth considering. 1.29 Service Sectors. A decline in agriculture's share of production and employment is the most universal feature of economic development, and one that will undoubtedly continue in China. The two issues are, rather, how fast this decline will occur, and what the relative sizes of the various nonagricultural sectors will be. Especially as regards employment, the two issues are inti- mately related, since, in the usual pattern of development, the main shift is from agriculture into services, with a more moderate increase in employment in industry, which is Less labor intensive. The conspicuous smallness of the service sectors in China at present raises questions about their future con- tribution: wiLl they continue to be unusually small; or will they - as in the past few years - expand unusually fast? 1.30 China's small commerce, finance, personal and enterprise support service sectors, partly offset by a large education and health sector, consti- tute a pattern that also emerges when the Soviet Union is compared with Western countries at comparable income levels. Neither in the Soviet Union nor in China can this pattern belexplained by the "concealment" of service activities within other sectors- As in other countries, the share of employment in services in the Soviet Union has risen with income, but has always been at an unusually low level, with a correspondingly above-normal share of employment in agriculture. There are variations aiso among develop- ing countries. But the striking contrast between the Soviet Union (and Eastern Europe) and virtually all other countries (including Japan, which has always had a large service sector) is more indicative of the range of options China has. 1.31 This contrast also strongly suggests that rapid service sector expansion (discussed further in Chapter 2) and reform of China's system of economic management are closely interrelated options. Expansion of commerce would go hand-in-hand with expansion of the role of market regulation and increased specialization of production units and localities. Expansion of 10/ C. Ofer, The Service Sector in Soviet Economic Growth, Harvard University Press, 1973. See also Annex E of this report. Nor can this pattern be explained by the classification in China and the Soviet Union of many services as not materially productive (and hence excluded from national income): the shortfall is in fact more conspicuous for material services (coimerce and enterprise support services) than for nonmaterial services (where the gaps in finance and personal services are offset by education and health).  - 12 - enterprise support services, including finance, accounting and law, would likewise be a corollary of increased enterprise independence and specializa- tion. And expansion of personal services would reflect a shift in the balance between planners' priorities and peoples' prefe-ences as criteria for allocat- ing resources. More generally, rapid and efficient expansion of the service sectors probably could not be accomplished through administrative directives and centralized resource allocation. International experience strongly sug- gests that growth of services should be pulled by demand, with government in a permissive rather than an active role. Individual and collective enterprises may also be better suited than state enterprises to providing high-quality, flexible and customer-oriented services. C. Microeconomic Change 1.32 The changes in macroeconomic structure ilLustrated in Figures 1.1- 1., and discussed in the preceding section, are sLow - at most, a few per- centage points per decade. Moreover, these broad structuraL shifts are more a consequence than a cause of growth. By contrast, microeconomic structural change (affecting individual products, processes, enterprises, and workers) is typically much faster; and its speed is a crucial determinant of the aggregate growth rate. 1.33 Microeconomic change contributes to growth mainly by enabling cheaper and becter products and processes - in agriculture and services, as well as in industry - to displace more expensive or inferior ones (Chap- ter 7). This involves introducing new technology - better seeds, faster machine tools, compucerized inventory control - and abandoning obsolete pro- ducts and equipment. It involves increasing specialization, which realizes technical economies of scale and allows people to become expert in particular sorts of work. It involves responding to needs and opportunities created by alterations in economic circumstances - energy and materials prices, the pat- tern of demand, transport facilities, and so on. It involves innumerable small modifications to existing products, processes and practices - in pro- curement and marketing, as well as production - which cumuLatively can reduce costs and improve the quality of products and services by striking amounts. 1.34 Much microeconomic change takes place within existing production units - farms and enterprises - sometimes involving radical transformation of the product mix. But microeconomic change is also associated with the birth of new enterprises, marked changes in the relative sizes of existing enter- prises, and the disappearance - through closure or merger - of particular pro- duction units (Box A). For individual workers, change may be less rapid, partly because workers with obsolete skills retire and are replaced by new entrants with new skills, partly because processes change less than products, and partly because changes in enterprise product mix reduce the need for work- ers to move among enterprises. But microeconomic change inevitably entails considerable relearning of skills and movement of workers among jobs and - in most countries - enterprises (Chapter 8). It may also affect the relative prosperity of different localities - sometimes in a persistent and even cumu- lative way - and thus create pressure for geographical movement of people and capital (Chapter 5).  - 13 - Dcx 11: SIZZ ND MM UGOK OF ErERUPISES XS MASTo E(rRlIIS A. I The seemingly disorderly process of are rapidly propelled into the large enter- enterprise births and deaths and widely dif- prise category. while an equally significant fering enterprise growth rates in most market number collapse altogether. economies has a strong underlying element of evolutionary natural selection: enterprises A.2 There is also a strong association, that provide cheaper and better products or independent of size, between profitability and services, and market them more effectively, growth; enterprises that are more profitable expand at the expense of their less efficient also grow faster. This is largely because rivals, yhich are forced to improve or to dis- they depend mainly on retained profits to appear:l/ Studies - mainly of American and finance lnvestment needed for expansion, but British enterprises - have also established also because higher profits make it easier to that this proc ss exhibits certain statistical borrow and issue new shares. Since greater regularities.._ Small enterprises are a large profitability generally reflects greater effi- and stable proportion of the total (whose ciency, this financial linkage between profit- overall size distribution is approximately ability and growth plays a crucial role in log-normal)_ On average, small enterprises enabling efficient enterprises to expand at grow just as fast, and are just as profitable, the expense of less efficient ones. Success as large enrerprises; but small enterprises and failure, moreover, are persistent: enter- display much more diversity of growth and pro- prises vith above-average growth and prof ita- fitability. which means that the most profita- bility in one (say five-year) period are ble and fastest growing enterprires, as well likely also to have above-average performance as the least profitable and slowest growing in the next period. But the evolutLonary pro- (or contracting) enterprises, are mainly the cess takes time: in every sector, there Is smaller ones. As a result, a significant pro- always a wide spectrua of enterprise profit portion of initially very small enterprises and loss rates. I/ Several relevant studies are surveyed in R. Nelson, "Research on Productivity Growth and Productivity Diferences", Journal of Economic Literature, September 1981, pp. 1059-61. / Many atudies are surveyed in R. Marris and A. Wood, The Corporate Economy, London, 1971, Appendices A-C. 1.35 The ingredients of microeconomic change described above can be dis- cerned, in some form or to some degree, in China over the past three dec- ades. Many new products and new processes have uieen introduced. New enter- prises have been born at high rates, especially since 1970 with the expansion of commune and brigade enterprises and (more recently) individual enter- prises. Largely as a result, the proportion of small enterprises in China is quite similar to that in a typical developing or developed country, and far higher than in Eastern Europe (Table 1.1). Individual enterprises have expan- ded at widely varying rates; some have merged, diversified and changed their product mix, such as one Shanghai engineering enterprise, whose expansion from a repair shop to a manufacturer - first of gear lathes, then of welding equip- ment, then of electrical cable equipment, then of pumps - resembles the his- tory of many engineering firms in other countries. Some enterprises have also been closed down, especially in recent years, including many small cigarette and machinery manufacturers.  - 14 - TabLe 1.1: SIZE DISTRIBUTION OF INDUSTRIAL ENTERPRISES (Percentage of total number of enterprises with 5 or more empLoyees) Size Class China UK US S. Korea Japan India Yugoslavia Hungary (No. of (1982) (1979) (1977) (1981) (1972) (1976-77) (1981) (1981) Employees) 5- 33 59.2 65.2 56.4 70.6 80.2 51.7 6.6 2.2 33- 75 19.5 15.7 20.3 14.4 10.7 35.3 15.8 4.8 75-189 12.2 10.8 12.4 9.2 6.1 7.8 32.1 18.7 189-243 8.5 1.4 3.8 1.5 0.8 0.8 12.0 9.2 243+ 0.6 6.9 7.1 4.3 2.2 4.4 33.5 65.1 Sources: For China, estimates based an data in the Statistical Yearbook of China, 1983. For other countries, various national sources. 1.36 But there have also been some basic differences between China and most other countries in the area of microeconomic change. Product and process innovations by existing enterprises appear to have been less frequent; the trend of increasing specialization has been much more muted; and producers have rarely sought new customers and markets (for example, smelters in China, unLike the US, normally produce only pure metaLs, with users having to make their own aLloys). Changes of product mix, diversification, merger, and spin- off of new enterprises seem to have been less common. Enterprise cLosure rates have been low. Workers have generalLy remained employed by the same enterprise throughout their career. 1.37 The competitive process by which better and cheaper products drive out worse and more expensive products likewise seems to have been slow in China. Especially in the state sector, backward enterprises have been protec- ted, and progressive enterprises held back - sometimes deliberately in order to maintain profits, capacity utilization and employment in less efficient enterprises. In recent years, for example, despite general agreement that there are too many bicycle producers, the output of enterprises producing good quality bicycles has been restricted by the Ministry of Light Industry in order to protect the producers of inferior bicycles. Similarly, a cotton mill in one province was recently obliged by the provincial authorities to stop sending its cloth elsewhere for dyeing and printing, because the province's own substandard printing and dyeing enterprise was underutilized. Subsidies to keep inefficient pLants and old equipment in operation are common, partly because enterprises play a key role in providing housing and social services.  - 15 - 1.38 Chronic shortages, strong emphasis on production targets, small price differentials between high and low quality products, and inadequate incentives have made product quaLity improvement, innovation, and cost reduc- tion occasional, rather than everyday, concerns of Chinese enterprises. The former policy of local self-sufficiency and deficiencies in the material sup- ply system have led local governments to construct many plants of suboptimal scale and inferior product quality to supply local needs for commodities such as steel. (This may explain why, by comparison with all the other countries in Table 1.1, China has so few enterprises in the largest size category.) 1.39 Deliberate and incidental obstacles to microeconomic change of course exist in other councries. Declining industrial sectors are protected by trade barriers; large firms in difficulty are helped out with loan guaran- tees and subsidies. Monopolistic practices, distorted prices, corrupt or incompetent bureaucracies, and inadequate recraining facilities, coo, retard the speed and efficiency of development elsewhere. Nonetheless, che obstacles to microeconomic change in China seem to have been unusualLy substantial and comprehensive, and could - if allowed to persist - prove incompatible with China's objective of sustained rapid growth. 1.40 tn all countries, the potential for "extensive growth" through duplication of existing production is limited, no matter how much capital is accumulated, by shortages of nacural resources and (ultimately) of labor, and by the unwillingness of consumers to buy ever-increasing amounts of the same goods. To attain a high level of economic development chus requires "inten- sive growth" based on innovation, cost reduction, and product improvement; and this in turn requires much faster microeconomic change. With close to 702 of the labor force still employed in agriculture, extensive growth alone could probably carry China up into the middle-income range over the next two decades or so, aLbeit at rather high cost. But, over the Longer term, intensive growth will be essential to China's catching up with the developed nations; and the experience of other countries such as the Soviet Union has shown the difficulty of switching to intensive growth at a late stage of development. 1.41 Faster microeconomic change in China will require - above all - changes in the system of lonomic management, many of which have already been implemented or proposed- System reform is thus a pervasive theme of this report: the general treatment in Chapter 10 builds on the discussion of spe- cific aspects of system reform in other chapters, especially those on technol- ogy, spatial issues, and human and financial resources. In all chapters, the need for difficult choices will emerge; this is partly because, beyond a cer- tain point, the objective of faster growth, with which this chapter has so far been solely concerned, conflicts with the objective of equity in income dis- tribution. ll/ In addition to the major reforms already carried out in rural areas, and numerous experiments in urban areas, a general strategy for reform of the urban economy is contained in the Communist Party of China's Central Committee Decision on Reform of the Economic Structure, October 20, 1984.  - 16 - D. Inequality 1.42 The distribution of income - and of living standards more generally - is of critical importance for a socialist country such as China, whose cen- traL economic objective is not merely rapid growth, but growth whose benefits are widely spread. Considerations of equity will inevitably play a central role in shaping China's future development strategy and system reform. Inequality in China in International Perspective 1.43 China's past strategy and present system have created an, on the whole, extraordinarily equal society. The first World Bank economic report dealt with this at length (Main Report, paras 3.46-102); its conclusions, which are essentially unaffected by more recent information and developments (to be described subsequently), merit summary recapitulation. 1.44 Inequality of urban incomes is uniquely low in China, with virtuaLLy no extreme poverty in urban areas. There is, however, a large gap - compar- able to that in other developing countries - between average urban and ruraL incomes, which in terms of living standards is magnified, as in some other countries, by concentration of social services on urban residents. In addi- tion, there is substantial inequality of rural incomes in China, largely because of wide differences in the quantity and quality of agricultural Land per person, coupLed with tight restrictions on geographical mobility. 1.45 Thus, because China is a low-income country, and because of substan- tial rural income inequality, a large minority of the population have extremely low incomes. These people, however, have a much higher standard of living than their counterparts in most other developing countries. Agricul- tural collectivization has prevented the emergence of a class of impoverished landless laborers; the state guarantees a minimum food supply; primary school enrollment is high; and basic medical care and family planning services are available to most people. As a result, the hunger, disease, high birth and infant death rates, general illiteracy, and constant fear of destitution and starvation that haunt very poor people in other countries have been more or less banished from China. Life expectancy, whose dependence on many other economic and social variables makes it probably the best single indicator of the extent of wl poverty, is on average in China outstandingly high (67 years in 1980 - ) for a low-income country; even in the poorest province, it is not far below the average for middle-income countries. 1.46 New Information. Since the first economic report was written, sub- stantial additional information has become available and further studies made. The preliminary results (10% sample) of the 1982 census have been published, as have more household survey results. Health, nutrition and 12/ World Bank estimate based on the 1982 census and other recently released demographic data: see Supplementary Paper 1 of China: The Health Sector (World Bank, 1984). The estimate for 1979 in the first economic report (64 years) was too low.  - 17 - population issues were examined by the World Bank in China: The Health Sector (1984). In general, this material has reinforced the conclusions of the first report. But a few changes of emphasis are required. 1.47 The distribution of income in rural areas is less unequal than the World Bank's first estimate (based on fragmentary data) had suggested. Rural household survey data, summarized in Table 1.2, show the Gini coefficient in 1979 to have been 0.26, rather than 0.31. This means that rural income inequality in China is significantly less than in other South Asian countries (with Cini coefficients of 0.30-0.35), rather than in the same range. 1.48 The same data, however, underscore the importance of geographical differences in rural incomes. Table 1.2 allows comparison of the rural income distribution for the whole of China with the distributions in a relatively rich province (Jiangsu) and in a poor province (Gansu); the differences are extremely striking, with four-fifths of the people in Jiangsu in the Y 200-500 plus range, and two-thirds of the penple in Gansu in the range below Y 200. Moreover, because Chinese provinces are large and internally diverse, these provincial statistics do not adequately convey the range of incomes among localities. For example, Gansu's average rural per capita income in 1983 was Y 228 (as compared with the national average of Y 310); but in Gansu's poorest county, Dingxi, the average was only Y 108; and in one of Dingxi's poorest townships, Dongye, it was only Y 55 - and this was the township's best-ever year for agricultural production. Table 1.2: RURAL INCOME DISTRIBUTION (Percentage of population in each income group) Income National average Gansu Jiangsu per capita 1979 1980 1981 1982 1982 1982 Below Y 100 19.6 10.1 4.9 2.9 21.1 ) )16.8 Y 100-200 54.0 53.2 39.5 25.6 45.2 ) Y 200-300 20.7 26.0 36.2 39.3 22.5 36.2 Y 300-400 4.2 7.7 13.5 20.2 8.2 27.6 Y 400-500 1.1 2.2 4.0 7.4 2.3 11.4 Over Y 500 0.3 0.9 1.9 4.6 0.7 8.0 Gini coefficient 0.257 0.237 0.231 0.225 na na Source: Estimates based on household survey data (Annex E, Appendix I). 1.49 Census data on educational attainment by sex and age group (Table 8.1) confirm that China's performance in basic education surpasses that of most low-income countries. Moreover, they clearly reveal the adverse effect of the Cultural Revolution on the proportion of younger people with  - 18 - postsecondary education. The same data aLso show a persistent (though narrow- ing) gap between male and female educational attainment. This, together with the census data on occupation by sex, suggests that the World Bank's first report may have overstated the degree to which China has reduced inequality between men and women, though it is still less than in most developing coun- tries. 1.50 Recent Developments. The data recently released also allow an examination of trends over the past 5 years - a period of important policy and systemic changes, including above all the replacement of collective agricul- ture by household farming. This, in conjunction with major increases in pro- curement prices and rapid development of rural nonagricultural activities, has caused an astounding surge in agricultural production (more than 7% per year since 1978) and rural incomes - up in real per capita terms between 1979 and 1983 by about 70% (Annex B, para. 1.33). Urban wages and employment have also increased and urban staple food prices have been held constant, which has raised real urban per capita incomes by about 40% over the same period; but the urban-rural income gap has still narrowed significantly. Rising agricul- tural production has also caused a sharp increase in food consumpcion, which, at about 2,700 kcal per person per day in 1982 (Annex B, para. 2.03), compares favorably with levels in most middle-income countries. Nonetheless food con- sumption in many localities still falls far short of the average, and the com- position of the Chinese diet is still that of a low-income country. 1.51 It was generally anticipated that the rural reforms, by making the incomes of individual households more dependent on their effort, skill and luck, would increase income inequality in rural areas. The statistics in Table 1.2, although they may have various biases (Annex B, para. 1.34), sug- gest that so far this has not happened, and that if anything rural inequality has declined. This could be due to increases in inequality within localities having initially been more than offset by reductions in income differentials among localities: the reforms spread fastest in poor and backward areas, some of which had also been especially disadvantaged by the former policy of con- centrating on grain production even when other activities would have been more economic. In any event, the lack of deterioration in the distribution of rural incomes has meant that the sharp rise in average rural incomes has greatly reduced the number of households with very low incomes. On the basis of the data in Table 1.2, and a poverty line based on food intake requirements of 2,185 kcal per day, it is estimated that the proportion of the ruraL popu- lation in poverty declined from 31% in 1979 to 13% in 1982 (Annex E, Table I.7). Though the latter figure still means 100 million people, and though people in places like Dongye township have gained little, the speed and scale of the improvement is probably unprecedented in human history. 1.52 In other respects, however, the rural reforms have had some unfor- tunate consequences for inequality. One has been a sharp decline in rural cooperative medical insurance, financed from collective welfare funds, to which individual households are reluctant to contribute. This has been parti- cularly pronounced in poor and backward areas (in richer provinces, the swell- ing profits of commune and brigade enterprises have by contrast led to improvements in collective medical and other social services). The drop in the proportion of production brigades covered by cooperative health insurance  - 19 - from 85% in 1975 to 58% in 1981 (with further declines since then) thus repre- sents a considerable step backwards, although the Government is making efforts to upgrade the quality of primary medical care, for example by giving barefoot doctors more training. 1.53 Similar problems have arisen in basic education, mainly because the household responsibility system has induced more rural parents - as in many other countries - to keep their children out of school to work on the family farm. As shown in Figure 1.4, the proportion of primary school age children enrolled (the net enrollment ratio) dropped in 1980-82, although subsequent government efforts seem to have reversed the decline; the gross enrollment ratio at the secondary level has dropped from about 46% to about 30%, below the developing-country average of 39% (Annex A, paras. 2.04-11). The declines have been particularLy marked in poorer areas, where enrollment ratios were already unsatisfactorily low (and actual attendance rates lower still); in richer areas, such as Jiangsu, rural parents see clearly that their children will not be able to join the rapid exodus into nonagricultural employment without a good education. The declines in enrollment have also been particu- larly marked among girls. Possible Future Trends in Inequality 1.54 China may not in the future be able to achieve such a low degree of inequality as in the past - and indeed, the Government's position is that past strategy was excessively egalitarian. The rural responsibility system, though thus far it has increased inequality only through its impact on education and health care, will probably also start to increase income inequality before too long. Systemic changes in urban areas could also increase inequality, by making diligent workers and successful individual business proprietors richer, and the idle or unlucky poorer. Increased inter-locality specialization and trade will surely benefit some localities more than others. 1.55 In addition, China may be affected by the forces that have tended to increase inequality during the low-to-middle-income transition in other coun- tries. The incomes of the poorest 40% of the population have usually increased more slowly than the average; the gap between agricultural and non- agricultur,l earnings has widened; and inter-regional income disparities have increased.-31 At higher income levels, though, all these trends have usually been reversed, as the modern sectors have come to dominate the economy and people have moved in search of higher incomes. And even in middle-income countries with an unusually high degree of inequality, most people have a much higher standerd of living than in a low-income country. 1.56 China will also have to deal with the losses that microeconomic change may inflict on certain families, occupations, enterprises and locali- ties - even though these may not be visible in aggregate measures of inequal- ity, and even though the gains of other groups may hugely outweigh these 13/ See, for example, World Development Report, 1980, Chapter 4; and Background Paper 3.  - 20 - Figure IA ENROLLMENT RATIOS IN FORMAL EDUCATION: CHINA (1965-83 AND TARGEFS FOR 2000); OTHER DEVELOPING COUNTRIES (1 96-80) 410.0 _ z EDUCA1K)N~~~~~~~~-DLCAK) 20.0 EDCAIO OlHB?U)CsY_ 7D ~ ~~ _ 4 0 ~~~~~~~~~ OHIER LC HIGHE R wS ~~~~~~~~~~EDIJCATION 300 -CHIN- . ~~~~~~~OTHER LDCs' - 4. _ _ UNIVERSI 3.0 _DrATO CHINAtNA o.2~ ~~~~~OTE D, ......=== 196Sa 1 0.7 ~ ~ ~ ~ ~ ~ ~ ~ E i n 1965 1970 1975 1980 1985 1990 1995 2000 YEAR 'We enro1lI0 ratrit ol eYpienes the nuTib8r of students enmolled in a Ie%OI of education as a percentage of the relevant age grtjp b Torgets for non-unwr*iealy hi"heducatiConf in 2000 are not Ovoiobe. Word Bank-27083  - 21 - losses. An example would be a backward enterprise suffering from competition from cheaper or better goods. Governments in all countries, especially socialist ones, feel impelled to reduce the social and political costs involved. Yet protecting potential losers, or routinely compensating them, also has a high cost, because protection and compensation can slow microecono- mic change and reduce growth. Ways of easing this dilemma will be discussed in subsequent chapters; but some compromise between conflicting objectives is unavoidable. 1.57 Despite all this, if appropriate poLicies are foLlowed (experience in other countries suggests the importance of support for agricuLture, labor- intensive industry and services, and small-scale economic activities), and if important features of the past system are preserved (including social owner- ship of land, widely dispersed access to basic education and health care, and a guaranteed minimum food intake), China should be able to maintain a much lower degree of inequality than in the typical developing country. In addi- tion, the new dynamism of the rural sector could further narrow the urban- rural income gap. And in the longer term, regardless of the precise pattern of distribution, the great majority of China's people will unquestionably gain substantially from rapid growth - although special attention will have to be given to people and localities excluded by isolation or inadequate resources from the general process of development.  - 22 - 2. ILLUSTRATIVE PROJECTIONS 2.01 To explore further China's long-term prospects for growth and struc- tural change, discussed in the previous chapter, and to provide a consistent numerical framework for the discussion of other topics in subsequent chapters, projections to the year 2000 have been made with an economic model. This chapter provides a simple description of the model, explains the main assump- tions underlying three alternative projections, and summarizes their results. (A fuller, technical account of the model and projections is in Annex D. The underlying data, incLuding an estimated input-output table, are in Annex E.) A. Description of the Model 2.02 The model is a simplified representation of China's whole economy, broken down into 20 sectors. There are two agricultural sectors (crops - including forestry - and animal husbandry - including fishing), four energy sectors (electricity, coal, oil extraction, and oil refining) and seven other industrial sectors (metallurgy, chemicals, machinery, building materials, food processing, textiles, and other manufacturing), as well as construction, transport, commerce, education and health, public administration and defense, miscellaneous services, and housing. The last four sectors are not regarded as "'materially productive" in China; they are included partly to permit calcu- lation of Western-style national accounts, but mainly to make it easier to handle some important components of investment, public expenditure and house- hold consumption. Chinese-style national accounts are also generated by the model, however, and in most respects t7e sectoral classification corresponds with that used in Chinese statistics.l! 2.03 For each of the 20 sectors, the model has production, use of materials, employment, investment and prices. And for each sector, several sources of demand are distinguished: intermediate purchases by other sectors, household consumption, public consumption, exports (less imports), and use for investment. The model projects all these variables - about 600 in total - into the future, simuLtaneously and in an internally consistent way, with allowance for the many accounting linkages among them. 2.04 To make a projection requires many assumptions, both about the future path of the numerous variables that are in the model but not determined by it (examples of such "exogenous" variables being population and energy con- servation rates in particular sectors), and about the numerous causal - as distinct from accounting - linkages governing the variables that are deter- mined by the model. By comparison with some recent models, the assumed causal linkages in this model are simple and mechanical, and not based on any parti- cular theory of household, farmer or enterprise behavior. This mat 3 them less precise and complete, but also easier to grasp and less controversial. 1/ Except that brigade enterprises are included in the relevant industrial sectors rather than in agriculture, and that transport includes both passenger and freight.  - 23 - Growth 2.05 Growth of national incoii.e in the model depends mainly on the level and efficiency of investment. The level of investment is determined by the proportion of national income saved and (to a much lesser extent) the amount of foreign borrowing, both of which are treated as government policy deci- sions. Investment is added to the existing capital stock, which increases production capacity in accordance with assumed linkages in each sector between capital stock and gross output (sectoral capital-output ratios). Because these ratios are not the same in every sector (a unit of capital adds more, for example, to the gross output of textiles than of metallurgy), the impact of investment on growth depends partly on its alLocation among sectors. it also depends on the ratio of net output (i.e. contribution to national income) to gross output in each sector, which in turn depends on efficiency in the use of materials, energy and other intermediate inputs. 2.06 Increased production in the model requires not onLy more capital and more materials, but also more labor. Sectoral labor requirements, and hence employment leveLs, are governed by assumed growth rates of labor productiv- ity. But the overall growth rate of the economy is not constrained by avaiLa- bility of labor. This is partly because skilled labor is not distinguished from unskilled labor, partly because there is assumed to be (over the next two or three decades) a pool of surpLus labor in agriculture, which can be drawn into empLoyment in other sectors with no reduction in agricultural output. 2.07 The determinants of growth in the model are a great simplification of reality. In particular, the sectoral capital-output and net-gross output ratios summarize and conceal a host of important underlying factors - product and process innovations, the quality of investment decisions, the training and effective use of skilled labor, and pressures on enterprises to improve pro- duct quality and to use less materials, labor, and fixed and circulating capi- tal. Growth in the model is also smooth. In particular, it is never held back by bottlenecks or shortages in individual sectors (such as metallurgy or transport), partly because imports and increased use of existing capacity are assumed to provide short-term flexibility, partly because capacity shortfalls in specific sectors are assumed to be remedied through increased investment within a year or two. These assumptions, though technically convenient, and reasonable for analyzing growth over a twenty-year period, make the model unsuitable for analyzing medium-term (say, five to seven year) growth pros- pects. Structure 2.08 Except for agriculture and energy, where natural resource cons- traints are important, the structure of production - the relative sizes of the different sectors, and the way in which these change as the economy grows - is determined in the model by the structure of demand. This is because the allo- cation of investment among sectors in the model is governed by the expected growth of demand for their output: in any given year, demand and production  - 24 - capacity in particular sectors may not balance;2- but sectors where there are shortages, and where the trend growth of demand is more rapid, receive larger shares of total investment, and hence experience faster growth of production. 2.09 rhe model's basic assumption that the Long-term structure of produc- tion is df;termined mainLy by the structure of demand means only that the Gov- ernment's production policies will be based mainly on assessment of needs. These production policies may be implemented directly, by quotas and adminis- trative decisions, or indirectly, by using economic levers to regulate the pattern of demand, and allowing farmers and enterprises to respond. The model is compatible with bo.h methods of implementation. 2.10 As regards final demand, the division of national income between saving (the main determinant of demand for investment goods), public consump- tion, and household consumption is treated in the model as a direct government policy decision, as is the composition of public consumption (mainly education and health, and public administration and defense). Household consumption is spread among sectors by assumed "income elasticities" that cause the share of food to decline and the shares of other goods and services - including housing - to increase at varying rates as income rises. The Government can of course influence these eLasticities by taxes and subsidies, or by rationing or other administrative means, but these policy instruments are not incLuded in the model. 2.11 Exports (less imports) are also a significant component of final demand in several sectors. They are determined over the Longer term in the model mainly by assumptions regarding the openness of each sector (the ratio of exports or imports to sectoral production), the degree of self-sufficiency in each sector (the sectoral trade deficit), and the composition of manufac- tured exports - all of which are treated as government policy decisions, although the instruments nccessary to implement these decisions are not expli- citly included. Sectoral exports and imports are also influenced in the model by the need to equate the overall outflow and inflow of foreign exchange, given the Government's chosen amount of foreign borrowing. 2.12 About half of totaL demand is not final, but intermediate. For any given sector, growth of intermediate demand depends partly on growth of pro- duction in the other sectors that use its output as an input (e.g., interme- diate demand in metallurgy is governed largely by machinery production), partly on changes in the pattern of intermediate use within those sectors (e.g., reductions in the amount of metal per machine). The composition of intermediate demand in the model thus depends on all the factors that govern the structure of final demand, plus the assumed pattern of constancy or change 2/ In sectors where there is foreign trade, temporary demand-production imbalances are accommodated in the model by adjustment of exports or imports, in other sectors by temporary changes in capacity utilization (reflected in temporary deviations of sectoral capital-output ratios from their assumed Long-term values). Demand-production imbalances do not affect prices. These assumptions are made more for simplicity than for accuracy, since the model is not intended for short-term analysis.  - 25 - in input-output coefficients, which may be influenced in various ways (not explicitly modelled) by government policies. B. Three Alternative Projections 2.13 Since a Large number of assumptions, both about the future values of exogenous variables and about specific causal linkages among variables, are required to make a projection, there is a colossal number of possible alterna- tive sets of assumptions (and hence different projections), even if each indi- vidual assumption is varied only within a limited, plausible range. After many experiments, three have been selected for discussion (some variants will also be mentioned). 2.14 These three projections - named QUADRUPLE, MODERATE and BALANCE - have no monopoLy of merit, but are useful for illustrative purposes. Although all of them lie within the range of international (including Chinese) exper- ience, they span a fairly wide range of possibilities, both for policy and for factors beyond government control. They also reveal a number of specific questions and problems, which are addressed later in this report. The present section outlines the assumptions of each of the three projections. The rest of the chapter describes their results. QUADRUPLE 2.15 The QUADRUPLE projection attains the Government's target of quadrupling the gross value of industrial and agricultural output (GVIAO) between 1980 and 2000. Specifically, given a reasonable set of assumptions - described below - about the efficiency of investment and the forces shaping the structure of the economy, the aggregate saving rate was varied until the rate that quadrupled GVIAO was found. 2.16 As regards the efficiency of investment, sectoral capital-output ratios in agriculture and energy were direct (though approximate) estimates of the likely investment cost of increasing gross output in t:lese sectors, draw- ing on WorLd Bank project experience in China and elsewhere. In other indus- trial sectors, and in construction, it was assumed that capital-output ratios would be equal to the average of the past 30 years in China (based on esti- mates of sectoral capital stocks in 1981), with reforms checking their past upward tendency. In transport, the capital-output ratio for new investment was set about 25Z above its past level, to correct for past underinvestment. In commerce, it was assumed that greater discretion in purchasing would cut circulating capital requirements by 202, and hence that the sector's future capital-output ratio would be lower than in the past. 2.17 Energy conservation rates in individual sectors in QUADRUPLE (and also in MODERATE and BALANCE) are those regarded as most likely - the conse- quences of more optimistic and more pessimistic assumptions are discussed in Chapter 4. Intermediate use of other materials and goods was assumed to change in accordance with average international experience. Specifically, the assumed pattern of change in input-output coefficients includes some substitu- tion of manufactured for agricultural materials, increased use of chemicals (especially plastics) and reduced use of metals (especially in machine- building). Total nonenergy intermediate requirements per unit of gross output in individual sectors (other than agriculture) were assumed to rpmain approxi- mately constant, as in most other countries.  - 26 - 2.18 The division of consumption between public and household consumption was assumed in QUADRUPLE to stay the same as in 1981, as was the proportionate allocation of public consumption among sectors. The share of household con- sumption allotted to food declines, despite a steep increase in expenditure on animal produc.- (c.e Chapter 3 and Annex B). The share of clothing in house- hold consumption was assumed Lo rise slightly, and the shares of other manu- factured goods, electricity, transport, miscellaneous services and housing to increase significantly - broadly in line with average internationaL exper- ience. Household consumption of fuel, though, was assumed to increase sLower than income, because of improvements in thermal efficiency (see Annex C). 2.19 Other assumptions of QUADRUPLE include growth of crop production at an average annual rate of 3.6Z, doubling of crude cil production (in accord- ance with the official target), attainment of the Government's targets for hydro and nucLear power, and coal production in 2000 of 1,400 million tons. The present orientation of foreign trade policy is assumed to be broadlvr main- tained, but the composition of manufactured exports shifts away from textiles toward machinery and other manufactures, and there is somewhat greater depen- dence (to save energy) on imports of metals and chemicals. Foreign borrowing increases gradually toward a Level consistent with a 15% ratio of debt service (interest and repayment) to exports. I.-ODERATE 2.20 The MODERATE projection maintains most of the assumptions made in QUADRUPLE, including the same aggregate saving rate, but takes a less optimis- tic view of the future efficiency of China's economy. Sectoral capital-output ratios in (nonenergy) industry and construction are assumed to increase gradu- ally, as in the Soviet Union and less efficient developing countries. Speci- fically, the amount of investment required per additional unit of industrial output is assumed to be 40Z greater than in QUADRUPLE. In addition, crop pro- duction is assumed to grow more slowly (an average annual rate of 2.9%), because of smaller increases in the efficiency with which investment and industrial inputs are used in agriculture. Lower efficiency in converting animal feed into meat is also assumed. Coal production in 2000 is 1,200 mil- lion tons. 2.21 Lower efficiency was also assumed to be associated with slower growth of labor productivity in nonagricultural sectors. In QUADRUPLE, on the basis of past experience in China and other countries, gross output per worker was assumed to increase at an annual rate of 5% in heavy industry (3% in coal mining), 4% in light industry, 3% in construction, transport and commerce, and 2% in miscellaneous services. In MODERATE all these rates were reduced by one percentage point. BALANCE 2.22 The BALANCE projection represents an alternative way of attaining the same growth rate of per capita national income as in QUADRUPLE, giving greater weight to the service sectors - specifically, to commerce and miscel- laneous business and personal services. This shifts the future structure of China's economy away from the Soviet pattern and toward the pattern of Japan and most other countries at comparable stages of development (para. 1.30).  - 27 - 2.23 In the model, the shift is accomplished partly by changing the pat- tern of household consumption. ReLative to household income, consumption of miscellaneous services and of commerce (retail trade is treated as part of consumption in the accounting framework used) grows faster in BALANCE than in QUADRUPLE. To compensate, the share of household income spent on food declines somewhat faster - aLthough this is partly offset by increased food consumption in restaurants, which are also part of commerce - and the share spent on manufactured goods increases somewhat more slowly. There is also faster growth of intermediate demand for services. Relative to gross output, the volume of intermediate commerce and material supply activities is assumed in BALANCE to increase by 3Z per year (it remains constant in QUADRUPLE); and expenditure on miscellaneous business services is assumed to increase gradu- ally in every sector from virtually nothing in 1981 to about 3% of gross out- put in 2000 - comparable to other countries. 2.24 Faster expansion of the service sectors involves costs - investment and use of energy, materials and labor in those sectors. In BALANCE, these costs are deliberately augmented by making the fixed capital required per unit of gross output in commerce three times its QUADRUPLE level, to reflect improvements in the size and quality of warehouses, shops, vehicLes and other equipment, which at present in China Lag Ear behind those in other coun- tries. In addition, to allow for improved staffing of large-scale commercial facilities, a larger share of more labor-intensive smaller facilities, and a shift in the composition of miscellaneous services toward more labor-intensive activities, the growth rates of labor productivity in commerce and miscella- neous services are reduced in BALANCE (to 1Z per year in both sectors). 2.25 Faster expansion of commerce and miscellaneous services would also have economic benefits - in addition to the gains that consumers would realize directly - especially in association with comprehensive reform of China's sys- tem of economic management (para. 1.31). For these sectors are indispensable to reducing costs and improving quality in material production through increased reliance on markets, more specialization and competition, and greater orientation of production toward customer requirements. 2.26 Capital. In this broader context, faster expansion of the service sectors could reduce investment requirements in several ways. A larger, better-equipped and more responsive commercial system could require - in light of experience elsewhere - smaller amounts of circulating capital, especially within the sector, but also in other sectors, which wouLd be less likely to produce unsaleable goods and would have less need to stockpile materials and semifinished goods (of which it is now difficult for Chinese enterprises to get regular supplies of suitable quality and specifications). A larger and better commercial system would also permit greater specialization in produc- tion, and hence larger-scale enterprises with lower capital costs, and a greater payoff to agricultural investment. 2.27 Business service enterprises can also make fuller use of specialized equipment (and personnel). This is obviously true of equipment leasing and rental enterprises. But it also applies to professional and technical servi- ces - advertising and market research, law, accounting, design, engineering, repair and maintenance, data prccessing - which can often be provided at lower cost and to a higher standard by specialized entities than by an enterprise's own staff. Even mundane business services such as catering, cleaning and  - 28 - trash removal can sometimes be undertaken at lower cost by specialized enter- prises. Finally, banks and other financial institutions can contribute to better investment decisions (Chapter 10). 2.28 The potential for reducing investment requirements in these ways cannot be accurately calculated or apportioned between fixed and circulating capital. However, on the basis of rough internationaL comparisons, it was assumed for illustrative purposes in BALANCE that faster expansion of the service sectors would enable circulating capital requirements to be reduced (below their QUADRUPLE LeveLs) by 30Z in nonenergy industry, construction and miscellaneous services, and by about 70Z in commerce. 2.29 Materials. It was also assumed in BALANCE that the increased inter- mediate purchases of commerce and business services would be offset by reduced purchases of manufactured intermediate goods (and hence indirectly of agricul- tural materials and energy, as well as other industrial materials). This is the pattern observed in the few studies that have been made in other countries - increasing service-intensity of production in individual sectors, but declining materials-intensity, with little systematic change (outside agricul- ture) in total intermediate purchases per unit of gross output. Part of the explanation apparently lies in the transfer of particular activities from manufacturing enterprises to more specialized service enterprises. But a more important factor appears to be constant improvements in product quality, with the cost of materials becoming a smaller part of the product price, and the value added by processing (including indirect processing, design, packaging and marketing) a greater part. 2.30 Of course, many materials-saving improvements in product quality are unrelated to service sector expansion. Independent technological advances, including methods of weight reduction and miniaturization, as well as general enterprise incentives to reduce waste and improve products, are crucial. But the service sectors stimulate and facilitate these improvements. Design agen- cies and technical consuLtants provide essential information, while a good commercial sector, supplemented by advertising and market research enter- prises, increases the flow of information between customers and producers and sharpens competition, making it both easier and more necessary for enterprises to improve their products. C. Growth and Demand 2.31 For each of the three projections, using both Western and Chinese measures, Table 2.1 presents aggregate growth rates over the period 1981-2000, while Table 2.2 shows the composition of both total final demand and household consumption. (More detailed breakdowns are available in Annex D. The projec- tions start in 1981 because this is the latest year for which all the neces- sary data exist; the projected values for 1982-84 do nct conform closely with actual developments in those years.)  - 29 - Table 2.1: AGGREGATE GROWTH RATES 1981-2000 (Annual average percentages, at 1981 prices) QUADRUPLE MODERATE BALANCE National Income GDP (Western measure) /a 6.6 5.4 6.6 NMP (Chinese measure) 7a 6.3 5.1 6.2 National Income Per Capita GDP 5.5 4.3 5.5 NMP 5.2 4.1 5.1 Gross Value of Industrial and Agricultural Output 7.2 6.0 6.4 /a GDP (Gross Domestic Product) is the net output of all sectors, including all services, plus depreciation. NMP (Net Material Product) is the net output of the materially productive sectors. Table 2.2: COMPOSITION OF FINAL DEMAND (Percentages) 1981 2000 QUADRUPLE MODERATE BALANCE Shares of National Income (Expenditure) /a Investment 28(29) 29(29) 29(28) 26(26) Public consumption 15(10) 15( 9) 15(10) 16(10) Household consumption 56(61) 56(62) 56(62) 59(64) Shares of Household Consumption /b Food (incl. processed food) 55 48 49 44 Manufactures 24 29 28 25 Services (incl. commerce) 18 20 20 27 Fuel, electricity, transport 3 4 4 4 /a Figures outside parentheses are Western measures, inside parentheses are Chinese measures. The Western measures may not sum to 100 because of external trade imbalances. /b On Western basis, at 1981 producer prices, with all commercial margins included in services.  - 30 - 2.32 In QUADRUPLE, as mentioned earlier, the saving rate was deliberately chosen so as to attain the target of quadrupling GVIAO in 1980-2000 (annuaL average growth of 7.2Z). The required domestic saving rate is 29%, although the investment rate is closer to 30% for most of the period because of foreign borrowing. This is higher than China's average investment rate in 1952-82 (28%), but lower than the 1970-82 average (32Z). It is also quite close to Chinese projections of the investment rate needed to quadruple GVIAO in 1980-2000 (26-29%), though somewhat higher, perhaps because QUADRUPLE assumes more of an increase - over China's yast leveLs - in capital requirements in agricuLture, energy and transport.- 2.33 National income - as a result of the many other specific assumptions made, rather than by deliberate choice - also grows in QUADRUPLE at a rate consistent with the Government's targets. Per capita GDP (which in China dif- fers only trivially from GNP) grows at 5.5% per year - from $300 in 1981 to $830 in 2000 - which is the minimum rate that China would need to sustain to catch up with the industrialized countries by the middle of the twenty-first century (para. 1.03). The Chinese measure of national income per capita grows slightly slower - 5.2% - because the relative size of the nonmaterial service sectors increases. 2.34 In MODERATE, with the same investment rate, the assumption of lower efficiency causes significantly slower growth. The growth rates of national income and of GVIAO are reduced by rather more than one percentage point, and per capita GDP rises to onLy $670 by 2000. To attain the same aggregate growth rates as in QUADRUPLE, the investment rate would have to be increased to about 36%. (All three projections, incidentally, assume that the Govern- ment 's target of a population of 1.2 billion by 2000 is attained. A higher population would cause a commensurateLy lower Level of per capita income.) 2.35 In BALANCE, with greater weight given to the service sectors, the saving rate was deliberately chosen to attain the same growth rate of per capita CDP as in QUADRUPLE, in line with the Government's long-term object- ives. The Chinese measure of national income in BALANCE grows only slightly slower than in QUADRUPLE, because most of the increase is in materially pro- ductive services (commerce and business services). CVIAO, however, grows sig- nificantly slower, because the greater share of services means smaller shares for agriculture and - especially - industry. Partly as a result, but also because of the assumed reductions in use of circulating capital and materials within individual sectors, the same growth rate of national income as in QUADRUPLE is attained in BALANCE with less investment - 26% of national income. The increase in consumption that this makes possible is divided pro- portionately between public consumption and household consumption (which in real per capita terms in 2000 is 9% higher than in QUADRUPLE). 3/ The Chinese projections are from Liu Guoguang (ed.), Zhongguo Jingji Fazhan Zhanlue Wenti Yanjiu, (Issues of China's Economic Development Strategy), Shanghai, 1983, pp. 405-406.  - 31 - 2.36 By most international standards, the growth rates projected for China are quite high. Even in MODERATE, per capita GDP growth is slightly above the 4.1Z 1960-82 average for upper-middle-income developing countries (3.2% for lower-middle-income countries), and similar to Western estimates of Soviet growth in 1950-75. It is well above the 2.6-3.24 range projected by the World Bank for middle-income countries in 1985-95._ On the other hand, the projected MODERATE growth rate of national income per capita is almost exactly what China achieved in 1952-82 (4.0% by the Chinese measure), and the rates in QUADRUPLE and BALANCE are no higher than those attained by China in periods of good economic management. These rates in turn have been surpassed by a few other countries, including Japan in 1953-73 (8.5%). 2.37 China's unusually rapid projected per capita income growth, as com- pared with most other developing countries, is due mainly to an unusually high investment rate (the 1982 average for middle-income countries was 24%) and unusually slow population growth (1% p.a., as compared with a projected 2.2% in middle-income countries), rather than to unusually efficient investment. This can be seen by looking at the aggregate ICOR (incremental capital-output ratio), which is the average amount of investment required to produce an addi- tional unit of output. It can be calculated by dividing the average share of investment in national income in a given period by the growth rate of national income. 2.38 The projected ICOR in MODERATE (calculated using Western national income measures) is about 5.5, much as in China in the two decades before 1978, and comparable also to the Soviet Union in 1950-75 and India in 1960- 82. In QUADRUPLE, the projected ICOR is substantially lower - about 4.5, which is similar to the 1952-82 average in China. In BALANCE, it is lower still - about 4.0, which is comparable to the average in Japan during 1950-80, but also to the average for all middle-income developing countries in 1960- 82. The assumptions on investment efficiency underlying the thrg7 projections thus seem to span a plausible range of international experience.- 2.39 It should be emphasized, though, that China's experience in the next two or three decades - in terms of growth rate and investment requirements - wiLl not necessarily lie within this range. The possibility of faster growth cannot be ruled out. To illustrate this, a variant of BALANCE was con- structed, with the saving rate increased to the same level as in QUADRUPLE. Per capita GDP in this projection increases at an annual average rate of 6.5% (one percentage point higher than in QUADRUPLE and BALANCE, though still 4/ World Development Report, 1984, Table 3.2. Most other developing country statistics in this chapter are from the Indicators at the back of this World Development Report. 5/ Total factor productivity (TFP) for the whole economy grows at 1.6% per year in QUADRUPLE (contributing 24Z of output growth), 0.8% in MODERATE (14%), and 1.9% in BALANCE (29%). The average TFP growth rate in a sample of mainly middle-income developing countries between 1950 and 1980 was 2%, and contributed 31% of output growth: Annex E, Table 3.16. In these TFP calculations, output is measured by CDP; for China, labor force and capital stock growth are given equal weight.  - 32 - slightly below the 6.82 of 1979-84), and reaches $990 by the year 2000. GVIAO in this variant of BALANCE more than quadruples (growth of 7.4Z p.a.), with a commensurate increase also in energy and materials use. 2.40 Equally, if internal or external misfortunes were adversely to affect China's saving rate or investment efficiency, growth could be slower than in MODERATE. In addition, it is possible that all three projections may have underestimated the backlog of past investment needs - especially in transport and in housing (the estimates for electricity are more reliable) - in which case their saving and investment rates would be too low, or their growth rates too high, or some combination of the two. In any event, exper- ience in China and elsewhere (including Eastern Europe) strongly suggests the need for cautious planning: the fluctuations and inefficiency caused by set- ting unreaListically high growth targets are usually much more serious than the problems caused by unrealisticaLly low growth targets, which can be gradually adjusted upward. 2.41 The projected changes in the composition of household consumption (Table 2.2) are broadly consistent with experience elsewhere. In all three projections the share of food declines: in QUADRUPLE and MODERATE the reduc- tion is unusually small, mainly because of rapidly increasing consumption of animal products; BALANCE is more normal by international standards (Chap- ter 3). In QUADRUPLE and MODERATE, most additional nonfood consumption is manufactures, although expenditure on housing and miscellaneous personal ser- vices also grows quite rapidly. In BALANCE, consumption of services (includ- ing retail commerce) grows abnormalLy fast, to bring China more into line with the usual pattern. The share of household income spent on manufactures thus increases slowly; but because - as mentioned earlier - aggregate consumption increases faster, real per capita consumption of manufactures in BALANCE in 2000 is only 5% less than in QUADRUPLE. D. Production and Investment 2.42 Table 2.3 shows gross output growth in the three projections, broken down among five broad sectors (the 20-sector breakdown is in Annex D). In all three cases, agriculture - discussed further in Chapter 3 - is the slowest growing sector, although its growth rate in QUADRUPLE and BALANCE is very high by international standards, partly because animal husbandry production is assumed to keep up with rapidly increasing consumer demand. Slower agricul- tural growth in MODERATE thus reflects both lower efficiency in crop produc- tion and slower overall growth, which reduces the demand for meat. (In these projections, the output of nonagricultural brigade enterprises is included in other sectors, rather than - as in present Chinese statistics - in agricuL- ture.)  - 33 - Table 2.3: SECTORAL GROSS OUTPUT GROWTH 1981-2000 (Annual average percentages, at 1981 prices) QUADRUPLE MODERATE BALANCE Agriculture 4.5 3.7 4.6 Heavy industry a/ 8.1 6.9 7.0 Light industry b/ 7.9 6.5 7.1 Infrastructure c/ 7.3 6.1 7.0 Services d/ 7.2 6.0 10.5 a! These broad sectoral definitions correspond only approximately to those used in China. Heavy industry includes metallurgy, coal and petroleum (including extraction), chemicals, building materials and machinery, but not electricity. b/ Food processing, textiles and clothing, wood, paper, other manufacturing. c/ Electricity, construction, transport (freight and passenger). a/ Comnerce, miscellaneous services, and housing; excludes education and health, and public administration and defense. 2.43 The projected growth rate of total industrial output in QUADRUPLE (about 8%) is in line with most Chinese yrojections, as is the slightly faster growth of heavy than of light industry.- Within heavy industry, the fastest growing sector (9.1Z) is machinery, pulled along mainly by investment demand and by household and public consumption, but also by foreign demand (the share of machinery output exported rises over the period from 4% to 11%). Chemicals production - pulled mainly by intermediate demand - grows almost as fast (8.8%), but metallurgy production grows slower (7.0%), partly because of increased efficiency of metal use in the machinery sector, partly because of increased net imports. Coal and petroleum are the slowest growing heavy industriaL sectors (on average, 4.8%). Within light industry, the fastest growing sector is food processing (8.7%), followed by wood, paper and miscel- laneous manufacturing (8.4%). Textiles and clothing, the biggest light indus- trial sector, grows more slowly (7.1%), partly because consumer demand grows only slightly fast.ir than income, partly because textile exports grow rela- tively slowly (but still by 6.0% per year in real terms). 2.44 In MODERATE, both heavy and light industry grow more slowly than in QUADRUPLE (because the whole economy grows less fast), but the difference between their growth rates is somewhat more pronounced, largely because, by assumption, coal production is not much lower and petroleum production remains the same. Gross output in other industrial sectors in 2000 in MODERATE is about 80% of its level in QUADRUPLE - the range is from 84% in chemicals to 77% in food processing. 6/ Most Chinese projections envisage this relationship for the whole period 1980-2000, but with faster growth of light than of heavy industry in the 1980s (see, for example, Liu Guoguang, op. cit., pp. 146-149).  - 34 - 2.45 In BALANCE, the assumed greater efficiency of investment and materials use - and the consequently lower investment rate - enables heavy industry to grow slightly slower than light industry, and significantly slower than in QUADRUPLE (even though national income grows just as fast). Machinery production in 2000 in BALANCE is only 80% of the QUADRUPLE level, metallurgy 74Z, building materials 85%, and chemicals 82%. Light industry in BALANCE also grows more slowly than in QUADRUPLE, partly because some consumer demand is shifted from manufactures to services, but mainly because of smaller stock building, more efficient intermediate use, and smaller net exports (because of less need to generate foreign exchange to pay for imports of energy, chemicals and metals). 2.46 The infrastructure sectors (electricity, transport, and construc- tion) in all three projections grow nearly as fast as industry, the main user of their outputs and supplier of their inputs. The link between infrastruc- ture and industry is particularly clear in QUADRUPLE and MODERATE - with slower industriaL growth in the latter projection causing an equiproportionate reduction in infrastructure growth. In BALANCE, relative to QUADRUPLE, the reduction in infrastructure growth is much less pronounced than the reduction in industrial growth, because of increased use of infrastructure by the ser- vice sectors. Electricity projections are discussed further in Chapter 4 (and Annex C), transport projections in Chapter 5 (and Annex F); in both sectors, the plausible range of outcomes in the year 2000 in each of the three projec- tions is quite wide (depending largely on the degree of success in reducing wasteful use of electricity and transport). The construction industry's out- put growth is determined mainly by investment demand: it therefore grows more slowly in MODERATE (6.1Z) than in QUADRUPLE (7.2Z), because of slower overall growth; it aLso grows somewhat more slowly in BALANCE (7.0%), because the investment rate is lower. 2.47 Excluding education and health, and public administration and defense, where gross output is a particularly elusive concept (employment in these sectors is discussed below), the service sectors in QUADRUPLE and MODERATE grow at about the same rate as the infrastructure sectors, and some- what more slowly than industry. Commerce grows at almost the same speed as GVIAO, while miscellaneous services and housing (both of which are propelLed by above-average growth of consumer demand) grow rather faster. In BALANCE, commerce grows faster (9.1%) than GVIAO (6.4%), with an increased amount of trade reLative to production, while consumer and business demand for miscella- neous services grows rapidly (15.3%). The service sectors thus grow faster than industry. 2.48 Table 2.4 shows broad sectoral shares of the economy's total capital stock (fixed and circulating) in the three projections. These depend, of course, on the sectoral allocation of investment, which in turn is related both to sectoral gross output growth and to sectoral capital-output ratios (investment requirements per unit of output). In all three projections, agri- culture's share of the capital stock does not change much over time, since the sector's below-average output growth is largely offset - as in other countries - by a rising capital-outpu: ratio. Its share in BALANCE is higher than in QUADRUPLE simply because the total capital stock is 10% smaLLer, due to the lower aggregate investment rate permitted by greater efficiency - reflected in the table in the lower ratio of circulating to fixed capitaL.  - 35 - Table 2.4: COMPOSITION OF CAPITAL STOCK (Percentage of total) 1981 2000 QUADRUPLE MODERATE BALANCE Sectoral Shares Agriculture 9.6 8.8 8.4 9.8 Heavy Industry a/ 29.5 33.4 37.3 28.4 Light Industry a/ 7.9 7.7. 8.5 6.4 Infrastructure a/ 14.1 18.1 16.4 18.4 Services bl 38.9 32.0 29.4 37.0 All sectors Fixed (net of depreciation) 65.3 67.5 66.9 74.8 Circulating 34.7 32.5 33.1 25.2 a/ For secroral definitions, see notes to Table 2.3. b/ Includes education and health, and public administration and defense, as well as commerce, housing and miscellaneous services. 2.49 In QUADRUPLE, industry's share of the capital stock increases over the period - implying that its share o total investment (about 42X) is greater than in the past (about 37%).7' This is the result of rapid output growth in sectors such as machinery and chemicals (which between them absorb around 20% of annual investment), coupled with rising capital-output ratios in slower-growing coal and petroleum. Light industry's small share of the capi- tal stock does not change much. In MODERATE, lower industrial investment efficiency causes the capital stock shares of both heavy and light industry to be even higher than in QUADRUPLE. In BALANCE, these shares decline slightly, because of gower industrial output growth and greater industrial investment efficiency.-' Industry's share of total investment (361) in BALANCE is still large, however - excluding mining and extraction, it is about the same as in Japan in 1951-65. 7/ These percentages include all mining (including oil extraction), but exclude electricity (included with infrastructure). The past estimate is based on the 1981 capital stock data in Annex E, which, as with the projected estimate, covers all investment in fixed and circulating capital, including nonstate and nonproductive investment. 8/ The share of light industry declines more than that of heavy industry partly because coal and petroleum capital requirements are not reduced, partly because greater investment efficiency is assumed to be reflected in lower circulating capital requirements (which are more important in light than in heavy industry).  - 36 - 2.50 In all three projections, the infrastructure sectors increase their shares of the economy's capital stock. In QUADRUPLE, transport absorbs about 10% of annual investment (excluding purchases of vehicles by farmers, house- holds and nontransport enterprises). Transport and electricity together account for a higher proportion (19%) of total investment than in the past (12%), but about the same proportion as in Japan in 1951-65. In MODERATE, these sectors' shares of the capital stock increase less, mainly because of greater absorption of investment by industry. In BALANCE, their capital stock shares are much the same as in QUADRUPLE, although in absolute terms somewhat Lower (because demand for the output of these sectors grows more slowly). 2.51 At the outset, the service sectors have a large share (39%) of the economy's capital stock - of which about one haLf is commerce (mainly circu- Lating capital), one third housing, and one sixth education, health, public administration and defense. In QUADRUPLE, this share declines, partly because circulating capital requirements in commerce are assumed to be somewhat lower than in the past, partly because capital-output ratios in other sectors are rising (they rise faster in MODERATE, which causes the service sectors' capi- tal stock share to decline even more). Nonetheless, these sectors absorb sub- stantial amounts of investment. In QUADRUPLE, about 12% of annual investment goes into housing and 11% into commerce (still mainly circulating capital), while education, health, public administration and defense between them absorb about 5% (the underlying assumption is a doubling of capital per worker over the period in these social sectors). Miscellaneous services absorb only 0.5% of investment. 2.52 In BALANCE, the capital stock share of the service sectors also declines slightly, but is significantly higher than in QUADRUPLE, partly because the total capital stock is smaller. In commerce, faster output growLh and higher fixed capital requirements are offset by lower circulating capital requirements, and the sector's capital stock in 2000 is 5% smaller than in QUADRUPLE (its share of annual investment is virtually the same). The social sectors and housing absorb slightly more investment because of faster growth of public and household consumption. MisceLlaneous services absorb three times as much investment as in QUADRUPLE, but still less than 2% of the total. E. Employmnent and Income 2.53 Table 2.5 shows projected sectoral employment shares (the 20-sector breakdown is in Annex D). These depend in the model on sectoral output growth rates and assumptions about sectoral labor productivity growth (para. 2.21) - except in agriculture, where employment is a residual, and in education, health, public administration and defense, where it is determined by public consumption.  - 37 - TabLe 2.5: SECTORAL EMPLOYMENT SHARES (Percentages of labor force) 1981 2000 QUADRUPLE MODERATE BALANCE Sectoral Shares a/ Agriculture 70 59 61 52 Heavy industry 9 11 11 9 Light industry 6 9- 8 7 Infrastructure 5 7 7 7 Services 10 14 13 25 Labor force (millions) 452 631 631 631 Population (millions) 990 1,196 1,196 1,196 a! Sectoral definitions as in Table 2.4. 2.54 In QUADRUPLE, nonagricultural employment increases at 3.5% per year. This is somewhat sLower than over the past 30 years in China - in 1952- 82, the rate was 4.52 - because of slower gross nonagricultural outpuL -rowth (7.8% in QUADRUPLE, excluding nonmaterial services, versus 8.9% in 1952-82). Average nonagricultural labor productivity in QUADRUPLE grows at the same rate as in 1952-82 (about 4Z), though significantly faster than in 1957-77 (about 3%). Labor productivity is assumed to grow fastest in heavy industry, with the result that this sector's employment share increases only modestly, aes- pite rapid output growth. Conversely, the share of employment in services increases more, despite slower output growth, because labor productivity does not rise so much. 2.55 Although faster growth of productivity in industry than in services is the normal international pattern, the share of employment in services in QUADRUPLE not only remains unusually low, but also increases slower than nor- mal (even by comparison with the low-service-share countries such as the Soviet Union). This is because of the prospective decline in the primary school age group, which will cause the number of teachers - now approaching a quarter of all service sector employment in China - to increase only modestly, even given higher enrollment rates (Annex A). Faster growth of employment in health, research, culture, and so on will to some extent offset this. But the projected increase of 15 million, or 80%, in education and health employment during 1981-2000 in QUADRUPLE - MODERATE and BALANCE are similar - may in fact be an overestimate, as may the projected increase of 5 million, or 502, in employment in public administration and defense (also currently approaching a quarter of total service sector employment). 2.56 Nonagricultural employment in QUADRUPLE increases between 1981 and 2000 by rather more than 120 million. The total labor force - assuming a slight decline in the participation rate of the adult population - increases by nearly 180 million. Employment in agriculture thus increases by 56 mil- lion, even though agriculture's share of employment drops from 70% to 59X.  - 38 - This pattern is typical of the low-to-middle-income transition elsewhere, with an absolute decline in the agricultural Labor force occurring only at a later stage of deveLopment. But the projected decline in the agricultural employ- ment share in QUADRUPLE is rather small by international standards (see Figure 2.1). 2.57 It is also much siialler than in most Chinese projections, which envis7ge a decline in the agricultural employment share by 2000 to under 40x.9 This would be norma' for a country with a per capita income roughLy double that projected for China in 2000 (Annex E, Table 3.8). Given the rates of gross output growth in QUADRUPLE, it could be achieved in China only if nonagricultural labor productivity increased at under 2% per year - half the past rate, and probably incompatible with efficient industrialization. In 1960-82, industrial labor productivity increased on average in all middle- income countries at rather more than 2% per year; but in the faster-growing upper-middle-income group at 3%, in Korea at 6%, and in Japan in the 1950s and 1960s at over 8%. (Productivity and surplus rural labor are discussed further in Chapter 8.) 2.58 Sectoral employment shares in MODERATE are similar to those in QUADRUPLE. This is because the assumed lower efficiency that causes slower output growth is also associated with slower growth of nonagricultural labor productivity (on average about 3%). But the nonagricultural labor force grows slightly Less than in QUADRUPLE, and hence the share of employment in agricul- ture does not decline quite so much. 2.59 In BALANCE, employment in services grows much faster than in QUADRUPLE - parcly because the demand for commerce and miscellaneous services expands faster, partly because improved quality and the changing composition of activities within these sectors are assumed to entail slower labor produc- tivity growth. By 2000, the service sectors, which employ 14% of the labor force in QUADRUPLE, employ 25% in BALANCE - almost exactly the 1980 average for lower-middle-income developing countries, with an average per capita income very similar to that projected for China in 2000. Because industrial output grows more slowly in BALANCE, and industrial labor productivity grows just as fast, industrial employment does not increase so much as in QUADRUPLE - and indeed increases very little as a proportion of the total labor force (although in absolute terms it rises by 35 million workers). Employment in the infrastructure sectors, however, increases only slightly less than in QUADRUPLE. 2.60 The greater increase in service sector employment in BALANCE much more than offsets the smaller increase in industrial employment, so that non- agricultural employment increases by over 40 million more than in QUADRUPLE (the annual growth rate of nonagricultural employment is 4.3%). Agricultural employment in BALANCE thus increases between 1981 and 2000 by only 13 million, and declines to 52% of the labor force. This would be quite normal for a large lower-middle-income country (Figure 2.1 and Annex E, Table 3.8). 9/ The most commonly quoted Chinese projection of agricultural employment in 2000 is 225 million - 50% of a rural labor force of 450 million. The projection cited in Annex B, Table 5.1, is somewhat higher - 250 million, or 39.6% of the World Bank's total labor force projection of 631 million.  - 39 - Figure 21: Share of Agiiculture in Total Ernplment 1960-1982 (China: Projecfions 1981-2000) I : hbo 90- P . Postan E : Egtpt T T -hofld PH Phlippk, TK Tukey 80 - TK K 5 Korea 80 _ _ * M Mob3ysb & ~~~~~~~B -azeil I IN 8 Y:Yuosaj ? Yugo_kiao 70- 60- ! ~~~~~~~~~~~~~~~PH \\ L 40- 30- 20 I I I II. 125 250 500 1.000 2.W0 4000 PER CAPITA INCOME (USS ot 1962) Legendt ChNno 1981-2000 Q6ZRLPLE xcenario - ---- ChNno 1981-2000 BALJaNCE scenario WOWd Bank-26818  - 40 - 2.61 Sectoral employment sliares largely determine sectoral shares of national income, although these also depend on the amount of capital per worker, on wage rates, and on prices - especially in relation to costs. The projections of sectoral national income shares in Table 2.6 (both Western and Chinese measures are presented) assume that sectoral prices change in parallel with sectoral costs, which means that the relative profitability of different sectors remains constant (Box A). Different assumptions about prices - for example, that they do not change, or that there is movement toward equaliza- tion of sectoral profit rates - would somewhat alter the projected shares of national income. 2.62 In QUADRUPLE, the share of agriculture in national income drops by 10 percentage points between 1981 and 2000, with increases in the shares of other sectors, especially heavy industry and infrastructure. In MODERATE, the movements are similar, althor.h the share of agriculture declines less. In BALA-NCE, there is a much greater increase in the share of services, especially using the Western measure (partly because some of the increase is in nonmater- ial services, partly because the material contribution of miscellaneous servi- ces is in the Chinese measure included in the output of the material produc- tion sectors rather than of services). The share of industry in national income in BALANCE increases less by the Chinese measure, and declines signifi- cantLy by the Western measure (the difference is again due mainly to the treatment of business services). Agriculture's share of national income declines more in BALANCE than in QUADRUPLE, although the difference is more pronounced by the Western measure. Table 2.6: SECTORAL SHARES OF NATIONAL INCOME a/ (Percentages) 1981 2000 QUADRUPLE MODERATE BALANCE Agriculture 35(40) 26(30) 27(32) 23(29) Heavy industry 24(26) 28(30) 29(31) 23(28) Light industry 15(17) 15(18) 15(18) 12(17) Infrastructure 11(10) 14(14) 13(13) 13(14) Services 16(7) 17(7) 16(7) 28(13) a/ Numbers outside parentheses are Western measures (shares of GDP), those inside parentheses Chinese measures (shares of NMP). In the latter case, infrastructure excludes passenger transport, and only commerce is in services; the contribution of business services to material production is included in the net output of other sectors. Other sectoral definitions as in Table 2.4. Figure 1.2 differs from the present table mainly because of price adjustments (see Annex E).  - 41 - Box 28: WAGRS. PROFrs AND PRIICS A.l Sectoral relative prices - which are tors with increasing use of intermediate producer (ex-factory or farm-gate) prlces - inputs (such as agrlculture), or where labor are determined in the model by sectoral costs, producttvity is rising more slowly than aver- including intermediate input costs, wage costs age (such as services), or where the capital- and capital costs. Intermediate Input costs output ratio Is risl.ng faster than average depend on usage rates - input-output coefft- (such as energy), relative prices rise. cients - and on input prices. Uage costs in each sector depend on labor productivity and A.4 Conversely, In sectors where inter- on wage rates, which in turn depend on the mediate - including energy - use is reduced, rate at which wages in general are rising and or with above-average labor productivity on the relationships among vage rates in dif- growth or below-average capital-output ratio ferent sectors. In the model, general wage Increases (many industrial sectors), relative increases are tied to household consunption prices fall. Some sectors are subject to con- increases; and in these projections, sectoral flicting cost pressures, which, moreover, may relative wage rates (agricultural earnings are vary among projections - for example, agricul- includea in wages) are assumed to remain con- tural prices rise less in BALANCE than in stant in proportionate terms - except in edu- QUADRUPLE because agricultural labor produc- cation and health, where the in