Report No. 21105-CHA China Overcoming Rural Poverty October 18, 2000 Rural Development and Natural Resources Unit East Asia and Pacific Region Joint Report of the Leading Group for Poverty Reduction, UNDP, and The World Bank CURRENCY EQUIVALENTS (As of September 2000) Currency Name = Renminbi Currency Unit = Yuan (Y) Y 1.00 = $0.12 $1.00 = Y 8.3 FISCAL YEAR January I - December 31 WEIGHTS AND MEASURES Metric System I ha= 15 mu Vice President Jemal-ud-din Kassum, EAP Country Director Yukon Huang, EACCF Acting Sector Director Mark D. Wilson, EASRD Task Team Leader Alan Piazza, EASRD CONTENTS Abstract ..............................v Preface ............................. vii Acknowledgments ............................. viii Abbreviations and Acronyms ............................. ix Executive Summary .............................. xi A. The Remaining Challenge ............................. xi B. The Macroeconomic Context ............................. xiv C. Poverty Reduction Programs ..................................... xvi I. TRENDS IN RURAL POVERTY IN THE 1990s . . A. Continued Reduction of Rural Poverty , , . ...I B. Location of the Poor ......................................3 C. Other Characteristics of the Poor .8 2. THE MACROECONOMIC CONTEXT . . 14 A. Introduction ..14 B. Macroeconomic Growth in China in the 1990s . .15 C. Trends in Employment .22 D. Inflation and Prices ...................................... 30 E. Fiscal Challenges in Poverty Reduction .32 F. Conclusions and Recommendations .36 3. POVERTY REDUCTION PROGRAMS IN CHINA . .38 A. Overview ....... 38 B. Institutional Arrangements, Strategy and Funding .39 C. Targeting .43 D. Poverty Programs and Instruments ..44 E. A New Approach: Implementing Programs to Meet the Needs of the Poor . 52 4. MEASURES TO INCREASE MOUNTAIN AREA PRODUCTIVITY .. 57 A. Poverty in Mountain Areas ..57 B. Strategy for Mountain Area Economic Development .58 C. Recommendations for Improving Agriculture Program Effectiveness .59 D. Market Analysis and Development ..67 E. Environmental Protection .68 F. Education, Health and Nutrition .69 G. Poverty Reduction Through Labor Mobility . .71 H. Poverty Reduction Through Voluntary Resettlement . .71 Annexes Annex I Quantitative Trends .77 Annex 2 Macroeconomic Statistics .87 Annex 3 Improved Technology, Research, and Extension Needs for Mountain Agriculture .92 Annex 4 Participatory Poverty Assessments .109 Bibliography ........ 115 ii Maps Map 1 Incidence of Rural Poverty in China by Province, 1991 Map 2 Incidence of Rural Poverty in China by Province, 1996 Map 3 China: 592 Nationally Designated Poor Counties Map 4 China: Poor Mountainous Region TABLES IN TEXT Table 1: Number of Rural Poor, 1990-98 ......................................................... xiii Table 1.1: Number of Rural Poor, 1990-98 ..........................................................2 Table 1.2: Incidence of Rural Poverty by Region and Province, 1988, 1989, 1991 and 1996 .....4 Table 1.3: Puding County (Guizhou) Township Disparity, 1992 .................................................8 Table 2.1: Regional Patterns: Growth versus Change in Rural Poverty, Initial Level of Income and Initial Poverty Rate ......................................................... 17 Table 2.2: The Importance of Agriculture for Poverty Reduction, 1991-96 .19 Table 2.3: Percent of Rural Household Net Income Per Capita from Agriculture . 19 Table 2.4: Percent of Rural Household Net Income Per Capita from Wage Earnings and Transfers .20 Table 2.5: Growth in China's Labor Force .23 Table 2.6: Growth in Rural Enterprise Employment .24 Table 2.7: Estimates of Migrant Employment .25 Table 2.8: Official Data on Employment in Agriculture (millions) .26 Table 2.9: Estimated Change in Labor Use in Agriculture, 1989-1996 .28 Table 2.10: Regional Growth Rates in the Labor Force and Off-farm Employment .. 29 Table 3.1: Central Government Poverty Reduction Funding (Y Million) .42 Table 3.2: Yunnan Poverty Program Funding - Sectoral Spending (Yuan Million) .47 FIGURES IN TEXT Figure 1: Number of Rural Poor, 1990-97. xi Figure 1.1: Number of Rural Poor, 1990-97 .2 Figure 2.1: Growth in GDP .15 Figure 2.2: The Relationship Between Per Capita GDP and Rural Poverty, 1990-97. 1 6 Figure 2.3: Provincial Growth Rates versus Rates of Poverty Reduction .17 Figure 2.4: Wages as a Share of Income by Income Group, 1996 .21 Figure 2.5: Indices of Agricultural Employment, 1989-=100 .27 Figure 2.6: Annual Change in GDP and Prices .30 Figure 2.7: Annual Change in Farm Output, Input, and Rural Consumer Prices .31 Figure 2.8: Government Revenues and Expenditures as a Percent of GNP .32 iii BOXES IN TEXT Box 2.1: Fiscal Challenges in Three Poor Counties ................................................................ 34 Box 3.1: Multisectoral Rural Development Poverty Reduction Projects .................................... 53 Box 4.1: Key Features of Agricultural Development in the Karst Areas .................................... 60 Box 4.2: Cunji Guihua - Working with Village Development Plans .......................................... 61 Box 4.3: Typical Technology Improvements for Karst Areas ..................................................... 64 Box 4.4: Impact of Project Approach - Analysis of Cashflows for Three Poor Households ...... 66 Abstract v ABSTRACT China is widely recognized for its achievements in reducing absolute poverty since the adoption of a broad program of rural economic reforms beginning in 1978. Based on the government's austere rural poverty line, official estimates indicate that poverty declined from more than 30 percent of the rural population in 1978 to less than 5 percent by end-1998. The Chinese government has a strong commitment to poverty reduction, and the scale and funding of its poverty reduction program, and the sustained dramatic reduction of absolute poverty over the last twenty years of reform, are exemplary by any standards. Estimates based on an international poverty line document an equally steep decline in the incidence of poverty in China. However, since the international standard is somewhat less severe than China's official poverty line, it indicates greater numbers of poor in all years, and that by end-1998 a much larger share of the rural population - about 11.5 percent or some 106 million people -remained in poverty. While China's austere poverty line was a useful standard when the incidence of extreme poverty was greater, the international standard has now become a more appropriate measure to gauge the extent of poverty and guide the government's poverty reduction program in the next century. Available evidence also shows that an increasing share of the remaining rural poor are now concentrated in China's western provinces, and mostly within remote and mountainous townships. The educational, health, and nutritional status of these remaining rural poor is deplorable, and minority peoples and the disabled are known to represent highly disproportionate shares of the rural poor. Assisting these remaining poor requires the continuance of the existing poverty reduction system. The key issue is not allocating more funding for poverty reduction, but is instead making more efficient and effective use of available funding. This can be achieved through a number of measures. First, available poverty reduction funding should be targeted to all poor townships. The current system of targeting the nationally-designated 592 poor counties results in the provision of very little assistance for the half of the poor residing outside the designated poor counties, and to a very substantial leakage of assistance to the non-poor within the poor counties. Second, financial monitoring and supervision of the use of poverty reduction funds must be greatly strengthened. At present, the weak fiscal situation in China's poor areas motivates local governments to divert a large share of poverty reduction funding to alternative uses, and the very limited supervision of the use of available funding often leads to poor quality of poverty reduction works and activities. A number of government agencies provide funding and other support for poverty reduction, and the overall coordination and accountability of this array of assistance is inadequate. Third, the effectiveness of funding to increase the productivity of upland agriculture (where the majority of China's poor attempt to eke out subsistence levels of production) could be greatly enhanced through (a) adopting a multi-year "project-based" approach with greater community participation in design and implementation, (b) developing appropriate applied agricultural technologies, and (c) completing realistic assessments of the market prospects for a wide array of niche crops which are now being planted extensively in China's upland areas. Fourth, greater efforts must be made to provide the poor with improved access to basic education, health, credit, water supply, and roads and other basic infrastructure. vi Abstract Fifth, past attempts to foster enterprise development in the poor areas through direct funding have had mixed results. Local government should instead focus on providing an enabling environment for rural enterprise development. Finally, China's poverty reduction work could be enhanced by forging stronger links with other government, academic and civic organizations. The next generation of poverty work could include contracting the implementation of some small projects to grass roots and civic organizations, which might enable the poverty program to try new and innovative approaches, and improve its outreach. Preface vii PREFACE China's "8-7" National Poverty Reduction Plan, which has been associated with sustained and significant reductions in the numbers of the extremely poor during the 1 990s, will conclude at the end of 2000. The primary purposes of this report are to assess the state of poverty in China at the end of the 1990s, and to contribute to the determination of the most effective and efficient means of overcoming remaining absolute poverty in the new decade. The report is intended for all those interested in poverty reduction in China and, in particular, govemment decision-makers with responsibility for designing the next stages of the national poverty reduction program. Available evidence indicates that the majority of China's absolute poor are increasingly concentrated in remote and mountainous townships and villages in the western provinces. The report focuses heavily on these remaining rural poor, and to a lesser extent on the ethnic minority groups and the disabled people who are heavily overrepresented among the poor. At present, less than one percent of China's registered urban population have income levels below the absolute poverty line. Although reform of state-owned enterprise and other developments may result in the emergence of urban absolute poverty as a major issue in China, the report does not cover that very important issue. The report gives considerable attention to the macroeconomic context, and finds that the trends in poverty reduction in the 1 990s have been at least partly determined by larger macroeconomic trends. The Asian financial crisis of the late-1990s dramatically confirmed that past achievements in poverty reduction could be quickly reversed. Although its effects on the poor in China were muted, it should be remembered that a future macroeconomic downturn could threaten China's success in poverty reduction. viii Acknowledgments ACKNOWLEDGMENTS This joint report was undertaken as a collaborative effort of the Leading Group for Poverty Reduction, United Nations Development Program, and The World Bank. The LGPR team was led by Su Guoxia, the UNDP team was led by Terry McKinley, and The World Bank team was led by Alan Piazza. The UNDP team, in collaboration with the International Labor Organization, has written an additional Supplement to this joint report. (The Supplement, which is available separately, provides a complete listing of the UNDP and ILO staff and consultants involved in the work of the joint report and the Supplement.) The report was written by Alan Piazza and Julia Li (East Asia and Pacific Region, Rural Development and Natural Resources Sector Unit), Enjiang Cheng, Claude Saint-Pierre, Terry Sicular, Bruce Trangmar, and Robert Weller (consultants). Major inputs were provided by Terry Mckinley (UNDP), Hu Tao, Kang Xiaoguang, Yang Qiulin, Zhao Yaqiao, and Richard Hardiman (UNDP consultants), Shaohua Chen and Martin Ravallion (Development Economics, Development Research Group), Li Guo (East Asia and Pacific Region, Rural Development and Natural Resources Sector Unit), Louise Beynon, Lisa Croll, and Zheng Baohua (DFID Consultants), and Lina Song and Wang Yihuan (The World Bank consultants). Additional inputs, comments, and review were provided by Halsey Beemer, Chen Xiaoping, Cheong Chup Lim, W. Hunter Colby, Yuri Dikhanov, Zafer Ecevit, He Jin, Paul Heytens, Janet Hohnen, Bert Keidel, Valerie Kozell, Kathie Krumm, Lena Lindberg, Lu Lei, Albert Park, Thomas Rawski, Arlene Reyes, Scott Rozelle, Mikiko Sasaki, Sari Soderstrom, Juergen Voegele, and Wu Guobao. The report is in large part based on findings of missions which visited Beijing, Anhui, Chongqing, Guangxi, Guizhou, Hunan, Qinghai, Sichuan and Yunnan during July 1998 and October/November 1998. These missions comprised a number of LGPR, UNDP and The World Bank staff and consultants, and benefited greatly from meetings and support from a number of staff, researchers, and officers of a variety of government agencies, donors, and domestic and international NGOs. In particular, the missions would like to gratefully acknowledge the time and generous assistance provided by the national, provincial, and lower level offices and various departments of the All China Women's Federation, China Disabled Persons' Federation, Chinese Academy of Agricultural Sciences, Chinese Academy of Sciences, LGPR's Foreign Capital Project Management Center, Ministry of Agriculture, Ministry of Science and Technology, Ministry of Water Resources, State Ethnic Affairs Commission, State Forestry Administration, State Statistical Bureau, CIMMYT, Ford Foundation, GTZ, IFAD, UNICEF, Winrock International, and WFP. The missions also benefited greatly from small group meetings with many poor farm households at the local level. Their assistance and participation is very much appreciated, and it is hoped that this report contributes to the further improvement of their well- being. Abbreviations and Acronyms ix ABBREVIATIONS AND ACRONYMS 8-7 Plan 8-7 National Poverty Reduction Plan ABC Agricultural Bank of China ADB Asian Development Bank CAS Chinese Academy of Sciences CDPF China Disabled Persons' Federation CEIS China Economic Information Service DFID Department for International Development FTA Farmer Technical Association FFW Food for Work FURC Foundation for Underdeveloped Regions in China GDP Gross Domestic Product GONGO Government Organized NGO ICIMOD International Centre for Integrated Mountain Development IFAD International Fund for Agriculture Development ILO International Labor Organization ILUS Integrated Land Use Systems Project LGPR Leading Group for Poverty Reduction MOA Ministry of Agriculture MOF Ministry of Finance MOST Ministry of Science and Technology MWR Ministry of Water Resources NGO Non-Governmental Organization PADO Poor Area Development Office PMO Project Management Office PPA Participatory Poverty Assessment PPP purchasing power parity PRA Participatory Rural Appraisal QBPRP Qinba Mountains Poverty Reduction Project RCC Rural Credit Cooperative SDPC State Development and Planning Commission SEAC State Ethnic Affairs Commission SEU Sheep Equivalent Units SFA State Forestry Administration SGWS Sichuan Grassland Working Station SOE State-Owned Enterprise SPC State Planning Commission (now SDPC) SSB State Statistical Bureau x Abbreviations and Acronyms SWPRP Southwest Poverty Reduction Project TVE Township and Village Enterprise UBE Universal Basic Education UNDP United Nations Development Program VDP Village Development Plan WFP World Food Programme WF Women's Federation WS Work Station Executive Summary xi EXECUTIVE SUMMARY A. THE REMAINING CHALLENGE 1. Twenty Years of Poverty Reduction. China is widely recognized for its achievements in reducing absolute poverty since the adoption of a broad program of rural economic reforms beginning in 1978. Broad participation in the subsequent reform-driven economic growth, together with a well-funded national poverty reduction program, have brought about a tremendous reduction in rural absolute poverty during the past twenty years. Official estimates indicate that rural poverty declined from roughly 260 million poor in 1978 to 42 million in 1998, or from one-third to about one-twentieth of total rural population. These official estimates are based on the government's austere poverty line equivalent to $0.66 per day (in constant 1985 purchasing power parity dollars). Figure 1: Number of Rural Poor, 2. The World Bank has developed 1990-97 (Source: Table 1) an international poverty standard of $1.00 per day (in 1985 purchasing power parity dollars) for cross-country 8250 comparisons of poverty. Estimates a. 200 based on this somewhat less severe '0- poverty line indicate substantially 50 greater numbers of absolute poor in China in all years, but do confirm the OZ 591, c 9 ON 9' lb 9 PO continuing remarkable decline in l poverty during the 1990s.1 As shown in Year Figure 1, the international standard l indicates that by end-1998 a much | Government Estimates ($0.66d larger share of the rural population - l International Standard ($1/day) about 11.5 percent or some 106 million people -remained in poverty.2 3. The government's austere poverty line was a useful standard when the incidence of extreme poverty was greater, and helped to target available poverty reduction funding to those most in need. However, the government's poverty line may no longer be such an effective tool in identifying the potential beneficiaries of national poverty reduction efforts since such a small share of the population now have income levels below it, and large numbers of transient poor have income levels just slightly greater than it. In addition, those with incomes consistently below The significance of the difference between the official government estimates of poverty and those based on the international standard is discussed in The World Bank (1996). As shown in Annex 1, the sharp decline in rural poverty during the 1990s is observed across a variety of poverty lines ranging between $0.5 to $2.5 per day. Available evidence indicates an increase in urban poverty during the 1990s. However, at present, less than one percent of the registered urban population have income levels less than the dollar per day international poverty line. Consequently, in 1998, the 106 million rural poor greatly exceeded the less than 3 million registered urban poor. xii Executive Summary the poverty line suffer extreme hardship and deprivation. Consequently, China should consider whether the international standard may now be a more appropriate measure to gauge the extent ofpoverty and guide its poverty reduction program. 4. The Chinese government has a strong commitment to poverty reduction, and most government ministries and agencies have special poverty reduction responsibilities and projects. The State Council's Leading Group for Poverty Reduction (LGPR) was established in 1986 in part to provide greater coherence to these many poverty reduction initiatives and, in particular, to expedite economic development in the poor areas. LGPR has emerged as the principal advocate of China's rural poor, and is now the key agency responsible for coordinating the nation's more than US$2 billion in annual funding for poverty reduction programs. This funding is organized under China's "8-7 Poverty Reduction Plan" (8-7 Plan), which was established by LGPR in 1994 to overcome remaining absolute poverty in the nationally-designated 592 poor counties. The scale and funding of China's 8-7 Plan poverty reduction program, and the sustained dramatic reduction of absolute poverty over the last twenty years of reform, are exemplary by the standards of any developing country. 5. Current Challenge. Even with the successes to date, there are still more than 100 million rural absolute poor, and in most cases this remaining poverty is both severe and relatively intractable. In the past, the broad incidence of poverty made it possible to achieve substantial reductions in poverty through general economic growth and through programs that were more broadly targeted. At present, the majority of the rural poor are concentrated in resource deficient areas, and comprise entire communities located mostly in upland sections of the interior provinces of northern, northwestern and southwestern China. Although these poor have land use rights, in most cases the land itself is of such low quality that it is not possible to achieve subsistence levels of crop production. Consequently, in most years the poorest of the poor consume grain and other subsistence foods beyond their own production levels. The poorest households are typically those further disadvantaged by high dependency ratios, ill health and other difficulties. Minority peoples and peoples with disabilities are known to represent a highly disproportionate share of the rural poor. Poverty also exacerbates society-wide problems of lower rates of female participation in education, higher relative female infant mortality rates, and higher rates of maternal mortality. Available evidence suggests that the severity of the remaining poverty worsened somewhat during much of the 1990s. As shown in Table I below, the squared poverty gap index increased during the 1990s, and in 1998 remained considerably greater than in 19903 6. The educational, health and nutritional status of these remaining absolute poor is deplorable. As many as half of the boys in many of China's poorest villages and, particularly in some minority areas, nearly all of the girls do not attend school and will not achieve literacy. Infant mortality rates and maternal mortality ratios in very poor counties exceed 10 percent and 0.3 percent respectively (or at least 50 percent to 100 percent greater than the national average), and are greater yet in the poorest townships and villages. Incidence of several infectious and endemic diseases, including tuberculosis and iodine deficiency disorders, is concentrated in poor and remote areas. Roughly half of children in households at or below the absolute poverty line are at least mildly malnourished (stunted), and iron, vitamin A, and other micronutrient deficiencies remain a severe problem among the poor. As many as 90 percent of poor children suffer chronic worm infections. 3 The squared poverty gap index is a measure of the income that would be necessary to bring a poor person up to the poverty line. A greater index value indicates greater severity of poverty. Executive Summary xiii Table 1: Number of Rural Poor, 1990-98 International of ficial Government Estimnates Standard ($ 1/day) Poverty Number of Share of Rural Squared Number of Share of Rural Line /a Rural Poor Population Poverty Rural Poor Population Year (Currenit Y) (million) (percent) Gap Index (million) (percent) 1990 300 85 9.5 0.43 280 31.3 1991 304 94 10.4 1.15 287 31.7 1992 317 80 8.8 0.85 274 30.1 1993 350 75 8.2 1.36 266 29.1 1994 440 70 7.6 1.38 237 25.9 1995 530 65 7.1 1.00 200 21.8 1996 580 58 6.3 0.58 138 15.0 1997 640 50 5.4 0.85 124 13.5 1998 635 42 4.6 0.67 106 11.5 /a The official government poverty line is the equivalent of $0.66 per day.. Note: Both sets of estimates are derived from per capita income data from the State Statistical Bureau's annual sample survey of rural households. 7. Regional Concentration of the Poor. The reduction in rural poverty has been greatest in China's coastal and central regions where rural economic growth has been greatest. Many of the rural poor in 1978 resided in less remote and less hilly areas in the coastal and central regions, where increased application of fertilizer, irrigation, better seed and other modem inputs could bring about rapid productivity gains, and so were better able to participate in the rapid agricultural growth of the reform period. Most of the residual poor have remained trapped in more remote upland areas where agricultural productivity gains have proven far more problematic. Available evidence shows that this trend has continued in the 1990s, with an increasing share of the rural poor now concentrated in China's western provinces. The number of poor in the western provinces (including both the northwestern and southwestern provinces) increased from less than half of all of China's rural poor at the beginning of the 1990s to more than two thirds (about 70 percent) by 1996. This increasing concentration of poverty in the western provinces is observed using both the official government poverty line and the international dollar per day poverty line. The change in the distribution of remaining rural poverty was driven by sharp declines in the incidence of poverty in the eastern and central provinces relative to the more modest achievements in the western provinces. Poverty reduction in the northwestern provinces appears to have been particularly limited, and the severity of poverty is also deepest in these provinces. 8. Mountain Area Poverty. In addition to this concentration of poverty in the western provinces, it is believed that the majority of China's poor reside in mountainous counties and townships. A number of observers have concluded that China's poverty problem mainly occurs in poor mountain regions, and China's poverty reduction programs have long focused on mountain areas. Available information and field visits in a number of poor areas have documented considerable disparity at the county level, and confirm that a very high proportion of the rural poor live in mountain townships. These poor mountain townships are located in the nationally- designated and provincially-designated poor counties, and in a number of other counties not specifically designated as poor. Administrative boundaries in mountain counties typically delineate better-off townships around the county seat and in wide valley floors ("baqu'), and poorer mountain townships in higher elevation outlying areas ("shanqu"). In Yunnan, for example, the provincial government selected 506 key townships for poverty reduction programs beginning in 1995. Almost all of these poor townships are distributed in high mountain ranges, in xiv Executive Summary steeply-sloped mountains, and in minority and border areas. Some 86 percent of Yunnan's remaining poor in 1996 reportedly resided in these 506 poor townships. B. THE MACROECONOMIC CONTEXT 9. Economic Growth. Overall economic growth explains much of China's record of success in poverty reduction since 1990. Both over time and across provinces, growth in per capita GDP has been closely associated with the pace of poverty reduction. Available evidence also confirms that the impact of aggregate growth on poverty in China has been substantially influenced by the regional and sectoral composition of that growth. Slower-than-average growth in poor regions explains in part the increasing regional concentration of poverty. In addition, uneven growth in agriculture, the main source of income for the rural poor, has contributed to differences in the rate of poverty reduction. Poverty reduction has been slower where agricultural growth has lagged, and faster where agricultural growth has more or less kept pace with that in other sectors. For these reasons, macroeconomic policies that promote growth, especially those that promote efficient agricultural growth and that target regions with high concentrations of poor such as the recent infrastructure investment program, should be seen as highly complementary to microeconomic poverty interventions. 10. Employment. Trends in employment further highlight the importance of agriculture. Agriculture is often viewed as a passive reservoir for China's surplus labor, and expansion of nonagricultural activities is seen as the main route for. absorbing this surplus labor. A closer examination suggests, however, a different understanding of agriculture's role. Estimates of agricultural employment show that in the 1990s labor use in agriculture has risen and suggests that agricultural employment responds actively to opportunities for diversification and to increased relative prices. Agriculture's potential to provide employment has been greatest in high- value, labor-intensive crops, horticulture, and livestock production. This potential can be promoted by ensuring that farmers have better access to water, credit, capital, and technology, and by allowing farners to produce what is most profitable. During the 1990s, off-farm employment has also been an important source of new employment, with migration and private and individual enterprises playing a growing part in generating jobs. Poverty programs should recognize these trends and assist the poor to gain access to migrant jobs and to participate in private businesses either as entrepreneurs or employees. 11. Fiscal and Financial Challenges. During the 1 990s the Chinese govemment has demonstrated a strong fiscal commitment to poverty reduction. This commitment has faced challenges on two levels. First, it has faced challenges at the central level because of ongoing fiscal deficits and the declining share of fiscal revenues in GDP. Second, it has faced challenges at the local level because of the extremely weak fiscal situation of poor counties. Despite its precarious fiscal situation, the central government has to date maintained and increased funding for poverty reduction. Maintaining funding, however, is only the first step in coping with fiscal pressures, because such pressures exist not only at the center but also at lower levels of government. Fiscal pressures are especially strong for poor counties, which face severe and persistent fiscal shortfalls. 12. The difficult fiscal situation of poor counties holds implications for the effectiveness with which central poverty reduction funds are used. The LGPR system is responsible for seeing that funding for poverty reduction is used in ways that benefit the poor. Poor county governments, however, face different incentives and have different objectives. First, they face pressing expenditure demands, including simply paying back wages to employees on the government Executive Summary xv payroll. Second, revenue-starved county governments face strong incentives to use poverty funds in ways that quickly generate local fiscal revenues. The result has been an incentive to spend poverty funds to cope with pressing budgetary problems or to promote rural industry and township and village enterprises (TVEs). The central government has recently responded to this problem by limiting the amount of poverty funds spent on rural industry. Yet even when poverty funds go to agriculture and basic infrastructure, the spending may not reach the poor. The returns to agricultural and infrastructure investments in nonpoor or even marginally poor townships tend to be higher, emerge more quickly, and be more easily taxable than the returns on such spending in the poorest townships and villages. Actions should be taken to address the underlying fiscal difficulties in poor areas and to provide incentives for local governments to assist the poor. Continued fiscal reforms are needed to broaden the local tax bases and generate more budgetary revenue, including from agriculture and the private sector. The distribution of the new tax system on inter-governmental distribution should be assessed, and changes made to improve net fiscal flows to poor areas. 13. Similar problems exist in the financial system. Most of poverty reduction funding is in the form of loans. Reforms in the banking system have caused the Agricultural Bank of China (ABC), which manages poverty loans, to behave increasingly as a commercial, profit-oriented bank. Commercial banks worldwide view poor farm households as poor credit risks. Risk of nonrepayment is high, and in China where land is not privately owned, poor households have little to offer in the way of collateral. Moreover, farm households are numerous and dispersed, the loan amounts are small, and administrative costs are high. ABC therefore has disincentives to direct poverty loans to poor farm households. These problems are compounded by the fact that the interest rate on poverty loans is low, so that such loans would be unprofitable even in the absence of the greater risk and transactions costs. In the past the government subsidized the ABC's losses due to the low interest rates on poverty reduction loans, but now government subsidies do not fully cover the losses. As the financial system moves towards the market, new methods are needed to provide credit to the poor. The policy requiring low interest rates on poverty loans creates perverse incentives and should be reevaluated Microfinance programs are now a significant component of China's poverty programs, and could be an effective way to provide credit to the poor and reduce poverty if they were better designed. 14. Outlook. Recent weak aggregate demand and slowing growth raises concerns about whether China's successful record of poverty reduction can be maintained in the future. A review of China's past macroeconomic performance suggests that China has followed a pattern of cyclical growth since the start of the reforms, and that the recent slowdown is a continuation of this cyclical pattern. International experience has shown that most countries undergo cycles in aggregate demand, and that poverty is sensitive to such cycles because of their impact on unemployment, inflation, the terms of trade, and fiscal capacity. Indeed, macroeconomic fluctuations can have devastating and long-term effects on the poor, as many of the poor live close to subsistence and typically have limited opportunities to insure against income shocks. For these reasons, the continued success of China's poverty reduction policies will require increased attention to the effects of macroeconomic cycles and fluctuations on the poor. Counter-cyclical macroeconomic policies can be beneficial to the poor, and microeconomic poverty interventions should continue to help poor households reduce production risk, diversify sources of income, and gain reliable access to sources of credit. Consolidation of China's past successes in poverty reduction will depend on efforts in this direction, because many of those who have recently escaped poverty are close to the poverty line and remain vulnerable. xvi Executive Summary C. POVERTY REDUCTION PROGRAMS 15. While recognizing China's exemplary success in reducing poverty, this report notes that international standards indicate that there are still more than 100 million rural poor concentrated in the western provinces and mountainous regions. To meet this challenge, it is important that China's existing poverty reduction program be extended, and funding continue at current levels, into the next decade (following the conclusion of the current "8-7 Plan" implementation period in 2000). Furthermore, it is essential that the effectiveness and efficiency of the existing program be enhanced through better coordination of the existing funding, institution building, more rigorous supervision of all poverty reduction works and activities, better targeting, and increased beneficiary participation. Adoption of a multisectoral project-based approach could also enhance program effectiveness. Coordination and Institutional Arrangements 16. Institutions. China's poverty reduction program comprises a wide variety of actors, programs and funding channels. The bulk of the central government's $2 billion of annual funding for the poverty reduction program is channeled through subsidized loans provided through ABC and the banking system, the Food for Work Program ("Yigongdaizhen") administered by the State Development Planning Commission, and grants administered by the Ministry of Finance (MOF).4 LGPR and its executive agency, the Poor Area Development Office (PADO) system which extends down to the county and township levels, has responsibility for the overall success of China's poverty reduction program and for the coordination of the large number of poverty reduction activities of these other government ministries and agencies. With the exception of some lesser programs under their immediate control, the LGPR system does not directly implement poverty reduction projects and activities. Instead, most poverty reduction projects and activities are managed or implemented by the sectoral agencies generally responsible for such activities. Rural roads constructed under the Food for Work Program, for example, are implemented by local staff of the Transport Bureau. 17. Implementation of the programs has been a challenge for the government, and it is believed that there is room for substantial improvements to program effectiveness. The subsidized loan program, which comprises half of all poverty reduction funding, has (a) found that its lending to enterprises has done little to reduce poverty, and (b) encountered significant difficulties in directing loans to poor households. These issues, compounded by low repayment rates on the program, convinced the Government to experiment with microcredit schemes in the second half of the 1990s, but these have also encountered serious difficulties. The Food for Work Program could be improved by more explicit targeting of the poorest areas, and by refocusing its work on the types of programs that bring the greatest benefits to the poor. Little is known about the MOF grant program, but it too is believed to have encountered a number of implementation difficulties. 18. The effectiveness of the entire poverty reduction program could be greatly improved through much stronger institutional arrangements. Despite its mandate to coordinate the nation's poverty reduction program, the LGPR system does not have (a) sufficient advance access to the detailed program informnation necessary to insure that all poverty reduction funding benefits the poor, or (b) the staffing to properly oversee the quality of poverty reduction project works and 4 The $2 billion of annual funding does not include additional funds and support for the poor including, most importantly, funding for poor area health and education administered by the Ministry of Health and the Ministry of Education. Executive Summary xvii activities. This limited control and inadequate staffing directly contribute to the significant leakage of available funding to works, activities and other uses which have little or no benefit to the poor, and to an unacceptably large share of the works and activities which do benefit the poor but which are of inferior or substandard quality. 19. At the provincial and lower levels, officials often complain that the LGPR system has been given responsibility for ensuring that poverty reduction goals are achieved, but not sufficient control over the funding or the necessary staffing to fully complete this important mission ("Fupinban you zeren, mei you quanli"). The LGPR system has only limited access to data on the projects and activities undertaken through the Food for Work and the MOF grant programs, and at best has only limited influence during the planning stage over the works and activities undertaken with subsidized loans provided through the Agricultural Bank of China. The current arrangement of relying on local bureaus and agencies to implement project works and activities must be maintained, but the LGPR system should play a greater role in the planning of howfunds are to be used, and in the supervision and monitoring of their usage. It is therefore strongly recommended that, in order to raise the impact of the funds on the poor and reduce the leakage of funding to alternative uses, the LGPR system should take on a greater planning and supervisory role for the use of the funds. 20. Roles of Other Institutions. Forging stronger links with government line bureaus, academic and civic organizations involved in poverty work would increase LGPR's effectiveness in setting policy and implementing programs. For example, LPGR's policy making function could be strengthened by contracting research work to organizations with specialized knowledge, and incorporating research findings into their policies and strategies. The Ministry of Agriculture (MOA), for instance, could make a far greater contribution to China's poverty reduction program by working with the LGPR system to create a development strategy for the karst region, and to initiate a program of applied agricultural research for mountain areas. 21. The next generation of poverty work could include contracting the implementation of some small projects to grass roots and civic organizations. For example, government organized NGOs (GONGO) have proven successful at implementing poverty projects in a number of areas. Yet most of their funding is self-raised, and official poverty alleviation funds are not channeled through them. Experimenting with channeling a portion of the poverty funds through GONGOs and other grass roots organizations could enable the poverty program to try new and innovative approaches, and improve its outreach. Such an approach has proven highly successful in other developing countries, and could be particularly valuable to work in China's minority areas. 22. Accountability. Strengthening LGPR's oversight and control over poverty reduction should be complemented by reciprocal measures to increase LGPR's accountability by improving monitoring of the impact of the poverty program and the use of poverty funds. The monitoring function should be contracted to an independent outside organization. Financial monitoring should trace the flow of funds, and determine whether the use of funds meets LGPR's guidelines. Impact monitoring could build on the State Statistical Bureau's rural household survey, and LGPR should rely more heavily on those survey data in targeting its programs and evaluating their effectiveness. 23. Multisectoral Project-Based Approach. Experience with a number of large- and small- scale projects during the 1990s in China suggests that one of the most effective means of assisting the absolute poor is through an integrated set of interventions in the form of a multiyear project. In most cases, the multisectoral rural development project model includes an integrated program xviii Executive Summnary of investments in (a) upland agricultural development, using menus of field and tree crop and livestock activities to increase upland agricultural productivity, (b) labor-intensive construction of rural roads, drinking water systems, small scale irrigation, agricultural drainage works, and other rural infrastructure, (c) provision of off-farm employment opportunities through a voluntary system of enhanced rural labor mobility for the upland poor, (d) institution building and poverty monitoring, and (e) rural enterprise development. Improved access to basic education and health, and separate microcredit components, are included in some donor supported projects. 24. The Chinese authorities have expressed their intention to roll out the multisectoral approach over a larger area, and some provinces have begun experimenting with this approach. Given the favorable experience with this new approach, it is recommended that China consider using a portion of the poverty reduction subsidized loan funds blended together with grant.funds to fund integrated multisectoral projects in the poorest areas. However, the multisectoral rural development projects are very demanding of design and implementation capacity and, at present, funding and institutional arrangements may not be adequate to enable an effective rollout of this approach. Better coordination between funding channels is necessary to enable the poverty program to direct sectoral components in an integrated manner to the poorest of the poor. Most critical to the successful implementation of this approach are strong institutional arrangements. This includes establishment of effective project management offices (PMOs) and work stations at the provincial, county, township and village levels, and implementation of a rigorous works supervision and acceptance process referred to as the "yanshou" system. At present, the number of staff involved in specific poverty work at the provincial and county levels are probably sufficient to staff functional PMOs. At the township level, however, staffing is inadequate. It is recommended that all township poverty reduction work stations comprise at least three full time staff. Assuming this model would be applied to all of China's poor townships, the estimated annual cost of strengthening township poverty reduction work stations to reasonable levels would be less than Y 100 million, or about 0.4 percent of total annual poverty reduction funding. Reducing direct poverty funding by this amount would not significantly reduce the scale of the program, and would likely have a significantly positive impact on the effectiveness of the program. Targeting 25. Since they were first established in 1986, the national and provincial lists of poor counties have played an important role in organizing China's poverty reduction program and helped concentrate available poverty reduction funding in some of the areas of greatest need. However, it is now evident that the current system of county-based targeting results in (a) a severe dilution of available poverty reduction funding for the half of China's remaining poor who reside in the nationally-designated poor counties, and (b) the near complete omission of central government poverty reduction funding for the other half of the poor who reside outside these poor counties. This severe problem could be largely resolved by switching to a system of township- based targeting which would direct the bulk of available funding to poor townships within and outside of the nationally-designated poor counties. 26. Just over half of China's remaining rural poor reside in the nationally-designated 592 poor counties. Most of the central government's poverty reduction funding is allocated to the nationally-designated poor counties, but is not specifically earmarked for the poor households or the poor townships within these counties. Instead, these funds are distributed both to poor and nonpoor townships within the poor counties. Therefore, it appears that available poverty reduction funding is equally distributed to all 200 million rural inhabitants of the 592 poor Executive Summary xix counties. Since only about 21 million absolute poor resided in the poor counties in 1998, it is evident that assistance to the most needy was diluted by roughly ten-fold by the leakage of benefits to the nonpoor. Just as importantly, very little central government poverty reduction funding is provided to the other half of the poor residing outside the poor counties. Most provinces use their own funding to assist the poor residing outside the poor counties. However, provincial funding levels are known to be very limited. Overall, therefore, the total poverty reduction assistance package available for the poor outside the nationally-designated poor counties is extremely limited. 27. Targeting All Poor Townships. Most of China's remaining poor reside in poor townships within and outside of the nationally-designated poor counties.It is therefore strongly recommended that, as a most crucialfirst step, all available poverty reduction funding should be channeled directly to poor townships within and outside of the 592 nationally designated poor counties, and poor villages should be guaranteed access to these funds. Pilot tests of this switch from county to township based targeting should be undertaken immediately, and need not wait until the completion of the 8-7 Plan. A switch from county to township targeting would increase the administrative costs of the program. The most obvious implication of township targeting would be the need for additional staff at the township level. Lack of staff at this level is currently one of the major weaknesses of the poverty loan program, and switching to township targeting would certainly necessitate additional staffing at the township level. However, the costs of doing so appear modest, particularly when set against the improvements in targeting and program effectiveness that are likely to occur with this switch. 28. Ethnic Minorities. Ethnic minorities are also heavily represented in the remote, mountainous townships and villages, and solving the problems of absolute poverty will help solve the most fundamental problems of minorities. Minorities constitute such a large proportion of the remaining absolute poor because past programs have often not reached out to the most remote areas where the obstacles to poverty relief are greatest. Therefore, concentrating project resources on the absolute poor will often automatically target minorities. Targeting all poor townships should also benefit the minorities who presently live outside the boundaries of both the poor counties and the minority regions, and thus currently receive neither poverty assistance nor special types of relief administered in minority regions. 29. Disabled Persons. Improving targeting of the rural poor will not necessarily help all of the needy disabled since, while the disabled are heavily overrepresented amongst the poor, even greater numbers of the disabled poor are found outside the poor areas. Poverty programs for the disabled will thus need to move beyond the traditional limits of officially designated poor counties. For those disabled who cannot work at all, the current welfare system does not effectively address the vicious circle that exists between poverty and disability. Very few of the disabled poor who are completely dependent on aid qualify as "Five Guarantee" households, and at most they receive only small amounts of aid, leaving them well below the poverty line. 30. The flagship program for the disabled who can work is financing for "rehabilitation poverty relief loans" (kangfu fupin daikuan). This program is currently budgeted at YSOO million annually. On the assumption that the funds already provided to the nationally designated poor counties will help all the poor there, the rehabilitation loan program only applies in non-poverty counties. However, disabled households do not always benefit from existing poverty relief programs in poor counties since such aid has tended to go first to those families most likely to show quick progress (and therefore often does not reach disabled households at all). This situation could be ameliorated either by extending the reach of the rehabilitation loan program to xx Executive Summary include all counties, or by building explicit measures aimed at the disabled into future poverty reliefprograms. 31. Gender and Poverty. Participatory poverty assessments are now providing greater understanding of the ways in which poverty disadvantages women in poor regions and poor households. First, the assessments draw attention to the heavy demands on female labor, and the rigid gender divisions of labor. Second, village and household assessments have shown how improving local infrastructure (most importantly, water supply, rural roads and paths, and electrification) can reduce the excessive demands on labor and improve the quality of life. Third, there is inadequate support for specialized training for women in agricultural technologies and for a variety of other knowledge and information that poor women require. Fourth, poor women and girls are particularly disadvantaged in access to basic education and health services. Finally, these assessments found that women's kin relations and their own organizations are important mechanisms for poverty reduction. 32. Participation. Development work around the world has found that allowing stakeholders a voice in project design, management, and evaluation improves results. Such approaches were not tried much in China before the 1990s, and still tend to occur primarily in programs supported by international organizations. Where they have been used, however, the results have been encouraging. Most of China's experiments with participation have been in project identification and preparation. Some projects are systematically developing strategies by creating a dialogue between experts and local beneficiaries. In many cases these result in a menu of options, which allows projects to adapt flexibly to local needs. Most recently, the results of participatory poverty assessments in several villages in Guangxi, Yunnan and Ningxia powerfully document the poor's own understanding of their experience with poverty, and highlight a number of significant weaknesses of poverty reduction activities and projects in their villages. 33. Less work has been done on building participatory institutions for project management, and the few experiments which have been undertaken have had mixed results. Local officials and experts are sometimes unhappy with what they perceive as a diminution of their authority, and China has very little precedent on which to model such independent institutions. On the other hand, there has been some success with institutions that combine local government and existing social networks, ranging from water management associations to small loan societies. Finally, monitoring and evaluation systems have very rarely tried participatory methods in China, and tend in general to lack qualitative input. The following recommendations, if applied to future poverty reduction programs in China, might further increase benefits to the rural poor: * Projects that bring management organization down to the village level typically respond more effectively to local needs. * The most successful community participation builds on existing social networks. Small loan societies, for example, work better in groups that already have close social ties than in ad hoc combinations with no other interests in common (as often happens in China). Participatory institution building will also work best when it builds on natural community leadership. * Local project officials need training programs both to learn participatory methods and to understand their goals. These methods are a significant departure from the usual cadre work style in China, and training programs need to show the benefits to local leaders. Executive Summary xxi * Attempts to build new participatory administrative units can be improved by building the interests of local government stakeholders into the project or by building more community input into existing institutions. * Minority community participation is especially important to help projects adjust to local cultural differences. Representatives from the local Ethnic Affairs offices should also be built into project management teams in minority areas. Program Design and Instruments 34. The greatest endowment of most of the rural poor is their own labor, and increasing the productivity of this labor is the main development objective of the poverty reduction program. Productivity can be increased through investment and support for (a) human capital, including education, health and nutrition, (b) farm and community level basic infrastructure including land improvements, roads, and power, (c) appropriate agricultural and other applied technologies, and their extension to the poor, and (d) improved access to microcredit and off-farm employment opportunities.5 35. Education and Health. A variety of programs have led to significant improvements to the poor's access to basic education and health services during the 1990s. The Ministry of Education and the Ministry of Health each administer special programs to improve the education and health status of the poor. In addition, Project Hope, Spring Bud, and other programs have funded the construction and repair of schools and the provision of desks and chairs and tuition assistance for the poorest children in many of China's poorest areas. Despite these achievements, extremely low levels of educational attainment, poor health, and malnutrition are still major contributors to, and at the same time partly the results of, absolute poverty in the rural areas. Achieving poverty reduction goals therefore requires that even greater levels of assistance be devoted to support for improved access to basic education, health and nutrition services for the absolute poor. International donor-supported poverty reduction programs have shown that activities to improve access to basic education and health services can be successfully integrated into larger multisectoral poverty reduction projects and are very warmly received by project beneficiaries. The successful integration of basic education and health services into these multisectoral projects had the additional important payoff of better mobilizing the local community in support of the overall project and significantly enhancing beneficiaries' participation in all aspects of the projects. 36. The Ministry of Education has established national implementation plans for achieving nine-year universal basic education (UBE) by 2010. However, in accordance with different regions' financial resources and abilities, the national plans allow for lesser degrees of UBE in China's poor areas. For the poorest 5 percent of the population, the national plan calls for only three- to four-year UBE. (For better-off areas, the national plan calls for achieving nine-year UBE by 2000.) In order to reduce inequalities in the provision of education, the Government should define a minimum level ofprovision for all children and should then ensure that there is adequate funding for this provision so that the disparity of needs and resources between richer and poorer areas is overcome. This would require that the central and provincial governments substantially increase the earmarked intergovernmental funds for compulsory education in the poor counties as part of the policy to achieve nine-year UBE throughout the country. The provincial governments should guarantee a minimum level of per-student funding for all counties in the 5 Rural infrastructure is discussed in the Supplement to this report. xxii Executive Summary province, at least at the primary and secondary levels, and limit the differences in per-student public spending between counties through equalization support (The World Bank, 1999c). 37. Similarly, the "Health for All in 2000" program calls for the re-establishment of the rural cooperative medical insurance system in all areas of China and an increase in the training of students from rural areas in secondary medical schools. However, these targets alone will certainly not be sufficient to upgrade basic health services and status in the poorest areas to acceptable levels in the near future. Instead, the current situation of inadequate public funding of essential services, even for cost effective prevention and treatment programs which have large public returns, must be addressed directly and urgently in the poorest areas. The central and provincial governments should increase their assistance to the poor areas in support of a limited set of health services directed at the principal causes of morbidity and mortality. At a minimum, this should include increased public funding for the control of infectious disease, overall disease surveillance and reporting, health informnation and education, and the strengthening of the basic infrastructure of the health system in the poorest areas. 38. Mountain Agriculture. China's poverty reduction programs have focused on mountain areas, and given that a large share of the remaining absolute poor are concentrated in core poor mountain townships or are scattered at high elevations and in difficult environments across other mountain townships, this geographic focus remains valid. However, there is no national framework for mountain area development, and existing sectoral strategies and mountain county governments have tended to underestimate both the problems (including resource constraints, market risks for specialty products, and watershed management costs) and the development opportunities facing the mountain areas. It is therefore strongly recommended that a long-term national strategy be established for the economic development of mountain areas. This would serve as a basis to (a) assess the impact of national trends and policy changes on mountain areas, (b) better prioritize national, provincial and donor investment, and (c) provide poor mountain counties with an appropriate framework within which to develop local strategies and programs. The Loess Plateau has already benefited from the development of such a regional strategy. The development of that strategy was subsequently complemented by coordinated investments into economic development, watershed management and agricultural research. A regional initiative should now be taken under the national strategy to fund agricultural research and development in the karst areas where perhaps as many as half of the rural poor now live. 39. Greater attention should be paid in the national and regional strategies to market development and the roles of the private sector. There is at present an urgent need to objectively assess the viability of markets for a number of specialty products now being introduced and supported in poor mountain areas. Particularly for some tree crops, medicinal tubers, and other specialty products, remote mountain areas face greater-than-average market risks. Since the investment cycle for many of these products requires substantial initial investments and a long (often several years) maturation period before harvest, there is the serious risk that markets will deteriorate after the investment has been made but before full or even initial production begins. In addition, China's several hundred poor mountain counties compete with 300 nonpoor mountainous counties and over 600 hilly counties in largely unsegmented markets for these products. Unfortunately, the decision to invest in these products in poor mountain counties often lags that in better-off areas by several years, thereby increasing the likelihood that poor mountain counties' production will begin just as the market decline hits. On the other hand, there appear to be some market opportunities which may merit further analysis and exploitation. These include: small niche consumer markets for specialty mountain products; improved processing capability (to maintain adequate levels of product quality); basic low-risk products (such as specialty maize, Executive Summary xxiii buckwheat, and pulse crops); development of local markets for fresh mountain products; and leather, silk, and products from fast-growing trees (including bamboo and poplar). Forestry products could be of increasing importance in mountain areas if production incentives for small holders result from the ongoing forestry sector reform. Market development and the provision of support services also requires the nurturing of farmer technical associations and the private sector. Private operators, including individual farmers and small and regional enterprises, should also be encouraged to invest in agriculture-related supply, service, processing, and marketing activities. 40. Most policy-makers, local governments and line agencies agree on the need for well- coordinated agricultural development programs, and agricultural development activities are increasingly being planned as multisectoral and multiyear programs. However, few of these programs are actually implemented in this way. Technical bureaus responsible for the implementation of single-sector programs tend to concentrate their limited budgets and human resources on small-scale pilot operations (often at the village-level) and provide support to demonstration farmers with good entrepreneurial capacity. More comprehensive and collaborative interventions are also constrained by government procedures, such as bureau responsibility and incentive systems which are related to their own technical area, the allocation of local project responsibility to one lead agency, and increasingly competitive commercial interests between bureaus. To overcome these weaknesses, development projects in poor mountain townships should apply a formal project approach to poverty reduction promoting (a) multiyear financial support to a diversified set of activities (including field crops, tree crops, animal husbandry, andforestry) reflecting the needs of the absolute poor, (b) an equitable share of investment to individual poor villages, and (c) the participation of all households, from the absolute poor in remote areas to farmers with entrepreneurial capacity. Administrative village development plans are a simple and efficient tool to achieve these objectives at the village level. It is therefore recommended that village plans should form the basis of township development projects in future poverty reduction programs. 41. The 8-7 Plan has responded to mountain agriculture's need for an integrated set of interventions including land development, small-scale water conservancy, and technical packages for food crops, animal husbandry, tree crops and cash crops. This sound overall framework could be further improved through the reprioritization of investments to better reflect the needs of the absolute poor. First, investments (such as drinking water systems and small tools) which allow households to reallocate labor to more productive tasks are crucial. However, such investments do not always receive adequate priority or are not routinely included in poverty reduction program budgets. Second, the share of investment devoted to animal husbandry does not reflect its key role in all mountain land use systems. Moreover, the need of many transient poor households to recover from a loss of animal assets is well known, but credit to purchase replacement animals is often not available and not specifically funded. Third, expanding the limited area of more fertile land (through terracing and other means adapted to each region, including supplementary irrigation) appears to be one of the best ways to improve food security. Land development, however, does not receive adequate priority relative to funding for technical packages for food crops. Fourth, these technical packages for food crops have typically been limited to the main staple grains (for example, maize in the southwestern provinces). A combination of new varieties and other improvements for a wide range of crops (e.g., for potatoes) is needed to ensure long-term food security. Fifth, tree or cash crop development is often encouraged at the expense of investment into other agricultural activities, and before market and financial analysis under local conditions have demonstrated their viability for poor farmers. xxiv Executive Summary 42. Access to training and to improved technology are key to increasing the efficiency of agricultural interventions. Current investment planning procedures, however, do not allow for significant interventions in these fields, and limited budgets severely restrict the scope and quality of such activities. Several domestic and donor-assisted programs have successfully demonstrated the value of higher-quality training provided to township technicians and to larger numbers of farmers, and of strengthened information exchanges through linkages with research, teaching institutions, and enterprises. The few existing research programs in China which are specifically targeted to mountain agriculture have clearly demonstrated that returns on investment into low- cost, low-risk, simple technical innovations are much higher than direct investment into production. It is recommended that a portion of public research funds be reserved for applied agriculture research in mountain areas, and there is a strong need to establish a specific program of grants for training, access to technical information, and agricultural research in mountain areas. 43. Microcredit. Microcredit programs have had notable success in directly providing poor households with affordable credit and encouraging their self-organization, and have thus become a favored method of poverty reduction by the government and donor organizations. In 1996, LGPR decided to try to channel most poverty reduction funds directly to households, and explicitly endorsed the microcredit approach and the Grameen model. China's poverty reduction microcredit program expanded rapidly since then, and by 1998 it was reported to have reached 200 counties and have comprised a total investment of about Y 800 million. Most government funded microfinance programs have used variations of the Grameen Bank approach, which has been employed successfully in other countries, and in some donor programs in China. Initially, the program was organized by the PADOs which lent the funds directly to households. 44. Relatively high arrears rates in a number of the government's programs led to a recent switch in the microcredit program, and regulations now require that microfinance experiments using subsidized loans be undertaken by ABC, rather than the PADOs. Loan contracts are to be signed between the ABC and households. The role of the PADOs will focus on organizing households into groups and centers and facilitating loan repayment. This move addresses weaknesses in financial management and supervision, which has been one of the central problems in the government schemes. However, ABC's current institutional capacity is likely to hinder its ability to undertake Grameen type microfinance experiments as many county ABCs do not have agencies and staff at the township level. Moreover, the loan commission to be paid by ABC to the PADOs for their work in organizing groups and repayment is most likely too small to sustain their interest in microcredit. 45. Well designed microcredit programs can avoid many of the pitfalls seen in subsidized credit programs around the world, and which have plagued China's poverty loan program. In designing any future role for microcredit in the Government's poverty program, a number of points should be emphasized. First, microcredit on its own is unlikely to meet the needs of the absolute poor, and should be combined with other types of interventions in the poorest areas. Second, improvedfinancial management, monitoring, supervision and internal auditing, backed up by intensive staff training, are key to the success of any microcredit program. Lax or non- existent systems made government experiments with microcredit vulnerable to spiraling repayment problems, and financial mismanagement. Third, programs should avoid the tendency to become top-down, and may wish to experiment with devolving responsibility for implementation to grassroots organizations. It probably would be advantageous for the PADOs to play a role in monitoring and supervising microcredit, but to have the actual programs managed by organizations outside the government structure. At a minimum, staff should be hired Executive Summary xxv by open recruitment and not be appointed by county or other officials. One positive aspect of the Grameen model is that its participatory methods promote initiative among the poor by encouraging them to form groups, choose group leaders, and decide their own investments. If programs side step this "bottom-up" approach, this aspect is lost, and targeting and repayment can also suffer. Fourth, government programs have tended to charge rates of interest well below the level that would allow them to cover operating costs. This interest rate policy should be reevaluated since it can lead to leakage of funds to the non-poor, and also threatens the sustainability of the programs. Finally, the current group based models can be quite costly, both to the poor and in terms of administrative costs (for example, through frequent group meetings and repayment). Impact assessment should be undertaken to better understand the costs and benefits to the poor of existing programs, and to assess whether variations on the current models might be better suited to some areas. 46. In the longer term, mechanisms to provide credit for the poor could take many forms. Developing savings services could be even more important for poor farners than credit services, but no savings should be organized until prudent regulation and supervision has been established. Reforming existing financial institutions in poor areas may be the most efficient way to reach large numbers of poor and near poor, but this should not preclude a future role for informal financial organizations that now lend to the poor. If they are able to meet appropriate performance and regulatory standards, some of the current temporary microcredit program offices could possibly be converted into more permanent institutions which provide credit and savings services to the poor. In addition, small, grassroots organizations have the potential to play a valuable role in piloting innovative techniques, and their strong outreach is invaluable. 47. TVE Development. Beginning in 1989, poor county governments have used a large of portion of their available poverty reduction program funding to support the growth of county and township run enterprises. In 1992 and 1993, for example, about half of all poverty reduction subsidized loans were lent to industrial enterprises. Unfortunately, beginning in the mid-1990s, many TVEs in poor areas have suffered consecutive annual financial losses, many now have a negative net worth, and a high proportion of their total assets are accounts receivable and inventories. It is evident that many of the rural enterprises supported in the past with poverty reduction funding have also been loss making firms, often with minimal or no poverty reduction impact. While the number of such firms is not known, it appears that poverty reduction program funding for poor county TVE development may have even led to a net decline in revenues in some poor counties. 48. Recognizing the shortcomings of providing direct financial support, the government recently decided to severely curtail poverty reduction program funding for poor county TVE development. Instead, the more appropriate role for local governments is to establish a more favorable environment for poor county TVE development. The experience of those TVEs in poor areas which have better survived the current economic turbulence helps clarify what local government can do to facilitate TVE development. Relatively small enterprises with a more clearly defined ownership structure, operations based principally on local comparative advantage (such as cheap labor and local raw materials), and relatively large enterprises combining local resources with technology and market access provided by enterprises from more developed areas, appear to have been most successful in weathering the current economic challenges. The government's decision to curtail poverty reduction program funding for poor county TVE development was appropriate, and it is recommended that local governments instead support TVE development through (a) reforming TVE ownership and management, (b) allowing those enterprises which have negative equity and which have suffered from financial losses for a xxvi Executive Summary number of years to go bankrupt, and (c) encouraging local enterprises, which have operations consistent with local comparative advantage, to establish joint ventures with enterprises from more developed areas. Past experience indicates that local governments in poor areas are not well-equipped to successfully establish and manage enterprises by themselves. These local governments should therefore instead focus on simplifying the procedures for the establishment of private and collective enterprises, reducing the taxes and other extra burdens on enterprises, improving local infrastructure, and providing more training for local accountants, auditors and technicians. Chapter 1. Trends in Rural Poverty in the l990s I 1. TRENDS IN RURAL POVERTY IN THE 1990s 1.1 Summary. China has experienced a tremendous decline in poverty over the last twenty years. Based on the government's austere rural poverty line, official estimates indicate that poverty declined from more than 30 percent of the rural population in 1978 to less than 5 percent by end-1998. For the purposes of cross-country comparisons, The World Bank has developed an international poverty standard. Estimates based on that international poverty line document an equally steep decline in the incidence of poverty in China. However, since the international standard is somewhat less severe than China's official poverty line, it indicates greater numbers of poor in all years, and that by end- 1998 a much larger share of the rural population - about 11.5 percent or some 106 million people -remained in poverty. While China's austere poverty line was a useful standard when the incidence of extreme poverty was greater, the international standard has now become a more appropriate measure to gauge the extent of poverty and guide the government's poverty reduction program in the next century. 1.2 Available evidence indicates that there has been an increasing concentration of remaining poverty in China's western provinces during the 1990s. The proportion of China's poor residing in the western provinces increased from less than half in the late-i 980s to more than two thirds by the mid-1990s. Most of these poor are concentrated in remote townships and villages, often in mountainous, low rainfall, or other lands with limited potential for even subsistence levels of production. It also appears that there has been a deepening of poverty during the 1990s, with the very poorest slipping further below the poverty line. The concentration of large numbers of extremely poor people in these areas suggests that geographic targeting of poverty reduction assistance to these households and communities should be intensified. A. CONTINUED REDUCTION OF RURAL POVERTY 1.3 China is widely recognized for its achievements in reducing absolute poverty since the adoption of a broad program of rural economic reforms beginning in 1978. These reforms, including the adoption of the production responsibility system, the dismantling of the commune systerni, agricultural product price increases and market liberalization, have been associated with dramatic rural economic growth. Broad participation in this growth, together with a well-funded national poverty reduction program, have brought about a tremendous reduction in rural absolute poverty during the past twenty years. Official government estimates indicate that the number of rural poor declined from roughly 260 million in 1978 to 42 million in 1998, or from one-third to about one-twentieth of total rural population. As summarized in Table 1.1 and Figure 1.1, this decline continued through the 1990s. 2 Chapter 1. Trends in Rural Poverty in the 1990s Table 1.1: Number of Rural Poor, 1990-98 International Official Government Estimates Standard ($1/day) Share of Share of Poverty Number of Rural Squared Number of Rural Line /a Rural Poor Population Poverty Rural Poor Population Year (Current Y) (million) (percent) Gap Index (million) (percent) 1990 Y300 85 9.5% 0.43 280 31.3% 1991 Y304 94 10.4% 1.15 287 31.7% 1992 Y317 80 8.8% 0.85 274 30.1% 1993 Y350 75 8.2% 1.36 266 29.1% 1994 Y440 70 7.6% 1.38 237 25.9% 1995 Y530 65 7.1% 1.00 200 21.8% 1996 Y580 58 6.3% 0.58 138 15.0% 1997 Y640 50 5.4% 0.85 124 13.5% 1998 Y635 42 4.6% 0.67 106 11.5% /a The official government poverty line is the equivalent of $0.66 per day. Source: State Statistical Bureau (SSB) for the official govemment estimates, and The World Bank for the international standard figures. Both sets of estimates are derived from per capita income data from the SSB's annual sample survey of rural households (see Annex I Tables I to 5). 1.4 These official estimates are based on the government's austere poverty line Figure 1.1: Number of Rural Poor, equivalent to $0.66 per day (in constant 1990-97 (Source: Table 1.1) 1985 purchasing power parity dollars). The government's poverty line reflects national 300 | perceptions of the minimal standard of 0 250 - __ - living, and those people with incomes less , 200 - than this amount would likely be unable to . 150 i -_ ____ secure subsistence levels of food and X clothing (in Chinese, this is known as the 0 |, "bu wenbao" poverty line). The World b X < C Bank has developed an international t , poverty standard of $1.00 per day (in 1985 Year purchasing power parity dollars) for cross- Yea country comparisons of poverty. Estimates |Govement Estmates based on this somewhat less severe poverty line indicate substantially greater numbers -.- Intemational Standard ($1/day) of absolute poor in China in all years, but I _ do confirm the continuing remarkable decline in poverty during the 1990S.6 Even L- 6 The significance of the difference between the official government estimates of poverty and those based on the international standard is discussed in The World Bank (1996). As shown in Annex 1, the sharp decline in rural poverty during the 1990s is observed across a variety of poverty lines ranging between $0.5 to $2.5 per day. Chapter 1. Trends in Rural Poverty in the 1990s 3 these estimates are considered to be conservative, since they rely on income, not expenditure, data. The government's austere poverty line was a useful standard when the incidence of extreme poverty was greater, and helped to target available poverty reduction funding to those most in need. However, the government's poverty line may no longer be such an effective tool in identifying the potential beneficiaries of national poverty reduction efforts since such a small share of the population now have income levels below it, and large numbers of transient poor have income levels just slightly greater than it. In addition, those with incomes consistently below the poverty line suffer extreme hardship and deprivation.S Consequently, China should consider whether the international standard may now be a more appropriate measure to gauge the extent of poverty and guide its poverty reduction program. 1.5 Although the incidence of rural poverty declined significantly, available evidence suggests that the severity of the remaining poverty worsened somewhat during much of the 1990s. As shown in Table 1.1, the squared poverty gap index increased during the 1990s, and in 1998 remained considerably greater than in 1990.9 In addition, it appears that rural poverty is increasingly concentrated in the western provinces, and that the depth of poverty is greatest in these western provinces. These trends in the location and severity of poverty in the l990s are discussed below. B. LOCATION OF THE POOR 1.6 Regional Concentration of the Poor. The reduction in rural poverty has been greatest in China's coastal and central regions where rural economic growth has been greatest. Many of the rural poor in 1978 resided in less remote and less hilly areas in the coastal and central regions, where increased application of fertilizer, irrigation, better seed and other modem inputs could bring about rapid productivity gains, and so were better able to participate in the rapid agricultural growth of the reform period. Most of the residual poor have remained trapped in more remote upland areas (particularly in the western provinces) where agricultural productivity gains have proven far more problematic. 1.7 Available evidence shows that this trend has continued in the 1990s, with an increasing share of the rural poor now concentrated in China's western provinces. As shown in Table 1.2, the number of poor in the western provinces (including both the northwestern and southwestern provinces) increased from less than half of all of China's rural poor in each year 1988, 1989 and 1991 to more than two thirds (about 70 percent) in 1996. This change in the distribution of remaining rural poverty in the 1990s was driven by the relatively sharp declines in the incidence 7 There are two reasons why this report's estimates of the incidence of poverty should be considered to be conservative lower bounds of the true number of poor. First, in order to be consistent with Government estimates and methods, all the estimates of the incidence of poverty presented in this report are based on the distribution of per capita income (from SSB's annual sample survey of rural households). However, per capita expenditure (consumption) data are considered to provide a superior basis for assessing the incidence of poverty, and should be used when available. As shown in Annex 1, estimates based on available expenditure data indicate much greater numbers of absolute poor in China. Second, there remains some uncertainty about the purchasing power parity (PPP) estimates for China. The price data underlying these estimates are limited, and some evidence suggests that the PPP dollar per day estimates may therefore significantly understate the incidence of poverty in China (The World Bank, 1994). For example, about half of the children of poor households are at least mildly malnourished (The World Bank, 1992). 9 The squared poverty gap index is a measure of the income that would be necessary to bring a poor person up to the poverty line. A greater index value indicates greater severity of poverty. 4 Chapter 1. Trends in Rural Poverty in the 1990s of poverty in the eastern and central provinces, from 7 and 14 percent respectively in 1988 to I and 3 percent in 1996. The incidence of poverty also declined (but by a much lesser margin) in the westem provinces, from 22 to 13 percent of the rural population in the western provinces in 1989 and 1996 respectively. This increasing concentration of poverty in the western provinces during the 1990s is observed using both the official government and the dollar per day poverty lines, and is illustrated by Maps 1 and 2 for 1991 and 1996.10 Table 1.2: Incidence of Rural Poverty by Region and Province, 1988,1989,1991 and 1996 1996 Poverty Incidence Rural (Number of Poor/Rural Population) Number of Rural Poor (Headcount) Share of National Total Population 1988 1989 1991 1996 1988 1989 1991 1996 1988 1989 1991 1996 (million) (%) (0/0) (%/6) (%) (million) (million) (million) (million) (%) (%) (%) (%) National 919.4 13.9 12.3 10.4 6.3 120 102 94 58 100 100 100 100 North 83.6 15.2 13.2 13.4 4.6 11.7 10.3 10.2 3.8 9.8 11.1 10.8 7.6 Beijing 3.7 0.5 0.2 0.1 0.8 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1 Tianjin 3.9 1.8 0.4 0.3 0.3 0.1 0.0 0.0 0.0 0.1 0.0 0.0 0.0 Hebei 53.2 14.0 13.0 13.4 3.9 6.9 6.5 6.7 2.1 5.8 7.1 7.1 4.1 Shanxi 22.8 22.1 17.4 17.0 7.5 4.7 3.8 3.5 1.7 3.9 4.1 3.7 3.4 Northeast 55.0 10.2 12.8 9.1 4.7 5.5 6.9 4.8 2.5 4.6 7.5 5.1 5.1 Liaoning 22.2 8.0 8.0 4.4 2.9 1.8 1.8 0.9 0.6 1.5 1.9 0.9 1.3 Jilin 14.4 8.8 12.2 8.7 4.7 1.3 1.8 1.2 0.7 1.1 1.9 1.3 1.3 Heilongjiang 18.4 13.6 18.3 14.3 6.7 2.5 3.4 2.7 1.2 2.1 3.6 2.8 2.5 East 272.0 7.0 4.9 7.7 1.2 17.9 12.7 19.1 3.2 14.9 13.7 20.3 6.4 Shanghai 3.8 0.3 0.0 0.1 0.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Jiangsu 53.1 4.5 3.4 4.1 0.1 2.4 1.8 2.2 0.1 2.0 1.9 2.3 0.1 Zhejiang 35.9 2.8 2.0 3.6 0.1 1.0 0.7 1.0 0.0 0.8 0.8 1.1 0.1 Anhui 49.8 10.3 7.7 27.8 2.7 4.7 3.6 12.9 1.4 3.9 3.9 13.7 2.7 Fujian 26.4 3.9 1.8 1.4 0.5 0.9 0.4 0.3 0.1 0.8 0.5 0.3 0.3 Jiangxi 32.0 8.5 5.0 0.3 0.7 2.5 1.5 0.1 0.2 2.1 1.6 0.1 0.5 Shandong 71.0 9.4 6.8 4.1 1.9 6.4 4.7 2.5 1.4 5.3 5.1 2.7 2.7 Central 233.3 13.6 8.7 10.3 2.6 27.8 18.0 21.2 5.7 23.1 19.4 22.5 11.3 Henan 77.5 25.0 16.5 21.7 4.3 17.6 11.8 15.8 3.3 14.6 12.7 16.8 6.6 Hubei 40.3 11.0 6.0 9.0 2.7 4.4 2.4 3.5 1.1 3.6 2.6 3.7 2.1 Hunan 53.2 7.9 6.2 3.1 1.5 4.0 3.2 1.6 0.8 3.4 3.5 1.7 1.6 Guangdong 57.5 2.7 0.9 0.3 0.2 1.3 0.5 0.1 0.1 1.1 0.5 0.1 0.2 Hainan 4.8 12.1 3.3 3.8 8.2 0.5 0.1 0.2 0.4 0.4 0.2 0.2 0.8 Southwest 198.3 20.5 14.2 13.2 10.5 38.4 27.0 24.1 20.7 31.9 29.2 25.6 41.2 Chongqing 6.6 1.6 3.2 Sichuan 93.9 16.7 11.2 11.2 7.0 15.3 10.4 9.6 4.9 12.8 11.2 10.2 9.7 Guizhou 29.8 23.0 17.8 23.5 12.8 6.3 5.0 6.2 3.8 5.3 5.4 6.6 7.6 Yunnan 33.5 23.8 19.0 17.3 22.9 7.4 6.0 5.5 7.7 6.2 6.5 5.9 15.3 Tibet 2.1 32.3 NA 11.6 10.1 0.6 NA 0.2 0.2 0.5 0.2 0.4 Guangxi 39.0 24.1 15.4 7.0 6.4 8.7 5.6 2.5 2.5 7.2 6.1 2.7 5.0 Northwest 77.2 26.2 24.0 19.8 18.6 18.9 17.6 14.7 14.3 15.7 19.0 15.7 28.4 InnerMongolia 14.2 17.3 23.5 13.7 9.3 2.4 3.3 1.9 1.3 2.0 3.6 2.0 2.6 Shaanxi 27.5 24.9 20.3 18.0 17.5 6.4 5.3 4.7 4.8 5.3 5.8 5.0 9.6 Gansu 19.7 38.4 34.2 29.0 22.7 6.8 6.2 5.1 4.5 5.7 6.7 5.4 8.9 Qinghai 3.3 22.4 23.7 17.8 17.7 0.7 0.7 0.6 0.6 0.5 0.8 0.6 1.2 Ningxia 3.7 24.7 18.9 22.7 18.5 0.8 0.6 0.8 0.7 0.7 0.7 0.8 1.4 Xinjiang 8.8 22.3 18.7 16.3 27.4 1.7 1.5 1.7 2.4 1.4 1.6 1.8 4.8 Source: SSB's annual sample survey of rural households income (see Annex I Tables 6-11). 0 Differences in the cost of living varies across provinces, and may bias these interprovincial poverty comparisons (Chen and Ravallion, 1996). Chapter 1. Trends in Rural Poverty in the 1990s 5 1.8 It also appears that the depth of poverty in the western provinces is now more severe than in the rest of China. As summarized below (and shown in detail in Annex 1 Table 12), the squared poverty gap was much greater in the northwestern provinces than it was in all other regions in 1991 and, particularly, in 1996: Regional Depth of Poverty: Squared Poverty Gap, 1991 and 1996 1991 1996 National 1.19 0.59 North 1.32 0.27 Northeast 1.36 0.52 East 0.76 0.03 Central 1.05 0.17 Southwest 0.77 0.49 Northwest 1.75 1.45 The squared poverty gap declined sharply in the north, northeast, east and central regions during 1991-96. On the other hand, the depth of poverty in the southwestern provinces was less than most of the other regions in 1991, but had declined by a much smaller margin by 1996. 1.9 Nationally-Designated Poor Counties. In 1986, using Ministry of Agriculture (MOA) county-level rural income data, the State Council's Leading Group for Poverty Reduction (LGPR) developed a national roster of 331 poor counties eligible for development assistance. A county was identified as poor if its 1985 average rural per capita income fell below poverty lines of Y 300, Y 200, or Y 150, depending on other locational or political factors. Average per capita grain production of less than 200 kg was also adopted as a second key indicator of poverty.11 In addition to the national roster, 368 provincially-designated poor counties, eligible for provincial funding, were selected on the basis of poverty lines determined by the provinces. Province- specific poverty lines ranged from Y 150 in Yunnan to Y 400 in Jiangsu. In addition to drawing up a roster of poor counties, most provinces have identified townships in otherwise well-off counties as eligible for special assistance. As evidenced by the wide difference in qualifying income levels (Y 150 to Y 400), the provincial rosters included a number of counties that are relatively poor by provincial standards but that would not qualify as absolutely poor by the national minimum standard. That provinces created their own special programs testifies to a concern with relative poverty and, in many instances, to provincial budgetary strength. On the other hand, the provincial rosters did, in some cases, compensate for the incorrect exclusion of counties with average per capita income levels below the national minimum standard. The Yunnan provincial roster, for example, included all the counties that should have received central government support for absolute poverty reduction (that is, 15 counties with average per capita income of between Y120 and Y150 in 1985). 1.10 While the LGPR list of poor counties did capture a large share of China's most severe rural poverty, there has been considerable criticism of the "poor county" approach to poverty Eighteen indicators were used to determine counties' poverty status (LGPR, 1989). In addition to per capita income and grain production, these indicators include (a) access to safe drinking water, road transport, and other basic infrastructure, (b) total and per capita arable land, (c) several demographic factors, and (d) other measures of economic production and fiscal strength. 6 Chapter 1. Trends in Rural Poverty in the 1990s monitoring and targeting.'2 The determnination of poor counties, and the subsequent allocation of poverty alleviation funding, was necessarily political. Counties with favorable political credentials and strong supporters-most notably the old revolutionary base areas-were included under the roster of poor counties despite having per capita income levels twice the level deemed to represent subsistence. The inclusion of politically favored counties forced the exclusion of many counties with per capita income levels well below the poverty line. Yunnan, for example, was forced to adopt a special cutoff of Y120, rather than the nationally mandated level of Y1 50, in order to reduce its number of poor counties. 1.11 Partly in response to these shortcomings, LGPR established a new list in 1993 of 592 nationally-designated poor counties. (The 592 nationally-designated poor counties are shown in Map 3.) The new list included a large number of previously provincially-designated poor counties and other poor counties which had been otherwise excluded from the original list. The new list of nationally-designated poor counties certainly does a good job of capturing most of China's poor counties. For example, of the 118 counties with average rural income levels of Y500 per capita or less in 1995, 90 percent (112) are included in the new list. Park et at. (1999) have noted that the new list includes a larger number of poor counties in the very poor southwestern provinces of Yunnan and Guizhou, while the much higher income coastal provinces of Fujian, Guangdong, Shandong and Zhejiang were net losers. However, while the new list better captures the poorest counties, it also nearly doubled the number of counties receiving central government funding for poverty reduction activities. This has contributed to spreading available poverty reduction program funding over a much greater population. The total rural population of the 593 nationally designated poor counties numbers at least 200 million. Since government estimates indicate that in 1997 roughly half (that is, about 27 million) of China's remaining 50 million absolute poor resided in the nationally designated poor counties, this means that only one in seven of these counties' rural inhabitants were absolutely poor. As discussed below in para. 3.19, it is believed that available poverty reduction program funding is distributed more-or-less equally across the entire rural population of these counties, so there is a very substantial leakage of benefits to the nonpoor. 1.12 Mountain Area Poverty. In addition to this concentration of poverty in the western provinces, it is believed that the majority of China's poor reside in mountainous counties and townships.'3 A number of observers have concluded that China's poverty problem mainly occurs in poor mountain regions (ICIMOD 1994, MFO 1995, Han 1998, Wang 1998), and China's poverty reduction programs have long focused on mountain areas. In 1986, LGPR commissioned a Chinese Academy of Sciences (CAS) study (Jiang, 1989) which identified 21 core poor zones covering 741 counties (see para. 4.2 and Map 4). The CAS study found that at least 90 percent of the rural population of these poverty zones resided in mountain counties, and LGPR has focussed government support for poverty reduction activities on 18 of the CAS poverty zones. All of the 12 Park et al (1999), for example, have noted that less than one third of the initial selection of 258 nationally- designated poor counties met the lowest standard of poverty (that is, only 83 of the designated counties had average rural income levels of less than Y 150 per capita). 3 Mountain counties are defined as those with elevations of at least 500 meters, sloped land comprising at least 60 percent of the total land area, the county seat located in a mountainous area, and a majority of the population living in mountainous areas (MOAb, 1997). Some 910, or about 40 percent, of China's counties are considered to be mountain counties. Chapter 1. Trends in Rural Poverty in the 1990s 7 18 LGPR zones are mountainous, and two thirds of the 592 nationally-designated poor counties are mountain counties (Annex 3 Table 1).14 1.13 Available information and field visits in many poor areas indicate considerable disparity at the county level, and confirm that a very high proportion of the rural poor live in mountain townships. These poor mountain townships are located in the nationally- and provincially- designated poor counties, and in some other counties not specifically designated as poor. Administrative boundaries in mountain counties typically delineate better-off townships around the county seat and in wide valley floors ("baqu'), with poorer mountain townships in higher elevation outlying areas ("shanqu").15 In Yunnan, for example, the provincial government selected 506 key townships for poverty reduction programs beginning in 1995. Almost all of these poor townships are located in mountainous, minority, and border areas. Four fifths of the poor townships are in Yunnan's 73 nationally-designated poor counties, and the others are located in the 17 provincially-designated poor counties and in some non-poor counties. (Simao County, for example, is not a designated poor county, but has two poor townships in the karst area with water availability problems.) In the Jiuwan Dashan mountains in Northeast Guangxi, Longsheng, Ziyuan and Xingan are three neighboring counties. Longsheng is a nationally-designated poor county, Ziyuan is a provincially-designated poor county, and Xingan is not designated as poor. The Longsheng County government reports that 4 of its 7 townships are poor, Ziyuan reports that 4 of its 8 townships are poor, and Xingan reports that 6 of its 11 townships are poor. All of the 14 poor townships are located at higher-elevations and have high proportions of Yao or Miao ethnic minority people. Similarly, Baojing County in Hunan has identified 5 poor townships out of its total of 24 townships. Four of the poor townships are in the southern part of the county on karst terrain, and the other is in the higher-elevation part of the northern hill range. The county has further listed 69 poor villages, of which 43 are within these 5 poor townships. 1.14 Another example is Puding County, in the karst mountains of Guizhou, which is divided in half by the Sanfen River. In Machang, Longchang, Chengguan, Baiyan, Huachu, and Maguan townships to the south of the river, there are more basins and agricultural production is based on irrigated crops. The poor townships of Jichangpo, Pingshang, Bulang, Houchang, and Maodong to the north of the river are typical core poor mountain townships. As summarized in Table 1.3, minority peoples comprise one third of the rural population of the northern townships, but only one eighth in the southern townships. Less than half of administrative villages in the northern townships were accessible by road in 1992, while more than three quarters were in the southern townships. Average per capita income was about 40 percent less in the northern townships, and per capita grain production was 10 percent less in the northern poor townships than in the southern townships. Moreover, corn comprised nearly half of the grain produced in the northern townships, but only about a quarter of that in the southern townships. The northem poor townships had on average only 0.24 mu per capita of irrigated land, or just slightly more than half that of the southern townships. 14 LGPR publications often mention that mountain counties comprise 497 out of 592 of national-poor counties. However, as shown in Annex 3 Table I of this report, this figure includes both mountain and hilly counties. Two types of poor townships occur: (a) "core" poor townships in which the absolute poor form the majority of the rural population, and (b) townships which have a lower proportion of absolute poor who are mostly found at higher elevations and in other difficult environments within the townships. The core poor townships are primarily found in the karst, loess uplands, and pastoral areas, while the other type of townships occur in all mountain types (see para. 4.2). 8 Chapter 1. Trends in Rural Poverty in the 1990s Table 1.3: Puding County (Guizhou) Township Disparity, 1992 Share of Administrative Per Capita Income and Output Agricultural Population Villages With Arable Land Per Capita Agricultural Total Minorities Road Access Total Irrigated Rainfed Income Grain Output Value (number) (%) (%) (mu) (mu) (mu) (Y) (kg) (Y) Southern Machang 33148 10.9 73.7 0.74 0.33 0.41 369 258 319 Longchang 27619 10.0 68.8 0.88 0.25 0.63 331 191 283 Chengguan 47967 23.6 87.8 0.71 0.39 0.32 419 267 360 Baiyan 27620 17.0 76.0 0.77 0.46 0.31 391 260 333 Huachu 44228 3.8 69.8 0.69 0.33 0.36 371 223 325 Maguan 44064 12.7 83.9 0.70 0.52 0.18 373 219 323 Northern Jichangpo 28300 36.7 70.8 0.8 0.25 0.55 248 216 259 Pingshang 22800 32.0 40.0 1.06 0.20 0.86 235 248 280 Bulang 19500 34.9 50.0 1.13 0.23 0.90 222 222 294 Houchang 19000 37.9 41.2 0.88 0.20 0.68 224 238 274 Maodong 25600 26.2 34.6 0.82 0.30 0.52 230 153 267 All Puding 339846 20.0 67.2 0.81 0.34 0.47 329 229 309 Southem Towns 224646 13.2 77.1 0.76 0.40 0.36 379 237 327 Northern Towns 115200 33.3 47.2 0.92 0.24 0.68 233 213 273 Source: Guizhou Poor Area Development Office. C. OTHER CHARACTERISTICS OF THE POOR 1.15 Since virtually all of China's rural population received land use rights as part of the implementation of the production responsibility system during the early-1980s, there are few if any landless laborers. Although these poor have land use rights, in most cases the land itself is of such low quality that it is not possible to achieve subsistence levels of crop production. Consequently, most poor consume grain and other subsistence foods beyond their own production levels, and are negatively affected by price increases for these products. The poorest households are typically those further disadvantaged by high dependency ratios, ill health and other difficulties. Minority peoples are known to represent a highly disproportionate share of the rural poor. Available evidence does not suggest that women are greatly overrepresented among the poor, though poverty certainly does exacerbate society-wide problems of lower rates of female participation in education, higher relative female infant mortality rates, and higher rates of maternal mortality. 1.16 The educational and health status of China's remaining absolute poor is deplorable. As many as half of the boys in many of China's poorest villages and, particularly in some minority areas, nearly all of the girls do not attend school and will not achieve literacy. The infant mortality rate and maternal mortality ratio in very poor counties -- which exceed 10 percent and 0.3 percent respectively -- are at least 50 to 100 percent greater than the national average, and are much greater yet in the poorest townships and villages. Incidence of several infectious and endemic diseases, including tuberculosis and iodine deficiency disorders, is concentrated in poor and remote areas. Roughly half of children in households at or below the absolute poverty line are at least mildly malnourished (stunted), and iron, vitamin A, and other micronutrient deficiencies Chapter 1. Trends in Rural Poverty in the 1990s 9 remain a severe problem among the poor. Up to 90 percent of poor children suffer chronic worn infections. 1.17 Distribution of Ethnic Minority Poverty. Minority peoples16 are known to represent a highly disproportionate share of the rural poor. Ethnic minority groups make up less than 9 percent of the total population, but are believed to account for about 40 percent of the remaining absolute poor in China, and often live in the deepest poverty.'7 Minority autonomous counties also account for more than 40 percent of the nationally-designated poor counties. 1.18 The single most obvious contributor to minority poverty is geography. Minorities are concentrated in the northwest and southwest regions and within these regions they tend to occupy the most remote mountain areas. China historically shows a dynamic of poorer uplands and better-off lowlands, where minorities are usually concentrated in the highland economy. For example in Guangxi, the Han tend to be concentrated in the coastal region and more productive southern part of the province, sinicized Zhuang occupy the less fertile northeast, less sinicized Zhuang dominate in the karst regions of the northwest, and Yao occupy the most remote areas of the karst region. Guizhou shows a similar pattern, with the Miao, or Buyi and Miao, occupying the most remote, mountainous areas.18 1.19 Minorities who live in these remote areas tend to be the poorest of the poor. Work sponsored by the State Ethnic Affairs Commission (SEAC) concluded that in the mid-1990s minority autonomous counties and regions accounted for three quarters of all nationally- designated poor counties with per capita incomes under Y 400, and more than four fifths of the nationally-designated poor counties with incomes under Y 300. Provincial accounts support this finding. Sichuan, for example, reports that its poorest counties are all in minority areas. The average 1997 per capita income in Sichuan's poor counties was Y 1,032, but ranged between Y 720 and Y 860 in the three poorest prefectures, all of them minority autonomous regions. Simao Prefecture in Yunnan, whose population is 61 percent minority, estimates that 90 percent of its remaining poor are minorities. 1.20 Information from The World Bank-supported project in Gansu's Hexi Corridor region also documents a higher percentage of minorities in the poorer areas: 16 Membership in China's 56 ethnic groups (including the majority Han) was unambiguously defined in the 1950s, and is in principle based on shared language, religion, cultural traditions, and history. The minority groups make up less than 9 percent of the total population, although they constitute almost the entire population in some localities. 17 LGPR collects systematic data on the number of absolute poor in China, but does not classify them by ethnic group. The State Ethnic Affairs Commission has such data, but only for autonomous counties and regions. Many autonomous counties or regions, however, actually have Han majorities, and many minorities do not live in autonomous counties or regions. SEAC estimates that 40 percent of the poor are minorities. s There are exceptions. In northern Guangdong and southern Fujian. for instance, Hakka (a Han dialect group) often concentrate in the uplands. Han who have arrived late in an area often occupy the very worst land and are also extremely poor. IO Chapter 1. Trends in Rural Poverty in the 1990s Concentration of Minority Peoples and Poverty in Gansu's Hexi Corridor (Figures are for rural households in 1995) Per Capita Per Capita Share Below Minorities' Grain (kg) Income (Y) Poverty Line Share China 500 922 9% 8% Gansu 390 563 22% 8% 41 Poor Counties 288 449 46% NA 11 Project Counties 221 348 49% 22% 95 Project Townships 173 287 59% 25% These figures highlight a situation of extreme poverty among minority peoples concentrated in poor counties and townships. 1.21 In these remote, mountainous areas, extremely difficult agroecological conditions have been exacerbated by periods of rapid deforestation and population growth. Until recently, poverty programs rarely reached into these areas, and this has slowed progress. Moreover, the lack of outside contacts for minorities, who often do not speak Han Chinese languages, means that isolated minorities lack information about off-farm jobs, markets, and investment opportunities, as well as connections that would enable them to make use of these opportunities. One consequence is that minorities have benefited less from labor mobility, an important source of income for poor families for most of the decade, than have Han Chinese. Participation in labor mobility among minority women appears to have been particularly low. 1.22 Disability and Poverty. Disabled people and households with disabled members represent a large, and rising, share of the poor.19 The chronic shortage of labor power that often afflicts families with disabled members is a major cause of the absolute poverty remaining in China. 1.23 The China Disabled Persons' Federation (CDPF) estimates that of the 60 million total disabled population in 1997, 17 million were absolute poor, and 12 million were rural absolute poor. Thus, these figures suggest that the disabled account for about one quarter of all rural poverty. There is a higher rate of disability among the absolute poor than among the entire population. The disabled account for 5 percent of the national population, but are estimated to represent close to 20 percent of the poor living in the nationally-designated poor counties.' In poor areas, disability correlates with absolute poverty through a vicious circle: family labor shortages due to disability increase the odds of falling into poverty, and the inadequacy of effective public health measures and medical care in poor areas increases the chances of becoming disabled. 19 China distinguishes five categories of disability: physical disability, mental illness, blindness, deafness, and mental disability. It includes disabled who are unable to work, and those who, as is often the case, lack the opportunity to work. 20 This estimate is based on CDPF's own survey of 1997-98. LGPR believes the number of disabled poor may be much lower. Counts of the disabled often understate their number through sampling errors and by emphasizing physical over mental disability. 21 This figure is based on the CDPF estimate that about 5 million of the disabled poor live in those counties, and SSB's estimate that about 27 million of the remaining 50 million absolute poor resided in those counties in 1997. Chapter 1. Trends in Rural Poverty in the 1990s I1 1.24 However, unlike the general pattern of poverty in China over the last decade, the disabled poor are not found primarily in geographically isolated areas. CDPF estimates that about two thirds of the rural disabled poor live outside the 592 nationally-designated poor counties, and are thus not reached by most existing poverty relief programs. In relatively wealthy provinces disability appears to be the single most important cause of poverty. In wealthy Jiangsu Province, for example, the disabled reportedly comprise more than 60 percent of the poor. 1.25 Gender and Poverty. A recent participatory poverty assessment (PPA) in a number of poor villages has helped document the ways in which poverty disadvantages women either in poor regions or poor households, and the repercussions poverty reduction projects have for women, their organization and gender relations (see Beynon, Zheng et al, 2000). First, the PPA team made a number of observations which draw attention to the heavy demands on female labor, the quite rigid gender divisions of labor, and the benefits of improving the quality of female labor. It was observed that the amount and quality of family labor is an important determinant of a household's income levels and avoidance and alleviation of poverty, and that women labored extensively and for long hours to undertake agricultural field work and other economic and domestic activities. In busy seasons women had little time to rest and relax, and often work from very early in the morning until midnight. The demands for and on female labor are greater in (a) female-headed or female-managed households due to the illness or absence of adult males due to death or labor mobility, (b) particular phases of the domestic life-cycle marked by larger numbers of dependent children or aged, and (c) where there are few labor-saving devices or infrastructure necessitating long hours fetching and carrying water and other resources or goods due to the absence of local water supply, roads and electricity. 1.26 The marked gender divisions of labor apparent in all the villages had important consequences for the role of women in the maintenance and development of domestic economies and activities. In single-headed households without women, domestic maintenance and economic activities such as livestock raising are likely to be left undone or underdeveloped. In the absence of adult males, households are likely to be poorer and women are likely to add to their labor burdens by undertaking ploughing and marketing in addition to their other activities. In the absence of either male or female adult members, girls rather than boys are likely to be taken out of school to undertake the activities of the absent adults. The PPA report confirmed the direct correlation between the quality of labor and poverty levels. Poor households were more likely to have male and female adult members who had (a) less than full labor power due to illness or disability, or (b) had little or no education, technical knowledge or productive expertise. 1.27 Second, the PPA report clarified the importance of improving local infrastructure in water storage, road communication and electricity for reducing demands on female and male labor, for poverty reduction, and for quality of life. The introduction of water supply systems resulted in (a) a reduction in the demands on women's labor for fetching and carrying water by as much as by four hours per day, (b) easier and increased land and agricultural productivity, and (c) improved public and personal hygiene. The construction of roads permitted animal, cart and tractor transport and better communications which resulted in (a) a reduction in time spent by women and men in carrying resources and goods by foot to and from the village, and (b) increased female and male access to services including schools, hospitals and markets. The introduction of electricity allowed for (a) an extension of the labor time available for women's domestic and income-generating activities, and (b) the introduction of labor-saving devices to reduce the labor demands on females and males. 12 Chapter 1. Trends in Rural Poverty in the 1990s 1.28 Third, the PPA team observed the lack of specialized support for female income generation activities and women's demand for information and new technical knowledge. The importance of new technical knowledge and training for women was particularly evident in activities such as domestic livestock raising, which is typically a female responsibility and where the high incidence of animal death has a number of direct repercussions for family income and poverty alleviation. Gender disparities in access to agricultural and technical services were apparent in the villages where two-thirds or more of farmers taking part in extension services were men. These gender disparities, which led to women falling behind the male members of the household in skill development and in income generation, were the result of a number of factors: (a) the location of training services in townships is likely to hamper women's attendance given the distance and time taken up by their domestic and agricultural responsibilities, and (b) the exclusion of women emphasized gender divisions of labor and familial attitudes and the premise that "men earn the money and women take care of the fields and household" (despite evidence to the contrary). It was also observed that there was little female migration from any of the villages visited by the PPA team. Young women were either excluded or excluded themselves because (a) of their domestic responsibilities in the household for child and elderly care, or (b) they were fearful of leaving their villages for distant places and finding safe and supportive employment environments. The fact that several young women had returned to the village because they had not felt supported and safe or had been cheated of their earnings by their employers pointed to the importance of providing supportive services for young women in leaving their villages, obtaining employment, and monitoring work and living conditions. 1.29 Fourth, it was confirmed that education and health outcomes are typically given a low priority in poverty reduction projects, and that the unmet needs of women and girls are most apparent and greatest in education and health. In education, the PPA team noted the high levels of awareness in poor villages of the correlation between poverty and the lack of education and familial importance accorded to children's education in the hope of reducing poverty. However, the burden of school fees, one of the major expenses of families causing hardship, constituted a major constraint to raising educational levels and to reducing poverty. In investigating education levels and qualifications in poor villages, it was clear that in all cases women had fewer years of schooling than the males of their families and communities and in some cases they had received no schooling at all. Among children too, it was apparent that girls received less schooling and had fewer educational opportunities than their male siblings and peers. Girls were less likely to attend school and more likely to start school later and drop out earlier either because of familial circumstances (including inadequate financial resources, sibling competition, demand for children's labor, and son preference) or disadvantageous school environments (including excessive school fees and distance, and the lack of female teachers). It was reconfirmed that school enrolment figures do not reflect the late start, drop-out rates, and low school attendance of children and especially girls in poor regions, and that a number of steps need to be taken to improve the education levels of poor children and especially girls. Possible solutions include (a) subsidies for school fees and other education expenses which are equitable in allocation and inclusive of both girls and boys, (b) help with safe and reliable transport to and from schools, (c) encouragement of girls' schooling via female teachers, female role models, and reduction of gender bias in the classroom, and (d) recognition and reduction of gender bias within the family where son preference, scarce resources, and practical demands for girls' labor can all disadvantage girls. 1.30 Health is widely perceived by villagers as essential for well-being, and ill-health was identified as a major factor in causing poverty. It was observed that (a) the health of children, despite immunization, was placed at risk by poor nutrition and lack of medical attention in the Chapter 1. Trends in Rural Poverty in the 1990s 13 case of illness, and (b) the lower health status of women was exacerbated by the debilitating strain of heavy physical labor, reproductive health problems, and their lesser attention to health care. The greater severity of women's general ill-health and reproductive-related problems outlined by health officials was confirmed by the PPA team's own observations. Other studies in these same regions have also noted the high incidence of reproductive tract infection and other reproductive health problems made more serious by lack of water, poor hygiene by both men and women and poor or delayed health care. The needs of women for accessible and good quality health care are largely unmet due to a number of factors including high costs, excessive distance, the lack of female doctors, and the lack of adequate pre- and post-natal care. Women tend to place the health needs of other family members before their own, and this means that they often wait until they have reached serious or chronic levels before seeking the required attention. 1.31 Lastly, the PPA team's investigation confirmed the significance of women's kin relations and their own organization as important mechanisms for poverty alleviation. Women's kinship networks are very important as sources of social support for poor families and the efficacy of this social capital in alleviating the effects of poverty. Given that men's kin relations were more village-centered and local, it was affinal relations of the wife's kin who were likely to live some distance and potentially have more resources at their disposal that were often the defining factor in determining a family's well-being. Most often, the better-off families in the villages could draw on such ties for credit, food and other support, while the poorer and more vulnerable families in the village were those without any extended family safety nets related either through the women or men of the family. In all the villages, the wife's parents or siblings provided the most significant sources of support in both production and in times of hardship. This support includes (a) access to credit, loans of tools and animals and labor help to overcome production constraints, (b) gifts or informal, short-term and no-interest loans to meet expenses associated with school fees, medical care and other cash costs, and (c) supplies of food and other safety-net support in times of shortage and crisis. 1.32 The PPA team also found that the Women's Federation (WF) has been marginalized or excluded from the design and implementation of some poverty alleviation projects. However, WF is one of the few government, quasi-government or nongovernmental organizations that have extensive nationwide networks, contacts, and communication channels which are capable of reaching and inclusive of all women. For this reason, WF is one of the most sought-after counterpart institutions in China. It is the experience of both donors and researchers that the levels of activity of WFs at the local level do vary, but usually in direct relation to the funding and other resources which they have at their disposal. Even where a local WF is hampered by lack of resources, it is one of the few networks and communication channels which has the potential to be activated at any time, and experience suggests that it is unlikely to miss an opportunity to benefit local women. WF leaders were adamant that some of the failures of poverty alleviation projects -- to increase the quality of services and personnel skills, and the availability of services to girls and women in education and health and agriculture extension -- were less likely to occur if their organization is involved in project design and implementation. 14 Chapter 2. The Macroeconomic Context 2. THE MACROECONOMIC CONTEXT A. INTRODUCTION 2.1 Since the start of the reforms twenty years ago, China's economy has experienced rapid growth and substantial structural change. These macroeconomic developments, and the reform policies that underlie them, have influenced trends in poverty. Economists have long debated the relationship between aggregate growth and poverty, and most now recognize that growth is an important, but not sufficient, condition for poverty reduction. International experience has shown that the impact of growth on poverty depends not only on the rate of growth, but also on the composition of growth and on structural aspects of the economy that determine whether and how growth reaches the poor (Bardhan, 1996; Islam, 1990; Lipton and Ravallion, 1995). 2.2 The evidence reviewed in this chapter confirms the importance of aggregate growth in explaining China's record of success in poverty reduction since 1990. Both over time and across provinces, growth in per capita GDP has been closely associated with the pace of poverty reduction. The evidence also confirms that the impact of aggregate growth on poverty in China has been substantially influenced by the regional and sectoral composition of that growth. Slower-than-average growth in poor regions explains in part the increasing regional concentration of poverty. In addition, uneven growth in agriculture, the main source of income for the rural poor, has contributed to differences in the rate of poverty reduction. Poverty reduction has been slower where agricultural growth has lagged, and faster where agricultural growth has more or less kept pace with that in other sectors. 2.3 Trends in employment further highlight the importance of agriculture. Agriculture is often viewed as a passive reservoir for China's surplus labor, and expansion of nonagricultural activities is seen as the main route for absorbing this surplus labor. The evidence presented in this chapter suggests a different interpretation of agriculture's role. Estimates of agricultural employment show that in the 1990s labor use in agriculture has risen, and the evidence suggests that agricultural employment responds actively to opportunities for diversification and to increased relative prices. During the 1990s off-farm employment has also been a source of new employment, with migration and private and individual enterprises playing a growing part in generating jobs. 2.4 Recent weak aggregate demand and slowing growth raise concerns about whether China's successful record of poverty reduction can be maintained in the future. A review of China's past macroeconomic performance suggests that China has followed a pattern of cyclical growth since the start of the reforms, and that the recent slowdown is a continuation of this cyclical pattern. International experience has shown that most countries undergo cycles in aggregate demand, and that poverty is sensitive to such cycles because of their impact on unemployment, inflation, the terms of trade, and fiscal capacity (Helleiner, 1987; Kanbur, 1987; Lipton and Ravallion, 1995). Indeed, macroeconomic fluctuations can have devastating and long- Chapter 2. The Macroeconomic Context 15 term effects on the poor, as the poor typically have limited opportunities to insure against income shocks.2 2.5 For these reasons, the continued success of China's poverty reduction policies will require increased attention to the effects of macroeconomic cycles and fluctuations on the poor. Consolidation of China's past successes in poverty reduction will depend on efforts in this direction, because many of those who have recently escaped poverty remain close to the poverty line and economically vulnerable. B. MACROECONOMIC GROWTH IN CHINA IN THE 1990S 2.6 Since the start of the reforms China's economic growth has been high but cyclical (Figure 2.1). This pattern of growth has continued into the 1990s, during which time China has experienced a secular upward growth trend while concurrently passing through almost a full business cycle. The decade began at the bottom of an economic cycle. In 1990 growth in GDP per capita was only 3.8 percent, the lowest rate in the twenty years since the start of the reforms. Growth had slowed following contractionary policies adopted in the late 1980s in response to economic overheating and unprecedented inflation. This period was also characterized by a slowdown in economic reform. 2.7 In late 1989 and 1990 the government implemented a range of Figure 2.1: Growth in GDP stimulatory policy measures, including an easing of credit constraints, increases in 16 agricultural prices, and a major devaluation of 1 the exchange rate. These steps were 1 2 accompanied by the revival of the reform 1 0 agenda, which was raised to center stage by 10 Deng Xiaoping's Southern Tour in early e 8 1992. The economy quickly rebounded, and & 6 growth in per capita GDP accelerated to rates 4 "Ii exceeding 12 percent in 1992, 1993 and 1994. 2 0 2.8 As had been the case in earlier cycles, 80 82 84 86 88 90 92 94 96 98 accelerating growth was accompanied by year some unwelcome side effects. In 1993-94 China experienced inflation, including rapid Source: SSB(a), various issues. price spikes for key agricultural commodities, overly rapid credit expansion, and an import surge leading to a trade deficit. China's leaders responded with a series of measures aimed at achieving a soft landing. Tightened bank credit was combined with selected administrative controls on new investment. An increased urban wage bill and improved conditions in agriculture counterbalanced these contractionary measures. Growth decelerated gradually, a departure from the sharp downturns of earlier cycles. This period also departed from the pattern of past cycles in a second regard: the government continued to actively move forward on major structural reforms even as aggregate growth slowed. 22 A recent empirical study on China by Jalan and Ravallion (1999) finds that the extent to which rural households can insure against income risk is strongly related to their income level, and that poor households are least able to maintain stable consumption in the face of income variability. 16 Chapter 2. The Macroeconomic Context 2.9 By 1996 these measures had successfully brought inflation under control while maintaining slower, but still robust GDP growth. Growth, however, continued to decline, to 8.8 percent in 1997 and further to 7.8 percent in 1998. Weak aggregate demand since 1997 mainly reflects domestic developments, and to a lesser extent external factors following the East Asian financial crisis. The urban state sector has suffered from rising inventories, excess capacity, and record losses. Ongoing restructuring of state-owned enterprises and the extensive downsizing of both the government bureaucracy and state enterprises have contributed to increased urban unemployment, with repercussions on equilibrium in the broader labor market. In rural areas incomes have been affected by falling prices for farm products and stagnant growth in non- agricultural rural employment. The government has responded to these developments with a stimulus package of fiscal spending on infrastructure investment and loosened monetary policy. 2.10 In the 1990s, then, the Chinese government has displayed an ongoing commitment to dampening macroeconomic fluctuations through coordinated, countercyclical fiscal and monetary policies. Such policies have to some degree moderated, but have not eliminated, the cyclical pattern of China's growth. 2.11 Growth and the Poor. The 1990s have seen both growth in GDP per capita Rgure 2.2: The Relationship between and a substantial decline in rural poverty PerCapita GDPand Rural Poverty, 1990- (Figure 2.2).7 These trends suggest that 97 aggregate growth is a key factor explaining China's successful record in 4.6 poverty reduction. Indeed, the elasticity of , 4.5 - rural poverty with respect to aggregate o 4.4 growth has been about -0.8, implying that T. for each percentage point growth in GDP > per capita, rural poverty has declined 0.8 4.2+ percent. ' 4.1 0 4_ 2.12 This aggregate relationship, 39 however, sheds little light on the 14.4 14.5 14.7 14.8 14.9 15.0 15.1 15.2 mechanisms through which growth has log of GDPpercapita benefited the rural poor. A better understanding of these mechanisms Note: GDP is in constant prices. Sources: Chapter I and requires examining (a) the regional SSB(a), various issues. dimensions of aggregate growth, as the distribution of both growth and poverty have been regionally uneven, and (b) the sectoral composition of growth, and its relationship to the main sources of income for the poor. 2.13 Regional Differences in Growth. Analysis of changes in the regional patterns of poverty in Chapter I showed that in the 1990s poverty has become increasingly concentrated regionally. This changing concentration of poverty is partly explained by the fact that, while growth in almost all provinces has been high by international standards, poor provinces have grown more slowly than wealthy provinces. Between 1991 and 1996 rates of growth in per capita GDP (constant prices) were less than eight percent in Qinghai and Guizhou but more than seventeen 23 Here and elsewhere in this chapter, the number of poor and poverty rates are calculated using the official govemment poverty line. Chapter 2. The Macroeconomic Context 17 percent in Fujian, Jiangsu and Zhejiang. Table 2.1 shows that, more generally, growth has been fastest in the East and Center, relatively rich regions, and slowest in the Southwest and Northwest, China's poorest regions. A notable exception to this pattern is the Northeast, which, despite starting in 1991 with high per capita GDP and a low poverty rate, experienced relatively slow growth. Table 2.1: Regional Patterns: Growth versus Change in Rural Poverty, Initial Level of Income and Initial Poverty Rate Rate of Change in Growth in Per the Rural Poverty Per Capita Rural Poverty Capita GDP (%), Rate (%), GDP (yuan), Rate (%), 1991-96 1991-96 1991 1991 East 15.9 -31.1 1938 7.7 Center 14.1 -24.1 1639 10.3 North 12.1 -19.3 2089 13.4 Southwest 10.8 -4.5 1103 13.2 Northeast 10.6 -12.4 2216 9.1 Northwest 8.7 -1.2 1412 19.8 Note: GDP growth is in constant prices. Sources: Chapter I and SSB(a), various issues. 2.14 A plot of provincial rates of rural poverty reduction against GDP growth reveals furtier that provinces with more rapid growth in per capita GDP have experienced more rapid declines in the number of rural poor (Figure 2.3). Simple regression analysis using these provincial data reinforces this conclusion: between 1991 and 1996 an increase in average annual per capita GDP Figure 2.3: ProvincialGrowth Rates versusRatesofPoverty Reduction 46 20.0% 10.0% E w i r 00.0% & o -100.0% coE-20.0% 0 -30.0% - E -40.0% U c -50.0% - * U 0.1 0.1 0.2 0.2 growth rato of porcapita GDP(constant prices) Sources: Chapter land SSB(a), various issues. 18 Chapter 2. The Macroeconomic Context growth of one percentage point accelerated the rate at which poverty declined by 1.8 percentage points.24 In view of these conditions, the increasing concentration of poverty is not surprising. 2.15 Sectoral Differences in Growth. The sectoral composition of growth can, like the regional composition, influence trends in poverty. China's aggregate growth rates have largely been driven by rapid growth in industry (see Annex 2 Table I 1). While agricultural growth in the 1990s has been relatively high, it has lagged behind that in nonagriculture, and consequently agriculture's share of GDP has declined over time. This raises the question of how poverty has been affected by the changing structure of the economy. Does the decline in poverty in the 1 990s mean that the poor have been able to benefit from growth in nonagricultural sectors and increasingly earn income from off-farm or wage earnings? Or has growth in agriculture, although slower than in other sectors, been the main force underlying poverty reduction? 2.16 Closer analysis of the data on sectoral changes and poverty rates reveals that poverty in China is highly sensitive to trends in agriculture. The rate of poverty reduction has been slower in provinces where agriculture's share of GDP has fallen markedly. Simple regression analysis shows that, holding constant the rate of overall GDP growth, a one percent decline in agriculture's share of GDP reduces the rate of decline in poverty by 1.9 percentage points25 2.17 This impact of agriculture on poverty reduction is illustrated in Table 2.2. Table 2.2 gives the regional shares of agriculture in GDP for 1991 and 1996. In some regions (the Center, East, and Southwest) agriculture's share fell substantially, that is, agriculture grew much more slowly than per capita GDP. In other regions (the North and Northeast) agriculture's share remained constant or declined only slightly. Here agricultural growth more or less kept pace with that in other sectors. 2.18 These differences influenced poverty. The last column of Table 2.2 contains estimates of the rate at which rural poverty would have declined if agriculture's share in GDP had remained constant, that is, if agriculture had grown at the same rate as all other sectors. These estimates indicate that, on average, if agriculture's share had remained unchanged, the rate of poverty reduction nationwide would have been nearly double the actual rate. For poor regions, rates of poverty reduction would have risen even more dramatically. This is especially true for the Southwest, where agriculture's share of GDP declined substantially between 1991 and 1996. 24 This regression used data from 26 provinces. Beijing, Tianjin, Shanghai, and Zhejiang, an outlier, were excluded from the sample. Data for this same sample are shown in Figure 2.3. For the regression, the average annual rate of change in the number of rural poor, 1991 to 1996, was the dependent variable, and the independent variable was the rate of growth in per capita GDP, 1991-1996, in constant prices. The estimated coefficient on growth was - 1.796, with significance at the 5% level. The adjtisted R2 was 0.15. The estimated coefficient on growth remains the same (or increases) in both magnitude and significance when the initial number of poor or initial poverty rate is included as a control. 25 This regression used data from 26 provinces. Beijing, Tianjin, Shanghai, and Zhejiang, an outlier, were excluded from the sample. For the regression, the average annual rate of change in the number of rural poor, 1991 to 1996, was the dependent variable. The independent variables were the rate of growth in per capita GDP, 1991-1996, in constant prices, and the change in the primary sector share of GDP. The estimated coefficient on per capita GDP growth now increases to -2.75 (significant at the 1% level). The coefficient on the change in the primary sector's share is -1.87 (significant at the 1% level). The adjusted R2 is 0.36. Thus adding the change in the primary sector's share of GDP as an independent variable increases both the explanatory power of the regression and the estimated coefficient on GDP growth. Including the initial number of poor or initial poverty rate as a control variable does not substantially alter these results. Chapter 2. The Macroeconomic Context 19 Table 2.2: The Importance of Agriculture for Poverty Reduction, 1991-96 Rate of Change in the GDP Share of GDP Share of Rate of Change Number of Poor if Agriculture Agriculture Change in in the Number Agriculture's Share Ilad 1991 1996 Agriculture's Share of Poor Remained Constant (percent) (percent) (percentage points) (percent) (percent) East 21.6 15.0 -6.6 -29.2 -41.5 Center 30.5 21.9 -8.6 -23.1 -39.2 North 13.1 12.0 -1.1 -20.7 -22.8 Southwest 38.0 28.8 -9.2 -3.0 -20.2 Northeast 20.2 19.1 -1.2 -12.2 -14.4 Northwest 30.8 26.3 -4.4 -0.6 -8.8 National 24.5 18.7 -5.8 -12.4 -23.3 Note: The regional numbers exclude Beijing, Tianjin, Shanghai, and Zhejiang, an outlier. The national numbers include all provinces, independent municipalities, and autonomous regions. The numbers in the last column are calculated using the estimated regression coefficient for the primary share of GDP (-1.87, see footnote 25), and with the assumption that between 1991 and 1996 agriculture's share of GDP remained unchanged at the 1991 level. That is, the numbers in the last column equal the actual rate of change in the number of poor plus 1.87 times the change in agriculture's share. Sources: Table 2.1, footnote 25, and SSB(a), various issues. 2.19 The importance of agricultural growth to poverty reduction is explained by the fact that agriculture remains the primary source of income for the rural population and for the rural poor. Evidence suggests that, despite the rapid growth of township and village enterprises and other nonagricultural activities in rural areas, the income of rural households is still predominately derived from agriculture, and especially from crop production (Table 2.3). In 1997 agriculture generated 56 percent of rural household income. Although agriculture's share of rural household income fell somewhat in the 1990s, the share of crop production showed no sign of a secular decline. The contribution of crop production to rural household incomes fluctuated from year to year, but remained within a range of 43 to 49 percent of rural household income. Available data suggest that the importance of crop income is even higher for the poor.6 Table 2.3: Percent of Rural Household Net Income Per Capita from Agriculture 1990 1991 1992 1993 1994 1995 1996 1997 Agriculture total 62.8 61.3 58.5 59.1 58.4 58.0 57.1 55.9 Crop cultivation 48.1 45.7 43.1 47.6 48.4 49.1 48.0 45.1 Other agriculture 14.7 15.6 15.4 11.5 10.0 8.9 9.1 10.8 Note: Other agriculture includes forestry, aquaculture and animal husbandry. Agriculture total is the sum of crop production and other agriculture. Source: SSB rural household survey data, SSB(a), various issues. 26 Available data for Guizhou, for example, show that on average in 1997 crop production contributed 53 percent of net household income, and that the share of crop production in income was higher for the poorest income groups. For households in Guizhou with per capita net income of less than 790 yuan per capita (constituting the bottom 17 percent of the income distribution), crop production contributed 66.4 percent of net income (Guizhou Statistical Bureau, 1998). 20 Chapter 2. The Macroeconomic Context 2.20 The 1990s have seen rapid growth in rural enterprises, as well as an unprecedented expansion of rural-urban and intra-rural migration. These developments have led some observers to conclude that off-farm wage earnings have played the major role in raising rural incomes and reducing rural poverty. Available statistics on the share of wages in rural household incomes do not give strong support to this conclusion (Table 2.4). During the 1990s the share of rural household income from wages fluctuated between 20 and 25 percent of net household income, with only weak indications of a secular increase (much of which may reflect the fact that rural off-farm employment was abnormally low in 1990). In the 1990s, then, wage income has not grown disproportionately, but at about the same rate as other sources of income. Table 2.4: Percent of Rural Household Net Income Per Capita from Wage Earnings and Transfers 1990 1991 1992 1993 1994 1995 1996 1997 Wages 20.2 n.a. 23.5 21.1 21.5 22.4 23.4 24.6 Transfers 3.6 n.a. 3.9 3.7 3.9 3.6 3.6 3.8 Sources: SSB rural household survey data, SSB(a) and SSB(b), various issues. 2.21 Data from the national rural household survey do not provide separate figures for remittances. Remittances are included in "transfer income.'i7 Transfer income has remained fairly stable at less than 4 percent of rural household income in the 1990s. This suggests that remittances, like wages, have not grown disproportionately. 2.22 These national statistics suggest that wage and remittance incomes' shares of net rural household income have remained relatively stable, but they give few clues about the situation for poor households. Information on the share of wages by income group for 1996 reveals a complex picture of wage earnings among the poor. The very poorest households appear to rely heavily on wage employment as a source of income (Figure 2.4). For households with per capita net income of less than 200 yuan, the wage share is nearly 40 percent, as compared to the national average of 23 percent. 2.23 As income rises, the wage share quickly falls. By the time per capita income has risen to Y 300-400, the wage share is well below the national average. Since the population shares in the bottom income groups are very small-only 0.35 percent for the poorest group and 0.43 percent for the second poorest group-this decline cannot be detected unless data at the bottom of the income distribution are highly disaggregated. Once per capita incomes reach Y 600-800, the wage share begins to rise with income. Hereafter the wage share increases gradually, equaling or exceeding the national average only after reaching the top 20 percent of the income distribution. 27 Data for Guizhou show that remittances from family members and relatives living away from home accounted for about 40 percent of "transfer" income, and less than 2 percent of net income in 1997. Chapter 2. The Macroeconomic Context 21 Figure 2.4: Wagesasa Share of Income by Income Group, 1996 40 r5wage share ofincome * 30 525ll_ share ofpopulation 20 15 20 nationalaverage wage 510 k sh a re 8 o 0 N oo 0 00 a S 5 income group 2.24 Similarly disaggregated data for Shaanxi and Guangxi are consistent with this national pattern and provide additional information about the very poorest (Annex 2 Tables 2 and 3). In these provinces, as in the national sample, the bottom income groups have higher-than-average shares of wage income. In Shaanxi, for example, the poorest 1.6 percent of the sample earns nearly half its income from wage employment. The data for Shaanxi and Guangxi also reveal, however, that the high share of wage income for the very poorest income groups is due not to relatively high wage earnings, but to low net income from household production. Surprisingly, their gross revenues from household production are fairly high, exceeding those of the next lowest income group or groups. Yet net income remains low, because the costs relative to the revenues of household production are unusually high .28 2.25 This pattern implies that the very poorest households-those in the bottom 0.5 percent or 1 percent of the income distribution-may differ from households in the next lowest income groups. These households appear to have high gross but low net income from household production. The reasons for this could be several. Transitory income shocks, such as natural disaster, loss or theft of production inputs, or family illness could abnormally reduce gross agricultural revenues or abnormally increase the costs of production. Alternatively, low net revenues from farm production could be due to indebtedness. In these data the costs of production include repayments of interest and principal on production loans. Consequently, households with relatively large debt burdens could experience low net income despite earning reasonably high gross revenues. If indebted households eventually can pay off their debts, then their inclusion in the poorest income group would be temporary. Regardless of whether these households are affected by an income shock or indebted, their wage earnings serve as an important source of income to offset their low net earnings from production. 28 Data on revenues and costs of household production are not available for the national sample, but here, too, the high share of wage income for the poorest income group is caused by unusually low non-wage income rather than by relatively high wage earnings. Wage income per capita for the poorest income group is 62 yuan, close to that of the next three lowest income groups, all of which have wage income of 60 to 70 yuan per capita. Non-wage income of the poorest group, however, is only 93 yuan per capita, while non-wage income for the second lowest group is double this amount, 185 yuan. Non-wage income continues to rise to 294 yuan, 384 yuan, and 470 yuan as one moves up from the second to the fourth poorest income group. Non-wage income for these groups is likely to be largely derived from household production. 22 Chapter 2. The Macroeconomic Context 2.26 While the above data do not provide information about changes over time in the importance of wage income for the poor, they clearly indicate that by the mid-1990s wages constituted an important source of income for certain subgroups of the poor. Yet even for these subgroups, agriculture remains the dominant source of income. This explains why trends in agricultural GDP continue to have a significant impact on poverty. C. TRENDS IN EMPLOYMENT 2.27 The 1990s have seen new developments in employment that, despite the economic slowdown in recent years, have created substantial new job opportunities for the rural population. These developments have contributed to the trends in rural income described above. The supply of labor has grown steadily but fairly slowly at rates of less than 1.5 percent annually overall, and of 1 percent or less in rural areas (Table 2.5). These growth rates are lower than those in the 1980s and reflect the impact of population planning policies on age cohorts now entering the labor force. The strict population policies initiated in 1980 should cause growth in the labor force to slow even further in the coming decades. 2.28 Nonagricultural Employment. While growth in the labor force has been relatively slow, macroeconomic growth in the early and mid-1990s, combined with policy liberalization, generated fairly rapid growth in labor demand. Growth in nonagricultural employment has received considerable attention, as it is seen as a way of providing employment for surplus rural labor. Trends in nonagricultural employment, however, are difficult to track because the official statistics are problematic. The official statistics count the number of workers whose main occupation is nonagricultural, but typically such workers also work part time in agricultural production. Thus the official statistics on the nonagricultural labor force overstate employment in nonagricultural activities (Zhou, 1997).29 Furthermore, due to problems in statistical work at local levels, the official statistics on growth in township and village enterprises (TVEs) in the 1990s have been overstated. For this reason, in 1997 the MOA adjusted the official TVE statistics downward by a large margin. This statistical adjustment explains in part the sharp decline in official figures for rural enterprise employment in 1997. 29 Zhou (1997), based on MOA survey of 4000 rural households in 8 provinces, estimates that 28 percent of the work time of so-called nonagricultural workers is spent in agriculture. Chapter 2. The Macroeconomic Context 23 Table 2.5: Growth in China's Labor Force Economically Active Rural Employed Population Persons (annual (annual (1OOOs) growth) (I 000s) growth) 1990 644830 n.a. 472950 n.a. 1991 653990 1.4% 478220 1.1% 1992 661840 1.2% 483130 1.0% 1993 670330 1.3% 487840 1.0% 1994 678790 1.3% 487860 0.0% 1995 687370 1.3% 488540 0.1% 1996 696650 1.4% 490350 0.4% 1997 705800 1.3% 493930 0.7% Note: The data in this table are the SSB's adjusted series, where the adjustments are based on the annual one percent sample surveys of the population. The series used in this table are jingii huodong renkou (economically active population) and xiangcun congye renyuan (rural employed persons). Sources: SSB(a), various years and SSB(1997). 2.29 The data for private and individual enterprises are also weak. Increasingly fuzzy ownership structures, as well as strategic switching of registration between collective and private ownership, means that an enterprise counted in one year as private may the next year have been counted as a TVE, or vice versa. Furthermore, treatment of small-scale, seasonal household nonagricultural activities has been uneven over time and among regions. At times such activities have been included in the numbers on individual enterprises, and at other times not. 2.30 Despite these problems, the official statistics give an indication of broad trends (Table 2.6). These numbers suggest that rural employment in collective, private and individual enterprises grew substantially in the early 1990s, surpassing one-third of the rural work force by the mid-nineties. Partly as the result of the statistical adjustment and partly as the result of the macroeconomic slowdown, employment in rural enterprises fell in 1997. Even with this decline, in 1997 employment in rural enterprises was 22 percent higher than in 1990. 2.31 A notable feature of this employment growth is that more than half of it was created by private and individual enterprises. While TVE employment accelerated and then decelerated with the macroeconomic cycle, employment in private and individual enterprises continued to rise. By 1997 these enterprises employed 25 million more people than in 1990, and their share of rural employment had risen from only 3 percent in 1990 to over 8 percent. Even if some of this growth reflects the reclassification of collective enterprises as private, the evidence points to a qualitative increase in the importance of private and individual business that is likely to continue in the future. 24 Chapter 2. The Macroeconomic Context Table 2.6: Growth in Rural Enterprise Employment Employment in Rural Enterprises Township and Private and Percent of Total Village Individual Rural Labor Enterprises Enterprises Force (annual (1 000s) growth) (1 OOs) (1000s) (%) 1990 108690 92650 16040 23.0 1991 113410 4.3% 96090 17320 23.7 1992 124870 10.1% 106250 18620 25.9 1993 145420 16.5% 123450 21970 29.8 1994 148840 2.4% 120170 28670 30.5 1995 163870 10.1% 128620 35250 33.5 1996 173670 6.0% 135080 38590 35.4 1997 132800 -- 91580 41220 26.9 Note: Prior to 1997 the official data reportedly overstated employment in TVEs. The TVE Bureau of MOA changed the coverage of the TVE data series in 1997, and so the 1997 figures are not fully comparable to those for earlier years. Source: SSB(a), 1998. 2.32 A second notable development in the 1990s is the expansion of migrant employment. While estimates of rural migrant labor vary, they all show substantial growth-a doubling or tripling-since the late 1980s. Table 2.7 gives two estimates of migrant employment. The relatively conservative figures in the first two columns show migrant labor rising from 2.2 percent of the rural labor force in 1989 to over 5 percent in 1995-96. Estimates of Rozelle, et al, 1999 indicate that between 1988 and 1995 migrant employment rose from 5 to 12 percent of the rural labor force. The 1997 agricultural census gives numbers close to those of Rozelle, et al. The census found that in 1996 12.9 percent of rural workers were employed outside their township of residence (National Agricultural Census Office, 1998). These trends suggest that rural residents are increasingly willing to travel to find employment opportunities. In the long term, such labor g30 movement will reduce geographical inequality in wage earnings. 30 Data on migrant labor include migrants employed in TVEs and other rural enterprises, and so they do not entirely reflect new job creation beyond that already captured by the data for rural enterprise employment. Chapter 2. The Macroeconomic Context 25 Table 2.7: Estimates of Migrant Employment Ministry of Agriculture Rozelle, et al 1,000s Percent 1,000s Percent 1988 20034 5 1989 8875 2.2 1992 13785 2.8 1994 22961 4.7 1995 24495 5.0 58624 12 1996 25191 5.1 Note: The figures for migrant labor from MOA are the numbers of people employed outside the village as contract or temporary workers (waichu hetong, linshi gong). The Rozelle, et al, numbers are from a nationally representative sample survey and include people who leave the village for at least one month during the year, but who retain ties with the village by returning during Spring Festival or during peak farming seasons. They exclude commuters who are employed outside the village but live at home. Note that these are counts of the number of people, not of standard labor days. Also, these numbers include migrants employed in rural enterprises, and so overlap with data in the previous table. Percentages are calculated by dividing by the number of rural employed persons (Table 2.5). Sources: MOA(a), various issues and Rozelle, et al, 1999. 2.33 Although in the long term migration will continue to grow, in the short term migration has, like TVE employment, been affected by the economic slowdown. Weak aggregate demand and extensive layoffs by state and collective enterprises, prompted both by macroeconomic trends and by SOE reforms, have reduced employment opportunities in urban and coastal areas. Rising urban unemployment has resulted in a growing intolerance for rural migration by urban governments and populations. So as to protect urban workers, urban governments have in some cases tried to segment the labor market by requiring urban employers to hire migrants only for unskilled, menial jobs. These developments illustrate the sensitivity of rural off-farm employment both to urban reforms and to macroeconomic fluctuations. 2.34 Agricultural Employment. Trends in agricultural employment have received less attention than trends in nonagricultural employment. Although it is generally acknowledged that agriculture is the major employer in rural areas, it is often viewed as a passive reservoir for surplus labor and expansion of nonagricultural activities is seen as the main route for absorbing this surplus labor. The official statistics on employment appear to support this view (Table 2.8). They show that the share of rural workers employed in agriculture has declined markedly, from nearly 80 percent of the labor force in 1990 to 70 percent in 1997. They also show that agricultural employment rose during the economic slowdown in 1989-91, which hit rural enterprises particularly hard, fell when overall growth and employment in rural enterprises rebounded in the early 1990s, and edged upward in 1997, when TVE growth again slowed. 26 Chapter 2. The Macroeconomic Context Table 2.8: Official Data on Employment in Agriculture (millions) 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 Employment in Agriculture 314.6 324.4 333.4 341.9 340.4 332.6 326.9 323.3 322.6 324.3 Percent of Rural Labor Force 78.5 79.2 79.4 79.3 77.7 75.2 73.2 71.8 71.2 70.6 Note: Workers are counted as employed in agriculture if agriculture is their main profession. The figures on agricultural employment are not adjusted to be in line with the one- percent sample population surveys, and so the data in this table are not comparable with those in Table 2.5. Source: SSB(a), 1998, pp. 387-88. 2.35 These numbers, however, tell only a partial story. The official data on agricultural employment count the number of people whose main occupation is farming, and they do not capture the reallocation of labor time between different agricultural and nonagricultural activities. In order to gain a more accurate estimate of agricultural employment, Rawski and Mead (1998) calculated alternative estimates of agricultural employment. Their estimates, which use data on labor input per unit output from national surveys of production costs, suggest that trends over time in the level of agricultural employment differ from those implied by the official statistics. While the official statistics show a rise in agricultural employment in the late 1980s followed by a decline in the 1990s, Rawski and Mead's estimates showed that agricultural employment, here measured in standard labor years, fell in the 1980s and then began to rise in the early 1990s. Estimates calculated using labor per unit land area by Zhou Qiren similarly found that agricultural employment rose, rather than fell, in the early 1990s " 2.36 The above studies are now several years old, and so a new set of estimates of agricultural employment have been calculated for this report. The estimates are based on a methodology similar to that of Rawski and Mead 32 These new estimates also show positive net growth in agricultural employment in the early 1990s: between 1989 and 1996 agricultural employment grew 15 percent, or 22 million labor-years (Figure 2.5).33 These estimates, which are conservative, suggest that between 1991 and 1996 agriculture generated roughly the same amount of new employment as did rural enterprises. 31 Zhou Qiren's estimates assume that the labor input coefficients remain unchanged at the 1986 levels, and so do not capture changes over time in labor productivity. Rawski and Mead's estimates, as well as the new estimates presented here, allow labor input coefficients to change over time. 32 Recent data on labor coefficients for some products are unavailable, and so the estimates for this study cover fewer farm products than Rawski and Mead. Thus the estimated absolute levels of employment are here lower than Rawski and Mead's. Changes over time are comparable under the assumption that for these products labor use per unit changed at the same rate as the average for products for which recent data are available. 33 The assumption of 300 workdays per year gives relatively conservative estimates of the increase in agricultural employment. Using 260 workdays per year gives an increase of 25 million labor years. Higher estimates also result if one assumes that for products with incomplete data, labor use per unit remained unchanged at the level in the last year for which data are available (1989 or 1991, depending on the product). In this case, net growth in agricultural employment between 1989 and 1996 is 33 million labor years. Chapter 2. The Macroeconomic Context 27 Figure 2.5: IndicesofAgnculturalEmployment, 1989=100 1.20 T 1.15 1 1.0o IN 1.05 ~~~~~~~~~~~~0 ffic itaI 1.00 s . t+ . - t +Rawski& Mead 1.00 - ~~~~~~~~~~~~estimates 0.95 - alternative estimates' 0.90 0.85 1986 1988 1990 1992 1994 1996 Sources: See note to Table 2.9 2.37 This net growth in agricultural employment occurred despite higher efficiency of labor in the production of most farm products-on average for products included in these estimates, between 1989 and 1996 output per unit labor rose 22 percent. The major factors behind the growth in agricultural employment were increased production of relatively labor-intensive crops and expansion of animal husbandry and aquatic production. This can be seen in Table 2.9, which gives estimated growth in employment by product group. Use of labor in the production of major cereals remained more or less unchanged, while labor use in the production of minor grains, tubers and legumes fell. In contrast, labor use in vegetable, fruit, animal, and fish production each rose by 50 percent or more. In absolute terms, the largest increases in labor use were for vegetables and animal products. 2.38 These estimates of agricultural employment suggest that labor use in agriculture is not simply "passive," but responds actively to changes in the relative returns to different farming activities, and also to the returns of farm relative to off-farm activities. Most of the estimated growth in agricultural employment occurred after 1993, which years saw important agricultural reforms including a major liberalization of grain policies in 1992 and 1993 (Lyons, 1998; Rozelle, et al, 1996; Sicular, 1995; Watson and Findlay, 1995). These years were characterized by rising farm prices and the rapid expansion of free market trade in farm products, developments that raised the returns to agriculture, permitted diversification into high-value crops, and encouraged the expansion of agricultural employment. 2.39 Employment and the Poor. While the poor have undoubtedly benefited to some degree from the broad trends in employment described above, they have faced special challenges. One challenge arises because the supply of labor in poor regions has grown more rapidly than the supply of labor in other regions. This is especially true in poor minority regions, which have had relatively liberal population policies. Thus while the national rural labor force grew 1.3 percent annually between 1990 and 1997, the labor force in Ningxia grew 3.3 percent, Qinghai 2.5 percent, Guizhou 2.3 percent, and Xinjiang and Guangxi 2.0 percent (Table 2.10). Rapid growth in labor supply means that these regions must maintain either faster-than-average job creation or higher-than-average outflows of migrant workers in order simply to maintain existing employment levels. 28 Chapter 2. The Macroeconomic Context Table 2.9: Estimated Change in Labor Use in Agriculture, 1989-1996 Percent Change in Absolute Change in Labor Use Labor Use (millions) Total 15.3% 21.9 Major grains 0.9% .6 Other grains, tubers and legumes -24.5% -3.7 Oil crops 26.9% 1.9 Vegetables 36.8% 6.3 Fruit crops 170.7% 2.9 Other cash crops 4.0% .7 Animal products 64.9% 11.7 Freshwater aquatic products 80.5% 1.6 Note: Products included in the calculations for total agricultural employment are: major grains (rice, wheat, corn); soybeans; tubers and other grains and legumes; oil crops (rapeseed, sesame, peanut); vegetables; fruit crops (apple, pear, citrus); other cash crops (cotton, ambary hemp, sugarcane, sugarbeet, flue-cured tobacco, tea, silkworm); animal products (pork, poultry meat, eggs, milk); and freshwater aquatic products. To avoid distortions due to inaccurate area statistics, estimates for most products are calculated using days of labor input per unit output multiplied by output quantity. Due to lack of data on input per unit output for vegetables and for tubers and other grains and legumes, estimates for these products use labor input per unit sown area times sown area. Days are converted to years by assuming 300 workdays per year. Note that many minor products such as beef, minor fruits, and ramie are not included, and so the estimated change in total labor use therefore implicitly assumes that the net change in labor use for such omitted products is zero. Sources: SSB(a) and SSB(b), various issues; Rawski and Mead (1998), p. 769; The World Bank (1992), p. 71; Economic Research Service, U.S. Department of Agriculture, (personal communication). 2.40 The record of off-farm job creation in poor regions has been generally weak. During the 1990s growth of TVE employment in many poor provinces was slower than that nationally. Provinces that lagged most noticeably were Gansu, Shaanxi and Qinghai. Sichuan and Chongqing performed slightly better than the national average. The only poor province with growth in TVE employment that significantly exceeded the national average was Guangxi, which benefited from its proximity to Guangdong. 2.41 Job growth for individual and private enterprises in poor provinces has also been below average: in southwest and northwest China all but one province experienced below-average growth in employment in private and individual enterprises during the early 1990s. Despite the fact that poor regions lagged, however, their growth in such employment exceeded growth in the labor supply, and often by a large margin. Indeed, in Yunnan, Inner Mongolia, Shaanxi, Gansu and Ningxia, annual growth in employment by private and individual enterprises was well above 1 0 percent. 2.42 The extent to which the poor have benefited from the expansion of agricultural employment outlined above is unclear. A recent study using data for 1990 found that grain was the dominant source of income for the poorest income groups (The World Bank, 1997a). Systematic evidence on trends since 1990 is, however, unavailable. This period has seen the substantial diversification in agriculture away from grain, and evidence from the field reveals that at least some poor households have taken advantage of these trends to escape poverty (see Chapter 4). Chapter 2. The Macroeconomic Context 29 2.43 While further research is needed to clarify how trends in agricultural employment have affected the poor, it is probably safe to say that these trends have varied among regions. Such variation would in part reflect differences among poor regions in their agricultural endowments. Some poor areas have such weak agricultural resource bases that, in the absence of subsidies, agricultural development is not financially viable. In these areas off-farm employment and migration provide more viable routes to escape poverty. In other areas, however, the problem is not the resource base, but economic constraints such as lack of credit to pay for inputs or necessary initial investments in water control, land improvement, and animal stocks; lack of appropriate technical information and training; and weak marketing support (Chapter 4). In combination with measures to overcome these constraints, agriculture has the potential to provide increased employment in such areas. Table 2.10: Regional Growth Rates in the Labor Force and Off-farm Employment Growth in Individual and Private Growth in the Growth in TVE Enterprise Labor Force Employment Employment 1990-97 1990-97 1992-97 NATIONAL 1.3% -0.2% 17.2% North 1.2% -4.3% 21.4% Northeast 1.7% -5.6% 29.1% East 1.0% 1.3% 14.6% Central 1.4% 0.3% 20.3% Southwest 1.4% 1.7% 10.4% Sichuan & Chongqing 0.7% -0.5% 10.8% Guizhou 2.3% -2.1% 8.4% Yunnan 1.9% -1.8% 14.9% Tibet 1.1% 0.0% 6.1% Guangxi 2.0% 10.8% 6.4% Northwest 1.7% -3.8% 14.8% Inner Mongolia 1.7% 1.4% 15.2% Shaanxi 1.6% -4.9% 18.2% Gansu 1.3% -8.3% 15.6% Qinghai 2.5% -3.2% 9.4% Ningxia 3.3% -2.5% 13.1% Xinjiang 2.0% -1.4% 3.8% Note: National and regional growth rates are weighted averages. Provincial data for employment in individual and private enterprises are not available before 1992. Sources: SSB(a), various issues; MOA, 1993; SSB(b), 1993. 2.44 Migration can offset these features of labor supply and demand in poor areas. Evidence on the extent to which the poor have participated in migration, however, is mixed. Some studies have found that migrants tend to come from poorer inland provinces and from poorer villages in those provinces, while others have found that migration is more likely from middle-income than 30 Chapter 2. The Macroeconomic Context from poor villages.34 Most studies acknowledge that the poorest households face obstacles in finding migrant work because they typically have difficulty raising the funds needed to cover the initial costs of transportation and job search, are hindered by low literacy levels, and have limited access to information about migrant work opportunities (The World Bank, 1997b). This may explain why some researchers have found that although migrants come disproportionately from low-income localities, they are usually not from the poorest households in those localities (Du and Bai, 1997). 2.45 Finally, poor households have been held back from taking advantage of new employment opportunities both in agriculture and non-agriculture by the fact that their labor time is often absorbed by necessary, low-income, tasks. For example, poor households are more likely than richer households to travel long distances to reach the nearest water supply (Liu and Wu, 1997). This is especially true in mountain areas, where during the dry season it is not uncommon for households to devote four or more hours daily to this task. Similarly, in mountainous and hilly areas poor households often producing grain for subsistence can absorb a large share of family time, because distances between plots are far, the time required for cultivation of sloped land are high, and yields per unit land and labor low. Even when off-farm work is available, then, poor households' ability to participate may be constrained. 2.46 Under such conditions, investment in water supply, land terracing and other forms of basic infrastructure can have positive effects on employment and incomes. These investments allow poor households to reallocate work time from low-return subsistence tasks to more lucrative activities. Field investigations have shown that in villages that have participated in international donor-supported rural development projects providing water supply systems and land terracing, households begin to expand their animal husbandry production, start sideline activities, and look for off-farm employment. When asked why, they respond that the land and water supply investments have liberated their time and allowed them to pursue such activities (para. 4.14). D. INFLATION AND PRICES 2.47 Cycles of aggregate growth in China have been Figure 2.6: Annual Change in GDP and Prices closely matched by inflationary cycles. As shown in Figure 2.6, 25.0% T rising inflation generally follows + t upswings in growth with a lag of 20.0% A about one year. Inflation is 1/a-iiVchange ina generally thought to reduce the 5. . , real incomes of groups that hold 1 -.0 .|.annualchnei cash-denominated savings or l \ l pe de have fixed nominal incomes, and 5.0% - - / to transfer income from lenders A . to borrowers. The rural poor 0.0% generally have low savings, are 1985 1987 1989 1991 1993 1995 1997 net borrowers, and, since most Source: SSB(a), 1998. of their income is from production, do not have fixed 34 Rozelle, et al, 1999, gives a summary of the literature on this question. Chapter 2. The Macroeconomic Context 31 nominal incomes. Thus inflationper se need not have a net negative effect on the real incomes of the rural poor (Cardoso, 1992). 2.48 Inflationary cycles can nevertheless cause difficulties for the poor because during these cycles not all prices move together. The timing and amplitude of price cycles differs among products. In China cycles in the level of prices for farm products have greater amplitude, and occur earlier, than do cycles in the prices of items that farmers buy. This can be seen in Figure 2.7, which shows the percentage changes in prices for farm products, farm inputs, and rural consumer goods. The rate of increase in rural output prices is higher during inflation peaks, and lower during troughs, than those of inputs and rural consumer goods. During troughs, for example, farm product prices often decline in absolute terms, a rare occurrence for inputs and consumer goods. The greater amplitude of farm product price cycles causes the terms of trade to move in favor of farmers during inflationary peaks, and against farmers during troughs. Rgure 2.7: Annual Change in Farm Output, Input, and Rral Consumer Piices 40.0%1o 30.0% 20.0% +~ , - V- 1L' \ - _- farm input prices 10.0% ~ A -K' . ' \\ _ BYZAPROVINCE, 1996 g1 (Percent of Rurail Populalion Below Government Poverty iJeeJ .- . .lff'- Head Caunil} - C GUANGXI 1i DUANDDCONG4 \ 0.00 100 ®I Nahounal Capiha - _ (ssC2 5.00 10 -.- Inlematianl Bound.d, e 7 VIETNAM A44t- § tO . , ;, 1000 -1500 C 00 -2000 0 200 400 6000M. C PEa E '. m" N2 t ° Ab 21 3bo4C rA20.00 - 4003 I)EP -. N.Dl 0 100 200) 300 4C'OM,16Ta-A IBRD 30488 7 0'\ 8 0 90' 10 110 1 20' I 30o 50'~ ~ ~ ~ ~ ~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~0 \ ~~~~~~~RUSSIAN FEDERATION , 50',. f~~~~~~~~~ KAZAKHSTAN 7 _ ~~~~- ) ~ K,Rn< -- - ;p --_ !'\ UZBEKISTAN / KYRGYZ R~EPUBLIC ' O e JYAONIN - DEM > ~ A'~ X I NJIATN G LEs - of. / t>t 0 801 <~ B iIG ?KOREA~ This map was produced BNR < 'EM.REP. OrF by the Map Design Unit ~1AX 'KOREAI o. The World Bank, ~,' The boundaries, colors, ''.,-" -\A*ON / cdenominations andi any 0 N G H A I Yelloss other information sho-n,' ,- e on thNIs map do notf s ' 2H imply, on tbe part of The ~si'lAA4NJ(f ~ \, jO 2 World Bank Group, any \.4Hs ~ judlgmrent on the legal ' ~HJ staitus of any territory, or - -.~ 5 0 any endorsementaor XlI Z A N G test aicceptance of such -30' a :.;ij.,0l 5'5nHtE boundaries. ', -' HUAN*n~;',"~ hn -~~~ A Is, ZHEI~~~S' UAA,G hSee NEPAL 00,,,, f ,e > "' MAP 3 I,AI INX CHINA ' UIN 5 92 NATIONALLY DESIGNATED POOR COUNTIES ,. ~~jTAIWAN NATIONAL DESIGNATED POOR COUNTIES (1 994) AfQdt POOREST OF THE NATIONAL DESIGNATED POOR N A "X1 -''''"",o" ~ ' COUNTIES (~0% OF THE POPULATION HAVING PER * NATIONAL CAPITAL nX2 VIETNA 100 200 300 400 BQ00RILES COUNTY (XIAN) BOUNDARIES 20 )LA PEOPLE'S20 40 60 OMKLMTR PROVINCE BOUNDARIES W" E REP. 7 ~ ltnSe '~ THAILAN I ' < HINANI INTERNATIONAL BOUNDARIES , ' K''t,10PHILIPPI BS OCTOBER 2000 IBRD 30489 70" 80' 90' 100 I to 120' 130I '50' RUSSIAN FEDERATION 00' 1. Loess Plateau f-, I X N 7 > f -1 I Central and eastern Gansu (Dinxi) KAZAKHSTAN ) *" L 1.2 Southern Ningxia (Xihaigu) -W 7 r ) S < -> f ~ ;- H S , 1.3 Northern Shaanxi Loess Hills JI-, r - \> '-.----, IEILONGJIAN 1 4 Luliang Mountains i -2 f > / / ~~~~~~~~~~~~~~~~~~~~HEILONGJIANG I 2. East-west plain, hills and mountain belt UZBEKISTAN M N G L I A 2 1 Bashang sand dunes 2.2 Taihang mountains ) KYRGYEF > L rr i>URU.01 > '\ _H,}- l H iil C HJ5 2 3 Qinba mountains KYRGYZ X?~j," I~ N uRJ aI 2 N G rJ7> ~ _ l - 4 3.1 Wumeng mountains 3 Southwest Karst Mountains X Il N J I_ A N k Ss;\-zt?................. G f 3.1 Wumeng mountains yr ~~~ REP~~~~ ~~ 3.2 Jiuwan Dashan mountains NO KOREA1'" t3.3 Northswest Guangxi mountains 3.4 Southeast Yunnan mountains This rmap ws produced 5\ t 3.5 H d mountains by the Mop DeZign Unit engon of The World Bask. The boundorien, colos, / 4. Eastern hills and mountains denomi-ation- and any G A other information uhosn 4.1 Yimeng mountains on thin mop do not imply,s oN tfh port of The A.2 DH b e5m2untains World Banku Group, any 4NN .3 Jinggang hilly mountains udgmet o_n the legal 4.4 Fujian and Goangdong hilly mountains any endorsement or XZ acceptance of such SIHU nS 5. Qinghai and Tibet high altitude mountains boundaries. Mouse. nusGznou ,' CcJi< 05' - 5.1 Tibet~~~~~~~~~~~~~~~~~~~~~~~~~~~51 Tbe NEPAL CHONG _________________5.2 Qinghai MAP A HUNA WiAN ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~6. Inner and Mongolia and Xinjiang arid areas 6.1 Sandy region of SE Inner Mongolia POOFR MOUZ N TAI N O U S |ffia;g6 2 West Xinjiang arid area POOR MOUNTAINOUS YUN O NATIONAL CAPITAL 0 100 200 300 400 500 MILES O PROVINCE CAPITALS V0 200 400 600 8u0 KILOMETERS PROVINCE BOUNDARIES I s INTERNATIONAL BOUNDARIES 0 LA PEOPLE'S ;$ D'EM. REP. / 0 0 S THAILAND N \,PLIPlS OCTOBER 2000