82122 THE WORLD BANK OPERATIONS EVALUATION DEPARTMENT An Evaluation of World Bank Assistance to China For Poverty Reduction in the 1990s A. R. Khan Director-General, Operations Evaluation: Gregory K. Ingram Director: Ajay Chhibber Manager: R. Kyle Peters Task Manager: Gene Tidrick This paper is available upon request from OED. It was prepared in 2002 as 2005 a background paper for the China Country Assistance Evaluation. The World Bank Washington, D.C. ENHANCING DEVELOPMENT EFFECTIVENESS THROUGH EXCELLENCE AND INDEPENDENCE IN EVALUATION The Operations Evaluation Department (OED) is an independent unit within the World Bank; it reports directly to the Bank’s Board of Executive Directors. OED assesses what works, and what does not; how a borrower plans to run and maintain a project; and the lasting contribution of the Bank to a country’s overall development. The goals of evaluation are to learn from experience, to provide an objective basis for assessing the results of the Bank’s work, and to provide accountability in the achievement of its objectives. It also improves Bank work by identifying and disseminating the lessons learned from experience and by framing recommendations drawn from evaluation findings. OED Working Papers are an informal series to disseminate the findings of work in progress to encourage the exchange of ideas about development effectiveness through evaluation. The findings, interpretations, and conclusions expressed here are those of the author(s) and do not necessarily reflect the views of the Board of Executive Directors of the World Bank or the governments they represent. The World Bank cannot guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply on the part of the World Bank any judgment of the legal status of any territory or the endorsement or acceptance of such boundaries. Contact: Operations Evaluation Department Knowledge and Evaluation Capacity Development (OEDKE) e-mail: eline@worldbank.org Telephone: 202-458-4497 Facsimile: 202-522-3125 http:/www.worldbank.org/oed Acknowledgments The author is grateful for the help that he has received from many members of the World Bank staff in Washington D.C. and Beijing; officials of the Government of the PRC and the Ningxia Provincial Government; and professional colleagues both in and outside China. In particular he would like to acknowledge the help received from Tamar Manuelyan Atinc, who provided help at the preparatory stage and sent comments on an earlier draft; Athar Hussain, who gave valuable suggestions on the draft as a peer reviewer; Alan Piazza, who gave many useful suggestions both during the preparatory stage and on the penultimate draft; Carl Riskin, who generously provided comments, suggestions and background material; and Gene Tidrick, whose invaluable help at every stage made the report possible. Deepak Bhattasali, Shahoua Chen, Qiang Cui, Uma Lele, Martin Ravallion, Geoffrey Read, Supee Teravaninthorn, Wang Sangui, Zhao Renwei, Zhu Ling and several members of the China Country Assistance Evaluation Team not mentioned above provided valuable comments and help. For all remaining errors and inadequacies the author alone is responsible. ACRONYMS CAE Country Assistance Evaluation CPI Consumer Price Index ESW Economic and Sector Work FAO Food and Agriculture Organization GDP Gross Domestic Product MLSS Minimum Living Standard Scheme NBS National Bureau of Statistics NGO Non governmental organization PRC People’s Republic of China PTI Poverty-targeted Intervention TVE Township and Village Enterprises WTO World Trade Organization Table of Contents 1 Introduction................................................................................................................1 2. China’s Performance in Poverty Reduction in the 1990s ..........................................1 3. Why was Poverty Reduction in the 1990s not Faster? ..............................................7 Increase in inter-regional inequality ..........................................................................8 Slow and disequalizing rural income growth.............................................................8 Regressive transfer to households and reduced transfer from rich to poor provinces..................................................................................10 Slow employment growth ..........................................................................................10 Discriminatory treatment of migrants........................................................................11 4. Response by the Government of PRC .......................................................................12 5. Bank Response to China’s Poverty Problem .............................................................16 Economic and sector work.........................................................................................16 Poverty monitoring ....................................................................................................16 Poverty-reduction strategy .........................................................................................17 Analytical research on poverty ..................................................................................18 Issues that remain to be dealt with.............................................................................18 The lending program..................................................................................................19 Are there enough best-practice projects for poverty reduction in the Bank portfolio ............................................................................24 The looming danger ...................................................................................................26 Conclusions................................................................................................................26 6. Future Priorities .........................................................................................................27 Poverty monitoring ....................................................................................................27 Adjustments in the poverty-reduction strategy ..........................................................27 Economic and sector work.........................................................................................28 The lending program..................................................................................................29 Tables 2.1 Selected Indices of Poverty Headcount .....................................................................2 2.2 Income Indicators during the 1990s...........................................................................4 2.3 Comparison of Poverty Using PPP$ Per Day ............................................................4 5.1 Bank Lending in China Since 1990 ..........................................................................21 5.2 Sectoral Composition of Lending ..............................................................................22 Annex 1: Poverty Related Economic and Sector Work since 1990......................................30 1 1. Introduction 1.1 This paper attempts an evaluation of World Bank’s assistance to China for poverty reduction during the 1990s. It begins with a summary of China’s performance in poverty reduction during the decade as a backdrop to the subject (Section II). This is followed, in Section III, by an analysis of the causes why China’s performance in poverty reduction was not faster during this decade of extraordinarily rapid growth. Section IV briefly considers the evolution of China’s strategy for poverty reduction. Section V, the central section of the paper, discusses the role of the Bank’s main instruments - economic and sector work; and lending operations – in helping China alleviate poverty. Section VI considers future priorities for Bank strategy in China to help accelerate the reduction of remaining poverty in the country. 2. China’s Performance in Poverty Reduction in the 1990s 2.1 Estimation of poverty trends in China, as in most developing countries, has to be based on imperfect statistical sources. Arguably, methodological inadequacies underlying the official and most unofficial estimates are greater than they need to be. The income definition used by China’s National Bureau of Statistics (NBS) in its surveys is flawed. Published survey data, which serve as the basis of estimates with minor exceptions, are available only for highly aggregated groups and not for individual households. The sample for urban China systematically excludes the migrant households who do not have resident permits. For the period under review, namely the 1990s, the official cost-of- living indices suffer from methodological biases. On top of these difficulties, the official poverty lines are highly questionable indicators of constant command over goods and services over time. These problems make it very difficult to form a clear view of the change in the incidence of poverty in China. 1 2.2 Ignoring these problems, and concentrating on the best estimates available from World Bank’s own work, what judgment does one form about the trend in the incidence of poverty in China during the 1990s? The main findings may be summarized as follows. 1 These issues are elaborated in section IV. 2 Table 2.1: Selected Indices of Poverty Headcount (Percent of Population below the Poverty Line) Year Rural Urban National (1) (2) (1) (2) 1990 31.3 42.5 1.0 14.0 31.5 1991 31.7 - - - - 1992 30.1 40.6 0.8 7.6 29.6 1993 29.1 40.6 0.7 8.2 29.4 1994 25.9 34.6 0.9 8.5 25.0 1995 21.8 30.8 0.6 5.8 22.0 1996 15.0 24.1 0.5 5.4 17.2 1997 13.5 24.0 1.0 5.4 17.0 1998 11.5 24.1 1.0 5.8 17.1 1999 11.2 24.9 0.5 4.1 17.4 Note: Rural (1) refers to the income-poverty line of 1985 PPP$ 1 (World Bank, China: Overcoming Rural Poverty, March 2001; for 1999 the figure is from unpublished World Bank source). Rural (2) refers to the expenditure-poverty line of 1993 PPP$ 1.08 (Shaohua Chen and Yan Wang, China’s Growth and Poverty Reduction, Trends between 1990 and 1999, July 2001). Urban (1) represents the expenditure-poverty line of 1993 PPP$ 1.08 while urban (2) represents 1.75 times the urban (1) poverty line (both from Chen and Wang, op. cit.). National estimate represents the expenditure-poverty line of 1993 PPP$ 1.08 (Chen and Wang, op. cit.). 2.3 Over the decade as a whole, there was a substantial reduction in poverty however one chooses to measure it. Thus the proportion of population below PPP$ 1 expenditure per day declined from 31.5 percent in 1990 to 17.4 percent in 1999. The decline in the proportion of population below PPP$ 1 income per day may have been greater. 2 2.4 Almost the entire reduction in the incidence of poverty took place in 1994, 1995 and 1996 – the second of the three three-year sub-periods into which the decade might be divided - when an expenditure-poverty threshold is used. There was very little reduction in the incidence of poverty between 1990 and 1993, the first sub-period. Between 1996 and 1999, the third sub-period, the proportion of population in poverty increased. Using an income-poverty threshold for rural China, one gets a less extreme difference in the rates of poverty reduction during the three sub-periods of the decade. However, the broad pattern still holds: the rate of poverty reduction is rapid during 1994-96 and far slower during the earlier and later years. 3 2 Bank’s income-poverty estimates are available for rural China only and these show a faster rate of poverty reduction than do the expenditure-poverty estimates (cols. 1 and 2 of Table 1). 3 At the time of preparing this background paper, World Bank estimates of poverty in available ESW reports did not go beyond 1999, which is the terminal year of analysis in this section. Unpublished World Bank estimates and work-in-progress show a decline in poverty in 2000. While this may indicate the 3 2.5 For the same poverty threshold (e.g., PPP$ 1 per day), and without a correction for cost-of-living difference between rural and urban areas, the incidence of urban poverty is minuscule in comparison with the incidence of rural poverty. The rate of reduction in urban poverty has however been slow throughout the period and, in particular, there has been no trend reduction in the second half of the decade. 2.6 Even after the substantial reduction in poverty brought about by two decades of rapid growth, the overall incidence of poverty at the end of the 1990s was higher in China than in many countries at the same average level of living. 2.7 A perusal of the above raises a number of questions concerning China’s performance in poverty reduction during the 1990s. Consider first the fact that there was very little poverty reduction in the early 1990s, the first sub-period, and in the late 1990s, the third sub-period (when there was an increase in expenditure poverty and a slow reduction in income poverty). During both these sub-periods China experienced extremely high rates of GDP growth which coincided with very rapid increase in per capita real personal income in each sub-period in both rural and urban China (Table 2.1). 4 That both these sub-periods witnessed slow or negative decline in poverty indicates that China failed to realize the potential for poverty reduction that was created by rapid growth due to a sharp increase in the inequality of distribution that adversely affected the relative welfare of the poor. 2.8 The magnitude of this failure is blurred by the divergent trends in poverty based on different methods of estimation. But, as Chen and Wang 2001 sharply brings out, “[I]n China, the poor benefited far less than the rich from growth”, whatever indicator one uses to measure the change in welfare of different income groups. For the decade as a whole, the rate of income growth for different income groups was strongly inversely correlated with the level of their per capita income. 2.9 Next, consider how China’s average incidence of poverty at the end of the 1990s, after two decades of poverty reduction, compares with that in countries with comparable average level of living, i.e., similar per capita income in PPP$. Table 2.3 compiles the relevant information for all 12 countries with per capita PPP$ income between 3,000 and 4,500 in 1999 for which poverty estimates for PPP$ 1 poverty line are available. In addition it shows the data for Indonesia, a large country with a significantly lower real income. Of all these countries, China has by far the highest proportion of population in poverty with the exception of El Salvador, Paraguay and Turkmenistan. 5 The first two of beginning of another, more favorable, cycle in poverty trend, this does not invalidate the points made above. 4 The annual average growth rate in per capita personal income in rural China during the third sub-period was slightly lower than 5 percent. The annual growth rate in per capita urban personal income in both first and third sub-periods and the annual growth rate in per capita rural income in the first sub-period exceeded 5 percent, in some cases by a big margin. 5 The poverty lines are based on household consumption, but in China household saving is high, even in low-income households. Using income rather than consumption, only 9.8 percent of the population would have been below the poverty line in 1999. 4 these exceptions are highly unequal, small Latin American countries while the third is a small former Soviet republic in transition. Table 2.2: Income Indicators during the 1990s (Percent change over the Preceding Year) Year GDP and Component Sectors Per Capita Real Personal Income Total Primary Secondary Tertiary Rural Urban 1990 3.8 7.3 3.2 2.3 9.2 8.4 1991 9.2 2.4 13.9 8.8 0.9 7.2 1992 14.2 4.7 21.2 12.4 5.7 9.7 1993 13.5 4.7 19.9 10.7 3.4 9.5 1994 12.6 4.0 18.4 9.6 7.4 8.5 1995 10.5 5.0 13.9 8.4 9.9 4.9 1996 9.6 5.1 12.1 7.9 13.1 3.8 1997 8.8 3.5 10.8 8.2 5.9 3.4 1998 7.8 3.5 8.9 8.3 4.5 5.8 1999 7.1 2.8 8.1 7.5 3.8 9.3 Note: Growth rates in GDP and component sectors are from World Bank, Overcoming Rural Poverty, March 2001, for the period until 1997 and estimated from the data in National Bureau of Statistics, China Statistical Yearbook 2000 thereafter. Rural Per Capita Real Personal Income is estimated by deflating the “Per Capita Annual Net Income of Rural Households” by “General Consumer Price Index for Rural Areas”, both shown in China Statistical Yearbook 2000. Urban Per Capita Real Personal Income is estimated by deflating “Per Capita Annual Disposable Income of Urban Households” by “General Consumer Price Index for Urban Areas”, both shown in China Statistical Yearbook 2000. Table 2.3: Comparison of Poverty Using PPP$ Per Day Country Per Capita PPP$ Percent in Poverty El Salvador 4,260 26.9 Turkemenistan 3,340 20.9 Paraguay 4,380 19.5 China 3,550 18.5 Peru 4,480 15.5 Guatemala 3,630 10.0 Indonesia 2,660 7.7 Sri Lanka 3,230 6.6 Jamaica 3,390 3.2 Egypt 3,460 3.1 Ukraine 3,360 2.9 Jordan 3,880 <2.0 Morocco 3,320 <2.0 Source: World Development Indicators 2002. 2.10 Consider next the view that China does not have a significant problem of urban poverty, a view which is not only held by the official Chinese policy makers but might 5 also be attributed to the Bank’s most recent poverty assessment of China. 6 The ostensible reason for this is that a negligible proportion of the urban population is below the PPP$ 1 poverty line. 7 But serious questions have been raised about the appropriateness of this measure as an acceptable living standard for urban China for which, in official discussions, a higher poverty line is commonly used. Perhaps the most careful study of urban poverty in China, prepared for the Asian Development Bank and based on far more detailed data than the estimates in Table 2.1 above are based on, finds that in 1998 the incidence of absolute poverty, for a rather stringent threshold, was quite significant: 4.7 percent for income measure and 11.9 percent for expenditure measure.8 It finds that small increases in the poverty line lead to large increases in the proportion of population in poverty. The study also finds that the incidence of poverty among the migrant urban population, who are left out of the regular NBS surveys, is fifty percent higher than among the registered urban population. The very fact that these migrants have been increasing steadily as a proportion of registered urban population indicates that the trend in urban poverty discernible from estimates reported in Table 1 and elsewhere, which are based on the NSB survey data, are biased downward. 2.11 Is China’s continued high incidence of poverty, according to the above indicators, due to the use of too high a poverty line? There is a view that the equivalent of PPP$ 1 is too high a standard of absolute poverty for rural China. Indeed this is the view underlying the official poverty line which in recent years has been about 70 percent of the equivalent of PPP$ 1. First, it should be noted that this is unlikely to alter any of the conclusions about the rate of change in poverty over time. While the official Chinese estimates of rural poverty are lower than the Bank’s, they actually show a lower rate of reduction than do the Bank’s estimates. 9 Secondly, by itself a lower poverty threshold is unlikely to 6 World Bank, China: Overcoming Rural Poverty, March 2001. The exclusive pre-occupation of this report – the only comprehensive poverty assessment for the 1990s - with rural poverty implies a presumption on its part that urban poverty is not a problem in China especially in view of an absence of any ESW on the urban poor. 7 It is useful to be clear about the procedure followed in making the PPP$ poverty estimates. The objective is to make poverty lines internationally comparable. The national average conversion factor between domestic currency value and PPP$ value of consumption is used to arrive at the domestic currency equivalent of PPP$ 1 (more recently PPP$ 1.08 which is estimated to be the approximate equivalent of the “national”, nutrition-based, poverty line for a set of ten poor countries for which direct PPP$ national accounts estimates are available) for some benchmark year (1993 in recent estimates). There is however no mention in Chen and Wang, 2001 that adjustment between urban and rural poverty lines is made to reflect cost-of-living difference. Indeed the procedure does not make any adjustment for the difference between the PPP$ conversion factor for the expenditure bundles of different income/expenditure groups. The method uses the average adjustment factor for all income/expenditure groups. It is highly doubtful that the average PPP conversion factor is the same as the one for the expenditure bundle of the poor. It has also been reported that recent research at the NSB has shown that the PPP$ conversion rate used by the Bank diverges by a wide margin from the rate estimated by using recent price data. 8 The study – Final Report: Urban Poverty in PRC, TAR:PRC 33448, prepared for the ADB by Athar Hussain and others – is yet to be published. It uses the household level data from the 1998 Urban Household Survey by the NSB. It calculates the food-poverty line for each of the 31 provinces by using the consumption pattern of the poorest quintile of population to obtain the “cost efficient” value of 2100 kilocalories and, to arrive at the poverty line, adds to it the proportion of non-food expenditure of the typical household that barely attains the minimum food consumption. 9 Official Chinese estimates of rural poverty, based on income measure, fell from 9.5 percent in 1990 to 4.6 percent in 1998. As col. 1 of Table 1 shows, the World Bank’s income poverty estimate for rural China 6 change the conclusion about the comparison of poverty in China with the poverty in other countries because an equivalent PPP poverty line would have to be used for other countries for such a comparison. But the most telling argument against using a lower poverty line for the purpose of public policy is perhaps based on the findings of a 1997 survey of six poor counties in southwest, northwest and central China by the Institute of Agricultural Economics of the Chinese Academy of Agricultural Sciences which shows that 50 percent of the people were below 2,100 kilocalories of food consumption per day whereas only 33 percent were below the official poverty line (which, as noted above, was approximately 70 percent of the equivalent of PPP$ 1).10 This finding is in conformity with the view emerging out of numerous studies that a significantly lower poverty line than the equivalent of PPP$ 1 would mean that many above the poverty threshold would suffer from nutritional inadequacy of the most basic kind.11 2.12 If broader indicators are considered, one encounters evidence of the persistence of substantial deprivation and a significant worsening of some indicators of deprivation. According to a survey by the NBS, among the 170,000 administrative villages in designated poor counties, 20 percent do not have roads of acceptable standard, 8 percent have no electricity, two-thirds have no telephone and tap water, 14 percent are without primary schools and one-fourth are without clinics or doctors. 12 FAO estimated that 11 percent of the population consumed less than 1920 kcal/day during 1996-98 13 and an independent study based on a detailed survey established that in 1995 17 percent of the rural population consumed less than 2100 kcal/day. 14 National surveys by the Health Ministry have found a 22 percent stunting rate among China’s children in 1998 and much higher rates in poor counties. 15 These indicators of deprivation are larger than the official estimates of income poverty in China. There is evidence that some of these broader indicators have become worse over time, especially in poor areas. For example, an influential World Bank study reports that the proportion of population having access to government health facilities declined from 71 percent in 1981 to only 21 percent in 1993. 16 Another study estimates that 25 million Chinese fell below the official poverty line in 1998 because of the costs of “catastrophic medical expenses”. 17 fell over the same period from 31.3 percent to 11.5 percent. Also note that Chen and Wang, 2001 shows that for the poverty line of PPP$ 0.75 the trends during the three sub-periods are exactly the same as for the poverty line of PPP$ 1.08. 10 These results were provided to the author by Dr. Wang Sangui, the Director of the Division of Poverty and Development of the Institute, from the unpublished report of the survey. 11 For an account of the literature arguing this view, see Albert Park and Wang Sangui, “China’s Poverty Statistics”, in China Economic Review 12 (2001). 12 This is reported in the forthcoming UNDP report, Policies for Poverty Reduction in China. 13 FAO, The State of Food Insecurity in the World, Rome, 2000. 14 Zhu Ling, “Food Security of Low-Income Groups in Rural China”, in Riskin, Zhao and Li (Eds.), China’s Retreat from Equality, M. E. Sharpe, 2001. 15 This is reported in A. Park and S. Wang, “China’s Poverty Statistics”, China Economic Review 12, 2001. 16 World Bank, China 2020: Development Challenges in the New Century, 1997, p. 25. 17 Story by Martin Fackler in AP Worldservice, December 31, 2001. See also Zhu Ling, “Rural Health Care Critical” China Daily, September 17, 2002. 7 2.13 Any conclusion based on the evidence cited above must be qualified because of the general uncertainty about the quality of data that underlie the evidence. Time-series data on household income and expenditure need to be subjected to much careful scrutiny before resolving the conflict among different indicators, e.g., estimates of poverty based on income and expenditure thresholds. Inter-country comparison of population below PPP$ 1 is similarly subject to many inaccuracies such as those underlying the estimates of PPP$ conversion rates (see footnote 7 above). Mindful of these qualifications, one might note the following broad features of China’s performance in poverty reduction during the 1990s. On the one hand, China’s success in pulling scores of millions out of poverty stands in sharp contrast to the increase in the number of poor in the rest of the developing world taken together. 18 All available indicators suggest that at the end of the decade the incidence of absolute poverty was substantially lower than at its beginning. Going by this indicator, any developing country would be justifiably proud of such performance. On the other hand, some of the evidence cited above raises questions concerning certain aspects of China’s performance in poverty reduction. Most of the indices of income and expenditure poverty suggest that, during early and late 1990s in particular, the reduction in poverty in China may have fallen short of the opportunity for poverty reduction that was created by its astounding growth record. International comparison, based on admittedly imperfect data, suggest the possibility that, for all its success in rapidly reducing poverty over the last two decades, China may still have at the end of the 1990s a higher incidence of poverty than many major countries at comparable living standard. Finally, there are reasons to fear that official Chinese strategy for poverty reduction has underestimated both the magnitude and the persistence of the problem of urban poverty. As noted above, all these evidence need further careful scrutiny before a definitive account of China’s poverty performance during the decade can be put together. And yet the mixed evidence on performance calls for an analysis of the reasons why China’s remarkable growth during the 1990s was not accompanied by a faster rate of poverty reduction. 3. Why was Poverty Reduction in the 1990s Not Faster? 3.1 Simulation exercises have shown that, with unchanged distribution of income, the rate of reduction in poverty in China during the period since the mid 1980s would have been far greater than the actual rate of poverty reduction. Some increase in inequality during the reform period was inevitable, even desirable. For example, the artificial and arbitrary egalitarianism in the structure of urban wages that reforming China inherited from the past was a serious obstacle to efficiency and incentives. An increase in inequality due to a more rational structure of wages, reflecting differences in productivity, would contribute to an increase in inequality though it is probable that, by increasing efficiency and growth, it would offset its effect on poverty. There are, however, many other sources of increase in inequality which can not be claimed to have contributed to efficiency and growth. 18 According to World Bank staff calculations, the number of people below the poverty line of PPP$1 per day, fell by 8 million worldwide between 1987 and 1998, but their number outside of China actually increased by 82 million. 8 Increase in inter-regional inequality 3.2 China’s economic growth has been regionally concentrated in the coastal and eastern provinces. In comparison, the central and western regions have grown at a much slower rate. 19 While the eastern provinces had certain natural advantages in attracting the location of export industries and foreign direct investment in the period of China’s rapid integration with the global economy since the mid 1980s, the difference in the growth performance between the East and the rest of the country was at least partly an outcome of discriminatory public policy. Special privileges designed to attract foreign direct investment were long limited to the coastal region. The disadvantage of the backward regions in competing for investment resources was exacerbated by arbitrary administrative strictures. Artificial depression of producers’ prices for grain in the past was detrimental to the growth of the rural areas in certain provinces in the central and western regions which had a comparative advantage in grain production. The pricing of natural resources proved more detrimental to poor provinces than to others. An example is the extraction of the vast quantities of coal, using up scarce fresh water, out of poor areas of Shanxi at extremely low ex-factory price. The mountainous areas of the poor provinces in particular suffered from the serious inadequacy of infrastructure and access to resources. Poverty in China is heavily concentrated in the poor central and western provinces, especially in the remote mountainous areas. Only after the mid 1990s signs of a reversal of some of these policies became discernible in China. Slow and disequalizing rural income growth 3.3 Poverty in China is overwhelmingly concentrated in rural areas both because they still accounted for nearly 70 percent of the population at the end of the 1990s and because of the very large difference between the average urban and the average rural incomes. Based on the imperfect NBS data, the ratio of per capita urban household income to per capita rural household income increased from 2.20 to 2.65 between 1990 and 1999. This widening disparity was largely due to the slow growth of agriculture and the discriminatory terms of trade for agriculture over much of the period. A major reason why China’s growth in the years immediately following the launching of reforms was highly poverty alleviating was that this period witnessed a rapid reduction in the disparity between urban and rural incomes: between 1978 and 1985 the ratio of per capita urban household income to per capita rural household income declined, again based on the NBS data, from 2.57 to 1.86. This was brought about by improved producers’ prices for agricultural goods and an increased flow of resources into agriculture. Available NBS estimates suggest that the terms of trade for agriculture started deteriorating after 1988. 20 19 According to official estimates, the eleven coastal provinces and Beijing on the average grew at 78 percent faster rate than did the western and central provinces during the 1990s. The absolute growth rate of GDP of the western and central provinces was still high by international standard. But the growth in personal income substantially lagged behind the growth in GDP everywhere in China and the inequality in the distribution of income and consumption increased over time. Together these effects seriously inhibited the ability of the western and central provinces to reduce poverty. 20 The terms of trade, referred to here, is the ratio of the “General Purchasing Price Index of Farm Products” to the “General Rural Retail Price Index of Industrial Products”. Both these indices are shown in NBS, China Statistical Yearbook 2000, p. 290. The index of terms of trade for agriculture, thus calculated, had the following values for selected years: 1978: 100; 1988: 177 (steady increase in the intervening 9 Between 1994 and 1996 there was a temporary improvement in agriculture’s terms of trade. After 1996 it started declining once again. At least in the period before the mid 1990s the unfavorable terms of trade for agriculture were not the outcome of market forces; these were largely a consequence of public policy of controlling the producers’ price of grains. That an avoidance of this policy would have resulted in a favorable outcome for the poor is illustrated by the events of the years 1994-96 when producers’ price for grains was revised upwards. There was a sharp increase in the personal income of the rural population and a large decline in the incidence of rural and overall poverty in those years (Tables 2 and 1). 3.4 The widening of urban-rural inequality after the mid 1980s was also fuelled by the reduction of public investment in agriculture, as compared to the growing need for public investment in order to overcome emerging constraints to agricultural growth, and by the disequalizing effect of the fiscal system and transfer payment (see below). Relative stagnation of agriculture and the failure of agricultural earnings to rise also meant that over time an increasing proportion of rural household income came to be derived from non-agricultural activities. While the equality of access to land has ensured a strongly egalitarian distribution of earnings from agriculture, rural household income from non- agricultural sources has been more unequally distributed because of the regional inequality in their distribution and the failure of employment in these activities to increase as rapidly as in the past, issues that are discussed more fully later. A final obstacle to improved living standard for agricultural households was the limitation of emigration out of rural areas (see below). 3.5 The failure to maintain improved terms of trade for agriculture after 1996 seems to be a complex phenomenon. During this period the preparatory reform for WTO accession appears to have limited the sustainability of high producers’ price for farm products which had been introduced in 1994. It appears that the government had hoped that the official monopoly of procurement by the grain bureau would enable the latter to pass on the high procurement price to the buyers. This proved impossible due to the de facto operation of market forces in grain trade. As a result the grain bureau ended up with huge debts. 21 The response of the government was to sharply reduce the volume of grain purchased at the guaranteed high price. The farmers were forced to sell the large remainder of the marketed output in the open market where price was far lower than the official support price. Currently the government is reforming the system of grain purchase by abandoning the official monopoly of the grain bureau. As an outcome of these consequences, the weighted average producers’ price fell below the official support price. With the exception of rice, domestic prices of grains are currently the same as or higher than the world price. It therefore seems that improving agriculture’s net barter terms of trade by raising producers’ price is not a sustainable policy. Some possibility of reducing input prices may exist in so far as the domestic price of fertilizer appears to be higher than the world price. Future improvement in farm income must however be based period, the index peaking in 1988); 1993: 154 (steady fall between 1988 and 1993, the year in which the index bottomed out); 1994: 184; 1995: 192 (sharp rise for two years which was followed by only a small decline in 1996); 1999: 151 (steady fall since 1996). 21 It is widely believed that the loss sustained by the grain bureau was also partly due to corruption in trading. 10 on improved single factoral terms of trade for agriculture by way of increased labor productivity. The success of this critically depends on the liberalization of the labor market so that agriculture can shed its surplus labor for employment elsewhere (see below). Regressive transfer to households and reduced transfer from rich to poor provinces 3.6 China’s fiscal system substantially contributed to the growing inequality and exacerbated the problem of poverty. First, the system of taxation and transfers by the state and the collectives is highly regressive. A detailed study based on the analysis of household level data for 1995 established that an average household in rural areas was subject to the payment of a “net tax”, amounting to 0.5 percent of income, to the state and the collective while an average urban household received a “net subsidy” of 11 percent of income. Within the rural and urban areas these taxes and subsidies were highly regressively distributed. Per capita net tax paid by the extreme poor households in rural areas was 36 times higher than the per capita net tax paid by the non-poor households. In urban China transfers accounted for 5.36 percent of the income of the extreme poor households as compared to 11.14 percent for the non-poor. 22 The regressive fiscal system is perhaps the most outstanding anomaly of China’s poverty alleviation strategy. It undoubtedly offsets the positive results of much targeted intervention for poverty reduction in China. 3.7 Secondly, the overall budgetary system in China has undergone changes that circumscribe the ability of the poor regions to finance the delivery of essential services that are critical for the welfare of the poor. In the past, fiscal surpluses (deficits) of provinces were transferred to (met by) the central government which meant a large transfer from rich provinces to poor provinces. By the early 1990s China’s state budget had come to be starved of revenue sources. Surpluses of rich provinces had become very small and deficits of poor provinces were forced to decline, thereby sharply reducing the magnitude of transfer between rich and poor provinces. Many poor provinces had to curtail basic services, such as, education and health, and reduce investment in economic development. The provinces, not the central government, pay the external lenders like the World Bank and bear foreign exchange risks for loans for many social-sector projects. As a consequence, the extremely poor counties are often not eligible to participate in the externally-funded poverty-reduction projects. Slow employment growth 3.8 Growth in employment is the most effective way of reducing poverty. Poverty- reduction effect of economic growth is larger the greater the labor intensity of growth. During the 1990s, economic growth in China created very little employment, thereby causing the breakdown of the conduit through which the benefit of growth is passed on to the poor whose principal resource consists of labor. The reason for this is complex. It has been alleged that strong biases against credit availability for small and private firms 22 See A. R. Khan and C. Riskin, Inequality and Poverty in China in the Age of Globalization, Oxford University Press, 2001, pp. 133-35. 11 and government biases in favor of allocating credit and resources to large, capital- intensive enterprises have contributed to slow employment growth in some cases. Even so, there is little evidence to suggest that the industries in which growth was concentrated were, on the whole, not adequately labor intensive. The problem was that in the past China’s state and collective enterprises, which accounted for an overwhelming proportion of total industrial employment, resorted to the creation of employment in excess of efficient labor requirement. Indeed this system of concealed unemployment insurance was the principal instrument for social protection. As these industries opened up to competition from industries under other forms of ownership and to global competition, this concealed system of unemployment insurance became unviable. As a result state and collective sectors began shedding labor. This phenomenon was particularly strong in urban China where, between 1995 and 1999, state and collective enterprises together reduced their employment by a total of 41.24 million or 29 percent of employment in all such enterprises in 1995. Urban unemployment increased rapidly and became a major reason for the persistence of urban poverty. The phenomenon was not limited to urban China. Township and Village Enterprises (TVEs), which were a major source of output growth, had sharply reduced their labor absorption since the early 1990s. Since 1996 they too started recording a net reduction in employment. Between 1996 and 1999 employment in TVEs declined by 6 percent. 23 3.9 China has so far not been able to replace the past system of concealed unemployment insurance in state and collective enterprises by an alternative, transparent and comprehensive system of social protection. As a consequence, a high proportion of the laid-off workers entered the category of the poor. Discriminatory treatment of migrants 3.10 As noted in the previous section, the incidence of poverty among the urban migrants is about 50 percent higher than among the registered urban residents. A good part of this is due to discriminatory treatment of the migrants who have been denied access to health and education services and the opportunity for seeking employment on equal ground as the registered urban residents. Restriction on migration has also prevented an outflow of labor from rural areas in response to market forces and has contributed to the persistence of high urban-rural inequality. 3.11 The above five sets of circumstances probably account for at least the entire increase in inequality during the 1990s. It is probable that if these adverse factors were completely offset, China’s income distribution would have become more equal over the decade, thereby directing more than the entire benefit of economic growth to poverty reduction. By usual standards, poverty in China would have been a part of history. The validity of this claim can not be completely demonstrated; but its plausibility can be illustrated by the following example. It has been demonstrated that if the distribution of net urban subsidies in 1995 were merely equal for all households on a per capita basis – a modest objective in comparison with the goal of redistribution that most societies set for 23 All the data in this paragraph are from NBS, China Statistical Yearbook, several recent issues. 12 such programs – the increase in the Gini index of urban income distribution between 1988 and 1995 would have been less than half of what it actually was! 24 3.12 While reform in the areas indicated above are desirable in any case and would help accelerate the rate of reduction in poverty in China, it is possible that the steep fall in the incidence of poverty in the 1980s lowered the possibility of reducing poverty in the 1990s when much of the remaining poverty came to be heavily concentrated in the poor provinces and regions with inadequate infrastructure and commercial contact with the rest of the economy. This may have resulted in “diminishing returns” to China’s poverty reduction effort. 4. Response by the Government of PRC 4.1 The evolution of the problem of poverty and the public response to it in China may be divided into four distinct time periods. The first period, from the late 1970s to approximately the middle of the 1980s, was the period of the most rapid reduction of poverty. This was also a period without a distinct poverty-reduction strategy on the part of the government. Growth strategy itself - based on agriculture as the leading sector - was by and large free of most, though perhaps not all, of the five sets of problems outlined in the preceding section. 4.2 The second period may be said to have begun around the mid 1980s and continued until 1994. This was the period when China’s development underwent a major change as its focus shifted from agricultural growth to export-led growth within the framework of rapid integration with the global economy. This is the period when the five sets of disequalizing forces gradually became prominent. During this period China began responding to these trends by setting up the Leading Group for Poverty Alleviation and starting to identify poor counties that require help. While the hope that growth would by and large automatically alleviate poverty in most of China persisted, the need for targeted assistance for the ecologically-disadvantaged poor areas began to be gradually recognized. 4.3 The third period began with the formulation of a stronger and well-articulated strategy with the commitment of much greater resources for direct intervention for poverty reduction. In January 1994, the State Council put into effect the National 8-7 Plan for poverty reduction, with the objective of lifting the then-officially estimated 80 million rural poor out of poverty in the remaining seven years of the century. 25 The important features of the Plan were as follows. 4.4 The Plan identified poverty as an exclusively rural problem. Poverty in urban China was considered to be marginal in magnitude, which required no specifically designed action for its eradication. The strategy identified 592 rural counties as poor, mainly on the basis of their per capita income. It focused attention on improving the 24 Khan and Riskin, Op. cit., pp. 134-35. 25 The figure 8 in the 8-7 program is in units of qianwan, or ten million, which is a common numeric unit in China. 13 average income of these poor counties through a package of microeconomic interventions, and not on targeted support to improve the capability of the poor households themselves to overcome poverty. A county would be deemed to have overcome poverty when its per capita income exceeded Y500 at 1990 prices and it also achieved certain goals for infrastructure development, education, training, sanitation and family planning. 4.5 The state provided funds for poverty reduction activities through three main channels: the Ministry of Finance, the State Planning Commission, and the Agricultural Development Bank. The latter provided the largest part of funds in the form of low- interest loans for a variety of production purposes. The State Planning Commission funds were used for Food for Work programs for the construction of infrastructure and drinking water projects. Funds appropriated through the Finance Ministry were used for training purposes. Government departments, NGOs and international agencies were encouraged to supplement state funds by allocating a part of their own budgets to help the development of these poor counties. 4.6 Official sources have claimed that the number of poor in the designated counties has declined sharply as a result of these interventions. It is however hard to judge the validity of this claim since no systematic evaluation of these programs is available. A number of inadequacies of this strategy is nevertheless obvious and needs to be brought out. 4.7 First, quite apart from the question whether the programs succeeded in bringing down the incidence of poverty in the designated counties, there are large groups of poor who are completely left out of their scope. These groups consist of the rural poor who live outside the designated districts; 26 and the poor who live in urban areas. Official NBS estimates put the rural poor outside the designated counties at a third of all rural poor while certain other estimates put the number at half. 27 At the time the National 8-7 Plan was adopted, the problem of laid-off workers was still not perceived to be a major phenomenon with adverse implications for the welfare of the poor. This however was no longer the case in the late 1990s. 4.8 A second limitation of the strategy is that it is focused on enhancing the average welfare of the designated poor counties. The programs are not specifically targeted to improve the welfare of the poor within the counties. Indeed there are elements of the program that would benefit the poor more than the others; but there is the distinct possibility that a large proportion of resources leaked out to the non-poor. 28 The criterion of success – raising the average of the county above a certain threshold – does not consider the actual extent of poverty reduction within the county, and it is hard to know the extent to which the success claimed glossed over this distinction. 26 As discussed later, Bank’s most recent poverty assessment makes this point. 27 See Carl Riskin et al., Rural Poverty Alleviation in China: An Assessment and Recommendations, report prepared for the UNDP, Beijing, 1996. 28 This point is also made in Bank’s most recent poverty assessment. See later. 14 4.9 The third major problem of the strategy is that its isolation from the overall development strategy makes it impossible to judge the extent to which the two have operated at cross purposes. For example, an adverse performance with respect to some of the five phenomena discussed in the preceding section could easily outweigh the positive effect of a lot of resources targeted to the designated counties. 4.10 The fourth period of China’s public response to poverty may be said to have started around the turn of the millennium, at least in partial recognition of some of the above problems. A “National Poverty Reduction Conference” was held in June 1999 and an “International Conference on China’s Poverty Reduction Strategies in the Early 21st Century” was held in May 2000 under the joint auspices of the government of the PRC, the World Bank, the United Nations Development Programme and the Asian Development Bank. Following these deliberations, the government issued a new program in October 2001 called the “Outline for Poverty Alleviation and Development of China’s Rural Areas (2001-2010)”, hereafter referred to as the Ten-Year Plan. This plan emphasizes agriculture and farm production; the provision of education and training to the poor; the use of science and technology to promote the productivity of the poor; and the facilitation of out-migration and voluntary resettlement of people from ecologically disadvantaged areas. 4.11 In February 2000 the State Council adopted the “Great Western Development Strategy” which initiated a new approach for the promotion of economic development in all the western provinces as well as the relatively poor provinces in the central region. The Leading Group for Western Region Development was established, with the Prime Minister as the Chair, and an executive body, the Western Region Development Office, with the State Development Planning Commissioner as the Chair. The program has led to a large increase in investment in infrastructure development in this region. While the strategy is not directly focused on poverty reduction in the western region, it is hoped that improved growth in these poor provinces would benefit the poor who constitute a higher proportion of population in this region than in the nation as a whole. 4.12 A third element of the redirection of poverty reduction strategy in the late 1990s is the adoption of a program for the protection of the urban poor. Three instruments designed for this purpose are: (a) a living allowance for laid-off workers, which is the largest, though a transitional, program; (b) unemployment insurance, which has been replacing the transitional living allowance for the laid-off workers; and (c) the Minimum Living Standard Scheme (MLSS), which is a subsistence allowance paid out of the general revenue of the government. In principle, these measures could amount to a comprehensive system of protection from poverty for the urban poor. They have not so far succeeded in achieving this goal because of inadequate resources allocated to these programs. As a result not sufficiently high a proportion of the eligible population have been identified for support; only a small proportion of those identified for support have received assistance; and the assistance has been smaller than what is needed to lift the recipients above poverty. 29 29 For details see, ADB, Final Report Urban Poverty in PRC, TAE:PRC 33448, cited above. 15 4.13 One important precondition for sound poverty reduction is an accurate system of poverty monitoring. Official poverty estimates, which serve as the basis of public policy, continue to suffer from serious deficiencies, despite some recent improvements. Arguably they do not make full use of the data that the NBS collects through its annual surveys. Some of the more serious deficiencies of the official system of poverty monitoring are as follows: (a) The urban household surveys suffer from the failure to capture the “floating migrants”. (b) The definition of income (and expenditure) that is employed in the NBS surveys excludes rental value of owned housing and has an incomplete coverage of non-wage income payments (including welfare payments and subsidies). Since these elements have systematically varied over time as a proportion of household income, their neglect creates a bias in the estimates of change in poverty. (c) Income is universally used to specify the poverty line although (because?) work within the NBS has shown that the use of expenditure to specify poverty line results in a higher rate of poverty. The use of both income and expenditure measures would have been enlightening for a clearer understanding of poverty trends. 30 (d) The poverty line is unrealistically low, a point that has been repeatedly emphasized by the Bank’s analytical reports. (e) At least for the urban poverty line the NBS uses the consumer price index (CPI) to make adjustment over time. The official CPIs in China have been Paasche indices which not only are likely to understate the actual increase in cost of living, but their use in updating poverty line renders the latter devoid of any meaningful interpretation unless the base year of the index coincides with the base year of the poverty line (which is not the case). 31 (f) The rural poverty line – the one that really matters because the rural poor constitute a very high proportion of all poor – does not even make a complete adjustment for the change in official CPI. 32 30 Albert Park and Sangui Wang, op.cit., reports that since 1998 the NBS began using a dual criteria by classifying a household as poor if its income falls below the poverty line and its consumption is less than 1.5 times the poverty line or if its consumption is below the poverty line and income is less than 1.5 times the poverty line. It is unclear what the consequence of this break in methodology has been for the estimated change in poverty over time. 31 Since 2001 the NBS has reportedly changed the method of the CPI from Paasche to Laspeyre index. 32 A. Park and S. Wang, in their paper “China’s Poverty Statistics”, in China Economic Review 12 (2001), show that “if the 1985 line is inflated by the rural CPI, the 2000 poverty line reaches 721 compared to the official 625”. The numbers refer to current yuan. 16 5. Bank Response to China’s Poverty Problem 5.1 What has been the effect of Bank operations in China on the poverty outcome? How did the Bank respond to the trends and circumstances with respect to the poor described above? This section makes an attempt to address these issues by considering the poverty focus of the Bank’s operational instruments, namely the economic and sector work (ESW) and lending operations. Economic and sector work 5.2 During the 1990s, Bank’s poverty-focused ESW consisted of two detailed poverty assessment reports; two other reports/papers dealing with trends and analysis of poverty; and at least seven analytical papers dealing with specific poverty issues, and a variety of additional reports which contain significant discussion of poverty issues although their focus may be broader (Annex 1). Together these reports and papers have dealt with a wide range of issues concerning poverty monitoring and policy design for poverty reduction. Poverty monitoring 5.3 As in so many other countries and in so many areas of measurement, Bank’s ESW work has had considerable influence in modernizing China’s poverty estimates. Together, these reports and papers have introduced the standard indices of poverty measurement (supplementing the headcount by proportionate poverty gap indices and, sometimes, by the squared poverty gap indices). They emphasized the importance of using the expenditure-based poverty indicators, in addition to the income-based indicators that are nearly universally used in China. They introduced the notion of variable poverty lines to test the sensitivity of poverty measurements to changes in the poverty threshold. In doing so, they emphasized that the official poverty line used in China is unrealistically low. Bank’s ESW also introduced methods of adjustment in the poverty line for change in cost of living over time that are, in principle, superior to the method used in official estimates. These are highly important contributions which have exerted influence beyond the officialdom in China and elevated the debate and discussion of poverty issues to a higher level than would otherwise be the case. 5.4 There are however significant areas in which further progress can be made. The first of these relates to the data on the basis of which poverty estimates are made. As noted in Bank’s 1992 Poverty Assessment, China has established a good household survey which is implemented annually. Unfortunately, with rare exceptions, the data from these surveys that have been available to Bank’s ESW are grouped for a handful of income categories. Household-level records have not been available to Bank’s ESW. This has seriously limited the quality and accuracy of analysis. Secondly, as discussed in the preceding section, there is an urgent need to improve the definitions, concepts and methods of estimation that underlie China’s official poverty monitoring system by finding a way to transfer the Bank’s great experience in this area to the NBS through an appropriate lending or technical assistance instrument. 17 Poverty-reduction strategy 5.5 The 1992 Poverty Assessment – China, Strategies for Reducing Poverty in the 1990s - spelled out a comprehensive strategy for poverty reduction for China in the 1990s. It recognized that poverty reduction had slowed down since 1985 and attributed it to “the absence of meaningful levels of agricultural growth and rural enterprise development in the upland areas” (p. xi). In retrospect, its justification of this – “the quick reductions of poverty through agricultural growth were largely exhausted by end- 1984. Most of the residual poor have remained trapped in more remote upland areas where agricultural productivity gains have proven far more problematic” (p. xi) – appears at best partly valid. Much of the slowdown of agriculture after 1984 was brought about by changes in public policy: a reversal and halt to the policy of improved terms of trade and a reduction in resources for agriculture. That a return to improved terms of trade for agriculture in 1994 sharply improved the rate of poverty reduction is convincing evidence that the potential of poverty reduction through improved incentives for agriculture had not been exhausted. 5.6 Be that as it may, the 1992 assessment outlined a sensible strategy which was a combination of a broad endorsement of the government strategy and the addition of elements and emphases to make it more comprehensive. While endorsing what at the time was the emerging government strategy – viz., extending assistance to the ecologically-disadvantaged poor areas - it argued for a broad program consisting of: increased funding for and delivery of social services; strengthening the agencies responsible for poverty alleviation; upgrading the system of poverty monitoring; improving labor mobility; and fostering the development of agriculture and non-farm enterprises in poor areas. 5.7 The 1992 Poverty Assessment found a very low incidence of urban poverty, an inevitable consequence of using the poverty lines in urban and rural areas that represent comparable living standard in spite of the enormous difference between the average living standards between the two areas. It nevertheless notes that “this optimistic portrayal, however, fails to reflect the status of the unregistered urban population” (p. x). It also foresaw that then-forthcoming reform of the state-owned enterprises could lead to some increase in unemployment, but ended on the optimistic note that it should be possible to make adequate provisions for unemployment insurance without creating a significant additional burden for the government. Later events proved it to have been an overly optimistic prognosis. 5.8 The most recent Poverty Assessment – China: Overcoming Rural Poverty, published in 2001- came out years after the 8-7 Plan had become the official poverty- reduction strategy. This report, in line with the official strategy, endorses the view that poverty is entirely a rural phenomenon, thereby going much further than the position taken by the 1992 report. At about the same time, the emerging problem of urban unemployment had driven the government to adopt a program of action to protect these new urban poor categories. In addition to its own analysis, the 2001 Poverty Assessment airs the views of a number of other international agencies. The policies outlined by the Bank and these other agencies together provide a more balanced strategy for poverty 18 reduction than the preoccupation of the government strategy of targeting a limited number of poor rural counties. 5.9 While giving general endorsement to the government strategy, the 2001 report clearly argues that nearly a half of the rural poor live outside the 592 targeted counties. It is also critical of the inadequate targeting of the poor by the official strategy. It argues that the government program of providing support to improve the average incomes of these targeted counties would result in a large leakage of benefits to the non-poor (p. xxi). It spells out a number of policies to improve the targeting of the poor: targeting the poor townships, a lower level of administrative unit than the counties, within and outside the 592 nationally designated counties (p. xxi); targeting the disadvantaged ethnic minorities, the disabled and the female poor. It makes a strong case for increased investment in health, education and nutrition; farm and community level basic infrastructure; development of agricultural and other technologies and their extension to the poor; and improved access of the poor to microcredit and nonfarm employment. These recommendations have substantially influenced the recent redirection of China’s strategy for poverty reduction. Analytical research on poverty 5.10 In addition to the comprehensive poverty assessments, Bank’s ESW during the 1990s included several important analytical pieces on different aspects of the problem of poverty. These studies have dealt with such issues as: spatial poverty, determinants of transient and chronic poverty, the externality of poor-area development, the possible bias in inequality estimates due to statistical practices followed by the NBS, and the implications of population aging for the pensions system for the protection of the old from poverty. These have been of analytical value not only to China but also other countries facing similar phenomena. A major externality of research in this category is that versions of a number of these papers have been or are likely to be published in professional journals, leading to a wide dissemination of these valuable ideas. Issues that remain to be dealt with 5.11 Altogether Bank’s ESW has served very useful purposes. It has strongly emphasized the importance of poverty reduction as the principal goal of Bank’s assistance to China. Its monitoring work has served as a polite dissent to official estimates at times when the poverty-reduction impact of China’s economic growth became weak. While strongly supporting the official strategy for poverty reduction, it has tried to sharpen its focus, improve its targeting and enrich its contents. However, there are areas in which its work needs to be extended. Some of these areas are suggested by the five sets of factors that are identified in section III above as explanations for the slow reduction of poverty in the 1990s. 5.12 The regressive outcome of the system of transfer and levies between the households on the one hand and the state and the collectives on the other is an obvious area of reform which could substantially improve the poverty-reduction performance of public policy. This is a subject in which the Bank has a great deal of experience from its 19 work in many developing countries. A comprehensive study of this issue and a set of operational recommendations deserve a place in Bank’s ESW agenda for China. 5.13 By now it is clear that the neglect of urban poverty is unjustified. Urban unemployment has grown to become far bigger a problem than was anticipated and public policy in China has started putting in place a program to protect the urban unemployed from poverty. A careful and comprehensive analysis of emerging trends affecting the welfare of the poor in urban areas deserves priority. The implementation of the one-off survey of the floating migrants conducted by the NBS in 1999 helped create an information base to include the migrants in an analysis of the urban poor. The possibility of getting the NBS to repeat such surveys should also be given serious consideration. The lending program 5.14 The lending program can contribute to poverty reduction by: (a) Directly improving the proportion of investment in projects that benefit the poor more than the non-poor; (b) Increasing the poverty focus of public policy through policy dialogue that accompanies lending; and (c) Providing the borrowing country with examples of best-practice poverty- reduction projects that can be replicated. 5.15 The importance of these different criteria of judging the poverty focus of Bank’s lending operations depends on the specific circumstances of the country concerned. Thus consider the first of the above criterion which calls for maximizing the proportion of lending for projects that directly benefit the poor, the so-called “full poverty-targeted intervention” (PTI) lending. 33 This would clearly have a decisive effect on the proportion of total investment benefiting the poor if Bank lending (along with the lending by like-minded donors) is a high proportion of total investment in the country concerned (for example, Bangladesh in the 1980s or many countries of Sub-Saharan Africa). In China this is not the case. During the decade under review, China emerged as the biggest borrower from the Bank, annually receiving an average of $ 3 billion of IBRD and IDA funds at the peak rate of borrowing, during the mid 1990s. 34 This however represents approximately one percent of gross annual investment in China during the late 1990s. A large difference in the proportion of these funds that benefits the poor would have an insignificant effect on the overall proportion of investment funds that is directed to projects that benefit the poor. Thus maximizing the proportion of PTI projects in total lending would by itself be a meaningless objective to pursue in China. A further reason 33 Projects are classified as “poverty targeted interventions” if they have a specific mechanism for targeting the poor and/or the proportion of poor among the beneficiaries is significantly higher than the proportion of poor in the total population. This is an assessment made at the time of Board approval. 34 This is the annual average of loans approved, not funds actually disbursed. 20 for not maximizing this proportion is that the method of assigning the PTI status to projects is far from a rigorous and reliable process. As is shown below, often the beneficial impact on the poor of a non-PTI project can be very substantial and, arguably, as great as that of a PTI project. 5.16 Next consider the second of the above criteria, namely using the lending program as an instrument for policy dialogue leading to conditionality to improve the poverty focus of country policy. This would be important in a country in which public policy is at variance with Bank’s own priorities for poverty reduction. In China, this does not appear to be the case. There is a broad convergence of country objectives and Bank objectives in poverty reduction. Under these circumstances, lending conditionality is not the best instrument for policy dialogue on poverty reduction. Furthermore, Bank’s lending operations in China almost entirely consist of project lending and virtually no policy lending. Admittedly a project loan could permit useful policy dialogue in the specific area of the project; but it is not a suitable instrument for a dialogue on broad policies for poverty reduction. The Bank has therefore preferred to influence government policy concerning the appropriate mix of instruments for poverty reduction by ESW, not by wielding the leverage of project lending which, in any case, is very weak. 5.17 The third of the above criteria - providing replicable examples of best-practice blueprints of projects that are efficient vehicles for the implementation of sensible poverty reduction strategies – is clearly an important indicator of the contribution that Bank lending can make to poverty reduction in China. The most effective way that Bank’s lending program can help poverty reduction in China is by including in the portfolio an adequate set of examples of poverty-reduction projects that embody the expertise based on the Bank’s world-wide work in this area. Has Bank lending in China fulfilled this task? 21 Table 5.1: Bank Lending in China Since1990 (Amounts in Million $) Fiscal Year All Projects PTI Projects No. IBRD IDA Total No. IBRD IDA Total 1990 5 - 590.0 590.0 1 - 150.0 150.0 1991 10 601.5 977.8 1,579.3 3 147.1 361.9 509.0 1992 16 1,577.7 948.6 2,526.3 6 30.0 693.6 723.6 1993 18 2,155.0 1,017.0 3,172.0 1 - 147.0 147.0 1994 14 2,145.0 925.0 3,070.0 2 - 260.0 260.0 1995 16 2,369.5 630.0 2,999.5 3 47.5 390.0 437.5 1996 16 2,490.0 480.0 2,970.0 3 - 300.0 300.0 1997 11 2,490.0 325.0 2,815.0 3 30.0 305.0 335.0 1998 16 2,323.0 293.4 2,616.4 5 262.0 293.4 555.4 1999 19 1,674.4 422.6 2,097.0 5 276.0 260.0 536.0 2000 8 1,672.5 - 1672.5 - - - - 2001 7 787.5 - 787.5 - - - - Total 156 20,286.1 6,609.4 26,895.5 32 792.6 3,160.9 3,953.5 PTI Lending as Percent of Total Lending Fiscal Year IBRD IDA Total 1990 - 25.4 25.4 1991 24.5 1992 1.9 73.1 28.6 1993 - 14.5 4.6 1994 - 28.1 8.5 1995 2.0 61.9 14.6 1996 - 62.5 10.1 1997 1.2 93.8 11.9 1998 11.3 100.0 21.2 1999 16.5 61.5 25.6 2000 - - - 2001 - - - 1990-2001 3.9 47.8 14.7 NOTE: - means zero or not applicable (i.e. there was no lending in the category) 22 Table 5.2: Sectoral Composition of Lending Sector Percent of Total Lending PTI Lending as Percent of Sectoral Lending IBRD IDA Total IBRD IDA Total 1990-94 Agriculture 15.5 55.9 32.0 14.6 44.0 35.5 Education - 9.2 3.8 - 31.6 31.6 Power and Energy 25.5 0.8 15.4 1.8 100.0 4.0 Environment 6.1 7.6 6.7 - - - Finance 3.1 2.8 3.0 - - - Health, Nutrition and 2.3 5.4 3.6 - 100.0 61.5 Population Oil and Gas 3.9 - 2.3 - - - Public Sector Mgmt. 1.3 1.1 1.2 - - - Telecommunications 3.9 - 2.3 - - - Transportation 34.7 2.9 21.7 - - - Urban Development 1.2 4.9 2.7 - - - Water Supply and 2.5 9.4 5.3 - 26.3 19.0 Sanitation TOTAL 100.0 100.0 100.0 2.7 36.2 16.4 1995-2001 Agriculture 20.0 38.1 22.4 13.4 65.9 25.4 Economic Policy 0.1 1.6 0.3 - - - Education 1.7 16.5 3.7 - 80.3 48.7 Power and Energy 24.2 - 20.9 - - - Environment 3.3 2.3 3.1 22.3 100.0 30.0 Finance 0.3 0.3 0.3 - - - Health, Nutrition and 0.1 16.0 2.3 58.8 94.2 92.5 Population Private Sector Dev. - 0.2 - - - - Public Sector Mgmt. 1.4 1.6 1.5 - - - Social Protection 0.4 10.5 1.8 82.6 88.9 87.6 Transportation 34.5 - 29.9 - - - Urban Development 3.9 4.8 4.0 27.5 4.4 Water Supply and 10.2 8.0 9.9 6.3 69.8 13.2 Sanitation TOTAL 100.0 100.0 100.0 4.5 72.0 13.6 Note: - means zero or not applicable (i.e., there was no lending in the sector). Names of sectors in which there was no lending in a sub-period have been excluded in the listing. 5.18 Table 5.1 summarizes the annual Bank lending both for all projects and for “full poverty targeted intervention” (PTI) projects for each year since 1990. Table 5.2 summarizes Bank lending since 1990 by sector both for all projects and for the PTI projects. PTI projects represent just less than 15 percent of lending in China, a far smaller proportion than the corresponding figure for India, the other major borrower for which similar estimates are available. 35 As noted above, maximizing the proportion of PTI projects in total lending is not an objective that by itself is worth pursuing in China. While the proportion of PTI lending has fluctuated a great deal without showing a clear trend until the end of the 1990s, in any given year during the decade there has been 35 In India 56 percent of lending approved between 1994 and 1999 represented PTI projects. 23 enough of lending in PTI projects for the Bank portfolio in China to satisfy the criterion of providing the country with enough examples of best-practice projects for poverty alleviation. Best-practice projects for poverty reduction: some examples 5.19 During the period under review a total of 156 projects were approved for lending to China. It is impossible for the CAE to make an ex-post evaluation of all or a majority of these projects in required detail to identify all worthwhile poverty-reduction projects. Even the 32 PTI projects can not be subjected to such an analysis, partly because many of them are still in operation, but more importantly because it would require a great deal of resources to carry out satisfactory ex-post analysis of the poverty impact of those of the PTI projects which have been completed. But a limited number of these projects have been subjected to detailed analysis by the Bank and recipients have often expressed unambiguous views about their perception of the poverty impact of some of the projects. While these evaluations and perceptions do not permit the quantification of the extent of best-practice projects for poverty reduction, they help identify some outstandingly successful projects of this category in the recent Bank portfolio in China. 5.20 A recent internal evaluation of selected Bank poverty-reduction projects in East Asia assigns very high rating to two Chinese projects both by absolute and relative standards. 36 These projects are: Southwest Poverty Reduction Project, a social protection project that illustrates the application of the Bank’s multisectoral approach to poverty reduction in poor regions, and Shanxi Poverty Alleviation Project, a more conventional agricultural project that attempts poverty alleviation by reducing production risk and increasing productivity of poor farmers. These projects have been highly praised by the PRC government and undoubtedly serve as best-practice examples of poverty reduction projects. The report had the following to say on the individual projects: • On the South-West Poverty Reduction Project: It shows clear benefits on the lives of the poor as well as eliciting important responses from them, and provides a laboratory for new, innovative approaches to poverty alleviation. The project’s underlying philosophy is that a multisectoral strategy is needed to help poor households in these relatively remote areas to escape poverty in a sustained way... [I]ts actual and potential impact at the highest policy levels can already be characterized as “best practice.. 36 Four projects, two each in China and Indonesia, were subjected to a detailed review by a panel of Bank experts. 24 • On the Shanxi Poverty Alleviation Project: Its strength lies, among other things, in the use of the mainstream of governmental and other existing institutions (e.g., the Agricultural Credit Bank) but the project tries to mold these institutions to achieve project objectives of increased agricultural productivity to alleviate poverty by reducing production risk and raising incomes. It thus provides a foundation for sustained improvements in income and social welfare. 5.21 In the discussions with the CAE mission, the Chinese counterparts strongly expressed their admiration for three projects – the Southwest Poverty Reduction Project (which is one of the two projects subjected to detailed internal review), the Qinba Mountains Poverty Reduction Project and the Western Poverty Reduction Project – as examples of poverty reduction interventions that deserve wide replication. 37 It is not altogether clear why the Chinese officials were limiting themselves to these three as the category of Bank-funded poverty projects. One possibility is that they were simply guided by the official titles of these multi-dimensional projects which labeled them as poverty-reduction interventions. 38 During interactions with the Chinese counterparts it appeared to the CAE mission that their (the Chinese counterparts’) designation of these three as poverty-reduction projects was not based on a careful analysis that had led them to the conclusion that these were the only or even the best of the poverty-targeted interventions in the Bank portfolio. Are there enough best-practice projects for poverty reduction in the Bank portfolio? 5.22 The above four projects are admittedly among the Bank’s best of the poverty- reduction projects in China, not a random selection of PTI projects. The question therefore is whether there are similar examples of best-practice projects in all sectors in which investment can potentially benefit the poor? A starting point in seeking an answer to this question is to see if lending in each of the traditionally poverty-oriented sectors includes a reasonable proportion of PTI projects. The proportion of PTI projects varies a great deal among sectors of lending. Of the three large sectors – agriculture; power and 37 In their discussion with the CAE mission, the Chinese counterparts were extremely eloquent about the great contribution of these three Bank-financed projects to poverty reduction in China. According to them, in the officially-estimated reduction of the number of rural poor from 80 million in 1993 to 30 million in 2000, eight million poor people benefited from these three projects. These three projects accounted for only $ 587.5 million of IBRD and IDA funds, just 2.4 percent of total IBRD and IDA lending of $ 24.4 billion during the decade of the 1990s. In contrast to this glowing tribute, it is worth noting that the official evaluation of the Qinba Poverty Reduction Project - a document titled World Bank Qinba Mountain Poverty Reduction Project in China: Poverty Monitoring Report 2000, dated August 2001 – based on a survey covering the benchmark year (1997) and the period up to 2000, shows very mixed results. Reduction in the headcount rate of poverty was a little higher in project villages than in non-project villages; but poverty gap index and the squared poverty gap index increased in project villages while they fell in non-project villages. Proportion of population consuming less than 150 kg of grain per year increased in both project and non-project villages, more so in the former than in the latter. Income increased faster in non-project villages; but improvement in infrastructure was faster in project villages. Improvements in education and health indicators were as often better in non-project villages as in project villages. 38 In Table 5 the first of these projects is classified as a social protection project while the other two are classified as agricultural projects. The CAE mission had an opportunity to visit certain locations of the Qinba Mountains Poverty Reduction Project. 25 energy; and transportation – which account for 70 percent or more of lending, power and energy has an insignificant proportion of PTI projects while transportation has none at all. Agriculture is the only large lending sector with a substantial proportion of PTI projects. Other sectors with a high proportion of PTI projects – health, nutrition and population; social protection; and education – have accounted for less than 8 percent of lending. 5.23 It seems reasonable to surmise that, prior to the year 2000, there were enough PTI projects in agriculture; education; health, nutrition and population; environment; and social protection that served as best-practice examples of poverty-reduction projects. Many of these projects are still operational or have been completed so recently as to serve as contemporary examples to replicate. However, for several large lending sectors - such as power & energy; and transportation - there have been few PTI projects. Power and energy may not be considered to be a sector in which investment is a direct instrument for poverty reduction. But investment in transportation can be an effective tool for poverty reduction. 39 Should the complete absence of PTI projects from transportation lending, the largest sector of Bank operations in China, be considered a failure to satisfy the objective of providing China with adequate examples of best-practice poverty interventions in this important sector? 5.24 Discussions with the transport sector staff persuaded the CAE mission that such an interpretation is unwarranted. Indeed there are a number of transport projects which have substantially benefited the poor but are not amenable to classification as PTI projects. Examples of such projects are as follows. 5.25 A team of Chinese experts carried out an ex-post evaluation of the poverty- alleviation road improvement (RIPA) component of the Second Henan Provincial Highway Project by comparing conditions in 6 townships covered by RIPA with 6 townships which were outside RIPA. 40 They found that RIPA townships experienced a substantially faster rate of poverty reduction due to a much higher rate of growth in income and in agricultural, industrial and services output due to their improved access to market, tourists and infrastructure facilities. The Second Shaanxi Provincial Highway Project includes a component of improving and constructing about 4,100 km of rural roads in a program to help alleviate poverty in poor counties. The Tri-Provincial Highway Project includes a component of constructing and improving 1,892 km of rural roads in poor counties in Inner Mongolia, Gansu and Ningxia. The SAR of the Guangxi Highway Project shows that the proposed highway will pass through and directly benefit relatively poor communities in four counties and also benefit another four poor counties in close proximity through which it will not pass directly. That transport sector lending shows no PTI project despite this careful targeting of the poor is either due to the combination of these poverty components with much larger components for which 39 That is the message of a major ESW report of the Bank – Hatim Hajj and V. Setty, Roads Improvement for Poverty Alleviation in China, EASTR Working Paper No.1, May 2000 – which forcefully argues that road improvement can be an effective method of helping poor communities by ending their isolation. 40 Post Evaluation Group of the Highway Bureau in Henan Communication Department and the Henan Social Sciences Academy, Deepening Study Report of Post Evaluation on the Poverty-alleviation Road Improvement Component (RIPA) of the Second Henan Provincial Highway Project Loaned by World Bank, Zhengzhou, Henan, P. R. China, August 2001. 26 benefits for the poor are not directly identifiable or due to other difficulties of satisfying the criteria of PTI classification. Had the RIPA components been packaged as separate lending projects, they would have been among the most clearly poverty-targeted interventions. The looming danger 5.26 It should be clear from the above that during the decade of the 1990s there were enough examples of poverty-reduction projections in most sectors in which investment constitutes an important intervention for poverty alleviation. Many of these projects are still in operation. For now they serve as replicable examples of poverty-reduction interventions and provide credibility to the commitment of the Bank’s mission to alleviate poverty in China. 41 Since the year 2000 Bank lending in China has been characterized by a complete absence of any PTI project. If this trend continues, the poverty focus of Bank lending in China will soon fade out of memory. 5.27 The explanation of the absence of PTI lending since 2000 is easy. None of the 32 PTI projects approved during the 1990s was a pure IBRD-funded project. They all were IDA funded while only 12 of them had supplementary IBRD funding. With China’s disqualification for IDA funding since 2000, this category of projects has been completely eliminated from the Bank’s lending to China. The plain fact, often repeated to the CAE mission by the Chinese counterparts, is that the poverty alleviation projects are not attractive to the government of PRC except on soft terms. Ways have recently been sought for a “third window” to soften the overall terms of lending by combining IBRD loans with bilateral grants. But these attempts have not come anywhere near compensating for the foregone IDA funds. Conclusions • Projects have been less important an instrument for the Bank’s poverty reduction strategy in China than in other comparable countries. • Even so, Bank’s project portfolio for China in the 1990s included outstanding examples of best-practice poverty-reduction projects. • It appears that in the 1990s such best-practice examples were available for all major, traditionally poverty-oriented, sectors despite the fact that some of these sectors were unable to classify the poverty-oriented components of their projects as PTI lending for one reason or another. • Since the year 2000, with China’s disqualification for IDA funds, Bank’s lending has failed to include any PTI project. This jeopardizes the role of the Bank as the provider of best-practice examples of poverty-reduction projects. It is not credible for Bank assistance to completely abandon projects as an instrument for poverty reduction in China, a country with a 41 One might wonder why the absence of PTI projects is a cause of concern if, as argued above and in Annex 2, PTI rating is not always accurate and if non-PTI projects have in the past provided substantial benefits to the poor. The answer lies in the reasonable assumptions that, notwithstanding exceptions, PTI projects are on the average more poverty alleviating than non-PTI projects and substantial benefits to the poor is an exceptional, not a normal, feature of non-PTI projects. 27 higher absolute poverty incidence than most countries with comparable real income. 6. Future Priorities 6.1 China has achieved substantial reduction in poverty during the 1990s although it appears that the rate of poverty reduction in this period could have been faster if the benefit of economic growth were evenly distributed between regions and among income groups. 6.2 During the 1990s the Bank made important contributions to China by disseminating methods of poverty monitoring, serving as a friendly critic of its poverty reduction strategy and offering a set of best-practice projects for poverty reduction. There however remain important additional areas in which the Bank’s analytical and operational instruments might provide effective assistance for further poverty reduction work in China. 6.3 Most of the areas of further work by the government of PRC and the Bank for an acceleration of poverty reduction have been brought out in the preceding sections. In this concluding section some of the major points are briefly highlighted. Poverty monitoring 6.4 The system of poverty monitoring in China needs to be improved. It should begin by making the household level survey data available both to the Bank and to independent researchers in China so that estimates of poverty become more accurate and transparent and a plurality of estimates promotes healthy debates and open discussions. There is an urgent need for an improvement of definitions of income and expenditure, and a corresponding improvement in the data collected by the surveys. Furthermore, it is essential to extend the survey to capture the urban migrants. The Bank has a great deal of expertise in these areas. Unfortunately, member countries are usually reluctant to borrow for investment in the improvement of the statistical system for poverty monitoring. The Bank should find a way of persuading the government of PRC to make the necessary investment in the enhancement of the NBS capability in these areas. An externally- financed technical assistance grant would act as a powerful incentive. Adjustments in the poverty-reduction strategy 6.5 China’s poverty-reduction strategy needs adjustment and broadening. Two main directions in which this needs to take place have been emphasized by the Bank’s recent poverty assessment: (a) the targeting of the poor should be improved within the designated rural counties; and (b) the poor who reside outside the designated rural counties should be brought under the umbrella of the poverty-reduction programs. A natural extension of the latter is the need to: (c) provide an official umbrella for the program of urban poverty reduction. 28 6.6 Already the Ten-Year Plan for poverty reduction has broadened the scope of policy interventions by highlighting support to agriculture; incentives for non-farm activities including TVEs; and increased importance of the delivery of health, education and technical services to the poor. These are in the process of being supplemented by the reversal of policies (e.g., privileges and tax concessions) which favor coastal regions and the redirection of public investment to poor regions. Indeed the favorable treatment of coastal regions has already been substantially eliminated. One hopes that the recently adopted Great Western Development Strategy would accelerate the process of redirecting resources to poor regions. Economic and sector work 6.7 Bank’s ESW agenda should include an early and comprehensive analysis of the system of tax on and transfer to households with a view to designing a program of making them at least mildly progressive (i.e., making poor households pay lower than average rate of net taxes and receive higher than average rate of net subsidies). A restructuring of state finance to increase transfer from rich states to poor states is a necessary concomitant of the large shift in the regional allocation of public expenditure that is contemplated by the new directions of the poverty-reduction strategy. This too is a priority topic for Bank’s ESW agenda. 6.8 The Ten-Year Plan has re-emphasized the importance of promoting agricultural development, especially in poor areas. Appropriate incentives to agriculture are under more severe constraints today than was the case in the past: agricultural producers’ prices are constrained by the WTO provisions with far-reaching implications for output, income and employment growth. This is an issue that has been dealt with in several recent studies both by the Bank and by outside researchers. 42 Conflicting views of the desirable strategy and the likely effect of WTO membership on agriculture have emerged from these studies. It would be highly useful for the Bank’s ESW agenda to sponsor a comprehensive and authoritative analysis of these issues with a focus on the effect on the poor and the needed adjustment in the poverty-reduction strategy. Hopefully this is a part of the major on-going Bank study of the effects of China’s WTO membership. 6.9 It is widely-anticipated that WTO membership will mean a restructuring of comparative advantage within agriculture and almost certainly a reduction of comparative advantage for the agricultural sector as a whole. If this turns out to be true, the need for relocating labor out of agriculture will become even more critical for rural poverty reduction than was the case in the past. This will call for an effort to direct more capital to non-agricultural activities in the rural economy and more labor out of the rural economy to urban employment. The liberalization of rural-to-urban migration – which is receiving limited support under the Great Western Development Strategy - will need to be broadened. 42 See Albert Nyberg and Scott Rozelle, Accelerating China’s Rural Transformation, The World Bank, 1999, Chapter 6; and Nicholas Lardy, Integrating China into the Global Economy, Brookings Institute, Washington, D.C., 2002. 29 6.10 Finally, the existing assistance for the urban unemployed should be developed into a comprehensive urban poverty-reduction program. The need for an improved system of urban poverty monitoring, which has been stressed above, is the starting point for such a program. The existing system of identification of the poor is not an adequate basis for the design of an urban poverty-reduction strategy. It leads to both the exclusion of the poor from the program and the inclusion of the non-poor in the program. 6.11 Existing schemes, outlined in section III above, provide a reasonable basis for the protection of the urban poor. The principal obstacle is the lack of resources at the disposal of these programs. Future Bank Poverty Assessment should deal with these issues while Bank advice to the government in the meantime should emphasize the urgency of these issues. 6.12 In the past China was able to reduce poverty due mainly to rapid economic growth which more than offset the poverty-inducing effects of increased inequality. It is essential that China maintains its high growth in the future as the principal source of poverty reduction. What are the dangers that the above redirection of the poverty- reduction strategy will have a negative impact on the rate of growth? The shift of resource allocation in favor of the poor regions will in the short run reduce overall investment productivity, if other things remain the same. In the long run this should pay off by the internalization of externalities which now justify the shift. To avoid a negative impact on growth, the shift should be accompanied by a concerted attempt to improve overall productivity of resources for which there is ample scope in China. Financing of urban insurance programs should not, in principle, increase the demand for public resources. The urban unemployed were on a concealed insurance scheme in the past. The resources that are saved by removing them from the payroll of public enterprises should become available to finance the new schemes for their protection. This again is a potential subject for Bank’s ESW agenda. The lending program 6.13 Bank’s lending program in China helped poverty reduction in the past by including in its portfolio a set of projects that provided large benefit to the poor per unit of investment, projects that served as examples of best-practice methods of investment for poverty reduction. Since the year 2000 PTI projects have been absent from the Bank’s lending operation. To maintain the Bank’s role as a source of best-practice poverty- reduction project blueprints and to preserve the credibility of the focus of the Bank’s strategy on poverty reduction, Bank lending in China in the future will need to make a special effort for the inclusion of such projects in its portfolio. This has clearly become a difficult task because of China’s aversion to borrow for the most effective poverty- reduction projects on IBRD terms despite the high financial and economic rates of return of such projects in the past. A combination of a “third window” and a serious effort at persuading China’s decision makers about the credibility of the rates of return of these projects might help overcome this problem. 30 ANNEX 1 Poverty-Related Economic and Sector Work Since1990 1. Comprehensive Poverty Assessments “China: Strategies for Reducing Poverty in the 1990s.” October 1992. World Bank. Report No. 11245. “China: Overcoming Rural Poverty”. March 2001. World Bank. Report No. 22137. 2. Other Reports Discussing Overall Trends in Poverty “Sharing Rising Incomes: Disparities in China”. September 1997. Report No. 17088. Chen S, Yan Wang. “China’s Growth and Poverty Reduction: trends between 1990 and 1999.” July 2001. World Bank. Policy Working Paper 2651. 3. Selected Analytical Reports and Research Papers Dealing with Specific Poverty-Related Issues “Population aging and pension systems: reform: options for China. May 1996. World Bank Policy Research Working Paper 1607. “Transient poverty in rural China”. June 1996. World Bank Policy Research Working Paper 1616. “Are there dynamic gains from a poor-area development program?” December 1996. World Bank Policy Research Working Paper 1695. “Are the poor less well-insured? Evidence on vulnerability to income risks in rural. December 1997. World Bank Policy Research Working Paper 1863 “Spatial poverty traps”. December 1997. World Bank Policy Research Working Paper 1862. “When economic reform is faster than statistical reform – measuring and explaining inequality in rural China”. March 1998. World Bank Policy Research Working Paper 1902. “Determinants of transient and chronic poverty: evidence in rural China. June 1998. World Bank Policy Research Working Paper 1936. Note: Papers dealing with poverty issues exclusively for years prior to 1990 have been excluded even if they were published since 1990. There are a number of sectoral poverty-related reports, e.g., the OED evaluation of forestry sector loans, China: From Afforestation to Poverty Alleviation and Natural Forest Management. Finally the CASs and many sectoral reports, though not primarily concerned with poverty, include significant analysis of poverty issues. Papers published by the Bank staff in professional journals, including different versions of papers originally issued as ESW work, have not been included in the list.