iN'1 'IL fj I IDip l"giz IDPi fit lFrl Oll]' I' j .IL I.l,lllb 4 . 1f 1r!l lrCalrIiz r u r ,; O 2 g j M : A W * 6ED /24583 a _, A a tT,t NU WFi'pe -L ... x- . , it .tt t- ja4- ..-. THE WORLD BANK OPERATIONS EVALUATION DEPARTMENT 24583 INDIA: Evaluating Bank Assistance for Poverty Reduction A Country Assistance Evaluation Jack van Holst Pellekaan Director-General, Operations Evaluation: Robert Picciotto Director: Gregory Ingram Manager: Ruben Larndany Task Manager: Gianni Zanini 2002 The author is a former Lead Poverty Specialist in the Africa Region at the The Wr World Bank (1993-July 1998). This paper is available upon request from OED. Washington, D. C. OPERATIONS EVALUATION DEPARTMENT ENTHANCIVG DEVELOPMENT EFFECTIVENESS THROUGH EXCELENCE 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 Partnerships & Knowledge Programs (OEDPK) email: ecampbellpage@worldbank.org email: eline@worldbank.org Telephone: 202-458-4497 Facsimile: 202-522-3125 http:/www.worldbank.org/oed 2 Abbreviations CAE Country Assistance Evaluation CAS Country Assistance Strategy CEM Country Economic Memorandum CIG Community Interest Group DPEP District Primary Education Project DPIP District Poverty Initiative Project ESW Economic Sector Work GDP Gross Domestic Product ICRISAT Intemational Crops Research Institute for the Semi-Arid Tropics IDA Intemational Development Association IEG Institute for Economic Grwoth JFM Joint Forest Management NCAER National Council for Applied Economic Research NSS National Sample Survey OED Operations Evaluation Department PDS Public Distribution System PRA Participatory Rural Appraisal PTI Poverty Targeted Intervention RRA Rapid Rural Appraisal UNDP United Nations Development Programme WRCP Water Resources Consolidation Project Contents Preface ...............i Executive Summary .............. iii 1. Situation, Achievements, and Challenges in Poverty Reduction . Measurement Issues .1 Levels of Poverty .2 Causes of Poverty .5 2. Choice of Poverty Reduction Strategy .6 Requirements for Poverty Reduction .7 Theory and Practice .7 3. Response By Government In The Nineties .8 Public Sector Programs .8 Impacts of Programs .9 Role of the Private Sector .10 4. Response By The World Bank .10 Non-Lending Assistance (Research, and Economic and Sector Work - ESW) .11 Research .11 Economic and Sector Work .12 Evaluation of Non-Lending Assistance .14 Lending Assistance .15 Strategic Focus of Poverty Reduction Strategy .15 Poverty-Targeted Interventions (PTIs) .17 Poverty Impact of the Lending Program .18 Current Moves to More Targeted Programs .20 Evaluation of Lending Assistance .21 5. Overall Evaluation Of The Poverty Focus Of The Bank's Assistance Program .22 6. Conclusion - A Vision For Poverty Reduction Strategies .23 Tables 1. Estimates of Measures of Poverty and Inequality over Time in India (1951-1997) .......................... 3 Figures 1. Summary of Strategic Poverty Reduction Focus of Bank Lending (FY94-FY99) .......................... 17 2. India - Relative Importance of Poverty-Targeted Interventions (FY94-FY99) .......................... 18 Annexes 1. Summary of Projects Approved from FY94 to FY99, their Poverty Focus, and PTI Rating .......................... 26 2. Summary of Discussion at CAE Workshop on Agricultural and Rural Development, and Poverty Reduction ..........................; 30 Preface This paper is one of the background papers prepared as an input to the India Country Assistance Evaluation (Task Manager: Mr. Gianni Zanini) by the Operations Evaluation Department (OED) of the World Bank. Findings are based on a review of project appraisal and completion reports, sector reports, and a number of other documents produced by the Borrower, the Bank, OED, and research papers. An OED mission visited India in April/May 1999. The mission interviewed current and retired government officials and Indian experts. Bank staff were interviewed at both headquarters and in the field office. Their valuable assistance is gratefully acknowledged. Peer reviewers were Messrs. Anandarup Ray (Consultant), and Sohail Malik and Ms. Alison Evans (OEDCR). An earlier version of this paper was reviewed by Ms. Valerie Kozel (SASPR) on behalf of the Bank's India poverty team, and Stephen Howes (ASAPR) and Martin Ravallion (DECRG) provided additional comments and advice. Poverty issues were also discussed in the context of a workshop on agricultural and rural development in New Delhi, on April 5, 2000 chaired by Professor Vijay Vyas and with the participation of central and state government officials, academics and members of research institutes, and other representatives of civil society. The author is grateful for all comments received; they have been taken into account in the preparation of this revised version. However, the views expressed in this paper remain entirely those of the author. They do not necessarily represent the views of OED or the World Bank. Executive Summary 1. The focus on poverty reduction by the Indian government goes back many years and those declarations continue to this day.' Declarations about the importance of poverty reduction are also found regularly in the World Bank's country assistance strategies. The objective of this paper is to assess the extent to which the Bank's analytical (non-lending) work and lending operations have had an impact on poverty reduction in India over the last decade with the main focus on the last five years. Disappointing Poverty Trends 2. Despite the declarations and many government programs for poverty reduction, it is not clear that poverty has been reduced in India during the 1990s. The National Sample Surveys (NSS) of Indian households, when the official price indexes are used to determine poverty lines, indicate that on average the incidence of rural poverty was the same in 1997 as it was at the end of the eighties, namely 36 percent. This was surprising because the nineties were generally years of strong economic growth. It was also in contrast to the steady decline in rural poverty during the eighties when its incidence came down from 45 percent in 1983 (the closest measure available) to about its current level. Urban poverty also declined steadily from 36 percent since 1983 but continued to decline to its estimated level of 30 percent in 1997. On a national basis the incidence of poverty declined from 43 percent in 1983 to 34 percent in 1997. The latest NSS data up to 1998 show a decline in poverty but because of problems with the questionnaire these results have been rejected. 3. In the light of the relatively slow decline in poverty up to 1997 based on estimates using official price indexes the unit values of consumption implicit in the consumption expenditures collected in household surveys have been used to re-estimate price indexes for 1987/88, 1993/94 and 1995/96. These revised price indexes were then used to re- estimate poverty lines and re-calculate the incidence of poverty. The re-estimated incidence of poverty was calculated to be 39.2, 32.9 and 31.6 percent in rural areas for the three years, and 22.6, 18.5 and 16.8 percent respectively in urban areas. This result. gives a much more favorable picture of the poverty trends in the nineties than when estimated using the official price indexes. Another recent estimate of poverty based on the private consumption component of national income claims that the incidence of poverty in India in 1997 was only 26 percent. 4. There is at present no consensus on the incidence or depth of poverty in India. The Bank therefore finds itself in the unenviable position of having a country assistance strategy aimed at reducing poverty but without adequate baseline information on the current incidence and depth of poverty at the national or state level. X See companion background paper by T. N. Srinivasan. India's Development in the Near Term: Constraints and Prospects, background paper for India CAE. draft. June 1999. iv 5. Regardless of which estimate of poverty is accepted, and even if it is concluded that a decline in poverty has taken place, the reduction in poverty is disappointing considering that average growth in GDP in the mid-nineties has been around 7.5 percent per annum following a respectable average 5 percent per annum during the eighties. GDP growth for agriculture, the sector that employs most of the poor, has been around 3.4 percent per annum during the nineties which was a modest increase compared with the average growth rate in the eighties of 3.1 percent per annum, but still at least a percentage point above the growth rate of population. There is evidence that, despite the higher growth, the agricultural sector had a reduced ability to maintain previous levels of employment. In addition, those traditionally employed in agriculture have had difficulties finding altemative remunerative employment outside agriculture. 6. Changes in poverty at the state level vary substantially compared with the national average reflecting their current levels of poverty and economic growth, as well as weaker primary social services. In states such as Bihar, Rajasthan and Uttar Pradesh, where many Indian poor live, the incidence of poverty (based on the use of official prices) has tended to increase over the decade to 1993/94 when the last comprehensive survey data are available. On the other hand in states where economic activity has been strong, where economic and social endowments are generally better, and where poverty was generally not so high such as in Andhra Pradesh, Madya Pradesh, Maharashtra, and Kerala, poverty has declined. The Punjab and Orissa are interesting exceptions. In the Punjab (a high income state), which had a relatively low level of poverty in 1983 (headcount index of 21.5 percent) experienced an increase in its headcount to 25.2 percent in 1993/94. On the other hand, Orissa (a poor state) had a headcount of 56.8 percent in 1983 but reduced this to 40.3 percent in 1993/94. Recent research sheds some light on the pre-conditions for poverty reduction in different states. These will be discussed later in the paper. 7. Moneymetric measures of poverty do not, of course, paint the complete picture of poverty and deprivation in India. Many reports by the govemment, the Bank and other institutions have illustrated the desperate condition in which about one-third of India's population live. While the situation is improving slowly, the UNDP Human Development Report for 1999 ranks India as 132nd out of 174 countries in terms of the human development index.2 World Bank Non-Lending Activities - An Improving Focus on Poverty Reduction 8. The Bank's Country Assistance Strategies (CASs) have always emphasized poverty reduction as the over-arching objective of the Bank's assistance program, but they have been unclear about the likely impact of assistance strategies on poverty reduction. In the early nineties the Bank strongly supported the govemment's focus on improving macroeconomic policies to achieve economic growth and poverty reduction. This was a necessary first strategy. There was, however, an inadequate focus by the Bank on the need to adjust national and state agricultural and rural development policies, as part of the 2 United Nations Development Programmne, Human Development Report 1999, Oxford University Press, 1999. The human development index is a composite index designed to measure human development. v growth and poverty reduction strategy, until the latter part of the nineties.3 On the other hand, projects aimed at broad based improvement of primary social services (e.g. national primary education and health programs), while a relatively small part of the lending program in the eighties, were an increasingly important second part of the strategy and lending program in the second half of nineties. Of course poverty reduction resulting from improved primary services takes many years. Third, projects with a narrowly targeted group of beneficiaries such as the poorest groups were a relatively unimportant part of the Bank assistance strategy until recently. 9. Overall there has, however, been a "sea change" in the attention to poverty reduction in the Bank's assistance programs from moderate interest in the late eighties and the first half of the nineties to acute concern evident in the CAS of 1997. The 1999 CAS update and the draft Comprehensive Development Review for India reflect a continued attention to poverty reduction in the Bank's assistance strategy. The focus by the Bank matches the concern in the Indian govemment and in public debate about the effectiveness of national and state poverty reduction programs. It also underlines a concern to achieve better governance and a more effective implementation of the poverty reduction programs through the Panchayats following the 73rd and 74th amendments to the constitution. 10. Bank research on poverty in India in the eighties and early nineties focused mainly on analysis of National Sample Survey data to evaluate trends in poverty and the decomposition of those trends into growth and distribution components. There was also considerable analysis of the govemment's targeted programs. This has been high quality work and highly valued. It is used regularly in Indian academic and policy circles, and in the Bank. In recent years research has become increasingly focused on analyzing policy issues which are relevant to public expenditure allocations for the reduction of poverty. 11. The Bank's economic and sector analysis on poverty over the last ten years included first a Review of Poverty, Employment and Social Services in 1989 (the country economic memorandum for that year). It discussed the substantial progress made in poverty reduction during the eighties, the constraints to a continued reduction in rural poverty, and the issues that lay ahead. The 1991 CEM laid out policies and public expenditure strategies that constrained the opportunities for income growth in the agricultural sector and drew conclusions on actions that needed to be taken to stimulate agricultural growth. These reports, with strong and clear messages about strategies for poverty reduction, were both virtually ignored within the Bank. They contained, for example, warnings about the difficulties of achieving continued poverty reduction in the nineties compared with the eighties. They also emphasized the necessity of stimulating rural development for poverty reduction. 12. After a hiatus in sector work on poverty issues, two poverty assessments were prepared in 1997 and 1998. They covered respectively the importance of growth and 3See companion background paper for the India CAE on agricultural and rural development. The Region's view is that achieving changes in policies at the central govemment level, while desirable, is fraught with the difficulties of an unpredictable political economy. It is, in the Region's view more direct and practicable to negotiate changes in state policies in the context of agreements on credits and loans at the state level. vi employment to poverty reduction, and the role of social services and targeting in poverty reduction. Both were informative relying heavily on Bank research, and interestingly written. The first report of 1997 focused on the importance of growth in general as the solution to poverty reduction but without any analysis of the pattern of growth that would be necessary to reduce poverty. It therefore added little value in terms of how to design Bank operations to achieve poverty reduction. The second report was more strategic in its discussion of options and recommendations for poverty reduction. A follow-up report on poverty in 1998 drew from the results of qualitative surveys in Uttar Pradesh and Bihar providing additional perspectives of poverty and drew attention to the importance of developing social capital as a way for the poor to insure against risks of social exclusion and as a basis for pursuing income generation. This work has recently been developed into an interesting poverty assessment for Uttar Pradesh. 13. There were also a number of country economic memoranda (CEMs) and a plethora of sector reports in the second half of the nineties. The CEM for 1995 included a chapter on rural development and poverty, but generally CEMs did not focus a great deal on poverty reduction strategies. There were sector reports on sugar and livestock (1996) primary education, health, and oilseeds (1997), and irrigation (1991 and 1998) which all could have focused to a large extent on poverty reduction. In the event their content (except for education) was narrowly sub-sector driven missing a broad perspective on major challenges such as poverty reduction. Indeed these sector reports reflect very well the overall conclusion that until the late nineties poverty reduction was not the central theme for the Bank's assistance strategy in India. 14. Finally, the report on rural development and poverty reduction in 1999, capping five years of background work, drew similar conclusions to those in the 1991 CEM. This report analyzed the linkages between rural development, the non-farm sector and poverty reduction. It promises to be a flagship product for intensifying the Region's focus on rural development and poverty reduction in India if its messages can be effectively disseminated and discussed with the central and state governments. The most important aspect of rural development is that it will involve an increased role for the private sector in contributing to poverty reduction. Growth Strategy for Poverty Reduction 15. A cohort of 53 projects approved in FY94 through FY99 was classified according to their generic objectives and implicit strategy for poverty reduction, namely stimulating growth, providing broad-based improvements in social services, or financing narrowly targeted services and assistance to the poor. For most of the second half of the nineties during which the Bank's strategy focused more explicitly on poverty reduction the project assistance was seen to intensify considerably 58 percent of project commitments was devoted to primarily supporting growth, namely projects addressing broad macroeconomic policy issues, increased output, or improving the enabling environment for increased output such as more reliable power supplies. About 34 percent of project commitments financed broadly targeted support for poverty reduction such as the development of primary services in education and health, and domestic water supplies. vii There was very little (8 percent) of the lending program devoted to narrowly targeted programs for poverty reduction. The forestry projects, which originally were not designed primarily to focus on poverty reduction, are the main projects in this third group. A surprising aspect of all projects was that, despite the rhetoric in CASs about the importance of poverty reduction being an important objective of the Bank's operations, there were no targets established for these projects that were directly relevant to poverty reduction and no systematic monitoring of the poverty reduction impact of the projects financed. By and large this was typical of the situation in many countries. Reflecting a renewed concem for monitoring the Bank's India field office has held a workshop in mid 2000 with the objective of improving monitoring and evaluation in the Region. 16. Another approach to assessing the poverty focus of the lending program is to use the Bank's classification of the expected impact of projects on poverty reduction (referred to as poverty targeted interventions or "PTIs") assessed at the time of Board approval. For India 56 percent of Bank-financed projects were classified as PTIs since FY 94 4, but it should also be noted that there is a healthy skepticism among many Bank staff about the veracity and usefulness of the PTI classification. It is no more than a rough guide to the focus of projects on poverty reduction. Evaluation of Main Assistance Programs 17. Non-lending (research, economic and sector work). This evaluation gives the greatest weight to poverty assessments, followed by other economic and sector work (ESW). The reason for this weighting is because poverty assessments and ESW generally are intended to be relevant to the Bank's operational work on poverty reduction. At the same time it is acknowledged that the Bank's excellent research work on India has been extensively used in preparing the two poverty assessments. The relevance and quality of the Bank's ESW work for poverty reduction in the first half of the nineties were respectively substantial and satisfactory. The impact on Bank operations, however, was negligible since the two economic reports most relevant to poverty reduction were virtually ignored because of the Bank's exclusive focus on macroeconomic issues at the time. 18. In the second half of the nineties research was highly satisfactory and substantially relevant. The 1997 poverty assessment was also substantially relevant to the Bank's poverty reduction strategy and operational work in India. While it adopted the conventional growth-led poverty reduction strategy, it was based on considerable research work (mostly by the Bank) and provided an important baseline of knowledge about poverty in India that the Bank had heretofore not been able to present. That was an important breakthrough in the Bank's collaboration with the Government. The generality of the 1997 poverty assessment no doubt stimulated the preparation of a specific Bank strategy for poverty reduction in the 1997 CAS. The next poverty assessment on social services and targeting in 1998 was more strategic, but it added relatively little to the Bank's existing strategy in the social sectors. The Bank had already placed considerable T Ihe classification of projects as PTIs or non-PTIs started in 1992. viii emphasis on putting development of the social sectors into high gear in India in the early nineties. Also, an explicit strategy had already been laid out in the earlier 1997 CAS. The chapter on targeting in the second poverty assessment was not relevant to the Bank's strategy and operations since it had no intention of providing assistance in this area. Nevertheless it was an excellent chapter and has no doubt generated many benefits outside the Bank. Overall the relevance of the second poverty assessment was also judged to be substantial. The quality for both poverty assessments was satisfactory, but their impact was modest since they contained few new strategic insights. The other major ESW output directly relevant to poverty reduction was the 1999 report on rural development and poverty alleviation was substantially relevant to poverty reduction, had relevant but weakly presented recommendations for agricultural policy change, and was of a satisfactory quality. 19. Lending. The outcome of projects in terms of poverty reduction for both the first and second halves of the nineties is likely to be moderately satisfactory, although there have been some improvements in project design (and therefore better prospects for poverty reduction outcomes) in the last half of the nineties. This assessment of outcome is based on substantial relevance of projects to poverty reduction, and modest efficacy since these projects were mainly directed at people with assets - namely not the poorest. In addition, because of the substantial surplus labor in rural areas, additional employment opportunities would have been limited. There was virtually no monitoring and evaluation of poverty reduction or employment generation in any of the Bank projects so the impact of projects on poverty reduction cannot be ascertained. The institutional development impact of projects for poverty reduction is expected to be modest but in some cases negligible because, apart from a few exceptions, project documents show that Bank- assisted projects have not developed long term institutional structures to address poverty. Sustainability of projects as mechanisms or models for poverty reduction is uncertain because in most cases a replicable approach for economic development and poverty reduction did not evolve. Overall Evaluation 20. On the basis of actual and expected results for poverty reduction, and giving most weight to the portfolio of project investments compared to non-lending work over the last five years, the overall outcome for the Bank's assistance program over the last decade is rated as moderately satisfactory. The basis for this evaluation is the substantial relevance of the program to poverty reduction. While the absence of poverty monitoring in most projects makes it impossible to assess their impact on poverty, it is expected that their efficacy in terms of poverty reduction will also be modest. Institutional development impact of Bank-assisted projects in terrms of poverty reduction programs is expected to be modest, and sustainability of projects or components of projects focused on poverty reduction, such as exist, is uncertain. I 1. Situation, Achievements, And Challenges In Poverty Reduction 1.1 The broad facts about poverty in India are well known thanks to considerable publicity, a large number of surveys, substantial research into poverty, and numerous statements about strategies for poverty reduction by Indian and international scholars over many years.5 At the same time the precision of poverty data has recently been questioned because of doubts about the analysis of the basic household survey from which the poverty data have been derived. Poverty continues to be the subject of intense discussion in India amid considerable controversy about both its measurement and the best strategies to reduce it. Measurement Issues 1.2 Controversies about the extent and depth of income poverty continue in India, despite the long history and relatively high quality of its National Sample Survey (NSS) of households. There are also many other surveys available but the NSS is by far the most comprehensive.6 One issue is the price data used to estimate the value of consumption expenditure. The prices used to estimate the value of consumption in the NSS are based on regional and state price series collected independently. Recently Angus Deaton and Alessandro Tarozzi calculated the unit values of consumption implicit in the consumption expenditures collected in household surveys for 1987/88, 1993/94 and 1995/96. These prices were then used to re-estimate the poverty lines. It was then possible to re-assess the poverty status of these households and summarize the results by states. Poverty estimates based on the latest NSS data have been rejected because of serious questions about the design of the questionnaire which would tend to inflate information on consumption expenditure. 1.3 Another issue is the long running controversy has been the divergence in the incidence of poverty between that measured through the NSS and that measured through estimates of private consumption in the national accounts. In recent years national accounts-based estimates of private consumption (higher than those estimated by the NSS) imply a more optimistic picture of poverty reduction in the 1990s than NSS household surveys.7 In 1993 an expert group appointed by the government concluded, inzter alia, that there was no reason to believe that consumption estimated through the national accounts were more reliable than the data collected through the NSS. Also, the See, for example Suresh D. Tendulkar, Indian Economic Policy Reforms and Poverty: An Assessment. in India's Economic Reforms and Development: Essays for Manmohan Singh, edited by Isher Judge Ahluwalia and I.M.D. Little, Oxford University Press, 1998. There is now increasing debate on the current incidence of poverty in India. It had been thought that this debate would be resolved when the latest complete household survey for 1999 is analyzed. Unfortunately the results of the latest NSS are in doubt because of questions about the questionnaire design. In the meantime projections of poverty measures have been made on the basis of results from smaller annual surveys: see for example Economic Reforms and rlre Poor, prepared by the National Council of Applied Economic Research, October 1998. 6 Surveys have been done by ICRISAT and NCAER. The draft of Bank report No. 19471 -IN, India: Comprehensive Development Review of Policies to Reduce Poverry and Accelerate Development, contains a discussion of these issues. 2 NSS data have a sound track record and are still the preferred source of data for the analysis of poverty. 1.4 In June this year Surjit Bhalla, who disagrees strongly with the conclusions of the expert group, calculated the incidence of poverty in India based on estimated private consumption in the national accounts and an elasticity of the change in the headcount index with respect to the change in real per capita income/expenditure. Using his estimated elasticity (-0.75), he claims that poverty in India declined more than estimated by the Government or the Bank.8 Levels of Poverty 1.5 Money-metric Measures. This paper will focus on estimates of poverty made in the Bank since they are the most complete and during the period for this CAE review were the data being used for analysis and strategic decisions. Table 1 summarizes the latest available information on poverty based on NSS data for the last two decades. The incidence of poverty in rural areas declined from 45 percent in 1983 to 36 at the end of the eighties and then to 36 at the end of the nineties, although in the intervening years it had increased under the impact of the stabilization and reform program but also as a result of other factors such as drought in 1992 when the incidence of poverty hit 44 percent. Over the same period urban poverty declined steadily from 36 percent to 33 percent to 30 percent. For India as a whole the incidence of poverty declined from 43 percent in 1983 to 36 percent in 1990/91 to only 34 percent in 1997. This means that for India as a whole, about 327 million people live below the poverty line.9 If a different benchmark is used such as a poverty line defined in terms of the internationally comparable purchasing power of $1.0 per day per head then 53 percent of India's population lived below that level in 1992. If the poverty line is defined as $2.0 per day then 89 percent of Indians were below that benchmark in 1992.10 Although these are probably over estimates because 1992 was a drought year and there were some negative impacts on consumption of the reform program, the basic issue is the enormity of the poverty problem that needs to be tackled. 1.6 In the interests of complete disclosure, it should be mentioned that the Deaton/Tarozzi paper estimated the incidence of poverty to have been 39.2, 32.9 and 31.6 percent in rural areas for 1987/88, 1993/94 and 1995/96. The incidence of urban poverty was estimated at 22.6, 18.5 and 16.8 percent respectively for the three years." Also, Bhalla estimated that the incidence of poverty had declined from 44 percent in 1987 to 26 percent in 1997 for India as a whole.'2 Obviously these results give a much more favorable picture of poverty trends in the nineties than the assessment by Gaurav Datt. This paper was not planned as a place to analyze the strengths and weaknesses of the u Siit Bhalla. Trends in World Poverrv - Ideology and Research., Oxus Research and Investments, June 28, 2000. mnimeo. 9 The poverty line for India was established in 1993 as monthly per capita expenditures of Rs. 49 in rural areas and RS. 57 in urban areas as at October 1973 - June 1974. 10 To be updated to latest available estimates. I I Note that the DeatonlTarozzi estimnate for rural areas in 1987/88 is the same as Datt's estimates, but for urban areas the DeatonlTarozzi estimate is lower than that by Datt. '- See Bhalla. op cit. Table 3. Note that Bhalla's estimate for poveTty in India for 1998 is even lower at around 15 percent. Even casual observation of Indian society suggests that this estimate is very optimistic. 3 three current estimates of Indian levels of poverty. It should be said, however, that there is no evidence that this review has uncovered to support Bhalla's view that national accounts are the best basis for estimating poverty in India, nor the charge in his paper that Bank estimates of India's poverty levels are ideologically based. 1.7 Regardless of the poverty estimates used, the trends and levels of poverty and inequality are disappointing given the strong growth in the economy during the nineties. In addition they contrast starkly aaainst the substantial reductions in poverty achieved in other large countries such as China and Indonesia over the last decade in urban and rural areas. In China, for example, the incidence of rural poverty (using a poverty line of $1 per day in purchasing parity terms) declined from 31percent in 1990 to between 13 and 14 percent in 1997 driven by a growth rate in agriculture of 4.4. percent per annum over that period. Urban poverty in China is minor compared to rural poverty, but it increased from 0.5 percent in 1990 to 0.8 percent in 1997.13 With poverty reduction so prominent in the agendas of the Indian government and the Bank it is worrisome that the reduction in poverty has been so elusive by comparison with the success of other countries. Leaving aside the actual level of poverty, this does then beg the question about the causes of poverty and these will be discussed in the next section. Table 1: Estimates of Measures of Poverty and Inequality over Time in India (1951-1997) (Percent) Survey Period Headcount Poverty Gap Squared Poverty Gini Index Index Index Gap Index Rural Urban Rural Urban Rural Urban Rural Urban August 51 - November 52 47.37 35.46 16.05 11.14 7.53 4.82 33.72 40.0 July 60 - August 61 45.40 44.65 13.60 13.84 5.53 5.83 32.48 35.57 July 70 - June 71 54.84 44.98 16.55 13.35 6.80 5.35 28.76 34.69 July 77 - June 78 50.60 40.50 15.03 11.69 6.06 4.53 30.92 34.71 January 83 - December 83 45.31 35.65 12.65 9.52 4.84 3.56 30.1 34.08 July 86 - June 87 38.81 34.29 10.01 9.10 3,7 3.4 30.22 36.75 July 87 - June 88 39.23 36.20 9.28 9.12 2.98 3.06 29.39 34.64 July 88 - June 89 39.06 36.60 9.50 9.54 3.29 3.29 29.51 34.8 July 89 - June 90 34.30 33.40 7.80 8.51 2.58 3.04 28.23 35.59 July 90 - June 91 36.43 32.76 8.64 8.51 2.9 3.1 27.72 33.98 July 91 - Dec 91 37.42 33.23 8.29 8.24 2.68 2.90 29.91 37.98 Jan 92 - Dec 92 43.47 33.73 10.88 8.82 3.81 3.19 29.88 35.11 July 93 - June 94 (A) 36.66 30.51 8.39 7.41 2.79 2.42 28.58 34.34 July 93 - June 94 (B) 36.32 30.29 8.11 7.12 2.49 2.25 28.13 34.13 July 94 - June 95 39.75 33.50 8.89 8.38 2.90 2.80 29.43 37.18 July 95 - June 96 37.46 28.04 8.31 6.78 2.64 2.22 28.62 35.53 Jan 97 - Dec 97 35.69 29.99 8.39 7.77 2.83 2.73 30.40 36.54 Sourre: Gaurav Dan. Has Povertv in India Declined During rile Post Reform Period. mimeo, Intemational Food Policy Research Institute, Februawy 1999. The different results for July 93 to June 94 are based on different estimation techniques. "A" is based on fitted Lorenz curves and "B" is based on fitted distribution functions. 13 Participants at the workshop were uncomfortable with the China/India comparison because Chinese agriculture had come from such a low base that growth rates and poverty reduction would inevitably be high.. It is true that yields in China in the early eighties were low, nevertheless the sustained increase in their yields over 20 years has been a remarkable perfonrance by any comparison and does raise the question as to why Indian agricultural productivity has been so slow to move. Note that Bhalla, op cit, estimated that the incidence of poverty in China declined from 47 percent in 1990 to 21 percent in 1997. 4 1.8 Social indicators. When money-metric measures of poverty are supplemented by various other (usually social) indicators the real extent of the deprivation of the poor emerges. For example, literacy and health services are still generally at low levels in India. Currently the average literacy rate is about 62 percent, with a level of 50 percent for females and 73 percent for males. While average literacy rates for females and males have risen over the nineties, the data show no change in the disparity between women and men. Infant (under five) mortality is about 80 per 1000 live births for India as a whole, but (based on 1992/93 data) the poorest quintile has double the infant mortality rate of the richest quintile. Access to safe water is found in about 62 percent of households, and access to toilet facilities is found in about 24 percent of households. Child immunization is between 80 and 90 percent for the whole of India. Communicable diseases, and pre- natal and maternal mortality account for 12.5 percent of deaths of rural women aged 15 to 45. This selection highlights the low level of India's social indicators. At the same time it should be acknowledged changing social indicators is difficult and requires a long-term effort. It should also be acknowledged that social indicators are improving. 1.9 A "silent emergency" in India is discussed in the second poverty assessment. It notes, inter alia, that India has 20 percent of the world's child population but 40 percent of the world's malnourished children. 14 Another paper in 1998 drew attention to information in the household surveys which shows that food and nutrition deficiencies in households are much larger than what one would conclude by looking at the income poverty estimates. It was concluded from the data that energy, protein and micronutrient deficiencies continue up to and including the seventh decile of the income distribution. 15 It is therefore not surprising that more than half of India's children are moderately or severely undernourished, and 30 percent of newboms are significantly underweight. 16 1.10 The 1999 Human Development report provides a summary index based on a number of social indicators which is indicative of the deprivation suffered by India's poor. It shows that India ranks 132nd out of 174 countries in terms of the human development index. In terms of this index India ranks just above Ghana and Cameroon and just below Zimbabwe and Equatorial Guinea. 17 1.11 Obviously an "all India" approach masks considerable variation between the states and communities. Changes in poverty at the state level are substantially different from the national average reflecting their current levels of poverty and economic growth, as well as substantially different levels of primary social services. In states such as Bihar, Rajasthan and Uttar Pradesh, where many of Indian poor live, the incidence poverty has tended to increase over the decade to 1993/94 when the last comprehensive survey data are available. On the other hand in states where economic activity has been strong, where economic and social endowments are generally better, and where poverty was generally not so high such as in Andhra Pradesh, Madya Pradesh, Maharashtra, and Kerala, poverty has declined. The Punjab and Orissa are interesting exceptions. The " For example, see Bank report No. 17881 -IN, India: Reducing Poverty in India: Optionsfor More Effective Public Services, op cit page 29. '" Abusaleh Shariff and Ananta C. Mallick, Dynamics of Food Intake and Nutrition According to Expenditure Class, 1998. 16 See Bank report No. 18667-IN. India: Wasting Away: tile Crisis in Malnutrition in India, December 15, 1998. '' Human Development Report 1999, op cit. 5 Punjab and Orissa are interesting exception. The Punjab, also a state with substantial endowments, had a low level of poverty in 1983 (headcount index of 21.5 percent) the headcount increased to 25.2 percent in 1993/94. On the other hand, Orissa (still a poor state) had a headcount of 56.8 percent in 1983 but reduced this to 40.3 percent in 1993/94. To summarize the variation another way, the latest full scale household survey in India in 1993/94 shows that incidence of rural and urban poverty varied between lows of.25 and 11 percent respectively in the Punjab and highs 64 and 40 percent respectively in Bihar. The average depth of poverty was about 5.2 percent and 1.8 percent respectively in rural and urban Punjab. whereas it was 17.3 percent and 9.8 percent respectively for rural and urban areas respectively in Bihar.'8 Causes of Poverty 1.12 It is income, endowments, and access to social services that distinguish the rich from the poor. The large variation in poverty among the states mentioned above almost certainly reflects the variation in the endowments of their households. Datt and Ravallion have referred to this as the impact of "initial conditions" when they examined the reasons for differences in the poverty rankings of Indian states between 1960 and 1990.'9 They found that there was a favorable effect (based on cross-state regressions) of the initial proportion of land under irrigation on the growth rate of average household consumption and therefore the rate of progress in poverty reduction. The rate of poverty decline was also significantly lower in states that started with lower female literacy rates. For example, a state such as Kerala which has the highest proportion of villages of any state with electricity, high road density, a relatively high proportion of land under irrigation, and by far the highest female literacy rate of any state, reduced the incidence of poverty by 2.3 percent per annum between 1957/58 and 1990/91. 1.13 A more recent (unpublished) paper by Ravallion and Datt underlines that "initial conditions," such as poverty itself, and low levels of education inhibit the prospects of the poor finding employment in the non-farm sector. In other words its difficult to find a job off the farm if you are already below the poverty line, without education and sound health and few other assets with which to change one's circumstances. 1.14 Gaiha and Kulkami came to similar conclusions drawing attention to circumstances in which poverty persists because growth has no impact on the welfare of the poor. The reasons, they concluded were that many of the poor lack physical and human capital assets or belong to lower castes (that is a state of "social exclusion"). In some circumstances the disadvantaged poor may live in remote regions which means that they cannot participate effectively in markets or take advantage of services. 20 These conclusions add emphasis to the importance of human capital development and infrastructure development as a means for the poor to escape from their low "level equilibrium trap". 18 See Bank report No. 19471 -IN, India: Comprelhensive Development review of Policies to Reduce Poverty and Accelerate Development, for more information on regional disparities in poverty. 19 Gaurav Datt and Martin Ravallion, Whiy Have Some Indian States Done Better tilan Otliers at Reducing Rural Poverty?, Economica, (1998) 65, 17-38. 20 See Raghav Gaiha and Vani Kulkami, Is Growthi Central to Poverty Alleviation in India? Journal of International Affairs, Fall 1998, 52, No. 1. pp 145-188. 6 1.15 The ability to address poverty is also diminished by the isolation the poor, and particularly women, feel as a result of not being included in a community and its economic and social activities. This aspect of poverty is perhaps more relevant for India than for many other developing country.2' The Mandal Commission found that caste was an important constraint to equitable development in India and as a consequence the government has since then taken steps to redress discrimination based on caste. This was an important step forward in the fight against poverty. For India, therefore, social development and rural infrastructure are just as important a part of a poverty reduction strategy as growth in income or improved primary services. 1.16 As noted already, the decline in poverty in rural areas stagnated even while the growth rate of agricultural GDP declined. The National Council of Applied Economic Research concluded that the main reason for the arrest in the decline in rural poverty during the eighties was the declining share of agriculture in total GDP without ZP ~~~~~~~~~~~~~~~~~~~~~22 compensating reductions in the number of people dependent on the sector. Clearly increased non-farm income would be the answer, but the growth of non-farm income opportunities has also declined leaving the poorest farners trapped. Traditionally the non-farmn sector has been the strongest where the working conditions are harsh and agriculture suffers uncertainty. Rajasthan is a classic example. The poor could usually find jobs in the past, but there is at present a gap between expectations and reality since employment prospects depend on having skills. Many of the unemployed do not have the skills, and training facilities are inadequate at present.23 1.17 Obviously this is a severely limited overview of the causes of poverty in India. The averages also gloss over the tremendous variation in access to services for the poor in particular, and in different states and communities. Nevertheless, they provide pointers for an action plan for achieving poverty reduction. The conundrum for India is that while there is a wealth of data and analysis on poverty and its solutions, progress on poverty reduction is painfully slow. 2. Choice Of Poverty Reduction Strategy 2.1 As in other countries, there has been a debate in India about the efficacy of growth-led poverty reduction versus a strategy based on improved social services and programs targeted at the poor. This is not the place to devote a lot of time to this but the main message from this debate in India about whether growth, improvements in social 21 Isolation has been found to be the most important problem faced by the poor in many African communities for which participative poverty surveys are available. People of lower caste in India must feel extremely isolated. It was also a point mentioned by Rakesh Kaushik in a written review of this paper following the workshop in New Delhi. 1 National Council for Applied Economic Research, Economic Reforms and the Poor, October 1998. 23 Rakesh Kaushik, personal commnunication. 7 services, or targeted programs are the most important strategy for poverty reduction is that there is no choice. They are all required under different circumstances if broad- based poverty reduction is to be achieved. The question is how to decide the balance among these strategies and choose that combination of strategies, which is the most relevant for each circumstance. The following brief references to some of the basic research provides a flavor of the arguments and the criteria for choice. Requirements for Poverty Reduction 2.2 Ravallion and Datt have concluded that "fostering the conditions for growth in the rural economy ... must be considered central to an effective strategy for poverty reduction in India."24 Growth in yields on farms has also had a strong positive impact on poverty reduction.25 The NCAER paper referred to above finds an association between the incidence of poverty and relative performance of the rural and agricultural sector in India.26 In a recent paper Ravallion and Datt find that the responsiveness of poverty to the growth of farm yields is important in all the states they studied which means that technological change, along with positive incentives to produce, will benefit the rural population in both poor and not so poor states and regions. Obviously higher incomes for farmers resulting from higher yields will create demand by the farm sector for products from the non-farm sectors. This interaction will be positive for growth generally. 27 These conclusions are consistent with those of Mellor and Johnston many years ago and more recently Hazell and Haggeblade, who insisted that rural led growth based on improved agricultural technology would benefit both urban as well as rural populations. Theory and Practice 2.3 It is necessary to match the theory with practical policy. Agriculture provides a highly relevant example. While the growth of agricultural output in aggregate has been positive (around 3.4 percent per annum) over the last decade, the rate of growth of productivity has declined. Reductions in the rate of growth of agricultural productivity have an impact on the rate of change of rural employment and poverty reduction. Research has shown (and common sense would confirm) that the impact of productivity change on poverty reduction is much more evident in the long run than in the short run. For example the elasticity of changes in the incidence of poverty reduction (measured by the headcount index) with respect to increases in average farm yield in India is -0.177 in the short terrn and -0.875 in the long term.29 Sustained growth in yield is therefore an important contribution to long tern poverty reduction. Obviously the declining rice and -4 Martin Ravallion and Gaurav Dant. How Important to India's Poor is tie Sectoral Composition of Economic Growtlz? The World Bank Economic Review, Vol. I0, (m) pages 1-25. s Gaurav Dant and Marnin Ravallion. Farnm Productivity and Rural India, Joumal of Development Studies. Vol.. 34, No. 4. April 1998, pp 62-85. 26 National Council for Applied Economnic Research, op cit, page 113. Ravallion, Martin and Gaurav Dant (1999): When is Growtih Pro-Poor? Evidencefrom tihe Diverse Experience of India's States, Policy Research Working paper WPS 2263, World Bank ' See John W. Mellor and Bruce Johnston, Th7e World Food Equation: Inter-relations among Development. Employment and Food Consumption, Joumal of Economic Literature 22 (June): pp 531-74: Hazell, P.B.R. and Steve Haggeblade, Rural Urban Growti Linkages in India, India Joumal of Agricultural Economics, 46 (4): 515-529. '9 Datt, Gaurav and Martin Ravallion, 1998, Farm Productivity and Rural Poverty in India, Joumal of Development Studies, Vol. 34, No. 4. April 1998. 8 wheat yields between the 1980s and the 1990s (discussed in the companion background paper on rural development) had negative consequences for poverty reduction and may be the reason for the increase in poverty in the Punjab mentioned in paragraph 1.11 of this paper. Of course, while the elasticities indicate that productivity growth has a relatively small impact on poverty reduction in the short term, the short term is crucial in human and political economy terms and hence other measures are required to achieve a comprehensive attack on poverty reduction. This underlines the need for short term targeted interventions in certain circumstances. 2.4 Other measures for addressing poverty include broad-based improvements in primary social services that have both short term (e.g. health, sanitation and water supplies) and long tern benefits (e.g. education). This set of measures is aimed at enhancing human capital development providing people with the assets to take advantage of the employment prospects offered by economic growth. Targeted programs are the other option but they face difficult design questions relating to the tendency for these programs to become very costly, and the difficulty of ensuring that those who need assistance benefit and those who do not need it do not get a "free ride." The rest of the paper is about how the govemment and the Bank have chosen programs for poverty reduction among the various priorities and how the programs have been implemented. 3. Response By Government In The Nineties Public Sector Programs 3.1 Well before the nineties the central and state governments had devoted considerable energy and resources to programs supporting poor consumers through targeted employment, food subsidy (public distribution schemes) and social service programs. The government also financed programs to assist producers with subsidies on inputs such as fertilizers, water and electricity, but the distribution of benefits from such subsidies often favored those who were already better off.30 Price supports for outputs were also (and still are) an important part of the programs,31 although it was to a large extent those with assets (land, irrigation, capital, technology) who reaped most of the benefits from these support programs. Government interventions have ranged from narrowly targeted programs (directed at a specific group based on the welfare of individuals such as the well-known range of employment programs) to broadly targeted programs (based on the welfare of a group). 30 It was estimated that electricity subsidies for agricultural consumers in 1995/96 were an average of Rs. 700 per caput in the high income states, but only Rs. 180 per caput in the low income states. See S.R. Hashim and T. Haque, Policv Reforms in Agricultural Domestic Sector - Role of tile Centre and tile States, Paper presented at the NCAER-IEG-World Bank Conference on Reforms in the Agricultural Sector for Growth, Efficiency, Equity and Sustainability, New Delhi, April 15-16. 1999. It should be added that rates for electricity for agricultural producers were raised by about 50 percent in the States of Rajasthan and Haryana during 2000 although the adjusted rates still do not cover costs and hence a subsidy remains. 31 See India: Towards Rural Development and Poverty Reduction, op cit, Volume One, page 8 for a review of the central and state government market interventions in agriculture and agro-industry. 9 3.2 Although adjustments have been made in government targeted programs, most of them still exist in the nineties in much the same way as they were in the eighties. Many of them (such as the Public-Distribution System) have, however, become increasingly unwieldy and cost ineffective.32 The solution is not to reject targeted programs, rather programs need to be designed that will have an increased impact on the poor and that can at the same time be sustainable on an institutional and fiscal basis. The basic drawback of these publicly managed programs is that they were managed top down with minimal community participation. 3.3 In the early nineties the Government of India tumed a comer and adopted a multi- pronged poverty reduction strategy and it involved a larger role than before for broad- based growth. Nevertheless, it continued to rely on a wide array of broadly and narrowly targeted anti-poverty programs (such as the Public Distribution System and various employment programs) with a renewed emphasis on human capital development. As a continuation of the reform program there have been changes in the PDS and the employment programs such as adjustment in issue prices and the levels of wages paid. ; Impacts of Programs 3.4 The policies of the early nineties included liberalization of the government's control over economic activity, fostering growth of the private sector, the liberalization of trade in manufactured products, reductions in tariffs and quantitative restrictions, and reductions in price controls. For the most part these policies, through their effect on growth and employment, benefited the urban poor, but in some cases also created costs and difficulties for them. For example an NCAER report in October 1998 concluded that the food price increases which accompanied the post 1991 reform period (which were of course important for farmers who produced a surplus), had a disproportionately detrimental effect on consumers in the lowest income groups compared with the effect of such price rises in earlier years.34 The reforms of the early nineties did not include the agricultural sector, except to the extent that policies such as exchange rate adjustments resulted in an improvement in the terms of trade for producers of some commodities, the mirror image effect of prices facing the urban consumers. The increased focus on exports resulted in structural changes and growth in agriculture. Joshi and Little found that concurrently there was a very large rise in food grain prices, resulting in a depression of real rural wages.35 This issue is discussed in the background paper on agriculture and rural development for the CAE. 31 This is not the place for a detailed review, but for a very careful review see for example, R. Radkhakrishna and K. Subbarao, India's Public Distribution Svstenz: A National and International Perspective. World Bank Discussion Paper No. 380. November 1997. See also Raghav Gaiha, Katsushi Imai and P.D. Kaushik, On tihe Targeting and Cost-Effectiveness of Anti-Poverty Programmes in Rural India, mimeo, February 1999: and Martin Ravallion, Appraising Workfare Programs, Development Research Group. World Bank to mention but a few. 33 It should be noted that PDS prices were raised substantially in the 2000 central government budget. 34 See National Council for Applied Economic Research, op cit, Table 3.4. 35 See Joshi, Vijay and I.M.D. Little (1994), India: Marroeconomics and Political Economy, 1964-1991' (Washington DC and New Delhi: World Bank and Oxford University Press). With respect to structural change, note that the Minister for Agriculture reported to the central Parliament in 1996 that some two million hectares of agricultural land had changed from food to cash crops during the post reform period which was meant to partly explain higher food prices. But this was only about 1.6 percent of the total area devoted to foodcrops in 1996, hardly enough to make a big difference to naional prices. This was reported in Poverty and Hunan Development- An India Profile, report prepared for CARE by Sandya Venkateswaran, January 1999. 10 Role of the Private Sector 3.5 A theme running through the previous paragraphs has been the dominant role of government programs in addressing poverty reduction. On the other hand, as the background paper for the India CAE on Openness of the Indian Economy and Private Sector Development states, "the private sector has remained a major source of national income growth as well as employment". A dynamic private sector stimulated by investment incentives should be able to provide a major part of the solution to poverty reduction in India. Indeed the evidence, discussed above, makes it clear that this has happened in urban areas where reductions in urban poverty have been sustained over many years to the present time. The reform programs of the nineties that stimulated manufacturing and trade have further strengthened the downward trend in urban poverty. On the other hand, while rural poverty also declined steadily in the seventies and eighties this was not sustained (or not sustained at the same rate) in the nineties following the introduction of the reform program. While Indian agriculture is certainly private sector based, it has not been a.dynamic source of growing employment opportunities. The reasons almost certainly include a slowdown in agricultural productivity, agricultural policies which reduced production incentives, the constrained capacity of small farms (which are rapidly getting smaller) to employ additional labor, the limited off-farm employment opportunities (see paragraphs 1.13 and 1.16) and the still quite rigid social constraints within the labor market which limit the employment opportunities of specific groups.36 4. Response By The World Bank 4.1 The Bank's impact on poverty reduction in India through its lending programs will inevitably be relatively minor in absolute terms because the Bank's assistance represents only a small fraction of total public investment. On the other hand the Bank can, on the basis of its partnerships with central and state govemments, have a significant impact on public policy by making public policy that has an influence on poverty reduction a central part of its dialogue on the nature and level of assistance. The dialogue should, of course, be backed up by sound research, and "economic and sector work," and technical assistance. Obviously this means that the Bank needs to be diligent and ensure that it has established a coherent assistance strategy for poverty reduction, and that it brings to the dialogue the accumulated experience from its work world wide on strategies that have and have not worked. The CAE mission to India was on many occasions advised that the wealth of Bank experience is not readily available and should be made more readily available during Govemment/Bank dialogue. ` The issues related to employment in agriculture are discussed in the background paper on agriculture and rural development. 11 4.2 It was not until the mid to late nineties that a clear Bank assistance strategy for poverty reduction in India emerged. The 1992 CAS mentioned poverty only once in the context of improving safety nets for the poor. There was little strategic focus on poverty reduction in CASs in the following years until the 1995 CAS that recognized poverty reduction as part of the unfinished agenda from the economic reform program. The renewed focus on poverty emphasized the importance of growth and extending the macroeconomic reforms to other sectors such as agriculture. The 1996 Progress Report on the CAS again emphasized the importance of growth and sectoral reform in agriculture and foreshadowed the District Poverty Initiatives projects as a targeted program that could complement broader reforms. 37 There was little strategic content such as a strategy for the long term pattern of growth. Neither was there much confidence in the contribution that agricultural sector reform could make to poverty reduction. The 1997 CAS contained.the first explicit "country assistance strategy proposals" for poverty reduction (Annex 2 of the CAS). This CAS made it clear that the Bank's assistance strategy was to address poverty on all three fronts suggested on the 1990 WDR. The 1999 CAS Update continued this strategy. As a result there is now considerable energy devoted to ensuring the poverty focus of Bank operations, although monitoring has been weak. Non-Lending Assistance (Research, and Economic and Sector Work - ESW)38 Research 4.3 The Bank's substantial body of research has had a significant impact on the debate on poverty in India. Some of the main highlights of this research and their conclusions have already been mentioned above in sections on the poverty situation and its causes in India. This is not the place to review this work in more detail, but it is useful to identify its main groupings. One group is measurement of poverty, the decomposition of changes in poverty into the growth and redistribution components, and the factors that cause the changes in poverty. The latest contribution in this grouping (mentioned previously) confirms the importance of growth to poverty reduction, but identifies the important factors and household characteristics which influence the ability of households to take advantage of macroeconomic and sectoral policies for poverty reduction. This research underlines the importance of human capital development and the value of household assets for poverty reduction. The results are important for policy purposes since they indicate the types of investment most likely to lead to poverty reduction. The second group of work has evaluated targeted programs such as public distribution programs, the well-known employment programs for the poor, and nutrition programs. The conclusions are that most programs have been excessively costly in fiscal terms and often led to the inefficient allocation of resources. This research is of high quality and 37 Two such projects were approved by the Bank in Andhra Pradesh and Rajasthan in FY 2000 38 There have been many papers and reports produced on issues which impinge on the welfare of the poor such as education, health, and infrastructure, not to mention conferences of various kinds that take place almost every month on social development and poveny and poverty reduction. For example the workshop on poverty reduction in New Delhi in February, 1999, and the subsequent workshop on Reforns in India Agriculturefor Growth, Effiriency, Equiry and Susrainabiliry, in April, 1999. It is not intended in this background note to summnarize the conclusions of all the studies and workshops. 12 used frequently as a background to policy discussions and analysis on poverty reduction strategies in government and academia in India. 4.4 One criticism of the Bank's research work mentioned during the CAE mission to India was that, while much of the research that focused on measurement was of high quality, after a point in time it had limited marginal benefit because the demand in India was more for an analysis of policy options for poverty reduction. It was also regretted that there was still not a robust model that policy makers could use to assess the impact on poverty reduction of different policies and strategies. An analysis of the impact of policy change is crucial, but recent research (such as the evaluation of factors that influence the poverty reduction impact of growth) has gone a considerable way to making the Bank's research more relevant to policy questions. Econzomic and Sector Work39 4.5 There were numerous pieces of economic and sector work on poverty and related issues over the last ten years.4 The first report was the Review of Poverty, Employment and Social Services in 1989.41 It was the country economic memorandum for that year and discussed the substantial progress made in poverty reduction during the eighties and the way ahead. It also included a remarkably good prediction of the difficulties that lay ahead for poverty reduction. It noted that the "complexion and context of Indian poverty have changed." The report listed four issues that were important for a continued poverty reduction in the rural sector. They were the importance of generating productive employment in industry and services because of the limited prospects for increased employment in agriculture, the increasing concentration of poor in some regions, the growing share of the landless, numerous wage dependent households among the poor, and slower growth in public spending on poverty reduction programs. 4.6 The 1991 CEM highlighted policies and public expenditure strategies that at the time constrained the opportunities for income growth in the agricultural sector and drew conclusions on actions that needed to be taken. That report, while emphasizing the serious shortcomings of agricultural policy, was sidelined during the overwhelming focus in the early nineties on macroeconomic and trade issues and had little impact. Except for the 1996 CEM that has a small section on poverty, there was a hiatus of 5-6 years after 1991 before further sector work on poverty was produced. 4.7 Poverty assessments were prepared in 1997 and 1998.42 The first tackled the growth dimension of poverty reduction. The second addressed the role of social services and targeted programs in poverty reduction programs. The chapter on targeted programs in the second poverty assessment was based on a forthright report that contained critical 39 "Economic and sector work" is more focused on operational issues than the Bank's research work. It consists of regular country economic memoranda. reports on specific sectors in the economy, and studies of specific issues. 40' There have been numerous examples of ESW which have been directly or indirectly relevant to poverty reduction such as reviews of the education and health sector, water resources, and forestry to name but a few. These have been discussed in other background papers and will not be discussed here. See, for example, the separate background paper on agriculture and rural development. 4' Bank report No. 7617-IN. India: Poverty, Employment and Social Services, Resident Mission, New Delhi. May 10, 1989. '2 Bank report No. 16483-IN, India: Achiievements and Challenges in Reducing Poverty, May 27, 1997; and Bank report No. 17881- IN, Reducing Poverty in India: Optionsfor More Effective Public Services, June 29, 1998. 13 conclusions on the heavy costs and low efficiency of the government's targeted programs.43 According to an unsolicited statement from one state govemment official, it was well regarded for its analysis and clear suggestions for follow up actions. The poverty assessments were well written and interesting, but essentially they were a re- packaging of previous research work by the Bank, with little material by others such as Indian scholars. The first in 1997 did not break new ground and did not indicate new operational modalities, strategies, or a re-balancing of strategies for the Bank or the government for poverty reduction. It should be recognized, however, that the 1997 report provided an important baseline of knowledge about poverty in India that the Bank had heretofore not been able to acquire. That was an important breakthrough. The second poverty assessment in 1998 was more strategic on the role of social services. The govemment has said that it appreciated the various pieces of ESW, but it is not clear what the value added of the two poverty assessments was to the Govemment's strategic thinking on poverty reduction. On the other hand, it is known that the government was not at first favorable to the Bank undertaking any poverty assessment for India and therefore the fact that such reports were produced at all is a credit to the increased collaboration between the government and the Bank in addressing difficult and sensitive issues. 4.8 A Bank-financed effort in poverty analysis of a completely different kind was a qualitative analysis of poverty in 30 villages in eastern Uttar Pradesh and north and central Bihar states. This was an innovative and collaborative approach aimed at bringing perspectives together from different disciplines, using qualitative research techniques drawn from the Rapid Rural Appraisal (RRA) and Participatory Rural Appraisal (PRA) approaches. The essential conclusion from this painstaking work was that poverty is not a simple one-dimensional condition. The study identified different types of poor44 and also unearthed a considerable amount of interesting information on the importance of social capital as a support mechanism which, if absent, left families at the mercy of outsiders for assistance and employment, thus making them extremely vulnerable to chronic poverty. 4.9 Another major report in the second half of the nineties was on Rural Development and Poverty Alleviation in 1999. It followed an intense five year preparation period which involved specific studies of issues such as grain marketing, livestock, sugar, oilseeds, water resources management, targeted programs for poverty reduction, and rural credit. When the report finally emerged it drew similar conclusions to those in the 1991 CEM about the "crisis in agricultural policy" and the costs of controls, the misallocation of resources, and the inadequate development of rural infrastructure. It also noted that, although some policies introduced by the government in the early nineties had improved the terms of trade for producers of some crops, many of the same policies in place in the early nineties still constrained agricultural development and poverty reduction. This report was discussed at a workshop in New Delhi in April, 1999 and received a positive 13 See Radkhakrishna and K. Subbarao, op cit. 44 Indeed there were two types of poor identified in this work, namely the destitute poor and stuctural poor. Assets (broadly defined to include human capital) characterized both types of poor. See Confronting Poverty in India - The Chzallenge of Uttar Pradesht, internal World Bank draft report, June 23, 2000. 14 response. It promises to be a flagship product for intensifying the Region's focus on rural development and poverty reduction and already some projects based on the report's conclusions are in preparation. If its messages can be disseminated effectively its impact will be substantial, but up to the end of the nineties there was no evidence at present of an active program of dissemination or dialogue with the central and state governments on the report's findings and follow-up. In 2000 there has been more focus on the issues raised in the rural development report because there has been a trend in the current central government as well as state governments to address a number of policy issues such as food price support levels and input subsidies (see paragraphs 3.1 and 3.3 above). 4.10 The OED mission to India for the CAE noted four criticisms of the Bank's approach to the analysis of poverty in its economic and sector work. First, the Bank has not developed an adequate mechanism for sharing the household data sets it has on India. This is surprising since the full data sets come from India and perhaps the real issue is the process through which these data are made available to Indian researchers. In addition the data used by the Bank are available on one of its web pages. Another criticism was that papers and reports prepared by the Bank are available in India only at a cost that is impossible to meet for the average researcher or student. This is certainly a problem and perhaps the use of electronic media is a solution. Third, the Bank has not adequately involved Indian researchers in the Bank's poverty research and ESW on a regular basis, or on as many occasions as may have been possible. This does need to be addressed and it should be possible to have the Bank's field office in New Delhi involved in identifying qualified researchers and to make arrangement for the appropriate levels of compensation. Finally, it was said, the Bank has not provided adequate financial support for analytical work on poverty in Indian institutes. Perhaps some of the Bank's projects could address this issue directly with components to support specific research. Evaluation of Non-Lending Assistance 4.11 In the first half of the nineties the relevance of the Bank's research to its poverty reduction strategy was modest because it was dominated by measurement issues and the evaluation of employment programs. This research was no doubt relevant to Indian governments that financed these targeted programs. They were generally not illuminating for the Bank in making choices and allocating resources among long-term investments, although it is known that the govemment in 1989 approached the Bank with a request for assistance with the financing of rural works programs. The Bank responded that no support could be provided unless there was an increase in the effectiveness and productivity of the programs. This response was no doubt given on the basis of the research into the public works programs. So far as the ESW is concemed, namely the 1989 and 1991 CEMs, their relevance and quality were respectively substantial and satisfactory but their impact was negligible because they were virtually ignored by Regional management. 4.12 - In the second half of the nineties the relevance of Bank research work on poverty became substantial because it was more focused on strategic questions such as the characterization of poverty and policy choices. The quality of research continued to be highly satisfactory. CEMs starting with the 1996 report, also became more relevant to 15 actions on poverty reduction, and rural development in particular. The two poverty assessments (the flagship documents on poverty) produced in the second half of the nineties were readable and interesting but were essentially re-statements of research results, and despite some discussion of options for action, contained little discussion of priority actions. The relevance of the 1997 poverty assessment was substantial. The 1998 report was more strategic but provided little further value added over the "country assistance strategy proposals" on social services presented in the 1997 CAS a year earlier (see paragraph 4.2). While it covered targeting issues very well, this area of activity was not relevant to the Bank's lending program. Nevertheless, the overall relevance of the 1998 poverty assessment is judged to have been substantial. The quality of the two poverty assessments was satisfactory. The impact of all ESW can be viewed in two ways. On the one hand they provided an excellent synopsis of the conclusions of major areas of Bank research work on India (with some new material) which has been widely read. On the other hand, their operational impact appears so far to have been modest. 4.13 Reports for sectors such as rural development, education, and nutrition contained important and substantially relevant analysis of satisfactory quality, and they successfully incorporated poverty issues into their analysis of agricultural price policy, ineffectiveness of agricultural subsidies, human capital development, food security and early childhood development. Except for the education report, however, they were somewhat muffled in their operational conclusions and hence the impact of most of the reports, without an intense dissemination program is uncertain. 4.14 The other sector reports, including the 1997 sector report on health, contributed very little to the debate on poverty reduction. Both health sector reports in 1995 and 199746 refer to the difficulties for the poor from "out-of-pocket" payments for health care and the need to target this group for exemptions. The solution, however, was to "strengthen the existing system for targeting the poor rather than create a new mechanism". But in the event the system proposed focused on the identification of those eligible and little else (see 1997 report, pages xiii and 57). Lending Assistance Strategic Focus of Poverty Reduction Strategy 4.15 The Annex table provides a list of all projects financed by the Bank by major sector for the years FY 94 to FY 99 (the primary focal period for this CAE). These projects have been classified according to whether they were likely to be making a contribution to poverty reduction through broad-based growth, broadly targeted social services or narrowly targeted services to the poor. 45 In reviewing the first draft of this paper. the Region made the point that there have been substantial improvements in lending services and that progress has been most appreciable over the past few years. The evaluation of the non-lending assistance in this revised version of this paper is much more positive than in the first draft. ' Bank report No. 1 3042-IN, India: Policy and Finanre Strategiesfor Strengthzening Prinary Healt)i Care Services, Population and Human Resources Division, South Asia Country DepartmenL May 15, 1995; and Bank report No. 15753-IN, India: New Directions in Healtih Sertor Developnment at thie State Level: An Operational Perspective, Population and Human Resources Division. South Asia Country Department, February 11, 1997. 16 4.16 Figure 1 summarizes the information in the Annex. The table shows quite a heavy concentration of Bank projects (on average 58 percent) over these six years are focused on the objective of achieving growth and implicitly poverty reduction. Another 34 percent are focused on broadly targeted social services as vehicles for poverty reduction. Figure 1 shows the increase in lending for broadly targeted assistance in FY95 and particularly in FY 96 that reflects the increase in lending to the health and education sectors. The emphasis on growth increased over the last two years due to a substantial number of new irrigation projects (including water resource consolidation projects), and the National Technology project that should contribute to long run growth in agricultural productivity. 4.17 Of course year to year changes (such as the substantial decline in lending during FY 99) should not be the focus, rather the proportions over time are more relevant. The general pattern indicates that the Bank's assistance program has focused mainly on supporting broad-based growth and employment generation through investments or by contributing to establishing the enabling policy or institutional environment for growth. Obviously a substantial part of the Bank's potential impact on poverty reduction will need to come from these projects since growth and employment generation are necessary conditions for poverty reduction, but this analysis says nothing about whether growth is pro-poor or not. 17 Figure 1: Summary of Strategic Poverty Reduction Focus of Bank Lending (FY94-FY99) 2500- fto Kioo E C.)L FY 94 FY 95 FY 96 FY 97 FY 98 FY 99 1* Growth M Broadly Targeted is Narrow ly Targeted Poverty-Targeted Interventions (PTIs) 4.18 Since 1992 the Bank has rated all investment projects submitted to the Executive Board as either a "poverty-targeted intervention" or not. Projects are given a "yes or no" PTI rating. Broadly, the rating reflects an assessment by the Bank about the expected impact of the project on poverty reduction .4 Most projects, if they qualify as a PTI, do so according to the second (weaker) criterion mentioned in the footnote, namely the proportion of poor among the project beneficiaries is higher than the proportion in the population as a whole. There is usually no mechanism to verify (or subsequently monitor) whether the potential beneficiary group has a larger proportion of poor than the population as a whole, nor whether benefits from the project will pass to the poor. The poor group may be quite broad. For example, the District Primary Education Projects (DPEPs) have been classified as "poverty targeted" and therefore it is presumed that the proportion of poor primary students among those enrolled in the program is greater than the proportion of poor in the total population. While these projects are focused on the poorest districts, there are no data available to verify the proportion of students from poor households. 47 Projects are classified as "Poverty targeted interventions" if "(a) they have a specific mechanism for targeting the poor and/or (b) the proportion of poor people among its beneficiaries is significantly larger than the proportion of poor in the total population". If projects qualify according to the '(a)" criterion they are referred to as showing narrow targeting, but if they qualify according to the "(b)" criterion, they are referred to as showing broad targeting. 18 4.19 The results for India for all years since this classification of projects started are shown in Figure 2 below. It shows that Indian projects have been rated quite high in terms of the PTI criterion, namely about 56 percent for the FY94-FY99 period. Indeed the India projects have scored well compared with all IDA projects for which the PTI percentage has typically been around 54 percent. It should be said that there is substantial skepticism regarding this classification within the Bank since many projects are seen as getting a "free ride" into the PTI classification because the criteria are quite generous and are often based on minimal verification. Moreover actual benefits for the poor are rarely measured. Figure 2: India: Relative Importance of Poverty-Targeted Interventions (FY94 - FY99) 2500 2000- 0 1500- z500 ioooEN O-WE FY 94 FY 95 FY 96 FY 97 FY 98 FY 99 j Pbverty Targeted a Not Poverty Targeted Poverty Impact of the Lending Program 4.20 There is little doubt that a number of the Bank-assisted projects will have a positive impact on poverty reduction, either through their impact on growth and employment, increased income of the direct beneficiaries in agricultural projects, or in terms of some direct effect through various types of targeting such as in the Bank-assisted rubber and sericulture projects. 4.21 Nevertheless, the precise impact of Bank-assisted projects on poverty reduction in India is impossible to measure because there has been no systematic monitoring of the 19 poverty reduction.48 Some agricultural projects such as irrigation projects contain inforrnation (reported in Project Status Reports) on changes in incomes of participating farmers, but this information is usually only directly relevant to those with assets such as land and only indirectly relevant to those without assets through a (usually unmeasured) employment effect. In forestry projects there are data on employment effects during implementation, but it is not clear whether this impact will be sustained after project completion. Projects aimed at improving social services such as the District Primary Education projects are clearly important to human capital development and they collect data on enrollment, drop outs and repetition rates. The following are some examples. * The Assam Rural Infrastructure project (approved in FY95, US$126 million) has, led to an acceleration of the growth rate in paddy and horticultural production. This is presumed to have had a positive impact on employment, but basically it is those with land and other assets who have benefited. Similar conclusions can be reached for other irrigation-based projects. * The Forestry Projects which have implemented Joint Forest Management (JFM) based on community involvement in forest management are expected to provide benefits for the poor. But, as the draft India Country Case Study for the OED Forest Sector review states, "... a large part of the (community) benefits to the poor have been through wages" during implementation of forest rehabilitation activities (page 40). The question is, will these benefits be sustained through other project activities once the rehabilitation work has been completed. It is generally acknowledged that poverty reduction was never the original main design goal in the series of forestry projects during the nineties, rather it was improvement of the public sector management of forest resources. * The District Primary Education projects (DPEPs) now have improved monitoring systems for primary school enrollments, repetition and dropouts that will provide information on the progress of participating district schools. So far the gains have been "more noteworthy for girls than for other socially disadvantaged groups, scheduled tribes, other minority groups, children with special needs and working children"49 This is a major achievement for gender and social balance, and certainly encouraging for long term poverty reduction since these projects are being implemented in the poorest districts in the poorest states. 4.22 There was, during the last ten years only one Bank-financed project designed to assist the government with its broad and narrowly targeted poverty reduction programs, namely the Social Safety Net project. * The Social Safety Net Project provided assistance to the govemment (approved in December 1992, $500 million) which was highly relevant as it aimed to help the central government protect its social spending and facilitate the ongoing structural A criticism of the Bank made at the "entry workshop" on poverty issues in New Delhi in May, 1999 by an invited Indian expert was that despite "pumping in huge amounts, the Bank had no monitoring and evaluation of projects." 49 Aide Memoire for Ninth Joint Review Mission (April 9-23. 1999), for India: District Primary Education Program. 20 adjustment program at a time of severe budget pressures. The central objective of the Social Safety Net (SSN) program was to finance key social programs such as primary education, primary health care, disease control, nutrition and for labor retrenchments in public enterprises. Without such additional assistance the govemment may not have been able to continue its targeted programs which at that time were seen as being even more necessary because of the secondary effects of the macroeconomic crisis (for example inflation had reached 17 percent per annum in August 1991 and there were considerable pressures on the foreign exchange account) and because of the necessary belt-tightening that followed. * The OED evaluation of this project concluded that the project had only a moderately satisfactory outcome and limited-impact, with likely sustainability but negligible institutional development. The project missed a golden opportunity to review the social and economic justification for a number of targeted programs. Instead, many inefficiently targeted programs still exist at substantial costs to the budget. Current Moves to More Targeted Programs 4.23 The Region has moved quite vigorously toward financing a number of targeted programs in three states (the District Poverty Initiatives Projects in Rajasthan, Andhra Pradesh and Madhya Pradesh) which have the hallmarks of social funds plus significant institutional development components. 50 These projects rely on the existing institutional structure to provide support to Community Interest Groups (CIGs) in identifying, designing and implementing projects demanded by the communities and then maintained by them. There is a formal but simple ex ante evaluation procedure that is passed through the village councils (Gram Panchayats) for review and to a District Office for approval.5' A recent review by a Quality Assurance Group panel in the Bank reviewed these projects. While accepting that there were some risks that the institutional framework for DPIP projects may not be sustained, the panel regarded them as carefully prepared in great detail, with sound "quality at entry," and strong govemment support. The panel concluded that they were well worth appraising. 4.24 This initiative took five years to prepare and so the details are many. An important recent paper which provides a framework for institution building for social fund projects, a joint effort between the social and agricultural groups in the Bank's New Delhi office, provides a logical framework for these projects. If successfully implemented as designed they have good prospects for making a substantial impact on poverty reduction. Fundamentally these projects aim at the development of social capital at the community and village level as the basis for making faster progress on poverty reduction among the poorest groups. '° DPIPs for Andhra Pradesh and Rajasthan were approved in April 2000 "' It is also possible for the CIGs to take a project directly to the District project unit. 21 Evaluationi of Lenldinlg Assistance 4.25 CASs insist that the Bank's assistance strategy has the over-arching objective of 52 C poverty reduction. It is, however, impossible to assess the outcome of the program in terms of poverty reduction because of the absence of any systematic monitoring of the outcomes on poverty or other welfare measures. While many of the projects have poverty reduction as an objective there is usually little connection between the objectives and the subsequent actions. For example in irrigation projects most of the benefits accrue to those who hold land and who by definition are not the poor. There are no projects in the current cohort which address land policy issues although in the context of the Sodic Lands Reclamation Project women were given co-owner status of land reclaimed and allocated to poor individuals as part of the project.53 In addition many of the specific measures to assist the poor and improve the efficiency of resource allocation for the benefit of the poor are lagging in projects.54 These are serious shortcomings in an otherwise increasingly well balanced, focused program of assistance and an increasing attention to poverty reduction. 4.26 On balance, the outcome of the Bank's completed and ongoing projects as a whole with respect to their achievement of poverty reduction is expected to be moderately satisfactory. The reason for this rating is that, while the program has been focused on outputs (including growth) from projects, its relevance to poverty reduction was by and large modest in the early nineties tending to substantial in the late nineties. Efficacy in terms of poverty reduction can only be assessed if there is monitoring. Obviously the Bank cannot shoulder the whole burden of achieving poverty reduction in India, but as the 1999 Rural Development report has stated, the 1990s saw a "less poverty alleviating growth."55 On the other hand the Bank was in the past not focused sharply and strongly enough on the necessary enabling conditions such as institutions, agricultural production incentives and agricultural technology which would have enhanced the prospects for the growth of rural employment and incomes, and poverty reduction. 4.27 Institutional development for many Bank projects was modest in most cases but sometimes negligible because of its penchant for establishing project implementation units.56 They cannot possibly develop the regular institutions for future work of the same kind. On the other hand there are projects (such as the series of forestry projects) which have focused on strengthening the existing forestry institutions and playing a catalytic role in bringing policy and institutional issues "to the table" for discussion and resolution. This strategy is making good progress. In general, however, the common problem in ongoing projects is institutional and therefore on balance the institutional development 52 The most recent example being the Executive Summary of the India: Country Assistance Strategy - Progress Repon, January 27, 1999. 53 This was an agreement reached with state authorities as part of the implementation the Sodic Lands Reclamation project in Uttar Pradesh. 5' See the background paper for.the CAE on Agricultural and Rural Development for details of components in irrigation projects such water users associations and special programs for the poor which are behind schedule in comparison with construction which is usually on schedule. 55 India: Towards Rural Development and Poverty Alleviation. op cit. page 30. 56 While project management or coordinating units are still being established. there is less emphasis on this in recent years. 22 for poverty reduction arising out of Bank projects is likely to be modest. For some projects such as the DPEPs the institutional development'is more likely to be substantial because these projects have established an institutional framework which is capable of focusing effectively on the primary education of the poor. The sustainability of projects in terms of poverty reduction is unlikely to uncertain because the loaical framework for a poverty focus has not been established. 5. Overall Evaluation Of The Poverty Focus Of The Bank's Assistance Program 5.1 The Bank's total assistance program over the last ten years, as reflected in project documents and project status reports, did not give an air of urgency about poverty reduction in India. It is true that CASs have always placed poverty reduction on the agenda as an important objective but it was not until the latter part of the nineties that any real fire was put behind this objective. Projects were negotiated earlier in the nineties that included poverty reduction in their objectives but the effort put into that objective and the outcome in terms of poverty reduction was only moderately satisfactory. Projects were certainly relevant to energy production, infrastructure construction, delivery of health and education services, and to agricultural production in the early nineties. There was certainly an increase in lending for primary social services in the first half of the nineties. Monitoring and evaluation systems to assess the results for poverty reduction were usually non-existent and therefore efficacy is more difficult to assess. Concurrently the Bank's non-lending work as a whole on poverty was satisfactory, but by and large (assigning greatest weight to ESW) the impact was modest since the work was largely ignored at the operational level. On balance, the outcome of the total Bank assistance program in terms of poverty reduction was moderately satisfactory in the first half of the nineties, since strategies were weak and most projects were more relevant to production increases without very much evidence that substantial increases in employment (and therefore poverty reduction) were being generated. 5.2 The second half of the nineties certainly saw improvements in the poverty focus of Bank assistance programs. CASs became more focused on the issues of poverty but non-lending assistance was less than satisfactory. The lending program was partly business as usual with large irrigation and infrastructure projects but it also had a substantial component of broadly targeted assistance such as primary social sector projects. 5.3 As mentioned earlier there was a "sea change" in the latter part of the nineties in respect to the strategy and focus of the India assistance program on poverty reduction. There was also increased attention by the Bank, along with the government, to the measurement of poverty at the state and national levels. As mentioned above, the result is now considerable uncertainty about the overall trends in poverty in India. This uncertainty is, of course, out of the Bank's control but it does have implications for claims the Bank may wish to make in its country assistance strategy about reductions in 23 poverty. The Bank is to be congratulated, however, on its assistance to the government to improve the accuracy of poverty measurement in India. These developments and the Bank's assistance are relatively recent and are mainly relevant to the future lending program. It does not change the evaluation for the whole of the nineties from a moderately satisfactory outcome in terms of poverty reduction. 5.4 The institutional development impact of projects for poverty reduction is expected to be modest but in some cases negligible because Bank-assisted projects have not always developed sustainable institutional structures to address poverty. Sustainability of projects as mechanisms or models for poverty reduction is uncertain because in most cases a clear replicable replicable model for economic development and poverty reduction has not yet evolved. 6. Conclusion - A Vision For Poverty Reduction Strategies 6.1 This paper has referred to a number of issues important to poverty reduction in India. In order of priority they are * Macroeconomic stability-to reduce the "inflation tax" on the poor. * Stimulate growth in poorer (and often less progressive) states and regions (rural and urban) and investment opportunities and establish programs to stimulate "pro poor" growth. * Rural development-growth in rural areas (farm and non-farm) to provide employment for all labor, but increased farm productivity and production will create many jobs for the poor. In this context the private sector has an important contribution to make in achieving widespread poverty reduction. * Improved social services on a wide front to improve human capital and provide the poor with the "initial conditions" to participate in the labor market for off farm jobs. * Support improvements in the structure, quality and sensitivity of governance at the district, community and village levels to enhance the rate of development of social capital so that effective and efficient targeted programs of assistance to the poor can be implemented that will provide benefits directly to the poor. * Poverty Monitoring and Evaluation is a necessary activity without which neither the Government's actions nor the Bank's assistance program for poverty reduction can be evaluated. 6.2 Macroeconomic stability. This is endangered at present by the government's serious fiscal problems. Addressing these problems is an important first order of business for the economy as well as the poor because inflation created by an increasing budget deficit will be highly damaging to the welfare of the poor and will undo progress on poverty reduction already made. The Bank should emphasize macroeconomic stability as one of the most important poverty reduction strategies open to the government. 24 6.3 Stimulating growth in poor states and regions. An increase in the rate of growth and employment is of course at the heart of the Government's strategies, and supported by the Bank. The issue for poverty reduction is that this growth needs to be "pro poor" and therefore take place in poor regions and cities. Hence the government's infrastructure development (supported by the Bank) needs to focus on backward states and regions which also have a policy framework which provides incentives for investment by the private sector. The Bank's current emphasis on supporting policy reform in the progressive states should be maintained and strengthened because of the importance of improving the incentives for private sector investment for the creation of employment and poverty reduction, but this effort should be complemented by an intensive dialogue with the less progressive states aimed at enhancing the "initial conditions" for poverty reduction. A strengthening of the focus on the poorer states could be based on supporting analytical work designed to reveal the substantial marginal benefits of policy change for the poor. 6.4 Stimulate rural development. A disappointing aspect of the recent economic performance has been the relative stagnation in rural poverty at around 36 percent, although as discussed above this may be too pessimistic an estimate. But even the more optimistic estimates of the incidence of poverty are depressingly high in the light of the substantial aggregate growth of the economy. To a large extent the ongoing high rates of poverty in rural areas are due to inappropriate policies in agriculture. Policies that need to be rectified are the responsibility of both central and state govemments and include those resulting in excessive regulation of trade creating a disincentive effect on investment and production, and those associated with the misallocation of public expenditures through subsidies for rural power and irrigation water which still exist despite some reductions. It is also necessary to foster research to stimulate improvements in total factor productivity (including the arrest of the declining growth rate of foodcrop yields) since improvements in productivity has been shown to have a strong positive impact on poverty reduction.. Finally, the serious shortfall in rural infrastructure development and maintenance needs to be addressed. All these factors have reduced private investment and the rate of growth of employment and poverty reduction in rural areas. They need to be urgently addressed. Conclusions from Indian academic and Bank analyses of the situation coincide and there is also a solid understanding of the problems within the Government. Given the consensus on what needs to be done, the Bank should not continue to lend in the agricultural sector without improvements in the policy framework. Now is the time to act in collaboration with the central and state governmerits on eliminating inefficient agricultural and rural policies to allow dynamic growth of agriculture, increased employment opportunities and poverty reduction. 6.5 Continue to improve social services. While India lags in its average level of human capital development, progress has been made. But the emphasis should be on accelerating this progress to ensure that those looking for work in the expanding Indian economy are qualified and ready to take part in its economic development and continued growth. Research clearly shows that deficiencies in assets such as literacy and technical skills are devastating in their negative effect on achieving "pro poor" growth. Improving a range of "initial conditions" for growth are necessary and the Bank should insist that 25 public expenditure programs should address these issues as essential'conditions associated with any development program focused on increasing the rate of economic growth and reducing poverty. 6.6 Improve social capital. One of the pervasive shortcomings in govemance at the district, community and village levels has been the ineffectiveness of efforts by the poor to have an impact on the allocation of financial resources. The Bank is now embarking on a series of programs with state govemments to empower community and village institutions with greater influence and decision making opportunities in the allocation of local resources. This program (District Poverty Initiatives projects) should be pursued, carefully monitored, and if shown to be successful expanded to other states. The Bank should take the risk on these types of projects to not only address poverty reduction but also to make local governance more "pro poor". 6.7 Poverty monitoring and evaluation. It is essential that central and state govemments (in collaboration with the Bank) agree to achieve a major improvement in poverty monitoring for each ongoing and future investment operation. It is not credible for the Bank to be investing, in partnership with the govemments, if neither partner is interested in monitoring the outcome of the most important objective of those investments. 6.8 Closing comment. These issues should be the basis for the country dialogue at the central and state govemment levels. A necessary condition for such a dialogue will be sound analysis of policy options through research, and economic and sector work, and the monitoring of outcomes of investments. This dialogue, analysis, investment for growth, rural development, human and social capital development, and improved governance are what the "second wave of reforms" should be based on. The Bank must, however, stand firm on the fundamental principle that financial assistance without an acceptable policy framework is not likely to yield results and is therefore not acceptable. There is enough evidence that central and state govemment policy analysts agree with this proposition but the political process for achieving policy change usually stalls before changes can be made. 6.9 Despite progress on human capital development during most of the nineties the Bank treated poverty reduction more as an after thought than the guiding light of its assistance strategy. It focused too much on assistance benefiting those who already have assets and who can take advantage of the new opportunities created by the structural reforms. This strategy generally ignored the poor with relatively few assets. There are strong signs that this approach has changed considerably over the last two years or so leading to a much stronger focus by the Bank on ensuring that development programs benefit the poor. 26 Annex 1: Summary of Projects Approved from FY94 to FY99, their Poverty Focus, and PTI Ratinga Project Project Name Net Approval Revised Program objective Poverty Focus Strategy Type PTI ID Commit FY Closing Rating US$M Date Ag ric ulItu re Growth Broadly Narrowly Targeted Targeted 41264 Watershed Mgt. Hills II 135.0 1999 3/31/05 Improve Agric. Productive Potl. in 5 states Yes Yes 50646 UP Sodic Lands il 194.0 1999 9/30/05 Environmentally Sustainable Dev. Yes Yes 10561 Natl. Agric. Technology 196.8 1998 12/31/03 Environmentally Sustainable Dev. Yes No 35824 UP Diversified Agric.Support 179.9 1998 3/31/04 Poverty Red. and Human Resource Dev. Yes Yes 35169 UP Forestry 52.9 1998 7/31/02 Environmentally Sustainable Dev. Yes Yes 49477 Kerala Forestry 39.0 1998 12/31/02 Environmentally Suslainable Dev. Yes No 36062 Ecodevelopment 28.0 1997 6/30/02 Environmentally Sustainable Dev. Yes Yes 35158 AP Irrigation iII 325.0 1997 1/31/03 Environmentally Sustainable Dev. Yes Yes 10529 Orissa WRCP 290.9 1996 9/30/02 Environmentally Sustainable Dev. Yes Yes 10506 MP Forestry 58.0 1995 12/31/99 Environmentally Sustainable Dev. Yes Yes 10503 Agric. Human Res. Devlpt. 59.5 1995 12/31100 Poverty Red. and Human Resource Dev. Yes_ No 10522 Assam Rural Infrastructure 126.0 1995 12/31/03 Economic Management Yes Yes 10476 Tamil Nadu WRCP 282.9 1995 3/31/02 Environmentally Sustainable Dev. Yes No 10449 AP Forestry 77.4 1994 9/30/00 Environmentally Sustainable Dev. Yes Yes 10448 Forestry Research/Educ. 47.0 1994 12/31/99 Environmentally Sustainable Dev. Yes No 9964 Water Resources Consol. 258.0 1994 12/31/00 Environmentally Sustainable Dev. Yes Yes Education _ | 45050 DPEP (Rajasthan) 86.0 1999 12/31/04 Poverty Red. and Human Resource Dev Yes Yes 27 Annex 1: Summary of Projects Approved from FY94 to FY99, their Poverty Focus, and PTI Rating @' Project Project Name Net Approval Revised Program objective Poverty Focus Strategy Type PTI ID Commit FY Closing Rating USSM Date 38021 DPEP iII (Bihar) 152.0 1998 9/30/03 Poverty Red. and Human Resource Dev. Yes Yes Education (cont'd) 50638 UP Basic Education 59.4 1998 9/30/00 Poverty Red. and Human Resource Dev. Yes Yes 35821 DPEP 11 425.2 1996 6/30/03 Poverty Red. and Human Resource Dev. Yes Yes 10464 District Primary Education 260.3 1995 3/31/02 Poverty Red. and Human Resource Dev. Yes Yes Power 49537 AP Power (APL I) 134.0 1999 8/31/03 Economic Management Yes No 35160 Haryana Power (APL I) 60.0 1998 12/31/00 Economic Management Yes No 35170 Orissa Power Sector 350.0 1996 12/31/02 Economic Management Yes No Finance 39935 ILFS Infrastructure Finance 205.0 1996 9/30/01 Private Sector Development Yes No 10563 Financial Sector Devlpt. 500.0 1995 10/31/02 Private Sector Development Yes No Environment 43728 Environmental Cap. Bldg. 50.0 1997 6/30/03 Environmental Management Yes No 108 ~T/AI____ 10485 JHydrology Project 142.0 1996 3/31/02 Environmental Management Yes No 10463 Indust. Pollution Prevention 166.0 1995 3/31/02 Environmental Management Yes No Population, Health and Nutrition 45051 Second Nati. AIDS/HIV 191.0 1999 7/31/04 Poverty Red. and Human Resource Dev. Yes Yes 50651 Maharashtra Health System 134.0 1999 3/31/05 Poverty Red. and Human Resource Dev. Yes Yes 35827 Women & Child Devlpt. 300.0 1998 9/30/03 Poverty Red. and Human Resource Dev. Yes Yes 28 Annex 1: Summary of Projects Approved from FY94 to FY99, their Poverty Focus, and PTI RatingX Project | Project Name Net Approval Revised Program objective Poverty Focus Strategy Type PTi ID Commit FY Closing Rating USSM Date 10496 Orissa Health System 76.4 1998 3/31/04 Poverty Red. and Human Resource Dev. Yes Yes Population, Health and Nutrition (cont'd) 10473 Tuberculosis Control 142.4 1997 12/31/02 Poverty Red. and Human Resource Dev. Yes Yes 10511 Malaria Control 164.8 1997 3/31/03 Poverty Red. and Human Resource Dev. Yes Yes 10531 Reproductive Health I 248.3 1997 3/31/03 Women in Development Yes Yes 44449 Rural Women's Devipt. 19.5 1997 12/31/03 Women in Development Yes No 35825 State Health System II 350.0 1996 3/31/02 Poverty Red. and Human Resource Dev. Yes Yes 10489 AP First Ref. Health System 133.0 1995 3/31/02 Poverty Red. and Human Resource Dev. Yes Yes 10455 Blindness Control 117.8 1994 6/30/01 Poverty Red. and Human Resource Dev. Yes Yes 10457 Population IX 88.6 1994 12/31/01 Poverty Red. and Human Resource Dev. Yes Yes Social protection 49385 IAP Econ. Restructuring 543.2 1998 3/31/04 Private Sector Development Yes Yes Transportation 49301 AP Emergency Cyclone 150.0 1997 7/31/00 Environmentally Sustainable Dev. Yes No 45600 TA State Roads Infrastruc. 51.5 1997 12/31/01 Environmentally Sustainable Dev. Yes No 9995 State Highways I (AP) 350.0 1997 1/31/03 Environmentally Sustainable Dev. Yes No 9870 Container Transport 94.0 1994 12/31/00 Environmentally Sustainable Dev. Yes No Urban Development 50637 |TN Urban Development II 105.0 1999 11/30/04 Environmentally Sustainable Dev. Yes No 34162 Maharashtra Earthquake 216.8 1994 12/31/98 Poverty Red. and Human Resource Dev. Yes No 29 Annex 1: Summary of Projects Approved from FY94 to FY99, their Poverty Focus, and PTI Rating X Project Project Name Net Approval Revised Program objective Poverty Focus Strategy Type r PTI ID Commit FY Closing Rating US$M Date Water SuPPlV & Sanitation 10484 UP Rural Water Devtpt. 59.6 1996 5/31/02 Women in Development Yes No 10480 Bombay Sewage Disposal 192.0 1996 12/31/02 Environmentally Sustainable Dev. Yes No 10461 Madras Water Supply II 86.5 1995 6/30/02 Environmentally Sustainable Dev. Yes No Mining 9979 Coal Sector Rehabilitation 517.0 1998 6/30/03 Environmentally Sustainable Dev. Yes No 43310 Coal, Env & Soc. Sec 63.0 1996 6/30/01 Environmentally Sustainable Dev. Yes Yes Mitigation_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __ __I I_II__ __ a/ PTI = Poverty Targeted Intervention. See paragraph 4.18 of the main text for an explanation. 30 Annex 2: Summary of Discussion at CAE Workshop on Agriculture, Rural Development, and Poverty Reduction April 5, 2000 Introductory presentation drew attention to: * Relatively little World Bank (and Government) attention to agricultural development issues and limited concern about poverty reduction in the first half of the 1990s. * A substantial pick up in the Bank's focus on agricultural development and poverty reduction in the second half of the 1990s. * Overall OED concluded that Bank-assisted agriculture and rural development programs had inadequate focus on poverty reduction during the 1990s despite sound results in forestry and watershed development. * On balance the outcome of the agriculture, rural development and poverty reduction programs were judged to be moderately satisfactory, with modest institutional development and sustainability. Discussion The chair noted three issues emerging from the paper: * The stagnation of agricultural growth; * The impact of this deceleration of growth on poverty reduction * What should have been done by the Bank and the Govemment during the 1990s. The chair also noted two important institutional changes that were not highlighted in the background paper, namely: * The emergence of decentralized government through the Panchyat Raj institutions * The recent revamping of the Public Distribution System. Agricultural Growth and Performance * Data suggested that growth in agriculture , while fluctuating annually, was similar in the 1990s compared with the 1980s. Nevertheless, many in the audience observed that agriculture was nevertheless performing poorly. For example, the growth in grain yields and the rate of growth of grain production had both declined, although it was recognized that these trends were part of an overall structural adjustment in agriculture resulting in less emphasis on traditional grain crops. * It was agreed that the decline in public investment in Indian agriculture during the 1980s was a contributing factor to slow growth in agriculture during the 1990s. * There was agreement on the slow down in growth of real rural wages and also the collapse of non-farm employment-with obvious impacts on incomes and poverty. Dialogue on Agricultural Policy * There was also agreement that the Bank had not interacted closely enough with either central Government/states on constraints of agricultural policy to rural development. 31 * No disagreement that the Bank's dialogue on agricultural and rural development policy with the states had, increased in the late 1990s. * It was suggested that the background paper should point out that agricultural development would probably have been much worse had it not been for the changes in macroeconomic policy such as exchange rates and trade in the early 1990s. Poverty Reduction * There was uncertainty about actual trends in rural poverty given that the NSS data (used in the background paper) showed a stagnation in rural poverty during the 1990s but very recent analysis by Deaton and Tarozzi showed that poverty had declined in some years during the 1990s. * Nevertheless, there was a consensus that if rural poverty had declined during the 1990s, the decline was relatively small because of droughts and other problems in the agricultural sector. * It went without saying that the overall national trends in rural poverty could not be tied to or changed by the Bank's assistance program because of its relatively small impact on India as a whole. The background paper on poverty should make this clear. Performance of Bank-Assisted Projects * It was suggested that the Bank should focus even more on supporting research, rainfed agriculture and forestry than at present. At the same time there was also support for irrigation projects. * On the evaluation of projects, they should be evaluated against their original objectives (Extension), namely support for irrigated grain production. Also, care should be taken with criticism in terms of short term results (Upper Krishna irrigation) which in the longer term should be seen as opportunities for learning rather than crucial failures. * Emphasis was placed on the importance of forestry and watershed projects for the future along with rural roads. * Bank should continue to focus on institutional issues for small farmers who characteristically had few supporting institutions. Future Strategy * There was no objection to the focused approach recommended in the paper. * Bank's focus on reforrning states should be addressed in terms of where there is demand for support. * Bank should see its support as synergistic in terms of the rural sector as a whole.