WORLD BANK DISCUSSION PAPER NO. 380 k in progress public discussion Naov |1qq Inidia's Public I)istribution Svstem R. IR.\J/ukri h ,,/l , , K. SK,//,,a/ - 'l it/ S . h,,,r,,ka,,,t ,,,,,/ C: Ra\,, Recent World Bank Discussion Papers No. 307 The Uruguay Round and the Developing Economies. Edited by Will Martin and L. Alan Winters No. 308 Bank Governance Contracts: Establishing Goals and Accountability in Bank Restructuring. Richard P. Roulier No. 309 Public and Private Secondary Education in Developing Countries: A Comparative Study. Emmanuel Jimenez and Marlaine E. Lockheed with contributions by Donald Cox, Eduardo Luna, Vicente Paqueo, M. L. de Vera, and Nongnuch Wattanawaha No. 310 Practical Lessonsfor Africa from East Asia in Industrial and Trade Policies. Peter Harrold, Malathi Jayawickrama, and Deepak Bhattasali No. 311 The Impact of the Uruguay Round on Africa. Peter Harrold No. 312 Procurement and Disbursement Manualfor Projects with Community Participation. Gita Gopal No. 313 Harnessing Information for Development: A Proposal for a World Bank Group Strategy. Eduardo Talero and Philip Gaudette No. 314 Colombia's Pension Reform: Fiscal and Macroeconomic Effects. Klaus Schmidt-Hebbel No. 315 Land Quality Indicators. Christian Pieri, Julian Dumanski, Ann Hamblin, and Anthony Young No. 316 Sustainability of a Government Targeted Credit Program: Evidencefrom Bangladesh. Shahidur R. Khandker, Zahed Khan, and Baqui Khalily No. 317 Selected Social Safety Net Programs in the Philippines: Targeting, Cost-Effectiveness, and Optionsfor Reform. Kalanidhi Subbarao, Akhter U. Ahmed, and Tesfaye Teklu No. 318 Private Sector Development During Transition: The Visegrad Countries. Michael S. Borish and Michel Noel No. 319 Education Achievements and School Efficiency in Rural Bangladesh. Shahidur R. Khandker No. 320 Household and Intrahousehold Impacts of the Grameen Bank and Similar Targeted Credit Programs in Bangladesh. Mark M. Pitt and Shahidur R. Khandker No. 321 Clearance and Settlement Systemsfor Securities: Critical Design Choices in Emerging Market Economies. Jeff Stehm No. 322 Selecting Development Projects for the World Bank. Jean Baneth No. 323 Evaluating Public Spending: A Frameworkfor Public Expenditure Reviews. Sanjay Pradhan No. 324 The Bangladesh Rural Advancement Committee's Credit Programs: Performance and Sustainability. Shahidur R. Khand- ker and Baqui Khalily No. 325 Institutional and Entrepreneurial Leadership in the Brazilian Science and Technology Sector:Setting a New Agenda. Edited by Lauritz Holm-Nielsen, Michael Crawford, and Alcyone Saliba No. 326 The East Asian Miracle and Information Technology: Strategic Management of Technological Learning. Nagy Hanna, Sandor Boyson, and Shakuntala Gunaratne No. 327 Agricultural Reform in Russia: A Viewfrom the Farm Level. Karen Brooks, Elmira Krylatykh, Zvi Lerman, Aleksandr Petrikov, and Vasilii Uzun No. 328 Insuring Sovereign Debt Against Default. David F. Babbel No. 329 Managing Transboundary Stocks of Small Pelagic Fish: Problems and Options. Max Aguero and Exequiel Gonzalez No. 330 China: Issues and Options in Greenhouse Gas Emissions Control. Edited by Todd M. Johnson, Junfeng Li, Zhongxiao Jiang, and Robert P. Taylor No. 331 Case Studies in War-to-Peace Transition: The Demobilization and Reintegration of Ex-Combatants in Ethiopia, Namibia, and Uganda. Nat J. Colletta, Markus Kostner, Ingo Wiederhofer, with the assistance of Emilio Mondo, Taimi Sitari, and Tadesse A. Woldu No. 333 Participation in Practice: The Experience of the World Bank and Other Stakeholders. Edited by Jennifer Rietbergen-Mc- Cracken No. 334 Managing Price Risk in the Pakistan Wheat Market. Rashid Faruqee and Jonathan R. Coleman No. 335 Policy Optionsfor Reform of Chinese State-Owned Enterprises. Edited by Harry G. Broadman No. 336 Targeted Credit Programs and Rural Poverty in Bangladesh. Shahidur Khandker and Osman H. Chowdhury No. 337 The Role of Family Planning and Targeted Credit Programs in Demographic Change in Bangladesh. Shahidur R. Khand- ker and M. Abdul Latif No. 338 Cost Sharing in the Social Sectors of Sub-Saharan Africa: Impact on the Poor. Arvil Van Adams and Teresa Hartnett No. 339 Public and Private Roles in Health: Theory and Financing Patterns. Philip Musgrove No. 340 Developing the Nonfarm Sector in Bangladesh: Lessons from Other Asian Countries. Shahid Yusuf and Praveen Kumar No. 341 Beyond Privatization: The Second Wave of Telecommunications Reforms in Mexico. Bj6rn Wellenius and Gregory (Continued on the inside back cover) WORLD BANK DISCUSSION PAPER NO. 380 India's Public Distribution System A National and International Perspective R. Radhakrishna K Subbarao with S. Indrakant and C. Ravi The World Bank Washington, D.C. Copyright © 1997 The International Bank for Reconstruction and Development/THE WORLD BANK 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. All rights reserved Manufactured in the United States of America First printing November 1997 Discussion Papers present results of country analysis or research that are circulated to encourage discussion and comment within the development community. To present these results with the least possible delay, the typescript of this paper has not been prepared in accordance with the procedures appropriate to formal printed texts, and the World Bank accepts no responsibility for errors. Some sources cited in this paper may be informal documents that are not readily available. The findings, interpretations, and conclusions expressed in this paper are entirely those of the author(s) and should not be attributed in any manner to the World Bank, to its affiliated organizations, or to members of its Board of Executive Directors or the countries they represent. The World Bank does not guarantee the accuracy of the data included in this publication and accepts no responsibility whatsoever for any consequence of their use. The boundaries, colors, denominations, and other information shown on any map in this volume do not imply on the part of the World Bank Group any judgment on the legal status of any territory or the endorsement or acceptance of such boundaries. The material in this publication is copyrighted. Requests for permission to reproduce portions of it should be sent to the Office of the Publisher at the address shown in the copyright notice above. The World Bank encourages-dissemination of its work and will normally give permission promptly and, when the reproduction is for noncommercial purposes, without asking a fee. Permission to copy portions for classroom use is granted through the Copyright Clearance Center, Inc., Suite 910,222 Rosewood Drive, Danvers, Massachusetts 01923, U.S.A. The complete backlist of publications from the World Bank is shown in the annual Index of Publications, which contains an alphabetical title list with full ordering information. The latest edition is available free of charge from the Distribution Unit, Office of the Publisher, The World Bank, 1818 H Street, N.W., Wash- ington, D.C. 20433, U.S.A., or from Publications, The World Bank, 66, avenue d'Iena, 75116 Paris, France. ISSN: 0259-210X R. Radhakrishna is a professor of economics/senior fellow at the Centre for Economic and Socail Studies, Hyderabad, India. K. Subbarao is a principal economist in the poverty group in the World Bank's Poverty Reduction and Economic Management Network. S. Indrakant is a reader at the Department of Economics at Osmania University. C. Ravi is a visiting fellow at the Centre for Economic and Social Studies in Hyderabad, India. Library of Congress Cataloging-in-Publication Data India's public distribution system: a national and international perspective / R. Radhakrishna K. Subbarao with S. Indrakant and C. Ravi. p. cm. - (World Bank discussion papers ; no. 380) Includes bibliographical references (p.). ISBN 0-82134090-5 1. Food supply-India. 2. Food relief-India. 3. Income distribution-India. 4. Poor-India. I. Radhakrishna, R. (Rokkam), [date]. II. Series: World Bank discussion papers; 380. HD9016.1421554 1997 363.8'56'0954-dc2l 97-35142 CIP Contents FOREWORD vii ABSTRACT viii ACKNOWLEDGMENTS ix EXECUTIVE SUMMARY x CHAPTER 1: INTRODUCTION The Background 1 Objectives and Outline of the Study 2 CHAPTER 2: POVERTY ALLEVIATION PROGRAMS: AN OVERVIEW Public Spending on PAPs 4 Self-Employment Programs 7 Wage Employment Programs 8 In-Kind Transfers 10 Summary 14 CHAPTER 3: PUBLIC DISTRIBUTION SYSTEM, FOOD SUBSIDY AND TARGETING Food Subsidy 16 FCI Costs and Subsidies 20 PDS Supplies 23 Regional Patterns in Allocations and Off-Take of Foodgrains 24 PDS Food Access to the Poor 27 CHAPTER 4: EFFICACY OF PDS AS AN ANTI-POVERTY PROGRAM Introduction 32 Data 32 Methodology 32 Income Transfers to the Poor: 1986-87 38 PDS Impact on Poverty 43 PDS Impact on Nutrition 46 Cost of Income Transfer-All India 47 Cost of Income and Calorie Transfers in Andhra Pradesh 49 iii CHAPTER 5: FOOD TRANSFERS: AN OVERVIEW OF INTERNATIONAL EXPERIENCE Regional Patterns in Program Choice 57 Reforms: Food Stamps, Self-Targeted Rations, and Other Approaches 58 Lessons from International Experience 65 Conclusions 69 CHAPTER 6: PDS: SUMMARY OF FINDINGS AND OPTIONS FOR REFORM Main Findings 71 Recent Trends in Administered Prices 72 Lessons from International Experience 73 PDS in the Context of a Changed Agricultural Scenario 74 Recent GOI Proposal for a Targeted PDS 75 Elements of a Reformed PDS 76 Reform Options to Improve Targeting 78 Conclusion 80 APPENDIX NOTE ON DATA ADJUSTMENTS 82 APPENDIX 1 84 REFERENCES 95 iv List of Tables, Boxes and Figures Tables: Public Spending on Poverty Alleviation Program in 1993-94 (Center and States) 5 Public Spending on Major Poverty Alleviation Programs in Selected States in 1993-94 6 Approximate Population Coverage in an ICDS Project 12 Estimated Cost of an ICDS Project 13 Cost of Transfer of Food in ICDS Project 14 Central Government Expenditure on Food Subsidy 17 Central Government Expenditure on Cereal Subsidy and Cost of Carrying Cereals 18 Details of FCI Cereal Sales and Cereal Subsidies During 1990-96 19 State-Wise FCI Supplies of Cereals and Central Subsidies in 1993-94 21 Details of Economic Cost and Subsidy for Distribution Operations (Rs per Quintal) 22 Allocation for PDS and Off-Take of Rice and Wheat in States and Union Territories of India 25 Per Capita Off-Take Levels of Cereals and its Correlates 27 Monthly per Capita Purchase of Cereals (Kgs) from PDS and their Percentage to Total Cereal Consumption 1986-87 28 Per Capita Monthly Consumption (Kg) of Rice from PDS in Andhra Pradesh and Kerala and Cereals from PDS in Bihar and Uttar Pradesh 1995 29 Population, Percentage of Poor and Very Poor and Poverty Gap Ratio by States in 1986-87 34 Number of Poor and Very Poor and Aggregate Monetary Shortfall by States in 1986-87 (Million) 35 Percentage Distribution of Very Poor, Poor and Aggregate Monetary Shortfall 36 Per Capita Purchases of Rice, Wheat and Sugar from PDS in 1986-87 (Kg/Month) 39 PDS and Open Market Prices of Rice, Wheat, Sugar and Edible Oils in 1986-87 (Rs/Kg) 40 Income Transfers to Poor, Non-Poor and All Income Classes PDS in 1986-87: Rural 41 Income Transfers to Poor, Non-Poor and All Income Classes PDS in 1986-87: Urban 42 Decline in Poverty Due to Subsidies: 1986-87 44 Effects of Subsidies on Gini Co-Efficient in 1986-87 45 Number of Persons Moved Out of Poverty Due to Subsidies in 1986-87 (Millions) 45 Percentage Distribution of Persons Moved Out of Poverty Due to Subsidies in 1986-87 46 Calorie Gains Due to PDS (All Commodities) in 1986-87 (K.Cal/Day/Person) 48 Cost to Central Government to Transfer One Rupee to Poor Through PDS Cereal Supply in India in 1986-87 49 Cost to Central Government to Increase Calorie Intake Through PDS Cereal Supply in 1986-87-All India 53 Cost of One Rupee Transfer to Poor Through Rice Subsidy Scheme in Andhra Pradesh 1986-87 54 Cost of Calorie Increase Through Rice Scheme in Andhra Pradesh in 1986-87 55 Cost Per Re 1 of Income Transferred by Various Programs (1988-90) 55 Regional Patterns in Food Transfer Programs 58 Incentive Costs - Rice Subsidy, Sri Lanka, 1980 59 Targeting and Transfer Efficiency for Selected Food Programs 66 Bangladesh: Cost-Effectiveness of Targeted Income Transfer Programs, 1993-94 67 v Functional Framework Under Existing and Modified PDStFood Stamps 81 Boxes Food Subsidies in Sri Lanka: Reforms And Impacts 60 Food Stamp Programs in Jamaica and Honduras: An Assessment 61 Tunisia: Reducing Food Subsidies Without Hurting the Poor 62 Food Rations And Food Stamps: Pros and Cons and Lessons Learned From International Experience 63 Figures Administrative Cost and Cost of Leakage of Food Subsidies, India 52 vi FOREWORD Reducing poverty in India is one of the major development challenges faced in the world. The Government of India has experimented with a number of direct anti-poverty programs. Of these, the Public Distribution System (PDS)-a quantity rationing-cum-food subsidy progran-is perhaps the oldest. It has evolved in its present form in the wake of national-level food shortages of the 1 960s. Since then, the national agricultural scene has vastly changed for the better, contributing to a substantial fall in the incidence of poverty. However, the PDS remained fundamentally unaltered, and the cost of food subsidy has grown enormously since the mid-1980s. Moreover, some stake governments have introduced food subsidy programs with additional resources of their own. Since 1990, India has introduced a wide range of economic reforms, including significant liberalization. Within these reforms, the PDS has been given added emphasis. Some changes towards better regional targeting are being considered. However, few studies have estimated the impact on poor households of PDS, in terms of reductions in the incidence and severity of poverty as well as nutritional improvements. Nor have the costs per unit of welfare gains been estimated. This study-a joint product of the PREM Poverty Anchor Group and the South Asia Department-fills this gap. In view of the important role played by safety net programs, both during normal times and in periods of major economic reform, it is hoped that this case study for India would be of interest to researchers and policy-makers engaged in poverty reduction. Michael Walton Director Poverty Reduction Poverty Reduction and Economic Management Network vii ABSTRACT We estimate the impact on poor households of India's Public Distribution System (PDS) in terms of income gains, reductions in the incidence and severity of poverty as well as nutritional improvements. We assess the cost-effectiveness of PDS, relative to other programs. We then evaluate the rice subsidy program in the state of Andhra Pradesh. Our findings suggest that the welfare gains of PDS in terms of income transfer were very meager and the impact on poverty and nutritional status minimal. Even the meager transfer benefits were realized at an exorbitant cost. In the light of international experience and India's own experience, we review the options for reform and, in this context, assess the recent moves of the Government of India towards a better targeted PDS. viii ACKNOWLEDGMENTS This is a background paper prepared for the India Rural Development Report currently underway in South Asia Rural Development Sector Unit (SASRD) of the World Bank. For comments and suggestions on earlier drafts, authors are indebted to Deepak Ahluwalia, Harold Alderman, Benoit Blarel, Ishrat Husain, Dina Umali-Deininger, John Williamson and Roberto Zagha. Successive drafts of this paper were processed by Precy Lizarondo. The authors alone are responsible for the findings and views expressed in this paper. ix EXECUTIVE SUMMARY Quantity rationing of essential commodities has been in existence in India since the inter- war period. The Public Distribution System (PDS) in its present form-a (producer) price- support-cum-consumer subsidy program-has evolved in the wake of critical, national-level food shortages of the 1960s. The program has probably contributed to a containment of upward pressures on food prices and ensured access of the food to urban consumers-the two principal objectives of the scheme at that point in its evolution. Although the national agricultural scene has vastly changed for the better, contributing to a fall in the incidence of poverty (from about 50 percent in the 1950s to 30 percent in the 1990s), the design of PDS has remained fundamentally unaltered, except that its outreach is now being extended to rural areas and tribal blocks and areas of high incidence of poverty. The food subsidy cost has trebled over the past two decades, and accounted for 0.7 percent of GDP in 1993-94. PDS is now "perceived" to be the main safety net to protect the poor from potential short-run, price-induced adverse effects of the economic reforms currently under way, but the perception appears to be far from reality. Notwithstanding the added emphasis given to PDS in the wake of economic reforms, few studies have estimated the welfare gains-in terms of income as well as nutritional impacts. Nor have the cost per unit of income or nutritional gains been estimated. This paper fills this gap. It estimates the gains and costs of PDS, and compares it with other similar programs. The impact of PDS on the incidence and severity of poverty is also estimated. Based on the country's own experience and the experience of other countries, the paper considers various policy options for reform. Main Findings Though it has been in operation for four decades, the access of the poor to PDS is still very limited. The access is particularly weak-almost nil-in the states with the highest incidence and severity of poverty. The access of the poor is much better in a few states, particularly Andhra Pradesh and Kerala. Yet, even in these states, the monthly cereal purchases tended to be regressive. The per capita income gain to the poor in 1986-87 from all (food and nonfood) consumer subsidies was no more than Rs 2.01 per month, or 2.7 percent of their per capita expenditure, in rural areas; in urban areas it was slightly higher at Rs 3.4 per month (or 3.2 percent of per capita expenditure). There were differences in the magnitude of income transfers between commodities and across the states, but the overall transfer gains were very meager indeed. Only in four states (Andhra Pradesh, Kamataka, Kerala and Gujarat), the income gains to the poor were substantially higher than the national average, but some of these states have expended additional resources of their own on the program. Not surprisingly, the impact of PDS on poverty and nutritional status was minimal. For the country as a whole, without PDS poverty would have increased by 2 percentage points; the x adverse impact would have been extremely small at 0.3 percentage points in rural areas of the states with the highest incidence of poverty. With the exception of Andhra Pradesh and Kerala, the nutritional impact of PDS has been minimal. Even the meager transfer benefits were realized at an exorbitant cost. When only the Central government costs are considered, an amount of Rs 4.27 was incurred to transfer one rupee of income to the poor. Typically, in addition to Central expenditures, state governments also incur costs of their own. When the combined state and Central level costs were estimated, one rupee of income was transferred at a cost of Rs 6.35 in Andhra Pradesh; the costs could be even higher in administratively weaker states which also account for a large share of the poor. Given this high level of costs, the program is unlikely to become cost-effective, unless drastic changes are introduced in its design. In comparison with other anti-poverty programs, PDS turns out to be the costliest. PDS delivered 100 kcal of nutrients at three times the cost incurred under the direct nutrition program of Integrated Child Development Services (ICDS). In terms of cost per rupee of income transferred, both the national PDS and the AP rice subsidy scheme are much less cost efficient than employment and nutrition programs. The main reason: while the two food transfer programs adopt income-based means tests, self selection occurs to a great extent in nutrition programs and to a lesser extent in employment programs. PDS and Changing Agricultural Scenario PDS in its present form had evolved in the 1960s when the agricultural scene in the country was completely different from today. The Indian agricultural situation has vastly changed for the better. Not only has foodgrains growth rate exceeded the effective demand for foodgrains, but more importantly, the regional location of accelerated growth-the eastern states-has been highly conducive for poverty reduction. There has been a substantial reduction in the incidence and severity of poverty. An untargeted, open-ended price-cum-consumer subsidy program hardly fits in with the realities of these recent developments. Relentless pursuit of a program basically at odds with economic realities and the changing profile of poverty has its inevitable costs. In order to pass the effect of the devaluation and economy-wide reforms, procurement prices and issue prices have had to increase. Without accompanying reforms in external trade and better targeting of PDS, the welfare implications of higher procurement and issue prices are likely to be adverse. For example, in response to highly favorable procurement prices (in relation to market prices), farmers unloaded huge stocks on to the Food Corporation of India (FCI). Large stocks, much in excess of those needed for strategic purposes, were accumulated. The carrying cost of buffer stocks now accounts for nearly one- third of the total food subsidy cost. At the same time, the poor did not benefit from the whole operation. PDS allotments have not been lifted in the states with the highest concentration of poverty because the (enhanced) issue price was close to the market price thereby rendering PDS virtually dysfunctional. In other states such as Andhra Pradesh, stocks were lifted, but only when a further subsidy was borne by the state governments, diverting resources from other priority sectors. xi Lessons from International Experience The cross-country experience reviewed in the paper suggests that practically every country that adopted open-ended programs has found the policy not only fiscally unsustainable but also, more importantly, significantly distortionary in its effects on the agricultural sector. Most countries have begun to restructure their systems of food transfer. While no particular targeting method was found to be perfect, even imperfectly targeted programs have proven better in reaching the poor and keeping the costs down than completely open-ended programs. Restricting the subsidy to commodities disproportionately consumed by the poor led to considerable self-targeting of the program in some countries; the costs were contained and the benefits reached poor consumers. In general, income-based means tests have proven difficult to administer. Leakages appear to be lowest in programs that selected beneficiaries contingent upon the beneficiary's (or their children's) participation in another activity, such as a child's attendance in a primary school, or registration at primary health clinics. Tying beneficiary participation (or time contribution) to food transfers lowers not only leakages but may also lower the economic costs associated with work disincentives. However, care needs to be taken that the poor's participation does not lead to excessive transaction costs to them. Elements of a Reformed PDS Any proposals for a revamped and restructured food transfer program for India needs to be made against the backdrop of the changing agricultural and institutional scenario in India. In particular, three aspects of change are stressed in the paper. First, the overall national-level shortages have abated; the country's cereal production is growing faster than the effective demand. The relatively poor eastern regions-West Bengal in particular-have seen a significant improvement in grain production. The incidence of poverty has steadily fallen since the 1950s. Second, except in tribal and a few dry land areas, the agricultural marketing infrastructure is well- developed and integrated. Third, following democratic decentralization, the Panchayat Raj Institutions (PRI) can now be entrusted with the responsibility of implementing all major poverty alleviation programs. The Government of India has recently made drastic changes in the Public Distribution System. The new Targeted Public Distribution System (TPDS) would most definitely correct the regional mnisallocation of FCI grain supplies, to the extent the incidence of poverty will be given a substantial weight in grain allocations between the states. However, TPDS expects the retail dealers at PDS outlets to supply the same variety of grain at half the price to the poor, and at the regular price to the nonpoor. The poor will be identified and given ration cards. The scope for abuse and diversion of supplies from the poor to the nonpoor in this system is likely to be high, given past experience with similar arrangements for kerosene and cooking oil. The GOI wants to enforce strict monitoring, but this is likely to be administratively costly. Except for better regional allocation of supplies, new TPDS virtually retains all other earlier characteristics of the system, albeit with increased scope for abuse at the retail outlet level. It is unlikely that the costs of operation of FCI will fall under the new TPDS. xii Given the changes in India's agricultural economy, and bearing in mind the weaknesses and constraints under which the FCI currently operates, the study recommends a reduced role for FCL by phasing out government controls over grain markets as well as procurement operations. FCI could be allowed to compete in the market, free from controls, but with the added advantage of scale economies. FCI may, over time, become an effective player in the market. Its new role may be to stabilize the price of foodgrains within a band, and to maintain strategic buffer stocks or equivalent foreign exchange reserves; it could intervene in the market by off-loading stocks. The study recommends that the subsidy cost saved (as FCI goes out of procurement operations) could be distributed to PRIs in the form of food stamps or vouchers or as a conditional cash grant. The poor-identified by PRIs-may be given these vouchers, to be exchanged for food either at PDS outlets or at regular private grain stores. To begin with, vouchers could be given to participants of JRY (and other employment programs) since the poor may be expected to self-select themselves into such programs. Careful planning would be needed before extending vouchers to other categories. Where such a voucher program is difficult to administer, PDS retail outlets may still supply grain, but introduce PRI-administered identification of the poor. One sub-option is to link the food voucher program with other programs depending upon availability-such as employment programs (JRY/EAS), or nutrition programs (ICDS). These measures are most likely to ensure targeting without incurring additional costs. At the same time, the poor may be given some choice in the type of grains as well as retail outlets. The existing PDS shops could still continue to function, but only as normal grain retail outlets (like any other private dealer) rather than as outlets supplying "subsidized" commodities as in the past. The above structural changes would ensure that the income transfer would go to the intended beneficiary without incurring a huge cost on administration as is the case at present. It would also minimize incentive costs to the extent it strengthens work-fare or nutrition programs that confer benefits to the poor in high measure such as JRY and ICDS, as well as democratic institutions such as the Panchayat Raj. xiii CRAPTER 1 INTRODUCTION The Background Of all the safety nets that are currently in operation in India, the most far reaching in terms of coverage as well as public expenditure (on subsidy) is the public distribution system (PDS). PDS has been in operation in some form or another since the inter-war period; thus it is probably the earliest publicly-funded safety net in the country. It aims to provide essential commodities such as rice, wheat, sugar, edible oil, soft cake and kerosene oil at subsidized prices. State governments, which manage the public distribution in their respective states, also supply other commodities like pulses, salt, coarse clothing etc. In the early years, PDS was conceived of mainly as an instrument of price stabilization as well as an alternative to private trade. It was widely believed that private traders indulged in speculative activities and a countervailing agency was, therefore, needed to control traders' speculative activities. Its coverage was confined to urban areas and to a few food deficit states. In the 1950s and 1960s, the country was experiencing serious, national-level food shortages; the upward pressure on food prices was particularly acute in urban areas. The PDS as it evolved during this period probably served the dual purpose of containing the upward pressure on prices of staple food, and protecting urban consumers, even if it did not serve the interests of rural poor. Also, during this period, employment programs were limited in scope so that during a drought period, PDS supplies were augmented basically to keep the market price of staple within reasonable bounds. By the 1980s, the whole agricultural scenario has changed (for the better). The country experienced a green revolution, and per capita food availability has significantly increased. Nevertheless, rural poor households (especially the landless poor) benefited little from the program. Not surprisingly, since the mid-1980s, the welfare component of PDS has gained importance and its coverage has been extended to rural areas in some states. Specially subsidized foodgrain distribution was introduced in 1985 in all the tribal blocks covering about 57 million people. PDS was later expanded under the Revamped PDS (RPDS) scheme to 1,752 blocks with a high incidence of poverty, covering 164 million people. Efforts were also made to increase the number of fair price shops (FPS) in rural areas. Subsidized foodgrains are also being distributed as a part of wages under the employment programs. The food subsidy given by the Central Government increased from Rs 6.5 billion (0.5 percent of GDP) in 1980-81 to Rs 55 billion (0.7 percent of GDP) in 1993-94. Some states like Andhra Pradesh, Karnataka and Kerala have their own subsidized food schemes for which they provide additional subsidies. PDS is now perceived to be the main safety net to protect the poor from short-run, price- induced, adverse effects of economic reforms (see Mooij, 1994). Hence there has been an added ernphasis given to PDS as a safety net since the introduction of reforms-the RPDS being one such initiative. The question of topical interest is: to what extent is PDS serving the objective expected of it? And at what cost? Most previous studies on the PDS in India have concentrated on the issues bearing on regional variations in the supply of foodgrains through PDS (Tyagi, 1990), urban bias in PDS (Howes and Jha, 1990), targeting of PDS (Jha, 1991; Dev and Suryanarayana, 1992), the growing cost of food subsidy, etc. Few studies have measured the welfare gains from PDS (Parikh, 1994). The welfare gains include the extent of income transfer to the poor through PDS and the consequential reduction in poverty and the extent of nutritional support to the poor. Virtually no studies exist comparing the welfare gains and costs of PDS with other anti-poverty, nutrition- based programs such as ICDS and income-support programs such as JRY which currently co- exist with PDS in India. This study aims to fill this gap. It is also useful to get a global picture of the poverty-reducing effects of PDS, taking into consideration the initiatives of the Center as well as the States. Given the inefficiencies of the Food Corporation of India (FCI), a number of alternatives including food stamps schemes have been suggested for regulating the entire network of procurement, storage and distribution. A review of international experience would be helpful in examining policy options. Objectives and Outline of the Study This study addresses the following questions. What is the extent of income transfer through PDS? How much reduction in the incidence and severity of poverty has taken place as a result? What has been the nutritional impact of PDS on the poor? What did it cost to realize the above-mentioned gains, and how do PDS costs and benefits compare with other similar anti- poverty, nutrition-based programs? What can be learned from international experience in food transfers? In the light of India's experience and international experience, what options exist for reforming India's PDS? In the second chapter, PDS is placed in the context of other anti-poverty programs (IRDP, JRY, ICDS, etc.). In addition, a brief review of the findings from previous research on PDS is provided. In a federal set up, the cost of PDS is borne by both the Central and State governments. The cost borne by the Central Government for selected years is analyzed in Chapter 3. A number of important developments with regard to PDS have taken place in the recent period. Notably, the gap between the subsidized and open market prices has been closing, while some efforts have been undertaken to improve the targeting of the PDS (e.g., geographical targeting through the Revamped PDS). This chapter reviews these recent developments and their likely impact on the poor. In Chapter 4, the extent of income transfer is measured. The impact of PDS on poverty is measured both in terms of incidence and severity. This was done for both rural and urban areas of India and for all the major states of India utilizing the NSSO data on PDS for the year 1986-87. The chapter also derives indirectly the possible nutritional impact of PDS. The chapter provides 2 estimates of the cost of income transfer and consequential nutritional gains and compares the results with another similar program (ICDS) for Andhra Pradesh. Chapter 5 provides an overview of the international experience with respect to food transfers with particular reference to lessons from various approaches to targeting and design of programs. Finally, in the last chapter (6), the options available for improving the effectiveness of PDS are discussed. 3 CHAPTER 2 POVERTY ALLEVIATION PROGRAMS: AN OVERVIEW The Government of India has launched several poverty alleviation programs (PAPs) during the last two decades. These programs fall into three broad categories: (i) family of livelihood creation (self-employment) programs such as Integrated Rural Development Programme (RDP), Development of Women and Children in Rural Areas (DWCRA), and Training of Rural Youth for Self-Employment (TRYSEM) Programmes; (ii) labor-intensive public works schemes under two national programs, namely, Jawahar Rojgar Yojana (JRY), and Employment Assurance Scheme (EAS) and a program funded and implemented by one state, viz., the Maharashtra Employment Guarantee Scheme (MEGS); and (iii) two programs of income transfers in kind, viz., the Public Distribution System (PDS) and the Integrated Child Development Services (ICDS). This chapter provides an overview of these programs. Public Spending on PAPs The available data on public spending for all major PAPs for the year 1993-94 are shown in Table 2.1. Three findings are worth noting. First, the total spending (Center and States) on all programs was 1.39 percent of GDP-not a high percentage in comparison with countries similarly placed (e.g., Sri Lanka). Second, roughly one half of total spending is on the Public Distribution System-the subject of this study. The second in importance are the wage employment programs (37 percent), and the much-researched self-employment program of IRDP ranks third in spending (about 9 percent). These three programs account for 95 percent of total spending. Although more recent data are unavailable, it is unlikely that the relative spending priorities would have changed much.' Detailed information on Central and state-level spending on major PAPs is hard to obtain. Available data for three states are shown in Table 2.2. Of these, Andhra Pradesh is a state with a significant program of its own-the two-rupee a kilo rice program. Karnataka has no food transfer program of its own but delivers the centrally-administered PDS involving the local authorities (Panchayats). Uttar Pradesh is a large state with a high incidence of poverty, but with no state-sponsored program of food transfer. For PDS, while the Center and Andhra Pradesh Government are spending roughly similar amounts on PDS, State's contribution is less than one fourth in Karnataka and Uttar Pradesh. Of all the PAPs, the contribution of the states and the center are roughly equal only for IRDP. XThe 1996 Budget presented in the Parliament by the new Governnent has proposed a further increase of 40 percent in total spending on PAPs. Considering that the price index would have risen by at least 20 percentage points since 1993-94, the real spending on PAPs, as a percent of GDP, would not have risen much. 4 Table 2.1 Public Spending on Poverty Alleviation Program in 1993-94 (Center and States) Program Expenditure % of Total % of Total % of Share of each (Rs million) govt. govt revenue GDP program in expenditure expenditure total spending on antipoverty programs A. In-Kind Transfers 1) Food Subsidy 55370 2.44 3.02 0.69 49.7 2) Integrated Child Development Scheme 5645 0.25 0.31 0.07 5.1 3) Subsidy on Controlled Clothing 160 0.01 0.01 0.00 0.1 B. Income-Generation Programs 4) Integrated Rural Development Program 9567 0.42 0.52 0.12 8.6 b. Wage Employment Programs 5) Jahawar Rojgar Yojana (I) 35902 1.58 1.96 0.45 32.2 6) Jahawar Rojgar Yojana (II) 2885 0.13 0.16 0.04 2.6 7) Jahawar Rojgar Yojana (1+11) 38787 1.71 2.11 0.49 34.8 8) Employment Assurance Scheme 1838 0.08 0.10 0.02 1.7 9) All Wage Employment Program 40625 1.79 2.22 0.51 36.5 (7+8) 10) All Income Generation Programs 50191 2.21 2.74 0.63 45.1 Including Self-Employment (4+9) C. All Major Poverty Alleviation 109300 6.51 8.07 1.36 100.0 Programs (A+B) D. All Poverty Alleviation Programs, 123.5 - - - per Capita Notes: Total Expenditure (Center + States) - 2273 billion. Total Revenue Expenditure (Center + States) - Rs 1834 billion. GDP at market prices - Rs 8010 billion. The central spending on ICDS includes only plan expenditure. The expenditure on food subsidy (Rs 55370 million) does not include States' spending on food subsidy. Sources: Ministry of Rural Development for IRDP, JRY (I & II) and EAS, Indian Public Finance Statistics, 1994, Ministry of Finance, Department of Economic Affairs. For Food Subsidy, Total Expenditures. 5 Table 2.2 Public Spending on Major Poverty Alleviation Programs in Selected States in 1993-94 (Rs million) Programs Andhra Pradesh Karnataka Uttar Pradesh 1. Food Subsidy 8273 2445 2153 a) Center 4193 1995 1653 b) State 4080 450 500 2. Integrated.Child Development Scheme 625 393 336 a) Center 464 352 000 b) State 161 41 336 3. Integrated Rural Development Program 881 402 2019 a) Centee 442 201 1009 b) State8 439 201 1010 4. JawaharRojgarYojana (I) 2857 1756 6953 a) Center 2286 1405 5562 b) State 571 351 1391 5. Jahawar Rojgar Yojana (II) 425 16 198 a) Centere 340 135 158 b) State8 85 34 40 6. Employment Assurance Scheme 257 68 891 a) Center 206 54 713 b) State 51 14 178 7. Major Poverty Alleviation Programs Total spending 13317 5235 12549 a) Center 7930 (59.5%/6) 4144 (79.2%) 9095 (72.5%) b) State 5387 (40.5%) 1091 (20.8%) 3454 (27.5%) 8. Major Poverty Alleviation Programs Per Capita Spending 200.2 116.4 90.2 a) Center 119.2 92.1 65.4 b) State 81.0 24.3 24.8 a Break-up of expenditure by Center and State is not available. The total expenditure has been split into Central and State expenditure using shares of Center and State in the total funds released for the scheme. Sources: State Budgets. Ministry of Rural Development. 6 Self-Employment Programs Integrated Rural Development Program (IRDP) The Integrated Rural Development Program was introduced in 1978-79. The program, currently involves a public spending of about 0.12 percent of GDP and 0.42 percent of public expenditure. It provides soft credit to the households below the poverty line (an annual family income of Rs 11010 at 1991-92 prices) for the purchase of income-earning assets in agriculture, animal husbandry, village industries, or services. The expectation is that the enhanced income levels of the beneficiaries would move them out of poverty. The assistance takes the form of capital subsidy of a share of the investment for income-generating assets by the Government, and concessional credit by the public sector banking system and allied institutions like the Regional Rural Banks. The extent of capital subsidy varies according to the socio-economic position of the beneficiary; it ranges between 25-50 percent of the investment subject to a ceiling ranging between Rs 4000-6000.2 The program is being implemented by the District Rural Development Agencies (DRDA)3 in all blocks of the country as a centrally-sponsored scheme funded on 50:50 basis by the Center and States.4 In 1993-94, the public expenditure on IRDP (Center and States) was Rs 9570 million which constituted 0.12 percent of GDP (Table 2.1) and the Central share accounted for about two percent of the Central expenditure. Over time IRDP has acquired a significant presence in rural areas. The number of beneficiaries increased from 2.73 million in 1980-81 to 4.25 million in 1987-88.5 The number subsequently declined to 2.57 million in 1993-94 mainly due to the increase in per capita subsidy. IRDP has been extensively researched. Studies suggest that the beneficiaries of IRDP are mostly near-poor or non-poor (NSS 42nd Round);6 its effectiveness in lifting the poor above the poverty line is dubious;7 it is biased in favor of land-owning households; and the program is least helpful to the poorest among the poor. Moreover, the long-term sustainability of benefits is in doubts (Kurien and Rajeev, 1995). The studies also point out the reasons for the generally poor performance of IRDP. For example, -in the choice of schemes, no attention has been paid to the local constraints in 2 In 1992-93, the per family investment was in the range Rs.6307-7613 and subsidy was in the range Rs 1989-2317 (Concurrent Evaluation Survey, 1992-93). 3 For drawing up poverty alleviation programs and execution of works, DRDA, with the Collector as its Chainnan, was created with relatively more flexibility than the Government agency at the district level. All officials of Line Departments assist the collector. 4 The Constitution of India lays down the division of work between Central and State Goverinments in three lists-Central, State and Concurrent. The Center sponsors development which falls in the State list and such schemes are called centrally sponsored schemes. The expenditure is shared by the Center and States. 5 Durinlg 1983-88, according to the NSS data, 6.8 percent of households received ERDP assistance. 6 A study on IRDP schemes in Anantapur district of Andlra Pradesh showed that 29 percent of the beneficiaries of IRDP were non-poor, and they comered one-third of IRDP investment (Galab, 1987). 7 The Concurrent Evaluation Survey of 1992-93 has shown that about 41 percent of the eligible families could move out of poverty (see Kurien and Rajeev, 1995). ' The Concurrent Evaluation Survey of 1992-93 has shown that only 15 percent of families assisted in 1989 under IRDP were not poor in 1992-93. 7 availability of markets, support services and skills as well as uncertainties of prices and input supply.9 Evidence further suggests that employment and income effects were better in developed regions than in backward regions;10 in the latter the scheme provided financial relief rather than long-term livelihoods. The program is subsidy-driven; a part of the capital subsidy accrued to rent-seekers. Training of Rural Youth for Self-Employment (7RYSEM) TRYSEM, introduced in 1979, trains rural youth (in the age group of 18-35 years and from families living below the poverty line) in necessary skills and technology to take up livelihood programs in rural areas. Since the Seventh Plan it has been integrated with IRDP. In 1987, the program's scope was enlarged to include wage employment for the trained beneficiaries. The trainees are paid a stipend, and the trainers an honorarium. A new self- employment scheme for the rural educated youth known as the Prime Minister's Roigar Yojana (PMRY) has recently been introduced. All these programs are very modest in terms of their coverage and public spending. Development of Women and Children in Rural Areas (DWCRA) DWCRA was launched in 1982-83 as a sub-program of IRDP with the specific objective of improving the status of poor women in rural areas."1 Women are organized into homogeneous groups of 15-20, and each such group is provided training in a chosen economic activity along with the necessary infrastructure. The activities include weaving, fish vending, broom and rope making, brick making, and pickle making. DWCRA and IRDP are not mutually exclusive. Women belonging to poor households can become members of DWCRA and also avail of subsidy and credit under IRDP subject to the overall subsidy ceilings. In 1992-93, 9,029 groups were assisted with a public spending (Center and States) of Rs 117 million. Wage Employment Programs Jawahar Rojgar Yojana (JRY) JRY seeks to attain the twin objectives of generation of gainful employment for poor households and asset creation in rural areas, though the emphasis is on the former. Preference is given to under-privileged groups ( such as scheduled castes and tribes); in addition, 30 percent of the employment opportunities are earmarked for women. Wages under JRY are paid at the rate notified in the prescribed schedule of employment under the Minimum Wage Act for the relevant work, partly in foodgrains and partly in cash. The Central Government bears 80 percent of the cost while the States contribute 20 percent. JRY differs from other programs in that it incorporates decentralized decision-making by the village panchayat. 9 The Concurrent Evaluation Survey of 1992-93 observed that 30 percent of existing assets generated no income. ° The performance of the scheme in one agriculturally developed district of Andlura Pradesh (East Godavari) is found to be very impressive because of its bias towards non-farm activities having linkages with the ongoing process of conmnercialization. (Omkamath, 1993) See also Subbarao (1985). DWCRA differs from IRDP in its strategy, while the family is the focus of IRDP, the group is the focus of DWCRA. 8 There are now three streams under JRY. The first stream is the continuation of the earlier JRY which has been in operation since its inception in 1989; the second and third streams were introduced from 1993-94. Since JRY created in a year only about 45 days of employment per worker-clearly inadequate in regions of high unemployment'2-4ntensified JRY (second stream under JRY) was launched in 120 identified chronically backward districts with high levels of unemployment and under-employment, labor out-migration and depletion of natural resources. Programs under the third stream of JRY consist of innovative schemes aimed at prevention of distress migration via special programs for drought proofing, watershed development, wasteland development, etc. Seventy five percent of the funds allocated for JRY is to be utilized for the implementation of the first stream throughout the country, 20 percent for the second stream and 5 percent for the third stream. In 1993-94, the first stream of JRY created 952 million person days of employment, and the second stream. 7.35 million personal days. Since underemployment is estimated to be of the order of 3000 million person days, about one-third of the total underemployment may have been reduced by JRY. Notwithstanding the substantive impact in the aggregate, the program's impact appears to be modest at the household level. The concurrent evaluation of JRY, conducted by the Ministry of Rural Development during January-December 1992, has brought out that a JRY worker got, on an average, 3.8 days and a family 5.15 days of employment from panchayat JRY works during the reference period of 30 days preceding the date of survey. In some backward states the impact was higher. A JRY worker in Orissa, for example, got, on an average, 12.32 days and a family 17.55 days of employment during the reference period; and in Assam 10.72 and 13.80, respectively. According to the concurrent evaluation, the average daily wage rate was Rs 22.80 for the male unskilled worker, and Rs 21.21 for the female unskilled worker, which were in general higher than the prevailing market wage rates. It works out that a JRY worker earned Rs 96.70 and a IRY family Rs 130.71 during the 30 days preceding the date of survey (or 15 percent of the poverty line threshold). In Orissa, the earnings were Rs 282.13 and Rs 401.90, respectively. Foodgrains distributed (as a part of wages) was negligible (Rs 0.21 worth of foodgrains per day per JRY worker)-there was no proper linkage between JRY and PDS. Using the results of the concurrent evaluation study, the public expenditure required to transfer one rupee to a household under the JRY was derived. In 1991, it required a public expenditure of Rs 1.87 to transfer one rupee of income to a rural household."3 The evaluation study has also revealed that 57 percent of households that benefited from JRY works were non- poor. Under JRY, the program wage almost always remained higher than the market wage. This factor may have been one of the reasons for attracting many of the non-poor to the program. 12 A problem with the first stream of JRY is that as each and every village is supposed to be covered (even those in prosperous Punjab) the average amount at the disposal of a village panchayat is liunited and in most cases too meager to reduce un(der)employment within the village substantially. 13 It can be inferred from the Concurrent Evaluation Report that in 1991 Rs 20152 million was spent on JRY works by the Panchayats, of which the wage component was Rs 10773 million. 9 Taking into account the extent of mistargeting, it can be estimated that in 1991 Rs 4.35 of public spending was required to transfer one rupee to a poor rural household. Employment Assurance Scheme (EAS) EAS was launched on October 2, 1993 in 2446 identified backward blocks in drought prone, desert, tribal, and hill areas.14 Under EAS, a maximum of two adults per household will be provided assured employment for 100 days during the lean season in a year. The wages to be paid will be the minimum wages for unskilled labor prescribed by the concerned State Governments. Persons in the age group 18-60 who need and seek employment under EAS have to register themselves in the village panchayat. Those seeking work are to be provided work within two weeks within the area of the block where they reside. However, unlike in the Employment Guarantee Program of Maharashtra, there is no compensation for failure to provide employment. The scheme is operated at the block level by the DRDA. Unlike in JRY, all works are to be executed departmentally. In 1993-94, the Government spent Rs 1838 million (Table 2.1) and created 49.47 mnillion personal days employment. In-Kind Transfers Public Distribution System (PDS) Of all the safety net operations that exist in India, the most far reaching, in terms of coverage as well as public expenditure on subsidy, is the public distribution system. PDS provides rationed amounts of basic food items (rice, wheat, sugar , edible oils) and other non-food products (kerosene, coal, standard cloth) to the rural and urban population at below market prices to consumers through a network of 3,24,000 fair price shops, the access to the system being almost universal. PDS has changed both qualitatively and quantitatively since the late 1970s. Till the late 1970's, PDS was mainly confined to urban areas and food deficit regions. The main emphasis was on price stabilization. Private trade was considered "exploitative," and PDS was considered the countervailing power to private trade. Since the early 1980s, the welfare dimension of the PDS has gained importance. Rural areas were covered in many States in the 1980s; more recently PDS is being linked to employment programs. Both the Central and State Governments participate in the procurement and distribution of foodgrains. A specific portion of grain output from producers or traders is sold to the Government at the minimum support price fixed by the Government on the basis of yearly recommendations of the Commission for Agricultural Costs and Prices (CACP). Much of the procured output is supplied to the States at a uniform issue price for supporting their public distribution systems. 14 These are also the districts in which the Revamped Puiblic Distribution System (RPDS) is also in operation. RPDS is discussed in the following sections. 10 The purchase, storage, movement and distribution of food are being carried out by the Food Corporation of India, a parastatal established in 1965. In addition, some States have food and civil supplies corporations or cooperative marketing agencies which make purchases and sales on behalf of FCI. FCI issues foodgrains to the public distribution system based on allocations made by the Central Government. It also supplies foodgrains to the State Governments for some of the anti-poverty programs such as JRY. The state governments lift their quota from FCI godowns and distribute them through a network of fair price shops (FPS). In most states, practically everyone is entitled to draw supplies from FPS. However, the access is restricted (targeted) in a few states. For instance, in Andhra Pradesh and Karnataka, cards of different colors are issued based on means tests. 5 The issue price of FCI is lower than the cost incurred by it. The difference is met through a Central Government subsidy. The retail price of foodgrains is fixed by each state government after adding its distribution costs and taxes to the FCI issue price. Some state governments may provide additional subsidy and fix the retail price at an even lower price than the price fixed by the Central Government. 16 In addition, foodgrains are allocated from the Central pool for employment programs, undertaken particularly during the off-season, and relief works undertaken during drought, floods, etc. In these programs a part of the wage is paid in the form of (subsidized) foodgrains. However, except in drought years, the distribution of cereals for employment programs is negligible. The food subsidy cost of the Central Government has increased manifold from Rs 6.5 billion (0.53 percent of GDP) in 1980-81 to 55 billion (0.7 percent of GDP) in 1993-94. In addition a state like Andhra Pradesh has spent an additional Rs 4 billion on its subsidized rice scheme. 17 The weakness, and the costs and benefits of PDS are analyzed in Chapters 3 and 4. Integrated Child Development Services (ICDS) India's most important nation-wide Centrally sponsored nutrition scheme is the ICDS which integrates supplementary nutrition with primary health care and informal education. It seeks to ensure for pre-school children (0-6), pregnant and nursing mothers, a package of health, nutrition and education (Table 2.3). By September 1992, about 15 million pre-school children and 2.87 million pregnant and nursing mothers were getting supplementary nutrition under ICDS, and nearly 8.87 million children aged 3-6 years were getting pre-school education services (Radhakrishna and Narayana, 1993). The total expenditure on ICDS was Rs 4424 million in 15 The Rice Scheme in Andhra Pradesh covered all households with an annual income below Rs 6000 p.a., those having less land than 1.5 acres irrigated under assured sources, 2.5 acres irrigated under other sources, 3 acres irrigated raising commercial crops or 5 acres unirrigated raising other dry crops. Such households were eligible for supply of 5 kgs. of Rice per person per month at Rs 2 per kg. subject to a maximum of 25 kgs. per month. In rural areas of the State, 80 percent of households were covered. The scheme covers poor as well as about 70 percent of the non-poor. It is clear that the concept of targeting is loosely applied. 16 For instance, in 1994-95 under the Rice Scheme, Andhra Pradesli Government distributed 2.25 million tonnes of subsidized rice purchased from FCI at a price of Rs 5.37 to the target group at a retail price of Rs 2. 17 For Andhra Pradesh, the Center's own subsidy cost accounted to Rs 4.2 billion in 1993-94 (see Table 2.3). 11 1993-94 or 0.1 percent of GDP (Table 2.1). The program is now specially focused on areas with high incidence of poverty. Table 2.3 Approximate Population Coverage in an ICDS Project Rural/Urban Project Tribal Project (Population 100,000; Villages 100) (Population 35,000; Villages 50) Target Group Services Total Target Percent Total Target Percent Population Population Population Population Population Population Children 0-6 years below 6 Supplementary Years Nutrition 6,800 40 4,462 75 hnmunization 17,000 17,000 100 5,950 5,950 100 Health Check- 17,000 100 5,950 100 Up 3-6 Years Non-Formal Pre-School Education 8,000 4,000 50 2,800 2,100 75 Expectant Supplementary Nursing Nutrition Mothers 4,000 1,600 40 1,400 1,050 75 Health Check- 4,000 4,000 100 1,400 1,400 100 up Inmuiuzation against Tetanus (Expectanit 2,400 2,400 100 910 910 100 Mothers) 3. Women Health and (15-45 Years) Nutrition 20,000 20,000 100 7,000 5,250 75 Source: National Evaluation of ICDS, NIPCCD, 1992. ICDS is organized through a chain of projects each of which is located at a community development block covering 1,00,000 population in rural and urban areas and 35,000 population in tribal areas. The coverage of a project is shown in Table 2.3. The Anganwadi is the focal point in the delivery of ICDS services at the village level. The project team of ICDS consists of one Child Development Officer (CDPO), 3-5 Supervisors with supporting staff for the project head office and one Anganwadi Worker (AWW) and one Helper for each Anganwadi. All of them are women except the CDPO who may be a male in some instances. The AWW is usually recruited from the same village, and is responsible for delivery and coordination of services as shown in Table 2.3. 12 Based on a case study of the projects in one backward district in Andhra Pradesh the annual cost of an ICDS project is estimated to be Rs 4.5 million (Table 2.4). Nine thousand two hundred twenty-one beneficiaries-7730 are children and 1491 are expectant and nursing mothers are covered under this project (Table 2.5). Each child is given 65 gms. of supplementary food per day for 25 days in a month while pregnant mothers are given 135 gms. per day (for 25 days in a month). Each ration (food supplement) contains 200 k cal. The cost of providing 100 k cal worked out to Rs 0.72 at 1995-96 prices or two cents per day per child. Table 2.4 Estimated Cost of an ICDS Project Items Cost per Project (Rs 000) Percentage L. Health 142.18 3.19 1) Salaries 102.78 2.31 2) Traveling Allowance & Petrol 9.40 0.21 3) Medical Kit 30.00 0.67 11. Welfare 1215.20 27.30 1) Salaries 990.60 22.26 2) Wages 3.60 0.08 3) Rents 34.86 0.78 4) Traveling Allowance 141.68 3.18 III. Nutrition 1) Ready to Eat Food 3093.60a 69.50 IV. Total Cost (I+II+III) 4450.98 100.00 a Includes transport cost. Source: Field Study in Andhra Pradesh, 1996. Recent evaluations of ICDS in four States by the National Institute of Nutrition has shown that (a) most of the ICDS beneficiaries come from very deprived socio-economic groups such as Scheduled Castes, Scheduled Tribes and lower rungs of the backward classes who are vulnerable to nutritional disorders; and (b) the coverage of National Health Program like immunization has been better in ICDS areas. Regarding the shortcomings, it has been found that (a) there has been irregular food supply; (b) the coverage of children below three years under SNP has been relatively low; (c) there is little community participation in running the ICDS; (d) the AWs do not have adequate building and function in an unhygienic physical environment; and (e) the interdepartmental coordination is poor (Radhakrishna and Narayana, 1993). 13 Table 2.5 Cost of Transfer of Food in ICDS Project I. Total CostRs 000 4450.98 II. Number of Anganwadis 127 Number of Children 7330 Number of Mothers 1491 Number of Rations (per day) 10312 III Cost Per Ration (Rs) 1.44 IV. Cost per 100 Kcal(Rs) 0.72 Source: Field Study in Andhra Pradesh, 1996. Summary India has long experimented with both transfer programs and income-generation programs. PAPs currently account for eight to ten percent of budgetary expenditure (both Center and States) of which over 60 percent goes to PDS and JRY. One weakness common to all PAPs pertains to the administrative structure. The delivery system is highly centralized and over-bureaucratized and, as such, is not flexible enough to adjust to the needs of the poor. The Center has a large direct role in financing and designing of PAPs, while the States and Panchayat Raj institutions play greater role in the implementation. There is little coordination between agencies involved in financing, designing, and actually delivering the programs. Better outcomes were noticeable wherever NGOs and the community were involved as had happened in some villages.8 The eleventh schedule of the 73rd Amendment Act lists twenty nine items including poverty alleviation programs under the purview of the panchayats. Once the elected bodies are in place, and appropriate steps are taken to devolve powers to them, the Panchayat Raj institutions can be expected to coordinate the various sectoral and area development programs with poverty alleviation programs. These changes at the grass-roots level may provide opportunities to render PAPs cost-effective, in addition to ensuring greater participation of the village community in their design and implementation. 18 Some of the successful experiments are presented in Pande (1991) and Deshpande (1992). They suggest that community involvemnent and effective participation of PRIs/NGOs institutions will enhance the potential of the poverty alleviation programs through a careful design and planning of projects towards long term benefits. Anna Hazare's experiment in Ralegan Shinde in Maharashtra shows how through conmnunity participation in the design and implementation of PAPs, the depth of poverty can be substantially reduced. Some of the other successful experiments are SEWA and Sadguru Sewa Sangh in Gujarat, Pani Panchayat and Manavlok in Maharashtra, and Bhagvatula Charitable Trust and Young India Project in Andhra Pradesh. 14 With the exception of ICDS, EAS, and RPDS which account for a small share of the total public spending on PAPs, all the other major safety net programs suffer from loose targeting and leakages. Our subsequent analysis would consider the cost-effectiveness and other aspects of the two fold transfer programs, viz., PDS and the AP rice subsidy scheme, and compare the findings with other transfer programs. 15 CHAPTER 3 PUBLIC DISTRIBUTION SYSTEM, FOOD SUBSIDY AND TARGETING FCI undertakes the functions of procurement including purchases under price support operations, as well as, storage, transportation, movement, distribution and sale of foodgrains.19 The Central Government fixes the minimum support/procurement and issue prices of foodgrains. FCI supplies foodgrains to state governments against the allocations of the Central Government to various schemes such as Public Distribution System (PDS), Revamped Public Distribution System (RPDS), Integrated Tribal Development Programme (ITDP), Jawahar Rojgar Yojana (JRY) and Nutrition Schemes and also sells in the open market. The quantity'of open market cereal sales and open market price are also fixed by the Central Government. FCI operates within the above parameters. This chapter provides an overview of food subsidy cost and its composition, changes in off-take and regional distribution of PDS supplies, and the access of the poor to the PDS. Food Subsidy The Central government expenditure on food subsidy increased eight fold in current prices (about two times in real terms) between 1980-81 and 1994-95, and in fact, in one single year, 1993-94, it nearly doubled (Table 3.1). The rate of growth of Central Government expenditure on food subsidy has slowed down after 1993-94. However, the share of food subsidy in GNP as well as in Central Government expenditure, has fluctuated, showing no strong secular trend one way or another. Recent changes in the composition of food subsidy are shown in Table 3.2. In 1994-95, out of the food subsidy amount of Rs 51000 million, subsidized cereals and buffer stocking accounted for Rs 42041 million (about 82 per cent) (see Table 3.2). A notable development is the sharp increase in the annual carrying cost of cereal buffer stocks-both absolutely, and relative to total food subsidy. It increased nearly three fold over the period 1990-91 to 1994-95. On the other hand, the expenditure on cereal subsidy has declined by 27 per cent between 1992-93 and 1994-95. The distribution of cereal subsidy between wheat and rice was more or less equal except in the year 1994-95 when the share of wheat has increased to 65 per cent. The subsidy on coarse cereals has been negligible. 19 FCI also handles the distribution of levy and imported sugar on behalf of the Central Government. FCI is maintaining Non-Statutory Sugar Price Equalization Fund. It draws from the fund whenever the cost of purchase and distribution of levy/imported sugar exceeds the receipts from sales and makes payments into the fund otherwise. The Central government periodically contributes money to the fund to meet the accumulated losses. For, instance, in 1994-95, Central government contributed Rs 5099 million to this fund. 16 Table 3.1 Central Government Expenditure on Food Subsidy Expenditure (Rs million) Expenditures as Percentage to Year At current At 1980-81 Total Government prices Prices GNP Expenditure 1974-75 2950 4238 0.44 3.01 1975-76 2500 3688 0.35 2.08 1976-77 5060 7029 0.66 3.85 1977-78 4800 6276 0.55 3.20 1978-79 5700 7316 0.61 3.22 1979-80 6000 6689 0.58 3.24 1980-81 6500 6500 0.53 2.89 1981-82 7000 6349 0.49 2.76 1982-83 7100 5961 0.45 2.33 1983-84 8350 6466 0.45 2.32 1984-85 11000 7927 0.53 2.51 1985-86 16500 11032 0.71 3.11 1986-87 20000 12511 0.77 3.12 1987-88 20000 11506 0.68 2.84 1988-89 22000 11703 0.63 2.70 1989-90 24760 12172 0.61 2.60 1990-91 24500 10863 0.52 2.33 1991-92 28500 11018 0.53 2.53 1992-93 28000 9903 0.45 2.22 1993-94 55370 17871 0.78 3.80 1994-95 51000 14920 0.61 3.01 Sources: FCI Performance Budget documents; Economic Survey, various years. 17 Table 3.2 Central Government Expenditure on Cereal Subsidy and Cost of Carrying Cereals (Rs million) Items 1990-91 1991-92 1992-93 1993-94 1994-95 1) Cereal Subsidy20 20718 28921 32237 31738 23507 (81.31) (86.94) (87.73) (71.81) (55.91) Wheat 10365 15017 18096 16112 15178 (40.68) (45.12) (49.24) (36.45) (36.10) Rice 10353 13904 14141 15626 8329 (40.63) (41.81) (38.48) (35.35) (19.81) CoarseCerealse 1 2 - 1 ( 0.00) (0.01) (0.00) (0.00) (0.00) 2) Canying Charge 4762 4327 4507 12453 18535 Buffer Stocks (18.69) (13.01) (12.26) (28.17) (44.09) Total (1+2) 25479 33238 36745 44193 42041 (100.00) (100.00) (100.00) (100.00) (100.00) 8 Incidentals on Coarse cereals paid to Maharashtra Government as subsidy. Source: FCI Performance Budget. As already noted, subsidized food is distributed through various poverty alleviation programs (PAPs) in addition to PDS. The distribution of food subsidy by antipoverty programs is shown in Tables 3.3. The share of PAPs in the cereal subsidy has declined in the recent years. PDS of course is the major program through which the food subsidy is delivered. In 1994-95, 56 per cent of wheat subsidy, 92 per cent of rice subsidy and 69 percent of cereal (wheat and rice) subsidy accrued to the PAPs. A part of FCI grains are sold also in the open market.21 20The expenditure on cereal subsidy is lower than the expenditure on the food subsidy given in Table 3.1 because the fonmer excludes the subsidy on sugar, edible oil etc. However, cereal subsidy includes the carrying cost of stocks held for the PDS operations. 21 Two caveats are in order here. First, although 69 percent of cereal subsidy was due to the allotments made to PAPs, not all of it may have actually gone to the poor since access to PDS remained universal. Second, it is not possible to assert that FCI open market cereal sales (of 29 percent) has not benefited the poor at all. To the extent such open market sales kept the retail price of food staples low, the poor (as much as the rich) may have indirectly benefited. An analysis of general equilibrium effects of food price interventions is beyond the scope of this paper. 18 Table 3.3 Details of FCI Cereal Sales and Cereal Subsidies During 1990-96 Quantity (million tons) Subsidy (Rs million) Item 1990-91 1991-92 1992-93 1993-94 1994-95 1995-96 1990-91 1991-92 1992-93 1993-94 1994-95 1995-96 Public Distribution System 13.28 17.09 13.71 10.78 8.63 10.76 16514 23264 23375 17069 8827 15647 (78.03) (80.2) (76.41) (57.83) (44.38) (40.9) (79.34) (80.17) (72.29) (55.26) (37.4) (41.75) Integrated Tribal Development 1.92 2.20 3.50 4.15 4.36 3.96 3018 4158 7694 8811 7110 7783 Programn/ Revamped Public Distribution (11.25) (10.29) (19.49) (22.26) (22.41) (15.05) (14.5) (14.33) (23.8) (28.53) (30.12) (20.77) Sytemn Jawvahar Rojgar Yojana 0.03 0.02 0.32 0.40 0.22 0.24 36 25 535 686 301 354 (0.21) (0.11) (1.78) (2.16) (1.14) (0.90) (0.18) (0.09) (1.66) (2.22) (1.27) (0.94) All Poverty Alleviation Programs 15.23 19.30 17.54 15.33 13.22 14.96 19569 27447 31605 26566 16238 23784 (89.5) (90.6) (97.68) (82.24) (67.93) (56.86) (94.01) (94.58) (97.74) (86.01) (68.8) (63.46) OpenMarket 1.28 1.01 0.05 2.98 5.68 5.16 491 152 111 3659 6413 6474 Sales (7.52) (4.76) (0.28) (15.99) (29.18) (19.62) (2.36) (0.52) (0.34) (11.85) (27.17) (17.27) Export 0.27 0.73 0.04 0.04 0.00 4.50 346 1137 10 8 1 4592 (1.56) (3.43) (0.21) (0.21) (0.02) (17.11) (1.66) (3.92) (0.03) (0.03) (0.01) (12.25) Others 0.25 0.25 0.33 0.30 0.56 1.69 311 184 511 557 854 2537 (1.42) (1.21) ( 1.84) ( 1.56) (2.87) (6.42) (1.50) (0.63) (1.58) (1.80) (3.62) (6.77) Total 17.03 21.29 17.96 18.65 19.46 26.31 20718 28921 32238 31738 23507 37386 (100.00) (100.00) (100.00) (100.00) (100.00) (100.00) (100.00) (100.00) (100.00) (100.00) (100.00) (100.00) Note: Figures in parentheses are percentage to total. Data for the year 1995-96 are revised estimates. Subsidies are estimated by multiplying the quantity sold by the difference betNveen economic cost and sales realization per quintal. Source: FCI Performance budget In recent years, there has been an increase in the supply of subsidized grains to tribal and poverty-stricken areas distributed through ITDP/RPDS. Their share in cereal subsidy rose from 14 per cent in 1991-92 to 30 per cent in 1994-95. The increase in the share of ITDP/RPDS suggests some improvement in the targeting of subsidized cereals. However, this increase in the share of ITDP/RPDS in the cereal subsidy should be seen against the overall declining trend of cereal subsidy going to PAPs as a whole from 94 per cent in 1991-92 to 69 per cent in 1994-95. The distribution of cereal subsidy across states is shown in Table 3.4. Nearly one-half of the cereal subsidy has gone to the four southern states of Andhra Pradesh, Kerala, Karnataka and Tamil Nadu and West Bengal which account for about one third of the poor in India whereas about one fifth has accrued to Bihar, Orissa, Madhya Pradesh and Uttar Pradesh which have about half of the poor. One would have expected the cereal subsidy to accrue disproportionately to states with a high incidence of poverty. In reality, the cereal subsidy has not accrued even proportionately to states with a high incidence of poverty.22 FCI Costs'and Subsidies The details of FCI costs and subsidy are given in Table 3.5. It can be seen that the economic cost exceeds the average sales realization23 and the difference is being met by the central subsidy. The subsidy per quintal fluctuated depending upon whether or not the increase in procurement price is passed on to the consumer through a corresponding increase in issue price. Thus in 1992-93, wheat subsidy increased by 40 percent when the procurement price of wheat was increased, but the issue price to the consumers was not raised immediately. In the following years, as the issue price was adjusted upwards, the wheat subsidy rate has declined somewhat, though it continues to be higher than in 1990-91. In general, the increases in procurement prices have been passed on to the consumers by increases in issue prices.24 As a result, the differences between FCI issue prices and market prices have been getting narrowed in recent years. 22 The regional mistargeting is more glaring for rice subsidy; two states, Andhra Pradesh and Kerala which account for less than 10 percent of the poor have received about 39 percent of the rice subsidy. 23 Te average sales realization from rice sale fell short of the procurement price of rice during 1990-92 and wheat in 1992- 93. 24 following regression equations estimated for the period 1986-95 (excluding 1987-88 for which data are not available) provide an indirect evidence. Wheat: In PR = 1.34 + 0.72 In PP + 0.02t, R2 = 0.96 (2.27) (0.03) Rice: InPR=-1.00+ 1.16InPp-O.Olt,R2=0.99 (2.83) (0.15) where PR is the average sales realization per quintal, Pp is the average procurement price and t is the time. t-values are given within parentheses. 20 Table 3.4 State-Wise FCI Supplies of Cereals and Central Subsidies in 1993-94 Quantity Distributed (million tons) Subsidy (Rs millions) Wheat Rice Total Wheat Rice Total Andhra Pradesh 0.239 2.304 2.543 389 3726 4115 (2.61) (24.28) (13.65) (2.41) (23.84) (12.96) Assam 0.274 0.509 0.783 522 838 1360 (3.00) (5.36) (4.2) (3.24) (5.36) (4.29) Bihar 0.663 0.13 0.793 1218 209 1427 (7.25) (1.37) (4.26) (7.56) (1.34) (4.5) Gujarat 0.363 0.243 0.606 728 428 1156 (3.97) (2.56) (3.25) (4.52) (2.74) (3.64) Haryana 0.621 0.026 0.647 811 48 859 (6.79) (0.27) (3.47) (5.03) (0.31) (2.71) Jammu & Kashnmir 0.188 0.197 0.385 337 305 641 (2.06) (2.08) (2.07) (2.09) (1.95) (2.02) Karnataka 0.42 0.633 1.053 742 1076 1818 (4.59) (6.67) (5.65) (4.6) (6.88) (5.73) Kerala 0.332 1.604 1.936 613 2400 3013 (3.63) (16.9) (10.39) (3.8) (15.36) (9.49) Madhya Pradesh 0.432 0.249 0.681 803 426 1229 (4.72) (2.62) (3.65) (4.98) (2.73) (3.87) Maharashtra 0.759 0.653 1.412 1370 1047 2416 (8.3) (6.88) (7.58) (8.5) (6.7) (7.61) Orissa 0.299 0.218 0.517 561 414 975 (3.27) (2.3) (2.77) -(3.48) (2.65) (3.07) Punjab 0.257 0.095 0.352 344 294 637 (2.81) (1) (1.89) (2.13) (1.88) (2.01) Rajasthan 0.657 0.025 0.682 1470 43 1513 (7.18) (0.26) (3.66) (9.12) (0.27) (4.77) Tamil Nadu 0.429 0.875 1.304 620 1360 1980 (4.69) (9.22) (7) (3.85) (8.7) (6.24) Uttar Pradesh 1.301 0.281 1.582 1952 506 2458 (14.22) (2.96) (8.49) (12.12) (3.24) (7.74) West Bengal 0.776 0.464 1.24 1491 770 2261 (8.48) (4.89) (6.65) (9.25) (4.93) (7.12) All India 9.147 9.489 18.636 16113 15626 30789 (100.00) (100.00) (100.00) (100.00) (100.00) (100.00) Note: Figures in parentheses are percentage distribution. Source: Performance Budgets of FCI. 21 Table 3.5 Details of Economic Cost and Subsidy for Distribution Operations (Rs per quintal) ffheat Rice Items 1990-91 1991-92 1992-93 1993-94 1994-95 1995-96 1990-91 1991-92 1992-93 1993-94 . 1994-95 1995-96 A. Average net 204.68 211.20 295.51 325.21 334.80 351.10 338.42 372.45 433.72 500.40 537.77 584.60 procurement price (10.93) (3.19) (39.92) (10.05) (2.95) (2.22) (5.72) (10.06) (16.45) (15.37) (7.47) (8.71) B. Procurement Cost 64.34 87.69 89.28 89.28 110.86 109.28 41.46 32.69 31.53 40.25 58.13 52.98 (41.44) (36.29) (1.81) (0.00) (24.17) (-1.43) (87.52) (-21.15) (-3.55) (27.66) (44.42) (-8.86) C. Total purchase 269.02 298.89 384.79 414.49 454.34 460.38 379.88 405.14 465.25 540.65 595.90 637.58 cost (A+B) (16.97) (11.1) (28.74) (7.72) (9.61) (1.33) (11.00) (6.65) (14.84) (16.21) (10.22) (6.99) D. Distribution cost 87.48 91.90 120.02 117.44 105.51 103.19 87.48 91.90 120.02 124.45 98.81 109.14 (14.61) (5.05) (30.60) (-2.15) (-10.16) (-2.20) (14.61) (5.05) (30.60) (3.69) -20.60) (10.45) E. Economic cost 356.50 390.79 504.10 532.03 551.17 563.57 457.52 497.04 585.27 665.10 694.71 746.72 (16.38) (9.62) (29.00) (5.54) (3.60) (2.25) (9.31) (8.64) (17.75) (13.64) (4.45) (7.49) F. Sales realization 239.95 251.68 279.36 355.88 407.89 415.87 330.02 365.58 442.40 500.42 600.75 610.70 (20.32) (4.89) (11.00) (27.39) (14.61) (1.96) (12.13) (10.78) (21.01) (13.11) (20.05) (1.66) G. Subsidy 116.55 139.11 224.74 176.15 143.28 147.70 127.50 131.46 142.87 164.68 93.96 136.02 (9.03) (19.36) (61.56) (-21.62) (-18.66) (3.08) (2.63) (3.11) (8.68) (15.27) (42.94) (44.76) H. Sales (million 8.89 10.79 8.05 9.15 10.59 13.50 8.12 10.58 9.90 9.49 8.86 12.80 tonnes) (15.09) (21.3) (-25.35) (13.60) (15.81) (27.44) (6.38) (30.26) (-6.41) (4.13) (-.59) (44.4) I. Subsidy amount 10365 15006 18096 16112 15178 19940 10351.72 13903.21 14141.27 15626.49 8328.61 17410.56 (Rsmillions) (25.48) (44.78) (20.59) (-10.96) (-.80) (31.38) (9.18) (34.31) (1.71) (10.50) (46.70) (109.05) Notes: Data for the year 1995-96 are revised estimates. Figures in parentheses are percentage change over previous year. For the years 1993-94 and 1994-95 purchase cost and distribution costs do not add exactly up to economic cost due to adjustment for stock prior costs etc. Net Procurement Price represents the prices paid for Wheat and Rice. Procurement Costs include i) mandi charges made up of market fee, auction fee rural development commission to kutcha Archaliyas, ii) payments to mandi labor iii) forward charges iv) internal movement charges v) cost of gunny used in bagging the grains vi) purchase tax vii) charges paid to State agencies for establishment storage and interest. Distribution Costs include cost incurred on transportation, interest, storage, grain losses, handling labor, administration overheads, charges paid to state agencies for procurement and stocks held by them till taken over by FCI. Sales Realization represents the annual total sales realization divided by the quantity sold Economic Cost is the sum of procurement costs and distribution costs. Central government subsidy is the difference betwveen economic cost and sales realized Source: Performance Budgets of FCI. PDS Supplies PDS supplies have increased rapidly since the mid-sixties; the average annual supply increased from 6.5 million tonnes during 1961-65 to 18.4 million tonnes during 1990-92. A noteworthy feature is the response of government supply to fluctuations in production; it has been higher in drought years than in normal ones. For instance, in 1979-80 and 1987-88, years associated with low foodgrain production, PDS supplies were substantially higher than in nortnal years. In both the years, the government increased the supplies through the employment programs in distressed areas.25 Studies suggest that the PDS operations are not very sensitive to inflationary situations (see Parikh, 1992). For instance, in 1993-94, the year in which the annual inflation rate (point-to- point) went on increasing from 6.9 per cent in April to 10.5 in March and the wholesale price of foodgrains increased by 7.8 per cent, the central pool foodgrain stocks had risen (from 11.8 mt in January, 1993 to 22.0 mt in January, 1994) and foodgrain off-take from PDS had dropped (from 17.0 mt in 1992-93 to 14.7 mt in 1993-94). Another recent trend is the steady decline of off-take of cereals from PDS. Between 1991-92 and 1994-95, wheat off-take from PDS fell from 8.78 to 4.83 million tonnes and rice off- take fell from 9.94 to 8.00 million tonnes. Subsidized wheat distributed through PAPs declined from 9.06 million tonnes in 1991-92 to 5.13 million tonnes and rice from 10.24 to 8.09 million tonnes.26 The decline in off-take can be attributed to the abnormal increase in the issue price of wheat and rice in recent years. These upward revisions of issue prices have reduced the differential between the issue price and the open market price, which in turn has resulted in a shift of consumers from PDS to the open market.27 The minimum support prices were raised substantially between 1990-91 and 1995-96 (69 per cent for wheat and 44 per cent for rice), resulting in FCI buying up more foodgrains than it can manage, since it has no option but to buy whatever produce is offered.28 Since off-take from PDS has also declined due to the upward revision of issue prices, Central buffer stocks have reached uneconomical levels. While the recommended stock of foodgrains on 1st July of every year is 22.3 million tonnes,29 the FCI had a foodgrain stock of 35.6 million tonnes in July 1995. FCI had an excess holding of about 7.72 million tonnes in 1994 and 13.4 million tonnes in 1995, which means that the carrying cost of buffer stocks would have increased by 45 per cent in 1994 25 The field study of Acharya (1988) carried out in the distressed areas of Rajasthan to assess the effects of drought shows that the per capita consumption of cereals could be maintained due to the distribution of subsidized cereals in the employment programs. But the per capita consumption of other food items did decline. 26 However, subsidized wheat distributed in rTDP/RPDS increased from 0.8 mt in 1991-92 to 1.8 mt in 1994-95 and subsidized rice distributed in ITDPIRPDS from 1.4 mt to 2.6 mt. The distribution of subsidized wheat and rice in the employment programs (JRY/EAS) tended to be small during 1991-95. 27 Kailash Sarap in a personal correspondence reports that in January 1996 in Sambalpur town (Orissa State), PDS rice was sold at Rs.6.22/kg. whereas the market price of rice ranged between Rs.6.00-6.50/kg. depending on quality. 28 Field investigations in Andhra Pradesh and Uttar Pradesh have revealed wide prevalence of bribing in the transactions. The bribing takes place in the form of FCI purchasing inferior quality cereals at higher prices fixed for better quality. The determination of quality is very crude and is left to the subjective judgment of Quality Inspectors of FCI. The process of weighing also results in bribing. Thus, both farners and FCI procurement staff gain at the cost of the government. 29 The minimum norm of 22.3 million tonnes on Ist July includes 12.3 million tomnes required for PDS and a holding of 10 million tonnes as an insurance against fluctuations in foodgrain production. 23 and 78 per cent in 1995. Since the carrying cost of one tonne of foodgrains in 1994-95 was Rs 1447 (FCI, Performance Budget, December 1995), the extra stock would have involved an additional Central Government expenditure of about Rs 9689 million in 1994-95. This is clearly an avoidable waste of resources. Recently, the Central Government has permitted the FCI to off- load the foodgrain stocks at the prices fixed by it. The prices fixed by the Government are generally higher than the Central issue price, but lower than the economic cost.30 To that extent, the losses on account of excess holding of stocks is not entirely elirninated. Regional Patterns in Allocations and Off-Take of Foodgrains A number of considerations determine the state-wise allocations of grains by the Central government from the Central pool, including the historical allocation patterns, foodgrain availability and prices in various states.31 The historical allocation patterns still retain the earlier emphasis of PDS confining its coverage to urban areas and a few deficit regions. In the period of overall food scarcity i.e., prior to the 1980s, the allocations from the Central pool were lower than the demands of the state governments32 and the states were lifting a high proportion of their quota. With the recent change in the overall foodgrain supply scenario from one of scarcity to a surplus situation, the allocations from the Central pool are no longer a binding constraint for the distribution of PDS grains within a state. Moreover, the recent upward revisions in the issue prices of wheat and rice which narrowed the difference between the open market and subsidized prices, have adversely affected the off-take levels.33 Consequently, in all the states, the off-take was lower than the Central allocation (see Table 3.6). However, in states such as Andhra Pradesh and Kerala which are implementing subsidized food schemes giving additional subsidies, the rice off-take was high and closer to allocations. In poorer states such as Bihar, Orissa, and Madhya Pradesh, the off-take was substantially lower than Central allocations partly due to the narrow difference between the Central issue price and open market price and partly because of the weak fiscal capacity of these states to finance any additional subsidy via lowering of issue prices. (see the data in Table 3.7). Besides these states have not yet built up effective institutional mechanisms to lift the quotas from FCI depots to fair price shops.34 Ironically, the incidence of poverty is high precisely in the states with the lowest off-take of FCI grains. 30 It may be noted that the economic cost does not include the return on fixed investment, and is not adjusted for subsidized transport and credit. 31 The criteria of allocations are not explicitly stated. The monthly allotments to the states are supposed to be determined by the Central Civil Supplies Department based on the requests received from the states. The state governments submit every month their requests based on the standard ration population of the state and the scale of ration. There are substantial differences between the states requests and the Central allocations. In the recent years, with the increase in the Central stocks, Central allocations have been on the liberal side. 32 Whenever the size of the Central pool was lower, the Central govermment reduced the state-wise allocations on a pro-rata basis making some marginal adjustments for state specific conditions. In periods of food scarcity, the states tended to exaggerate their PDS requirements and reduce their procurement targets. 33 It was reported in Indian Express (March 20, 1994) that in Karnataka the quantity of rice purchased by ration card holders from the fair price shops declined in 1994 because the issue price of rice was raised to Rs.7.05/kg. whereas they could buy in the open market good quality rice at Rs.7.50-8.00/kg. The Karnataka govenunent was compelled to surrender a huge quantity of rice to the Food Corporation of India. 34 Numerous studies have brought out the deficiencies in the administration of PDS as well as disfunctioning of fair price shops in poorer states (see Kabra and Ittyerah, 1986 and Tyagi, 1991). Tyagi observed that casual workers in Uttar Pradesh who purchase grains on a day to day basis did not buy from PDS because ration quantities were supplied once or twice in a month and the day of supply was also uncertain. Since there was no demand, the fair price shop dealers were not procuring the grains. Kabra and Ittyerah showed how the fair price shops were financially unviable on account of low withdrawals. 24 Table 3.6 Allocation for PDS and Off-Take of Rice and Wheat in States and Union Territories of India Figures in '000 Tons Wheat State/U. T 1993-94 1994-95 1995-96 Allocation Lifiing Allocation Lifting Allocation Lifting Andhra Pradesh 172.60 117.00 180.00 110.20 192.00 102.70 Arunachal Pradesh 7.20 6.50 15.20 12.80 7.20 5.20 Assam 260.00 231,90 310.00 276.00 360.00 351.90 Bihar 725.20 427.80 714.00 222.40 705.60 227.20 Goa 37.20 22.90 40.30 17.30 42.40 21.40 Gujarat 642.00 306.20 642.00 378.90 835.50 424.90 Haryana 141.90 74.10 151.80 30.90 209.48 62.50 Himachal Pradesh 121.00 107.60 136.00 117.50 144.00 98.20 Jammu & Kashmir 240.00 126.00 350.00 360.00 360.00 116.00 Kamataka 295.00 257.20 360.00 262.40 360.00 219.50 Kerala 350.00 315.50 445.00 386.70 585.00 557.30 Madhya Pradesh 492.00 247.40 560.94 142.70 583.92 136.70 Maharashtra 960.00 528.50 960.00 463.60 1100.00 610.60 Manipur 34.40 24.70 32.40 13.20 32.40 31.80 Meghalaya 32.00 29.70 24.00 24.40 28.00 27.70 Mizoram 17.30 16.50 21.50 16.80 24.00 23.40 Nagaland 40.50 33.10 70.00 59.30 18.20 20.00 Orissa 270.00 221.00 415.00 182.40 420.00 238.60 Punjab 240.00 12.80 210.00 2.40 155.00 7.60 Rajasthan 1140.00 577.30 1443.69 528.20 1453.92 459.30 Sikkim 7.20 6.10 10.00 8.30 12.30 10.10 Tamil Nadu 245.00 235.70 300.00 155.10 310.00 162.40 Tripura 21.60 13.90 21.60 7.90 UttarPradesh 972.60 391.60 1185.60 203.20 1185.60 225.90 West Bengal 992.00 821.80 1035.00 751.60 1098.60 842.20 Andman & Nicobar Islands 0.40 8.40 -- 9.00 12.60 Chandigarh 21.60 11.00 21.60 0.90 21.60 0.90 D & N Haveli 2.40 -- 2.40 -- 2.75 0.50 Daman & Diu 1.80 0.30 1.80 -- 2.15 -- Delhi 864.00 558.80 936.00 205.00 840.00 153.30 Lakshadweep 0.50 -- 0.50 0.10 0.50 -- Pondicherry 9.00 -- 9.00 -- 9.00 -- ALL INDIA 9368.60 5723.30 10613.73 4690.40 11129.72 5147.40 25 Table 3.6 (cont..) Allocation for PDS and Off-Take of Rice and Wheat in States and Union Territories of India Figures in '000 Tons Rice State/lU. T 1993-94 1994-95 1995-96 Allocation Lifting Allocation Lifting Allocation Lifting AndhraPradesh 2282.50 2172.90 2230.00 2198.80 2620.00 2159.31 Arunachal Pradesh 124.20 87.40 90.70 74.50 102.70 90.60 Assam 510.80 397.30 465.80 310.70 568.00 436.30 Bihar 299.20 86.40 372.00 39.50 381.60 23.50 Goa 54.00 39.80 63.00 39.40 78.00 44.80 Gujarat 414.00 238.90 414.00 187.60 409.00 208.60 Haryana 36.00 16.20 36.00 5.90 53.56 8.40 Himachal Pradesh 80.60 77.20 108.00 38.90 131.00 45.20 Jammu & Kashmir 434.40 162.90 520.20 154.20 528.00 249.20 Karnataka 828.50 597.90 1307.34 678.00 1443.12 942.90 Kerala 1825.00 1578.30 1800.00 1118.10 1800.00 1170.50 Madhya Pradesh 490.80 190.90 559.74 162.40 580.16 204.20 Maharashtra 858.00 578.00 858.00 284.50 858.00 359.50 Manipur 120.00 49.50 120.00 30.60 120.00 33.00 Meghalaya 126.00 111.20 136.00 113.40 172.00 164.20 Mizoram 103.20 98.40 94.00 79.70 94.00 93.63 Nagaland 112.60 94.90 84.00 75.90 72.50 70.80 Orissa 464.40 185.30 543.60 192.30 790.00 365.80 Punjab 18.00 5.20 17.25 1.30 16.65 1.80 Rajasthan 84.60 18.40 46.00 14.40 52.00 9.10 Sikkim 54.00 42.70 56.10 33.80 57.60 43.80 Tamil Nadu 878.80 856.70 1200.00 1224.20 1590.00 1587.92 Tripura 194.40 146.80 194.94 125.80 194.40 148.60 Uttar Pradesh 535.60 232.50 549.60 197.30 549.60 209.50 WestBengal 967.20 528.90 932.40 434.30 856.00 457.10 Andman & Nicobar 36.00 -- 31.25 -- 30.00 -- Islands Chandigarh 3.60 2.00 3.60 3.80 3.60 1.10 D & N Haveli 6.00 1.10 6.00 - 6.00 Daman & Diu 6.00 1.20 6.00 5.10 6.70 Delhi 240.00 143.00 240.00 50.80 240.00 26.40 Lakshadweep 6.30 3.30 6.30 6.70 6.30 Pondicherry 24.00 3.42 24.00 2.90 24.00 1.90 ALL INDIA 12218.70 8749.02 13115.30 7884.80 14434.50 9164.56 Source: Food Corporation of India. The per capita off-take was low in food surplus states with a low incidence of poverty such as Punjab, Haryana, as well as in poorer states (with low fiscal capacity) such as Bihar, Orissa, Madhya Pradesh and Uttar Pradesh. The off-take was high in food-deficit but relatively (fiscally) richer states such as Gujarat and Maharashtra. 26 Table 3.7 Per Capita Off-Take Levels of Cereals and its Correlates Off-Take from FCI Per Capita State Government's (Kg. per annum, 1993-94) Poverty State Domestic (Rs during 1993-94) Product States Per Capita Per Rice Wheat Total 1986-87 1993-94 Expenditure Capita Receipts AndhraPradesh 31.34 1.69 33.02 27.20 6651.00 1520.22 1558.01 Assam 16.84 9.83 26.67 36.84 5916.00 1534.56 1475.88 Bihar 0.95 4.70 5.65 53.37 3650.00 927.24 921.52 Gujarat 5.54 7.11 12.65 32.33 7600.00 1951.24 2004.24 Haryana 0.93 4.27 5.21 16.63 10359.00 2369.68 2401.09 Jammuu&Kashmir 20.04 15.50 35.54 23.20 4244.00 3189.78 3038.13 Karnataka 12.81 5.51 18.32 38.14 7029.00 1733.26 1761.48 Kerala 52.49 10.49 62.98 32.08 6242.00 1708.76 1741.85 Madhya Pradesh 2.75 3.56 6.31 43.40 5485.00 1279.01 1268.25 Maharashtra 7.00 6.40 13.39 40.10 10984.00 1934.30 Orissa 5.60 6.68 12.29 55.61 4726.00 1347.70 1368.90 Punjab 0.25 0.61 0.86 12.70 12319.00 2484.87 2366.84 Rajasthan 0.40 12.47 12.87 34.60 5220.00 1604.74 1590.55 TamilNadu 14.98 4.12 19.10 45.13 7352.00 1759.34 1781.36 Uttar Pradesh 1.60 2.70 4.30 41.99 4744.00 1121.97 1080.46 West Bengal 7.46 11.60 19.06 43.99 6055.00 1125.01 1109.20 Note: The State Governments' expenditures include both cufrent and capital expenditures. The State receipts include both tax and non-tax as well as grants from the Center. Sources: Computed by the authors; Statistical Abstracts, 1990-95; Export Committee on Poverty Report. PDS Food Access to the Poor Many studies have pointed out the biases in the inter-regional distribution of PDS supplies, and the very minimal distributional impact of the program. How far do the above features still persist? These findings are based on slightly dated information (42nd NSS round, 1986-87). There has been no recent nationwide survey to obtain information on the distributional incidence of PDS. Based on the results of some selected village studies on PDS in 1995 in four states, one can provide a crude answer to the above question by comparing the class specific purchases of cereals from the PDS estimated from the NSSO (Table 3.8) with those of more recent village level data (Table 3.9). 27 Table 3.8 Monthly per Capita Purchase of Cereals (kgs) from PDS and their Percentage to Total Cereal Consumption 1986-87 Rural Areas Urban Areas S. No. States Very Moderately Poor Non- All Very Moderately Poor Non- All Poor Poor poor Classes Poor Poor poor Classes Andhra Pradesh 2.33 2.49 2.43 2.56 2.53 2.63 2.83 2.72 2.23 2.44 26.15 19.92 21.99 17.37 18.19 26.04 24.35 25.28 18.83 21.44 Assam 0.93 0.79 0.84 0.63 0.72 1.78 1.20 1.40 1.25 1.28 9.71 6.64 7.58 4.27 5.43 18.22 10.96 13.27 10.15 10.76 Bihar 0.04 0.05 0.05 0.07 0.06 0.16 0.06 0.12 0.38 0.26 0.36 0.36 0.40 0.39 0.39 1.45 0.46 1.01 2.76 2.02 Gujarat 1.43 1.59 1.52 1.21 1.30 1.25 1.13 1.18 0.71 0.91 18.97 16.58 17.43 10.05 11.76 15.36 12.65 13.79 7.91 10.34 Haryana 0.02 0.02 0.02 0.01 0.01 0.10 0.08 0.08 0.08 0.08 0.18 0.15 0.16 0.07 0.07 0.86 0.75 0.74 0.76 0.75 Jammu&Kashmir 1.82 2.92 2.61 1.89 2.08 6.08 5.58 5.70 7.54 7.16 13.21 19.74 18.00 9.91 11.67 53.10 42.89 45.13 49.80 48.97 Kanataka 0.89 0.99 0.94 1.16 1.08 1.15 1.53 1.31 2.02 1.70 9.50 8.31 8.86 7.92 8.21 13.19 16.22 14.54 17.57 16.38 Kerala 4.31 4.39 4.35 4.52 4.46 4.29 4.79 4.47 4.02 4.24 57.93 49.05 52.16 40.47 43.39 56.30 47.19 52.34 38.10 44.40 Madhya Pradesh 0.26 0.28 0.27 0.27 0.27 0.47 0.54 0.50 0.52 0.51 2.15 1.96 2.04 1.61 1.80 4.14 4.47 4.27 4.44 4.36 Maiarashtra 1.05 1.00 1.02 1.21 1.13 1.26 1.80 1.49 1.51 1.50 11.50 8.67 9.85 9.34 9.54 14.84 18.54 16.56 16.06 16.23 Orissa 0.01 0.00 0.01 0.08 0.04 0.13 0.23 0.17 0.51 0.35 0.08 0.00 0.07 0.43 0.25 0.97 1.67 1.25 3.55 2.50 Punjab 0.19 0.02 0.06 0.00 0.01 0.06 0.05 0.05 0.04 0.04 2.62 0.22 0.71 0.00 0.08 0.76 0.58 0.59 0.38 0.39 Rajasthan 1.58 1.12 1.32 0.68 0.86 0.39 0.21 0.29 0.24 0.26 12.59 7.87 9.76 3.79 5.15 3.34 1.70 2.41 1.71 1.95 Tamil Nadu 1.20 1.27 1.24 1.34 1.30 1.18 1.26 1.22 1.17 1.19 12.86 11.20 12.09 9.77 10.59 14.59 13.07 13.82 10.84 12.00 UttarPradesh 0.11 0.13 0.12 0.27 0.21 0.20 0.17 0.19 0.43 0.32 0.95 0.97 0.95 1.53 1.33 1.89 1.48 1.81 3.45 2.72 WestBengal 0.87 1.00 0.94 0.96 0.95 2.59 3.45 3.05 3.42 3.31 7.85 6.88 7.26 5.52 6.16 24.98 28.73 27.14 27.23 27.22 AllIndia 0.70 0.81 0.76 0.95 0.88 1.19 1.36 1.27 1.39 1.34 6.51 6.27 6.38 6.08 6.17 12.65 12.73 12.59 12.30 12.26 Notes: Monthly per capita purchases from PDS. PDS purchases as per cent of total cereal consumption. Poverty lines used in classification of households are the state specific poverty lines adopted by the Expert Group on Estimation of proportion and number of poor, appointed by the Government of India. Households having per capita expenditure less than three quarters of poverty line are classified as very poor. Source: Computed by the authors from the NSS 1986-87. The evidence shows that even now, the efficacy of PDS in distributing food to the poor seems to be as bad as in 1986-87 and that some of the disquieting features persist-the virtual exclusion of backward states such as Bihar and Uttar Pradesh from the PDS network and the universal character of PDS in states in which PDS off-takes were significant. None of the four villages surveyed in Bihar received any PDS supplies, nor did three out of four villages surveyed in Uttar Pradesh. 28 Table 3.9 Per Capita Monthly Consumption (KG.) of Rice from PDS in Andhra Pradesh and Kerala and Cereals from PDS in Bihar and Uttar Pradesh 1995 Andhra Pradesh Expenditure Class Village I Village 2 Village 3 Village 4 Village 5 (Rs/month/per capita) Nettampadu Singtham Jaggasagar Narsyapalem Machavaram Less than 175 2.28 1.07 2.81 4.81 2.81 (22.96) (12.71) (36.34) (48.73) (30.34) 176-225 2.27 2.57 2.27 2.53 2.91 (21.43) (22.69) (20.99) (24.85) (27.39) 226-275 1.81 0.00 0.94 3.38 2.51 (14.77 (8.44) (27.95) (32.77) 276-350 0.00 0.00 0.47 0.39 1.66 (1.86) (7.52) (14.53) Above 350 0.00 0.00 0.00 0.43 11.26 (1.63) (9.64) All Classes 2.42 1.41 1.66 2.32 2.77 (22.92) (14.07) (18.05) (24.33) (28.97) Kerala Expenditure Class Village I Village 2 Village 3 Village 4 (Rs/month/per capita) Anand Anjengo Edavaka Peruthoor Lessthan 175 3.40 1.01 6.17 5.00 (46.01) (14.45) (80.33) (100) 176-225 3.17 1.95 4.28 5.18 (46.55) (28.89) (69.59) (68.07) 226-275 4.11 2.12 4.78 3.81 (53.03) (21.2) (53.95) (44.82) 276-350 5.01 3.12 3.86 5.63 (44.49) (29.46) (42.37) (66.47) 351-500 6.39 4.08 3.86 4.52 (58.46) (36.82) (26.62) (54.19) 501-1000 4.74 3.89 3.75 6.13 (40.83) (23.50) (38.38) (54.10) Above 1000 7.43 0.00 0.00 0.00 (79.89) Village 4.77 2.70 4.21 4.73 (48.47) (26.06) (41.35) (55.52) 29 Table 3.9 (cont..) Per Capita Monthly Consumption (KG.) of Rice from PDS in Andhra Pradesh and Kerala and Cereals from PDS in Bihar and Uttar Pradesh 995 Bihar Uttar Pradesh Expenditure Class Village ) Village 2 Village 3 Village 4 Village I Village 2 Village 3 Village 4 (Rs/month/ Kaghar per capita) Bharat Feta Parsa Pathraha Jamgain Jha ra Sadar Khalispu Less than 175 0.00 0.00 0.00 0.00 0.00 1.75 0.00 0.00 (37.23) 176-225 0.00 0.00 0.00 0.00 0.00 1.90 0.00 0.00 (37.33) 226-275 0.00 0.00 0.00 0.00 0.00 1.86 0.00 0.00 (44.93) 276-350 0.00 0.00 0.00 0.00 0.00 3.03 0.00 0.00 (58.16) Above 350 0.00 0.00 0.00 0.00 0.00 2.15 0.00 0.00 (44.38) All classes 0.00 0.00 0.00 0.00 0.00 2.08 0.00 0.00 (43.24) Note: Quantities in kgs. Figures in brackets are the percentage of Cereals from PDS in the total cereal consumption. The data relates to 1995. Source: UNDP Research Project on Strategies and Financing for Human Development. Empirical evidence for urban bias seems to be weak (see Table 3.8). Out of the eight states (Andhra Pradesh, Gujarat, Jammu & Kashmir, Karnataka, Kerala, Maharashtra, Tamil Nadu and West Bengal) in which PDS network has spread, only in Jammu & Kashmir, Kamataka, Maharashtra and West Bengal, were the per capita purchases from PDS as well as their share in the market purchases of cereals higher in urban areas. However, among these states, the urban bias was extreme in Jammu & Kashmir and West Bengal. An interesting aspect is the coverage and targeting of PDS in Kerala. In rural Kerala, 1986-87 NSS data show that (i) per capita monthly PDS cereal purchases (4.3-4.5 kg.) tended to be regressive across the expenditure groups, and (ii) per capita monthly PDS cereal purchases constituted 40-58 percent of cereal consumption, and 43-58 percent of the market purchases. In the case of the very poor, while the per capita PDS cereal purchase was low, its share in market purchases as well as in cereal consumption was high because the overall spending (market + PDS) levels were low. The dependency of the very poor on PDS was very high. In regard to urban areas of Kerala, the extent of coverage and inter-class patterns in PDS purchases were more or less similar to those in rural areas with the exception that the inter-class distribution of subsidized cereals was slightly less regressive than in rural areas. In Andhra Pradesh, the monthly per capita purchases in 1986-87 was 2.23-2.56 kg. in rural areas and 2.23-2.63 kg. in urban areas. The share of PDS cereal purchases in cereal consumption was 17-26 percent in rural areas and 18-26 percent in urban areas. Inter-class 30 variations were similar to those in Kerala. Even though there was some targeting in both Kerala and Andhra Pradesh, in reality the coverage was almost universal. The poor did benefit from the PDS in both the states, but so did the nonpoor. In contrast to the impressive coverage of PDS in the states of Kerala and Andhra Pradesh, the coverage was low in All India; the monthly per capita purchases was 0.88 kg. in rural areas and 1.34 kg. in urban areas. The low PDS cereal purchases in All India can be attributed to the extremely low cereal purchases from PDS in the prosperous states such as Punjab and Haryana, as well as in the states with high poverty levels such as Bihar, Orissa, Madhya Pradesh and Uttar Pradesh. The monthly purchases from PDS was lowest for the very poor uniformly across all the states, both in rural and urban areas. As already noted, this was the case even in Kerala and Andhra Pradesh. It suggests that impressive coverage and/or additional state-level spending on subsidy is no guarantee that the very poor are better served. The conclusion is inescapable: PDS has remained an expensive and largely untargeted program. The central issue, therefore, is: how to improve the efficacy of PDS in transferring food to the poor cost-effectively. The policy initiatives should distinguish between the very poor and moderately poor, and attempt at improving the efficacy of PDS in transferring food to the former since the ultrapoor suffer not only from chronic food insecurity but are also severely exposed to the risk of uncertainty both in the food and labor markets. 31 CHAPTER 4 EFFICACY OF PDS AS AN ANTI-POVERTY PROGRAM Introduction In this chapter, an attempt is made to estimate the extent of income transfer through PDS to the poor, and the consequent reduction in poverty in terms of percentage and severity. The analysis has also been extended to assess the nutritional impact of PDS on the poor as well as the cost per rupee of income transferred to the poor. The analysis is done at the all-India level covering the rural and urban areas of 16 major States (which together account for 98 per cent of the rural population and 95 per cent of urban population), and for Andhra Pradesh state. Data The data on utilization of PDS collected by the National Sample Survey Organization (NSSO) during July 1986-June 1987 (42nd Round of NSSO) have been used. The PDS was defined to cover ration shops, fair price shops and control shops. During the survey data was collected on quantities and values of different consumer goods purchased by households from PDS as well as other sources for rice, wheat, jowar, bajra, other cereals, pulses, edible oil, sugar, coal, kerosene, and clothing. The Indira Gandhi Institute of Development Research (IGIDR) has tabulated the survey data for decile classes of households based on their monthly per capita expenditure (PCE). Our analysis is based on the IGIDR-processed data as well as on the original data on consumer expenditure collected by NSSO during the forty-second round survey on household consumer expenditure (through Schedule 1.0 of NSSO). Methodology The subsidy transfer or income gain to a household from PDS is defined as the difference between the expenditure that the household would have incurred in the absence of PDS and the actual expenditure under PDS. It is measured by multiplying the quantity of purchases from PDS with the difference between open market and PDS price. Hence, income gain to a household (Ay) is defined as: Ay = qR (PM - PR) where PM and pR are the open market and subsidized prices, and qR is the quantity purchases from PDS. The open market price and subsidized price have been estimated from the PDS survey data as: PM = eM/qM PR = eR/qR where qM and eM are the quantity and value of purchases from open-market, and qR and eR are the quantity and value of purchases from PDS. 32 For the purpose of analysis, sample households have been classified into the following categories on the basis of monthly per capita expenditure and the poverty line. Households having per capita expenditure (PCE) less than three-quarters of the poverty line are treated as 'vely poor'; PCE higher than three quarters of the poverty line but less than the poverty line as moderately poor'; and PCE higher than the poverty line as 'non-poor'. The poor constitute the sum of the very poor and moderately poor. The poverty lines for the year 1986-87 have been estimated separately for rural and urban areas of the 16 major states and for all-India by following the procedure suggested by the Expert Group on Estimation of Proportion and Number of Poor, appointed by the Government of India. The State-specific poverty lines have been estimated by evaluating a common basket of consumer goods at the prices prevailing in the respective States. The State-specific cost of living index for agricultural laborers has been used in the estimation of State-specific poverty line for rural areas, and the cost of living indices for industrial workers and for urban non-manual employees have been used in the estimation of the State-specific poverty lines for urban areas. The estimated poverty lines for 1986-87 are presented in Table 4.1. These poverty lines are very close to those of the Expert Group for the year 1987-88. In the present study, the extent of poverty is measured by head count ratio, i.e. the percentage of poor in the total population. The severity of poverty is measured by the poverty gap ratio and FGT (Foster, Greer, and Thorbeck) ratio. For estimating the poverty measures we have used the beta type Lorenz curve specified as: L(p) = p_QpG (1 _p)d (1) where, p and L(p) are the population proportion and share in income, and Q, G and d are the parameters. The poverty ratio (head count) H is obtained by solving L'(H) = z/m (2) where z is the poverty line and m the mean income. The poverty gap ratio (PG) can be derived as: PG=1-mL(H)/z (3) The Foster, Greer and Thorbeck poverty FGT2 is given by FGT2=(l1 m/z) (2PG)-(1m-/z)H)+Q2 (mVZ)2 [G2B(,2G-1,2d-1) -2Gd(H,2G,2d+d2B(H,2G+ 1 ,2d- 1 [ where B(k,s,r,) = f k ps (I-P)r-I (4) 0 33 Table 4.1 Population, Percentage of Poor and Very Poor and Poverty Gap Ratio by States in 1986-87 Rural Urban States Popula- Poverty Percentage Percentage Poverty Popula- Poverty Percentage Percentage Poverty tion Line of Very of Poor Gap tion Line of Very of Poor Gap (millions) (Rs 0.0) Poor Ratio (millions) (Rs 0.0) Poor Ratio AndhraPradesh 45.12 82.90 9.09 22.55 0.0552 13.43 146.45 24.14 42.61 0.1241 Assam 14.48 116.23 14.57 41.03 0.0837 1.70 128.05 8.01 23.19 0.0466 Bihar 65.25 103.34 25.25 48.23 0.1291 9.31 148.10 27.58 45.87 0.1442 Gujarat 23.27 101.28 12.64 29.88 0.0757 9.85 161.26 19.94 43.18 0.1055 Haryana 11.49 107.87 5.92 17.81 0.0364 3.55 127.40 3.58 19.09 0.0293 Jammu & Kashmir 5.22 111.91 7.56 27.07 0.0495 1.17 131.63 4.93 20.54 0.0346 Karnataka 27.70 93.87 18.89 36.70 0.0993 11.89 151.66 26.31 44.99 0.1373 Kerala 20.10 123.14 12.80 31.65 0.0764 4.67 163.97 31.49 49.66 0.1621 Madhya Pradesh 44.59 95.74 23.57 48.36 0.1236 10.86 160.25 24.64 46.76 0.1290 Maharashtra 42.51 106.02 21.05 43.22 0.1141 21.45 166.38 22.72 39.13 0.1183 Orissa 24.85 103.73 29.32 52.90 0.1546 2.96 155.36 26.81 47.82 0.1353 Punjab 12.54 107.87 3.77 15.88 0.0268 5.05 126.96 2.33 12.73 0.0193 Rajasthan 29.52 99.34 12.16 28.47 0.0676 7.35 149.30 17.12 35.65 0.0901 TamilNadu 34.03 106.68 22.39 41.80 0.1184 17.66 159.99 23.27 44.66 0.1246 Uttar Pradesh 108.64 96.42 16.69 36.02 0.0890 22.32 141.94 26.31 45.91 0.1363 WestBengal 44.11 118.90 20.45 44.46 0.1115 14.05 138.60 14.18 30.46 0.0753 AllIndia 553.42 16.82 36.31 0.0911 157.27 20.71 39.09 0.1078 Note: Poverty line is defined for per capita per month. Source: Computed by the authors. Table 4.2 Number of Poor and Very Poor and Aggregate Monetary Shortfall by States in 1986-87 million Rural Urban All Very Poor AMS Very Poor AMS Very Poor AMS Poor (Rs/mo) Poor (Rs/mo) Poor (Rs/mo) AndhraPradesh 4.10 10.17 206.47 3.24 5.72 244.08 7.34 15.90 450.56 Assam 2.11 5.94 140.87 0.14 0.39 10.14 2.25 6.34 151.01 Bihar 16.48 31.47 870.51 2.57 4.27 198.82 19.04 35.74 1069.34 Gujarat 2.94 6.95 178.41 1.96 4.25 167.58 4.91 11.21 345.99 Haryana 0.68 2.05 45.12 0.13 0.68 13.25 0.81 2.72 58.37 Jammu & Kashmir 0.39 1.41 28.92 0.06 0.24 5.33 0.45 1.65 34.25 Kamataka 5.23 10.17 258.20 3.13 5.35 247.58 8.36 15.52 505.78 Kerala 2.57 6.36 189.10 1.47 2.32 124.13 4.04 8.68 313.23 MadhyaPradesh 10.51 21.56 527.65 2.68 5.08 224.50 13.19 26.64 752.15 Maharashtra 8.95 18.37 514.24 4.87 8.39 422.20 13.82 26.77 936.43 Orissa 7.29 13.15 398.51 0.79 1.42 62.22 8.08 14.56 460.73 Punjab 0.47 1.99 36.25 0.12 0.64 12.37 0.59 2.63 48.63 Rajasthan 3.59 8.40 198.24 1.26 2.62 98.87 4.85 11.02 297.11 TamilNadu 7.62 14.22 429.83 4.11 7.89 352.05 11.73 22.11 781.88 UttarPradesh 18.13 39.13 932.28 5.87 10.25 431.81 24.00 49.38 1364.09 WestBengal 9.02 19.61 584.78 1.99 4.28 146.63 11.01 23.89 731.42 All India 100.09 210.97 5539.38 34.39 63.79 2761.58 134.47 274.76 8300.95 Note: AMS: Aggregate Monetary Shortfall (Rs million/month). AMS is the aggregate income necessary to bring the income of all the poor to the level of poverty line. Source: Computed by the authors. 35 Table 4.3 Percentage Distribution of Very Poor, Poor and Aggregate Monetary Shortfal Rural Urban All Very Poor AMS Very Poor AMS Very Poor AM4S Poor (per Poor (per Poor (per month) month) month) Andhra Pradesh 4.10 4.82 3.73 9.43 8.97 8.84 5.46 5.79 5.43 Assam 2.11 2.82 2.54 0.40 0.62 0.37 1.67 2.31 1.82 Bihar 16.46 14.92 15.71 7.47 6.69 7.20 14.16 13.01 12.88 Gujarat 2.94 3.30 3.22 5.71 6.67 6.07 3.65 4.08 4.17 Haryana 0.68 0.97 0.81 0.37 1.06 0.48 0.60 0.99 0.70 Jammu& 0.39 0.67 0.52 0.17 0.38 0.19 0.34 0.60 0.41 Kashmir Karnataka 5.23 4.82 4.66 9.10 8.39 8.97 6.22 5.65 6.09 Kerala 2.57 3.02 3.41 4.28 3.64 4.49 3.01 3.16 3.77 MadhyaPradesh 10.50 10.22 9.53 7.78 7.96 8.13 9.81 9.70 9.06 Maharashtra 8.94 8.71 9.28 14.17 13.16 15.29 10.28 9.74 11.28 Orissa 7.28 6.23 7.19 2.31 2.22 2.25 6.01 5.30 5.55 Punjab 0.47 0.94 0.65 0.34 1.01 0.45 0.44 0.96 0.59 Rajasthan 3.59 3.98 3.58 3.66 4.11 3.58 3.61 4.01 3.58 Tamil Nadu 7.61 6.74 7.76 11.95 12.36 12.75 8.72 8.05 9.42 Uttar Pradesh 18.12 18.55 16.83 17.08 16.06 15.64 17.85 17.97 16.43 WestBengal 9.01 9.30 10.56 5.79 6.71 5.31 8.19 8.70 8.81 All hidia 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 Source: Computed by the authors. The IGIDR-tabulated data provide information on the subsidies accruing to various quintiles of households formed on the basis of household per capita expenditure. Though IGIDR data also contains per capita total consumer expenditure by quintiles, they are found to be unreliable. In the survey on availability and utilization of PDS, information on per capita consumer expenditure of a household was collected by posing a one-shot question to the respondents. The information on PCE collected in the NSSO survey on PDS can be used only as a classificatory variable. The NSSO collected data on consumer expenditure during 1986-87 (Schedule No. 1.0 of NSSO). We have used it for estimating the PCE of the quintile. IGIDR data and NSSO 36 consumer expenditure survey data have been matched by adopting the following procedure. We have estimated the Lorenz curve using the household distribution as given in the consumer expenditure survey, viz., L(h) = h-ahb (1-h)'. From this estimated Lorenz curve, we have estimated L(hi), the share of the i'h quintile by household in total expenditure. Using the NSS forty-second round data, we have estimated the Lorenz curve based on distribution of persons as specified in (1) and obtained pi, the proportion of population in the ih quintile, by solving L(hi) = L(pi). The quintile-wise PCEs are estimated by using the equation PCE=L(pi)m/pi. The data set for all-India has been obtained by pooling the 20 quintile-wise data of the 16 major States. The 320 observations, so obtained, were collapsed into 20 classes after assigning weights in proportion to the population of the respective States. In the NSSO data, rationed quantities are expressed at subsidized prices and hence per capita expenditure and poverty line would also reflect the subsidized prices. The expenditure of ith person can be represented as Yi(pM,pR) and poverty line Z(pM,pR). Addition of income gains from all subsidies accruing to ith person, AYi to Yi(PM,PR) would give the per capita expenditures at market prices, Y;(pM). The poverty line at market prices Z(pM) can be estimated from the distribution of L(./Y(pM)) so that the head count ratio would be the same as that estimated from the distribution of L(./Y(pM,PR)) using the poverty line Z(PM,PR). The poverty measures estimated from L(-/Y(pM)) using poverty line Z(pM) would form the base scenario. The changes in the poverty measures when the income gain due to a subsidized commodity is subtracted from Yi(pM), give the welfare effects of the subsidy on the commodity (see Appendix Tables for detailed results). The calorie intake in 1986-87 has been obtained as a sum of calories derived from cereal consumption and calories derived from food items other than cereals. The quantities of cereals consumed (as reported in the Report on Consumer Expenditure Survey, i.e. Report No.351/I and 352/2) have been used to estimate the calories derived from cereals. The calories derived from non-cereal food items is obtained as a product of expenditure on non-cereal food items and calorie contained in one-rupee expenditure on non-cereal food items. The estimates of calorie content in one-rupee expenditure on non-cereal food items are taken from the work of Radhakrishna and Ravi (1990). 37 The increase in calorie intake Ac due to subsidy transfer is obtained from Ac/c=j Ay/y where 11 is calorie elasticity and Ay is income gain. The class-specific estimates of calorie elasticity obtained by Radhakrishna and Ravi (1990) have been used in the present study. It is also assumed that the all-India estimates of elasticity hold good for the States as well. Income Transfers to the Poor: 1986-87 The per capita income gain from all subsidies in 1986-87 is estimated to be Rs 2.30 per month (1.6 per cent of PCE) in rural areas of India, and Rs 3.68 per month (1.7 per cent of PCE) in urban areas (Tables 4.6 and 4.7). The gain was slightly lower for the poor: Rs 2.01 (2.7 per cent of their PCE) in rural areas and Rs 3.40 (3.2 per cent of PCE) in urban areas. Although the per capita income transfer to the poor was less in absolute terms, it was higher when expressed as a percentage of PCE. Income transfer via food subsidy contributed to an insignificant decrease in inequality: Gini coefficient declined by 0.004 in rural areas and 0.005 in urban areas (Table 4.9). Subsidies on rice, sugar and clothing together accounted for 75 per cent of income transfer to all classes in rural areas, and 58 per cent in urban areas. In the case of the poor, the contribution to the income transfer was more: 84 per cent in rural areas, and 70 per cent in urban areas. Despite substantial budgetary support to wheat subsidy, its contribution to income transfer was meager. This is understandable, given the small differences between the ration and open market prices of wheat; ration price was less than the open market price by 9-11 per cent only. Consequently, wheat purchases from PDS were negligible. By contrast, in the case of rice, the price difference between the ration and open market prices was large (ration price was lower than open market price by 30-37 per cent) and, therefore, sizable purchases from PDS contributed to the higher level of income gain due to rice subsidy. In the case of sugar, edible oil and clothing, although the price differences were more advantageous than on rice, low releases from PDS resulted in lower levels of income transfers. The income transfer through consumer subsidies was biased against the States with a high incidence of poverty such as Bihar, Orissa and Uttar Pradesh. In these States the per capita income increase due to subsidies was Rs 0.40-0.48 per month in rural areas, and Rs 0.96-2.01 in urban areas. The income transfer was much higher in States with a moderate incidence of poverty (e.g. Kerala, Andhra Pradesh and Karnataka). The factors underlying the income gain differ among the States. In Kerala, the income gain was mainly due to rice subsidy; in Andhra Pradesh it was due to a combination of rice and clothing subsidies; in Karnataka, it was subsidized clothing in rural areas and subsidized (diverse items of) food in urban areas; and in Gujarat it was different types of subsidized food. 38 Table 4.4 Per Capita Purchases of Rice, Wheat and Sugar from PDS in 1986-87 Kg/month Rural Urban State Rice Wheat Sugar Edible Oil Rice Wheat Sugar Edible Oil AndhraPradesh 2.51 0.00 0.25 0.09 2.13 0.30 0.46 0.15 Assam 0.67 0.03 0.26 0.00 1.26 0.00 0.32 0.00 Bihar 0.01 0.04 0.13 0.00 0.01 0.25 0.37 0.00 Gujarat 0.49 0.63 0.48 0.21 0.32 0.58 0.51 0.23 Haryana 0.01 0.00 0.35 0.00 0.08 0.00 0.36 0.00 Jammu & Kashmnir 1.64 0.25 0.32 0.00 6.24 0.55 0.42 0.02 Karnataka 0.75 0.22 0.21 0.02 1.29 0.39 0.50 0.11 Kerala 4.07 0.39 0.43 0.02 3.68 0.56 0.43 0.04 Madhya Pradesh 0.12 0.24 0.24 0.00 0.25 0.27 0.44 0.00 Maharashtra 0.39 0.67 0.32 0.10 0.67 0.81 0.39 0.15 Orissa 0.03 0.02 0.10 0.00 0.03 0.31 0.45 0.00 Punjab 0.00 0.01 0.31 0.00 0.04 0.00 0.29 0.00 Rajasthan 0.01 0.82 0.31 0.00 0.03 0.22 0.30 0.00 TamilNadu 1.18 0.11 0.23 0.02 0.93 0.25 0.40 0.09 Uttar Pradesh 0.12 0.10 0.12 0.00 0.14 0.17 0.54 0.00 West Bengal 0.50 0.44 0.28 0.02 1.53 1.78 0.34 0.08 All India 0.62 0.24 0.23 0.03 0.84 0.48 0.43 0.08 Source: Computed from NSS, 1986-87. 39 Table 4.5 PDS and Open Market Prices of Rice, Wheat, Sugar and Edible Oils in 1986-87 Rs/kg. Rural Urban Rice Wheat Sugar Edible Oils Rice Wheat Sugar Edible Oil PPPDS) P(OM)) P(PDS) P(A'I P(PDS) P(OM) P(PDS) P(OM) P(PDS) P(OM) P(PDS) P(OM) P(PDS) P(OM) P(PDS) P(OI) Andhra Pradesh 2.03 3.15 2.27 2.97 4.96 6.16 12.07 21.24 2.15 3.51 2.19 3.16 4.71 5.99 13.16 20.25 Assam 2.35 3.54 0.73 2.93 5.01 6.84 19.05 3.83 - 2.97 4.57 7.10 18.51 Biar 2.97 3.58 1.85 2.38 4.92 6.76 19.87 2.90 3.85 2.09 2.51 4.69 6.31 12.00 21.62 Gujarat 2.45 4.05 1.79 2.53 3.45 6.39 11.78 20.77 2.62 4.64 1.98 2.62 3.39 6.45 11.78 20.32 Haryana 3.14 4.20 - 2.04 4.33 6.56 20.72 3.28 4.19 - 2.02 4.58 6.22 22.58 Janmnu & Kashmir 2.61 3.41 2.09 2.21 4.38 7.31 20.82 2.50 3.63 2.05 2.55 4.69 7.06 15.20 20.76 Karnataka 2.21 3.67 1.79 2.81 4.74 6.37 16.03 22.06 2.67 4.14 2.01 3.17 4.67 6.28 12.57 21.40 Kerala 2.74 4.16 2.20 3.00 4.63 6.20 18.97 27.65 2.77 4.13 2.18 3.18 4.32 6.29 15.35 27.29 Madhya Pradesh 2.40 3.12 2.18 2.24 4.01 7.68 18.96 2.95 3.64 2.21 2.37 4.06 6.53 20.33 20.80 Maharashtra 2.99 3.38 2.20 2.80 4.79 6.35 11.82 21.29 3.02 4.24 2.20 3.07 4.13 6.30 13.20 22.23 Orissa 2.74 3.07 2.25 2.32 4.81 6.71 11.56 21.26 3.26 3.27 2.25 2.61 4.39 6.79 21.10 Punjab - 3.61 2.00 1.92 3.87 6.80 21.57 2.95 4.18 - 2.03 4.75 6.81 20.33 Rajasthan 2.87 4.34 1.77 2.20 4.38 6.45 20.66 2.88 4.81 2.11 2.28 4.66 6.49 21.65 TamilNadu 2.35 3.83 2.15 2.84 4.38 5.96 12.74 21.78 2.39 4.11 2.08 3.24 4.01 6.03 12.02 21.33 Uttar Pradesh 2.95 3.04 2.14 2.00 4.69 6.44 19.92 2.78 3.69 2.06 4.74 6.53 20.99 West Bengal 2.71 3.57 2.08 2.52 4.48 6.15 12.87 20.43 2.92 4.04 2.23 2.50 4.55 6.49 11.33 21.22 All India 2.42 3.45 2.02 2.21 4.48 6.56 12.65 20.78 2.62 3.91 2.17 2.41 4.39 6.40 12.74 21.26 Note: P(PDS): Price in Public Distribution System P(OM1): Price in Open Nlarket Source: Statistical Abstracts of States, Food Corporation of India. Table 4.6 Income Transfers to Poor, Non-Poor and AU Income Classes PDS in 1986-87: Rural (Rs 0.00 per capita per month) Income Gain Due to PDS supply of State Income Rice Wheat Cereals Pulses Edible Sugar Food Kerosene Cloth Non- All Group Oils food AndhmaPradesh Poor 2.70 0.00 2.70 0.00 0.57 0.18 3.46 0.18 2.03 2.21 5.67 Non-Poor 2.84 0.00 2.84 0.00 0.90 0.33 4.07 0.25 1.81 2.06 6.13 All 2.81 0.00 2.81 0.00 0.82 0.30 3.93 0.24 1.86 2.09 6.02 Assam Poor 0.91 0.11 1.07 0.04 0.00 0.41 1.52 0.04 0.29 0.33 1.84 Non-Poor 0.71 0.04 0.78 0.07 0.00 0.52 1.36 0.06 0.38 0.46 1.82 All 0.79 0.07 0.90 0.06 0.00 0.47 1.42 0.05 0.35 0.41 1.83 Bihar Poor 0.01 0.02 0.03 0.00 0.00 0.19 0.22 0.09 0.08 0.17 0.39 Non-Poor 0.02 0.03 0.05 0.00 0.00 0.27 0.32 0.12 0.06 0.18 0.50 All 0.01 0.02 0.04 0.00 0.00 0.23 0.27 0.11 0.07 0.18 0.45 Gjarat Poor 0.81 0.53 1.50 0.00 1.80 1.32 4.62 0.02 0.27 0.29 4.92 Non-Poor 0.78 0.44 1.28 0.00 1.97 1.43 4.69 0.05 0.05 0.10 4.79 All 0.79 0.46 1.35 0.00 1.92 1.40 4.67 0.04 0.12 0.16 4.82 Haryana Poor 0.03 0.00 0.03 0.00 0.00 0.78 0.81 0.27 0.00 0.27 1.08 Non-Poor 0.02 0.00 0.02 0.00 0.03 0.77 0.81 0.25 0.04 0.30 1.11 All 0.02 0.00 0.02 0.00 0.02 0.77 0.81 0.26 0.04 0.29 1.11 Jammu & Poor 1.65 0.03 1.76 0.00 0.03 0.76 2.55 0.02 0.16 0.18 2.74 Kashmir Non-Poor 1.17 0.03 1.25 0.00 0.04 1.01 2.30 0.05 0.04 0.09 2.39 All 1.30 0.03 1.39 0.00 0.04 0.94 2.37 0.04 0.07 0.12 2.48 Kamnataka Poor 0.98 0.19 1.21 0.00 0.08 0.25 1.54 0.05 4.31 4.36 5.90 Non-Poor 1.17 0.24 1.47 0.00 0.21 0.39 2.07 0.09 3.96 4.05 6.12 All 1.10 0.22 1.37 0.00 0.16 0.34 1.88 0.07 4.09 4.16 6.04 Kerala Poor 5.85 0.15 6.00 0.00 0.15 0.62 6.77 0.69 0.85 1.54 8.31 Non-Poor 5.72 0.32 6.03 0.00 0.27 0.70 7.00 0.88 0.66 1.53 8.53 All 5.76 0.26 6.02 0.00 0.23 0.67 6.92 0.82 0.72 1.54 8.46 Madhya Pradesh Poor 0.09 0.00 0.09 0.00 0.02 0.77 0.88 0.06 0.50 0.56 1.44 Non-Poor 0.09 0.02 0.11 0.00 0.04 0.98 1.13 0.07 0.58 0.67 1.80 All 0.09 0.01 0.10 0.00 0.03 0.88 1.01 0.06 0.54 0.61 1.62 Maharashtra Poor 0.13 0.37 0.52 0.00 0.95 0.46 1.93 0.07 0.49 0.56 2.49 Non-Poor 0.18 0.42 0.61 0.00 1.03 0.53 2.17 0.10 0.03 0.13 2.30 All 0.16 0.40 0.58 0.00 0.99 0.50 2.07 0.08 0.23 0.32 2.38 Orissa Poor 0.00 0.00 0.00 0.00 0.05 0.13 0.19 0.03 0.22 0.25 0.44 Non-Poor 0.03 0.00 0.03 0.00 0.00 0.24 0.27 0.03 0.22 0.25 0.52 All 0.01 0.00 0.02 0.00 0.03 0.18 0.23 0.03 0.22 0.25 0.48 Punjab Poor 0.00 0.00 0.00 0.00 0.00 0.66 0.66 0.11 0.11 0.22 0.88 Non-Poor 0.00 0.00 0.00 0.00 0.04 0.96 1.00 0.18 0.29 0.47 1.46 All 0.00 0.00 0.00 0.00 0.03 0.91 0.94 0.17 0.26 0.43 1.37 Rajasthan Poor 0.03 0.51 0.54 0.00 0.00 0.59 1.12 0.04 0.22 0.26 1.38 Non-Poor 0.03 0.28 0.31 0.00 0.02 0.66 0.99 0.09 0.29 0.37 1.37 All 0.03 0.35 0.37 0.00 0.01 0.64 1.03 0.07 0.27 0.34 1.37 TamilNadu Poor 1.71 0.04 1.75 0.00 0.14 0.23 2.12 0.26 0.00 0.26 2.38 Non-Poor 1.76 0.07 1.84 0.00 0.34 0.46 2.63 0.41 0.00 0.41 3.04 All 1.74 0.06 1.80 0.00 0.26 0.36 2.42 0.35 0.00 0.35 2.77 UttarPradesh Poor 0.00 0.00 0.00 0.00 0.00 0.13 0.13 0.09 0.01 0.11 0.24 Non-Poor 0.01 0.00 0.01 0.00 0.00 0.27 0.28 0.17 0.03 0.21 0.49 All 0.00 0.00 0.00 0.00 0.00 0.22 0.23 0.14 0.02 0.17 0.40 WestBengal Poor 0.47 0.17 0.65 0.00 0.10 0.39 1.13 0.39 0.50 0.89 2.02 Non-Poor 0.40 0.22 0.62 0.00 0.18 0.52 1.33 0.54 0.49 1.05 2.37 All 0.43 0.20 0.63 0.00 0.15 0.46 1.24 0.47 0.49 0.98 2.22 All India Poor 0.62 0.10 0.72 0.00 0.21 0.35 1.29 0.12 0.60 0.73 2.01 Non-Poor 0.79 0.12 0.91 0.00 0.31 0.52 1.74 0.22 0.51 0.72 2.47 All 0.73 0.11 0.84 0.00 0.27 0.46 1.58 0.18 0.54 0.73 2.30 Source: Computed by the authors. 41 Table 4.7 Income Transfers to Poor, Non-poor and All Income Classes PDS in 1986-87: Urban (Rs 0.00 per capita per month) Income Gain Due to PDS Supply of Income Rice Wheat Cereals Pulses Edible Sugar Food Kero Cloth Non- All Group Oils -sene food Andhra Pradesh Poor 3.50 0.14 3.65 0.00 0.78 0.38 4.82 0.33 2.70 3.03 7.85 Non Poor 2.47 0.41 2.87 0.00 1.31 0.74 4.93 0.52 0.57 1.09 6.02 All 2.91 0.29 3.20 0.00 1.09 0.59 4.88 0.44 1.48 1.92 6.80 Assam Poor 1.32 0.00 1.34 0.08 0.04 0.74 2.20 0.04 0.00 0.05 2.25 Non Poor 1.19 0.00 1.20 0.06 0.00 0.84 2.10 0.06 0.00 0.06 2.15 All 1.22 0.00 1.23 0.06 0.01 0.82 2.12 0.05 0.00 0.06 2.18 Bihar Poor 0.00 0.05 0.05 0.00 0.00 0.47 0.52 0.16 0.02 0.18 0.70 NonPoor 0.02 0.15 0.17 0.00 0.02 0.73 0.92 0.19 0.06 0.25 1.17 All 0.01 0.10 0.12 0.00 0.01 0.61 0.74 0.18 0.04 0.22 0.96 Gujarat Poor 0.85 0.48 1.34 0.00 2.14 1.68 5.17 0.02 0.00 0.02 5.18 Non Poor 0.49 0.29 0.79 0.00 1.80 1.46 4.05 0.02 0.04 0.07 4.11 All 0.65 0.37 1.03 0.00 1.95 1.56 4.53 0.02 0.02 0.04 4.58 Haryana Poor 0.08 0.00 0.08 0.00 0.00 0.74 0.82 0.14 0.00 0.14 0.96 Non Poor 0.07 0.00 0.07 0.00 0.00 0.56 0.63 0.21 0.66 0.87 1.50 All 0.07 0.00 0.07 0.00 0.00 0.59 0.67 0.19 0.53 0.73 1.39 Jammu Poor 5.25 0.33 5.77 0.00 0.11 0.74 6.61 0.09 0.54 0.64 7.25 & Kashmir NonPoor 7.48 0.27 7.93 0.00 0.13 1.07 9.12 0.23 0.07 0.29 9.42 All 7.02 0.28 7.48 0.00 0.12 1.00 8.61 0.20 0.16 0.36 8.97 Karnataka Poor 1.41 0.38 1.81 0.00 0.73 0.60 3.14 0.14 0.05 0.20 3.34 Non Poor 2.29 0.52 2.81 0.00 1.12 0.97 4.90 0.24 0.18 0.47 5.37 All 1.89 0.46 2.36 0.00 0.94 0.81 4.11 0.19 0.12 0.35 4.46 Kerala Poor 5.41 0.46 5.86 0.00 0.45 0.84 7.15 0.54 1.75 2.29 9.44 Non Poor 4.65 0.55 5.20 0.00 0.65 0.87 6.72 0.90 1.32 2.22 8.94 All 5.03 0.50 5.53 0.00 0.55 0.85 6.93 0.72 1.53 2.26 9.19 Madhya Poor 0.18 0.04 0.22 0.00 0.10 0.96 1.28 0.04 0.45 0.50 1.78 Pradesh Non Poor 0.17 0.04 0.22 0.00 0.03 1.21 1.46 0.05 0.46 0.63 2.09 All 0.17 0.04 0.22 0.00 0.07 1.09 1.37 0.05 0.45 0.57 1.94 Maharashtra Poor 0.66 0.78 1.45 0.00 0.98 0.75 3.19 0.39 0.00 0.39 3.57 Non Poor 0.91 0.65 1.56 0.00 1.60 0.90 4.07 0.85 0.01 0.86 4.92 All 0.81 0.70 1.52 0.00 1.36 0.84 3.72 0.67 0.00 0.67 4.39 Orissa Poor 0.00 0.08 0.08 0.00 0.00 0.79 0.87 0.07 0.68 0.78 1.65 Non Poor 0.00 0.20 0.20 0.00 0.05 1.33 1.58 0.31 0.25 0.75 2.33 All 0.00 0.14 0.14 0.00 0.03 1.07 1.24 0.20 0.45 0.77 2.01 Punjab Poor 0.07 0.00 0.07 0.00 0.00 0.49 0.56 0.41 0.00 0.41 0.97 Non Poor 0.05 0.00 0.05 0.00 0.00 0.62 0.68 0.59 0.06 0.67 1.34 All 0.05 0.00 0.05 0.00 0.00 0.60 0.66 0.57 0.05 0.64 1.30 Rajasthan Poor 0.02 0.04 0.06 0.00 0.00 0.58 0.64 0.05 0.17 0.22 0.85 NonPoor 0.10 0.03 0.12 0.00 0.01 0.54 0.67 0.08 0.11 0.20 0.87 TamilNadu Poor 1.80 0.19 1.99 0.00 0.61 0.63 3.23 0.68 0.00 0.68 3.91 Non Poor 1.44 0.36 1.82 0.00 0.95 0.95 3.71 1.07 0.00 1.07 4.79 All 1.60 0.28 1.89 0.00 0.80 0.81 3.50 0.90 0.00 0.90 4.40 UttarPradesh Poor 0.09 0.00 0.09 0.00 0.00 0.80 0.89 0.30 0.11 0.42 1.31 NonPoor 0.16 0.00 0.16 0.00 0.00 1.12 1.29 0.48 0.24 0.72 2.01 All 0.13 0.00 0.13 0.00 0.00 0.98 1.11 0.40 0.18 0.58 1.69 West Bengal Poor 1.84 0.38 2.22 0.02 0.69 0.56 3.49 0.36 1.13 1.56 5.05 NonPoor 1.65 0.53 2.18 0.03 0.85 0.71 3.76 0.50 0.50 1.14 4.90 All 1.71 0.48 2.19 0.03 0.80 0.66 3.68 0.46 0.69 1.27 4.95 All India Poor 1.17 0.23 1.40 0.00 0.50 0.72 2.62 0.28 0.49 0.78 3.40 NonPoor 1.10 0.32 1.42 0.00 0.76 0.92 3.10 0.49 0.23 0.75 3.86 All 1.13 0.28 1.41 0.00 0.66 0.84 2.91 0.4' 0.33 0.76 3.68 Source: Computed by the authors. 42 Income gains were higher in states where access to PDS was relatively universal. For instance, in Kerala, the per capita monthly income gain was Rs 8.31 for the poor and Rs 8.53 for the non-poor in rural areas and Rs 9.44 and Rs 8.94, respectively in urban areas; in Andhra Pradesh, the corresponding figures for the poor and the non-poor were Rs 5.67 and Rs 6.13 in rural areas, and Rs 7.85 and Rs 6.02 in urban areas. The poor did not receive a more than proportionate absolute amount of income transfer as compared to the nonpoor. The percentage gain was more for the poor because of their lower per capita expenditure. Hence, the reduction in inequalities in per capita total expenditure (as measured by the Gini coefficient) was higher in these states than for all-India (Table 4.9). The upshot is that the gains in terms of income transfer from PDS were negligible to the poor for the country as a whole largely because of negligible impact in states with a high incidence of poverty. The per capita monthly income transfer was Rs 1.84 to the very poor and Rs 2.17 to the moderately poor in rural areas of India and Rs 3.27 and Rs 3.60, respectively in urban areas. In Andhra Pradesh and Kerala, the poor did receive substantive income transfer, but so did the nonpoor; and the impact, therefore, is regressive. PDS Impact on Poverty Considering India as a whole, the impact of all consumer subsidies on poverty was very moderate; subsidies reduced poverty (as measured by the Head Count Ratio) by 1.66 percentage points in rural areas and 1.71 percentage points in urban areas (Table 4.8). About 12.1 million persons (9.4 million in rural and 2.7 million in urban areas) may have moved out of poverty in 1986-87 due to income transfers from PDS (Table 4.10). The impact of cereal subsidies was much less: no more than 4.6 million may have moved out of poverty due to cereal subsidies (Table 4.10). Given the large size of the absolute number of poor (274 million in 1986-87), these numbers are small.35 = 351t is worth noting that our analysis is based on the assumption that the market price in the absence of PDS intervention was equal to the market price witi PDS intervention. If the market price in the absence of PDS were to be lower than the actual market price with PDS (as argued by Hayami, Subbarao, Otsuka) the income transfer gains as well as poverty impacts would have been even smaller. Thus, our estimates of income transfer gains and poverty impacts must be considered as upper limits, i.e. the gains could not have been higher. 43 Table 4.8 Decline in Poverty Due to Subsidies: 1986-87 Rural Urban State Poverty Measure Cereals Food All Cereals Food All Andhra Pradesh Head Coumt 2.21 2.86 4.64 1.63 2.28 3.24 Poverty Gap 0.73 0.95 1.64 1.03 1.36 2.18 FGT 0.30 0.39 0.71 0.57 0.75 1.26 Assam Head Counit 0.77 1.24 1.50 0.63 1.01 .1B03 Poverty Gap 0.36 0.52 0.63 0.26 0.41 0.42 FGT 0.15 0.21 0.26 0.10 0.16 0.17 Bihar Head Count 0.04 0.21 0.30 0.04 0.29 0.37 PovertyGap 0.02 0.11 0.18 0.02 0.16 0.22 FGT 0.01 0.05 0.10 0.01 0.10 0.14 Gujarat Head Count 1.18 3.74 3.85 0.68 2.65 2.67 Poverty Gap 0.43 1.36 1.43 0.37 1.43 1.43 FGT 0.19 0.60 0.64 0.17 0.69 0.69 Haryana Head Count 0.01 0.45 0.56 0.04 0.44 0.64 PovertyGap 0.00 0.15 0.19 0.01 0.11 0.12 FGT 0.00 0.05 0.07 0.00 0.03 0.03 Jammu & Kashmir Head Count 1.28 1.98 2.06 3.59 4.07 4.31 Poverty Gap 0.39 0.59 0.63 0.98 1.13 1.19 FGT 0.13 0.19 0.21 0.29 0.34 0.36 Karnataka Head Count 0.95 1.26 4.33 1.01 1.75 1.88 Poverty Gap 0.45 0.58 2.20 0.53 0.94 1.01 FGT 0.22 0.28 1.14 0.28 0.52 0.57 Kerala Head Count 4.08 4.58 5.49 2.20 2.68 3.62 Poverty Gap 1.55 1.76 2.16 1.69 2.07 2.75 FGT 0.67 0.77 0.96 1.06 1.31 1.73 Madhya Pradesh Head Count 0.08 0.91 1.43 0.12 0.66 0.92 Poverty Gap 0.04 0.45 0.74 0.06 0.38 0.54 FGT 0.02 0.21 0.36 0.03 0.21 0.30 Maharashtra Head Count 0.46 1.66 2.05 0.67 1.50 1.75 Poverty,Gap 0.22 0.79 0.98 0.34 0.76 0.86 FGT 0.11 0.39 0.48 0.18 0.40 0.45 Orissa Head Count 0.01 0.20 0.30 0.06 0.51 0.69 Poverty Gap 0.00 0.09 0.21 0.02 0.26 0.50 FGT 0.00 0.04 0.13 0.01 0.14 0.30 Punjab Head Count 0.00 0.42 0.55 0.02 0.23 0.44 Poverty Gap 0.00 0.10 0.13 0.01 0.06 0.10 FGT 0.00 0.03 0.04 0.00 0.02 0.03 Rajasthan Head Count 0.30 0.73 0.91 0.02 0.28 0.40 PovertyGap 0.15 0.32 0.40 0.01 0.16 0.21 FGT 0.07 0.14 0.18 0.01 0.08 0.10 TanilNadu Head Count 1.25 1.61 1.81 0.97 1.71 2.14 Poverty Gap 0.68 0.82 0.93 0.55 0.90 1.10 FGT 0.37 0.44 0.49 0.30 0.47 0.57 UttarPradesh HeadCount 0.00 0.13 0.22 0.07 0.54 0.82 Poverty Gap 0.00 0.05 0.08 0.03 0.29 0.44 FGT 0.00 0.02 0.04 0.02 0.16 0.24 West Bengal Head Count 0.50 0.98 1.67 1.06 1.71 2.36 Poverty Gap 0.24 0.43 0.74 0.49 0.79 1.11 FGT 0.12 0.20 0.36 0.22 0.36 0.52 AllIndia HeadCount 0.61 1.12 1.66 0.69 1.36 1.71 Poverty Gap 0.25 0.46 0.72 0.37 0.70 0.90 FGT 0.12 0.21 0.34 0.19 0.36 0.47 Note: The above figures show decline in poverty measures multiplied by 100. Source: Computed by the authors. 44 Table 4.9 Effects of Subsidies on Gini Co-efficient in 1986-87 Decline in Gini Co-efficient due to subsidy on Rural Urban Rice Food All Rice Food All Andhra Pradesh 0.006 0.007 0.013 0.007 0.008 0.013 Assam 0.002 0.002 0.003 0.002 0.003 0.003 Bihar 0.000 0.000 0.001 0.000 0.001 0.001 Gujarat 0.003 0.009 0.009 0.003 0.009 0.009 Haryana 0.000 0.001 0.002 0.000 0.001 0.001 Jammu & Kashmir 0.003 0.004 0.004 0.007 0.008 0.009 Karnataka 0.002 0.003 0.012 0.002 0.004 0.004 Kerala 0.011 0.012 0.014 0.010 0.012 0.016 Madhya Pradesh 0.000 0.002 0.003 0.000 0.002 0.003 Maharashtra 0.001 0.004 0.005 0.002 0.004 0.005 Onissa 0.000 0.000 0.001 0.000 0.001 0.002 Punjab 0.000 0.001 0.001 0.000 0.001 0.001 Rajasthan 0.001 0.002 0.002 0.000 0.001 0.001 Tamil Nadu 0.004 0.004 0.004 0.003 0.005 0.006 Uttar Pradesh 0.000 0.000 0.000 0.000 0.001 0.002 West Bengal 0.001 0.002 0.003 0.004 0.006 0.008 All India 0.001 0.002 0.004 0.002 0.004 0.005 Source: Computed by the authors. Table 4.10 Number of Persons Moved out of Poverty Due to Subsidies in 1986-87 Millions Cereal Subsidies All Subsidies Rural Urban All India Rural Urban All India Andhra Pradesh 0.998 0.219 1.218 2.094 0.436 2.529 Assam 0.111 0.011 0.122 0.218 0.017 0.235 Bihar 0.024 0.004 0.028 0.195 0.035 0.230 Gujarat 0.274 0.067 0.342 0.895 0.263 1.159 Haryana 0.002 0.002 0.003 0.065 0.023 0.087 Jammu & Kashmir 0.067 0.042 0.109 0.107 0.050 0.158 Karnataka 0.265 0.120 0.385 1.200 0.223 1.424 Kerala 0.821 0.103 0.924 1.104 0.169 1.273 Madhya Pradesh 0.037 0.013 0.051 0.636 0.100 0.736 Maharashtra 0.195 0.144 0.339 0.873 0.375 1.247 Orissa 0.003 0.002 0.005 0.075 0.020 0.096 Punjab 0.000 0.001 0.001 0.069 0.022 0.091 Rajasthan 0.089 0.002 0.091 0.270 0.029 0.299 Tamil Nadu 0.424 0.171 0.595 0.617 0.379 0.996 Uttar Pradesh 0.001 0.015 0.016 0.242 0.184 0.426 WestBengal 0.222 0.149 0.371 0.737 0.331 1.068 All India 3.532 1.065 4.597 9.396 2.657 12.052 Source: Computed by the autlhors. 45 The impact on both the extent (Head Count Ratio) and severity (Poverty Gap Rates and FRG) of poverty, was better in the States which had higher per capita income transfers from PDS. In Kerala, a decline of 5.5 in the percentage of the poor in rural areas and 3.6 in the percentage of the poor in urban areas can be attributed to the income transfers from all consumer subsidies. In Andhra Pradesh a decline of 4.6 percent in rural areas and 3.21 percent in urban areas was due to income transfers from PDS; a moderate decline could be seen in Tamil Nadu, Karnataka, Maharashtra, Gujarat, West Bengal and Jammu & Kashmir. In the rest of the States, the impact of subsidies on poverty was more or less insignificant. The minimal impact of PDS on the poverty levels in Bihar, Orissa and Uttar Pradesh needs to be noted; the decline in poverty in these States was less than 0.3 percentage points in rural areas and 0.8 percentage points in urban areas. As expected, the share of these three States among the poor who moved out of poverty due to PDS was 6.2 per cent (Table 4. 11) whereas they accounted for one-third of the poor in India (Table 4.3). By contrast, the share of Kerala and Andhra Pradesh in the poor who moved out of poverty was 32 per cent (Table 4.11) whereas they together accounted for less than 10 per cent of the poor in India (Table 4.3). Table 4.11 Percentage Distribution of Persons Moved Out of Poverty Due to Subsidies in 1986-87 Cereal Subsidies All Subsidies Rural Urban All lndia Rural Urban All India Andlira Pradesh 28.26 20.59 26.49 22.28 16.40 20.99 Assam 3.14 1.01 2.65 2.32 0.66 1.95 Bihar 0.67 0.35 0.60 2.07 1.31 1.90 Gujarat 7.77 6.31 7.43 9.53 9.91 9.61 Haryana 0.04 0.14 0.07 0.69 0.85 0.72 Jammu & Kashmir 1.89 3.94 2.36 1.14 1.90 1.31 Karnataka 7.49 11.32 8.37 12.77 8.41 11.81 Kerala 23.24 9.65 20.09 11.75 6.36 10.56 Madhya Pradesh 1.06 1.27 1.11 6.76 3.77 6.11 Miaharashtra 5.51 13.57 7.37 9.29 14.11 10.35 Orissa 0.08 0.17 0.10 0.80 0.77 0.79 Punjab 0.00 0.08 0.02 0.73 0.83 0.75 Rajasthan 2.53 0.15 1.98 2.87 1.10 2.48 Tamil Nadu 12.00 16.08 12.95 6.57 14.25 8.26 Uttar Pradesh 0.02 1.39 0.34 2.58 6.92 3.54 West Bengal 6.30 13.98 8.08 7.84 12.47 8.86 All India 100.00 100.00 100.00 100.00 100.00 100.00 Source: Computed by the authors. PDS Impact on Nutrition The increase in calorie intake due to PDS is estimated to be 41 kcal/day for the very poor; 29 kcal/day for the moderately poor, and 11 kcaVday for the non-poor in rural areas of India. The 46 corresponding figures were 44, 27, and 12 kcalUday for urban areas of India. Although per capita income transfers to the very poor were the lowest, the increase in their calorie intake levels was larger because of higher calorie elasticity in both rural and urban areas. For a similar reason, the increases in calorie intake were higher in rural than in urban areas. The impact of PDS on the calorie intake of the very poor and moderately poor was significant in both rural and urban areas of Andhra Pradesh and Kerala (123-132 kcallday increase in rural areas, and 101-110 kcalday in urban areas for the very poor). The increase in calorie intake was also sizable in rural Karnataka and urban Jammu & Kashmir. But these increases in calorie intake could bridge the calorie gap only marginally even in Kerala and Andhra Pradesh. Taking into consideration a calorie norm of 2000 kcal, the calorie gap for the poor in rural Andhra Pradesh was 530 kcal/day and in rural Kerala 637 kcallday even after the gains due to PDS were taken into account. Clearly, the calorie gaps in these two States would have been much less had there been better targeting. It is worth observing from Table 4.12 that the calorie intake levels of the poor were lower in Andhra Pradesh, Kerala, Karnataka and Tamil Nadu, states with a better functioning PDS than in the poorer States of Bihar, Orissa and Uttar Pradesh. This could be attributed to the differences in the composition of their consumption baskets; the consumption baskets of the poor in the latter States contained proportionately more cereals which were cheap sources of calories. Cost of Income Transfer-All India The total income gain to the consumers from the subsidized distribution of wheat and rice in PDS, ITDP and JRY schemes in 1986-87 is estimated to be Rs 8474 million (Table 4.13). The central government expenditure on the cereal subsidy was Rs 12,114 million.36 Some of the state governments had also incurred expenditure on cereal subsidy. Hence, a part of the income gain could be due to the states' expenditures on cereal subsidy. Considering only the central government expenditure, it may be inferred that in 1986-87 the Central government spent Rs 1.43 (which includes the amount of transfer) to transfer one rupee to the consumer. (It is worth recalling that the 'consumers' included both the poor and the nonpoor since there was no real targeting). If the states' expenditure is also considered, the cost of transfer to the government (Central and State governments) would be higher. However, since only a few States were providing additional cereal subsidy, the cost of transfer might not be substantially higher than the above figure.37 The income transfer to the poor in 1986-87 was estimated to be Rs 2840 million and the central government expenditure required to transfer one rupee worked out to Rs 4.27. The high cost is largely due to leakages to nonpoor, since the share of the poor in the total income gain was only 34 percent, and due to systemic inefficiencies (leakages due to pilferage, etc.). 36 This amount is only for rice and wheat, and is somewhat lower than the figure shown in Table 3. 1, which shows the subsidy for all conmnodities. 37 The next section provides a case study of Andlira Pradesh: one of the few states providing substantial additional subsidy. 47 Table 4.12 Calorie Gains Due to PDS (All Commodities) in 1986-87 Kcal/day/person Rural Urban States VPoor MPoor All Poor NPoor All VPoor MPoor All Poor NPoor All Andhra Pradesh 132 84 104 27 40 110 64 82 19 28. (12.71) (5.32) (7.62) (1.25) (2.03) (7.92) (3.63) (5.27) (0.83) (1.43) Assam 29 20 24 7 12 38 15 22 7 9 (2.17) (1.17) (1.52) (0.34) (0.59) (2.74) (0.95) (1.43) (0.32) (0.41) Bihar 9 5 6 3 4 11 6 8 4 5 (0.60) (0.26) (0.40) (0.10) (0.18) (0.68) (0.30) (0.45) (0.17) (0.21) Gujarat 80 61 70 19 29 66 36 47 13 18 (7.35) (4.05) (5.25) (0.84) (1.47) (4.60) (1.97) (2.87) (0.56) (0.92) Haryana 29 13 19 4 7 11 9 10 6 6 ( 1.77) (0.67) (0.98) (0.16) (0.26) (0.64) (0.49) (0.58) (0.24) (0.27) Jammu & Kashmir 44 40 45 11 17 95 66 79 36 40 (2.46) (1.97) (2.29) (0.37) (0.65) (6.15) (3.58) (4.45) (1.41) (1.67) Karnataka 115 89 101 28 43 39 25 30 17 18 (9.67) (5.48) (7.20) (1.23) (2.22) (2.96) (1.56) (2.11) (0.74) (0.93) Kerala 123 78 96 25 38 101 76 85 25 34 (11.67) (5.66) (7.50) (1.17) (2.06) (8.64) (4.21) (6.03) (0.99) (1.71) Madhya Pradesh 34 23 28 9 13 28 13 19 7 9 (2.19) (1.18) (1.58) (0.34) (0.61) (1.64) (0.65) (1.01) (0.28) (0.40) Maharashtra 40 36 39 9 16 37 31 33 14 16 (3.06) (2.11) (2.53) (0.41) (0.81) (2.57) (1.69) (2.05) (0.59) (0.77) Orissa 10 7 8 3 4 27 13 19 9 10 (0.63) (0.34) (0.46) (0.10) (0.19) (1.51) (0.63) (0.97) (0.29) (0.39) Punjab 13 10 12 5 7 13 7 9 4 5 (0.99) (0.63) (0.76) (0.20) (0.30) (0.98) (0.45) (0.58) (0.18) (0.21) Rajasthan 36 20 26 7 10 14 7 9 3 4 (2.25) (1.00) (1.42) (0.23) (0.40) (0.79) (0.34) (0.49) (0.11) (0.16) Tamil Nadu 42 28 34 12 17 40 28 33 14 16 ( 3.47) (1.79) (2.47) (0.53) (0.92) (3.24) (1.68) (2.26) (0.61) (0.85) Uttar Pradesh 5 4 4 2 3 19 12 15 7 8 (0.31) (0.21) (0.25) (0.09) (0.13) (1.20) (0.64) (0.86) (0.29) (0.38) West Bengal 35 25 30 11 16 63 46 53 15 19 ( 2.45) (1.36) (1.78) (0.43) (0.74) (4.36) (2.61) (3.28) (0.63) (0.87) All India 41 29 35 11 16 16 9 12 4 5 ( 2.91) (1.65) (2.15) (0.45) (0.76) (1.10) (0.49) (0.71) (0.15) (0.22) Notes: Figures in parentheses are percentage calorie gain due to PDS. Calorie elasticity with respect to income for Very Poor, Moderately Poor, All Poor, Non-Poor and All Classes are 0.9151, 0.6449, 0.7688, 0.3207 and 0.4555, respectively in Rural areas and are 0.8103, 0.5339, 0.6459,0.3803 and 0.4161 respectively in Urban areas. Source: Computed by the authors. 48 Table 4.13 Cost to Central Government to Transfer One Rupee to Poor Through PDS Cereal Supply in India in 1986-87 Central Government Spending on rice and wheat subsidy Rs 12114 million Annual income gain to all consumers Per capita Rs 11.64 Aggregate Rs 8474 million Central Government expenditure required for 1 rupee to the Rs 1.43 consumers (all) Annual Income gain to the poor Per capita Rs 10.56 All poor Rs 2840 million Central Government expenditure required for 1 rupee gain to the poor Rs 4.27 Central Government expenditure required for I rupee Rs 1.43 gain to the under perfect targeting Notes: Central Government expenditure on cereal subsidy includes expenditure on the supply of subsidized wheat and rice in PDS, ITDP and JRY schemes. The States expenditure on subsidy could not be considered due to lack of data. The cost estimates under the perfect targeting were based on three assumptions: (i) leakages other than those due to mistargeting would remain, (ii) entire income gain would accrue to the poor, (iii) perfect targeting would not involve additional costs. Source: Computed by the authors. In 1986-87, the increase in the per capita daily calorie intake due to PDS was estimated to be 174 k.caVmonth for all consumers and 370 k.calVmonth for the poor (Table 4.14). It cost the Central Government Rs 1.01 to increase the calorie intake of the poor by 100 k.cal. Were all public spending to be targeted to the poor, the cost of transfer would have decreased from Rs 1.01 to Rs 0.33. The evidence suggests that potential benefits from PDS to the poor could not be realized cost-effectively due to weak targeting and leakages. The cost of income transfer was high mainly because the program was open-ended and never targeted. How did the program fare in Andhra Pradesh which incurred a heavy subsidy cost on its own additional initiatives? Cost of Income and Calorie Transfers in Andhra Pradesh A subsidized rice scheme known as 'Two-rupee-a-kilo' rice has been in existence in Andhra Pradesh since 1983. Although targeting is based on a means test, the scheme is covering 70 per cent of the State's population while the poor estimated to be about 30 per cent. The 49 covered households are supplied 5 kg. of rice per capita per month subject to a ceiling of 25 kg. per household at a retail price of Rs 2 per kg. of rice. About 2 million tonnes of rice are distributed through a network of retail outlets. In the reference year, 1986-87 official records show that 2.24 million tonnes of rice was distributed under the scheme. It involved a government subsidy of Rs 3490 million-Rs 1769 million State subsidy and Rs 1721 million Central subsidy (see Table 4.12). This had resulted in a per capita income gain of Rs 34 per annum and total income gain of Rs 1990 million per annum. Hence, the cost of one rupee increase in consumer income under the scheme would work out to Rs 1.75. The total amount of income transferred to the poor under the scheme was Rs 550 million (Table 4.15). Hence, under the scheme a public spending of Rs 6.35 (which includes the transfer itself) was involved to transfer one rupee to the poor. The high cost of rupee transfer was due to weak targeting. If the additional costs involved under perfect targeting are assumed away, it would have cost only Rs 1.75 to transfer a rupee to the poor, and the same amount of income transfer (Rs 550 million) could have been achieved with a public spending of Rs 962 million. Even if targeting costs are considered, this might not exceed Rs 1200 million. However, the probability of leaving out the very poor would be higher under perfect targeting due to its strict restrictions. In 1986-87, the increase in per capita daily calorie intake due to PDS was estimated to be 104 kcal for the poor, 22 kcal for the non-poor and 40 kcal for all classes in rural areas. The corresponding figures were 82, 19 and 28, respectively in urban areas. As can be seen from Table 4.16, Rs 2.82 per capita per month income gain due to PDS resulted in an additional consumption of 510 kcal per capita per month to all consumers (rural + urban) in Andhra Pradesh. This means that to increase calorie intake by 100 kcal, an additional income of Rs 0.55 was required in 1986- 87. It has already been shown that it would cost Rs 1.75 to transfer one rupee to all consumers. Hence, to increase calorie intake by 100 kcal, a public spending of Rs 0.97 was required in 1986- 87 in contrast with Rs 0.31 under ICDS. For the poor, the per capita calorie gain was 1350 kcal/month, and the per capita income gain Rs 2.97 per month. Since public spending of Rs 6.35 was required to transfer Re I because of weak targeting, it involved a public spending of Rs 1.40 to increase their calorie intake by 100 kcal. If all public spending were to be targeted to the poor, and if targeting were costless, the poor would have enjoyed an additional income of Rs 10.75 per month as compared to Rs 2.97 under weak targeting, and an additional 4590 kcal per month as compared to 1350 kcal. If fine targeting involved no additional costs, it would cost Rs 0.41 to increase the calorie intake of the poor by 100 kcal. 50 The distribution of costs under the centrally-administered (all-India) PDS scheme, and the AP rice scheme are shown in Figures IA and 1B. When the programs are "weakly targeted", the costs of leakage were 54 to 57 percent, and the costs of administering the programs were 26 to 27 percent. If more refined targeting were introduced (with vigorously administered meanstests) it is most likely that the cost of leakage would fall, but the costs of administration might increase. Overall, it seems unlikely that fine targeting would make much of a difference to total costs under the existing program of quantity rationing either for the Central program or for the AP rice scheme. Approaches other than quantity rationing, including self-targeting and other alternatives such as food stamps, need to be considered in order to deliver food transfers to the needy cost- effectively. (The last section reviews the options for reform.) We have so far analyzed the cost per Re 1 of income transfer to the poor, and noted that it is currently exorbitant particularly for the AP rice subsidy scheme. The total costs incurred by the state are also so high as to cause concern. The cost of the subsidized scheme is shared by the Central and State Governments. The cost to the Central Government depends on per unit subsidy (the difference between economic cost and issue price) and quantity of rice utilized from the Central pool. The cost to the State Government depends on the difference between the issue price plus distribution cost and the retail ration price, and the quantity distributed under the scheme. There has been a steep increase in the Central issue price due to upward revisions of procurement prices of rice and, consequently, the subsidy/cost to the State Government has increased manifold since it fixed the retail ration price of rice at Rs 2. According to the revised budget estimates of FCI, in 1995-96 the economic cost of rice was Rs 7.47 per kg. and average sales realization was Rs 6.11 per kg. which involved a Central subsidy of Rs 1.36 kg. The distribution cost of rice incurred by the State Government was Rs 0.28. Hence, distribution of rice at Rs 2 per kg. involved a subsidy of Rs 4.39 per kg. to the State Government. Since about 2.2 million tonnes were distributed under the scheme, it would cost the State Government Rs 10000 million (7.3 per cent of the State Government total expenditure). It is important to ask what would be the opportunity cost of this large amount. It appears that the subsidy was at the expense of investments in irrigation and power. This may have adverse long term effects. Clearly, the welfare gains to the poor from the scheme can be enhanced from the existing amount of public spending by fine targeting. Alternatively, resources can be saved for the other anti-poverty programs such as ICDS, MDM and EGS. 51 Administrative Cost and Cost of Leakage of Food Subsidies, India Figure 1-A. Centrally-administered Public Distribution System (Rupees) 1.43 El Income gains to the poor 2.84 ;U: . - l lE Cost of leakage for the nonpoor 0 Cost of administering the scheme Figure 1-B. Rs. 2/kilo Scheme Administered by Andhra Pradesh State Government (Rupees) 1.75 . _ Income gains to the poor 3.6 U | @S Cost of leakage for the 3.6 . nonpoor El Cost of administering the scheme Source: Computed by the authors. 52 Table 4.14 Cost to Central Government to Increase Calorie Intake through PDS Cereal Supply in 1986-87 All India All consumers Per Capita Calorie Increase Due to PDS Cereal Supplies 174 k.callmonth Per Capita Income Gain Due to PDS Cereal Supplies Rs 0.97 per month Central Government Expenditure Required to Transfer Rs 0.97 Rs 1.39 Cost of Transfer of 100 k.cal Rs 0.80 Poor Per Capita Calorie Increase Due to PDS Cereal Supplies 370 k.callmonth Per Capita Income Gain Due to PDS Cereal Supplies Rs 0.88 per month Central Government Expenditure Required to Transfer Rs 0.88 Rs 3.75 Cost of Transfer of 100 k.cal Rs 1.01 Under Perfect Targeting Per Capita Calorie Increase Due to PDS Cereal Supplies 1128 k.cal/month Per Capita Income Gain to the Poor Due to PDS Cereal Supplies Rs 2.62 per month Central Government Expenditure Required to Transfer Rs 2.62 to the Poor Rs 3.75 Cost of Transfer of 100 k.cal Rs 0.33 Notes: Central Government expenditure on cereal subsidy includes expenditure on the supply of subsidized wheat and rice in PDS, ITDP and JRY schemes. The States expenditure on subsidy could not be considered due to lack of data. The cost of estimates under perfect targeting were based on three assumptions: (i) leakages other than those due to mistargeting would remain, (ii) entire income gain would accrue to the poor (iii) perfect targeting would not involve additional costs. The difference between calorie intake under perfect targeting at the existing level of PDS cereal supplies and the calorie intake in the absence of PDS is taken as the increase due to PDS cereal supplies under perfect targeting. The increase in calorie intake has been estimated using (i) increase in per capita income due to PDS cereal supplies in rural and urban areas and (ii) calorie elasticities given in Table 4.12. Source: Computed by the authors. 53 Table 4.15 Cost of One Rupee Transfer to Poor Through Rice Subsidy Scheme in Andhra Pradesh 1986-87 Item Unit Cost 1) Total Annual Public Spending on Rice Scheme (a+b) Rs million 3490 a) Central Govt. Spending Rs million 1721 b) State Govt. Spending Rs million 1769 2) Annual Income Gain to All Consumers a) Per capita Rs 00.00 34.00 b) Aggregate Rs million 1990 3) Public Spending Required for transfer of 1 Rupee (1/2b) Rs 00.00 1.75 Per capita to All Consumers 4) Annual Income Gain to the Poor a) Per capita Rs 00.00 35.64 b) Aggregate Rs million 550 5) Public Spending required for I Rupee (1/4b) Rs 00.00 6.35 Income Gain to the Poor 6) Annual Income Gain to the Poor Under Perfect Targeting a) Per Capita Rs 00.00 129.06 b) Aggregate Rs million 1990 7) Public Spending Required for I Rupee Income Gain to the Poor Under Perfect Targeting Rs 00.00 1.75 Source: Computed by the authors. 54 Table 4.16 Cost of Calorie Increase Through Rice Scheme in Andhra Pradesh in 1986-87 All Consumers Per Capita Calorie Gain 510 K.cal/month Per Capita Income Gain Rs 2.83 per month Public Spending Required Transfer of Rs 2.83 per Month Rs 4.95 per month Cost of Transfer of 100 K.cal per Day Rs 0.97 Per Capita Calorie Gain to Poor 1350 K.cal/month Per Capita Income Gain to Poor Rs 2.97 per month Public Spending Required Transfer of Rs 2.80 to Poor Rs 18.86 Cost of Transfer of 100 K.cal Rs 1.40 Per Capita Calorie Gain to Poor 4590 K.cal/month Per Capita Income Gain to Poor Rs 10.75 per month Public Spending Required Transfer of Rs 10.85 to Poor Rs 18.81 Cost of Transfer of 100 K.cal Rs 0.41 Source: Computed by the authors. How do the food transfer schemes (both all-India PDS and Andhra's rice subsidy scheme) compare with other in-kind transfer programs in the country? We estimated the cost-effectiveness of alternative in-kind transfer schemes (Table 4.17). Both the PDS and Andhra's rice subsidy emerge as very expensive. Table 4.17 Cost per Re 1 of Income Transferred by Various Programs (1988-90) Total Cost PDS 5.37 AP Scheme 6.35 JRY 4.35 MEGS 3.1 ICDS 1.8 Source: Computed by the authors. 55 Why are the rice subsidy schemes so expensive? International experience suggests that the choice of the targeting method makes a lot of difference to cost-effectiveness (Subbarao, et. al. 1997). The rice subsidy scheme has adopted income-based means-tests. Given poor reporting of incomes, and the difficulty of verifying incomes, it is not surprising that income-based meanstests have resulted in excessive coverage. Nearly 70 percent of the population is engaged in informal sector activities and agriculture where historically income-reporting is non-existent. Table 4.17 shows that in terms of cost-effectiveness, the employment programs are better, largely because of the potential for self-targeting in these schemes. By far self-targeting is greatest in the Integrated Child Development Services program (Radhakrishna and Narayana, 1993) the principle reason why this scheme turns out to be the least expensive in terms of cost per rupee of income transferred. Incentive Effects Adverse incentive effects of transfer programs are hard to quantify. There has hardly been any such study for any transfer program in India. Some useful lessons can be drawn from the experience of a neighboring country, Sri Lanka. In Sri Lanka, a four-decade old quantity rationing system was replaced by a food stamp scheme. In the initial years following this change, the program continued to be very generous. Careful research has shown that the program resulted in reduced work effort of recipient households (Page 59 Table 5.2). Sri Lanka's experience underscores the importance of keeping the size of transfer modest in order to avoid adverse incentive effects on labor supply. How significant is the size transfer effected from the rice subsidy scheme? The mean consumption per household (for a five-member household) for 1993-94 is estimated to be about Rs 1110 per month. For a 20-kilo rice take-up at Rs 3.50 a kilo (when the market price is about Rs 7), the income transfer amounts to Rs 70 or about 6 percent of household expenditure, or 12 percent of household expenditure on food, assuming a poor household incurs 50 percent of total expenditure on food. In terms of work days, it amounts to wages of two workdays per household per month. The transfer appears to be modest, at least as the scheme is operating at present. The main problem with the food subsidy scheme appear to be their high financial costs per rupee of transfer, which is largely due to leakage of benefits to the nonpoor. The high transaction costs incurred by the poor, relative to the size of the transfer benefit, in the less-developed states and villages-the main reason for exclusion errors-remains a source of concern. In these respects, what lessons can be drawn from international experience? This is attempted in the next section. 56 CHAPTER 5 FOOD TRANSFERS: AN OVERVIEW OF INTERNATIONAL EXPERIENCE Regional Patterns in Program Choice Food subsidies have existed for more than four decades in some Asian countries; these have also been the subject of much evaluation and analysis. Three variants of food subsidy programs are in vogue: open general subsidy, quantity rationing, and food stamps. In addition, some programs have addressed the problem of nutritional deficiencies of specific groups such as pregnant women and children. Some regional patterns are worth noting. Quantity rationing has been the preferred form of intervention in South Asia, (See Table 5.1), food with work requirements in Latin American countries, and open general price subsidies (often with heavy involvement of parastatals) in some African and MENA countries.38 Nutrition and school feeding programs exist in all regions, though these predominate in Latin American countries. It is difficult to explain why countries differ in the choice of programs. A convincing demonstration of country variations in program choice would require data on program-specific coverage and size, poverty incidence, infrastructural development, and some index of administrative capacity. It is hard to obtain such detailed information even for a few countries. Available broad patterns suggest that, apart from country-specific historical circumstances, one reason for program choice could be that the administrative and physical infrastructure required to implement programs such as food stamps is much lower in African than in South Asian or Latin American countries. Food-for-work programs are currently prevalent only in a few countries with a low level of infrastructural development and a high level of headcount poverty ratio (e.g., Bangladesh, Ethiopia), whereas food stamps exist in countries which are infrastructurally more developed, and have a lower incidence of poverty.39 Political economy as well as administrative considerations also determine program choice: it is both politically infeasible and administratively difficult to run finely targeted programs in countries where the incidence of poverty is high, and where households requiring assistance are not sharply and easily identifiable from those which can make do without such assistance.40 Changes in the character or coverage of programs over time have been driven by considerations such as fiscal tightness (e.g. Sri Lanka, Tunisia), poor cost-effectiveness (e.g. the Philippines, India), and high inclusion and/or exclusion errors in targeting (e.g. Egypt, Brazil). 38 The objective of protecting the consumer, and the belief that private marketing chaimels are exploitative, led to the creation of parastatals in much of Africa. The adverse impact of administered price regimes implemented by the parastatals for producer as well as consumer welfare and equity is well-known. Refonn and reduction if not elimination of parastatals is an important component of the ongoing adjustment policy framework of many African countries. 39 It is not implied that Food for Work programs are easier to organize and to target cost-effectively than food stamps. -40 This is not to argue that a generalized price subsidy, lowering staple food prices for all consumers, is the best way of approaching the problem. Indeed, such a policy framework has had the opposite effect in some African countries: by enforcing lower farm prices via parastatals and blunting incentives to producers, it exacerbated supply response and aggravated the food problem in these countries. 57 Most countries initially introduced some form of food-based intervention (that was not necessarily targeted to the poor) in response to shocks that disrupted household-level food security. Upon choosing a particular form of food-based intervention, countries generally remained with it until one of the above mentioned considerations dictated a change in the program. The actual pattern and extent of change in the characteristics of the program depended in large measure on political economy considerations. Some governments have responded by either (a) modifying the existing program or replacing it by a new program, or (b) by eliminating the program altogether. In some countries, attempts to totally eliminate a program resulted in riots, eventually forcing governments to compromise by resorting to a much less targeted program than was initially planned before the eruption of riots. The history of state provision of cheap food in Egypt was punctuated by popular disturbances even during the 14th and 15th centuries. India, Pakistan, Bangladesh, Sri Lanka and Egypt, all had universal ration programs, initiated in the early 1940s (Alderman 1988, 1991; Ahmed 1988; Edirisinghe 1987; Subbarao 1989). Table 5.1 Regional Patterns in Food Transfer Programs Region Country General Pice Quantity Food Feeding Food with Subsidy Rationing Stamps Programs Work Requirement Sub Saharan Africa Ethiopia x x Gambia x Tanzania x Zimbabwe x Middle East and North Africa Algeria x Tunisia x Morocco x x Egypt x Jordan x (till 1992) x (1993-) South Asia Bangladesh x x India x x x Pakistan x x x SriLanka x x East Asia Philippines x x Indonesia x Latin America and Caribbean Brazil x x Chile x Colombia x x Costa Rica x Honduras x x Jamaica x x x Haiti x Peru x Uruguay x Mexico x Source: K. Subbarao, et. al. (1997). Reforms: Food Stamps, Self-Targeted Rations, and Other Approaches Almost all of these countries have begun to restructure their programs, some more thoroughly than others. For example, in Sri Lanka, the universal ration program prior to 1979 was costing the government up to 5 percent of GDP. The unsustainably high cost was the main 58 reason for the government's move towards food stamp programs which cut costs to 1.3 percent of GDP. However, the food stamps program in Sri Lanka is not well-targeted: many nonpoor are included, and some deserving poor are excluded (see Box 5.1). Moreover, the scheme has generated adverse incentive effects on labor supply. Research has shown that even after reforms, the scheme may have reduced both female and male labor supply (Table 5.2). Tunisia also cut costs from 4 percent to 2 percent of GDP by moving from a universal price subsidy to a more targeted program. As can be seen from Box 5.2, Tunisia's design differed from Sri Lanka's. Unlike Sri Lanka, Tunisia retained rations, but adopted self-targeting as a means to reach the poor. Table 5.2 Incentive Costs - Rice Subsidy, Sri Lanka, 1980 Before After Labor Supply Male 25.9 23.4 Female 23.9 21.1 Income Transfer/Household Gross 91.0 Net change in income 59.0 Source: Sahn and Alderman, 1995 A food stamp program was introduced in Jamaica in 1984, directed at pregnant and nursing mothers and young children, regardless of household income (Alderman, 1991). Although the Jamaican program was not specifically targeted to poor families, more than half of the recipient households belonged to the poorest quintile, while only 6 percent of those in the wealthiest quintile obtained stamps (Table 5.1). The Jamaican program differed from the Sri Lankan program in its design: unlike the latter, Jamaica did not adopt means-tests; instead, the stamps could be collected (by all) but only at the primary health clinics in exchange for immunization of children and pregnant women and nutrition counseling of mothers (Box 5.3). Since the nonpoor did not frequent these clinics, the program reached the poor effectively, and the administrative cost was 4 percent of program cost (Grouch, 1994). A food stamp program was pilot-tested in Colombia, but it never progressed beyond this stage (Urine, 1986). Mexico successfully replaced its urban maize subsidy program with the tortilla program, targeted to urban households with incomes lower than twice the minimum wage. Plastic cards were issued to households for identification. The program could be administered largely because it was an urban program. Kennedy and Alderman (1987) provide a summary of leakage estimates for various food intervention programs. Food stamp programs experienced low to moderate leakage, ranging from 3 to 10 percent, where the programs were targeted by the health status of vulnerable households and/or visits of households to clinics/primary schools (as in Jamaica). Where income-based means-tests were adopted (as, for example, in Sri Lanka and the United States), leakage ranged from 10 to 30 percent. 59 Box 5.1 Food Subsidies in Sri Lanka: Reforms and Impacts In 1979, Sri Lanka replaced a four-decade old quantity rationing system by a food stamp scheme. Prior to this change, food subsidies accounted for about 5 percent of GDP or 15 percent of total government expenditure. After the switch to food stamps, the share of food subsidies in GDP and as a percent of total government spending fell to 1.3 percent and 3 percent, respectively. Two features of this food stamps program have considerably reduced its effectiveness in helping the most vulnerable households. First, the value of food stamps was not indexed to inflation. The immediate consequence is an erosion in the real value of transfer as staple food prices (rice in particular) escalated soon after the switch to a food stamp program. Second, the principal method for identifying eligible households was self-reported household income. The value of stamps was adjusted for household composition: each member older than 12 would receive stamps worth Rs 15; each child between ages 8 and 12 would receive stamps worth Rs 20, and each child less than 8 years would receive stamps worth Rs 25. The information requirements for administering and monitoring the program were formidable. To ensure proper targeting of the program, it was necessary to continuously monitor incomes as well as household composition. Not surprisingly, household incomes were under-reported. Whereas the poverty ratio in the country was less than 30 percent, over 50 percent of the households were covered under the food stamp program-inclusion error was significant. A large share of food stamp outlay "leaked" to the upper-income households which reduced the cost-effectiveness of the program. Assuming that the responsibility of the government was to ensure adequate nutrition only to the lowest quintile, the cost to the government of providing a given amount of calories to the households in the lowest income quintile was 250 percent of the cost the households themselves would have incurred. Evidence suggests that the generous coverage may have enhanced the "real economic cost" of the program, as it may have reduced labor supply at margin. The impact of the food stamps program on the nutritional status of children would depend on the pattern of intra-household food allocation. A special survey of 480 children revealed that, controlling for all other factors, an increase in income due to food stamps increased the calorie consumption of children by 5.4 percent, whereas the calorie consumption of all other members of the households increased by 10 percent. The implication is: if the objective is to protect the nutritional status of children in very poor households, a nutrition program directly targeted to such children would have perhaps proved a more effective intervention than the food stamp program. Thus, the switch from quantity rations to food stamps did result in a reduced fiscal burden to the government, and enhanced the viability of the private sector in the foodgrain retail system. Yet the switch did not guarantee better targeting, nor did it protect the welfare of the most vulnerable. The lessons: (a) It is not enough to switch from one program to another; it is important to get the design of the program right to avoid inclusion and exclusion errors. (b) Before a change is initiated, it is useful to first decide on which objective is to be realized from a program, consider all potential alternatives, and then select the most cost-effective option. Sources: Edirisinghe, N. (1987) The food stamp scheme in Sri Lanka: Costs, Benefits and Options for Modification. (Washington DC. International Food Policy Research Institute.); H. Alderman (1991) "Food Subsidies and the Poor" in G. Psacharopoulos ed. Essays on Poverty and Equity and Growth. Oxford, Pergamon Press; D. Sahn and H. Alderman (1995) "The Effect of Food Subsidies on Labor Supply in Sri Lanka", in D. van de Walle and K. Nead (eds) Public Spending and the Poor, Baltimore and London, The Johns Hopkins University Press. 60 Box 5.2 Food Stamp Programs in Jamaica and Honduras: An Assessment In 1984, Jamaica switched from an expensive general food subsidy to a food stamp program. The program targeted four categories of vulnerable groups: pregnant and lactating women, children under six years of age, individuals who already qualify for poverty relief and public social assistance (mostly the elderly and the handicapped), and low-income single-parent or poor households. Food stamps are legal tender, and can be used to purchase cornmeal, rice, powdered skim milk, dark sugar, flour and meat. The value of the stamps is JS45 for pregnant and lactating women, J$60 for persons receiving poverty relief and public assistance and single-person households, and J$105 for multiple-headed households. Registration of recipients at a primary health clinic is necessary; stamps are issued every two months and can be collected in public areas (e.g., police stations, churches, etc.). Overall, the program is relatively well-targeted. Linking benefits to other requirements such as registration at clinics has considerably reduced inclusion errors, though some benefits certainly reach the upper quintiles. Thus, as of 1991-92, while there is a severe shortfall in the registration of pregnant and lactating women, single-person households and poor households have exceeded registration targets, implying substantial inclusion errors. Presumably the relatively higher value of benefits to households may have encouraged the nonpoor to enter the program. Although the rolls have been "cleaned" since the introduction of the scheme (once in 1987 and again in 1989), and recipients means-tested repeatedly, the system does not yet have exit mechanisms that are better defined and implemented. Although the poor were favored, the share of benefits across quintiles did not become progressive over time. According to the Survey of Living Conditions Report of 1991, the poorest 40 percent of the population comprised 58.2 percent of the beneficiaries-about the same as previous years. The main problem is the undercoverage of pregnant and lactating women-the most vulnerable group. The reason: the procedures for registration at health clinics is complicated and far from automatic: implementation problems persisted at all stages from application to registration to disbursement. Weak coordination between the staff of the Ministry of Labor and the Ministry of Health, and shortage of staff in the latter, have been cited as the main reasons for the observed deficiencies. Honduras provides another example of a delivery system that promoted self-targeting. In 1993, Honduras implemented a food stamp program that distributed stamps through schools and health centers. Under the Women Head of Household Coupon Program, food stamps are distributed through schools to poor mothers and their children attending grades one through three and who are shown to be at risk of malnutrition. In addition, at schools where the results of the annual nutrition survey shows a high incidence of malnutrition equal to or higher than 60 percent, all first graders are eligible for the program. Food stamps are distributed three times a year, and are estimated to cover 20 percent of household food expenditures. Food stamps are also delivered through health centers in the Maternal Child Coupon Program. The target group are low-income children under five and pregnant and lactating mothers. The intervention is concentrated on the earliest stages of infancy and childhood, and includes assistance to pregnant mothers to improve the chances of reaching children before malnutrition causes permanent damage. Beneficiaries must meet health surveillance requirements in order to maintain eligibility. The program also provides nutrition education and involves the private sector. Successful dissemination of information has resulted in 100 percent of merchants accepting the coupons. Local commercial banks are also willing to redeem coupons directly; this leaves no room for merchants to demand 'servicing fees" in redeeming coupons against goods. Beneficiaries can redeem coupons from commercial banks within four months of their issue (with redemption not contingent on purchasing of food). The program has been relatively more cost-effective than other nutrition programs in reducing poverty, and it has been widely accepted by participating retailers, and banks. Source: The World Bank (1994): Jamaica: A Strategyfor Growth and Poverty Reduction. 61 Box 5.3 Tunisia: Reducing Food Subsidies Without Hurting the Poor In Tunisia, universal food subsidy had existed since 1970. By the early 1980s, it became obvious that universal subsidy was simply not fiscally sustainable; the subsidy was costing the government 4 percent of GDP (and 10 percent of total government expenditure). In much of the Middle East, subsidized bread is often taken for granted; so the government was faced with the dilemma of reforming its subsidy program while protecting the poor, but in a politically acceptable way. The government retained the principle of "universality" but resorted to quality-differentiated food subsidy. The principal design of the program was as follows. A careful and disaggregated analysis of household consumer expenditure data revealed significant differences in consumption across income groups. The first step was to draw up a list of "inferior goods"-goods which were "perceived" to be inferior by the nonpoor because they possessed certain unattractive features in their packaging or ingredients. One type of bread disproportionately consumed by the poor was chosen for subsidization; subsidy on all other types of bread and pasta, disproportionately consumed by the rich, was eliminated. Similarly, "superior" food products were identified. Although food markets were otherwise tightly controlled in Tunisia, authorities have liberalized the sale of all superior-quality food products at market-determined prices-a policy in keeping with the general liberalization policy. All subsidies on pasta and fine flour-disproportionately consumed by the rich-were abolished. A similar approach was adopted for cooking oil. The household survey revealed that the rich consumed cooking oil that was properly bottled and labeled, whereas the poor were willing to buy unmarked, generic oils. While subsidy was retained on generic oils, the government liberalized the sale of superior oils. Attempts to improve the targeting of sugar were less successful. The government attempted an "inferior" goods approach by introducing a heavily subsidized less-refined brown sugar that was thought to be unattractive to the rich. In reality, even poorer households considered brown sugar "dirty" and failed to lift the supplies. However, subsidy on "cube" sugar, disproportionately consumed by the rich, was removed. Tunisia's experience suggests that self-targeting has considerable potential as a means to restrict subsidy to the poor in some countries. However, information requirements are quite demanding. Any country wanting to adopt this approach should have fairly detailed, quality-differentiated products listed in the household consumption survey. It is best if such surveys are small-scale and more frequent. Market studies are also required for testing consumer acceptance prior to introducing new (subsidized) products. Source: The World Bank (1996). Republic of Tunisia: From Universal Food ,Subsidies to a Self-Targeted Program. Washington DC. 62 Box 5.4 Food Rations and Food Stamps: Pros and Cons and Lessons Learned from International Experience Type of Program Pros Cons Food Price Keeps the price of food staples Has potential for high leakage to Subsidy (and Rations) low and avoids inflationary nonpoor. pressures. Allows for self-targeting via May have adverse effects on commodity selection. agricultural production. Has no information requirements Is expensive and fiscally for beneficiary identification and unsustainable in the medium to monitoring. long term. Food Stamps Facilitates targeting of transfer to Scope for fraud; stamps can be the poor. duplicated. Are prone to urban bias, feasible only if households rely mainly on market for food purchases. Allows for distribution channels Requires some form of that can be used as mechanism screening, i.e., difficult to for inducing behavioral changes. administer in countries where identification of beneficiaries is expensive or difficult. Can use existing retail networks Can have significant for distribution. disincentive effects depending on level of transfer and Impact may be limited in highly inflationary economies. Difficult to get people off this program. Supplementary Feeding * Targets high-risk individuals or * Can only reach limited number groups of individuals. of the poor at a time. No work disincentive effects Is administration-intensive, requiring a moderate amount of infrastructure and logistical support. Has minimal leakage to the * Requires coordination of health nonpoor. and/or education infrastructure at the grassroots level, usually a difficult task. Lessons Learned 1. Untargeted food transfers, whatever the form, are bound to be fiscally unsustainable, and should be avoided. 2. A range of targeting approaches exist. In general, approaches that impose an obligation on the part of the recipient are best in screening the needy. But such obligations should not be so onerous as to cause significant transaction costs to the poor. 3. Though administration-intensive, food transfers targeted to women and children, along with other services (e.g., immunization), can be a very effective means of supporting the poor with minimal distortions, provided communities are involved and the approach is demand-driven. Source: Adapted from K. Subbarao, et. al. (1996) Social Assistance and Poverty-targeted Programs: A Sourcebook. Poverty and Social Policy Department. The World Bank. 63 The pros and cons of various food rations and food stamps are delineated in Box 5.4. Food stamps have some definite disadvantages. First, they can be duplicated. In Zambia, large- scale counterfeiting led to the virtual abandonment of the program (von Braun et. al, 1992). However, fraud was prevented in many other countries. Second, food stamps not indexed for inflation are not a good option in countries experiencing hyper-inflation. Third, it is difficult to get people off food stamps once their incomes increase; this necessitates periodic means-testing which is expensive. There have been very few cases of innovative targeting methods employed under food rations programs. One such example is the Food for Education Program (FFEP), which was implemented in 1993 in Bangladesh. Eligible beneficiaries for the program are chosen from poor households who send their children to primary schools. Enrolled children must attend 85 percent of the total classes in a month. If a household has only one primary school-age (6 to 10 years old) child in a school then that household is entitled to 15 kg of wheat per month. If the household has more than one primary school-age child and sends all such children to school it is entitled to 30 kg of wheat. Targeted households must meet one of the following criteria: (a) they must either be landless or own less than 0.5 acres of land; (b) household head must be a day laborer, (c) household must be headed by a female (widowed, separated from husband, divorced, disabled husband); or (d) household head must be engaged in one of the low-income professions (such as fishermen, potters, blacksmiths, weavers, cobblers). Households that meet these criteria but not covered by the other targeted food programs qualify for assistance. Preliminary results from the program indicate that there have been significant increases in school attendance (from 63 percent in 1993 to 77.6 percent in 1994). Yearly drop-out rates declined from 18.5 percent in 1993 to 10.9 percent in 1994. For girls, drop-out rates declined by 9 percent and for boys it declined by 6 percent. These rates compare favorably with non-FFEP schools, where there was not a statistically significant decline in drop-out rates. All beneficiaries of the program were among the poor. However, leakage to the nonpoor of wheat allocated to the program was estimated at 6.5 percent. Many countries operate food supplementation and nutrition interventions for women and children, (a combination of immunization, oral rehydration and growth monitoring, food supplementation and fortification), not necessarily as temporary safety nets for poor families, but more broadly as direct measures to improve the nutritional status of vulnerable individuals. Countries vary widely in the range and effectiveness of their efforts to address the general problem of malnutrition. Countries also vary greatly in the amount of resources devoted in combating malnutrition. Three types of programs are to be distinguished: (a) general maternal and child nutrition programs (covering all pregnant and lactating mothers and pre-school children, not necessarily targeting the neediest of them (MCHN programs), (b) targeted nutrition programs (aimed only at mothers and children at nutritional risk), in combination with nutrition education and growth monitoring, and (c) school feeding programs. 64 Lessons from International Experience Targeting Efficiency Whether or not a program reaches the poor depends on how effectively the programn is targeted or how much of the program benefits leaks to non-intended beneficiaries. Leakage is measured by the amount of transfer going to the nonpoor as a percent of total transfer. Data for 10 countries are presented in Table 5.3. Generally the proportion of leakage to the nonpoor is higher under price subsidies and quantity rationing, modest under food stamps, and lowest under programs involving food with work requirements. However, targeting efficiency depends not only on the type of food program adopted, but also on its design. For example, Tunisia was able to greatly improve the targeting efficiency of a universal price subsidy by subsidizing only those foods consumed by the poor (see Box 5.2). In Bangladesh, leakages to the nonpoor in the ration program were greatly reduced by tying distribution to primary school attendance (Ahmed and Billah 1994). In the case of food stamps, targeting was improved in Honduras and Jamaica by linking stamp delivery to health clinics-see Box 5.3 (Subbarao, Ahmed, and Teklu 1995).41 Thus, cross-country and cross program experiences suggest that leakages can be prevented not so much by adopting a particular program per se, but by its innovative delivery ensuring self- selection of the poor into the program. Moreover, very narrowly targeted programs may not only enhance the costs of administration, but also result in the program losing political support and public funding, hurting the poor in the long run. Self-targeted approaches appear to be most suitable in many country situations. Cost-effectiveness Have the poor been assisted cost-effectively? Program costs are made up of administrative costs and food costs. Various measures are adopted to measure program costs including total annual cost per beneficiary, cost as a percent of GDP, and cost per unit of nutrients (calories) transferred. It is extremely difficult to get comparable cross-country data on these measures. Available data on the cost of transferring 1,000 calories per day per person, and per poor person, are shown in Table 5.3. All data are shown in US$ PPP-corrected 1987 numbers. Though it is hazardous to draw firm conclusions from this small sample, two findings are worth noting. First, the cost of transferring 1,000 calories is much lower for food stamps-whether or not means-tested--than for quantity rations. Second, the cost of transferring 1,000 calories per poor person, is indeed very high for quantity rations. The reason, of course, is obvious, viz., the high leakage of benefits to the nonpoor. (The difference between the cost per person and per poor person represents the additional cost due to spill-over of benefits to the unintended beneficiaries.) Thus, if the objective of public policy is to provide nutritional support to poor families, quantity ration (without any built-in design for self-selection--see Bangladesh example below) is quite clearly not a cost-effective option. An interesting example is from Bangladesh, where five different programs are currently in operation. These programs are compared for their cost-effectiveness (see Table 5.3). 41 In most voucher and stamp schemes leakage (defined as the value of vouchers/stamps going to the nonpoor) is low. However, voucher/stamps ate tradable, and the real test is how much of the voucher and stainp value reaches the poor? 65 Unfortunately, data on the size of the population covered under each program are not available. Nevertheless, one general finding holds good: typically, programs which have some work requirement, or some other obligations that ensure self-selection of the poor into the program (such as sending children to school) are a lot more cost-effective than food rations without any such restrictions (see row 3, Table 5.4). Table 5.3 Targeting and Transfer Efficiency for Selected Food Programs (1980s and early 1990s) Leakage t Cost o Transfer Cost to Transfer Cost as Nonpoor as 1000 Calories 1000 Calories to the Percent of Percent of perDay (1987 PoorperDay (1987 GDP Country Program Targeting Date Total US$)b US$) Transfer Brazil Feeding Geographic 1980 0.36 0.11 Colombia Feeding 1982 low' 0.46 0.67 Dominican Feeding 1982 lowa 0.24 0.45 Rep Pakistan Feeding Self-selection 1982 lowa 0.44 0.69 Bangladesh Ration Geograplic 1991 74.0 1.21 2.11 Colombia Ration Geographic 1981 0.41 0.41 Egypt Ration Universal 1982 66 0.22 0.36 15.0 Honduras Stamps Health Center 1991 Negligible 0.13 0.13 0.2 Honduras Stamps Schools 1991 Negligible 0.2 Sri Lanka Stamps Means test 1982 31 0.20 0.32 1.3 Brazil Subsidy Universal 1974 81 0.46 0.84 Mexico Subsidy Milk 1986 0.49 0.2 Bangladesh Food-for- Self-selection 1991 31 Work Bangladesh Food-for- Individual 1994 7 Education Assessment and child enrollment Note: a Although leakage to the nonpoor is low, leakage out of the program has been estimated to be quite high (i.e., 45.9 percent for Colombia Feeding, 85.8 percent for Dominican Republic Feeding, and 56 percent for Pakistan Feeding). This arises mainly from leakage to other members of household and through the substitution of program food for food normally consumed in the household. b Calculated as cost per beneficiary per day (assuniing 365 feeding days), multiplied by 1000 and divided by the calorie transfer. c Accounts for leakage to the nonpoor-cost to transfer 1000 calories times (I + the percentage leakage to the nonpoor). Source: Calculated from various coumtry case studies, Colhumns 5-8. As cited in K. Subbarao, et. al., Safety Net Programs and Poverty Reduction: Lessons from Cross-Country Experience, 1997. In terms of cost-effectiveness, targeting efficiency and benefits (both in terms of per capita transfer per annum, and in terms of nutritional impact), maternal and child nutrition programs have been generally found to be much superior in Latin America largely due to their potential for self selection of the neediest groups in the program (Musgrove, 1989, Horton, 1992). The superiority of nutrition programs notwithstanding, in most countries the total spending on maternal and child nutrition programs was less than 0.5 percent of GDP. Moreover, countries spending more or less the same amount differed much in the effectiveness of that spending. A targeted nutrition program backed by nutrition education and growth monitoring has been highly successful in cost-effectively combating malnutrition in India. India's Tamil Nadu Integrated Nutrition Project programs confirms this finding-mothers who brought their first child into the 66 program, and who were given "nutrition education," did absorb the lessons as was evident in the way they brought up their second and subsequent children (Sekhar, 1995). Unfortunately, very few such detailed studies exist for other countries. Table 5.4 Bangladesh: Cost-Effectiveness of Targeted Income Transfer.Programs, 1993-94 (per ton of grain) Rural Rationing Vulnerable Group RMP Food-for-Work Food-for-Work Food-for- Development CARE WKP Education (rice) (wheat) (cash)" (wheat) (wheat) (wheat) Costs GrainPurhase Coste $272 $129 $129 $129 $129 $129 - Sales Receipts $207 - Net Cost per Ton $ 65 Gov't. Contribution $ 16 Administration -DGFood $ 60 S 54 $ 27 $ 54 $ 54 $ 54 -CARE/WFP S I S 13 $ 42 S I - Ministries S 2 S I $ 8 5 8 $ 8 TotalCostperTon $125 $186 $170 $233 $192 $191 Ideal Actual Ideal' Actual Actual Ideal Actual Ideal Actuarl Ideal Actual Ideal Actual (IFRI) (IFRI) (WFP) (IFRI) (CARE (BBDS/ (IFR) FR) Income Transfer to Vulnerable Households Lekage 0% 70% 0% 14% 8% 0% 0% 0% 36% 0% 28% 0% 7% Income Transfer per Ton if Monetized at WorldPrice $ 73 5 19 $129 Slll $119 $129 $129 $129 $ 83 $129 $ 93 S129 $120 Costhncome Transferred (Id/2b) 1.73 6.55 1.44 1.68 1.56 1.32 1.32 1.64 2.81 1.49 2.06 1.48 1.59 The WGTFI estimated the cost-effectiveness of targeted programs at the then prevailing world price of $195 per ton of wheat. Currently, Bangladesh imports wheat under the U.S. Export Enhancement Program (EEP) and pays a CIF price of$ 129 per ton of wheat This report uses the current EEP wheat price. b Rural Maintenance Program (RMP) pays cash wages to its workers for maintaining rural roads. However, the cash is generated by monetizing wheat at the official ration price. Therefore, the purchase price of wheat, handling, and monetization costs are included in cost calculations for RMP. Note: "Actual" refers to the current situation. "Ideal" refers to the situation (simulated) with zero leakage to the nonpoor. Vulnerable households defined to include the population at risk of under nutrition and poverty. They constituted about 55 percent of the population in 1991-92. Source: As cited in K. Subbarmo, et. al., Safety Net Programs and Poverty Reduction: Lessons from Cross-Country Experience, 1997. As with other programs, the design aspects appear to be crucial for targeting and cost- effectiveness both for nutritional programs as much as for food stamps. Thus, clinic-based programs that combine health inputs with nutrition supplementation operate with lower administrative costs than school-based programs (World Bank, p.23). However, school-feeding programs may achieve other objectives such as promoting school enrollments particularly of girls, or improving the learning abilities of students. If this is the case, then the program of school 67 feeding should be assessed, not as a nutrition intervention, but as a program meant to realize other social objectives. Economic Costs and Disincentivesfor Work While assessing the effectiveness of a program, it is not enough to examine the program costs alone. Economic costs in terms of reduced work effort are also important. However, economic costs, by their very nature, are difficult to quantify; so it is not surprising that very few empirical estimates of economic costs are available. Yet, economic costs can be quite significant. For example, a recent study of the generous food stamp program in Sri Lanka found a substantial reduction in work effort of individuals receiving food stamps.42 Presumably targeted nutrition programs are least likely to generate disincentive effects, though no study has been done on this aspect. Another advantage of some nutrition programs-not allis their ability to combat intra-household discrimination in countries where such discrimination is pervasive. But such programs which successfully combated big leakages of food supplements and intra-household fundability are few and far between. Political Economy of Food Transfers Of all types of transfers, food transfers are governed very much by political economy considerations. Broad targeting or universal benefits is often due to the politicians' desire to maintain clienteles constituency. The most visible symptom of such clientelistic distortion is the urban bias food subsidy programs have entailed so long in many countries. Not only do such policies benefit the politically vocal urban population who are not at nutritional risk, but they hurt the rural poor (the landless in particular) who face food prices that are both higher and more unstable. The continuation of transfer policies benefiting both the poor and the rich inevitably involve heavy subsidies. As programs expand in response to the demands of political coalitions for rents from the state, explicit trade-offs soon emerge between spending on safety net programs and investments in growth. The resulting fiscal drain and macroeconomic disequilibrium inevitably set in motion the difficult political process of scaling down programs. In addition to India, countries such as Pakistan, Bangladesh, Egypt, Sri Lanka, Tunisia, and Morocco, are currently at various stages in this difficult political process towards scaled down, better targeted food subsidy programs for the poor. Political economy seems to prevail even in programs aimed at the welfare and protection of children. Thus, although growth monitoring and targeted food supplementation of children under three years of age and pregnant and lactating women is proven to be cost-effective with better overall impacts on the nutrition situation, the less effective school feeding programs are often taken up purely for political economy reasons. A most recent example is India. A nation- wide program of school-feeding at an enormous cost has recently been announced. Yet extensive 42 Sahn and Alderman found that in the rural areas of Sri Lanka, the reduction in work effort had a corresponding value of 50 percent (for males) and 40 percent (for females), of the value of the subsidy. 68 research on India's nutrition programs has pointed to the ineffectiveness of school feeding as a nutrition intervention, especially since much malnutrition is prevalent among pre-schoolers (Subbarao, 1989a). Conclusions Lessons from international experience. Four general lessons emerge: * First, practically every country that adopted open-ended generalized price subsidies or quantity rationing as a way to protect the poor has found the policy not only fiscally unsustainable, but also significantly distortionary in its effect on the agricultural sector.43 * Second, many countries have resorted to targeted approaches. No particular targeting method has been found to be perfect; yet programs which are imperfectly targeted have proven to be better in reaching the poor and keeping costs down than no targeting at all. Moreover, leakages appear to be lowest in programs that allowed the self-selection of beneficiaries into the program, based on beneficiary (or their children's) participation at primary schools or clinics. Income-based means tests have almost always proven expensive and difficult to administer cost-effectively." Tying beneficiary participation (or time contribution) to food transfers, not only lowers leakages, but may also lower economic costs associated with work disincentives. However, it is important to ensure that self-selection procedures (such as tying a transfer to school attendance) do not increase transaction costs to the poor significantly. - Third, while the cost of transferring income to the poor is lower for food stamps and targeted food and nutrition programs, than for quantity rations or general subsidies, it is difficult to rank the targeted programs in terms of their cost effectiveness and incentive costs because of the significant variations in the design of programs. The cost- effectiveness of a program depends very much on the choice of a particular targeting method and the design and delivery of the program which, in turn, depend on the needs of beneficiaries and on country-specific constraints. 43 Although not discussed here, some of these policies that tax agriculture have had deleterious effects on production, particularly in Sub-Saharan Africa. 44 The reference here is not to school feeding, but to programs such as food ratious which are given subject to the household's willingness to send children to schools. 69 * Fourth, political economy considerations are a major country-specific constraint that need to be considered in designing sustainable food transfer programs. In particular, too much of fine tuning in targeting may result in the erosion of political support for food transfers.45 45 Gelback and Pritchett (1995) argue that too much of indicator-based targeting may result in expenditure cuts which offset the benefits of better targeting. Self-targeting may survive politically more than otlier fonns of targeting. This will always be the case because the politician can say the individual decided to participate/not to participate. It was not a govenmment decision. 70 CHAPTER 6 PDS: SUMMARY OF FINDINGS AND OPTIONS FOR REFORM Introduced in response to very critical food shortages in the mid-1960s, PDS has evolved into a price support-cum-subsidy program. The program has continued more or less unchanged although the food scenario at the macro- (national) level has changed from one of scarcity to a surplus situation. PDS subsidies account for substantial budgetary expenditures at the Central level; public spending levels are exceptionally high in states which have their own subsidy programs. Moreover, the country is implementing an economic reform program. At no time is the question of the efficacy of PDS as a safety net more relevant than now, especially since the Government has recently embarked on a drastic reform of the system. Although PDS is an over- researched subject in India, few studies have estimated the welfare gains and costs per rupee of income transferred. Almost no studies exist that compare the benefits and costs of PDS relative to other similar programs. This paper's aim was to fill this gap. The main findings of the paper are summarized in this chapter. Recent developments with respect to prices and off-take are then assessed. In the light of the past performance of the PDS, current developments, and international experience, the chapter concludes by suggesting viable options for a reformed and cost-effective food transfer program for India. Main Findings 3 Regional mistargeting. There is a regional mistargeting in the distribution of foodgrains through PDS. The off-take by the by states like Andhra Pradesh and Kerala, which are implementing subsidized food scheme, is high. On the other hand, the off-take by the poorer states like Bihar, Orissa and Madhya Pradesh is low. 3 Poor targeting and access in most states. Despite being in operation for four decades, the access of the poor to PDS is still very limited. The more recent (1995) data based on selected village surveys only reconfirm the poor's limited access in Bihar and Uttar Pradesh to PDS found earlier from the larger household data set for 1986-87. * Regressive and too generous in some states. The access of the poor to PDS is certainly much better in a few states, particularly Andhra Pradesh and Kerala. Yet, even in these states, per capita monthly PDS cereal purchases tended to be regressive. The conclusion: with few exceptions, PDS remained an untargeted program, with the poor's access being limited at best and nil at worst. * Negligible welfare gains. The per capita income gain to the poor from all consumer subsidies was no more than Rs 2.01 per month, or 2.7 percent of their per capita expenditure, in rural areas; in urban areas it was Rs 3.4 per month (or 3.2 percent of per capita expenditure). There were differences in income transfers between the commodities and across the states, but the overall transfer gains were very meager indeed. 71 * Cost-ineffectiveness. Even the meager transfer benefits were realized at an exorbitant cost. When the Central Government costs alone are considered, an amount of Rs 4.27 was needed to transfer one rupee of income to the poor. But typically the states would also incur additional costs. When both the Central and state-level expenses are properly accounted forand this study was able to do this for Andhra Pradesh-one rupee of income was transferred to the poor at a cost of Rs 6.35. That this level of spending was incurred in states such as Andhra Pradesh which have historically much better administrative structures than states such as Bihar, suggests that the program could hardly be expected to become cost-effective unless drastic changes were introduced in its design. * Minimal impact on poverty and nutritional status. Not surprisingly, the impact of PDS on poverty and nutritional status has been minimal. Actually, with the exception of Kerala and Andhra Pradesh, the impacts on poverty and nutritional status was negligible. For the country as a whole, there would have been a decline of barely 2 percentage points in the poverty ratio due to the combined incidence (income gains) of food and non-food consumer subsidies. Comparison with other programs. Of all poverty-targeted programs, PDS and the AP rice scheme appear to be the least efficient in terms of transfer gains and costs. Employment programs fare better than food transfer programs in terms of cost-effectiveness, though ICDS appears to transfer income to the poor at lowest cost. Recent Trends in Administered Prices The minimum support price was raised by 69 percent for wheat and 44 percent for rice between 1990-91 and 1995-96. The administered prices were substantially higher than what the closed market would otherwise bear-so the farmers unloaded huge stocks on the market. Since the FCI has no choice but to buy whatever is offered at the minimum support price, it has had to buy up more wheat and rice than it could manage efficiently. In order to minimize the losses arising out of an enhanced procurement price, the issue prices were also correspondingly raised. Due to the large increases in issue prices, the off-take from PDS has declined. Consequently, buffer stocks have reached uneconomic levels, far exceeding the norms suggested by the technical group on buffer stock management set up by the Government. The carrying cost of huge buffer stocks is placing an excessive burden on the government. In 1994-95, out of the Central food subsidy cost of Rs 51 billion, the carrying cost of buffer stocks were Rs 18.5 billion (or 36 percent of total food subsidy cost). Because of poor management, while the cereal subsidy (defined as FCI economic cost minus its sales realization) itself has declined since 1992-93, the total central food subsidy cost has risen. The share of cereal subsidy going to poverty-targeted programs (such as JRY) has also been reduced. Increase in total food subsidy cost seems to bear no relation to the benefits accruing to the poor. 72 The sale of wheat in the open market is emerging as the major activity of the FCI. Such sales increased from 0.02 mt in 1992-93 to 5.19 mt in 1994-95. The open market sales of rice have also recently gathered momentum. Whether or not such sales have eventually reduced the total food subsidy depends very much on what is happening to the carrying cost of buffer stocks. The moot question is one of the counterfactual (within the framework of a closed economy): what would have been the open market price had the same quantity which the FCI had off-loaded in the open market not been procured, stocked and sold by the FCI, but allowed to flow through the normal market channels? It is arguable that the outcome would have been the same, if not better, for the taxpayers money would have been saved to that extent. The upward revision of issue prices has reduced the price differential between the open market price and the central issue price. Only the states which are prepared to give a further subsidy, such as Andhra Pradesh and Karnataka, and the food deficit state of Kerala, have lifted their allotments from the FCI. The PDS allotments have not been lifted in most other states, particularly in the states of Bihar, Uttar Pradesh, and Orissa. Because the issue price and open market prices are almost identical in these states, PDS has virtually become dysfunctional. Following economic reform, the retail prices of foodgrains have risen all over the country. The poor in the eastern belt were not protected at all by the PDS. The relatively richer states such as Andhra Pradesh did succeed in keeping the ration price lower than the open market price and the central issue price, but at a huge subsidy cost that may have crowded out other critical investments. Thus, PDS was dysfunctional in states with a high concentration of the poor, while in other states PDS provided a cushion for the poor in the short-run but at the cost of excessive fiscal strain and possible diversion of public resources from other priority areas such as health and education, the consequences of which are bound to harmful to the poor in the medium-run. The question that arises from the foregoing is: are there no better ways of providing food security to the poor? What lessons can be drawn from available international experience with respect to food transfers? What are the options for reform? Lessons from International Experience Many countries in Asia, Africa, Latin America and the Middle East have experimented with different types of food transfer programs. The cross-country experience reviewed in Chapter 5 suggests that practically every country that adopted open-ended generalized price subsidies as a way to protect the food consumption levels of the poor has found the policy not only fiscally unsustainable but also, more importantly, significantly distortionary in its effects on the agricultural sector. Most countries have begun to reform and restructure their systems of food transfer. Some have retained the "generalized" nature of the program but introduced self-targeting by keeping the subsidy only for those commodities which are consumed disproportionately by the poor. Others have shifted to programs that lend themselves to targeting. Evidence suggests that no particular targeting method has been found to be perfect; yet programs which are imperfectly targeted have proven to be better in reaching the poor and 73 keeping costs down than no targeting at all. However, merely shifting to a targeted program such as, for example, food stamps, does not automatically ensure that all of the poor gain access, and leakages to the nonpoor are avoided. In general, income-based means tests have proven extremely difficult to administer. Leakages appear to be lowest in programs that self-select beneficiaries into the program. Self-selection can be promoted in many ways. One way is to subsidize only those varieties of grain disproportionately consumed by the province. Another way is to select beneficiaries contingent upon the beneficiary's (or their children's) participation in another activity, such as a child's attendance in a primary school, or registration at primary health clinics for all preventive services. Tying beneficiary participation (or time contribution) to food transfers lowers not only leakages but may also lower the economic costs associated with work disincentives. However, care needs to be taken that the poor's participation does not lead to excessive transaction costs to them. PDS in the Context of a Changed Agricultural Scenario A major conclusion of this study is that an unreformed PDS (of the 1960s vintage) would not fit in with the dynamic changes of the 1990s. It is important to bear in mind that India's agricultural economy has vastly changed since the early 1960s when the PDS in its present form had evolved. Four aspects of change are worth stressing: - As already noted, overall, national-level food supply scenario has changed from one of scarcity to one of surplus. Cereal production grew at 3 percent during the 1980s 3/4 faster than the growth of effective demand for cereals (2.6 percent). The expectation is that the surplus situation will be sustained. In fact the effective demand for cereals is projected to grow at an even lower rate of 2.45 percent during 1995-2000, and 2.24 percent during 2000-2010 (Radhakrishna, 1996). Drought situations cannot be ruled out, but the scenario of catastrophic national-level shortages are only a remote possibility. - Another noteworthy recent change is with respect to the pattern of agricultural growth. There has been an improvement in the foodgrains growth rates in all the eastern states; the growth has been spectacular in West Bengal. Widespread agricultural growth, because of its labor-absorbing character, tends to have strong poverty-reduction effects, thereby obviating the need for wide coverage of PDS, and underscoring the need for its targeting. * Except in tribal and a few dry land areas, the market infrastructure for foodgrains in India is well-developed and the sub-markets are well-integrated (Subbarao, 1989b). * Following the recent amendment of the Constitution, Panchayat Raj institutions (PRIs) may be entrusted with the responsibility of planning and implementing the poverty alleviation programs. These responsibilities will require the PRIs to identify the poor 74 assisted under the anti-poverty programs, one of which is PDS. The targeting of PDS will improve if the PRIs are involved in the selection of beneficiaries for PDS.46 Recent GOI Proposal for a Targeted PDS GOI has announced its intention to introduce sweeping reforms in the Public Distribution System (PDS) in an effort to improve targeting. According to recently issued guidelines, the Targeted PDS (TPDS) would offer two separate distribution channels: one aimed at households below the poverty line, and the other for the population above the poverty line. Under the first channel targeted to the poor households, the central govermnent would transfer to state governments wheat and rice at about half the issue price set for the PDS. The monthly ration under the TPDS would be set at 10 kg per poor household in the state. The number of poor households in a state would be the one determining by the recently approved Expert Group's methodology. This would effectively determine for each state a maximum entitlement based on the number of poor households in the state-a vast improvement in relation to the present system in which the amount of foodgrain that a state can drawn from the PDS is left at the discretion of the state and thus leads to situations where states with a high incidence of poverty utilized PDS much less than states with a low incidence of poverty. The Central Government would leave to State governments the responsibility of designing and implementing targeting mechanisms for reaching the poor, and corresponding guidelines were recently issued by the Ministry of Food. State government would need to be in a position to identify the poor, issue special cards, and deliver foodgrain to the intended beneficiaries. The central government would monitor the states' performance in identifying and delivering fo.odgrains to the beneficiaries, for which reporting requirements have been developed. These features of the TPDS will encourage the states to improve targeting or else, at least in theory, their access to the TPDS could be discontinued. Under the second and non-targeted channel, which the guidelines indicate would be phased out gradually, the central government would transfer to state governments wheat and rice at an issue price which would remain close to the market price. Access to this non-targeted TPDS channel would be universal. It is proposed that-as an interim measure-the quantities to be allocated to each state be based on the average lifting of wheat and rice over the last ten years by the states. The proposed TPDS, however, does not resolve some basic issues, and raises some new problems. * Allocation of 10 kg. per family under the poverty line is bound to have a regressive effect. We know from the poverty profile for India that there is a significant correlation between family size and poverty. The proposed TPDS delivers proportionately smaller transfer (on a capita basis) to families with a large number of children who, on balance, tend to be poorer than small families. 4' The Kamnataka experience of involving PRIs in the distribution of subsidized rice has led to substantial improvement in its targeting (Thiinnaiah, 1996). 75 * The same retail outlet catering to families both below and above the poverty line and selling the same commodity at substantially different prices renders the scheme extremely difficult to implement in practice, rigorous monitoring notwithstanding. Similar attempts earlier for kerosene and vegetable oil have led to widespread hoarding and diversion of commodities simply because profiteering from such a diversion were enormous, monitoring difficult and administratively expensive. The TPDS offers no incentives to retail dealers to implement the program as proposed. * The GOI has proposed rigorous enforcement and monitoring. But the cost of enforcement and monitoring is likely to be high, as is evident from international experience of schemes involving strict administrative enforcement. It might be preferable, instead, to consider alternative transfer mechanisms that promote self-targeting and mitigate against leakages and provide incentives for enforcement. With these considerations in mind, some farther reform options to TPDS are discussed in the next sections. Elements of a Reformed PDS Improve the Performance of Cereal Markets Proposals for a revamped and restructured food transfer program should be recommended in the context of the evolving agricultural scenario and on-going institutional and economy-wide policy changes. In a country where the living standards of the poor are overwhelmingly influenced by the price of food staples, the policy instruments chosen for any food transfer should be consistent with price stabilization objectives. Besides, it would seem desirable to support the role and performance of markets in moving foodgrains from surplus to deficit regions and between harvest and lean seasons. Given the overall surplus situation, there is little possibility of a few high income (urban) areas cornering the supplies. With the comfortable availability of foreign exchange, short term production shortfalls and hoarding behavior can be avoided by imports, underscoring the need to free the external trade regime and promote infrastructure investments to import and move foodgrain domestically. Such dangers may have been present in the 1960s, but not in the 1990s. Therefore, in order to reform the PDS, it would seem necessary to (a) promote the performance of foodgrain markets by phasing out (domestic and external) trade restrictions and policy disincentives to marketing infrastructure investments, (b) devise instruments which improve food security of the poor while at the same time letting grain markets function freely, (c) relieve the price policy from pressure groups, and (d) improve the targeting and cost-efficiency of food transfer. New, Supportive Role of FCI Given the weaknesses and constraints of FCI, the pervasive role of well-developed foodgrain markets, and bearing in mind the improved supply situation, it would seem appropriate to phase-out government controls as well as current procurement operations. FCI could be allowed to compete in the market, free from controls, but with the added advantage of scale economies and comparative advantage in inter-state trade (BICP, 1992). Thus, FCI may, over 76 time, become an effective market player and indeed promote the development of efficient and competitive markets by adopting commercial practices (e.g., by doing away with the (rice) levy system and its pan-seasonal and pan-territorial price policy), and expanding the sub-contracting of its storage and transport activities to private operators. FCI may also act as countervailing power and stabilizing influence to private trade by progressively limiting its functions to a holder of strategic buffer stock (or foreign exchange reserves), intervening in grain markets only in periods of critical, national-level food shortages. How Should the Poor be Identified? Recent moves towards targeting have resulted in ration cards being issued after an income-based means tests at least in urban areas. In backward rural areas, PDS has been extended but the access is contingent upon participation in public employment programs. Neither approach has had much of a success. One reason for the failure of distribution of cereals through employment programs can be attributed to the lack of coordination at the grass roots between the PDS agencies and the staff involved in employment programs. This study recommends that the Panchayat Raj institutions be entrusted with the responsibility of identifying the poor based on household-specific characteristics. Several characteristics can be used for including a household into the program, or excluding a household from the program. For example, all households participating in employment programs may be included; another category could be single mothers with children, or widows with no support. Similarly, some principles for exclusion could also be developed. For example, all income tax assesses, or households owning more than 5 acres of irrigated land or 10 acres of unirrigated land could be excluded. It would be relatively easy for a Panchayat Raj institution to administer such criteria. It would be help if this activity is overseen by an NGO where possible. The important point that we wish to stress is: avoid exclusive reliance on income-based means tests, and use village-level panchayat institutions or gram sabhas (village councils) to identify the poor for program benefits. Urban areas may require a different approach to targeting. One approach is careful selection of areas: slums, areas occupied by traditionally poor communities such as potters or cobblers, construction workers, etc. This can be supplemented by an exclusion principle: exclude households who are income tax assesses, or possess expensive consumer durables and services (cooking gas connections), etc. How Should the Program be Redesigned? Three issues bearing on the design are: 1. The Government is currently incurring a huge subsidy cost. Once the FCI goes out of the operations of procurement and supply to PDS shops, a substantial proportion of subsidy is saved. The saved food subsidy could be distributed to PRIs in the form of food stamps or vouchers. The allocations to PRIs can be based on already available information on the incidence and severity of poverty. (NSS data are drawn from a fairly representative sample 77 across the country, and poverty ratios can be derived according to agro-climatic zones (see Rao, Ray and Subbarao); such disaggregated information on the varying incidence of poverty should provide useful for the allocation of subsidy. Such a procedure would also allow for groups of villages with a higher incidence of poverty to receive higher allocations and vice versa. Alternatively, criteria similar to those used in the allocation of financial resources for JRY/EAS may be used. Once a PRI receives the amount of subsidy (in terms of vouchers or stamps), it would follow the procedure described in the foregoing paragraph for identifying the poor for subsidy (ensuring that vouchers/stamps be given to the women). The poor would have the option to buy from among a list of essential commodities (food, fuel or clothing). Since most villages have retail outlets-some PDS outlets and some with regular private small-scale grocery stores-it should be possible to exchange the vouchers for the needed commodities. Vouchers could be exchanged either with PDS outlet stores or with private stores. Such a program would allow for greater flexibility in operation and choice of outlet and commodity composition to the household. 2. It may not be feasible to adopt the above program in the entire country. Some villages in India may not have a good network of private shops. Food stamp or voucher program may not become feasible in the immediate future. The alternative is to continue the same PDS cutlets, but introduce PRI-administered identification of the poor in place of the current, bureaucratically administered identification based on means-tests. 3. The linking of PDS with employment and other nutrition programs is another component of the proposed design. It is workable only if coordination among the agencies involved is ensured. India has been experimenting with part payment of wages in food under the JRY and the EAS scheme. Many villages also have a child nutrition program (the Integrated Child Development Services). Households utilizing the services of ICDS outlets may be enrolled for PDS. In the case of the very poor who are unable to participate in public employment programs due to old age or disability, PRIs may have to identify such individuals for assistance. Reform Options to Improve Targeting The proposed reform options and design changes summarized in Table 6.1 will have the least distortionary effects, involve almost no additional requirements, and at the same time free the FCI of some of its functions. Even if foodgrains actually reach the states in proportion to the incidence of poverty, the absence of an innovative delivery system can hamper the access of the poor to the system. The GOI may consider three reform (design) options: Option 1: Whether under PDS or TPDS, the fundamental problem is one of identification of the poor and improved household targeting. PRIs are in a better position to identify the poor households. Therefore, the responsibility of identifying the target households and issuing ration cards may be assigned to PRIs. This modification is likely to improve targeting and minimize leakages. 78 Even on political economy considerations, the scheme is unlikely to be met with much resistance since it makes no radical departure from the proposed TPDS. There is some sort of consensus among the policy makers that the involvement of panchayat raj institutions would improve the functioning of the system. However, the structural weaknesses associated with the present administered price regime may remain. Option 2: The Central government allocates the food subsidy (in the form of cash) as a conditional grant to States on the agreed formula under the TPDS. FCI would then be relieved of its task of procurement and distribution of subsidized cereals. However, FCI would still be expected to stabilize price within a narrow range by purchase and sales in the open market and would be given an autonomous status without any government controls. FCI would then not be involved in the actual procurement and distribution of subsidized cereals. The State governments would be free to buy from FCI or open market to meet their PDS requirement without interfering with the free movement of commodities within their states. The State Civil Supply Corporation would distribute the subsidized cereals within a State (the State may supplement the Central subsidy). The PRIs would be assigned the task of identifying the poor households and issuing ration cards. This option, besides improving the targeting of PDS, is expected to make FCI more efficient and contribute to improved performance of domestic grain markets. In this option, the State government would be the major player and the Central government's role would be limited to the allocation of Central food subsidy in cash to the States. If the inter-state allocation of subsidy is based on rational criteria, the regional mistargeting may not be a serious problem. However, the inefficiencies of the State government agencies involved in the purchase, storage and transport of foodgrains may increase the cost of income transfer to the poor. Given the growing importance of federalism in India, this option is likely to be politically feasible. Option 3: This would be a radical departure from the existing system. The Central government would distribute food coupons to States under an agreed formula as proposed in the TPDS, and they in turn to PRIs. The panchayats would identify the poor and distribute the coupons to poor households. The role of FCI and State Civil Supply Corporations would be minimal. Private retail dealers would cash the coupons at any branch of any recognized financial institution. Food grain prices would remain undistorted at their market-clearing levels. The Food Stamp Scheme would be the better option in urban areas as well as infrastracturally developed rural areas and the decentralized PDS (option 2) in the lagging regions where markets are weak. A combination of food stamps which uses existing market channels and decentralized TPDS is worth considering for India. As market infrastructure develops in the lagging regions, the food stamps scheme may gradually replace the other scheme. 79 In some very poor States like Bihar, Orissa and Madhya Pradesh whose administrative structures are poor and where local institutions are very underdeveloped, special programs which tie food distribution to wage employment programs (JRY/EAS), nutrition programns (ICDS/MDM)) and welfare programs for old and disabled persons may receive high priority, as proposed under the TPDS (Section IX, GOI 1997). Conclusion PDS reform is long overdue. TPDS is proposing to make great strides towards an improved targeted system. However, further improvements are needed to reach the poor cost effectively under the TPDS. India's own experience and international experience offers substantive guidance to approaches for reform. Some of the approaches to reform and their implications for the institutions involved are outlined in this paper. Given the vast size of the country, it is not necessary for one particular reform approach or design to be adopted throughout the country. Ideally, different approaches should be tried in different regions/states, and the experience gained be used for further restructuring. 80 Table 6.1: Functional Framework under Existing and Modified PDS/Food Stamps Central Government Food Corporation of State Government State Civil Supply Panchayat Raj NGOs/ Women Ministry ofAgriculture and India Food and Civil Corporation Institutions Organizations Ministry of Food and Supplies Department Consumer.Affairs TPDS Fixation of: Procurement, storage, Issue of ration cards; Lifts foodgrains (existing) minimum support prices, transportation of fixation of ration retail from FCI godowns procurement targets, foodgrains, open price, ration scale; and transports it to centrl issue prices, etc. market sales, exports allotment to districts. its own godoNvns maintenance of buffer and to fair price Allotment of quota to: stocks. shops. TPDS, States, JRY/EAS, ITDP. TPDS state allocations. Buffer stock operations. Option I Fixation of: Procurement, storage, Fixation of ration Lifts foodgrains Identifies the poor Participates in Gram Modified TPDS minimum support prices, transportation of retail price, ration from FCI godowns and issues ration Sabha meetings of procurement targets, foodgrains, open scale, allotment to and transports it to cards. village panchayats in central issue prices, etc. market sales, exports districts. own godowns and to identifying the poor. maintenance of buffer fair price shop. Allotment of quota to: stocks. TPDS, States, JRY/EAS, ITDP. TPDS state allocations. Buffer stock operations. l Option II Allocation of food subsidy (in Open market purchases Fixation of ration Free to buy grains Identifies the poor Participates in Gram Decentralized cash) to States. and sales. retail price, ration from FCI or from and issues ration Sabha meetings of TPDS scale, allotment to the open market w/o cards. village panchayats in districts. interfering w/ free identifying the poor. movement of grains. Option III Allocation of food stamps to Open Market purchases Allocation of food Identifies the poor Participates in Gram Food Stamps States. and sales. stamps to PRIs. and issues food Sabha meetings of stamps. village panchayats in identifying the poor. APPENDIX NOTE ON DATA ADJUSTMENTS National Sample Survey Organization (NSSO) in its 42nd round (i.e., 1986-87) collected information on the utilization of Public Distribution System (PDS). The Survey covered 73,564 households spread over 8400 villages and 4465 urban blocks all over India. Indira Gandhi Institute for Development Research (IGIDR) has tabulated the household data of this survey by 20 quintiles formed on the basis of per capita consumer expenditure. Their tabulations provide for each quintile information on per capita quantity and value of rice, wheat, jowar, bajra, other cereals pulses, edible oil, sugar, kerosene, coal, and clothing purchased both from PDS and open-market. The tabulated data also contain data on per capita total expenditure. Critical analysis of the data showed that the estimates were unreliable.4' Hence they cannot be used in poverty analysis. In the same year (i.e., 1986-87), NSSO carried out a survey on household consumer expenditure. NSSO Reports (Number 351/1 and 351/2) provide per capita consumer expenditure on various items of consumption for different expenditure classes as well as distribution of households and persons by expenditures classes. This data could complement the decile tabulated data of IGIDR. Lorenz curve has been estimated from the data on distribution of households by per capita expenditure. From the estimated Lorenz Curve quintile-wise per capita expenditure and distribution of persons by quintile classes have been derived. This information has been matched with that of IGIDR, which contains information on PDS. The income gain due to subsidy on a commodity has been estimated from the IGIDR data set by using: Ayi = qRi (pM-pR) Ay; is per capita income gain for ii" quintile, qRi is per capita PDS quantity purchased by ih quintile, PM is open market price, and PR iS ration price. There are outlying observations in the tabulated data of IGIDR which were eliminated. The estimates of poverty and class specific income gains due to subsidies obtained without as well as with eliminating outliers are found to be very close to those obtained without eliminating the outliers. Poverty estimates are also comparable to the estimates obtained by the Expert Group on In the NSS survey on the utilization of PDS unlike in the surveys on consumer expenditure, the information on per capita total expenditure was collected by posing one-shot question to the respondents. Comparison of this data with those consumer expenditure survey in other rounds showed that this data is not useful for poverty analysis and can be used only for clarificatory purposes. 82 Estimation of Number and Proportion of Poor for the year 1987-88. For the present study, the estimates obtained by eliminating outliers have been used. The quintile-wise per capita expenditure and distribution of persons by quintile classes (households) obtained from the NSS household consumption data and quintilewise income gains estimated from the IGIDR data constitute the basic data set for the analyses of the effects of subsidies on poverty. The extent of subsidy transfer by items to various income groups and the estimated parameters of Lorenz curve under alternative scenario for rural and the urban areas of 16 major states of Indian Union are presented in the Appendices. The Scenario PCE pertains to the actual situation prevailing in 1986-87 (i.e., inclusive of all subsidies). The Scenario PCE adjusted for rice subsidy indicates the situation that would have prevailed in the absence of rice subsidy. Similar interpretation may be given for other scenario. The scenario PCE adjusted for 'All' refers to a situation that would have prevailed in the absence of all subsidies. 83 APPENDIX 1 Institutional Arrangements in Selected States Introduction The Central and State Governments are responsible for ensuring the smooth functioning of the public distribution system; while FCI procures, stores and transports the subsidized foodgrains to Central godowns, the State Governments lift foodgrains from Central godowns and distributes them through a network of Fair Price Shops. Every year the Central Government fixes the procurement/minimum support prices and also fixes the procurement targets for each State.48 FCI implements the procurement policy of the Central Government. The State Governments/their agencies and Producer Co-operative Societies procure foodgrains on behalf of FCI. These agencies have the freedom to choose the mode of procurement. Generally, rice is procured by imposing a levy on rice millers while wheat is purchased from the mandis (wholesale foodgrain markets). FCI makes available subsidized foodgrains to the State Governments at a uniform price. The distribution of foodgrains is under the jurisdiction of the State Governments and some State Governments are pursuing a vigorous Public Distribution Policy. Every month, each State Government makes an estimate of its requirement of foodgrains and requests the Central Government to release the requisite quantity from the Central pool. The Central Government, depending upon the availability and the requests made by different State Governments, allocates the monthly foodgrain quota to the various States. It is the responsibility of the State Government to lift the allotted quota and distribute the same to the people through the Fair Price Shops (FPS). In each State, the food and civil supply department looks after the delivery system. Some States have set up State Corporations to lift foodgrains from the FCI godowns and transport it to the fair price shops. There are considerable inter-State variations in the institutional structure. For instance, while in Andhra Pradesh the Civil Supply Corporation-a State enterprise-undertakes the actual movement of foodgrains, in Kerala, private traders are involved in the movement of foodgrains under the supervision of the Food and Civil Supply Department. Similarly, the relative roles of State and private sectors in storing foodgrains differ across States. As we shall see, the scale of ration, the retail ration price, the identification of target group and discrimination within the target group vary across the States. 48 Wheat procurement takes place mainly in Punjab, Haryana and Uttar Pradesh (together they account for almost the entire wheat procurement). Rice procurement is slightly less concentrated. Andhra Pradesh, Haryana, Punjab, Tamil Nadu and Uttar Pradesh account for about 90 per cent of the rice procurement. 84 PDS in Andhra Pradesh Political Economy of Subsidized Rice The State Government introduced the 'two-rupee-a-kilo' rice scheme in December 1982. The scheme is applicable to families with an annual income of less than Rs. 6,000. The scheme covers about 70 per cent of the total households and provides 5 kg. per month per person subject to a ceiling of 25 kg. per month per family at Rs. 2 per kg. The then ruling party (Telugu Desam) which was different from the one at the Centre (Congress), included the scheme in its election manifesto in 1982. Soon after it came to power, it implemented the scheme and kept the retail ration price of rice at Rs. 2 per kg. even though the open market price had been rising. Subsequently, the scheme was modified by the Congress Government in 1991. The ration price was increased to Rs. 3.50 per kg. and the ceiling was reduced to 20 kgs. per family. After its victory in the 1994 elections, the Telugu Desam Government restored the ceiling to 25 kgs. per month and the ration price was again lowered to Rs. 2 per kg. from 1 January 1995. Due to fiscal tightness, from August, 1996 the Telugu Desam Govermnent has increased the ration price to Rs. 3.50 per kg. and reduced the ceiling per family to 20 kgs. per month. Administrative Set Up In the State, the Food and Civil Supply Department looks after PDS, dealing with the issue of ration cards, commodity coverage, periodicity, ration scale, licensing of fair price shops, fixation of profit margins, transport, allotment, purchases from other States, storage, etc. The District Supply Officer at the district level and the Mandal Revenue Officer at the 'mandal' (about 20 villages) level look after the PDS. The Andhra Pradesh State Civil Supply Corporation with its branches at district and mandal levels look after the operations relating to finance, purchase, storage, transportation, and distribution of foodgrains and other items within the guidelines laid down by the State Food and Civil Supply Department. Fair Price Shops Over 37,000 fair price shops (FPS) are functioning in the State, of which over 30,000 of FPS are in rural Andhra Pradesh. On an average, there is one fair price shop for every 1,750 persons as compared to the national average of one fair price shop for every 2,000 persons. Each fair price shop, on an average, caters to 380 card-holders. The fair price shops are owned by private individuals. The fair price shop dealer is given a commission of Rs 7.38 per bag of rice (100 kg.), Rs. 4.67 per bag of sugar and 12 paise per litre of kerosene. The average gross earnings of an FPS dealer is about Rs. 1,000 per month (inclusive of free gunny bags). 85 Procurement of Rice At present, 3.5 million tonnes of rice are being procured within the State by the FCI and 2.5 million tonnes are given back to the State to meet its PDS requirement. Rice is being procured by imposing a levy on rice millers.49 The present levy policy is that for every one quintal of rice supplied to the FCI, the miller/trader is entitled to sell half a quintal outside the State and another half within the State. The rice millers in the surplus districts are readily cominng forward to sell rice to the FCI for three reasons50: (a) the difference between open market and the FCI procurement prices is very narrow; (b) the buyers get cash immediately after the transaction; and (c) they get permission to sell rice outside the State. These rice millers are not able to fully utilize their quota of open-market sales within the State as prices are not remunerative. Ration Cards and Scale of Issue The Government regarded families with an annual family income below Rs. 6000 or land up to 1.5 acres of wet land under assured water sources or 2.5 acres of wet land under other irrigation sources like tanks, wells, etc., or 3 acres of dry land under commercial crops like chilies, tobacco, etc., or 5 acres of dry land under other crops as poor and issued white cards to them. Other households were issued pink cards. In 1995, a household with a white card was entitled to 5 kgs of rice per person per month subject to a ceiling of 25 kgs. The ration price of rice was only Rs. 2 per kg. while the open market price of rice of comparable quality was Rs. 7-8 per kg in 1995. Besides rice, they are entitled to sugar and kerosene. A household with a pink card is not entitled for subsidized rice but is entitled to sugar and kerosene. Occasionally, edible oil is also supplied to the card-holders through the fair price shops. Earlier, the households were required to purchase the entire monthly rice quota in the first week of every month. Since many of the poor were daily wage earners, they could not utilize their entitlement. The Government modified this monthly distribution and allowed the rural poor to purchase their quota of subsidized rice in four installments in a month. At present, there are around 10 million white card-holders and 4.2 million pink card-holders in the State.5' The subsidized rice scheme covers nearly 70 per cent of the population while about 30 per cent of the population is estimated to be poor. Thus, the benefit of the scheme is also enjoyed by the non-poor households.52 49 The Rice procurement order says that every miller/trader who processes paddy into rice should surrender half of his produce as levy to the Government. Presently in Andhra Pradesh, movement of paddy is free within the State, but the transport of paddy from Andhra Pradesh to other States is prohibited. 50 Our field investigations in Andhra Pradesh indicate that millers/traders want the FCI to stay in spite of the controls because they tend to benefit from its operations. In the absence of the FCI, they fear that open market prices in the State would crash. 5' Venugopal (1992) has estimated that the number of white and pink cards issued exceeded the number of families by 1.5 per cent. This is due to the prevalence of bogus cards. 52 On the basis of a recent household survey of five villages, Indrakant (1995) has observed that in the developed villages all the poor households were covered but over 70 per cent of the white card holders were above the poverty line and in the backward villages 40-60 per cent of the poor households did not have ration cards and 40-60 per cent of the white card holders were non-poor. 86 PDS in Karnataka Political Economy of Subsidized Rice The present subsidized rice scheme was introduced in Karnataka by the Janata Dal Government in November 1985 soon after the implementation of the subsidized rice scheme in Andhra Pradesh. The scheme provided subsidized rice and rag to households whose annual income was below Rs. 3,500 and covered 2.3 million households. It was supervised in rural areas by elected Panchayat Raj Institutions (PRIs). This enabled the PDS to spread in rural areas. However, the Congress (I) Government, which came to power in 1989, dissolved the elected Panchayats; consequently, the PDS lost its effectiveness. In the 1995 Assembly elections, the Janata Dal Party's supplementary manifesto contained a promise of supplying subsidized rice at Rs. 3.20 per kg. to the poor. From March 1995, the Janata Dal Government has been providing 10 kg. of rice per month at Rs. 3.20 per kg. to families whose annual income is below Rs. 6,400. The scheme covered 5.17 million families involving an expenditure of Rs. 2,200 million in 1995- 96. Administrative Set Up The administrative set up is more or less similar to that of Andhra Pradesh except for the participation of the Cooperative sector. The Department of Food and Civil Supplies looks after the activities like issuing of ration cards to households, issuing of license to FPS dealers and vigilance over hoarders, etc. The Karnataka Food and Civil Supplies Corporation (KFCSC) implements the Government policies like purchasing/procuring the foodgrains and distributing the same through FPS. Besides, it also distributes wheat, sugar and kerosene oil. The cooperative sector is also involved in the distribution of foodgrains. Of the 257 wholesale licensed dealers, 130 belong to KFCSC and 127 to the Karnataka Cooperative Marketing Federation (KCMF). Foodgrains are transported to the doorsteps of Fair Price Shops at Government cost as in the case of Andhra Pradesh.53 Fair Price Shops (FPS) In 1995, there were 19,124 fair price shops in the State serving 9.83 million households. Among them, 5174 fair price shops were in urban and 13,950 in rural areas. KFCSC has opened shops in notified slum areas of Bangalore city to improve the distribution of kerosene oil to the poor people. For hilly and backward areas where necessary, mobile vans have been introduced for the distribution of foodgrains. 53 See Thimmaiah (1996). 87 Procurement of Foodgrains KFCSC and KCMF are acting as agents of FCI to procure rice, ragi, jowar and maize from the growers. Rice is procured by imposing a 25 per cent levy on the rice millers and paddy, ragi, jowar, and maize have been collected under the support price scheme by opening purchase centers. Unlike in Andhra Pradesh, there are no restrictions on inter-State movement of rice and paddy. Since it is a foodgrain deficit State, it contributes a small quantity of rice (100 tonnes per annum) to the Central pool. FCI allots to the State 126 thousand tonnes of rice per month and 30 thousand tonnes wheat per month. Ration Cards Three types of cards are in circulation in the State. They are: 1) Tri-colour Card/Green Card; 2) Saffron Color Card; and 3) Photo Card. Tri-colour Cards are issued to the families whose annual income is below Rs. 6,400 in rural areas and also to notified urban slum dwellers and weavers' families in the State. Presently, there are 5.17 million tri-colour cards in circulation. Saffron color cards are issued to the non-poor households in rural areas. Presently, 2.25 million saffron color cards are in circulation. Photo cards are issued in urban informal rationing areas54 and in towns. Under this scheme 2.40 million urban households have been covered. In all, 9.83 million households are being served by PDS in the State. There is a complex system of varying scale of ration quantity and prices for the different types of cardholders. 54 Fifty Cities with a population of 40,000 and above have been notified as Informal Rationing Areas (IR Areas). 88 Table Al.1 Issue Price of Essential Commodities (1994-95) Type of Card Item Unit Tri-colour Saffron Photo RPDS Others Rice Super fine Rs. per kg 4.45 6.45 7.05 7.05 Fine Rs. perkg 4.35 6.15 6.75 6.75 Common Rs. per kg 4.10 5.35 5.95 5.95 Wheat Rs. per kg 2.95 3.95 4.45 4.45 Sugar Rs. per kg 9.05 9.05 9.05 9.05 Kerosene Rs. per ltr 2.74 to 3.10 depending upon distance Source: Food Corporation of India. Scale of Issue Rice and Wheat Each tri-colour cardholder per month is entitled to 6 to 7 kgs of rice, 3 to 4 kgs of wheat (totally 10 kgs) at subsidized rates. If the tri-colour cardholder requires more than 10 kgs of foodgrains, he will be issued up to 20 kgs at the saffron card rate. For the saffron cardholders of ITDP and DPAP areas, the scale of issue of rice is between 10 to 15 kgs and that of wheat is between 3 to 5 kgs. In non-ITDP/DPAP areas, the scale of issue of rice is between 5 to 12 kgs per card and wheat is also distributed depending upon the availability. For photo cardholders in urban informal rationing areas and in towns the scale of issue is 4 to 6 kgs of rice and 2 to 3 kgs of wheat per card. If excess stock is available the same will be issued up to 20 kgs, to these cardholders. In the Bangalore city IR Area, the scale of issue of foodgrains depends upon the number of persons in each family. According to this, an adult is considered as two units and a child below 11 years as one unit. The scale of issue per unit is 1.5 kg of rice subject to a ceiling of 21 kgs. Wheat is being distributed at the rate of 4 kgs per card. The scale of issue will be increased as per the availability. 89 Sugar In the Bangalore I.R.Area, the scale of issue of sugar per card is 4 kgs. In other towns and I.R.Areas, the scale of issue of sugar is 3 kgs per card. In rural areas, the scale of issue is only 1.25 kgs. Kerosene Oil In the Bangalore I.R.Area, a cardholder without (LPG) Gas is entitled to 13 litres of kerosene per month. In other towns and I.R.Areas, the quota is 8 to 10 litres per card. In rural areas, the limit is 3 litres for cardholders without a gas connection and for cardholders with LPG to 2 litres per card per month. From April 1995, the Government started supplying 10 kgs of rice per month at Rs. 3.20 per kg. to tri-colour card holders, involving an additional subsidy of Rs. 1,260 million per annum. RPDS has been introduced in the State in 94 blocks consisting of 23 ITDP blocks and 71 DPAP blocks. The Central Government is issuing cereals at less than the FCI issue price by Rs. 50 per quintal. Further, the scale of issue in RPDS is higher than in the non-RPDS areas. 90 PDS in Kerala Kerala, a food deficit State, has built a comprehensive and effective PDS over the past twenty five years, which covers almost the entire population. The political conditions that ruled the State have conceived PDS as a measure of controlling foodgrain prices (Narayanan Nair and Sivanandan, 1995). PDS in Kerala entirely depends on the Central allocation. There is no procurement of foodgrains in the State. Though the per capita availability of rice from internal production has declined over time, the increased availability through PDS has resulted in an increase in the per capita availability of rice in the State. Administrative Set Up The role of the State in the movement of foodgrains is limited. The Food and Civil Supplies Department authorizes certain wholesale dealers and retailers to move the foodgrains. Kerala Food and Civil Supplies Corporation sells various other items through its retail outlets. The authorized wholesale dealers lift the stocks from the specified FCI sub-depots to their godowns, from where the authorized retail dealers move the material to their shops. The dealers are allowed a commission for performing this function, which varies from item to item. It also depends upon the distance between godown/shop and FCI depots. These commissions are revised from time to time. Earlier the commissions were fixed in absolute terms. On the request of the dealers, presently, the commissions are fixed as a percentage of the Central Issue Price of rice and wheat. Hence, the amounts of commission are automatically revised whenever the Central Issue Price is revised. The commission of a wholesale dealer for common variety of rice in non-RPDS areas ranges from 2 to 2.52 per cent. Wholesale dealers, whose godowns are at a distance of over 64 km from FCI depots are compensated by an additional commission at the rate of 0.011 per cent per kilometer. Likewise, the commission to the authorized retail dealer ranges between 2.8 and 2.32 per cent. Retail dealers who are located farther are provided additional commission. In the case of wheat, the wholesale dealers' commission ranges between 2.65 and 3.34 per cent while that of retail dealers between 2.36 and 2.71 per cent. For dealers operating from a distance of over 64 kms, additional commission at the rate of 0.015 per cent per km is allowed. Coverage In 1991, PDS covered over 90 percent of the population. The retail outlets of PDS have reached even the remotest areas in the State. Around 13,700 retail shops cater to 5.53 million ration cards, and 92 per cent of them are run by private dealers and the rest by co-operatives. These retail shops are served by 307 wholesale shops. Out of these, 83 percent is owned by licensed private agencies and the rest by the co-operatives. FCI has twenty sub-depots in the State for supplying ration commodities to the wholesale shops. The licensed wholesale and retail 91 agencies get conunission fixed by the State Government from time to time, and work under the supervision of the Civil Supplies Department. Scale of Ration The commodities distributed regularly through PDS are rice, wheat, sugar and kerosene. Households producing sufficient rice will not be entitled to rice ration. The State fixes the scale of issue of rice after taking into account internal production and its requirement. The rice entitlement per day per adult unit, with some fluctuations, has increased from 130 gms. in 1976 to 220 gins. in 1995 (Narayanan Nair and Sivanandan, 1995)." For other items, the scale has remained unchanged. The State has been lifting almost fully the allotted quota of rice. In the case of wheat, the off-take was very much below the allotment in the initial years. But in recent years, the off-take has matched with the allotment. During the sixties and the seventies the off-take of wheat from the PDS was inversely related to the off-take of rice. But with the substantial improvement in the availability of rice, such a relationship has disappeared in the 1980s.56 5 The difference between open market price and ration price, although it has narrowed down in the recent period, is still not negligible. The ration price of rice was Rs. 6.50 kg. while the open market price was Rs. 8.50 per kg. However, the quality of the PDS rice is inferior. Hence, off-take of rice is lower than the entitlement in the recent period (see Narayanan Nair and Sivanandan, 1995). 56 It is worth noting that the people of Kerala, over a period of time, have incorporated wheat into their dietary pattern. 92 PDS in Uttar Pradesh Uttar Pradesh is a major contributor of foodgrains to the Central pool. Although a system of informal rationing in respect of rice, wheat, sugar, and kerosene exists in Uttar Pradesh through a network of fair price shops, the distribution of subsidized foodgrains is on a meager scale and PDS supplies are mostly confined to areas which are deficit in cereal production or have weak marketing facilities. However, PDS is being used by all the people for sugar in all the areas. Administrative Set Up The Food and Civil Supplies Department looks after the PDS, dealing with issue of ration cards, commodity coverage, ration scale, licensing of fair price shops, movement of foodgrains, storage etc. Unlike in Andhra Pradesh and Karnataka, fair price shops (FPS) have to lift their monthly quota from the godowns of FCI or the godowns of the Food and Civil Supplies Department.57 The Provincial Cooperative Federation (PCF) supplies sugar. The FPS owners lift their monthly quota from the godowns of PCF. Oil companies through wholesale dealers supply kerosene oil to FPS. Procurement There are various procurement agencies operating in Uttar Pradesh. FCI directly purchases from the farmers of road-side villages. It authorizes the Food and Civil Supplies Department, Uttar Pradesh Essential Commodities Corporation58 and Producer Co-operative Societies to procure wheat from the farmers. It is not uncommon to see 2 or 3 agents of FCI operating in the same village. Generally, the Producer Cooperative Societies, because of their network and indirect control over farmers, are able to procure a substantial quantity of wheat. Ration Cards and Scale of Issue A system of informal rationing in respect of rice, wheat and sugar operates in Uttar Pradesh through a net-work of around 55,700 FPS. Ration card, which is issued to every household, is required for making purchases from FPS.59 57 The fair price shops are unviable in Uttar Pradesh because of lack of proper transport arrangement between FCI godowns and FPS as well as low volume of foodgrains turnover. 5S The Uttar Pradesh Essential Commodity Corporation works under the administrative control of the Civil Supplies and Food Department. It was set up to supplemeiit the PDS by procuring, storing and selling essential commodities like vegetable oil, imported sugar, rice, clothes, gas, tea and pulses. 59 There were 27.4 million card holders in Uttar Pradesh in 1990; about 16 per cent of the cards were bogus (ryagi, 1991). 93 Table A1.2 Entitlement of Essential Items Item Unit Plain Regions Hilly Regions Urban Rural Urban Rural Rice: Per Adult Unit 6 kgs 6 kgs 8 kgs 8 kgs Wheat Per Adult Unit 8 kgs 8 kgs 10 kgs 10 kgs Sugar Per Adult Unit 1 kg 1/4 kg 1 kg 1 kg Kerosene Per Card 10 Its 5 ltr 10 ltr 5 ltr Source: Government of Uttar Pradesh. The scale of issue in Uttar Pradesh is high but the per month allotment (and the lifting) of rice and wheat to the State is as low as 5 per cent of the total requirement on the basis of the households entitlement. Due to the narrow difference between open-market price and retail ration price, the actual off-take is much lower than the allotment. The monthly allotments of sugar and kerosene to the State are 53,000 tonnes and around 75,000 litres, respectively. With these allotments, the State Government cannot meet the full entitlement of all the cardholders. Controlled Cloth Janata Cloth is being supplied by the Government of India for the benefit of poor people. Of the total controlled cloth (113 million meters) allotted to the State, the Uttar Pradesh Government Employees Welfare Corporation handles 75 per cent and the remaining 25 per cent is taken care of by the Uttar Pradesh Essential Commodities Corporation. The controlled cloth is made available to the consumers through the network of sale points of these two agencies. Imported Edible Oils The two agencies, viz., the Uttar Pradesh Essential Commodities Corporation and the Uttar Pradesh Government Employees Welfare Corporation distribute the imported edible oils in the ratio of 70 and 30 per cent respectively. These two agencies lift the edible oils from the State Trading Corporation of India and Hindustan Vegetable Oil Corporation godowns and supply the same to FPS in rural and urban areas. The demand for imported oil is generally low with the result that against an allocation of 48,000 tonnes the actual off-take was less than 25,000 tonnes. 94 REFERENCES Acharya, S.S. (1989), Socio-Economic Impact of 1987 Drought in Rajasthan, Department of Agricultural Economics, Rajasthan Agricultural University, Udaipur. Ahmned, Akhtar and K. Billah. 1994. "Food for Education in Bangladesh: An Early Assessment," International Food Policy Research Institute, Washington, D.C., unpublished. Ahmed, Raisuddin. 1988. "Structure, Costs, and Benefits of Food Subsidies in Bangladesh." in Food Subsidies in Developing Countries: Costs, Benefits, and Policy Options, Pinstrup-Andersen, ed., Johns Hopkins University Press, Baltimore. Alderman, H. 1988. "The Twilight of Flour Rationing in Pakistan," Food Policy. Alderman, H. 1991. "Food Subsidies and the Poor." in G. Psacharopoulos, ed. Essays on Poverty and Equity and Growth. Bureau of Industrial Costs and Prices. 1991. Report on the Operations of The Food Corporation of India, Ministry of Industry, 1991. Deshpande, Vasant. 1992. Islands of Development: Role of Voluntary Action in Rural Transformation, Published by Raju Deshpande, Madhav Baug Housing Society Ltd, Pune. Dev, Mahendra S., and Suryanarayana, M.H. 1992. "Is PDS Urban Biased and Pro-Rich? An Evaluation," Economic and Political Weekly, Vol. XXVI, No. 41. Edirisinghe, N. 1987. The Food Stamp Scheme in Sri Lanka: Costs, Benefits, and Options for Modifications. Research Report 58. International Food Policy Research Institute, Washington, D.C. Galab, S. 1987. Effectiveness of Integrated Rural Development Programme in a Drought Prone Region: The Case of Anantpur, Center for Economic and Social Studies, Hyderabad. Government of India. 1997. Focus on the Poor: Guidelines for the Implementation of the Targeted Public Distribution System. Ministry of Civil Supplies, Consumer Affairs and Public Distribution. New Delhi, India. Grosh, M. 1994a. "Administering Targeted Social Programs in Latin America - From Platitudes to Practice." World Bank, Washington, D.C. Grosh, M. 1994b. "Administering Targeted Social Programs in Latin America." World Bank Regional and Sectoral Studies Series. 95 Howes, Stephen and Tha S. 1992. "Urban Bias in the Public Distribution System," Economic and Political Weekly, Vol. XXVII, May 9. Horton, S. 1992. "Unit Costs, Cost-Effectiveness and Financing of Nutrition Interventions." Policy Research Working Paper, WPS 952. World Bank, Washington D.C. Jha, S. 1991. Consumer Subsidies in India: Is Targeting Effective? Indira Gandhi Institute of Development Studies, Bombay. Kabra, K.N. and Ittyerah, A.C. 1996. The Public Distribution System: A Report of Target Group Orientation and the Viability of Retail Outlets, Indian Institute of Public Administration, New Delhi. Kennedy, E.T. and H. Alderman. 1987. "Comparative Analyses of Nutritional Effectiveness of Food Subsidies and Other Food-Related Interventions." International Food Policy Research Institute and Joint WHO-UNICEF Nutrition Support Programme, Washington D.C. Kurien, N.J. and P.V. Rajeev. 1996. "Self-Employment and Capacity Enhancing Programmes," Indian Journal of Labour Economics, Vol. 39 No.1, January-March. Mooij, Jos E. 1992. About Rationcards and Policy Rationalization, (Mimeo). Mooij, Jos E. 1994. "Public Distribution System as Safety Net. Who is Saved?" Economic and Political Weekly. Musgrove, Philip. 1991. "Feeding Latin America's Children, An Analytical Survey of Food Programs." Latin America and the Caribbean Technical Department, Regional Studies Series Report, No. 11. World Bank, Washington, D.C. Narayanan, K. and P. Sivanandan. 1995. Food Security and Public Distribution System in Kerala, Research Project on Human Development, Center for Development Studies, Thiruvananthapuram. Onkamath, G. 1993. Rural Labour Processes: A Case Study in Coastal Andhra, Unpublished Ph.D. thesis, Jawaharlal Nehru University, New Delhi. Pandey, S.R. 1991. Community Action for Social Justice: Grassroots Organization in India, Sage Publications, New Delhi. Parikh, K.S. 1994. "Who Gets How Much from PDS: How Effectively Does it Reach the Poor," Sarvekshna, Vol. XVII No. 3, January-March. 96 Prahlad Rao, N. 1990. "Constraints in Nutritional Programme," IICW New Bulletin, Vol XXIII, No. 4. Rao, V.M. 1996. "Food Security in a Liberalized Region," Invited Paper, 32 Annual Conference, Indian Econometric Society, Bangalore, March. Radhakrishna R. and S. Indrakant. 1987. Effects on Rice Market Intervention Policies in India: The Case of Andhra Pradesh. Center for Economic and Social Studies, Hyderabad. Radhakrishna, R. and C. Ravi. 1990. Food Demand Projections for India, Center for Economic and Social Studies, Hyderabad. Radhakrishna R. and C. Ravi. 1992. "Effects of Growth, Relative Price and Preferences on Food and Nutrition," Indian Economic Review, Volume 27, Special Number. Radhakrishna, R. and K.V. Narayana. 1993. Nutritional Programs in India: Review and Assessment, Center for Economic and Social Studies, Hyderabad. Radhakrishna, R. 1996. "Food Trends, Public Distribution System and Food Security Concerns," Indian Journal of Agricultural Economics. Volume 15, Nos. 1 and 2, Jan-June. Sahn, D. and H. Alderman. 1995. "The Effect of Food Subsidies on Labor Supply in Sri Lanka," in D. van de Walle and K. Nead (eds.), Public Spending and the Poor, Baltimore and London, The Johns Hopkins University Press. Shekar, M. 1994. "The Tamil Nadu Integrated Nutrition Project (TINP-1) Revisited: An Evaluation Perspective," OED, The World Bank, Washington, D.C. Subbarao, K., A. Bonnerjee, J. Braithwaite, S. Carvalho, K. Ezemenari, C. Graham and A. Thompson. 1997. Safety Net Programs and Poverty Reduction: Lessons from Cross- Country Experience. Directions in Development Series. The World Bank, Washington, D.C. Subbarao,K, et. al. 1996. Social Assistanlce and Poverty-targeted Programs: A Sourcebook. Poverty and Social Policy Department. The World Bank, Washington, D.C. Subbarao, K., A. Ahmed and T. Teklu. 1995. "Philippines Social Safety Net Programs: Targeting, Cost Effectiveness and Options for Reform," Discussion Paper No. 317, World Bank, Washington, D.C. Subbarao, K. 1989a. "Improving Nutrition in India," World Bank Discussion Paper No.49. . 1989b. Agricultural Marketing and Credit: A Survey of Research. New Delhi: Indian Council of Social Science Research. 97 _ 1985. "Regional Variations in the Impact of Anti-Poverty Programs," Economic and Political Weekly. October. Thimmaiah. 1996. "The Political Economy of Populist Programmes," Journal of Indian School of Political Economy, Vol. 8 No.1, January-March. Tyagi, D.S. 1990. Managing India's Food Economy: Problems and Alternatives, Sage Publications, New Delhi. Tyagi, B.N. 1991. Public Distribution in Uttar Pradesh, Center for Advanced Development Research, Lucknow. Venugopal, K. R. 1992. Deliverance from Hunger. The Public Distribution System, Sage Publications, New Delhi. 98 Distributors of COLOMBIA GERMANY ISRAEL NEPAL PORTUGAL SWEDEN Infoenlace Ltda. UNO-Vedag Yozmot Lderature Ltd. Everest Media Intemational Services (P) Ltd. Livrada Potugal Wennergren-Williams AB W orld Bank Carrera 6 No. 51-21 Poppelsdorfer Allee 55 PO. Box 56055 GPO Box 5443 Apartado 2681, Rua Do Carmo 70-74 P0. Box 1305 Apartado Aereo 34270 53115 Bonn 3 Yohanan Hasandlar Street Kathmandu 1200 Lisbon S-171 25 Solna Publications Santalf de BogotA, D.C. rel: (49 228) 949020 Tel Aviv 61560 Tel:(9771)472152 Tel:(1)347-4982 Tel:0(46 ) 705-97-50 Prices and credit terns varyfrom Tel: (57 1) 285-2798 Fax: (49 228) 217492 Tel: (972 3) 5285-397 Fax: (977 1) 224 431 Fax: (1) 347.0264 Fax: (46 8) 27-00-71 country to contry. Consult your Fax: (571)285-2798 URL: hhtpilwwwuno-ver1ag.de Fax: (9723)5285-397 E-mail:mail@wwi.se local distribuitor before placitig an COTE DIVOIRE E-mail: unoveldag@aol.com NETHERLANDS ROMANIA order. COTE D'IVOIRE R.O.Y Intemational D~~~~~~~~~~~~~~~~~~~e LindeboornilnOr-Publikaties Comparsi De Librarii Bucsresti S.A. SWITZERLAND order. Center d'Edition et de Dlfusion Africaines GREECE PO Box 13056 PO. Box 202, 7480 AE Haaksbergen Sir. Lipsocani no. 26, sector 3 Librairie Payot Service InstRutionnel ARGENTINA (CEDA) Papasotiriou S.A. Tel Aviv 61130 Tel: (31 53) 574-0004 Bucharest Cotes-de-Montbenon 30 ARGENTINA del Libro Intemacionl04 B.P 541 35, Stoumara Str. Tel: (9723)5461423 Fax: (31 53) 572-9296 Tel: (40 1)613 9645 1002 Lausanne Oliiavdel Liro Int1macional Abidjan 04 10682 Athens Fax: (972 3) 5461442 E-mail: lindeboo@worldonline.nl Fax: (40 1)312 4000 Tel: (41 21) 341-3229 1120 Buenos Aires Tel: (225)246510:246511 Tel: (30 1)364-1826 E-mail: royit§nelvision.net.it URL:http://www.wordonline.nV-lindeboo Fax: (41 21) 341-3235 Tel: (54 12B 815o8354 Fax: (225) 25 0567 Fax: (30 1) 364-8254 RUSSIAN FEDERATION Tel: (54 t() 815-8354 Faa: (225(250567 Fan:(301(364-8254 Palestinian Authoaty/Middle East NEW ZEALAND lsdatelsivo cVes Mir> ADECO Van Diermen EddionsTechniques Fax: (541) 815-8156 CYPRUS HAITI Index Inlormation Services EBSCO NZ Ltd. 9a, Lolpachniy Pereulok Ch. de Lacuez 41 AUSTRALIA, FIJI, PAPUA NEW GUINEA Ceenter frApplied Research CuhureDifusion O.8. 19502 Jernsalem Private Mail Bag 99914 Moscow 101831 CH1807Bbonay SOLOMON ISLANDS, VANUATU, AND 'Cyprus Cottage 5, Rae Capois Tel: (972 2(6271219 New Market Tel: (7 095) 917 87 40 Tel: (41 21) 043 2673 WESTERN SAMOA 6, Diogenes Street, Engomi C.P257 Fax: (972 2) 6271634 Auckland Fax: (7095)9179259 Fax: (41 21) 9433605 D.A. Information Se.vjces PO. Box 2006 Por-au-Prince Tel: (64 9) 524-8119 D.A. WormationrServices Nicosia Tel: (509)239260 ITALY Fax: (649)524-8067 SINGAPORE, TAIWAN, THAILAND 6 rieorse 3 o Tel: (357 2) 44-1730 Fax: (509) 23 4858 Licosa Commissionaria Sansoni SPA MYANMAR, BRUNEI Central Books Distribution Michra 3132 Fax: (357 2) 46-2051 Via Duca Di Calabna, 1/1 NIGERIA Asahgate Publishing Asia Pachic Pte. Ltd. 300 Silom Road Vlctora 3 9210 7777 HONG KONG, MACAO Casella Postale 552 University Press Limled 41 Kallang Pudding Road #04-03 Bangkok 10500 Fa: (61) 3 9210 7738 CZECH REPUBLIC Asia 2000 Ltd. 50125 Firenze Three Crowns Building Jericho Golden Wheel Building Tel: (66 2) 235-5400 E-mail: service ( dacirectcomau National Information Center Sales & Circulalion Department Tel: (55) 645-415 Private Mail Bag 5095 Singapore 349316 Fax: (66 2) 237-8321 URL: hfp://wwvv.dadirect.corv.au prode)na, Konviktaka 5 Seabird House, aflit 1101-02 Fax: (55) 641-257 lbadan Tel: (65) 741-5166 CS -11357 Prague 1 22-28 Wyndhdm Street, Central E-mail: licosa64 tbcc. Tel: (234 22) 41-1356 Fax: (65) 742-9356 TRINIDAD & TOBAGO, AUSTRIA Tel: (42 2) 2422-9433 Hong Kong URL: hfipixwww.ftbcc.iUlicosa Fax: (234 22) 41-2056 E-mail: ashgate@asianconnecl.com AND THE CARRIBBEAN Gerold and Co. Fan: (42 2) 2422-1484 Tel: (852) 2530-1409 Systematics Studies Uni Weibbrggasse 26 URL: htp://www .nis.cz/ Fax: (852) 2526-1107 JAMAICA NORWAY SLOVENIA 9 Wafs Street E-mail: sales@ asia2000.com.hk tan Randle Publishers Ltd. NIC Info AVS Gospodarski Vestnik Publishing Group Curepe A-1(01-17W1en DENMARK URL:h9p.ilwww.asia2000.com.hk 20601 dHope Road, Kingston 6 Book Department, Postboks 6512 Elerstad DunajskacestaS Trinidad, West Ilndies Tel: (431) 512-47-31-029 SamfundsLitteratur Tel: 876-927-2085 N-0606 Oslo 1000 Ljubljana Tel: (809) 662-5654 URL htp1/vwwgeoIdco)'etonlne RosenoemsnAld I11 HUNGARY Fax: 876-977-0243 Tel: (47 22) 97-4500 Tel: (386 61( 133 83 47;132 12 30 Fan: (809) 662-5654 URL: h p-J/www .gerold.co/al.online DK-1970 Frederiksberg C Euro Into Service E-mail: irpl@cntiso.om Fan: (47 22) 974545 Fax: (386 61) 133 80 30 E-mail: tobettnidad.net BANGLADESH Tel: (45 31) 351942 Margaszgeti Europa Haz E-mail: repansekj@gvestnik.si Micro Industries Development Fax: (45 31) 357822 N-1 138 Budapest JAPAN PAKISTAN UGANDA Assistance Society (MIDAS) URL: hftr:1/www.sl.cis Tel: (3601)111I 6061 Eastemn Book Service Mirzs Book Agency SOUTH AFRICA, BOTSWANA Gustro Ltd. House 5, Road 16 Fax: (361)3025035 3-13 Hongo 3-chome, Bunkyo-ku 65, Shahrah-e-Quaid-e-Azam For single tines: PO Box 9997, Madhvani Building Dhannondi R5, rea 6ECUADOR E-mail: euroinfo@mail.matav.hu Tokyo 113 Lahore 54000 Oxford Universiy Press Southem Africa Plot 16/4 Jinja Rd. Dhaka 1209 Libn Mundi Tel: (81 3) 3818-0861 Tel: (92 42) 735 3601 Vasco Boulevard, Goodwood Kampala Tel: (880 2(326427 Libreria Intemacional INDIA Fax: (813)(318-0864 Fax: (92 42) 576 3714 PO. Box 12119, Ni City 7463 Tel: (256 41) 251 467 Fan: (880 0)11188 PO. Box 17-01-3029 Allied Publishers Ltd. E-mail: orders@ svt-ebs.co.jp Cape Town Fax: (256 41) 251 468 Juan Leon Mera851 751 Mount Road URL:htpJv/wwwv.bekkoame.orjp/-svt-ebs Oxford Univenly Press Tel: (27 21)595 4400 E-mail: gus@swifluganda.com BELGIUM QulO Madras - 600 002 5 Bangalore Town Fax: (27 21) 595 4430 Jean DL Tel: (593 2) 521-606; (593 2) 544-185 Tel: (91 44) 852-3938 KENYA Sharae Faisal E-mail: ofxord@oup.co.za UNITED KINGDOM Lea u onnoy Fax: (593 2) 504-209 Fax: (91 44) 852-0649 AMrica Book Sexrre (E.A.) Ltd. PO Box 13033 Microinlo Ltd. Ax. do Roi 202 E-mail: librimu1 @librimundi.com.ec Quaran House, Mfangano Street Karachi-75350 Fbr subscription orders: PO. Box 3, Ahon, Hampshire GU34 2PG Tel: (32n2s538-516 E-mail: ibrimu2@1ibrimundi.oom.ec INDDNESIA PO. Box 45245 Tel: (92 21) 446307 Intemational Subscription Service England Fax: (32 2) 538-0416 Pt. Indira Limited Nairobi Fax: (92 21) 4547640 PO. Box 41095 Tel: (44 1420) 8848 Fox: (322(538-0841 EGYPT, ARAB REPUBLIC OF Jalan Borobudur 20 Tel: (254 2) 223 641 E-mail: ouppak@TheOOfice.net Craighall Fax: (44 1420) 89889 BRAZIL Al Ahram Disitbution Agency PO. Box 181 Fax: (254 2) 330 272 Johannesburg 2024 E-mail: wbank@ukminfo.demon.co.uk Publicacres Tecnicas lItemacionais da. AlGalaa Street Jakara 10320 Pak Book Corporation Tel: (2711)880-1448 URL:hftp://www.microinfo.co.uk Rua Peixoto Gomidv, 209 Ld. Cairo Tel: (62 21( 390-4290 KOREA, REPUBLIC OF Aziz Chambers 21, Queen's Rood Faa: (27 11( 880-6248 01489 Saxo Pob, SP. Tel: (20 2) 578-6083 Fax: (62 21) 3904289 Daejon Trading Co. Ltd. Lahore E-mail: iss@is.co.za VENEZUELA Tel: (55 11) 259-664 Fax: (20 2) 578-6833 PO. Box 34, Youida, 706 Seoun Bldg Tel: (92 42) 636 3222; 636 0885 Tecni-Ciencia Libros, S.A. Fan: (5511)258-6990 IRAN 44-6 Youido-Dong, Yeongchengpo-Ku Fax: (92 42) 636 2328 SPAIN Centro Cuidad Comercial Tamanco E-mail. poStmasterttimol ) b5 The Middle East Observer Ketab Sara Co. Publishers Seoul E-mail: pbc@brain.net.pk Mundi-Prensa Libros, S.A. Nivel C2, Caracas E-mal: pstmaster@pti.uol.br 41, Sherid Street Khaled Eslamboli Ave., 6th Street Tel: (822) 785-1631/4 Castelb 37 Tel: (58 2) 959 5547; 5035; 0016 UR.: hgp:llwww .uot.br Cairo Delafmooz Alley No. 8 Fax: (82 2) 784-8315 PERU 28001 Madrid Fax: (58 2) 959 5636 CANADA Tel: (20 2) 393-9732 PO. Box 15745-733 Ediorial Desarrollo SA Tel: (341) 431-3399 RenoAf Publishing Cs, Ltd. Fax: (20 2) 393-9732 Tehran 15117 MALAYSIA Apartado 3024, Lima 1 Fax: (34 1) 575-3998 ZAMBIA 5369 Canotek Road Tel: (98 21) 8717819, 8716104 University of Malaya Cooperative Tel: (51 14) 285380 E-mail: libreria@mundiprensa.es Universty Bookshop, University of Zambia GOtawa, Ontario K1J 9,J3 FINLAND Fan: (98 21) 8712470 Bookshop,l-imited Fan: (51 14) 286628 URL: hfip:/Iwww .mundiprensa.es/ Great East Rood Campus Tel:561369 7452665 RAkateeminen Kirjakauppa E-mail: kelab-sara@neda.net.ir Po. Box 1127 PO. Box 32379 Fax: (613) 745-7660 PO. Box 128 Jalan Pantai Baru PHILIPPINES Mundi-Prensa Barcelona Lusaka E-muil: rdor .dept- renufbooks.com FIN-00101 Helsinki Kowkab Publishers 59700 Kuala Lumpur Intemational Booksource Center Inc. Consell de Cent, 391 Tel: (260 1) 252 576 UAL: htnp:/rwww.renounbooks.com Tel: (3580)121 4418 PO. Box 19575-511 Tel: (60 3)756-5000 1127-A Antipolo St, Barangay, Venezuela 08009 Barcelona Fax: (2601)253952 rFax: (3580)121-4435 Tehran Fax: (603) 755-4424 Makati City Tel: (343)488-3492 CHINA E-mail: akatilaus@stockmann.ti Tel: (98 21) 258-3723 Tel: (63 2) 8966501; 6505; 6507 Fax: (343) 487-7659 ZIMBABWE China Financial & Economic URL: htp:llwww .akateeminen.comx Fax: (98 21) 258-3723 MEXICO Fax: (63 2) 896 1741 E-mail: barcelona@ mundiprensa.es Longman Zimbubwe (Pvt.)Ltd. INFOTEC Tourle Road, Ardibennie Publishing House FRANCE IRELAND Av. San Femando No. 37 POLAND SRI LANKA, THE MALDIVES PO. Box ST125 8, Da Fo Si Dong Jie Wodrid Bank Publications Government Supplies Agency Col. Tondello Guerra International Publishing Service Lake House Bookshop Southerton Beijing Tel: (86 10) 6333-8257 66, venue dIsna Oitig an tSolathair 14050 Mexico, D.F Ul. Piekna 31/37 100, Sir Chiltampalam Gardiner Mawatha Harare Fax: (86 10) 6401 5 75116 Paris 4-5 Harcoud Road Tel: (52 5) 624-2800 00-677 Warzawa Colombo 2 Tel: (263 4) 6216617 rax:(8610)6401-7365 Tel: (331)40-69-30-56/57 Dublin 2 Fax (525)624-2822 Tel: (402)628-6089. Tel: (941)32105 Fax: (2634)621670 Fax, (33 1)40-69-30-68 Tel: (353 1)661-3111 E-mail: infotec@rn.net.mx Fax: (48 2)621-7255 Fax: (941)432104 Fax: (353 1)475-2670 URL: htp:lr1n.net.mx LE-mal: boohs%ipscikp.awt.pVipl E-mail: LHL@sri.lanka.net URL: hotp:/fwwwrv.ipscg.waw.pVips/exporslWi~ Recent World Bank Discussion Papers (continued) Staple No. 342 Economic Integration and Trade Liberalization in Southern Africa: Is There a Rolefor South Africa? Merle Holden No. 343 Financing Private Infrastructure in Developing Countries. David Ferreira and Karman Khatami No. 344 Transport and the Village: Findings from African Village-Level Travel and Transport Surveys and Related Studies. Ian Barwell No. 345 On the Road to EU Accession: Financial Sector Development in Central Europe. Michael S. Borish, Wei Ding, and Michel Noel No. 346 Structural Aspects of Manufacturing in Sub-Saharan Africa: Findingsfrom a Seven Country Enterprise Survey. Tyler Biggs and Pradeep Srivastava No. 347 Health Reform in Africa: Lessonsfrom Sierra Leone. Bruce Siegel, David Peters, and Sheku Kamara No. 348 Did External Barriers Cause the Marginalization of Sub-Saharan Africa in World Trade? Azita Amjadi Ulrich Reincke, and Alexander J. Yeats No. 349 Surveillance of Agricultural Price and Trade Policy in Latin America during Major Policy Reforms. Alberto Valdes No. 350 Who Benefitsfrom Public Education Spending in Malawi: Resultsfrom the Recent Education Reform. Florencia Castro-Leal No. 351 From Universal Food Subsidies to a Self-Targeted Program: A Case Study in Tunisian Reform. Laura Tuck and Kathy Lindert No. 352 China's Urban Transport Development Strategy: Proceedings of a Symposium in Beijing, November 8-10, 1995. Edited by Stephen Stares and Liu Zhi No. 353 Telecommunications Policiesfor Sub-Saharan Africa. Mohammad A. Mustafa, Bruce Laidlaw, and Mark Brand No. 354 Saving across the World: Puzzles and Policies. Klaus Schmidt-Hebbel and Luis Serven No. 355 Agriculture and German Reunification. Ulrich E. Koester and Karen M. Brooks No. 356 Evaluating Health Projects: Lessonsfrom the Literature. Susan Stout, Alison Evans, Janet Nassim, and Laura Raney, with substantial contributions from Rudolpho Bulatao, Varun Gauri, and Timothy Johnston No. 357 Innovations and Risk Taking: The Engine of Reform in Local Government in Latin America and the Caribbean. Tim Campbell No. 358 China's Non-Bank Financial Institutions:Trust and Investment Companies. Anjali Kumar, Nicholas Lardy, William Albrecht, Terry Chuppe, Susan Selwyn, Paula Perttunen, and Tao Zhang No. 359 The Demandfor Oil Products in Developing Countries. Dermot Gately and Shane S. Streifel No. 360 Preventing Banking Sector Distress and Crises in Latin America: Proceedings of a Conference held in Washington, D.C., April 15-16, 1996. Edited by Suman K. Bery and Valeriano F. Garcia No. 361 China: Power Sector Regulation in a Socialist Market Economy. Edited by Shao Shiwei, Lu Zhengyong, Norreddine Berrah, Bernard Tenenbaum, and Zhao Jianping No. 362 The Regulation of Non-Bank Financial Institutions: The United States, the European Union, and Other Countries. Edited by Anjali Kumar with contributions by Terry Chuppe and Paula Perttunen No. 363 Fostering Sustainable Development: The Sector Investment Program. Nwanze Okidegbe No. 364 Intensified Systems of Farming in the Tropics and Subtropics. J.A. Nicholas Wallis No. 365 Innovations in Health Care Financing: Proceedings of a World Bank Conference, March 10-11, 1997. Edited by George J. Schieber No. 366 Poverty Reduction and Human Development in the Caribbean: A Cross-Country Study. Judy L. Baker No. 367 Easing Barriers to Movement of Plant Varieties for Agricultural Development. Edited by David Gisselquist and Jitendra Srivastava No. 368 Sri Lanka's Tea Industry: Succeeding in the Global Market. Ridwan Ali, Yusuf A. Choudhry, and Douglas W. Lister No. 369 A Commercial Bank's Microfinance Program: The Case of Hatton National Bank in Sri Lanka. Joselito S. Gallardo, Bikki K. Randhawa, and Orlando J. Sacay No. 370 Sri Lanka's Rubber Industry: Succeeding in the Global Market. Ridwan Ali, Yusuf A. Choudhry, and Douglas W. Lister No. 371 Land Reform in Ukraine: The First Five Years. Csaba Csaki and Zvi Lerman No. 373 A Poverty Profile of Cambodia. Nicholas Prescott and Menno Pradhan No. 374 Macroeconomic Reform in China: Laying the Foundation for a Socialist Economy. Jiwei Lou No. 376 Poverty, Social Services, and Safety Nets in Vietnam. Nicholas Prescott No. 377 Mobilizing Domestic Capital Marketsfor Infrastructure Financing: International Experience and Lessonsfor China. Anjali Kumar, R. David Gray, Mangesh Hoskote, Stephan von Klaudy, and Jeff Ruster (Am THE WORLD BANK ISIS Ii St \.. \\. \\;ishliigot nl 1).( 21l13, t S Iclcph)L,,: 2012-477-1234 I:-, iIiIc: 21)2-,,I 7 d II 1 H v [ i 0 \ K it I I I I t -,,,.,il: (R CIII t, I I C 1 , '_ L tl t I C t I1 I I I1 09 \ ,IlldthL lIII \,SiI A Rtil. A I')Ltt\CI1)1CI SLCtill .tTl 9 780821 340905 ISBN 0-82134090-5