India - Achievements and challenges in reducing poverty

Report No. 16483-IN India Achievements and Challenges in Reducing Poverty May 27, 1997 Country Operations, Industry & Finance Division Country Department II South Asia Region Document of the World Bank CURRENCY Rs/USS Currency Of IWilkil Market Prior to June 1966 4.76 June 6, 1966 to mid-December 1971 7.50 Mid-December 1971 to end-June 1972 7.28 1971-72 7.44 1972-73 7.71 1973-74 7.79 1974-75 7.98 1975-76 8.65 1976-77 8.94 1977-78 8.56 1978-79 8.21 1979-80 8.08 1980-81 7.89 1981-82 8.93 1982-83 9.63 1983-84 10.31 1984-85 11.89 1985-86 12.24 1986-87 12.79 1987-88 12.97 1988-89 14.48 1989-90 16.66 1990-91 17.95 1991-92 24.52 1992-93 26.41 30.65 1993-94 31.36 1994-95 31.40 1995-96 33.46 1996-97 35.50 Note: The Indian fiscal year runs from April I through March 31. Source: IMF, International Finance Statistics (IFS), line "rf"; Reserve Bank of India. 'A dual exchange rate system was created In March 1992, with a free market for about 60 percent of foreign exchange transactions. The exchange rate was reunified at the beginning of March 1993 at the free market rate. Vice President Mieko Nishimizu Directors Robert S. Drysdale, Edwin Lim Division Chief Luis E. Derbez Staff Member Zoubida Allaoua, Senior Economist  ABBREVIATIONS AND ACRONYMS AERC Agricultural Economics Research Center ANM Auxiliary nurse Midwife APP Anti-Poverty Program AWW Anganwadi Worker BRAC Bangadesh Rural Advancement Committee CEM Country Economic Memorandum CPIAL Consumer Price Index for Agricultural Labourers CPIIW Consumer Price Index for Industrial Workers DPEP District Programme for Elementary Education EAS Employment Assurance Scheme EGS Employment Guarantee Scheme FCI Food Corporation of India GDP Gross Domestic Product GOI Government of India HIV Human Immune Virus HYV High Yielding Varieties ICDS Integrated Child Development Services ICRISAT International Crops Research Institute for the Semi-Arid Tropics IMR Infant Mortality Rate IRDP Integrated Rural Development Program JRY Jowahar Rojgar Yojana NAS National Accounts Statistics NCAER National Council of Applied Economic Research NGO Non-Governmental Organization NRY Nehru Rozgar Yojana NSS National Sample Survey NSSO National Sample Survey Organization OXFAM Oxford Committee for Famine Relief PDS Public Distribution System PRIs Panchayati Raj Institutions SC/STs Scheduled Castes/Scheduled Tribes SDP State Domestic Product SEWA Self-Employed Women's Association SHGs Self-Help Groups TB Tuberculosis TPDS Targeted Public Distribution System WDR World Development Report u Contents Currency Abbreviations and Acronyms Acknowledgments Economic Development Data Executive Summary Chapter 1 Poverty: Where does India stand? 1 A. Introduction 1 Poverty measurement 2 B. Evolution of poverty in India 3 How much progress has India made in fighting poverty since the 1950s? 3 Regional poverty incidence 8 C. Who are the poor? 10 Access to land, sources of income and poverty 10 Caste, ethnicity, religion and poverty 15 Demographics, family structure and poverty 16 Chapter 2 Growth, human development, and anti-poverty programs 17 A. Introduction 17 B. Growing out of poverty 17 C. Capital for rural growth 21 D. And for the rural poor 22 E. Poverty-reduction: The missed target 26 F. Conclusions and recommendations 34 Annex 1 Poverty measurement: the debate on how best to capture the extent of deprivation 39 Annex 2 Village studies 45 Statistical appendix 47 References in text 69 List of tables Table 1.1 Poverty in India: 1993-1994 3 Table 1.2 Trends in literacy rates and life expectancy 7 Table 1.3 State rankings on the basis of alternative indicators of well-being 9 Table 1.4 Rural poverty by agro-climatic region 10 Table 1.5 Changes in rural poverty by agro-climatic region 11 Table 1.6 Characteristics of the rural poor 13 Table 2.1 How responsive is poverty to economic growth in India? 18 Table 2.2 Central plan budgetary expenditures on anti-poverty programs 27 Table 2.3 Public distribution system: allocation of subsidy on cereals between states 28 List of boxes Box 1.1 What could someone living at India's poverty line consume per day? 3 Box 1.2 Poverty measurement 5 Box 1.3 Did India's macroeconomic stabilization increase poverty in 1992? 5 Box 1.4 Land reforms and poverty reduction 12 Box 1.5 Labor market imperfections and the distributional impact of non-farm employment in India 14 Box 2.1 Poverty reduction: states performance 20 Box 2.2 Growth-enhancing public investment 22 Box 2.3 Child labor 24 Box 2.4 Resource requirements for quality education 25 Box 2.5 How well does the PDS serve the needs of the poor? 28 Box 2.6 The targeted public distribution system 29 Box 2.7 How well have ICDS interventions worked? 30 Box 2.8 Decentralization and outreach to the poor: evidence from social assessment studies 33 Box 2.9 The centrality of growth with equity to poverty reduction 35 Box 2.10 NGOs and poverty reduction 36 Box 2.11 Transient poverty in rural India 37 List of figures Figure 1.1 Incidence of poverty in India 1951-94 4 Figure 1.2 Change in poverty indces: 1951-94 4 Figure 1.3 Change in poverty indices for rural areas 4 Figure 1.4 Change in poverty indices for urban areas 4 Figure 2.1 Cumulative change in headcount index in India 18 Figure 2.2 Sectoral decomposition of headcount index change in India 19 Figure 2.3 Progress in poverty reduction by states 20 Figure 2.4 Subsidies continue to crowd out productivity enhancing expenditures 21 ACKNOWLEDGMENTS This Poverty Assessment was prepared by a team Tzanninis and Martin Muhleisen (IMF), and Ajit led by Zoubida Allaoua. It draws on the results of Gosh (ILO). The report benefited from Luis a three-year World Bank project directed by Emesto Derbez's (Division Chief) continuous Martin Ravallion and summarized in several support and Robert Drysdale's (Director) guidance papers by Martin Ravallion and Gaurav Datt. It and comments. Roberto Zagha contributed to the also draws on work by Peter Lanjouw (village report and provided direction. Richard Skolnik studies and poverty profile), Valerie Kozel and (Division Chief), and Shawki Bbarghouti Ruslan Yemtsov (survey of anti-poverty (Division Chief) also provided helpful comments. programs), Ravi Srivastava (survey of existing The document benefited from discussions held in literature), Shreelata Rao-Seshadri (NGOs), Keith May, 1997 with the Indian authorities. Hinchliffe and Mark Schlagel (human resources). The reviewers Lionel Demery and Soniya Arrangements for missions were made by Padma Carvalho provided very helpful comments and Gopalan and Sheni Rana. The report was suggestions. The report also benefited from useful desktoped by Lin Chin. Financial support for comments from John Williamson (Chief preparatory work was provided through the Economist), Rui Coutinho (SASVP), Benu Bidani Netherlands Poverty Trust Fund. (ASTHR), Meera Chatterjee (SA2RS), Dimitri « ECONOMIC DEVELOPMENT DATA GNP Per Capita (US$, 1995-96): 350 a Gross Domestic Product (1995-96) Annual Growth Rate (% p.a., constant prices) % of 70-71- 75-76- 80-81- 85-86- 91-92 92-93- US$ Bin GDP 75-76 80-81 85-86 90-91 95-96 GDP at Factor Cost 294.6 89.7 3.4 4.2 5.4 5.9 0.8 6.4 GDP at Market Prices 328.3 100.0 3.3 4.2 5.6 6.2 0.4 6.3 Gross Domestic Investment 86.1 26.2 5.3 3.7 5.7 9.5 -11.0 12.8 Gross National Saving 79.8 24.3 4.4 2.6 3.5 8.7 -0.3 10.7 Current Account Balance -6.4 -1.9 - -- -- -- -- - Output, Employment and Productivity (1990-91) Value Added Labor Force b V. A. per Worker US$ Bln. % of Tot Mill. % of Tot. US$ % of Avg. Agriculture 82.5 31.0 186.2 66.8 443 46.4 Industry 78.0 29.3 35.5 12.7 2198 230.2 Services 105.7 39.7 57.2 20.5 1848 193.7 Total/ Average 266.2 100.0 278.9 100.0 954 100.0 Government Finance General Government Central Government Rs, Bin. % of GDP Rs. Bin, % of GDP 95-96 95-96 90-91-95-96 95-96 95-96 90-91-95-96 Revenue Receipts 2174,1 19.8 19.6 1101.3 10.0 10.1 Revenue Expenditures 2539.5 23.1 23.3 1398.6 12.7 13.2 Revenue Surplus/ Deficit (-) -365.3 -3.3 -3.7 -297.3 -2.7 -3.1 Capital Expenditures d 416.2 3.8 4.2 305.1 2.8 3.4 External Assistance (net) e 3.2 0.0 0.6 3.2 0.0 0.6 Money, Credit, and Prices 89-90 90-91 91-92 92-93 93-94 94-95 95-96 (Rs. billion outstanding, end of period) Money and Quasi Money 2309.5 2658.3 3170.5 3668.3 4344.1 5308.0 6018.4 Bank Credit to Government (net) 1171.5 1401.9 1582.6 1762.4 2039.2 2224.2 2574.1 Bank Credit to Commercial Sector 1517.0 1717.7 1879.9 2201.4 2377.7 2896.6 3409.0 (percentage or index numbers) Money and Quasi Money as % of GDP 50.6 49.6 51.4 52.0 53.7 55.7 54.8 Wholesale Price Index (1981-82 = 100) 165.7 182.7 207.8 228.7 247.8 274.7 294.8 Annual Percentage Changes in: Wholesale Price Index 7.4 10.3 13.7 10.1 8.4 10.9 7.3 Bank Credit to Government (net) 20.3 19.7 12.9 11.4 15.7 9.1 15.7 Bank Credit to Commercial Sector 14.4 13.2 9.4 17.1 8.0 21.8 17.7 a. The per capita GNP estimate is at market prices, using World Bank Atlas methodology. Other conversions to dollars in this table are at the prevailing average exchange rate for the period covered. b. Total Labor Force from 1991 Census. Excludes data for Assam and Jammu & Kashmir. c. Transfers between Centre and States have been netted out. d. All loans and advances to third parties have been netted out. e. As recorded in the government budget. Balance of Payments (US$ Millions) Merchandise Exports (Average 1990-91-1995-96) 1993-94 1994-95 1995-96 US$ Mil % of Tot. Exports of Goods & NFS 27,947 32,760 39,636 Tea 404 2.1 Merchandise, fob 22,683 26,857 32,467 Iron Ore 487 2.5 Imports of Goods & NFS 29,798 38,150 48,540 Chemicals 1,891 9.6 Merchandise, cif 25,069 31,840 41,405 Leather & Leather products 1,439 7.3 of which Crude Petroleum 3,407 3,285 3,442 Textiles 2,708 13.8 of which Petroleum Products 2,244 2,396 3,759 Garments 2,731 13.9 Trade Balance -2,386 -4,983 -8,938 Gems and Jewelry 3,753 19.1 Non Factor Service (net) 535 -407 34 Engineering Goods 2,832 14.4 Others 3,423 174 Resource Balance -1,851 -5,390 -8,904 Total r 19,667 100.0 Net factor Incomea -3,775 -3,621 -4,945 External Debt, March 31, 1996 Net Transfersb 3,825 6,200 7,480 US$ Mill. Balance on CurrentA -1,801 -2,811 -6,369 Public & Publicly Guaranteed 79,725 Private Non-Guaranteed 6,618 Foreign Investment 4,235 4,895 4,347 Total (Including IMF and Short Term) 93,766 Official Grants and Aid 368 472 416 Net Medium & Long Term Capital 3,122 1,153 -1,036 Debt Service Ratio for 1995-96 Gross Disbursements 8,247 6,800 6,689 Principal Repayments 4,027 4,828 6,780 % curr receipts Public & Publicly Guaranteed 21.8 Other Capital Flowse 1,516 2,330 -308 Private Non-Guaranteed 14.7 Non-Resident Deposits 1,097 818 945 Total (Including IMF and Short Term) 28.1 Net Transactions with IMF 189 -1,174 -1,719 IBRD/ IDA Lending, March 31, 1996 (USS Mill) Overall Balance 8,538 6,858 -2,005 IBED IDA Change in Net Reserves 8,727 5,684 -3,724 Outstanding and Disbursed 9,767 17,499 Gross Reserves (end of year)d 15,476 21,160 17,436 Undisbursed 4,116 4,583 Outstanding incl. Undisb. 13,883 22,082 Rate of Exchange End-March 1997* US$ 1.00 = Rs. 34.80 - Not available. a. Figures given cover all investment income (net). Major payments are interest on foreign loans and charges paid to IMF, and major receipts is interest earned on foreign assets. b. Figures given include workers' remittances but exclude official grant assistance which is included within official loans and grants, and non-resident deposits which are shown separately. c. Includes short-term net capital inflow, changes in reserve valuation and other items. d. Excluding gold. e. The exchange rate was reunified at the market rate in March 1993. f. Total exports (commerce); net of crude petroleum exports. India Social Indicators Latest single year Some region/income group 1970-75 1980-85 1990-95 South Asia Low-income POPULATION Total population, mid-year (millions) 613.5 765.2 929.4 1,243.00 3,179.90 Growth rate (% annual average) 2.3 2.1 1.7 1.9 1.6 Urban population (% of population) 21.3 24.3 26.8 26.4 28.6 Total fertility rate (births per women) 5.6 4.8 3.2 3.5 3.2 POVERTY (% ofpopulation) National headcount index* .. .. 35.04 Urban headcount index .. .. 30.51 Rural headcount .. .. 36.66 INCOME GNP per capita (US$) 180 280 350 350 430 Consumer price index (1990=100) 36 70 165 Food price index (1990=100) .. 66 174 INCOME/CONSUMPTION DISTRIBUTION (%l of income or consumption) Lowest quintile 5.9 8.1 8.5 Highest quintile 49.4 41.4 42.6 SOCIAL INDICATORS Public expenditure (%/ of GDP) Health .. .. 0.7 Education 2.1 2.5 2.9 Social security and welfare .. .. Net primary school enrollment rate (% of age group) Total Male Female Access to safe water (% of population) Total 31 54 63 63.2 53 Urban Rural Immunization rate (A under 12 months) Measles .. .. 84 80 77 DPT .. 41 92 84 80 Child malnutrition (% under 5 years) .. .. 63 61 42 Life expectancy at birth (years) Total 50 55 62 61 63 Male 51 56 62 61 62 Female 49 55 63 62 64 Mortality Infant (per thousand live births) 132 108 68 75 69 Under 5 (per thousand live births) .. .. 95 106 104 Adult (15-59) Male (per 1,000 population) .. .. 229 239 244 Female (per 1,000 population) .. .. 219 230 211 Maternal (per 100,000 live births) .. 460 437 * Data for 1993-94 Source: World Development Indicators CD-ROM, World Bank, February 1997 and India: Poverty Assessment Report. s EXECUTIVE SUMMARY 1. The reforms India started in 1991 hold the percent. These patterns hold for both rural and urban promise of considerable improvements in the living poverty. Indicators that measure the depth and severity standards of the country's 300 million poor. During the of poverty suggest that the decline of poverty did not last few decades, India's inward-looking and public touch only those just below the poverty le while sector driven industrialization strategy led to rates of leaving the remaining poor unaffected. Rather, the growth and poverty reduction far more modest than process through which poverty was being reduced those witnessed elsewhere in the world, particularly in included those whose consumption levels were far South East Asia. The economy has responded well to below the poverty line. the reforms, and the government has made it an explicit objective to accelerate the development of the 3. Nonetheless, because of India's rapid country's human resources. By maintaining its population growth rate, even this rate of poverty commitment to economic liberalization, and reduction has not been sufficient to reduce the absolute redirecting towards infrastructure, health, and number of poor which increased from about 164 education the large resources now absorbed by million in 1951 (when almost half the population was subsidies inter-alia for power, irrigation, and fertilizer, below the poverty line) to 312 million in 1993-94 (35 India can give its long battle to reduce poverty a new percent of the population). In addition, India's progress impetus. in fighting poverty is modest when compared with some of its Asian neighbors. Between 1970 and 1993, Where Does India Stand? the proportion of Indonesia's population living in poverty dropped from 58 percent to 8 percent, an 2. Poverty is clearly declining, but slowly, and annual decline of nearly 10 percent--a greater decline remains widespread Poverty imposes an oppressive in a shorter period of time. weight on India, especially in the rural areas where almost three out of four Indians and close to 80 percent 4. Other social indicators of well being record a of the Indian poor live. Although that burden lessened history of progress that has, like the decline of poverty somewhat during the past four decades, significant itself, been steady but slow. Health, education, and easing did not come until the second half of that nutrition indicators describe a country which has made period. Unsurprisingly, poverty in India is responsive substantial gains against widespread deprivation over to the country's economic health. As the economy the 50 years of its Independence but has not achieved grows, poverty diminishes, faster when growth is the momentum needed to lift the great majority of its strong, slower when it is not. The proportion of the poor into the economic mainstream. Infant mortality population living in poverty declined at an annual rates, as one example, fell from 146 deaths per average of just 0.9 percent over 1951-75--when per- thousand births in the 1950s to 80 at the start of this capita income grew by a modest 1.7 percent. In the decade. Nevertheless, the Indian rate is still one of the years of stronger per-capita growth--2.5 percent-- highest in the world. Life expectancy at birth, now between the mid-1970s and late 1980s, however, the twice the 30 years that was the Indian average in 1947, average reduction was almost three times as fast--2.4 remains well below that of China (69 years). In 1991, n Executive Summary adult literacy rates for Indian females (34 percent) and scheduled tribe or caste, a person who lives in a for males (64 percent) were almost identical to those landless household or is dependent on wage earnings, for Sub-Saharan Africa and far behind those in China-- all face a significantly higher than average risk of 85 percent for women, 96 percent for men--ten years poverty. A recent survey of rural households indicates earlier. Maternal mortality (437 per 100,000) is high that the incidence of poverty was 68 percent among and accounts for a quarter of maternal deaths landless wage-earners, 51 percent for members of worldwide. scheduled castes and scheduled tribes, and 45 percent for members of household in which no-one was 5. The data also reveal sharp disparities within literate. India, between development oriented states and laggards, between women and men, between adults and 7. Education can be a powerful instrument to children, and between city and countryside. Different reduce poverty, but two-thirds of all women and two states have progressed at differing paces and, even fifths of all men--poor and non-poor--remain illiterate. within states, different regions have achieved markedly Among the most disadvantaged social groups, the varied results. Even more noticeable than geographic ratios are even worse: literacy rates of just 19 percent differences in poverty reduction are the inequalities among scheduled-caste women and of 46 percent for that persist across gender, caste and ethnic groups. men. One positive note, however, is that female Social indicators for women--literacy, for example--are literacy rates rose twice as fast as those for men during markedly lower than for men, and the level of the 40 years since 1950-SI when only nine percent of scheduled castes and tribes in both economic and social all Indian women could read and write. Continuing that achievements are still well below the national average. progress depends on getting girls into primary and The range in poverty reduction among the states is so secondary classrooms. Unfortunately, on average in wide that Kerala's progress in lowering the headcount India, half of all girls aged 5-14 years do not attend index of poverty (2.4 percent per year, on average, school. And where poverty is deepest, female literacy between 1957-58 and 1993-94) is more than 120 times is exceptionally low, 18 percent in the state of Bihar that of Bihar and more than four times that of and 21 percent in Uttar Pradesh. In Bihar, only 20 Rajasthan. These differences reflect more than natural percent of 5-9-year-old girls go to school and only 29 advantages or disadvantages. They mirror, instead, percent of the girls aged 10-14. Enrollment rates for conscious decisions on investing on the poor, their male counterparts were 33 percent for the specifically in improving their health and their younger boys and 55 percent for the older ones. Had education and, consequently, their opportunities to Bihar begun developing in 1960 with human resources participate in economic progress. As a result of these comparable to those in Kerala--where male and female decisions, Bihar--the poorest among India's 17 largest literacy rates are the highest in India and infant states which account for over 90 percent of the mortality is the lowest--it would have achieved an population--has the fifth lowest level of male life annual 1.2 percent decline in poverty incidence instead expectancy, the lowest level of male literacy, the of the marginal 0.2 percent it actually achieved. second lowest level of female literacy and the seventh highest level of infant mortality. Kerala, by contrast, 8. Landlessness is closely associated with rural has the lowest infant mortality rate and the second poverty. Not only is Indian poverty predominantly highest rates of male and female literacy. rural in absolute numbers, its incidence is more frequent in the countryside than in cities by a six-to- Who are the poor? five ratio. And in an agrarian economy a good indicator of well being is land ownership. The "risk" of 6. Gender, literacy, landownership, employment poverty among the landless is considerably greater than status, and caste are closely associated with poverty. for the population as a whole. Although in 1994 only Thus, an illiterate rural woman, a member of a 37 percent of the rural population was estimated to be Executive Summary ru in landless households, nearly half of the poor fell in and there is some evidence that the traditionally that category. And the incidence of poverty was dominant non-cultivating castes are being overtaken in highest of all--68 percent--among the landless wage economic terms by households of other castes who earners (one fifth of rural households) who provide have taken advantage of the new opportunities in unskilled labor in markets where the prevalence of agriculture. However, such changes do not affect all long-term contracts has been declining and wages castes. An important conclusion of the village studies remain too low to escape poverty. Again, rural women is that, in general, the lowest castes remain easily are more likely than men to depend on daily wages distinguishable from the rest of village society in terms from casual employment. More importantly perhaps, of both their very low material well-being and their village studies find little evidence of a broad rise in limited opportunities to improve their living standards. landlessness over time. In fact, many studies observe Even as poverty recedes, traditional inequalities largely some decline in landlessness, often coinciding with an remain. Paradoxically, in some households where expansion of small-holder cultivation. increasing affluence leads to the adoption of higher caste rituals and customs, women's ability to engage 9. Both secondary and village micro-study socially and economically outside of the household evidence indicate that over time there has been a declines. "casualization" of the rural labor force (a growth in the size of the daily-wage labor force). This is Reducing Poverty: What Matters Most? probably a consequence of the decline in the traditional artisanal occupations, and also the gradual decline of 11. Governments in India have relied on two the prevalence of permanent, long-term labor contracts. approaches to reducing poverty since Independence- Casual wage labor is a last resort which households the first based on the anticipation that the effects of turn to only when other options have been exhausted. aggregate rural growth would spread to all groups in As such, it is a good indicator of the relative position society such that poverty reduction is achieved side by of a household in the overall income distribution. side with increases in economic growth, and the second While casual labor is seen as a last resort, this trend that specific anti-poverty programs are required. The does not mean that rural poverty is increasing. Real reliance placed on each of the two objectives has agricultural wages have been increasing. The varied over time. Overall, the slow reduction in expansion of off-farm employment in the last few poverty through the 1950s and 1960s in particular, years and growing labor mobility, migration and reinforced skepticism regarding the strength of any commuting may have contributed to this outcome. trickle-down effect. Since the mid 1970s, however, the What this implies is that the dependence of the poor on faster poverty decline alongside a higher rate of agricultural wages has been increasing and that, more economic growth both in India and in the developing than before, the poor stand to gain from increases in world has led to a greater appreciation of the agricultural wages. contribution of growth to lasting poverty reduction and--equally importantly--reinforced the need for 10. While there is clear evidence of gender and investing more in human resource development since social inequalities in rural India, other inequalities that these investments not only contribute to faster long- are seen in many developing countries do not appear in term growth, they also increase the capacity of the poor acute form in India. Economic inequality varies little to benefit from it. Accordingly, the government's from the poorest regions to the more affluent, and it is strategy has over the recent past increased its emphasis also remarkably stable over time. This blurs, however, on providing the conditions for accelerated and the social reality of a nation where social stratification sustained labor-intensive growth, while expanding based on caste, ethnicity, and gender act to impede investment in human capital development. mobility. Rural growth has led to some erosion of the traditional caste structure in many parts of the country iv Executive Summary 12. Economic growth reduces poverty. The poor reductions in poverty, comparable to the progress made have, gained from economic growth, and lost from in a number of East Asian countries. From the results contraction and inflation. Results from a World Bank of this research, the lessons for the future are clear: research project using household surveys spanning promote growth and invest in human capital and forty years demonstrates that overall growth accounted infrastructure. for the lion's share of poverty reduction: 80 percent of the decline in the percentage of households below the 14. A more open trade regime since 1991, a poverty line between 1951 and 1970 and almost 100 significant reduction in distortions, and a liberalized percent since 1970. Improvements in distribution economy where the private sector plays a larger role (which occurred independently or as a result of have led to higher growth (7 percent in the last three deliberate government policies) achieved their greatest years); improved agricultural incentives; and are impact by the mid-1960s specially on reducing the creating the conditions for labor intensive intensity and severity of poverty, but with little impact industrialization. This will not only improve the on the incidence of poverty. The results of this exercise country's growth prospects, it will enhance the poverty clearly refute any presumption of "immiserizing reducing effects of industrialization and the role of growth"--that is growth that tends to marginalize or urban areas in this process. impoverish significant segments of the population--in India over this forty year period. Overall, two main 15. Investment in infrastructure and human messages emerge. First, India's development strategy resources, not subsidies, reduce poverty. Roads, well- over much of the period since Independence, focusing designed irrigation systems, flood control, and on capital intensive industrialization behind high tariff electrification all make inroads against rural poverty. barriers, brought negligible gains to the vast number of So also do investments in people, in their health, rural poor. The jobs it created required skilled labor education and nutrition and, in general, in their greater few of them could provide. This may explain why, ability to shape their own economic and social destiny. much more than urban-oriented growth, rural growth Since the bulk of the required investments in these has been the dominant force in poverty reduction. forms of physical, human, and social capital come Second, investment in agriculture, by contrast, has not from public rather than private sources in India, it is only raised average living standards in rural India but important that poverty-reduction strategies strike and has done so without accentuating inequalities. Green maintain the right balance between investments that revolution technology, initially suspected of conferring return significant dividends and subsidies that can be systematic advantages to larger landholders, has captured by better off sections of the population and generally proven scale neutral. distort markets. India has not yetfound that balance A recent government study estimates subsidies on 13. The same research project confirms that "non-merit" goods at 11 percent of GDP whereas between 1960 and 1993 some states reduced poverty spending on health and education is less than 5 percent much faster than others. In addition to differences in of GDP, and about 4 percent of GDP on infrastructure. growth, the study has shown that differences in initial Subsidization of current agricultural inputs has led to endowments of human capital and physical major micro-economic distortions that encourage infrastructure and in subsequent efforts at developing overexploitation of water and misuse of other these resources appear to have been an important factor resources. During the 1980s, agricultural subsidies in the differences in long-term rates of poverty grew at over seven percent a year or three times the reduction. An important message that has emerged is rate of expenditures on infrastructure and support that no Indian state effectively combined both policies services. Since 1991, the central government has to encourage growth and develop human resources halved the rate of growth of subsidies but what it has and physical infrastructure. The study suggests that if achieved, the state governments have undercut. any state had done so, it would have achieved rapid Executive Summary v 16. Cross country evidence confirms that the required to meet essential public health and clinical universalization of schooling and literacy are the pre- standards would require doubling the present public requisite for an equitable distribution of the benefits spending level on primary care of 0.6 percent of GDP. from higher growth. The evidence is similar between the states of India. There are strong relationships 19. With the exception of the Infant Mortality Rate between educational coverage, other indicators of the (IMR), India's social indicators are not lower than quality of life and poverty reduction. It is no those found in countries at similar levels of income. coincidence that Kerala has enjoyed the highest levels This reinforces one of the main messages of this report of both male and female literacy and the fastest decline that it is through rapid growth that India will be able to in the incidence of poverty over the past forty years. In reduce poverty and generate the resources to invest in spite of this evidence, around 33 million of the 105 the health and education of its people-who will in turn million children aged 6-10 years in India are still not sustain this growth. In the case of the IMR, the reason attending school; many more girls than boys are not in which may explain why India is an outlier is associated school; both boys and girls are in fact working (about with the composition of health spending, which may 17 to 44 million of them). not be reaching the poor, and gender discrimination. Innovative approaches to delivering social services, 17. Education alone does not raise the potential particularly to the poor, are therefore required. productivity of labor. Disease, maternal illness and malnutrition are all debilitating and are prevalent 20. Anti-poverty programs have not been cost among the poor. Communicable diseases and prenatal effective in reducing poverty. India's anti-poverty and maternal mortality cause four times as many deaths strategy comprises a wide range of anti-poverty per 100,000 population in India as in China and programs, many of which have been in effect for tuberculosis alone kills 500,000 people a year. Half of decades. Their share in total central budget expenditure all children under the age of five are malnourished and programs has been increasing--from 5.4 percent in one third of all newborn babies are underweight. 1990-91 to 7.7 percent in 1995-96 with similar trends Variations between states again demonstrate the ability at the state level. Most evaluations of India's anti- of governments to improve on these aspects of life to poverty programs--whether done by Government or which the poor are disproportionately exposed and independently--agree that these programs are not which reduce their capacity to benefit from increased performing as well as hoped. In particular, the extent to economic opportunities. which each program is accessed by the poor and by the non-poor differs widely. One of the better targeted 18. Investments in education and health have been programs is the Integrated Child Development Services below those necessary to accelerate growth prospects (ICDS), a nutrition and general development program and poverty reduction. Much of the fiscal focused at children aged 0-6 years and pregnant and responsibility for these activities lies with state lactating mothers and initially introduced in the governments and inter-state variations are large. country's poorest blocks. However, within these Nationally, however, real outlays on education as a groups there is no further targeting while the objective share of GDP have been falling. Programs to accelerate of rapidly universalizing the program may reduce educational development have begun, through the further its effectiveness. District Primary Education Program, (and there is a commitment to raise the allocation for education from 21. Public works have been relatively more the current 3.7 percent of GDP to 6.0 percent). In successful at targeting the poor. In particular, and at health, government spending is equal to just 1.3 least until recently, the Maharashtra Employment percent of GDP and covers only around 22 percent of Guarantee Scheme (EGS) was successful and relatively all health expenditures. To raise primary health care cost-efficient: it improved significantly the living from the current US$2-3 per capita to the levels standards of a large number of the extreme poor at a vi Executive Summary relatively low cost. Recent developments, however, Reducing Poverty: The Challenges Ahead may have eroded the effectiveness of the scheme. For example, the EGS wage which was on a par with 24. In the half century since its independence, India prevailing agricultural wages up to 1988, and so has accomplished many notable social and economic attracted only those unable to find employment achievements. Among these are the eradication of elsewhere, was doubled and thus eroded the program's widespread famine, a reduction in population growth, self selecting features. In addition, public works some lowering of caste barriers to economic programs including the EGS suffer from conflicting opportunity and the creation of a large pool of objectives (ensuring both gainful employment for poor technical and scientific talent. While it has also households and the expansion of infrastructure) which managed to reduce poverty in that period, only since have led to substantial leakages and a high 1975, when growth accelerated, has the decline been participation of non-poor. fairly steady. The pace, moreover, remains both slow and uneven--faster in the southern states than the 22. In contrast, a recent study of the PDS suggested northern ones, more likely to empower men than that only 40 percent of the total quantity of wheat and women. Government efforts to reduce poverty through about 47 percent of total quantity of rice supplied direct anti-interventions have yielded mixed results. through the PDS is consumed by the poorest 40 percent Many of those programs, in fact, have missed their of people in India. Moreover, the high administrative supposed targets--the poor--and delivered their benefits costs raise questions about the cost effectiveness of to the economically more advantaged. As India moves these programs. In general, it costs the government ahead with the economic liberalization that has yielded between 2 and 7 rupees (with the highest value a higher platform for growth and therefore the potential reported for the PDS) to provide one rupee to the poor. for a higher level of welfare, it has an opportunity to Several attempts have been made to reform the PDS. reexamine its approach to reduce poverty. The latest--the targeted PDS (TPDS) announced in the 1997 Budget--would grant all households with incomes 25. The centrality of Growth. The last five years below the poverty line up to 10 kg of cereals at half the have shown the rates of growth that India could economic cost. Subsidies for households with higher achieve with market oriented development policies and income levels would gradually be phased out. While a better integration into the world economy. This is a this proposal is promising, it is too early to assess its promising development because the last few decades effectiveness especially since the target group under have shown the extent to which the poor stand to gain consideration seem to still include the non-poor. from an acceleration in growth. The latter widens opportunity, provides the resources needed to invest in 23. To summarize, assessments of the current anti- human development, and creates the very foundation poverty programs indicate that there are a number of that will increase returns to human capital--and thus systemic problems: only partial success of targeting the families' willingness to send their children, including poor, such that estimates of non poor beneficiaries girls, to school, have fewer of them, or in multiple range from 20 percent in the ICDS and some rural other ways, invest in their future. development programs to 60 percent in the PDS; under-funding of better performing ones; often 26. Priority for human capital Neighboring complex administration; high administrative costs such countries in Asia which have made a point of that in many schemes they are higher than the value of combining pro-growth development policies with the resulting benefits; ill-defined multiple objectives investments in the health and education of their people which reduce quality and accountability; and have seen economic growth and poverty reduction inadequate monitoring often using indicators which follow. India, however, has not accorded sufficiently distort programs' true objectives, high priority to the education of the poor and 33 million of its 105 million 6-to- 10-years-oldsare not in Executive Summary vii school. These youngsters are not offered the programs while many of the non-poor benefit from opportunity to develop the skills needed for upward them. There is a consensus that.public works programs mobility. Along with the neglect of primary education have been among the most successful attempts at goes that of gender discrimination which condemns a reaching the poor. much greater proportion of girls and women to illiteracy and to ill-health. Reducing the gender 29. India therefore urgently needs to formulate an inequality among the poor requires a determined effort anti-poverty strategy that is fiscally sustainable and to focus on improved health care for women, maternal more finely targeted to those who truly cannot benefit health care in particular, combined with basic from the opportunities offered by growth. To increase education. India's health system needs to put a new their cost effectiveness and extend their outreach to the emphasis on basic care. Doing so will improve not very poor, safety nets need to be targeted to those who only the life span and well-being of poor women and either cannot participate in the growth process or face their ability to determine what family size they want, continuing exposure to risks, which are outside of their but it will also contribute to the economic health of control. Rural households are largely uninsured against their families, and consequently of India's society. agricultural yield shocks, for example. Effective safety And since poverty is not the only source of India's nets that insure rural poor against income fluctuations-- gender disparities, a determined government effort to such as public works programs--are essential in eradicate such disparities is urgently needed. overcoming an important market failure. 27. There is also scope for more involvement of 30. The challenge that India faces in reducing India's Non-Government Organizations (NGOs). poverty in the future faster and more effectively than it NGOs show great potential in helping the has in the past is an enormous one. India has met and implementation of many government schemes, such as mastered other major challenges and is laying the the IRDP, Watershed Development, Total Literacy foundation for higher growth with the continued Campaigns, to mention a few examples. Typically, liberalization of the economy. Building on this unlike in Bangladesh or in Pakistan where NGOs have foundation for faster growth would require that the developed national networks, in India, NGOs continue challenge of poverty reduction remains at the forefront to work in very limited areas. Yet many of them have of the official concern and becomes the subject of the proven effective in improving the poor's access to same imaginative, thorough, change-minded actions as financial resources, education and health services, the country's experiments in stabilization and structural government programs, or simply information, adjustment. In particular, the liberalization of the economy needs to focus on deregulating sectors on 28. Reforming anti-poverty programs. There is which depends the livelihood of many of the poor little evidence that anti-poverty programs have yielded (agriculture, particularly agro-industry, small scale gains in the living standards of the poor commensurate industries), and on improving the provision of with the significant resources that the country allocates infrastructure and social services to the to such programs. Many recipients of their benefits are underprivileged. Resources can be generated by widely recognized as amongst the poor. At the same reducing subsidies, and by reevaluating existing anti- time, many of the poorest people do not use these poverty programs. x 3 4 s « POVERTY: WHERE DOES INDIA STAND? A. Introduction 146 deaths per thousand births in the 1950s to 80 at the start of this decade. Nevertheless, the Indian rate is still 1.1 Poverty imposes an oppressive weight on India, high and two Indian states, Orissa (124 per thousand in especially in the rural areas where almost three out of 1991) and Madhya Pradesh (117 per thousand in four Indians and 77 percent of the Indian poor live. 1991), even recorded proportionally more infant deaths Although poverty has been reduced during the past than the sub-Saharan average (104 per thousand in four decades, it remains painfully high. Unsurprisingly, 1991). Life expectancy at birth, now twice the 30 years the strongest across-the-board progress coincided with that was the Indian average in 1947, remains well the years of rapid growth in the Indian economy as a below that of China (69 years.) Adult literacy rates for whole. As average annual increases of 3.4 percent in Indian males (64 percent) and for females (39 percent) GDP in the first half of the 1970s accelerated to rates in 1991 were almost identical to those for sub-Saharan of 5.9 percent in the last half of the 1980s, three widely Africa and far behind those in China--96 percent for used poverty measures--headcount; poverty gap; and men, 85 percent for women--ten years earlier. squared poverty gap indices (interpreted as indicating respectively the incidence, depth, and severity of 1.4 Indicators of health, education, nutrition and poverty)--recorded notable decreases in poverty. economic opportunity discussed in this assessment of National rates dropped by anywhere from one third to poverty in India describe a country which has made below one half their initial levels (Figure 1.1 and substantial gains against widespread deprivation over Appendix Table A. 1). the 50 years of its independence but has not achieved the momentum needed to bring the great majority of its 1.2 Because of India's rapid population growth rate, poor into the economic mainstream. Whichever way even that advance, however, has not been sufficient to one chooses to think about living standards, there is no reduce the absolute number of poor which increased question that in India in the 1990s a significant fraction from around 200 million in the 1950s to 312 million in of the population remains severely deprived. The data 1993-94 (most recent Five Year Survey). This leaves also reveal sharp disparities within India, between India with the largest concentration of poor people in women and men, between castes, between city and the world, particularly in the villages--fewer than countryside, between prospering regions and 5,000 people--where 60 percent of all Indians live. impoverished ones. Staggering as the overall numbers remain--240 million rural poor and 72 million urban poor--they do not tell 1.5 This assessment concludes that growth- the full story of change. Social indicators of well- enhancing public policies are the sine qua non of being, for instance, record a history of progress that lasting progress in reducing poverty. Such growth can has, like the decline of poverty itself been steady but both provide the Indian poor with opportunities to raise slow. themselves above deprivation and can enable governments at all levels to promote that rise through 1.3 Among those indicators, three illustrate this sustained investment in health, education, nutrition, point. Infant mortality rates, as one example, fell from family planning and other human capital development 2 Poverty: Where Does India Stand? policies. These conclusions are consistent with the first understanding of the effect of past development Bank poverty assessment for India issued in 1989, based policies on the incidence of poverty, and of the on data up to 1983-84. The report's main conclusions important role played by initial endowments in human then were: that both growth and redistributive policies capital and physical resources in reducing poverty over played a role in reducing poverty over much of the the long-term. In this context, after reviewing the 1970s and 1980s; and that while more rapid growth led profile of the poor (Chapter 1), the report takes a longer to significant declines in poverty, the agricultural, term perspective by focusing on determinants of industrial and labor regulations prevailing at that time persistent chronic poverty to inform public policy diluted the effect of growth on poverty reduction. The design to reduce and possibly eliminate it. It report also emphasized the importance of targeting anti- recognizes, however, that transient poverty (due to poverty programs because of India's unsustainably high unforeseen fluctuations in income such as in drought fiscal deficits at the end of the 1980s and the need to years or the death of the bread winner) is just as protect poverty programs (which the report argued were important and is the basis on which is built a rationale required to protect those who, because they lack the for effective safety nets policies (see chapter 11). human or physical capital, or live in chronically depressed regions, cannot benefit directly or 1.7 This report is largely, although not exclusively, immediately from the growth process) which were likely concerned with rural poverty. This is partly because at to bear the brunt of inevitable expenditure cuts. present there is a much stronger analytical underpinning to our understanding of rural poverty and 1.6 Since the publication of that report, India has its causes. It is also clear that while population trends radically changed its development strategy through point to a continued rise of the urban sector in the reforms that have significantly liberalized the economy future, current poverty incidence as well as numbers of and improved its growth performance (growth rates poor remain considerably higher in rural than in urban have been at around 7 percent in the last three years). areas. Further, it is important to recognize that the Also, much more is now known on India's poverty. definition of urban encompasses a wide range of town The Bank has conducted an extensive research and city sizes. A large share of the urban population program on the causes and evolution of poverty resides in small and medium-sized towns and cities, (Ravallion and Datt, 1996a, Ravallion, 1996). At the and much migration from rural to urban areas goes to same time, the debate over the real number of poor such centres. These towns generally retain strong links intensified as it became increasingly clear that the to the rural economy, and therefore poverty is likely to official methodology was underestimating poverty be strongly influenced by developments in rural incidence. The Government set up an Expert surroundings. While we are able to say relatively little Committee (whose recommendations were recently about the causes of, and solutions to, poverty in India's endorsed by the Government) which suggested ways of large urban centres, the analysis and recommendations remedying some of the methodological shortcomings in this report are lely to be of some relevance to the as well as further areas of research to refine further the small cities in the urban sector. That said, there is an estimation of poverty incidence in a country as vast obvious need for further analytical research leading to and diversified as India (see Annex I). This report better insights on how the workings of the urban extends the analysis contained in the 1989 report to economy affect the poor, and the public policies include data from the 1993-94 survey--last year for needed to improve the future of the urban poor. which nationally representative data are available. By analyzing India's past performance in these areas, this Poverty measurement assessment hopes to contribute to renewed and reinvigorated efforts to reduce poverty in the world's 1.8 What is poverty? In India, poverty is officially second largest nation. The analysis provides a stronger linked to a nutritional baseline measured in calories empirical underpinning of the central role of growth (food-energy method). The Planning Commission over distributive policies for poverty reduction, a better defines poverty lines as a per capita monthly Poverty : Where Does India Stand? 3 expenditure of Rs. 49 for the rural areas and Rs. 57 in urban areas at 1973-74 all-India prices. These poverty Box 1.1: What Could Someone Living at India's lines correspond to a total household per capita Poverty Line Consume Per Dayl expenditure sufficient to provide, in addition, to basic Based on average (all-India) prices. fhe aerage rood non-food items--clothing, transport--a daily intake of expendture he compoihion corresponds to the all- 2400 calories per person in rural areas and 2100 in Ini awrage% for the 30-40 percent fractile in that urban areas (Box 1.1). Individuals who do not meet 'ear) ofa person living in rural areas at India's po%erty line in 1993-94 would ha%e enabled the consumption these calorie norms fall below the poverty line. This report uses the poverty lines as refined by the 1993 recommendations of the Planning Commission's Consumption/dU Expert Group on Estimation of Number of Proportion (61s, rice. o hr) 40 gis of poor (see Annex I for a fuller discussion of poverty Milk 70 ml measurement) but corrects for across states and survey Eggs 0.2 (no.) periods price differentials (Table 1.1 gives results 40l ou t i 10 omtr under both methodologies). VegeEables 020b p013LO. 191a onion: 170% brinjal: 120o tomatoes) 1201 gins 1,1: Poverty in India: 1993-1994bananas 13 - 1 coconU L) 0.1 (no I FSjlac ksdothDatR alinriThuolT Dried chile 4 gins Rural Urban TotalTealeaves 3 gs Head-Count 36.7 30.5 35.0 (37.3)' (33.7)' (36.3)' Nfter buying such a bundle of food items, the person Number of poor (mill.) 240.5 71.5 312 Percent of total 77 23 100 %,ould ha%e left over about Rs.?/da\ to put to%ard Poverty gap 8.38 7.40 8.13one third of Pqa overty gap 8.38 7.40 8.13 Itndia's population cannot afford e%en this frugal Squared poverty gap 2.79 2.42 2.69bude a Corresponding figures based on the expert group methodology (government sources) Source: Datt (1997). of the National Sample Survey Organization household survey, covering a period from 1951 to 1993-94 (Datt, B. Evolution of Poverty in India 1997, Datt & Ravallion, 1996a, Ravallion & Datt, 1996a, 1996b). These national household surveys are suitable How much progress has India made in fighting for tracking the poor's living conditions since the poverty since the 1950s? consumption data which have been collected in these surveys are reasonably comparable. 1.9 The reduction of poverty has been a major concern of the Government of India since 1.10 As of the most recent (1993-94) household independence in 1947. As such the analysis of poverty survey conducted by the National Sample Survey enjoys a long-standing tradition with an extremely rich Organization and based on the poverty lines calculated literature and data base. India has one of the longest by Ravallion and Dart (1996c), and Dart (1997), 36.7 series of national household surveys--spanning over 40 percent of India's rural population and 30.5 of its city- years--suitable for tracking living conditions of the dwellers lived in poverty-a national average of 35.0 poor over time. Such a dynamic perspective also helps percent (Table 1.1, Figure 1.1). The evolution of to understand what factors have been important in poverty, measured on the basis of the headcount, determining poverty outcomes. The role of initial poverty gap, and the squared gap indices (Box 1.2) at the conditions, policies and institutions, can be drawn out, all-India level and separately between rural and urban and the lessons learnt can provide an input into the areas can be seen to have followed three phases (Figures formulation of forward-looking policies. A recent World 1.1-1.4). First, the period from the early 1950s to the Bank research project assembled and analyzed 35 rounds mid-1970s was characterized by fluctuations in poverty 4 Poverty: Where Does India Stand? Figure 1.1: Incidence of Poverty in India 1951-94 Figure 1.3: Change in Poverty Indices for Rural Areas: 1951-94, Relative to Base (1951) Level 65 6o 60 1.70 55 50 Nation Rural I - 0 45 . 40 0.30 35 Li0.0- 0.50 eHead count index -Poverty gap index Soure: Drt, 997.- - - Squared poverty gap index Source: Datt, 1997. ne Source: Calculation based on Datt, 1997. without a clear trend in either direction. The average head count index was 53 percent in 1951-1955, about Figure 1.: Change in Poverty Indices for the same as the average over 1970-74. It thus took 20 Urban Areas: 1951-94, Relative to Base (1951) years before poverty in India dropped consistently below Level the level observed in the early 1950s. Second, from 1971 1.0 to 1986-87, poverty entered a phase of steady decline. It 1.40 is noteworthy that the decline in poverty is more I--0 pronounced the more distribution-sensitive the poverty 0.80 measure (that is the fall in squared poverty gap is more 0.40 pronounced than those of the headcount and poverty gap ______LN _M ___ -Head count index indices). Thus, the decline of poverty was not simply a Poverty gap index process whereby a segment of the population which had Squrd poverty gap index previously been located just below the poverty line was Source: Calculation based on Datt, 1997 able to lift itself above the line, while the remaining poor were left unaffected. Rather, the process through which around a level considerably lower than that which poverty was being reduced included also those whose prevailed in the 1970s. There also emerged some consumption levels were far below the poverty line dissonance between the trends observed across poverty (Figure 1.2). Finally, after 1986-87, poverty appears to measures and across the rural and urban sectors. For have entered into a new phase of fluctuation, although example, while the rural headcount index rose between 1986-87 and 1987-88, the squared poverty gap index Figure 1.2: Change in Poverty Indices: 1951-94 continued to decline (Figures 1.3-1.4). What is important Relative to Base (1951) Level is that as average Indian living standards rose during 1 s.90. the 40 years since 1951 and particularly after the mid- 1.60 1970s, the poor did not get poorer. IA0 - - - 1.20 - A , -- ______ 1.11 The direction of change is unmistakable. Over 0.600 080 -----.'=--~---the roughly 15 years between 1970-71 and 1986-87, 0JA .. .. . ...0. , the incidence of poverty in India as a whole declined-- a ~at arate of about 2percent per year--from 53 percent Heado 38 In rural areas the decline was Poverty gap index Squared poventy ap index more rapid, about 2.2 percent per year while urban Source: Calculation based on DartF 1997. poverty incidence declined at a slightly lower rate-4.8 percent per year. This differential growth rate led to Poverty : Where Does India Stand? 5 production, and inflation) which one would expect to Box 1.2. Poverty Measurement be influenced by the effects of stabilization policy on Threethe poor (Box 1.3). The subsequent decline in poverty index (H), the poverty gap index (PG), and the squared poverty gap index (SPG)--which belong to reforms led to higher poverty. the general class of poverty measure introduced by Foster. Greer and Thorbecke (1984) are now widely used to capture alternate dimenions of po%ertN corresponding to its incidence. depth. and se%erit) Bo% 1.3: Did Indias Macroeconomic respecttmeh. H. is et%en by the proportion of the Stabilizadon Increase Poiert in 1992? population "hose per-capita consumption or income letel emeasured standard of li%inc) is below the The National sample Sure%s of 1990-91 and po,ern line whose real talue is f.ed o%er time H 1991A2 indicate that the incidence of po%ert\ measures the incidence of po%t-, Howe%er. H has increased from 16 percent of the population to 41 its limitation because it does not Po e an indication percent. To imlate the etfects or the poor of the of how the poor population is distributed along the %arious econom.-"tde factors, research for this consumption or income spectrum below the poverty report e\amined ho% Indias rural po%err% line. The other two measures provide such measures ha' e responded in the past to changes it information. The PG, is the average distance below keN economic 'ariables such as real agricultural the poverty line in the population (counting the non- state domestic product (SDP. per hectare of so"n poor as having zero poverty gap); expressed as a area in the state, real non-agricultural SDP per percentage of the po%ert% line The PG captures the capita rural rate of inflAon, per capita re.l state depth of po%en The last measure SPG. which de%elopment e\penditure. and real male measures the se'erity of po%ery airiong the poor. is aricultural "age Time-series measures of the based on the indildual po%ert gaps raised to a incidence, depth and 4e%eritN of rural po% er were power of 2 estimated u4ing ;tate-le%el data from 19 rounds of the National Samnple Sur'ey ' spanning 1960-61 to 1992 Thes e %ariables accounted tor 90 percent ot some decline in the gap between rural and urban the 'ariance itmeasured po%erN The nodel was povertyrates.then used to assewss %%hat role thos,e same variables. poverty rates. ma% ha%e pta%ed in the increase in po%erv, in 1992 A sub-et of these %ariables can be identified 1.12 The jump in measured poverty between the late as likel% channels through "hich stabilizaton 1980s and 1992 has generated considerable public %ould affect li%ing standards of the poor Those debate on the social consequences of stabilization and the inlton ra state de%elopnient e%penditure. hinltorae reform--as it coincided with the introduction of a and the real male agriculmral %age. Of course comprehensive program of stabilization and reforms. these %ariables are changing for other reasons. Establishing a causal link between these two events is including the effects of the crisis preceding the reforms and c urrent e\ogernous shocks I such as the difficult however without a counterfactual showing effects of the bad agricultural .ear on real %%age what might have happened to poverty if the 1991 in aericulturei stabilization program had not taken place. There is also the possibility that the rather high poverty rates crisis and stabilization accounted for at most 36 observed for 1992 are partly the result of sampling and percent of the increase in the po%erty rate in 1992. non-sampling errors in the data (The 1992 results come and a smaller share of the increase in the depth and from a "thin" survey, covering just over 13,132 attrita o The ret o ndt shu-run households and a shorter time frame for interviews as impacts ofpolic% changes. at least as the' atTected opposed to the five-year surveys with a sample of 64,300 the key economic %arables %hich matter to the households, see Datt and Ravallion, 1996e). Research poor Ho%e%er. the stud; "as less successful in e\plamning %%hat did in fact account for the carried out for the 1996 Country Economic increase in pomet' in I902. That remains, in large Memorandum, for example, finds that only about a pan. a myster One possible e%planation could be third of the increase in measured poverty over this the small sizeot the 1991-1992 NSS suney period could be attributed to variables (such as wages Sozrt- Gaura% Dan and Narfin Ravallion. 1996e for unskilled agricultural laborers, agricultural 6 Poverty: Where Does India Stand? 1.13 The magnitude of decline in poverty of the last harvest fluctuations. Some village studies point out that two decades is significant but not dramatic. While the on the basis of other indicators of well-being decline of poverty since the early 1970s has been (consumption, wealth) the moves into and out of sizable (from an incidence of 56 percent to 35 percent poverty are less frequent. In addition, even in the in 1993-94), India's progress in fighting poverty has presence of fairly high mobility there are often been modest when compared with some of its Asian particular sub-groups within villages who remain neighbors. Between 1970 and 1993, for example, the below the poverty line over long periods of time. In proportion of Indonesia's population living in poverty broad terms, however, the message from villages dropped from 58 to 8 percent, an annual decline of studies conforms with all-India survey results, i.e. there nearly 10 percent. is widespread and pervasive poverty which has been slowly falling over time. 1.14 As of 1993-94, India's poverty continues to be predominantly rural although rural poverty declined 1.17 Even when incomes do not appear to have faster than urban poverty over 1951-88 (see last row of improved much over time, other indicators of well- Table A. 1 and figures 1.3-1.4). Moreover, the decline being may point to improved longer-term living in national poverty seems to have been driven mostly standards. A well known example is a study of two by the decline in rural poverty--not surprising given Rajasthani villages, first surveyed in 1960-62 and then that 74 percent of India's population lives in rural revisited in 1980-82 (Jodha, 1989). This study found areas. These findings--based on 40 years of nationally that real per capita incomes were lower than in 1980- representative household surveys--are reinforced by a 82 than twenty years earlier. But there were significant host of multi-disciplinary village studies. improvements in other, qualitative, indicators of economic well-being. For example, Jodha observed: 1.15 Evidence from Micro-Studies. Alongside the (i) expanding economic opportunities, including a nationally representative household survey data, there broader range of employment activities and sources of has been a rich tradition of village studies in India over credit; (ii) increased consumption of goods with high the past decades. These studies have been undertaken by income elasticities (e.g. refined rather than unrefined investigators from a wide range of disciplinary sugar); (iii) investment in lumpy consumer durables; backgrounds. By definition, these village studies have and (iv) reduced reliance on patrons. Two lessons taken place in rural areas, and as such do not provide emerge. First, the study serves as a reminder that insights into the evolution of urban poverty. For the comparisons of only two observation points can give a purpose of this report, a review of longitudinal village misleading impression of change over time, studies offers some glimpses of how poverty has particularly when the indicator being used is current evolved over time from a perspective which is quite income--a measure which is very sensitive to local distinct from that available through household surveys harvest conditions. Jodha's study suggests that (Jayaraman & Lanjouw, 1997). consumption levels actually improved between 1960- 62 and 1980-82 even though income was lower in 1.16 Most of the village studies surveyed point to 1980-82. Second, the study also suggests that it is gradual declines in poverty over time. However, possible that certain aspects of well-being, such as poverty rates often fluctuate from year to year, as a perceived reliance on patrons, or simply an expansion result of variation in the quality of harvest and in the of choice, might evolve over time in a way which is movement of prices. It is also important to note that not necessarily directly linked to income levels. village studies show a fair amount of movement into and out of poverty over time. This high mobility is 1.18 Many village studies suggest that the poor particularly noticeable when current incomes are used perceive themselves to be better off now than in to identify the poor (as opposed to indicators such as previous decades. Factors which influence perceived consumption or wealth which reflect longer-term living well-being, but which are unlikely to be well-reflected standards), and is certainly related to the impact of in income or consumption figures, include the Poverty: Where Does India Stand? 7 increased sense of freedom and self-respect associated terms of infant mortality rates: per 1000 live births in with some weakening of caste-based restrictions on India, 79 babies perish. In China the comparable figure occupation and behavior. At the same time, however, is 31, in Thailand it is 26, and in Sri Lanka it is 18 some village studies indicate that these new freedoms (Dr&e and Sen, 1995). In addition, while India has can instill a sense of vulnerability, particularly among made significant improvements in food availability and those who are least well-placed to take advantage of distribution, rendering famines, even in drought, a new opportunities (due to low education levels, for thing of the past, India's rates of malnutrition among instance). children and women are among the highest in the world. Analyses of cross-country patterns (Anand and 1.19 Although the broad pattern of poverty decline Ravallion, 1993) indicate that the largest factor is reflected both in village studies as well as in accounting for the differences in social indicators is national statistics, the village studies do also point to India's per capita income. In fact, with the exception of pockets of increasing impoverishment. In general, the Infant Mortality Rate (IMR), India's social living standards are observed to be determined by a indicators are not lower than those found in countries complex interaction of various counteracting forces. at similar levels of income. This reinforces one of the On the one hand, population growth and degradation of main messages of this report that it is through rapid the natural environment puts increasing pressure on growth that India will be able to reduce poverty and per-capita food availability and per-capita incomes. On generate the resources to invest in the health and the other hand, expanding employment opportunities education of its people--who will in turn sustain this (farm and non-farm) and technological change in growth. In the case of the IMR, the reason which may agriculture raise wage rates and farm production. On explain why India is an outlier is associated with the balance the evidence suggests that latter forces have composition of health spending, which may not be been able to hold their own against, and even offset, reaching the poor, and gender discrimination. the former. But this has clearly not been the case in all settings. Moreover, numerous other factors come into 1.21 In terms of gender-related indicators, India, in play. For example, traditional social support networks fact, lags behind. Alongside the low female education may have eroded during the process of economic achievements (Table 1.2), other indicators show a transformation, leaving certain individuals, e.g. the similar pattern. The female share of the labor force in elderly and disabled, in a more vulnerable state than India, at 29 percent is lower than in all other Asian before. countries except Pakistan at 14 percent (Dr6ze and Sen, 1995). In terms of the ratio of females to males in the 1.20 How do the poor fare regarding other population, the figure for India as a whole, 93 percent indicators of well-being? We have already seen that at is somewhat lower than is observed elsewhere in Asia the time of the 1991 Census, literacy rates among (101 in Indonesia, 99 in Sri Lanka, and 94 in China). women and men (aged 7 and above) were 39 percent However, this masks considerable regional variation. and 64 percent respectively. For the population as a In the state of Uttar Pradesh, for example, the female- whole the corresponding percentage is 52. While these male ratio is as low as 88 percent, while in Kerala the v ery low figu res are in lin e w ith th o se ob serv ed in T able_ 1 .2__ T rends_ in_ Literacy _R ates _a _d_Life other South Asian states (except Sri Lanka where the Epectacy population literacy rate is 89 percent), they lie far Limy rtes Lire expectancy behind the achievements of countries such as China, (percent) at birth (years) Thailand and Indonesia at 78 percent, 94 percent and Female Nale 84 percent, respectively. Life expectancy in India as a 1960-61 1 40 41 whole at the time of the 1990-91 Census was 59 years. 1970-71 22 46 45.6 This compares against 72 years in Sri Lanka, 69 in 1980-81 30 56 50.4 China and 69 in Thailand. A similarly poor record 1990-91 39 64 59.2 (Sources: Drze and Sen (1995). relativegnificantrimprovementsrinsfood availabilitydand 8 Poverty: Where Does India Stand? ratio is 104 percent (closer to those observed in the urban poverty reduction followed by West Bengal, industrialized countries). In general, very low female- Madhya Pradesh, Tamil Nadu and Bihar (see Appendix male ratios are observed across the north of the Table A.2 which shows the annual trend rates of country. change in mean consumption, the three poverty measures and Gini indices across states over the period Regional poverty incidence 1957-58 to 1993-94). 1.22 The more rapid progress of rural poverty 1.24 State-wise poverty incidence is a hotly debated reduction and the four-decade average annual rate of issue in India because the central government's decline conceal very diverse patterns of regional transfers to the states are partly based on state relative development and well being. Regional disparities in poverty. The use of poverty rates to guide fiscal standards of living remain large. For instance, during transfers can be controversial, particularly since 1993-94 the proportion of rural population of the state statewise rankings of poverty are not robust as of Bihar in absolute poverty was 64 percent--more than illustrated by Table 1.3. Seventeen of the major states two-and-a-half times higher than the proportion (25 of India are ranked in this table in terms of the percent) of the (combined) states of Punjab and incidence of rural poverty, based on the 1993-94 NSS Haryana. Some of these differences appear to have data and using the Expert Group methodology for persisted historically: Punjab-Haryana was one of the setting the poverty line. States are ranked in terms of richest regions during the late 1950s. But there were increasing well-being (i.e. lower poverty). In the first also differential trends across regions. Looking back column it can be seen that Bihar, Orissa and Assam over time, the more striking feature of the Indian appear as the three poorest states, while Punjab, experience in reducing poverty has been the markedly Andhra Pradesh, and Gujarat represent the least different rate of progress across states. impoverished ones. The second column indicates 1.23 For instance, the headcount index shows Keralaof the scoring an annual rate of decrease in rural poverty of 2.4 percent between 1957-58 and 1993-94. It is the top capita income from a rural household (33,000 performer among the fifteen major Indian states--during perf,ormer aoted foee majorcet Indias- the first half of 1994 (NCAER 1996) rather than per which, together accounted for 97 percent of India's caiaonuponfmth1934NSO ste population in 1991--by this measurement as well as by welfa indiorom tha the the poverty gap, squared poverty gap and Gini indices. Attwo methods are not that different, the rankings of experienced increases in rural poverty rates on three whil iahes Prdeh is o the middexinptat out of four measures above, and Bihar's annual rankings according to the NSSO data, it was found to average reduction in the headcount index of rural be the third poorest state under the NCAER data. poverty amounted to just 0.02 percent. In between the Similar re-rankings are likely to occur with alternative extremes, progress against rural poverty in Tamil Nadu, Orissa, Gujarat and West Bengal varied from 60 poverty meaes o atate povert li en toother indicators of well-being are used (columns 3-8). Pradesh, the second best by all indices, recorded a This--when combined with other methodological headcount index only 88 percent as high as Kerala's. shortcomings in measuring state-wise poverty-- Performance is equally diverse within the urban sector. explains the complexity of the measurement issue and Punjab and Haryana had the highest rate of decline in the fact that any rebasing of fiscal transfers to states on urban poverty; the other good performers were Jammu the basis of rankings on any one poverty criterion (or and Kashmir, Kerala, Assam, and Gujarat. any ad-hoc combination of criteria), is unlikely to Maharashtra, on the other hand, had the worst record in receive widespread agreement. Poverty : Where Does India Stand? 9 Table 1.3: State Rankings on the Basis of Alternative Indicators of Well-Being Po%erty incidence Life expectancy Infant Literacy rate Female/ NSS data NCAER Male Female mortality Male Female Male ratio Bihar I -4 5 - 7 1 2 5 Orissa 2 I 3 2 I 8 6 12 Assam 3 12 - - 5 6 8 7 Uttar Pradesh - 5 2 3 3 4 3 2 West Bengal 5 2 8 8 13 10 10 6 MadhNa Pradesh 6 (1I I 2 5 4 8 Maharashtra 7 9 12 11 1-4 15 15 95 TamilNadu 8 10 9 9 15 13 13 14 Himachel Pradesh 9 3 - - 10 14 14 15 .ammu and Kashmir 11. - - - 11 5 - - - Karnaraka II II 10 7 6 9 9 11 Harvana 12 15 11 10 95 II 7 Rajasthan 13 - 4 4 4 2 1 4 Kerala 14 14 14 13 17 16 16 16 Gujarat 15 8 6 6 11 5 12 11 9 5 Andhra Pradeih 16 16 7 5 95 3 5 13 Punjab 17 13 13 12 16 8 12 3 1. Incidence of poterr\ in 1993-94 E%pert Group Calculations from the 50th Round of the NSS Sune), Srinitasan (1997), and NCAER (1996). 2. All other indicators are taken from Dr6ze and Sen (1995) and refer to the early 1990s. 1.25 While the relative position of a state is rarely what they mean and how they perform, is often most identical across different indicators of well-being, it is informative. nevertheless possible to observe certain regularities across the columns of Table 1.3. States such as Bihar, 1.26 Similarly, even within states, there may be wide Orissa, Uttar Pradesh, and Madhya Pradesh are nearly geographic variation. Many states in India have larger always clustered at the low end of the rankings. populations than entire nations elsewhere. The state of Similarly, Kerala, Punjab and Haryana cluster near the Uttar Pradesh, for example, has a population of over 160 top for most indicators. It is interesting to note that the million, larger than that of any country in the world other female/male ratio (last column) is fairly independent than China, Idonesia, Brazil and the United States. The from both material well-being indicators (proxied by states are also large and varied in terms of their poverty measures) and indicators such as infant geographic characteristics. It is therefore of interest to mortality or education outcomes. While Orissa scores disaggregate poverty data beyond the aggregate state fairly badly on all indicators, it is positioned near the level poverty incidence. The quinquennial NSS surveys top in terms of this measure of gender discrimination. have large enough samples to allow such a At the other extreme, states such as Punjab and disaggregation. The most recent survey for which this Haryana which do well in terms of poverty measures, has been done is the 1987-88 (Dr6ze and Srinivasan, literacy rates, life-expectancy and infant mortality 1996). In this study India was broken down into 61 rates, do very badly in terms of gender relations regions based on agro-clinatic conditions. Tables 1.4 (proxied by this female/male indicator). It is this type and 1.5 summarize some of the findings on rural poverty of independence across indicators which makes it from that study. Of the ten poorest regions in 1987-88, particularly difficult to interpret indices which five were in either in Bihar or Orissa. The observation, at aggregate together different dimensions of well-being. the state level, that these two states were very poor is That is, when different indicators of well being do not strongly supported by the regional data, and poverty move in the same direction, it becomes difficult to seems to be fairly evenly distributed across regions in interpret what the final outcome is. Therefore, keeping these states. clear the distinction between the indicators, in terms of 10 Poverty: Where Does India Stand? 1.27 In Orissa's southern agro-climatic region, Table 1.4: Rural Poverty by Agro-Climatic Region according to 1987-88 data, more than three of every Region Head-count Gini four rural inhabitants lived below the poverty line. coefficient Further examination of regional performance in poverty LH%g&est inddence reduction between 1972-73 and 1987-88 confirms again M inland the particularly deprived position of the rural population eastern 48.8 0.264 in Bihar. Of the three agro-climatic regions into which UP southern 50.1 0.255 Bihar can be divided, two (Bihar central and Bihar Bihar central 51.9 0.249 northern) have recorded among the smallest TN coastal northern 52.9 0.287 improvements in poverty rates over this period. Of the Bihar northern 53.0 0.262 southern states of India, only northern Tamil Nadu Orissa northern 53.7 0.286 WB eastern plains 54.2 0.247 stands out as a particularly poor region, in terms of both Rajasthan southern 61.1 0.327 high poverty in 1987-88 and a very weak performance in Orissa southern 77.0 0.251 reducing poverty between 1972-73 and 1987-88. 11 Lowest incidence oe pawMrt in 1987-8 UP Himalayan 8.4 0.288 1.28 While Uttar Pradesh is among the poorest states Haana western 8.7 0.268 on aggregate, in the Himalayan region of Uttar Pradesh Punjab northern 9.3 0.297 in 1987-88, poverty afflicted only 8.4 percent of the Karnataka coastal and population, the lowest incidence in all of India (Sec. 11 ghats J&K Jhelum valley 13.4 0.280 of Table 1.4). Moreover, between 1972-73 and 1987-88, Punjab southern 13.4 0.304 this region recorded the single largest decline in the J&K mountains 16.9 0.323 incidence of poverty over time (Sec. II of Table 1.5). APcoastal 17.0 0.291 Hrana eastern 18.7 0.3 12 Interestingly, the Himalayan region of West Bengal AP south western 18.9 0.309 achieved a similarly strong pace of poverty reduction Source: Dr6ze and Srinivasan (1996). (56.5 percent between the 1972-73 and 1987-88 surveys), a pace more than ten times that of the nearbytribes, northern region of Bihar (Table 1.5). respectively, C. Who are the Poor? 0 45 percent for households in which all members were illiterate, compared to 27 percent for 1.29 Factors such as population density, ecological households where both male and female adults conditions and the availability of irrigation and read and write. transport account for some of these differences among India's states and even within them. Other conditions Access to land, sources of income andpoverty affecting the rural poor--gender, literacy, land ownership, employment status, caste and family size-- 1.31 If a to ean andonoic create a more consistent pattern. Thus, an illiterate ooic go and hand, so do onrs rural woman, a member of a scheduled tribe or caste, a person living in a landless household or dependent on wage earnings, all face a significantly higher than nubr,isncdce smoefqetinh. aveage eris of l poverty.ifcnlyhgerta countryside than in cities by a six-to-five ratio. And in averagean agrarian economy a good indicator of well-being is 1.30 According to the NCAER survey, the incidence land ownership (Box 1.4). Table 1.6 provides evidence of poverty was (Table 1.6): on this relationship based on the NCAER survey. In this survey, poverty rates are estimated using income as * 52 percent for the landless as a whole, the indicator of well-being, rather than consumption (as * 68 percent for landless wage-earners, in the NSS surveys). In 1994, 39 percent of the rural population as a whole had a per capita income below the Poverty : Where Does India Stand? 11 Expert Group poverty line (by coincidence this is almost that 49 percent of the poor are landless, while only 2 the same as the incidence of rural poverty of 36.7 percent of the poor are large farmers. percent based on the 1993-94 survey--there is no reason to expect such close correspondence between an income- 1.32 Although only about 37 percent of the rural based estimate and the one that is based on per capita population in 1994 were estimated to be in landless consumption). The first column of Table 1.6 indicates households, nearly half of the poor fell in that the incidence of poverty amongst the population with a category. Among the roughly 63 percent of the given household characteristic. This might be loosely population categorized as landowners, by contrast, interpreted as indicating the "risk" of poverty for those fewer than one household in three fell below the with a given characteristic and can be contrasted with the poverty line (Table 1.6). This provides confirmation of average "risk" of poverty in the population as a whole the close association of land with rural living standards. (39 percent according to the NCAER data, see bottom It also indicates that landlessness could be a valuable row). The "risk" of poverty among the landless is 52 correlate of poverty for targeting purposes, particularly percent and declines steadily across landownership when combined with information on the principal source groups to as low as 11 percent for large landowners. of household income. While the landless are clearly poor on average, Table 1.6 also makes clear that this group is heterogeneous. 1.33 The incidence of poverty was highest of all--68 Landless wage-earners face a particularly high risk of percent--among the landless wage-eamers who provide poverty (68 percent) while other landless households (for largely unskilled labor in markets where the prevalence example those who earn salaries) face a lower than of long-term contracts has been declining and wages average risk of poverty. The last column of Table 1.6 remain too low to lift casual laborers from the bottom indicates what percentage of the poor is made up of rungs of the ladder. Again, rural female laborers are persons with a given household characteristic, and may more likely than men to depend on daily wages from therefore be interpreted as the "contribution" to poverty manual employment. For every hundred women thus of a given characteristic. For example, the table shows employed, there are only 85 males earning their living L Sm2ulest percentage reduction in ooan 1972-i73 and 198w7-88 Haryana eastern 18.6 16.0 +16.5 UP southern 50.2 44.5 +12.7 UP eastern 42.7 41.6 +2.6 J&K helum valley 13.4 13.1 +2.4 Bihar central 51.9 52.9 -1.9 Rajasthan western 28.3 29.1 -2.6 Rajasthan north eastern 29.2 30.1 -3.1 Karnataka inland southern 31.9 33.5 -4.8 Bihar northern 53.0 56.3 -5.9 TN northern 52.9 56.3 -6.0 . Lagest percentaghe reduction in .o33 T973-i74nandcw98e7-88 Gujarat Saurashtra 16.9 30.2 -44.2 Maharashtra inland western 30.2 54.3 -44.3 Kerala southern 26.7 48.9 -45.5 Maharashtra coastal 29.2 61.6 -52.6 Gujarat eastern 33.4 71.5 -53.3 AP inland northern 21.5 46.9 -54.2 WB Himalaya 26.5 61.0 -56.5 AP coastal 17.0 39.8 -57.2 Karnataka coastal and ghats 10.7 39.6 -73.1 UP Himalayan 8.4 42.1 -80.1 Source: Drze and Srinivasan (1996). 12 Poverty: Where Does India Stand? in the same, marginal way, even though men outnumber women in India by a ratio of 1000 to 929. Box 1.4: Land Reforms and Poverty'Reduction In terms of earning power, men are more than twice as likely as women to hold salaried jobs in the large and Indepedn e pora of land reforms medium-sized towns that are increasingly important abolition of landed intermediaries. achie%ed limited centers of economic life in the Indian countryside. and %ariable success The consolidation of land holdings. also concei%~ed as. a component of ihe land 1.34 There is little evidence of rising landlessness redsrbt ontnm mt %thonidale-sucei over time. Secondary data and village studies find little in man" staic Moderate land reforms. as in Kera" evidence of a broad rise in landlessness over time. In Wesi Bengal and Kamaiaka. h3%e helped to diffuse fact, several studies observe some decline in economic and political concentration and proticed a more adequate basis for promoting panicipation. landlessness, often coinciding with an expansion of mpro%jng the prospecis for labor and accelrating small-holder cultivation. In general, these studies po%erti reduction But ambitious programs of tenancy indicate that while there is evidence of considerable reform and implementation of land ceilings have been movement in the distribution of land over time, much large u e.;4ul in mof ies Ohe her of this movement can be accounted for by poor record in human de%elopment and po%errN demographic factors such as household partitioning, reduction iBihar. Utiar Pradesh. MadhNa Pradesh. rather than land purchases or sales. Evidence from Rajasthan) are characterized b% unequal ox%nership of land. ah5entee land Lmncrship anid insecure tenancies. village studies suggests that land markets are rather In these states. legalising tenancie- (as ad%ocated in thin, and where there are land transactions, these are Lhe Ninth Plan Approach Paper) or ficilir3ting the often in the direction of small farmers purchasing land enin of the rural poor into the rural land market %%ill impro%e agriculiural gro%%ih and reduce po%ern, from large landowners. The village studies also show (Shanku. forthcomingi One srud BandopadhN3N. that sometimes land-to-the-tiller tenancy reforms and 19861 shos%s that an important factor affecting labor land ceiling legislation have led to an increase in market participation in recent decades has been the landlessness. Land legislation which seeks to provide toate ac th Rond the 1s ac0prcent permanence of tenure and ownership rights to tenants increase in lando%ning families compared to a 26 often prompts landlords to evict their tenants prior to percent increase in the number tot owning an% land Nearh~ one-fifth of the ne%% holdings were acquired the legislation taking effect. Similarly, land ceiling through modest lard retirms While man% ofthe new laws can have the indirect effect of inducing holdings %%ere small. theN ne%erthelesi enabled the landowners from joint-families to evict their tenants poor to achiete a modicum of securit% (in the firm of and to separate their properties into smaller proprietary a marketable asset and sm cerlaint, in income flows) and social sL3rus Even a small hutmeni lot units among family members. However, carefully with space tor a gardei can make a qualimi%e designed land reforms may have a role to play (Box difference in a household's li%elthood strateg). A 1.4) stud% b ICRISAT found that increased access to land plaNed a large role in the real income gains of the poor in the senii-arid zone in the 1970s and ihc early 1.35 As poverty declines rapidly with 1980s (World Bank 19901 More recent e%31uations landownership, it is not surprising to observe in Table carried out h. the National Academ% of 1.6 that the incidence of poverty among cultivating AdministraiOn ha%e confirmed the imporiance of households is slightly below the average "risk" of ef the oor poverty. However, because about half of the rural population belongs to households whose principal source of income is cultivation, more than a third of the 1.36 As discussed above, the incidence of poverty poor belong to cultivating households. From Table 1.6 it can be seen that poverty among cultivators is clustered in the marginal and small farmer landholding households which were landless and depended on categories, agricultural wage income, and those which were landless but not agricultural wage earners. When Poverty: Where Does India Stand? 13 dividing non-cultivating households into those reliant vulnerable segments. Landless agricultural, and non- mainly on casual wage income (both agriculture and agricultural, wage laborers, and marginal farmers form non-agriculture) and those for whom salaries are the the segments who are at considerable risk. This principal source of income, the incidence of poverty typology of occupational group and poverty--based on among the former group is as high as 65 percent while the NCAER survey data--confirms what has been that for the latter is 16 percent (Table 1.6). In terms of observed in village studies. Viflage studies offer, in their relative "contribution" to total poverty, salary addition, several other insights into the link between earners and wage earners, represent roughly 5 percent poverty and occupation. First, many village studies and 43 percent of the poor respectively. point out that casual wage labor is a "last resort", which households turn to only when other options have been 1.37 The rural population can thus be divided into exhausted. As such it is a good indicator of the relative four distinct groups based on access to land and position of a household in the income distribution. principal source of income. Large landowners and Second, households reliant on agricultural wages are salaried households are clearly two of the less generally least likely to experience significant upward Table 1.6: Characteristics of the Rural Poor Characteristic Incidence or poterty Approximate incidence Approximate percent in population of total poor' LandtHolding Landowners 0.31 0.63 0.50 Landless 0.52 0.37 0.49 Landless wage-earter 0.68 0.18 0.31 Marginal farmer 0.45 0.32 0.37 Small farmer 0.27 0.19 0.13 Medium farmer 0.16 0.10 0.04 Large farmer 0.11 0.06 0.02 Landless others 0.37 0.16 0.15 Occupational Group Cultivators 0.31 0.46 0.37 Salaried 0.16 0.13 0.05 Wage earners 0.65 0.26 0.43 All others 0.38 0.15 0.15 Social Group Scheduled tribes 0.51 0.11 0.15 Scheduled castes 0.50 0.22 0.28 Other Hindus 0.32 0.50 0.41 Religion Hindus 0.39 0.84 0.84 Muslims 0.43 0.10 0.11 Christians 0.27 0.03 0.02 Other minorities 0.34 0.04 0.04 Adult Literacy Group None literate 0.45 0.29 0.33 Female literate 0.43 0.03 0.03 Male literate 0.39 0.32 0.32 Both literate 0.27 0.35 0.24 All rural India 0.39 SIncidence in the population and percent of the poor are estimates only, based on two assumptions: First, that the firaction of sample households with a given characteristic provides an acceptable measure of the total incidence of households with that characteristic. This 'assumption had to be made, as sampling weights were not reported in NCAER (1996). It was possible to carry out some checks of this assumption (from information in other tables) and it appeared fairly sensible. Second, that the fraction of households is a reasonable proxy of the proportion of the population with a given characteristic. Source: NCAER (1996) Table 3.2, 3.16. 14 Poverty: Where Does India Stand? mobility over time. Although there have been some income which are available (Box 1.5). studies indicating that relative positions in the rural income distribution can vary markedly from year to 1.38 Both secondary and micro-study evidence year, agricultural labor households are generally less indicate that over time there has been a well-placed to achieve sustained upward mobility over "casualization" of the rural labor force (a growth in time. Such upward mobility, when it is observed, usually the size of the daily-wage labor force). This is occurs when a family member finds employment in probably a consequence of the decline in traditional, some regular, salaried, non-agricultural occupation. artisanal occupations, and also the gradual decline of Third, access to employment in well-remunerated non- the prevalence of permanent, long-term labor contracts. agricultural jobs depends on one's circle of contacts. As Does this finding, combined with the observation that part of the gradual transformation of the rural economy, casual labor is a "last resort", imply that rural poverty village elites have become more interested in non- is rising? Not necessarily. Although casual wage labor agricultural sources of income, so that the poor now face is a last resort activity (and therefore indicates that a stiffer competition for those non-agricultural sources of household is at the bottom of the village income Box 1.5 Labor Market Imperfections and the Distributional Impact or Non-Farm Employment in India The non-farm secior in rural India can beoan impoveant hicle for porert% aJle%iation Since the mid-I970s there has been a noticeable decline in the propontion of the "orkforce emiplo%ed in agricutiure in rural India This trend could be drisen b% rwo distinct processes On ihe one hand. igricultaral gromth ini rural areas and e%pandins relationships beneen urban emploers and rural wrkers. stimulates non-agr icult ural acikines through %anious forn%ard and back-.ard linkages. On die other hand. population grovth outstmpping epanaion of cultirtable land mi h push people Out of agniulture and into a residual. lo- producik it. non-farm -;ector. Both processes occur m rural India but on balance the e tpanding non-farm sector is more commonly characterized as a dNynamic onepu/ing rural %orkers ota of agricultture. than as a residual sector into ahtch desoitute agricultural eorkers hae been pioied. Ho%e%er. the degree to %%hich eien a dynamic non-farm sector helps the poor. depends on the facors which influence access ito non-Farm come E% idence from detailed studies in Gujlurar andtiar Pradesh. suggest that the poor are not generalt cm ell-placed to take adntages of such new opportunities In a nnd of si n urges in Nlehsana district. GuJarat. in 1995-96. Unni (19971 eannes the influence of indiidual, household and communitbl characterisics on the likelihood that an ndi%idual wrould e employed in the non-farm sector, as d%elf as on the eaings of that ind.idual from such empilment. She diingurelns between %\age and self-employmen sources of incomes and finds that both ctpe are associated with higher erage ncOMeS than agriculture Her findings indicate that education is a critical factor in influencing both opportunities for emplo me t as ell as eamngs from non-farm actities. Hoefer. mnni il997) identities se%eral factors which act as impediments io such employment opportunities. een aer controlling for the influence ofeducation and a host of other factors. Firsm although female participaton rates in the district are generally %er high. "omen are highly underrepresenied in the non-Carm sector Second. those without a circle of contacts i the non-farm sector seem ht be less wellpl3ced to gain access to such enployment opportuniues Finally, lower caste %illagers are sign-icanin less like to engage in non-adriculural aelf-enplo%ment aciities. and een if they do. they earn considerably less. A longitudinal study or one ilillage in Nloradabad district. Unar Pradesh. documents the striking expansion of regular non- farm emplo.ment opportunities beteen 1973 and 1993 Laajou and Stem. 1997 . Employment in. and earnings from such actities is posimel related to education and the number of adult males in the household, and negaticuek linked to landholditgs ndi%duals ho belong to the loest caste in the illage not only appear to face greater bamers to access. but also ear signiticaniy less if they do manage to gain enrlr Labor market imperfections thus seriously dampen the role that the non-farm sector has played in directy reducing p0oert). Een so. the sector is likeh to pla an important positae role in poertr alleviation. throua is indirect impact of hising age raite in casual agricultur l labor--the sector Uh many of India's rural poor are concentrated Source: Lanjouw, P and Stem, N. 199. Unni. 1J. 1997. Poverty: Where Does India Stand? 15 distribution, i.e. poor in relative terms), agricultural of the urban population are employed, along with wages have been rising over time. While non- employment in insecure, casual labor (Sen, 1996). agricultural wages are more difficult to track (since they derive from a much more heterogeneous set of 1.41 While there is clear evidence of gender and activities), NCAER (1996) offers some evidence that social inequalities in rural India, other inequalities of non-agricultural wage rates are somewhat higher than consumption or income appear relatively muted in agricultural wage rates. This is probably the comparaison with other developing countries. As consequence of agricultural intensification and the measured by the Gini coefficient applied to the 1987- expansion of non-agricultural employment 88 NSSO data, economic inequality within regions opportunities in rural areas. Material living conditions varies little from the poorest regions to the more of the poor can rise alongside the increase in their fortunate. In ten areas where the incidence of rural number. It is thus possible that the rise in the rural poverty was highest, the average Gini index was 0.26. wage labor force is the outcome of factors "pulling" In ten other regions where poverty was lowest, the laborers into this occupation rather than the corresponding figure was only 0.29, not a wide gap. consequence of distress "pushing" people into casual However, the Gini coefficient does not capture the labor. An important implication of these processes, gender and social inequalities which persist in India. however, is that over time dependence on casual wage These inequalities severely constrain the extent to labor has risen, and that more than in the past, the poor which certain groups in the population are able to stand to gain from increases in agricultural wages. participate in and benefit from the process of economic growth. 1.39 The NCAER survey also provides some recent evidence on the extent of employment among children Caste, ethnicity, religion andpoverty (see Box 2.2). According to this survey about 4 percent of males and 3.5 percent of females aged 6-14 reported 1.42 Inequality measures based on income or wage employment on a regular basis in 1994. This consumption blur the social reality of a nation where number could well be an underestimate of the true stratification based on caste, ethnicity, and gender extent to which children contribute to household remain significant. The NCAER survey reveals that income, as it is based on a fairly restrictive definition scheduled tribes and scheduled castes are both highly of regular wage employment. Work participation rates vulnerable to poverty in India. In 1994, roughly half of among children decline sharply with household the population of scheduled tribes aid scheduled castes incomes, landholdings, and education levels, were below the poverty line (Table 1.6). Together, these suggesting that this phenomenon is indeed a clear two groups represent only about one third of the rural indicator of household distress. population but about 43 percent of the poor. The vast majority of the rural population are of the Hindu 1.40 Because the employment data from the 1993-94 religion, and taken as a whole, this segment of the NSS survey are yet to be published and the NCAER population runs an average risk of poverty. Among the survey focused on rural areas, there is no recent non-Hindu religions, the Muslims residing in rural areas evidence on the occupational characteristics of the are the poorest, with an incidence of poverty of 43 urban poor. However, examination of the 1987-88 percent (Table 1.6). round of the NSS indicates that the principal occupations of the urban poor in 1987-88 were casual 1.43 Evidence from village studies indicate that labor and some self-employment activities. Self- caste relations have undergone signyl-cant evolution employment is obviously a highly heterogeneous over time. Rural growth, and in particular the occupational category, comprising both highly paid integration of the village economy with the wider rural professional occupations as well as informal sector economy, has led to some erosion of the traditional low-paying activities. It is in the latter that the poorest caste structure in many villages--in general there has been some relaxation in a number of the restrictions on 16 Poverty: Where Does India Stand? caste interaction which held in the past. In addition, in small households. For example, one radio might give there is a discernible breakdown in the correlation the same enjoyment to all family members, irrespective between ritual status and economic status. In many of whether the family is large or small. But the per capita villages, the traditionally dominant, non-cultivating cost of that radio is, of course, lower for the large family. (but landowning) caste is being overtaken in economic terms by households of other castes, usually those who 1.45 Dr ze and Srinivasan (1995) undertake a were traditionally cultivators and who have taken careful sensitivity analysis of poverty rates, based on greatest advantage of the new opportunities in the 1987-88 round of the NSS survey, using alternative agriculture. However, while such changes can have a sets of equivalence scales and possible economies of significant impact on the village social structure, they scale parameters. Their conclusions are quite do not impinge on the life of all castes in the same suggestive. First, when no adjustments are made they way. An important conclusion emerging from many confirm the finding mentioned above that large village studies is that the lowest castes remain easily households are more likely to be poor than small distinguishable from the rest of village society in terms households. They note also that widow-headed and of both their very low levels of material well-being, other female-headed households which are typically and the very limited opportunities they have to small, under conventional practice, are significantly improve their living standards. These studies further less likely to be poor than the average household (and show that even as poverty recedes, traditional particularly extended households). Second, inequalities largely remain. Paradoxically, in some experimenting with a wide range of equivalence scales, households where growing affluence leads to the they find that these conclusions are not qualitatively adoption of higher caste rituals and customs, women's altered. Third, however, they find that these freedom actually shrinks. conclusions are very sensitive to even small adjustments for economies of scale in consumption. Demographics,family structure and poverty Therefore, when even small adjustments for scale- economies are introduced, large households are no 1.44 Conventional approaches to the measurement of longer more likely to be poor, and single widows and poverty often indicate that poverty is concentrated widow-headed households appear to be at a high risk among large households. The NCAER survey suggests, of poverty. for example, that the incidence of poverty among those belonging to households of 8 or more members is 43 1.46 It should be emphasized that there exists no percent while those belonging to households of less consensus whether, or how much, to adjust incomes or than 4 members run only a 26 percent "risk" of consumption for economies of scale. However, the poverty. There are two caveats which should be observation that extended families are relatively less attached to these results. First, equivalence scales likely to be poor than average, and that female-headed should, ideally, be introduced to adjust for the fact that (particularly widow-headed) households, are highly different family members have different needs. And of vulnerable to poverty, confm findings from course, there may be some relationship between demographic and other studies of rural poverty household size and family composition such that (including village studies). For example, demographic adjustment with equivalence scales eliminates the link studies indicate that mortality rates among widows in between poverty and household size. Second, it is India are almost twice as high as among married possible that there exist economies of scale in women of the same age (Dr ze and Srinivasan, 1995). consumption, such that the per capita cost of reaching a At a minimum, caution should be exercised whenever certain welfare level is lower in large households than poverty is linked to demographic variables. 4 GROWTH, HUMAN DEVELOPMENT, AND ANTI-POVERTY PROGRAMS A. Introduction 2.3 This chapter focuses on rural poverty, examines its determinants both at the national and state levels, 2.1 The extent to which poor people share in and the relative performance of anti-poverty programs. economic growth has been one of the most contested This focus on rural poverty is not intended to overlook issues both inside and outside India. Some observers the importance of urban poverty in India which have argued that "distribution must get worse before it according to some is worsening. Urban poverty is gets better" in developing countries, and that this puts a much less understood and articulating policies to severe brake on the prospects for pro-poor economic address it requires further research in this area. growth. There have also been debates about the effects of growth in specific sectors. For example, some have B. Growing Out of Poverty argued that the benefits of the "green revolution" (which resulted in substantial gains in agricultural 2.4 Expansion in some sectors of India's economy yields through new seed varieties and irrigation) were can mean little or nothing to the reduction of poverty in captured by relatively well-off farmers, and brought the cities or the countryside, but broad consumption little or no gain to the rural poor--and that, in some advances in rural India bring benefits to the poor there, cases, more capital intensive modes of production to the urban poor, and to the economy as a whole. In displaced labor and actually accentuated poverty the nearly four decades from 1950 through 1988, for ("immiserising growth"). Others have pointed to farm- instance, the secondary (mainly manufacturing) sector output growth as the key to poverty reduction. grew at an average annual rate of 5.3 percent but, probably because of the capital intensity of this growth, 2.2 In India, the growth performance of the first this sector generated relatively little new employment two decades after independence (aggregate GDP per for the poor. Agriculture growth, though only about capita grew at the rate of 1.5 percent per annum half as fast--2.3 percent a year--accounted for the bulk between 1950-51 and 1973-74) and the concurrent of the decline in poverty, with improvements in rural weak pace of poverty reduction led to widespread living standards also having an effect on those in the skepticism regarding the "trickle down" effect of urban sector. growth on poverty. Since 1973-74, however, poverty reduction has been associated with a higher rate of 2.5 Results from a World Bank research study growth (between 1973-74 and 1993-94, GDP per (Ravallion & Datt, 1996a, 1996b, and Datt 1997) using capita grew at 2.0 percent per annum). Moreover, the NSS data show that overall growth accounted for the earlier view that growth (industrial or agricultural) was lion's share of poverty reduction: 80 percent of the both immiserising and inequalising has given way to a decline in the headcount index over a 40-year period greater appreciation of the contribution of growth to and almost 100 percent since 1970 (Appendix, Table lasting poverty reduction and of the links between A.5 and Figure 2.1 and 2.2). Improvements m human resource development and growth. distribution (which occurred either through deliberate government policies or independently) achieved their 18 Growth, Human Development, and Anti-Poverty Programs Figure 2.1.: Cumulative Change in Headcount Index in India Total Change and Growth and Redistribution Components 15.0 10.0 -5.0 + -10.0 - - - - Cun. 6ro%%ih Componen - Cum Redisrr Loiponeni -15.0 \/ Cumularie Change in IleadCouni -20.0 ;, 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 Source: Datt, 1997. greatest impact by the mid-1960s especially on reducing the intensity of poverty (but with little impact Ronsh is Peyon? on poverty incidence as measured by the headcount index). The gains to the rural poor since 1970 have Percentage change in the poverty measure attributable to a 101 been due almost entirely to growth; neither urban percent incTease in mean net growth nor rural-to-urban migration had any Index domestic product discernible effect on the standards of living of the rural Headcount -9.9 poor. On the other hand, the fortunes of the poor in urban areas are also linked to those in rural areas through trade, migration and remittances. As a result of Squared Poverty Gap -18.9 these links, rural economic growth also helped reduce 2.7 Two main messages emerge. First, India's past urban poverty. The results of this research clearly development strategy--focused on a strategy of capital refutes any presumption of immiserizing growth in intensive industrialization behind high trade barriers India over this forty-year period. brought negligible gains to the nation's poor. The jobs 2.6it created required skilled labor few of them could sector was responsible for three fourths of the long- growTh a gothasn th doinan frin term decline in poverty but that strong rural growth povty ructi (aabein tt, d force S n would have a dramatic impact on Indian poverty. A investmetin iRovin arc t producond ten-percent increase in mean net domestic product, for employment, by con aveintuol rad ing instance, was projected to bring a 10 percent drop in stadarsen ra cndat have one s wihou the proportion of people below the poverty line and, accentuating inequalities among farmers. A study (Dat given India's population growth rate, a corresponding 10-11 percent drop in the actual number of the poor showed a consistent relationship between per-hectare (Table 2.1). agricultural output and the incidence, depth, and Growth, Human Development, and 19 Anti-Poverty Programs Figure 2.2: Sectoral Decomposition of Headcount Index Change in India Total Change and the Rural, Urban Components 10 -5 S -5 -10 -Cum Rural Component -Cum. Ulrban Component \ - -Cum. Populanon hifi -15- - - Cum Interacuon Component - Cum. Change in HeadCountj -20. I 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 Source: Datt, 1997. severity of poverty. Green revolution technology, households in the informal sector, or food-deficit initially suspected of conferring systematic advantages farmers. Almost 76 percent of their budgets, on to larger landholders, has proven neutral in terms of average, is spent on food, a slightly higher proportion inequality and this is confirmed by a range of village than in comparable urban households where, however, studies. It has, in fact, contributed to a significant the negative impact of inflation tends to be increase in rural non-farm employment (cottage and significantly harsher. Because nominal agricultural agro-industries, services and commerce, primarily wages do not rise as rapidly as inflation (only about outside the organized sector) which may have played a two-thirds as fast), a 20 percent increase in prices in role in reducing rural poverty through its effect on real rural areas in a given year would produce a 5-percent wages by tightening the rural labor markets (Sen, rise in poverty as measured by the headcount index and 1996). a 9 percent jump on the squared poverty gap scale. 2.8 Two negative factors--low rates of female 2.9 The dfferences in the rate of poverty reduction literacy, high rates of inflation--also affected progress across states are also striking (see Chapter t). Between against poverty. Where fewer women were literate, the 1960 and 1993 some states reduced poverty much decline of poverty was slower. A study (Datt & faster than others (Figure 2.3). What accounts for these Ravallion, 1t996b) shows that a 20 percent increase in differences? Initial endowments in human capital and female literacy adds one-tenth of a percent to annual physical infrastructure and subsequent efforts in rates of decline in the headcount index, or 0.2 percent developing these resources appear to have been an to the drop in the squared poverty gap index. Higher important factor. A study by Datt & Ravallion (1996b), prices have an even more dramatic effect on poor rural shows that two main approaches to reducing poverty households. This is not surprising since the bulk of the can be identified from India's experience. One is poor are wage dependent households, self-employed through rural economic growth which when combined 20 Growth, Human Development, and Anti-Poverty Programs Figure 2.3: Progress in Poverty Reduction by States I']Average over 1957-60 70% - Average over 1990-94 60% 50% + 40% 2 30% 20% 10% 91. 00/ Source: Datt, 1997 with good initial conditions in physical infrastructure, produced significant reductions in poverty in states such as Punjab and Haryana. The second approach relied on human resource development. This has Differing performance, in rural po%erty reduction allowed for instance Kerala to reduce its poverty %%ere due to both differing initial conditions in human incidence--through increased exports of relatively reure ot agripultur al instactue skilled labor and the resulting inflow of remittances incidence of po%ertN in Bihar declined at an annual despite modest economic performance. However, some rate 2.1 percentage points belo%% that in Kerala. but states such as Bihar failed on both counts: there was onl% about hAf of that diferential is explained b) the statesinitial conditions. Other factors. particularly the slow too little growth and human resources were gro%%th in agricultural output per hectare. ha%e been underdeveloped. These states made the least progress important in e\plaining Bihars unimpressive in the fight against poverty. The important message performance Iris nonetheless notable that if Bihar had sttarted otf %%iih Kerals. level of human resource here is that no Indian state effectively combined both de%elopmeni in ihe 1961is. the differential in the rates approaches. The study suggests that if any state had of po, erN reduction bemeen the two states could done so, it would have achieved rapid reductions in haiie been narro%ed to less than half their observed poverty, comparable to the progress made in a number le%e. Als. the implicit trade-offs can be large. For Bihar to oLerconle the adverse effects of its initially of East Asian countries (Box 2.1). disad,antageous human resource de%elopment relative to Kerala %.'ould ha%e required that its agricultural 2.10 From the results of this research, the lessons yields gre%% annually at a rare 3.4 percentage points for the future are clear. A more open trade regime hiheranala's The t% rels inhibit since 1991, a significant reduction in distortions, and a its rare of po)eN reduction. Suppose that Kerala had liberalized economy where the private sector plays a the sami trend P-rotih rates in farm yields as Punjab- larger role have improved agricultural incentives and Haryana The study indicates that Keralas trend rate of largr rle ave mprvedagriultralreduction in the po%erv, rate Lwsuld have been 3.1 are creating the conditions for labor intensive percent per Near. rather than 2.3 percent. industrialization. This will not only enhance the country's growth prospects but also increase the Source. Dart and Ravallton. 1996b poverty reducing effects of industrialization and the Growth, Human Development, and 21 Anti-Poverty Programs role of urban areas in this process. After the buffeting contributed to the eradication of once chronic famines. of the first years of structural adjustment, GDP growth The results were particularly noticeable in the 1980s, has now averaged 7 percent in the last three years. when, for the first time in India's history, agriculture These are healthy rates. Along with robust agricultural grew at a significantly higher rate than the population. performance, they open the way for a concerted effort Since 1980 agricultural growth has increased and aimed at maximizing the beneficial impact of growth spread across regions (to the East), crops and itself on rural poverty and most likely on urban agricultural activities. Because it was scale neutral, and poverty. The ingredients of that effort--the necessary it increased real rural wages, it contributed to an investments in agriculture, in infrastructure and, above important reduction in poverty. Government financial all, in human resources--are discussed in the following support was an important element of this expansion. section. 2.13 It is clear, however, that subsidization of C. Capital for Rural Growth... current inputs (as well as of capital costs in irrigation and power) is a major problem During the 1 980s, 2.11 Rural investment make inroads against rural agricultural subsidies grew three times faster than poverty. Important examples include roads, well- growth enhancing expenditures (which include all designed irrigation and drainage systems, flood public spending on economic services in agriculture control, and electrification. So also do investments in and its allied sub-sectors, and on new irrigation people, in their health, their education, their nutrition investments) and now form the largest share of and their ability to take part in shaping their economic agricultural spending (Figure 2.4). During 1991-1996, and social destiny. Since the bulk of these investments the growth rate of agricultural subsidies was halved, in physical and human capital come from public rather but remained high at 3.6 percent per annum in real than private resources in developing countries such as terms. Power and irrigation subsidies that are largely India, it is important that poverty-reduction strategies under the state governments' jurisdiction account for strike and maintain the right balance between more than half of all these agricultural outlays while investments that return significant dividends and the other subsidies continue to grow at a rate of 3.6 subsidies that can distort markets and divert financing percent a year. This pattern persists despite the central from human needs. India has not yet found that government's success in virtually eliminating explicit balance. A recent government study has put subsidies credit subsidies by 1994 and reducing those for on "non-merit" goods at 11 percent of GDP (of which fertilizer from 0.9 percent of GDP in 1990-91 to 0.6 4 percent of GDP are accounted for the center and the percent in 1995-96. It also held food subsidies in those remaining 7 percent by the states)--whereas total public years to 0.5 percent of GDP. But what central spending on health and education is less than 5 percent authorities achieved, state governments undercut. As of GDP and even less on infrastructure. The2 resait sFigure 2.4: Subsidies Continue to Crowd Out when,aforsthofiProductivity Enhancing Expenditures, in the world spends--over one-fourth of agricultural Rs, billion 1981-1982 GDP (8 percent of GDP)--as much on agriculture asa S200 India. Farmers benefited from large public spending for infrastructure (irrigation schemes, rural roads), support services (research and extension) and subsidies S50 (for fertilizer, credit, water and power for irrigation(as pumping). Intensive public spending on agriculture led inter alia to the spread of irrigation from coverage ofi 22 million hectares in 1951 to 78 million in 1996. g Subsidy Productivity enhancing Safety Nets These outlays helped agricultural growth and Source: Ministry of Finance, Union budget documents. 22 Growth, Human Development, and Anti-Poverty Programs their share of subsidy financing rose from 40 percent Underground aquifers are being depleted in regions of the total in the early 1980s to 56 percent currently, such as the semi-arid states of Rajasthan and Punjab, the states allowed power subsidies to grow at annual India's breadbasket, by unsustainable use of water to rates of 14 percent in real terms between 1990-91 and irrigate wheat and rice fields where high-yield crops 1994-95. State financing for irrigation expanded much that mature early permit two harvests a year (State of more slowly--at a yearly pace of just 1.2 percent--but the World 1997, Worldwatch Institute). Fertilizer use the credits too often went to schemes of dubious which generally rises with the availability of irrigation quality and reliability. can become a major source of groundwater pollution and, unless calibrated to soil nutrient levels, can easily 2.14 These resource-allocation decisions matter a go beyond the point of diminishing returs. While great deal to poverty reduction in rural India because India's fertilizer use is still below optimum, untargeted overall public investment in agriculture has actually subsidy programs carry the seeds of long-term stagnated for approximately the last decade and a half. ecological harm. Its average growth since the early 1980s has been only 2 percent a year in real terms, and capital formation, D. ...and for the Rural Poor having peaked then at 14 percent of all agricultural GDP, dropped to 7.7 percent in 1993-94. Under such 2.16 In the short term, most importantly, subsidies circumstances, agricultural subsidies which do not that go to these activities represent funds withheld achieve their intended effect or which, all too often, from investment in infrastructure and human favor richer states and well-to-do cultivators over resources. Box 2.2 highlights the importance of poorer regions and farmers, handicap efforts to reduce expanding the coverage of rural infrastructure. This poverty. They channel scarce resources away from report has shown the importance of literacy, for their most productive uses and slow or impede growth instance, on poverty-reduction. China, it can be which is essential to poverty reduction (Box 2.2). argued, laid the foundation for the success of its economic reforms with programs that, by 198 1, raised literacy rates for 15-19-year-olds to 96 percent for Box: 2.2: Growth-Enhancing Public Inestment males and 85 percent for females, compared to India's percentages in the same categories of 66 and 43. By the ioernment in%esiment in agriculture in less faorable same year, life expectancy at birth in China stood at 68 areas is likely to have a greater impact on both agricultural growth and poverty reduction. According to a study by Fan and Hazell (1996), the marginal infant mortality rate was 37 per thousand live births. productivity effects of additional government India's was 110. Between 1978 and 1990, as reforms investments in HYVs, roads, and irrigation, have took hold, the headcount index of poverty in China (by declined in the past decade in, irrigated areas and the marginal benefits from such investment are now China's poverty line, which need not have the same greater in rainfed areas. Further disaggregation of purchasing power as India's) dropped from 33 to 11 rainfed areas into high and low potential areas showed percent, while the decline in India was a much more that the poverty reduction effects of increases in modest one: from 55 percent in 1977-78 to 42 percent HYVs, irrigation and education are higher in low than in high potential areas. Therefore. in rainfed areas. in 1988-89. government invesIments no[ only impro%e productivity but also reduce rural po%erti. In iddition. 2.17 Education can be a powerful weapon against the study finds that iniproed literacN has a potet reducing effect no mater Lhe source of n%aier (irrigation or rainfall) for crops. poverty indicators similarly show the slowest declines. Two-thirds of all Indian women--poor and non-poor-- 2.15 In many areas one particular harmful effect of remain illiterate compared to 37.6 percent of all men. unchecked subsidization of fertilizer use and irrigation Among the most disadvantaged social groups, the schemes has been environmental degradation. ratios are even worse: literacy rates of just 19 percent Growth, Human Development, and 23 Anti-Poverty Programs among scheduled-caste women and of 46 percent for another area where the gender gap in India is acute, men. Even within India, the link between social and making Indian women, 80 percent of whom give birth economic progress is manifest in a state such as Kerala at home, 200 times more vulnerable to maternity- where high female life expectancy (74 years at birth) related death than their counterparts in the developed and literacy (86 percent) accompany the country's world. Almost half of all pregnant women in 1992-93 strongest performance in poverty-reduction--an annual gave birth without having a single antenatal checkup, decline of 2.4 percent in the headcount index from and 24 percent of under two years old children of 1957-58 to 1993-94. Bihar, the weakest of the large illiterate women received immunizations. states in long-term poverty-reduction, spent barely one- Unsurprisingly, while only 2 percent of Keralan third of Kerala's per capita rate on education in 1986- women did not get antenatal checkups, 46.6 and 54.5 87 and achieved a female-literacy level of just 18.2 percent of those in Madhya Pradesh and Uttar Pradesh percent. Rajasthan, with only a 20 percent level of did not, respectively. In Bihar, 54 percent of the adult female literacy, turned in one of the worst 16- children went without immunization, and among the year performances in reducing poverty, an average offspring of illiterate women only 5.4 percent had full annual decline of only 0.5 percent. And Madhya immunization coverage (National Health Survey, Pradesh, where female life expectancy is 20 years 1995). below Kerala's level, managed only a 0.4 percent annual reduction in poverty from 1957-58 to 1993-94. 2.20 Communicable diseases and prenatal and maternal mortality, which account for 12.5 percent of 2.18 The relation between education and well being the annual deaths of rural women aged 15-45, cause is too obvious from the history of development to be a about 470 deaths per 100,000 population in India--a subject of debate. One Indian study (Loh, 1995) even rate 4 times that of China and 2.5 times that of the found that the impact of a single additional year of world as a whole. Tuberculosis alone kills about workforce education was a 15-percent increase in 500,000 Indians a year, a staggering figure, reflecting economic output. And while higher levels of education the fact that as many as 12 million people in the may have a greater development impact, primary and country may suffer from active TB at any given time. secondary schooling are important in improving well- Another 2.5 million are believed to be infected with being. Yet, around 33 million out of 105 million Indian HIV now, and the number could rise to 12 million in children aged 6-10 years do not go to school. Most of the year 2000. them, in fact, work--sometimes for their families at home or in the fields, sometimes as wage earners. 2.21 Some of those statistics reflect failures in the Estimates of the numbers in the child labor market vary field of nutrition, where India has made great with definitions and range from 17 million to 44 progress in rendering famines a thing of the past but million with the latter figure including youngsters paid has not brought regular food supplies to the poor up in kind, rather than in cash (Box 2.3). The poorer the to proper levels. Thus, half of all children under the household, the more likely it is to use children to bring age of five are malnourished and, because their home wages and the more likely that labor is to be mothers often are as well, one-third of India's performed by girls. The decision to send a child to newborns are of low birthweight. Maternal mortality school or to work turns on largely a calculation of the rates are associated with the anemia that affects about return one activity or the other will bring to the three fourths of all pregnant women (and 60 percent of household. In the case of Indian girls, it appears that in the non-pregnant, as well). poorer families there is next to no contest. 2.22 Given the importance of health and education to 2.19 Especially for women, educational advance is growth prospects and poverty reduction, India has paid known to lead to better health, to smaller family size less attention than it should to improving the financing and hence to the health of children. But health is of these sectors. Much of the fiscal responsibility for 24 Growth, Human Development, and Anti-Poverty Programs Box 2.3: Child Labor In India, 33 million out of 105 million children aged 6-I') %ears (World Bank. 19 71 do not go to school Aany of those not in school are working: nearly 18 million as esiimjted bN the 1991 Census enter the labor market. while a study sponsored by the Labor Ministry reports a child labor force of 44 million Other esitnuaes of child labor in India, taking into account all household work of children. place the number of working children to as high as 115 million. Different micro studies su%gest that whereas boys are more likel% io be in%oked in wage occupations. girls in general ha%e a larger share in the o%erall burden of child labor. As a result female children are more often remo%ed from schools to do household %tork The majorai of child laborers are emploed in agnculure. but child labor also constitutes a backbone of seeral industries, particularly in carpet w%eating. local cigarettes production and silk %%eating employing together 850.000 children. Child laboris a widespread phenomenon associated with potent . Economic anal%,is and international experience suggest possible approaches to this issue. Front the econonic perspecte. the incomes and the number of children in the household determine the potential supply ofchild workers. As the welfare of the household increases. the income effect of parents' wages will outweigh the incenmises to suppli child labor by poor households This suggests that equitable economic growth will eventually reduce child labor Also on the suppr side, child labor can be part of a strateg to minimize the risk of job loss by a famil; member. or a failed har%est Attempts at its forced abolition are likelN to fail unless accompanied by provision of an adequate safetN net On the demand side. children substitute for adults because of significant wage differentials due to labor market imperfections Actions to break local monopolies on the labor market and to enforce labor laws would have a direct impact on the pre% alence of child labor. The existence of child labor could also be regarded as a result of discrepanc between the percet%ed prto ate returns to education and social returns. It might be a result of a failure in the market for education. or. een without a market failure. it might be determined bN discrimination against children within the household Bilth reasons can be affected b\ simultaneouslN banning child labor and enforcing compulson educatior.. and economic intentites for families to enroll children in schools. In the large spectrum of man% forms of child labor the mosi onerous are bonded child labor and child labor in hazardous occupaiions. Both are formally forbidden in India. But despite the ban on bonded labor in the Bonded Labor Sy stem Abolition Act of 1976 there are an oitimated 15 million children in bondage. %orking almost as slaves who serve as mortgage or repay.ment of their parent)' debts It is well known that bonded labor is concentrated in particular industries and geographical regions Thus the first step %ould be the proper enforcement of etisung laws The Child Labor Prohibition Act of 198t restricts the employment of chilJren in selected industries. but otherwise permits children of ant age tc. %%ork in agriculture. small-scale industries and in sert ice sector e%en in hazardous occupations. The tightening of the e\istine laws. H ould be necessar for limiting the most debilitating forms of child labor. But so long as the numbers of children in the labor force are a- high as they are. a legal ban on emplc.yment. though helpful. is ltkelN to hate only limited impact on child labor. Inierioanonal e\perience -uggests that enioreing compulsor% education was a necessar% condition for the reduction and abolition of child labor Countries such as Taiwan. South Korea, Indonesia and Japan--that hate achie\ed success in drasticall\ reducing child labor--enforced compulsorN education laws when their per-capita incomet were lo%%. po%err\ udespread. school facilites inadequate and parents would hate employed their children had they been permitted to do so. Sources: Human Rights Watch, 1996, Grootaert, C. and Kanbur. R. 1995 both health and education lies with the states, and the Further, not only did the share of health care costs range of their performance is predictably wide. But, on within total household budgets increase several fold in a national basis, real outlays actually have fallen since the 25 years since 1963, the burden grew heavier and the mid-I 980s. faster in poorer states. In Uttar Pradesh, for instance, the share of household expenditure on health increased 2.23 Much of the cost of health care--78 percent-- from 2.1 percent to 7.2 percent between 1963-64 and comes out of private pockets (World Bank, 1997a). 1987-88, with the percentage spent on medicines alone Household health care expenditure imposes a heavier more than doubling from 1.7 to 3.9 percent burden on the poor--for whom it has emerged as the (Srivastava, 1997). largest source of indebtedness--than on the well-to-do. Growth, Human Development, and 25 Anti-Poverty Programs 2.24 Even care in government hospitals--roughly 2.25 India has recognized the need for reform in half as expensive as private ones for a day of both education and health. Central government hospitalization--can be onerous, with a single episode (nominal) expenditures for elementary education did of illness requiring hospitalization eating up three to increase by about 190 percent between 1990-91 and six months of a bread-winner's earnings. The burden 1995-96. Moreover, there is a commitment to raise the of hospital treatment (as a percentage of annual per total budgetary allocation to education from the current capita consumption expenditure) equals 230 percent 3.7 percent of GDP to 6 percent by 2002--an objective for the poorest tenth of the population in states such as that can only be met with significant reforms at the Uttar Pradesh, Punjab, Rajasthan and Bihar even when state level. And along with special efforts to enroll the hospital is a public one. In contrast, the burden in female children and eradicate child labor, the Kerala for those at the bottom of the income scale is government is steadily expanding a District just 10 percent of annual per capita expenditure. And Programme for Elementary Education (DPEP) that for India's wealthiest 10 percent, the cost of treatment began in 41 districts marked by low female literacy. in private rather than public hospitals in all states But much more is needed just at the primary level, as except Bihar ranges as low as 3 percent of per capita much as Rs. 196 billion in 2007 according to a World consumption and only as high as 40 percent Bank study (Box 2.4). A steady economic growth rate (Srivastava, 1997). of 5 percent per year over the next ten years could generate those resources and more to improve primary Bo( 2.4n: Resource Requirements for Qualie Education Provi-oonal e .tiniaies uf re;ource rcquiremers f%)r pro% iding reasonable qualirN schooling for all children aged 6-10 Nears b% the Near 2(n). have been miade b [he World Bank i\%orld Bank 1997 draring hea- A on research carried out b% the DPEP bureau of the Department 11 Education based cr three oblectioes maintainin the e\i in system to accommodate the children age 6 to 10 year;. in ';L0. expanding the Nstern. Lind i mpro%ing 3he qualip o of schooling b the last oear of the Tenth Five Year Plan h 2soi r The cost of simply rehaNlitating the c,.i!ming s-,4tem %vuld require an increane from Rb 80 billion in 1996 to more than Rs. 121 billion in 2007 (11Q3 pricc ) The proicctions belo%% s;ho%% a set of scenarios, under different assumptions for GDP growth and the additional eNirnaed 0-ri Otepandine [lie semgto acommodate all studentai aged 6-10 Nears currently out of school. Based on reast-nable a.S rrnptions regardirip the cost of classrooms to he built, teachers' salaries (annual real salary increase of4 percenti. pro% or of book' and Lithk:r inceniies. total requirements are projected at Rs. 195 billion in 2007 at 1993 prices. This include- impro% ng schooling qualir. through rebuilding or rehabilitation of classrooms in need of major repairs, provisicr of drinking %aer and toiler facilities. pro, ision of instructional material and teacher training. Approximate estimation -ug ets that ihe required resources are mothin reach if India grows b 5 percent in real terms and share of GDP spent on education incrvases Ifthe econom%. erox%'s slo%ver, there will he. of course, a Financing gap. Resource Requirement. A%ailability and Financing Gaps for limproed Primary Education System b) the )ear 2 Bnk oinder Different Scenarios GDP Feachers d growth salaries Resource Resource F h inancing gap rate growth rate RequirementP availability or surplus pere n I ncRs hillion Scenario I I 4 4 195 182 -13 Scenario? 2 4 09 2i7 +12 Ssenadio 3 6 4 195 2 44 +49 Notes: 1. Cost of epanding primar educaon s se iitn impro ed quatr of schooling 2. This assume risine total budmepn allocations or education trom the current 3 7 percent o6 percent of blDP bc 2002 and a constant ;hare of the total education budget going to prima) education Source: Primary Education in India kk orld Bank. 1997 26 Growth, Human Development, and Anti-Poverty Programs education, but even that infusion of resources would E. Poverty-Reduction: The Missed Target not enable the states where 75 percent of all of India's youngest children are now out of school to achieve 2.28 India's current poverty reduction strategy universal enrollment and higher quality education in comprises a wide range of numerous programs 2007. To meet that goal, Andra Pradesh, Uttar Pradesh, financed through the central and state budgets. Some Bihar, Rajasthan, Madhya Pradesh, and West Bengal are relatively new. Others have been in effect for would have to increase their real annual outlays for decades. Few programs are abandoned, and new primary education by about 13 percent. programs appear with each budget cycle or change in g overnment. The end result is an overwhelming array 2.26 A similar dilemma confronts the states with the of large, complex, and sometimes innovative programs lowest levels of health care and--to a degree--all of that involve a range of (primarily government) India, where per capita government spending of US$2- agencies at the federal and state level, and an even 3 a year on primary health care is too low to meet the wider range of actors (local government, cooperatives, 50-year-old promise of insuring universal access to private distributors, NGOs and voluntary care in Primary Health Centers (for preventive and organizations) at the local level. basic services), district hospitals and specialized referral (tertiary) facilities. Health budgets are being 2.29 While referred to as anti-poverty programs, squeezed throughout the country, but the pinch is often many of India's largest and longest-lived initiatives tightest in the poorest states where, as in most others, have a broader set of development and stabilization the costly services of the tertiary sector soak up a objectives. Each program defines objectives, disproportionate share of resources. Also in most beneficiaries and their needs in its own manner and states, outlays for salaries run at such high levels that only in some cases (i.e. the Integrated Rural operations and maintenance activities are chronically Development Program--IRDP) does this include under-funded. For its part, the central government has poverty reduction and the poor per se- Actual made marginal shifts in its support programs to favor beneficiaries may include both poor and non-poor preventive health care and the control of households. One well known example is the Public communicable diseases. Providing a package of Distribution System (PDS)--initially not intended as an essential public health and clinical services would need anti-poverty program. Anyone with a ration card can a doubling of the current 0.6 percent of GDP obtain PDS supplies at subsidized prices, and ration government spending on primary health care. cards are--in theory--available to all people in India. In practice, the poor are no more likely to get PDS 2.27 However, in the fiscally straitened subsidized goods than middle-class consumers. circumstances of many states, it may be more realistic Similarly, the bulk of input subsidies go to larger in the short term to aim for a phased reorientation of farmers who have marketable surpluses. outlays so that three fourths of them go to the primary and secondary sectors with priority given to the 2.30 Table 2.2 presents central budgetary provision of medicines and essential supplies. Also, expenditures for the broad set of programs officially increasing user charges could boost revenues from classified as anti-poverty. Typically, state-level them from present levels of 3 percent of current spending represents an additional 25-50 percent. expenditure to 15-20 percent. Other reforms stressing According to these estimates, the share of the central improvement in the quality of care and availability of budget devoted to anti-poverty programs increased basic health services would make it easier to raise such from approximately 5.4 percent in 1990-91 to 7.7 user fees at the same time that the private sector is percent in 1995-96. Most of the increase occurred encouraged to expand its role in the provision of between 1992-93 and 1993-94. Food subsidies, rural specialized (and costly) tertiary services. employment programs, and social welfare and nutrition programs have contributed disproportionately to this Growth, Human Development, and 27 Anti-Poverty Programs Table 2.2: Central Plan Budgetary Expenditures on Anti-Poverty Programs (APP): 1990-91 to 1995-96 ipercent of total expenditure) Food Rural IRDP Social Welfare Total Center Subsid%' Emplo-, ment and Wel fare & NRY of SC/STs Center APP and Programs RelaItcd Nutrition APP Basic Needs" 1990-91 2.3 1.9 0.4 0.3 0.1 0.3 5.4 6.2 1991-92 2.6 1.6 0.3 0.4 0.1 0.4 5.4 6.4 1992-93 2.3 2.1 0.3 0.4 0.1 0.4 5.6 6.3 1993-94 3.9 2.8 0.5 0.4 0.1 0.4 8.0 9.1 1994-95 3.2 2.9 0.4 0.9 0.0 0.4 7.9 9.0 1995-96 3.0 3.0 0.4 0.8 0.0 0.4 7.7 9.0 1.Non-plan expenditure. 2.Basic needs programs include outlays for elementary education, adult education, accelerated rural water supply program, rural sanitation, urban water supply and sanitation, child welfare (incl. mid-day meals program), backward area development programs (drought prone, desert, wastelands). Source: Based on Government data. rise in central government spending, with income quantity of wheat and nearly 47 percent of the total transfer and welfare programs--in particular, food quantity of rice supplied through the PDS is consumed subsidies and to a lesser degree nutrition and welfare by the poorest 40 percent of people in India. This programs--accounting for over half of all spending on would be the expected pattern if the PDS were not at the Anti-Poverty Programs by the central government. all targeted to the poor but rather distributed its rations The Public Distribution System, one of India's oldest randomly. In addition, the program is not a cost and most far-reaching public safety net programs, effective way of transferring incomes to the poor as distributes rice, wheat, edible oil, kerosene, and sugar compared with other nutrition programs (Radhakrishna at subsidized prices through a network of some & Subbarao, 1997). Box 2.5 summarizes available 400,000 fair price shops in urban and rural areas evidence. throughout India. Subsidies in the PDS have stabilized at around 0.5 percent of GDP in recent years. 2.32 There is a long running debate in India on how Administrative costs in real terms have more than to make the PDS more efficient and cost-effective, how doubled over this period. Retail prices are fixed by to improve targeting to make it more pro-poor, and each state government after adding local distribution more generally how to lower its costs. A new PDS costs and retailers margins to the Food Corporation of scheme was launched in 1992--the Revamped PDS--to India (FCI) issue price. However, some states provide provide additional food subsidies in 1752 additional subsidies (IRs/Kg in Orissa and Karnataka, disadvantaged blocks. A second, major revision has and 5Rs/Kg in AP on rice in 1995-96). been proposed under the current budget that would grant poor households up to 10 kg of cereals at half 2.31 If the PDS functions as designed, the consumer the prevailing economic cost. Subsidies for households is protected against price fluctuations in the open with higher income levels would gradually be phased market, the farmer is protected against excessive losses out. Under the TPDS, each state allocation (and so the in periods of surplus production, and the public at large PDS subsidy) would be directly linked to that state's is protected against food shortages and famine in poor share in total poverty (Box 2.6). harvest years. But what of India's poor? States making the greatest use of the PDS are not necessarily the 2.33 The government also is considering targeting poorest (Table 2.3). Nor are the poor any more likely PDS benefits to the poor without additional cost to to consume basic foodgrains distributed through the persons who utilize other anti-poverty programs where PDS. One of the many studies on the PDS (Ahluwalia, self-selection works well. For example, the low wage 1993), estimates that only about 40 percent of the total rates of employment programs make it very likely that 28 Growth, Human Development, and Anti-Poverty Programs Table 2.3: Public Distribution System: Allocation of Subsidy on Cereals Between States Actual States' Shares in Would Be shares Based on Population Shares PDS Subsidy, 1993-94 States' Poverty incidence Percent in 1991 All India 100 100 100 Andhra Pradesh 13 5 8 Assam 4 3 3 Bihar 5 16 10 Gujarat 4 3 5 Haryana 3 1 2 Karnataka 6 5 5 Kerala 10 2 3 Madhya Pradesh 4 9 8 Maharashtra 8 9 9 Orissa 3 5 4 Punjab 2 1 2 Rajasthan 5 4 5 Tamil Nadu 6 6 7 Uttar Pradesh 8 18 16 West Bengal 7 8 8 Other States/Territories 12 4 4 Source: Based on Government data. participants will be poor--or at least poor enough to accept very low wages. Thus, some PDS subsidies Box 2.5:, How Well Doe% the PDS Serve (he Needs could be targeted to participants in employment of the P programs like the Employment Assurance Scheme Analysis suggests that the access of India's poor to (EAS) or Jawahar Rojgar Yojana (JRY). In fact, in commodities Napplied through the PDS is %en some states, EAS and JRY wages are already in part limited NanN of the poorest states (Bihar. 1.nar paid in food grains.choose to draw a much paid n fod grins.lo%%er share of .PDS subsidies than %%hat the% are entitled to (Table 2.31 and than their share -if 2.34 Although small in comparison to food po%em. For e%ample. Bihar accounts for 164 subsidies, spending on other nutrition and welfare percent of the poor and onl% 4.6 percent of PDS programs has also increased in recent years. For subsidies %%hile LP accounts for 18.4 percent of the poor and 8 percent ot total subsidies. In contrast. example, the Integrated Child Development Services Kerala accounis for onlN 2 4 percent of India*s poor (ICDS), a nutrition and general development scheme and 98 percent of PDS subsidies. %%hile Andhra targeted at children 0-6 years of age, is growing in Prade-h accounts for 4.8 percent of the poor and national importance. The program provides an integrated package of services that includes growth Targeting %hin states is like%%ise %eak. A careful monitoring, health education for mothers, early anal) sis IRadhahsistna & Subbarao. 19971. based on childhood education, health check-ups, immunization, rh e iN shte, th the arS sharePo referral, and supplemental feeding. The primary subsidies The per-capita income gains to the poor service provider is the Anganwadi Worker (AWW), a from all (food and non-food consumer subsidies in female "volunteer" who is government trained and rural areas %as an estimated RQ per month, or 2 7 percent of per-capita e%pendihures-. ir, urban areas compensated with a small honorarium. The AWW gains %tre onl slightl. higher at Rs 3.4 per month. coordinates her activities with those of the Auxiliary or 3 2 percent of per-capita expenditures. With the Nurse Midwife (ANM), attached to the local Health exception of Andhra Pradesh and Kerala. it is not Subcenter. As designed, the ICDS aims to attack the nutriin tha ee miima An ern malnutrition-disease cycle by serving as a link between meager benefits %%ere realized at a high cost. When the nutritionally deprived child and the formal health orth central go%ement costs are considered, an system. Through its pre-school education program, the amount of Rs.4.35 hatas incurred to transfer one rupee of income to the poor in 1986-87. Growth, Human Development, and 29 Anti-Poverty Programs Box 2.6: The Targeted Public Distribution System GOI has announced its intention to introduce weeping reforms in the Public Distribution System (PDS) in an effori to raise its cost-effecti%ene>s in reachinr the poor. According to recent% iksued guidelines. the Targeted PDS (TPDS) %would offer tw%o separate distribution channels one aimed at households helow the po%ert, line. and the other for the population abo%e the po%erR line. Under the first channel targeted to the poor households,. the central go%ernment would transfer to state go%ernments w%hear and rice at about half the issue price 4et for the PDS Tte monthly ration under the TPDS would be set at 10 kg per poor household in the state The number or poor household in a state "ould be the one determining by the recently appro%ed Expert Group's rnethodolog%. This 'ould eflecutelk determine for each state a maximum entitlement based on the number of poor households in the state--a tas,i improteient in relation to the present system in which the amount of foodgrain that a state can dra%% from the PDS is left at the discretion of the state and thus leads to situations where states wtith a high incidence of poerr utilized PDS much less than states with a lo% incidence of po%ertN The Central Go%ernment would leate to State eoernmcnts the re'ponsibili of designing and implementing targeting mechanisms for reaching the poor. and corresponding guidelines tere recenty issued b' the llinistr of Food. State go%emments %%ould need to be in a positon to idn.,t6 the poor. issue special cards. and debter foodgrain to the intended beneficiaries. The central government %wuld mmnior the suates performance in identif%ng and deli%ering foodgrains to the beneficiaries. for "hich reporing r.quirements haie been deteloped These features of the TPDS %till encourage the states to improve targeting or else. at least in theory. their access to the T PDS could be discontinued. Under the second and non-targered channel. %hich the guidelines indicate w%ould be phased our gradually. the central goernment wkould transfer t- state gierruments "heat and rice at an issue price which %%ould remain close to the market price Access to ihis non-targeted [PDS channel %%ould be uni%ersal It is proposed that--as an interim measure--the quantittes to be allocated to each st.ite be based on the j%erage lifting of "heat and rice toer the last ten years by the states ICDS also prepares children for formal schooling, and projects. The National Social Assistance Program, the antenatal care provided by the ANM contributes to introduced in 1995, provides for the first time a the prevention of malnutrition by decreasing the mechanism for central goverment to assist states in incidence of low birth weight. With its focus on early providing social protection to the poor (old age intervention, the program seeks to maximize both cost- pensions, survivor benefits, matemity relief). While effectiveness and nutritional well-being. evaluations of the national programs are yet to be made, evaluations of similar state-level schemes, show 2.35 Recent evaluations suggest that ICDS is mixed performance with considerable variability across relatively well targeted towards the poor and other districts and states (Prasad 1995, Rajan and Jayakumar disadvantaged groups, but the performance of the 1992). program has been mixed--program objectives are ambitious, resources are limited, and services are often 2.37 Central goverment spending on rural stretched thin on the ground. Efforts are currently employment-public works programs increased even underway both to improve the program and to widen more rapidly through the first half of the 1990s than coverage (see Box 2.7 for details). spending on food subsidies to the point that current budget shares are roughly equal. The primary objective 2.36 Other nutrition or welfare schemes are less well of the largest program, the Jawahar Rojgar Yojana targeted. The recently introduced mid-day meals (JRY) is to ensure gainful employment for poor program for school children is significantly larger in households. A secondary objective is to expand size than the ICDS, but may be less well-targeted. infrastructure and other social capital in backward Programs for tribals and scheduled castes provide a areas. At present, central government bears 80 percent wide variety of services, ranging from cash and in-kind of associated costs while states contribute 20 percent. transfers to employment generation and investment The program is administratively complex--three 30 Growth, Human Development, and Anti-Poverty Programs Box 2.7: How Well Have ICDS Interventions Worked? A 1992 official evaluation (National Institute of Public Cooperation and Child Development: "National Evaluation of Integrated Child De.elopment Ser ices") showed posimie results in child sur%ikal and nutritional status in ICDS blocks These blocks %%ere found to compare fa%orably "nh non-ICDS blocks on mator indicators such as the IMR. immunization coterage. utilization of health sertices. and scholastie enrollment and performance. Ne%eriheless. weaknesses %ere identified in the implementation of elie program in many% locations For example. oter one-third of AWWS iere found to be unable to perform the grotth monitoring funcion that is critical to identitying children in need of nutritional supplementation Links with the health system were likewise found to be weak in some cases. although ANNIs provided immunization support to 80 percent o'f Anganiadis. few receiked support in growth monitoring. identif"ing children at risk. or health education. .A backlog of 20 percent in the training of AWWs %as also re%ealed In some cases. an o%erburdened and undertiained \WAk has had to neglect a ke% service ksuch as outreach or health education) in order to fulfill the basic feeding. %eiwhing and childhoud education requirements of her position The sinergistic character of the ICDS program suffers whk-n any ot it components are abandoned. A number of strategies to address these problems are now on triaL Mans of them center around the participation of community members or NGOs in lvcistics and serices deliver. For example. the participation of "omen's groups or adolescent girls in gro,th monitoring has helped relie'e the AWW's burden of multiple tasks in some Anganadis. A scheme is in planning for lKerala w%herein the AWW would assume responsibilir% for a larger population in return for double her usual compensation. and the community would pro%ide the additional funds. Considerable scope still e,ists for increased local participation. panicularl.\ in the areas of planning and monitoring. On the other hand, the ICOS appears generally well-targeted towards the disadvantaged groups of the population. Ihe program reles hea l\ on geographical targeting in that protect blocks are prefereniallk selected ftom areas characterized by emreme po%erts or a high concentration of ccheduled tribe or caste households In addion. AWWs are rrained to perform outreach to lo%%-income huseholdc. and se%erelh malnourished children ir-picallk from poor householdsi are gi,en double the usual food ration Self-targeting is operable- social status considerations together with the coarse-grain qualities of the food pro%ided discourage wealthier families from extensike use of the centers A 1490-92 e%aluation study ire' ealed that the profile of study sample households conformed to the guidelines for benefician selection and that 50 percent of household heads were marginal or landless laborers. .1n on-going World Bank research iCharterlee 1997) of ICDS effectiveness in two states Maharashtra and Raiasthan re%ealed that while those "ho are co% ered by ICDS ser ices are predominantly poor. not all the poor are actually co%ered by the program Data from the field study sho% that 15-25 percent of the population in blocks co%ered bs ICDS lite in %illages that do not ha-e .A%s. and thus do not ha%e access to senices pro%ided through the program Thus. in rural Mahara,hua in tribal ICDS blocks at least half of the "uncovered' population Ike in poorest %illages Perhaps the more serious problem is the under-co%erqce that takes place w'uhin the angan%adi area This %illage level underco%erage appears largel) to be skewed agaici the poor Spectricall\. distant hamlets. Scheduled Caste clusters, and dispersed marginal households tend to be lct out. since \WCs are located in 'central" imost likely bener-off) sections of the village This suggests that ICDS should now focus its effrs on the poor. rather than further apanding coverage to new blocAs in the countrr. quarters of total JRY funds are earmarked for various 2.38 When carefully designed, public works schemes rural infrastructure schemes (i.e. the Million Wells like the JRY have tremendous potential for protecting Scheme, housing construction under the Indira Awas the poor from severe destitution as well as seasonal Yojana--both targeted to SC/ST households--and social income shocks. However, recent government forestry projects. Moreover, the program is unique in evaluations as well as independent studies suggest its level of decentralization--work plans and contracts that the JR Y has been beset by a number ofproblems. are administered by the village panchayats, subject to Employed workforce is predominantly male. In some basic program guidelines. An estimated one billion states, there have been reports that projects have been days of employment are generated each year under the contracted out by elected local officials to private JRY, which indicates that some 30-40 percent of contractors, who are free to hire whomever they choose potential beneficiaries may participate in the JRY. and failing to comply with the program's guidelines Growth, Human Development, and 31 Anti-Poverty Programs that unskilled laborers are supposed to be employed in productive assets for agriculture, animal husbandry, preference. Effective monitoring of the work being village industries, or service activities so that they done is often poor, and micro-studies have documented experience a sustained increase in income and welfare substantial leakages through false muster-rolls or levels. The assistance takes the form of a capital material purchase. The official minimum wage is often subsidy (shared equally by state and central quite different from the prevailing agricultural wages government) and directed credit from rural financial which makes the work attractive to semi-skilled rather institutions. IRDP has assisted over 40 million than strictly unskilled workers The participation of households since its inception in 1978-79 and currently non-poor workers (about 18 percent in 1993-94 benefits 2.5 million households annually. according to an estimated income criterion) in the JRY could be accounted for by the fact that the official 2.41 While targeting under IRDP relies on list of wage is well above the local market wage in several poor households (means-tested) that were developed areas. at the village level, not all poor households are included in these lists, and not all households on the 2.39 In addition to JRY, the Employment Assurance lists are poor. Estimates of "ineligible" beneficiaries Scheme (EAS) was launched in 1993-94 in drought range from 5 to 40 percent in different studies. prone, backward areas as well as tribal and hill areas. Furthermore, the regional distribution of IRDP The EAS is designed to guarantee 100 days of resources does not take into account incidence or employment each year in the slack season to a severity of poverty. Initially the program aimed at maximum of two adults per household. Like the JRY, providing assistance to the poorest and most destitute the EAS is based on self-targeting but as workers are households; however, experience has shown that these paid at the state minimum wage, the impact of the beneficiaries are less likely to retain their assets or targeting may be diluted. However, unlike the JRY, the realize income increases that last. As a result, there is a scheme is not decentralized to the panchayats and is preference for lending to household which are just implemented by sectoral agencies and departments. below the poverty line. Loan amounts for investment Moreover, the scheme has been slow to take off and are also now generally recognized as inadequate. Like only 50 million days of employment were generated many rural credit programs, the repayment record for last year. A major concurrent evaluation of the EAS is IRDP loans is poor. In comparison, credit-cum-subsidy now under way in the Planning Commission. A third assistance to women's groups (the Development of public works/employment program targeted towards Women and Children in Rural Areas Program urban areas--the NRY--is financed equally by central (DWCRA)) have been relatively more successful. government and the states. The NRY has three major Recent initiatives for IRDP reform include mechanisms components: wage employment for the urban poor for supporting self-help Groups (mostly organized and through construction of public facilities, housing and run in rural areas by women) and better identification shelter upgrading (including training of urban of rural micro-entrepreneurs. Some of these reforms construction labor), and investment subsidies for are broadly in line with interational experience. micro-enterprises. 2.42 Most evaluations of official anti-poverty 2.40 Yet another field of government-financed anti- programs-whether in-house by the Planning poverty activity is the relatively small amount Commission or done independently- agree that they budgeted for self-employment programs, of which the are not performing as well as hoped, particularly in largest and best known is the Integrated Rural terms of their impact on thepoor. A number of causes Development Program (IRDP), providing credits at can be identified, some amenable to change and others concessionary terms to households below the poverty not. In terms of the latter, two issues deserve mention: line. The intention is to help poor, rural households, first, the problems that the programs are designed to with a one-time infusion of capital to purchase address (i.e. widespread poverty, malnutrition, mass 32 Growth, Human Development, and Anti-Poverty Programs illiteracy) are too massive to be resolved through poor APP beneficiaries range from a low of 15-20 public resources allocated to these programs, however percent (EGS, ICDS) to an estimated 50-60 percent efficient the use of these resources is; and second, (PDS). many anti-poverty programs have become highly politicized; India's complex socio-political 0 Program Design, Inputs, and Implementation. environment makes it very difficult to change existing Many anti-poverty programs are well-designed on programs. paper, but serious problems arise in implementation. 2.43A nmbe ofsysemicprolem ca be The lack of adequate resources is a serious concern-- 2.43 A number of systemic problems can be mn rgasaeudrudd n okr r identified across programs that are more amenable to uneraid, underd nd pol otivae change and improvement. udrad netand n ory mtvtd chane an impovemnt.Program administration is often highly centralized and Objectives. For many of India's anti-povertylines--within respective programs, there is a need to simplify and prioritize rn ment instittisar inolved i e program objectives. Many of the largest and longest- ani-pover msnt stte ad cnrlvel nd running programs have multiple objectives; their cntion between agencie indeqate Mod relative importance is ill-defined and has changed over time. And many of the programs do not serve as an effective safety net for the poor, although clearly this participatory fashion. In recent years, however, some should be a key objective of these types of public significant initiatives have been taken to provide intevenions Whn pogras ty t do oo any greater flexibility in the design and management of interventions. When programs try to do too many prgastsaean lorlvl.Th Cnrl things (particularly with limited resources), quality government prvie aditona l entral suffers, outcomes are difficult to measure, and staermet even aicned ient ifie t accountability is problematic. In some cases, different m tat io n ith t aes Thete aten program objectives can even be conflicting. A typical free tonixtprioit and d es progra s ti the example is the PDS which tries to provide minimum set f bro btives A retr wotal se support prices to producers and cheap foodgrain and goron structes bee ted in te sugar to consumers. Another, though less glaring Constitution through the 73rd/74th Amendments. example, is that of the JRY which has the primary These amendments have put a number of important objective of providing gainful employment, which lead to the setting of standardized labor to non-labor cost norms, and a secondary objective of rural asset move towards decentralization at the national level. creation, which may demand higher capital costs than warranted by these norms. m warrnte bythes noms.slow in most' states and till, date, very few have undertaken 'the mandatory, financial devolution. * Beneficiaries. Most anti-poverty programs have Moreover, the scope for decentralization is still limited clearly identified beneficiaries--either the entire by a number of factors which include scarcity of population (PDS) or specific subgroups within it (JRY, resources, poor management and administrative ICDS). Bureaucratic selection and partial self-targeting capacity, corruption, and inadequate participation in are used in a number of schemes to try to ensure. that decision-makin by the economically and socially poor beneficiaries are included. Little effort is pevaluations show th therefore made to target the poor directly, and no panchayats has been more encouraging in states or consistent definition is used across programs to areas where these factors have been overcome to a identify intended beneficiaries. Targeting efficiency is arge extent, but is still poor in areas where they, or low for the majority of programs-Mthe number of non- Growth, Human Development, and 33 Anti-Poverty Programs other similar constraints, continue to predominate (Box programs--analysis suggests that it costs between 2 and 2.8). 7 rupees (with the highest value reported for the PDS) for anti-poverty schemes to provide one rupee of benefits to the poor (Radhakrishna & Subbarao, 1997). Box 2.8: Decentralization and Outreach to the Additionally, guidelines tend to be over-refined, Poor: Evidence from Social Assessment Studies complex, and often beyond the understanding of In preparation for a forthcoming World Bank District typically illiterate beneficiaries. The design of many Poverty Initiative Project (DPIP), social: assessments programs exemplifies factors typically associated with were carried out to gather information on the systems prone to corruption. For example, discretion is mechanisms through which poverty alleviation programs are reaching the poor at the village level. Results from such assessment in Madhya Pradesh granting or refusing a benefit are often not known to showed that: the applicant and change frequently. There is no Locally elected officials are controlled by their local transparency--no reasons need be given for refusal, and constituencies (consisting mainly of the upper classes there is generally no appeal except to the same person of village society) and respond to their vested who took the initial decision interests. At the same time the poorest of the poor residing in marginal communities, remote and inaccessible areas, have no say in the decisions. To a great extent, the problem of excessive T% e\mpes f pogamsmi,maageen %%re bureaucratic control has arisen from the TCo eaamples of programs mismanagement were quite re%ealing. Earlier, the local ofticials were mainIa responsible for the selection of beneficiaries of IRDP. against corruption. Instead of any real reduction in Not all the poor w%ere properik identified. but an upper corruption, however, the effort has produced limit on the number of households under the po%ert sluggishness, lack of adaptability and needless line pre%ented those lists from being inflated. In contrast to this. after the responsibilitN was transferred confusion and complexity. Decisions of program to %-illage-le%el pancha\ais. the lhts ha%e become managers are overdetermined by a number of rules; grossI inflated. In some places IRDP has tended to however, evaluations reveal considerable variation include anone and eer\one who were prospectie between earmarking on paper for different purposes and fa%orable %oters in the pancha\aL election. In the stud> areas in one district of NladhN1a Pradesh (Sagar. and different target groups, on the one hand, and what the IRDP list ended up including all rural households! happens in the field on the other. For example, the In the case of JRY. the funding w%as allocated between 1992 Planning Commission's concurrent evaluation of %illage le%el panchaNats so thinl%. that the program could not- achiee a sustainable impro%ement in JRY reports an average wage component of 53 percent infrastructure an\%%here The annual budget per at an all India level, as against the stipulated minimum %iflage panchaNat. ranging between Rs. 20.000 and of 60 percent. State-level variations ranged from 21 Rs 30,0111.. was ne%er sufficient to :onpleie one job ukthin one calendar \ear. B the time the next percent (Punjab) to 83 percent (Assam.) installment became due, the earlier construction usually started falling apart. * Program Impaciv. Impacts must be measured Source: Forthcoming Social Assessment Studies, against objectives and intended beneficiaries. For most World Bank, 1997. anti-poverty programs, a series of indicators identified for measuring success or failure are often (but not always) monitored on a regular basis and used in GOT * Another common weakness is high administrative evaluations to judge program performance. cost. Despite multiple layers of control, leakages are Unfortunately, the success of programs is judged too significant--ranging from 5 to 40 percent for JRY and often solely on the basis of fulfilling predetermined 20 percent to 50 percent for IRDP. Ineffective targeting physical targets. The sole use of a narrow set of mechanisms and high administrative costs raise indicators to evaluate performance may have questions about the cost effectiveness of these unforeseen negative impacts not only on the accuracy 34 Growth, Human Development, and Anti-Poverty Programs of the evaluation, but also on the workings of the more likely to empower men than women and far more programs themselves. In India there is a pervasive and responsive to the overall growth of the economy than troublesome tendency to attach overwhelming to various government efforts to direct resources to importance to performance indicators: instead of reduce poverty. Some of those programs, in fact, have treating indicators as simple measures of progress in largely missed their supposed targets--the poor--and specific areas, project staff tend to view indicators as delivered the bulk of their benefits or subsidies to the the outcomes of the program themselves, often politically or economically more advantaged. distorting the program's true objectives in the process. For example, India's current poverty reduction strategy 2.45 As India moves ahead with the economic tends to emphasize the number of people crossing the liberalization that has brought with it new surges of poverty line and thus shifting from poor to non-poor growth and a potential for achieving a higher level of status (e.g. a reduction of headcount index or incidence welfare, it has a new opportunity to reexamine rural of poverty) as a key measure of successful poverty--more than three-fourths of the total--and interventions. Given the very large number of people extend the reforms to unleash productivity growth in close to poverty line on both sides and the instability in agriculture which historically has contributed most to their incomes, movements across the line may be poverty reduction. The following brief neither meaningful nor sustainable. The exclusive use recommendations could be the basis on which to build of a partial success indicator--for example, the this new strategy. headcount index--can distort the evaluation of program performance. A classic example is provided by the 2.46 The Primacy of Growth and Human Capital results of various evaluations of the Maharashtra Development The strongest impetus to poverty Employment Guarantee Scheme (EGS), which is alleviation in India has come from economic growth. It generally considered as a program well-targeted to the is not that investments in health and education do not very poor, but criticized because it has very little bring real gains in poverty reduction, but that impact on the headcount index. The minimal impact on economic growth widens economic opportunities, the headcount index is due to the fact that its provides the resources needed to invest in human beneficiaries are in fact the very poor, and EGS development, and creates the very foundation that will earnings, while typically not sufficient to move them increase its returns--and thus families' willingness to above the official poverty line, do make significant send their children to school, have fewer of them, or improvements in the standard of living for some of the treat their girls as respectfully as they would their boys. state's most destitute households, one of the Economic growth in India has widened opportunities at fundamental objectives of the program the bottom as well as near the top of society, has seen wages rise in the countryside--where landless wage F. Conclusions and Recommendations earners are among the economically most disadvantaged--and, since liberalization began, has 2.44 In the half century since its independence, India improved returns to farmers. Along with those has accomplished many notable social and economic increased earnings went declines in poverty. achievements in a democratic political setting. Among the accomplishments are the eradication of famine, the 2.47 The last five years have shown the rates of limitation of population growth, the erosion of caste growth that India could achieve with market oriented barriers and the creation of a large pool of technical development policies and a better integration with the and scientific talent. While it has also managed to world economy. There is much that remains to be done reduce poverty in that period, only since about 1975 to sustain this growth. Reforms are particularly needed has the decline achieved a fairly steady momentum. to reduce the country's persistently high fiscal deficits, The pace, moreover, remains both slow and uneven-- overcome its tremendous infrastructure problems, faster in the southern states than the northern ones-- improve the efficiency of its financial system, and Growth, Human Development, and 35 Anti-Poverty Programs liberalize parts of the economy that remain heavily investments in the health and education of their people regulated--such as agriculture, small scale industry, have seen economic growth and poverty reduction urban land markets. These are worth the efforts (Box follow. India, however, has not given the education of 2.9). the poor the kind of priority it deserves, and 33 million of its 105 million 6-to-10-year-olds are not in school. Most of them, instead, are working for extremely low Box 2.9: The Centrality of Growth with Equity to wages which will not allow them to develop the skills Poverty Reduction needed for upward mobility. Distributional data from the NSSO was used to simulate (as indications of possible trends) the 2.49 Along with the neglect of primary education as evolution of poverty measures if growth were to be an economic and social dynamo goes a gender sustained at over 6 percent under both distributionally discrimination that condemns women, and non-distributionally neutral growth. For 1993-94, India's o%erall poxerLN rate %%a; 35 percent. Assuming, disproportionately, both to illiteracy and ill-health. that iii consuimption per capita grows at a compound Bridging the gender gap among the poor requires a growth rate of 4 2 percent bemeen 1993-94 and 2005- deterined effort to focus on health care for women . ) all le%e e of conunpilon grow at the same rate (that is the Lorenz curve does not change--roughly speaking "inequality" does not change); and (iii) the in particular, with basic education in literacy. The real value of the poverty line does not change. Then Indian health system needs to put a new emphasis on India's poverty rate (headcount) will fall to 6.3 percent basic care. Doing so will bring improvement not just in by 2005-06. by 205-06.the life span and well-being of poor women and in What if inequality rises? While this has not happened their ability to determine what size family they want in India (Gini index over the last 30 years was but also in their contribution to the economic health of relatively constant), it is of interest to simulate this scenario since some people are concerned about the their families and the nation. A recent study (Fihmer, possibility of rising inequality. India's Gini index for King, and Pritchett, 1997) reinforces the view that 1993-94 is 30 percent. For the sake of argument, determined public policy action to eradicate gender suppose it rose to 46 percent, the value in Ecuador in 1994 (based on consumption data). Assuming that the whole of India's Lorenz curve shifts to that of Ecuador collection of data assembled from the Demographic by 2005-06--which would result in a dramatic Health Surveys from a large number of countries and deterioration in distribution--, the same growth rates the National Family Health Surveys from India's imply that the poverty rate in India xtould fall to 27.5 percent. states, concluded that unlike many indicators of standard of living and even many social indicators, What if the real value of the poverty line rises, to gender disparity in school enrollment and child reflect changing perceptions of w%hai "po%ertN" in India means? For a low-income counirs and con.;isren tith mortality are not correlated with the level of income. evidence on cross-country compari;-ons of poertN While the level of female child mortality falls with lines with average incomes, an ela-ilicin of the po%ert) rising incomes around the world, including in India, line with respect to average con amption of 0 2 ik assumed. Under this scenario, the poverty rate will fall to 10.3 percent given a growth rate of 4.2 percent per be related to income. Poverty is therefore not the only capita private consumption over 1993-94 to 2005-06, source of gender disparity. and assuming that inequality does not change. If inequality worsens and the poverty line increases, 2.50 Large opportunities for greater private sector making the same assumptions as above, the poverty involvement remain unexplored. The private sector rate then falls to 31.9 percent. plays a vital role in the provision of selected aspects of health services. There is a wide scope for 2.48 Priority for Human Capital. Neighboring strengthening the public-private partnership in the countries in Asia which have made a point of delivery of health care services. This partnership could combining pro-growth development policies with include: enhancing the scope and importance of the 36 Growth, Human Development, and Anti-Poverty Programs private health sector, while improving the quality of majority share a strong grass-roots orientation and services; encouraging private sector involvement in focus on community development (Box 2.10). While preventive and promotive aspects of health care rather NGOs' programs are valuable in their own right, they than solely in individual curative care; finding the are limited in scope. However, recent years have appropriate mix between direct provision versus public witnessed growing NGO involvement in the financing of some activities performed by the private implementation of various government schemes. This sector; promoting partnership between the public, trend is encouraging; it creates positive synergies and private, and voluntary sectors; and improving the makes use of the strengths of both the NGO and existing arrangement for regulating and monitoring government capacities in reaching the poor. private health care. In addition, government efforts need to focus on districts where private sector 2.52 Reforming Anti-Poverty Programs. Anti- provision is weak and incidence of poverty is poverty programs were designed to have an important significant. but supplementary role to the twin pillars of the government's poverty reduction strategy--accelerated 2.51 Along with improving the capacity of state and and sustained labor-intensive growth, and investment Panchayati Raj Institutions, significant synergies can in human capital development. However, and in spite be reaped from greater involvement of India's Non- of the significant resources that the country allocates to Government Organizations (NGOs). While there is such programs, their performance in yielding tremendous heterogeneity amongst NGOs, the vast significant gains in the living standards of the poor has Box 2.10: NGOs and Poverty Reduction NGOs are increasingly, in%oled in the implementation ofgoierment schemes directed at reducing poveN and improving conditions as well as programs directl\ funded by multilateral and bilateral aid agencies. The implementation of mand go%emment schemes, such as the JRPP. Watershed De%elopment. Total Lteraco Campaign. to quote a NeG examples, has been facilitated by NGOs There are se\eral thousand N6gs in India seeking to improe the poor's access to resources-- physical. huiman skills, and financial The Self-Emplo\ed WVomen's Association iSE WA). an established NGO %%Ith a hision of successf2l promotion of micro-enterprise schemes for poor ioren. is no% e%perimenting %%ith etending social insurance to its members. Others seek to impro%e their members' access to credit, education or to knowledge nitah international NGOs such as ONFANI joining the national NG bs. and considerably strengthening them both financiallr and through ner-torking suppoi NGOs can contribute significantly to improving the qualir%. sustainabilir and effectiveness Of SoCl programs by pit developing and experimenting %ith innovati%e approaches. uit promoting popular paricipation and communit ownership: (iiis encouraging program uptake: and iaim extending benefits to those segments ofthe population that are most difficult to reach. The contributions o1 SEWA iii micro-enterprise development: the Chipko rmovement in UJP in the protection of forests, the Shahdol group in layiing the fogndation tor enironmenal assessment, and the dazdoor Kisan Shakti Sangathan in its campaign for a right to information have all clearly had a multiplier effect, and help to set a broader development agenda Ho%%e%er. NGC)N in India continue to %%ork in %vet' limited areas- a fcvv illages. a block or a district. Moreover, the distribution of effecte NGOs i g ' uneen across the countire While man NG Os enjoy strong communit ties, and are credible and competent. others ma be inexperienced. inefficient or sironger on rhetoric than in action. NGO often lack organizaTional and financial skills, and need significnto inputs ith regard to capacit building A number of interational NGOs are no working closely with national NGOOs and heop to strengthen them both financially and through net,orking suppon In adition. wih the Increasing devolution of responsibility for program implementation down to the Panchayai Raj Intituftions. NGOs need to coordinate more closeil their activities f ith the elected bodies. Whtle the benefits of'closer collaboration bemween go%ernmemn and NGOs are clear. the relationship ber%ween NGOs and all levels of govement has had its ups and dovns. ith each side tr%ing to find creative "asin hich they could work bener together. In September 1994 an institutional arrangement for a "collaborative relationship" between all levels of gorernment and NGOs "as set up. particularly for NG s invoed in poer eradication programs and Gover ment secretaries directl invaol%ed ith anti-pov ert programs. to facilitae the NGO in t mobilizing and organizing the poor to become active paicipants in the process of development, and it) deliherineri eices by implementing arious development programs. Growth, Human Development, and 37 Anti-Poverty Programs been disappointing. While many recipients of benefits from safety net programs are poor, many of the poorest Bo% 2.11: Transient Po%ertv in Rural India people do not use the programs and when they do, the cost of delivering assistance to the needy is much killage-le%el data rho" %de Near-to-)ear %ariations higher than the value of the aid itself. The programs in incomes. leding to pronounced fluctuations in had multiple and sometimes conflicting objectives consumption mn that he olis lutatio i between serving the poor, stabilizing commodity prices po%ert) in bad sears bui also a substantial flo% out of or building rural roads, irrigation and power pOvert in a good Near One longitudinal sur%e% of generators. Because of their unsustainable fiscal cost hol n ofrua populato is poo at and their lack of outreach to the poor, the government one Nea out of the eight consecutke Nears under already has started to look at options of seriously consideration. Thii sugge;ts that the poor are largely uphauling the functioning and targeting of these uninsured again;t agricultural ield shocks. Thus intertemporal consumption smoothing possibilities are programs as is currently the intent under the redesigned limited, or too cosilN for the poor Efrecti%e saferN PDS. Similarly, the ICDS might make a greater impact nets that insure rural poor against income fluctuations- by strengthening its activities in areas where it is -such Public %%orki pro,rami-are iherefore already established, rather than try to spread to the 1600 additional blocks for which no funds have been appropriated. 2.55 The challenge that India faces in reducing poverty in the future faster and more effectively than it 2.53 The ICDS could usefully focus even more has in the past is an enormous one. India has met and closely on the most needy and be more flexible in mastered other vast challenges and is laying the delivering its menu of services. The ICDS program foundation for higher growth with the continued also offers an excellent ground for expanding liberalization of the economy. Building on this community participation in the management of local foundation for faster growth would require that the affairs. Many aspects of India's anti-poverty efforts challenge of poverty reduction remains at the forefront and of its social development in general are poised for of the official concern and becomes the subject of the the infusion of new energy and commitment that come same imaginative, thorough, change-minded actions as from entrusting people with greater decision-making the country's experiments in stabilization and structural responsibility. For the poorest of the poor, for women adjustment. In particular, the liberalization of the in particular, the sense of empowerment to come from economy needs to focus on the deregulation of sectors much greater involvement in shaping programs and on which depends the livelihood of many of the poor activities would also be a stimulus to individual and (agriculture--particularly agro-industry, small scale collective progress against poverty, industries), and on improving the provision of infrastructure and social services to the 2.54 Safety nets have a key role in achieving poverty underprivileged. Resources can be generated by reduction, but India urgently needs to formulate an reducing the large share of resources absorbed by anti-poverty strategy that is fiscally sustainable. To subsidies, and by scrutinizing existing poverty increase their cost effectiveness and extend their programs, and eliminating those that are ineffective. outreach to the really poor, safety nets need to be targeted to those who cannot participate in the growth 2.56 Poverty measurement, poverty analysis and process (such as for reasons of physical handicap data accessibility. The analysis of poverty can, and combined with poverty) or face continuing exposure to should, occur along various lines at the same time. uninsured risk, which may well also jeopardize their Different approaches need not compete with one ability to participate in the opportunities offered by another for intellectual supremacy. Rather, they should growth (Box 2.11). Further research in this area is be recognized to shed light on different aspects of what crucially needed to design such a strategy. is clearly a complex problem. Together, these various 38 Growth, Human Development, and Anti-Poverty Programs approaches may combine to offer a fuller perspective, relationship of gender, caste and illiteracy to the and may thereby help in formulating well-conceived persistence of deprivation. For policymakers seeking to policy interventions. analyze the incidence by socio-economic groups of subsidies in public spending to determine the best use 2.57 Central to that effort are high-quality and of resources, such knowledge is essential. Yet access to timely data. In many respects, India stands out from existing data resources remains highly restricted. For other countries in terms of its tradition of data example, few researchers in India have received collection and its pioneering of many of the techniques permission to analyze NSSO data at the unit-record of data analysis which have now become common level. There are an enormous number of urgent currency throughout the world. Studying the questions which could be investigated if such data were information that has been released, analysts have been accessible, and there is certainly no shortage of well- able to detail the regional disparities in the incidence qualified researchers in India who could undertake and reduction of poverty, the role of growth as opposed such studies. to redistribution in alleviating this poverty, the Annex 1 39 Poverty Measurement: The Debate on How Best to Capture the Extent of Deprivation 1. The first widely-endorsed assessment of the themselves. In addition, it is suggested that basing the extent of poverty in India as a whole was produced in poverty line on a simplified nutritional norm 1962 by the Perspective Planning Division of the (expressed, usually, in calorie terms only) combined Planning Commission.' This exercise produced a with the availability of only relatively crude measures national poverty line, expressed in terms of the total of household consumption or income, fails to capture expenditure required by an average household to differences in access to services such as health and ensure that all family members were able to meet a education, and ignores inequalities in power and specified minimum nutritional norm. Households with influence. expenditures below this threshold were judged to be poor. This approach to the analysis of poverty has long 3. The conventional practice is often contrasted historical antecedents in countries such as the United with an alternative, broader, approach. This perspective States and the United Kingdom, and has also remained views incomes or expenditures as a means rather than the basis of the majority of poverty studies for India. an end, and well-being is seen in terms of both what However, the emphasis on calculation of poverty rates individuals can do, as well as what they actually do.' and on the representation of living standards by In this view, it is not sufficient to know whether an incomes or expenditure levels is sometimes questioned individual can afford to consume a given consumption in India.2 Some of the objections raised are worth bundle, but also to understand what factors might addressing up front. prevent the person from doing so. For example, if the individual is not educated enough to recognize the 2. Critics argue that conventional practice induces value of the bundle, or is constrained by local power a mechanistic approach to the alleviation of poverty. relations from choosing it, or is in such poor health as Insisting that the measurement of poverty should be to be unable to derive the supposed benefits from its based on an objective, externally-imposed poverty line, consumption, then that person can be considered they claim, leads to the simplistic perception that the fpoor" even though he or she might be able to afford poor are a passive group who can be objectively the bundle. Moreover, the approach emphasizes that counted and measured in various ways, and can then be consumption of a particular bundle of commodities targeted to receive various public transfers and does not define an individuals well-being. The subsidies. These criticisms are often best seen as a individual's functionings needs to be scrutinized, and commentary on the way in which poverty analysis is incomes or consumption levels should be viewed in sometimes undertaken, rather than as an objection to terms of how they contribute to these fiictionings. the approach per se. While there is no question that This broader approach thus suggests that a number of poverty studies vary widely in terms of their indicators, in addition to income or expenditure, should methodological rigour and the care with which be examined when assessing well-being. Examples of conclusions are drawn, best practice studies would such indicators might include education levels, presumably escape such critiques. Other criticisms are indicators of health, and proxies of gender more fundamental. It is argued for example, that discrimination. conventional practice ignores the views of the poor pAlthough the analysis of poverty was already well-established This approach, denoted the "capabilities" approach, was even before independence. For example, in the late 1930s the introduced by Amartya Sen, and is described funther, with Bombay Labour Enquiry Committee developed a poverty line applications to India, in Sen (1981) and Dr6ze and Sen (1995, for the purpose of measuring poverty in Sholapur city 1997). (Thakur, 1985). 4 Drze and Sen (1989) provide examples of how this broader Beck (1994) provides an interesting critique, approach might be applied in practice. 40 Annex I 4. Basing poverty measurement on consumption expenditure need not imply rejection of this broader Box A 1.1: Targeting the Poor Lsing a Synthetic perspective. A solid grasp of the many dimensions of Household Welfare Index in Ecuador deprivation is clearly critical to appreciate how poverty In 1994 the National Statistical Institute of Ecuador can be tackled. Yet it is clear that conventionally deloped a houschold-lewl indicator of v6elI-being measured poverty will represent a large part, albeit not from census data. This indicator consisted of a the total, of an overall assessment. In addition, it is not %%eighied compo'ite offi%e %armbles capturing access to %%ater. ;anitaiion and %%aste disposal sem ces. obvious how the other dimensions are best captured, education imeisured throueh the educaton le%el of nor how this additional information should be the head of the householdi and a cro%%ding index combined to reach an overall assessment of poverty. In (measured b% the number of people in the household fact, in India experience shows that the different d%eabN the n ed oring toEachjofbthese indicators have often told a similar story. This and qualiiN. The points assigned to each set' ice %%ere implies, and some evidence is presented in Chapter 1, essentialIN arbttrarN. The household %%elfire index %,a: whenpovety cmparson basd onthe orethen simplh t"~en as the total sum oft points across the that when poverty comparisons based on the more that ftne %ariables The loiwer the %alue of points per restrictive approach are properly carried out, they are household. the poorer the household %as percei%ed to often consistent with those which emerge from scrutiny be of dimensions other than income or expenditure. It is, A recent household survey for Ecuador allows one to of course, essential to check for such consistency and e\arine ho% %%ell the household %%elfare indicator to examine carefully those circumstances where performs in ident6tg the poor. Rankings or disagreements do arise. households ba:ed on't1&s crterion can be compared Iagainst rankings based on per capita consumption. As the household suvN %%as multi-modular in design. it 5. Data limitations may prompt a simplified included the same %ariables as "ere available to the approach. In some countries, particularly in Latin Statistical Inqiute from the census, thereby allo%%ie a America, it is common to construct a household similar household % elfare index to be constructed from the sur%e%. The surve\ also permitted the "welfare-index" in cases where income or consumption data happen to be unavailable. Such an household consumption indicator usually combines a range of indices based on, for example, quality of housing, access to public Suppose one %%anted to use the household welfare inidex to target the poorest 20 percent of households in services, household demographics, and so on, into an the population. It %as found that only 40 percent of overall household-level index. The selection of households ientified using this crterion %%ere also variables is generally guided by data availability, included among the bonom 20 percent of households under a consumptiokn criterion The leakage of usually drawing on sources such as the census. Unlike resources from such a h%pothetical transfer %ould be the capabilities approach, where additional indicators %er% high-613ercent of resources would go to non- are examined because they offer insight into particular intended beneficiaries capabilities which incomes do not reflect, the More ehborate houiehold ielfare indices might indicators here are selected in an ad-hoc way, on the perform better than this simple example. But as long basis of their perceived correlation with well-being, as the %eighting scheme across the sub-coiponents of however defined. Although systematic evaluations are condetrofathis.rOnerany itnisrdimiihtlb to scarce, there is some evidence suggesting that results construct empirical models to estimate the %%eight: for may not be very reliable (Box A 1.1). such an index 6. This report thus takes the position that S H h . o.. conventional poverty measurement techniques, I_I properly applied, offer a tractable and fairly reliable way in which to make comparisons of poverty in aong Ih poult subgroups port rs India. Such comparisons can provide a useful c ee d on the f ath is ep rls perspective on how poverty has evolved over time ando eth r an e Annex 1 41 evidence regarding other dimensions of well-being. In 9. Since 1979-80, official estimates of the number addition, it is possible to compare the findings derived of poor in India have been based on a definition of the from secondary data to those from detailed micro- poverty line recommended by a Task Force convened studies. While micro-studies do not permit ready by the Planning Commission. Following criticisms that generalization, the level of detail which they provide the official poverty line was underestimating the extent can help in the interpretation of results from large-scale of poverty incidence, the Planning Commission survey data. Such studies also offer a means to check convened a second Expert Group to examine this issue. whether findings from statistical data accord with In 1993, the "Expert Group on Estimation of Number experience at the ground level (Annex II). and Proportion of Poor" published its report in which methodological improvements were suggested with 7. In order to make meaningful poverty respect to two broad sets of issues (Box A1.2). First, comparisons, even within the relatively narrow the Expert Group indicated how the poverty line framework of conventional poverty measurement, it is calculation should take explicit account of differences critical to ensure that the methods employed are sound. in the cost of living across different parts of the This involves choices: an appropriate welfare indicator country. This implies that although derived from the must be selected; a decision must be reached regarding same base, the poverty line in each state will vary the application of equivalence scales; the poverty line according to the different cost of purchasing a given must be specified; adjustments must be made for cost bundle in each state. In addition, the Expert Group felt of living variation; and choices have to be made that comparisons of poverty over time require that the regarding the preferred poverty measure. poverty line be updated using an inflation adjustment based on changes in prices of those goods which are 8. Income and consumption data from household relevant to the poor. surveys are usually collected at the level of the household rather than the individual. This means that 10. A second set of issues which the Expert Group in attributing to individuals within the household their addressed concerned the prevailing practice of share of household resources, an adjustment based on adjusting NSSO consumption figures so that these some allocation rule must be imposed. The most aggregated up to consumption figures from the commonly applied "rule", and one which has also National Accounts. The Expert Group strongly conventionally been applied in India, is that all family recommended that this practice be discontinued members receive an equal share of household because of the arbitrary basis of the adjustments and resources. This allows household consumption or the lack of convincing evidence for such practice. The income to be divided through by household size to Group did note that discontinuation of this practice obtain a measure of per-capita consumption (or would result in a fairly sizable one-off change in income). Although this allocation rule is simple, it is official poverty rates. not necessarily appealing, as it seems unlikely that individuals need the same amount of resources to reach 11. The recommendations by the Expert Group the same welfare level. And, even if they do require the significantly strengthen the methodological basis of same amount, it is unlikely that they in fact receive poverty measurement in India, and can be regarded as equal shares. Alternative allocation rules, known as a major step towards "best practice" (even if there does equivalence scales, are often proposed. Although there remain room for further improvement--see Box A1.2). exists little guidance for choosing among the wide Following publication, the Expert Group's range of possible scales, it is important to examine the recommendations were widely adopted in the academic sensitivity of poverty comparisons to the particular and research community, but it wasn't until December allocation rule chosen. This issue is particularly 1996 that the Planning Commission declared its important when one is concerned with demographic intention to also accept these recommendations. characteristics of the poor (see chapter 1). 42 Annex I 12. A final issue to address prior to calculating distributed along the income or consumption spectrum poverty rates, concerns the selection of a specific below the poverty line. In general, when a particular poverty measure (or set of poverty measures). By far poverty comparison (such as comparing poverty the most widely used measure is the headcount. This is between two time periods) is very sensitive to even simply the proportion of the population with a per- small changes in the location of the poverty line or capita consumption or income level below the poverty changes in choice of poverty measure, then one would line. Although the headcount is attractive because it is not like to attach too great a significance to the so easily interpreted, it also has its limitations. Notably, outcome of that particular comparison. it is entirely insensitive to distance below the poverty line (see Box 1.2). Other poverty measures are 13. This report has broadly followed the Expert available which do capture how the poor population is Group method in the measurement of poverty with one Box Ab.2: Expert Group on Estimation of Proportion and Number of Poor The Expen Group (1993) focused on m o broad areas in vhich the prevailing official" ieihodolog could be improved Firm, the Group made a strong cas e for careful adtiu-tments to the poertN line to capture both spatial and temporal pnce variation. %kith paricular emphasis on diflcretice in price; %%hich are releant to the poor. Ity recommendations in this regard can be briefta summarized asp aollors a) The all-india po%em~ line corresponding it1 the 1973-74 o Thicial po%e. line was selected as a base. Thia pofert line was based on a calorie norm o2400 kcal per person in rural ares and amounted to Rs t9 per capita per month In urban areas it %Bas based on a calone nomi of 2100 kcaK and amounted to Rs 57 per capita per month he poert\ line represented the estimated aerage total per capita nionth ependiures at it hich the calorie noms iere picall y aried in the rural and urban sectors. respecti%eIl%. b) Instead of usig a unilon consumpion price deflator for the rural and urban areas deied aom the National Accouens Statistics (NAS), tihe Consumer Price Wnder or Agricultural Labourers. th consumpion meights deried tom ihe expenditure classes around the poterm line %%ere iaken for rural areas. For urban areas, a simple a, erage of Consumer Price Indices f r Industrial Worers and to r Non-anual Employees %ith consumpion %eights deried from e\penditure class around the poert\ line ere selected C) Spatial cost of lIing adjustments 0 ere made on the basisofsa7e-leel Fisher price Thdice calculated for 1960-61. Secondt the Expert Group came out irongl\ against the preailing practice of adrusring NSSO consumption tigures to eliminate the discrepanco betmeen the NSSO data and National Accounts data The Group pointed out that the ad-hoc scaling up NSS Figures. b a uniform f3cor. pas no consistent with aailable bidence thai discrepancies haried sharpl across commodir groups (fhr eample. for certain food iAems the consumptson figures obtained b the NSSO exceed those obtained by the NAS) soreouer. the NS consumption figures are calculated as a residual br deducring from etaged production of ianous goods and sen ices Jadjusted for foreign trade) the estimated use for capital formation and public consumption This calculation relies on incomplete infoi mation. There is thus little reason to suppose that the NAS consumption~ figures are more reliable than the NSS ones Although the adoption of the Expert Group recommendations represents a significant impro%ement in the credibilit of poier estimates in India. there do remain areas for further reseach Thi was emphasized bN the Group itself in its report E aples of important research direct ions include a) Further refinement of the pocert line to better incorporafte temporal and spantal pnce varieion ahich is particularl relant io the poor. The EperS Group indicated that its recommendations were deduen in part by the limited aiulabilt of data Fuher iproements are arranted, but require access to additional data and further empincal ork b) In a countrN the size of India it is clear that consumption patos %ar significantly. Additional ork is required to deise a manner in mhich to take this into account itohout jeopardizing the corparability ahich underpins the curen approach. Source Expert Group 1993). Annex 1 43 main difference, however. The Expert Group used the Consumer Price Index for Agricultural Labourers usual Consumer Price Index for Agricultural (CPIAL) for the rural sector. The adjustment carried Labourers. It is, however, well recognized that this out to the CPIAL was for the price of firewood which index is flawed, due to an odd practice of never has been held constant in the official CPIAL series updating the nominal price of firewood. The studies on since 1960-61. The nominal state-level distributions which this report is based have corrected for this were further normalized for inter-state cost of living problem. Thus, they constructed a consistent set of differentials estimated separately for urban and rural price indices across states and survey periods, using areas. For further details on the construction of the monthly data on consumer prices indices from the prices indices, see Ozler, Datt and Ravallion (1996), labour Bureau. The primary deflators were the Ravallion and Datt (I 996b), Datt (1997). Consumer Price Index for Industrial Workers (CPIIW) for the urban sector, and the adjusted all-India  Annex II 45 Village Studies 1. There has been a rich tradition of village studies 3. The studies included in this survey typically in India--some dating back to before India's comprised at least one re-visit to the village, although independence--undertaken by investigators from a in some cases, reports from single visits were included wide range of disciplinary backgrounds. In addition to when either the investigator was in the field for a very the anthropologists who might be expected to specialize long time, or where there was a specific effort to report in such studies, agronomists, demographers, economists, on aspects of change over time. This reduced the geographers, historians, political scientists and population of applicable village studies considerably (the sociologists have studied village economies. Many of library of the Agricultural Economics Research Centre these studies are detailed, assembling a combination of (AERC) of Delhi University, for example, is stacked both qualitative and quantitative information, collected from floor to ceiling with reports of village studies, most on the basis of a whole range of methodologies, of these are single visits however). Even so, there was a including formal questionnaires, informal interviews, daunting array of studies on which to base this group discussions, personal observation, archival investigation. searches. These studies typically offer a rounded perspective on the life of rural households, 4. While the study of an individual village by relationships with their surrounding community and definition concentrates on a very specific location, this the role of village institutions; information which review attempted to comment on developments in rural cannot always be easily distilled from larger-scale India more broadly. There is clearly no statistical basis surveys. for generalizing beyond the village studies which are being examined. Although care was taken to examine 2. A survey of village studies, which was prepared a number of studies, covering a wide range of (Jayarama and Lanjouw, 1997) for this poverty locations, the studies were not selected randomly. Nor assessment, focused on the subset of Indian village are the data collected in the villages of an identical studies which have been concerned with the analysis of form, permitting ready comparisons. In fact, one of the change over time. Such an exercise is worthwhile for challenges in undertaking this review was in the at least four reasons. First, it could indicate whether interpretation of findings from the different village the findings from large-scale surveys accord with the studies, undertaken by investigators with dissimilar observations and experiences at the village level. disciplinary backgrounds, varying methodological Where they do not, it is clearly warranted to investigate frameworks, and differently tinted spectacles. possible reasons. Second, although village studies across a country as large and heterogeneous as India 5. How comfortably can one extrapolate findings would be expected to display a wide variety of from these village studies to rural areas at large is thus experiences, there may also exist some commonalities. ultimately a question of judgment and personal Both differences and commonalities could shed useful inclination. Caution is clearly warranted. It is only light on the processes, and operation of institutions, When clear patterns are observed across village studies, which govern change-- aspects which large surveys that the authors propose that these might be part of a may be less well equipped to investigate. Third, where general pattern. The focus is also on identifying and there are common findings across village studies, one understanding possible diferences across regions, and may be able to develop indicators which could be clearly avoid generalizing when these are observed. usefully added to future large-scale survey questionnaires. Fourth, a review of village studies 6. A review such as this is difficult to undertake findings may raise new questions for further without some soi of benchmarks from which to investigation. itrtrpretw the results. The authors took advantage of 46 Annex II their familiarity with the experience of economic least to some degree, the nature of the questions asked. development in the village of Palanpur, in western Lanjouw (1992) wrote his doctoral dissertation at the Uttar Pradesh, and used it as a sounding board against LSE on the evolution over time of living standards in which results from other village studies were Palanpur. compared. They also allowed Palanpur to dictate, at STATISTICAL APPENDIX 48 Statistical Appendix TABLE OF CONTENTS Table A. I Poverty in India 1951 - 1994 (with correction for CPIAL) Table A.2 Trend annual rates of growth in mean consumption, poverty measures and Gini indices in Indian States, 1957-58-1993-94 Table A.3 Change in living standards in rural areas, 1957-60 to 1990-94 Table A.4 Change in living standards in urban areas, 1957-60 to 1990-94 Table A.5 Growth-redistribution decomposition of changes in poverty, 1951-1994 Table A.6 Sectoral decomposition of changes in poverty, 1951-1994 Table A.7 Poverty and Distributional Measures for Indian States Andhra Pradesh Assam Bihar Gujarat Jammu & Kashmir Karnataka Kerala Madhya Pradesh Maharashtra Orissa Rajasthan Tamil Nadu Uttar Pradesh West Bengal Punjab (including Haryana) Statistical Appendix 49 Table A.1: Poverty in India 1951 - 1994 (with correction for CPIAL) (Poverty line = Rs. 49 per capita per month at Oct 73 - Jun 74 rural prices) NSS Survey period Headcount index Poverty gap index Squared poverty gap index round Rural Urban National Rural Urban National Rural Urban National 3 Aug 51-Nov $2 47.37 3546 4531 16.050 11.141 15.199 7.531 4.824 7.062 4 Apr 52-Sep 52 43.87 3671 42.63 14.637 10.914 13.990 6705 4.408 6306 5 Dec 52-Mar 53 48.21 40.14 46.80 16.290 13.246 15.760 7.562 5.955 7.282 6 May 53-Sep 53 54.13 4277 52.15 19.030 13.831 18.123 9118 6.286 8.624 7 Oct 53-Mar 54 61.29 49.92 59.30 21.946 17.238 21.123 10.263 7.739 9,822 8 Jul 54-Mar 55 64.24 46.19 61.07 25,041 15,763 23.413 12503 7024 11.542 9 May 55-Nov 55 51.83 43.92 5044 18443 14648 17.775 8.804 6.398 8.381 10 Dec 55-May 56 48.34 43 15 47.43 15.646 13.342 15,240 6.710 5.411 6,481 11 Aug 56-Feb 57 58.86 51.45 57.55 19.449 18.162 19.221 8.496 8.509 8.498 12 Mat 57-Aug 57 62.11 4888 59.77 21.685 16.308 20.732 10005 7246 9.516 13 Sep 57-May 58 55.16 47.75 53 84 19.011 15.950 18.467 8.778 6.997 8.462 14 Jul 58-Jun 59 53.26 44.76 51.75 17.736 13.748 17.025 7.882 5.871 7.524 15 Jul59-Jun60 50.89 49.17 50.58 15.289 15.829 15.386 6.129 6.749 6.240 16 Jul 60-Aug 61 45.40 44 65 45 27 13.601 13.842 13.644 5.532 5828 5.585 17 Sep 61-Jul 62 47.20 43.55 46.54 13.601 13.789 13.635 5.314 6.046 5.447 18 Feb 63-Jan 64 48.53 4483 47.85 13.883 13,291 13.774 5486 5174 5.429 19 Jul 64-Jun 65 53.66 4878 52.75 16.083 15.241 15A926 6.602 6.383 6.561 20 Jul 65-Jun 66 57.60 5290 56.71 17.968 16.821 17.751 7603 6.981 7,486 21 Jul66-Jun67 6430 52.24 62.00 22.010 16.810 2101 10.010 7.187 9.472 22 Jul 67-Jun 68 63.67 52.91 61.60 21.802 16.930 20.863 9852 7220 9.345 23 Jul 68-Jun 69 5900 49.29 57.11 18.956 15 539 18.291 8.165 6.542 7.849 24 Jul 69-Jun 70 5761 4716 55.56 18.237 14.316 17.466 7729 5.863 7,362 25 Jul 70-Jun 71 54.84 44.98 52.88 16.545 13.347 15.910 6.798 5.349 6.510 27 Oct 72-Sep 73 55.36 45 67 5337 17.348 13.459 16.548 7328 5 262 6,903 28 Oct 73-Jun 74 55 72 47.96 54.10 17.175 13.602 16.430 7.128 5.219 6.730 32 Jul 77-Jun 78 50.60 40 50 48.36 15,025 11.687 14.284 6.057 4.526 5.717 38 Jan 83-Dec 83 4531 35.65 4300 12.649 9.517 11.901 4.841 3.557 4.534 42 Jul 86-Jun 87 38.81 3429 37.69 10.013 9.100 9.787 3.700 3395 3.625 43 Jul 87-Jun 88 39.23 36.20 38.47 9.275 9.121 9.237 2.982 3.056 3.000 44 Jul 88-Jun 89 39.06 3660 38.44 9.504 9,537 9.512 3.291 3.293 3292 45 Jul 89-Jun 90 34.30 33.40 3407 7.799 8.505 7.979 2.575 3.038 2.693 46 Jul 90-Jun 91 36.43 32 76 35.49 8.644 8,509 8609 2.926 3 121 2976 47 Jul 91-Dec 91 37.42 33.23 3634 8.288 8.244 8277 2.680 2.902 2.737 48 Jan 92-Dec 92 43.47 33 73 4093 10.881 8.824 10.345 3810 3-191 3.649 50 Jul93-Jun94 36.66 30.51 35.04 8.387 7405 8.128 2.792 2.417 2.693 Trend rate of growth during -0.86 -075 -0.88 -1.84 -1.43 -1 79 -2.65 -2.05 -2.56 1951-94 (% per year) Note: All poverty measures are expressed as percentage. Source: Datt, 1997 50 Statistical Appendix Table A.2: Trend annual rates of growth in mean consumption, poverty measures and Gini indices in Indian States, 1957-58-1993-94 (percent per year) State Rural Urban Mean Headco Poverty Squared Gini Mean Headcount Poverty Squared Gini consumption unt Gap Poverty index consumption index Gap Poverty index (Rs/person/ index index Gap index (Rs/person/ index Gap index month) month) Andhra Pradesh 1 14 -212 -3 51 -4.53 -021 0.92 -1,50 -2.11 -2.69 0.17 Assam -037 0.46 038 0.35 -060 0.47 -163 -311 -4.93 0.18 Bihar -0.03 -0.02 -0.90 -1.62 -0.96 0,57 -0.86 -1.61 -2.28 -0.14 Gujarat 070 -1.49 -2 76 -3.69 -058 075 -1 28 -227 -3 07 -0.05 Jammu & 0.34 -051 -0 89 -1.16 0.47 1.32 -3A0 -3.75 -413 0.18 Kashmir Karnataka 0.18 -0.64 -1.15 -1 52 -0.56 0.91 -1.42 -2.01 -2.55 0.06 Kerala 1.63 -2.41 -4.07 -5.24 -021 1.67 -206 -3.37 -428 0.30 Madhya Pradesh 0.21 -0.40 -1.18 -1.79 -0.44 0.45 -0.87 -1.52 -2.02 -0.15 Maharashtra 0.82 -099 -1.60 -2.06 023 0.20 -0.46 -0.65 -0.79 -0.10 Orissa 0.67 -1 47 -2.67 -3.71 -0.35 0,84 -0,95 -1.79 -2 56 -0.09 Punjab (incld. 0.41 -1.92 -2.97 -3 88 0.83 -3 06 -4 55 -5.82 Haryana) Rajasthan 0.16 -054 -1 01 -1.41 -0 52 0.65 -1 42 -2.10 -2.82 -0.16 Tamil Nadu 1.04 -1.45 -244 -3.20 -0.03 0.72 -1.10 -1.53 -1.84 0.16 UttarPradesh 0.33 -070 -1.31 -1.85 -0.22 062 -1.11 -2.12 -3.00 -0.40 West Bengal 0.77 -1.74 -2.74 -3.57 -0.06 0.44 -0.56 -0.99 -145 0.25 Note: These are annual (unconditional) trend rates of growth estimated from regressions of log poverty measures on the mid-point of the survey period for each NSS round. Source. Datt, 1997. Table A.3: Change in living standards in rural areas, 1957-60 to 1990-94 Average over 1957-58 to 1959-60 Average over 1990-91 to 1993-94 Mean H PG SPG Gimu State State Mean H PG SPG Gini consum consump ption tion 76.82 3124 8.376 3.437 Punjab-Haryana Punjab-Haryana 8327 20.64 4.020 1,179 65 19 37.53 8.318 2.510 2639 Jammu & Jammu & 7070 31.20 7.288 2.609 27.87 Kashmir Kashmir 65.45 39.92 9.064 2.882 28.47 Assam West Bengal 68 13 31.51 6.509 1.931 25.75 66.70 4636 14.033 6,043 3647 Rajasthan Kerala 7332 33.01 7962 2.767 30,67 6065 47.38 14.321 5.871 3094 Uttar Pradesh Orissa 66.32 34.66 7.432 2.284 26.29 5635 5033 13.684 5.110 2736 West Bengal Andhra Pradesh 6834 35.89 7.695 2.433 28.39 59.59 54.08 17.476 7.384 3478 Karnataka UttarPradesh 62.57 41.72 10.656 3.799 28.09 58.68 5668 19.394 8.516 3667 Madhya Pradesh Gujarat 5927 41.77 9.635 3.403 24.07 52.45 56.77 18.658 8 119 29.65 Gujarat Tamil Nadu 6393 41 80 11.018 4.090 29.39 53.89 6086 20.308 8.925 29 87 Orissa Rajasthan 6002 45.79 12.540 4.804 2798 47 64 64.52 22.419 10.203 29.76 Bihar Kamataka 57.69 46.88 12.939 4.985 2646 48.44 6514 22.867 10.249 3115 AndhraPradesh Assam 5263 49.33 10799 3.410 19,27 47.71 65.93 21.886 9.503 2921 Maharashtra MadhyaPradesh 6056 49.79 12.737 4.503 30,53 46.08 69 18 28.102 13.845 3466 Kerala Maharashtra 58 76 50.50 14.394 5.553 30.02 44.49 69.65 26.183 12.626 3095 Tamil Nadu Bihar 48.60 63.20 16.401 5.954 22.36 Note. Mean consumption is measured in Rs. per person per day at October 1973-June 1974 all-India rural prices. All poverty measures and the Gin! index are expressed in percentages. The averages are over the first and the last three available NSS rounds for any given state, weighted by the duration of the survey period of the rounds. The states are ranked by ascending order of the headcount index. Source: Datt; 1997. Statistical Appendix 51 Table A.4: Change in living standards in urban areas, 1957-60 to 1990-94 Average over 1957-58 to 1959-60 Average over 1990-91 to 1993-94 Mean H PG SPG Gini State State Mean H PG SPG Gini consumption consumption 8161 15.53 3.530 1553 22 10 Assam Assam 97.75 11.95 1.497 0323 28.94 82.32 28.34 6.905 2.377 32.28 West Bengal Punjab-Haryana 93.41 13 63 2339 0.581 7042 3309 6846 2.036 2834 Jammu & Jammu & 9936 1401 2.702 0755 28.45 Kashmir Kashmir 74 19 36.51 9348 3208 Punjab-Haryana West Bengal 9280 23.79 5.107 1.497 34.37 69.68 43.47 13 575 5 646 35 54 Maharashtra Rajasthan 75.90 29.50 6.770 2.149 29.61 6376 45 18 12.423 5.209 32.03 Rajasthan Andhra Pradesh 7866 3059 7.672 2 644 32.50 65.36 4660 14 178 5 794 3428 TamilNadu Kerala 89.53 30.62 7.775 2.734 37.16 66.98 4763 15.482 6.516 3699 Madhya Pradesh TamilNadu 8414 31.87 8.289 3092 3682 60.28 4803 15.238 6.575 31.73 Andhra Pradesh Karnataka 79.13 34.09 9.301 3.396 34.63 60.17 53.19 17.298 7.431 33 79 Kamataka Gujarat 68.27 37.33 8.947 3 087 29,52 5462 5421 18.429 8.120 30.51 Kerala Maharashtra 74.18 37.47 11.182 4.513 34.86 5747 54.84 18.135 7.843 33.85 Gujarat Madhya Pradesh 71.80 38.17 10.274 3850 33.76 54.64 58.05 24.455 13.330 3841 Orissa Uttar Pradesh 70.63 39.35 10.116 3.478 32.75 5681 5944 21.262 9.658 3755 Uttar Pradesh Bihar 6722 42.39 11.341 4042 31.72 51.84 60.26 20.468 8927 3123 Bihar Orissa 75.06 43.31 13.097 5.147 37.83 Note: Mean consumption is measured in Rs. per person per day at October 1973-June 1974 all-India urban prices. All poverty measures and the Gini index are expressed in percentages. The averages are over the first and the last three available NSS rounds for any given state, weighted by the duration of the survey period of the rounds. The states are ranked by ascending order of the headcount index, Source Dat, 1997 Table A.5: Growth-redistribution decomposition of changes in poverty, 1951-1994 Poverty measure Units Sources of change in poverty Total Growth component Redistribution change during 195 1- component 55 and 1993-94 Headcount index (%age points) -14.87 -2.25 -17.11 (percent) 86.87 13.13 100.00 Poverty gap index (%age points) -6.34 -3.65 -10.00 (percent) 63.44 36.56 100,00 Squared poverty gap index (%age points) -3 40 -2.53 -5.93 (percent) 57.39 42.61 100.00 Note: The Table shows the decomposition of the cumulative change in poverty between May-September 1953 (NSS round 6) and 1993-94, and the cumulative contributions of the components up to 1993-94. We selected NSS round 6 as the reference date for the first decomposition in the cumulative series, since the poverty measures for this round reasonably approximated the average poverty measures for 1951-55. The decomposition numbers can be interpreted broadly as the cumulative change in poverty (and its components) since the mid-point of the period 1951-55. Source: Datt, 1997. 52 Statistical Appendix Table A.6: Sectoral decomposition of changes in poverty, 1951-1994 Poverty measure Units Sources of change in poverty Total Change in Change in Inter-sectoral Covariance change since rural poverty urban population term 1951-55 poverty shift Headcount index (0/oage points) -1395 -3.01 -0.69 0.02 -17.62 (percent) 79.13 17.06 3.93 -0.12 100.00 Poverty gap index (%age points) -8.84 -1.53 -0.23 0.02 -10.57 (percent) 83.59 14.46 2.15 -0.20 100.00 Squared poverty gap index (%age points) -5.23 -0.85 -0.10 0.01 -6.16 (percent) 84.78 13.73 1.70 -0.20 100.00 Note: The Table shows the decomposition of the cumulative change in poverty between 1951-55 and 1993-94, and the cumulative contributions of the components up to 1993-94. Source: Datt, 1997. Statistical Appendix 53 Table A.7: Poverty and Distributional Measures for Indian States 1: Andhra Pradesh N Survey Rural Urban S S period R 0 Mean Head- Poverty Squared Gini Mean Head- Poverty Squared Gini u consump count Gap Poverty index consump count Gap Poverty index n tion index index Gap tion index index Gap d (Rs/pers/ index (Rs/pers index mth) imth) 13 Sep 57-May 58 4988 63 99 22.899 10.404 33.04 59.44 48.52 14.844 6.207 31.35 14 Jul 58-Jun 59 46.89 6672 22.791 9.692 28.61 61.39 45.48 14.099 6.275 31.06 15 Jul 59-Jun 60 4890 64.41 22.920 10.689 32.28 59.80 50.22 16.672 7.150 32.67 16 Jul 60-Aug 61 51 59 6369 19.571 8 004 31.12 61.28 50.93 15.73& 6.470 34.51 17 Sep 61-Jul 62 52.57 58.58 20,188 9047 32.19 56.54 5438 16.493 6,422 30A6 18 Feb 63-Jan 64 53.33 59.77 17.720 7.126 30.49 $9.87 52.29 15.437 6.008 33.04 19 Jul 64-Jun 65 56.23 55.14 16.468 6.632 31.34 59.19 52.56 15.215 5.739 382 20 Jul 65-Jun 66 52.81 62.19 18.735 7541 3136 59.38 50.40 14782 6.111 31.90 21 Jul 66-Jun 67 48.47 63.08 20.652 3.755 28.11 57.28 52.83 15.557 6.247 30.96 22 Jul 67-Jun 68 49A6 63.47 19.373 8.007 28.38 58,61 5235 14.653 5,785 32.01 23 Jul 68-Jun 69 49.10 61.43 19.795 8.636 28.26 61,23 50:85 14.186 5.156 32.29 24 Jul 69-Jun70 50.91 57.44 19.244 8641 29.25 62.70 47.97 14.299 580W 33.4L 25 Jul 70-Jun 71 52,22 57.16 16.579 6,506 27.30 62.76 45,79 12.701 4.911 31,41 27 Oct 72-Sep 73 4794 6448 20,308 8665 27.74 58.50 49.66 11,949 $309 29.68 28 Oct 73-Jun 74 53 64 56 84 17.029 6921 29.52 58.29 51.02 13.763 4.969 29A9 32 Jul 77-Jun 78 59.73 47 84 13.606 5.477 30.25 65.75 43.53 12,196 4.879 3323 38 Jan 83-Dec 83 6740 3799 9.555 3.291 29.72 71.81 35.63 9.376 3.537 31.95 42 Jul 86-Jun 87 70 65 33.96 8.775 3.545 30.92 73.29 38.63 10.99 4.209 35.12 43 Jul 87-Jun 88 72.37 34 02 7.538 2 270 30.50 74.04 39.98 10.113 3.380 35.65 45 Jul 89-Jun 90 6992 31.85 6.844 2.046 27.60 80.92 31.96 8,037 2.33 35.22 46 Jul 90-Jun 91 69.07 36.90 7.843 2.351 29.46 78.40 2791 7398 2,759 31.3 48 Jan 92-Dec 92 61.97 41 85 9,422 3.148 26,78 78,12 33,0$ 8,776 3.141 33.67 50 Jul 93-Jun 94 73.99 2893 5A19 1.801 28,92 79.47 30.82 6.842 2.032 32.29 Note- In this and following tables for other states, mean consumption is expressed in Rs. at October 1973-Jun-1974 il-India rural and urban prices respectively All poverty measures and Gini indices are expressed as percentages. Source: Datt, 1997. 54 Statistical Appendix Table A.7: Poverty and distributional measures for Indian States (contd.) 2: Assam N Survey Rural Urban S period S R 0 Mean Head- Poverty Squared Gini Mean Head- Poverty Squared Gini U consump count Gap Poverty index consump count Gap Poverty index n tion index index Gap tion index index Gap d (Rs/pers/ index (Rs/pers/ index nth) mth) 13 Sep 57-May 58 65 72 37 32 7878 2224 25.98 7980 16.49 1.875 0.333 2541 14 Jul 58-Jun 59 6476 3875 9661 3 332 28.04 87 92 12.34 1 881 0.446 18.92 15 Jul 59-Jun 60 65.93 4305 9357 2926 30.78 76,67 18.01 6.419 3.575 2281 16 Jul 60-Aug 61 68,06 32 13 5969 1 524 2552 84.37 41.55 5254 1.030 3682 17 Sep 61-Jul 62 58.49 4273 9026 2619 2257 8608 23.19 6.159 2 308 32 27 18 Feb 63-Jan 64 61.82 3634 7354 2 143 2238 9263 14.83 1453 0.213 3241 19 Jul 64-Jun 65 5883 3499 6621 1878 19.55 76.16 24.90 4.124 1 031 27.67 20 Jul 65-Jun 66 55.13 4469 9590 3 103 21 04 79.47 25 11 4616 1.161 2952 21 Jul 66-Jun 67 5050 6207 15815 5416 2455 8421 35.39 9.229 3.624 41.55 22 Jul 67-Jun 68 49 51 5503 11.746 4037 18.08 67.60 33.52 7.379 2.281 27.11 23 Jul 68-Jun 69 4699 62 77 15.531 5 399 19.46 7023 27.63 5.102 1 495 25.12 24 Jul 69-Jun 70 5244 4920 10415 3 184 1894 8149 21.33 4,133 L266 2857 25 Jul 70-Jun 71 51.45 5065 10479 3 174 18.02 82.10 17.29 2.692 0582 2598 27 Oct 72-Sep 73 4913 5827 12927 4101 1876 6989 32.13 6.495 1799 27.08 28 Oct 73-Jun 74 4982 56,00 13.751 4913 2084 6857 40.15 9.078 2708 3046 32 Jul 77-Jun78 4743 64.21 14480 4685 1881 72.83 3758 8562 3037 3451 38 Jan 83-Dec 83 5467 46,20 9 190 2635 19.85 77 56 21.68 4.663 1.434 26.54 42 Jul 86-Jun 87 5630 44.26 9610 2972 2166 8933 2781 6960 2.292 34.41 43 Jul 87-Jun 88 5859 43 01 9.078 2 680 2324 95 83 12.26 1.554 0.296 28.04 45 Jul 89-Jun 90 5886 41.90 7.937 2 144 21 85 91 24 12,16 1.749 0.348 2633 46 Jul 90-Jun 91 56.28 4240 8850 2748 2026 89.12 19.34 3,448 0868 28.79 48 Jan 92-Dec 92 49.05 5661 13914 4770 1966 101.41 6.55 0.214 0.009 2905 50 Jul 93-Jun 94 52.57 4899 9632 2711 1789 10273 995 0828 0092 2898 Statistical Appendix 55 Table A.7: Poverty and distributional measures for Indian States (contd.) 3: Bihar N Survey Rural Urban S period S R 0 Mean Head- Poverty Squared Gini Mean Head- Poverty Squared Gini u consump count Gap Poverty index consumap count Gap Povrty index n tion index index Gap tion index index Gap d (Rs(pers index (Rs/pers iodtx mth) mtlt) 13 Sep 57-May 58 4605 65 36 24723 12596 31 65 5524 6030 22.849 10.627 37.28 14 Jul 58-Jun 59 4804 6622 23.171 10393 31.28 52.20 58.63 18.696 7.774 29.65 15 Jul 59-Jun 60 48.43 62.19 19939 8218 26.83 48.92 61.56 20.455 8804 2826 16 Jul 60-Aug 61 7154 47 34 14.563 5,736 39.26 62.44 48,13 16,918 6,939 33,76 17 Sep 61-Jul 62 52.40 56 86 17.039 6940 28,14 78.23 46,14 15.170 6.618 44.82 18 Feb 63-Jan 64 54.15 5496 16.158 6.508 2894 60.87 52.26 15579 5.90 32.92 19 Jul 64-Jun 65 50.98 59.79 18.717 7.981 29.13 56.68 55.03 16.958 6.856 31.99 20 Jul 65-Jun 66 47.81 6751 22.803 10.108 30.61 50.81 62.88 21 696 9.647 32.47 21 Jul 66-Jun 67 35.78 8031 35.481 19589 31.36 47.62 67.85 25.750 12.422 34.42 22 Jul 67-Jun 68 3768 7708 32.790 17401 30.30 51.08 62.36 22.742 10.24 34.33 23 Jul 68-Jun 69 44.64 6753 23.811 11 164 28.15 57.31 53.15 17.128 7.350 32.62 24 Jul 69-Jun 70 46.59 6602 22.216 9962 28.51 58.13 53.42 1633 6.912 3292 25 Jul 70-Jun71 45.04 67,29 22.677 10.109 26.89 60,31 52.5$ 16.342 6.786 34.56 27 Oct 72-Sep 73 44.98 6919 24317 11202 29.25 58.46 52.61 16.636 6911 3219 28 Oct 73-Jun 74 4473 6954 23618 10.672 28.02 51.58 57,35 17.056 6.889 27.10 32 Jul 77-Jun 78 4625 66.21 21.007 8895 26.22 58.39 51.90 15.651 6.262 31.65 38 Jan 83-Dec 83 44.61 6994 22.477 9.535 26.01 59.19 50.32 14.701 5.680 30.82 42 Jul 86-Jun 87 5175 56.45 15.575 5,719 24.99 67.65 42.78 12.995 5.342 34.15 43 Jul 87-Jun 88 51 80 5857 15.063 5 194 24.91 59.3Q 52.94 13.745 4.603 30.34 45 Jul 89-Jun 90 4997 58,57 15,422 5674 23.34 75.39 42.29 11.45& 3.987 37.12 46 Jul 90-Jun 91 51.13 5829 12,292 3.875 18,90 72.35 41,13 11.213 3,974 34.65 48 Jan 92-Dec 92 47.22 6781 19663 1665 2513 60,70 46,32 13,028 4,932 29.58 50 Jul 93-Jun 94 47.46 63 51 17.249 6.323 22.45 68.40 39.72 9382 3,22 30.91 56 Statistical Appendix Table A.7: Poverty and distributional measures for Indian States (contd.) 4: Gujarat N Survey Rural Urban S period S R 0 Mean Head- Poverty Squared Gini Mean Head- Poverty Squared Gini U consump count Gap Poverty index consump count Gap Poverty index n tion index index Gap tion index index Gap d (Rs/pers/ index (Rs/perst index mth) mth) 13 Sep 57-May 5 14 Jul 58-Jun 59 4885 65 48 23.251 10553 32.32 57.14 54.24 18.135 7.886 33 30 15 Jul 59-Jun 60 5549 5602 18302 1780 3230 5274 60.83 20.522 8925 3289 16 Jul 60-Aug 61 52.93 4994 15,027 6322 25.08 61 82 50.22 16088 6.879 3515 17 Sep 61-Jul 62 5252 57 12 16.057 6059 26.89 64.16 42.42 12089 4769 3120 18 Feb 63-Jan 64 52.46 5993 18248 7385 2999 64.45 44.92 11.807 4365 33.05 19 Jul 64-Jun 65 48.13 6882 21.641 8 959 29.70 51.37 60.08 16.596 6 165 2667 20 Jul 65-Jun 66 4591 67 62 23079 10 353 28.70 5265 5805 17.963 7.368 2947 21 Jul 66-Jun 67 44.48 6926 25.507 12049 29.89 50.84 61.06 18.967 7.921 29.31 22 Jul 67-Jun 68 4584 6493 23.000 10549 28.35 52.07 5749 17.322 6.917 2790 23 Jul 68-Jun 69 50.33 5824 19.124 8532 2851 48.89 62.22 18.279 7.323 26,00 24 Jul 69-Jun 70 47.17 6582 21.262 8.972 2759 52.17 58.11 17,669 7.143 2863 25 Jul 70-Jun 71 4953 61.38 18816 7673 27.20 5646 53.03 13541 4.685 27.63 27 Oct 72-Sep 73 52.03 61 60 19.140 7851 3079 51 73 57.69 15241 5427 24.96 28 Oct 73-Jun 74 5043 5807 15 529 5 740 23 94 4991 59.99 17.095 6.674 2555 32 Jul 77-Jun78 56.24 5527 17.717 7447 32.55 65.04 45.31 11.09 4020 32.12 38 Jan 83-Dec 83 62.58 3923 8.735 2.721 2586 63.56 41.65 9.527 2.953 27.73 42 Jul 86-Jun 87 6274 4263 12206 4958 30.59 62.89 46.10 12.019 4.290 30.83 43 Jul 87-Jun 88 59.08 42.93 9341 3074 2423 65.68 38.51 8.813 2764 28.08 45 Jul 89-Jun 90 6232 36.98 7.813 2.396 23.57 7927 37,55 8,027 2.397 37.72 46 Jul 90-Jun 91 57.51 43.13 8,006 2.148 20.40 61.46 44,19 11548 4,255 2859 48 Jan 92-Dec 92 56.85 4678 13528 5745 2781 69.75 37.14 8.536 2865 30.83 50 Jul 93-Jun 94 63.46 35 39 7371 2.316 24 73.59 30.66 6.706 2.14 29 13 Statistical Appendix 57 Table A.7: Poverty and distributional measures for Indian States (contd.) 5: Jammu & Kashmir N Survey Rural Urban S period S R 0 Mean Head- Poverty Squared Gini Mean Head- Poverty Squared Gini u consump count Gap Poverty index consump count Gap Poverty index n tion index index Gap tion index index Gap d (Rs/pers/ index (Rs/pers/ index inth) math) 13 Sep 57-May 58 14 Jul 58-Jun 59 15 Jul 59-Jun 60 16 Jul 60-Aug 61 6705 3746 8046 2321 2760 6846 35.91 7.713 2.347 28.37 17 Sep 61-Jul 62 63.41 4044 9.152 2 815 2677 66.64 34,90 7.172 2.402 26.73 18 Feb63-Jan64 64.65 3494 7869 2450 2463 7617 28.13 4.987 1336 2979 19 Jul 64-Jun65 60.88 3656 7009 1899 21.97 6662 36.05 7.031 1.991 2678 20 Jul 65-Jun 66 62.04 33 07 7 100 2 280 22.62 60.41 40,09 8.043 2.268 2349 21 Jul 66-Jun 67 56.92 41.89 9589 3 205 21.97 64.05 40.44 8.084 2.273 2663 22 Jul 67-Jun 68 6216 2990 6.136 1.887 20.49 7126 28.06 4.420 1.020 25.45 23 Jul 68-Jun69 6971 2354 4693 1359 2251 71.74 25,50 3.975 0.898 23.91 24 Jul 69-Jun 70 67 18 2685 5.432 1597 2274 69.63 27,47 5.017 1297 2497 25 Jul 70-Jun 71 74.78 2129 3993 1059 23.94 74.08 26.57 4.957 1282 27 13 27 Oct 72-Sep 73 62 16 3525 7.670 2433 2366 71.49 27.50 4.750 1.131 2549 28 Oct 73-Jun 74 53 35 5225 12.882 4354 23 24 70.38 24.95 3.883 0 841 2327 32 Jul 77-Jun78 58 14 4287 8.671 2584 2246 77.29 26.88 5.067 1.524 30.56 38 Jan 83-Dec 83 67.16 28 15 4.753 1126 22.68 78.28 17.76 2.346 0.428 24.54 42 Jul 86-Jun 87 7260 30.67 5.904 1 665 28.13 84.41 19 18 3,573 0.927 28.46 43 Jul 87-Jun 88 74.50 30 78 5.830 1 586 2973 90.45 15.38 2,374 0.567 28.43 45 Jul 89-Jim 90 79.25 2002 2650 0552 2756 101.16 7,86 1.402 0.349 2209 46 Jul 90-Jun 91 58.35 42.81 13.383 5.690 2631 10646 18.79 4,331 1349 3482 48 Jan 92-Dec 92 50 Jul 93-Jun 94 58 Statistical Appendix Table A.7: Poverty and distributional measures for Indian States (contd.) 6: Karnataka N Survey Rural Urban S period S R 0 Mean Head- Poverty Squared Gini Mean Head- Poverty Squared Gini u consump count Gap Poverty index consump count Gap Poverty index n tion index index Gap tion index index Gap d (Rs/pers/ index (Rstpers/ index mth) mth) 13 Sep 57-May58 65.95 48.80 15.789 7.017 38.11 55.22 51.87 17.727 8.056 30.55 14 Jul 58-Jun 59 5949 53.84 17.352 7.174 3433 63.32 47.67 15.045 6.668 31.01 15 Jul 59-Jun 60 54.92 58.27 18.865 7.870 32.74 60.72 59.69 19.229 7.723 38.99 16 Jul60-Aug61 59.19 47.42 13.669 5.237 2876 72.0$ 48.82 14.640 5.584 4070 17 Sep 61-Jul 62 69.96 44.58 12.234 4.762 16.92 61.89 52,03 17207 7 438 3642 1 Feb 63-Jan 64 53.70 5774 16491 6.332 2898 55.36 57.57 19092 8413 3408 19 Jul 64-Jun 65 48.50 62.89 20.410 8.701 28.18 59.50 49.53 16.003 6.984 3269 20 Jul 65-Jun 66 42.36 72.56 28.104 13870 30.91 52.34 61.81 20.273 8.514 32.15 21 Jul 66-Jun 67 44.80 67.61 25.512 12.382 30.68 49.02 65.29 21.558 9.222 30.41 22 Jul 67-Jun 68 4673 66.84 23329 10666 30.23 51.60 59.61 20,208 9.004 31.35 23 Jul 68-Jun 69 46.90 69.73 24.345 11 170 32.00 56.52 56.20 18.436 8.013 34.02 24 Jul 69-Jun 70 47.92 64.28 20.024 8 181 26.79 61.96 48.38 14,997 6.469 33.82 25 Jul 70-Jun 1 51.94 58.00 17440 7.063 28,16 64.50 46.00 12981 4.890 32.67 27 Oct 72-Sep 73 51.89 56.57 17.153 7222 2790 59.51 51.72 15850 6.410 3283 28 Oct 73-Jun 74 50.77 6097 18.799 7.660 28.67 55.92 53.78 15.882 6.241 29.85 32 Jul 77-Jun 78 5622 5422 17.661 7725 32.56 63.34 48.85 15.462 6.65S 3561 38 Jan 83-Dec 83 61.84 44.68 12.978 5.157 30.40 73.67 37.46 11.071 4.480 34.46 42 Jul 86-Jun 87 59.24 46.15 13.815 5618 28.98 69.68 42.21 13.096 5.384 35.07 43 Jul 87-Jun 88 62.31 43 49 11.272 4.000 28.94 69.40 42.04 10.990 3.790 33.37 45 Jul 89-Jun 90 53.63 54.43 15.246 5720 26.63 79.46 34.46 9,012 3.138 3522 46 Jul 90-Jun 91 58.63 42.73 13304 5 587 2629 82.46 34.S0 9,613 3 499 3679 48 Jan92-Dec 92 51.93 56.94 15.759 6.023 2612 75.11 37.76 10818 4148 3522 50 Jul 93-Jun 94 62.52 40.97 9.754 3345 26.97 79.82 29.71 7.472 2.542 31.87 Statistical Appendix 59 Table A.7: Poverty and distributional measures for Indian States (contd.) 7: Kerala N Survey Rural Urban S period S R 0 Mean Head- Poverty Squared Gini Mean Head- Poverty Squared Gini u consump count Gap Poverty index consump count Gap Poverty index n tion index index Gap tion index index Gap d (spers/ index (Rs/pers/ index mth) mth) 13 Sep 57-May 58 4727 6688 26948 13455 3496 6059 52.60 18.178 8.019 3554 14 Jul 58-Jun 59 4588 68.86 29.125 14571 35.15 51 49 56.31 17.639 7220 27,01 15 Jul 59-Jun 60 4540 7121 27.944 13410 3395 5326 5331 19.40S 9.095 3023 16 Jul60-Aug 61 46.64 6865 25655 11765 3220 53 60 6194 23.564 11,075 3676 17 Sep 61-Jul,62 52.25 5920 21 103 9861 3302 56.08 58.36 20991 9899 3682 18 Feb 63-Jan 64 49.50 6324 21070 9140 30.18 5691 5780 21011 9886 3741 19 Jul 64-Jun 65 45.21 6940 27.314 13.644 33.57 52.95 64.71 24 689 11.990 3900 20 Jul 65-Jun 66 38 10 7976 32.929 16400 2983 40.33 80.01 33.474 9.999 34.34 21 Jul 66-Jun 67 40.04 7737 30752 14944 30.13 52.11 64.69 28.765 15.880 4298 22 Jul 67-Jun 68 43.35 74 13 27.803 13 329 31.85 43.75 74.35 31.164 15968 36.63 23 Jul 68-Jun 69 48A9 7376 30.231 15205 41.38 46.19 70.95 29.646 15321 37,83 24 Jul 69-Jun70 4017 78 15 30.913 15 168 30.92 52.45 66,64 26.866 13.513 41,26 25 Jul 70-Jun71 4379 7265 28.391 13.926 3304 5579 6103 23380 11.461 39.77 27 Oct 72-Sep 73 4765 6679 23487 10665 3140 5631 6030 23552 11.518 3949 28 Oct 73-Jun 74 50.46 6206 21 570 9810 3202 53.69 62.72 23.840 11.253 37 57 32 Jul 77-Jun78 59.75 52.66 17.898 8167 35.86 6008 56.05 21.594 10.719 4037 38 Jan 83-Dec 83 6770 4370 11.687 4296 33.88 72.54 44,72 13.496 5.549 39.36 42 Jul 86-Jun 87 72 58 3966 9.996 3.508 34.82 7267 48.81 16.089 6.881 43.33 43 Jul 87-Jun 88 7290 3467 8.284 2714 31.70 7822 38,02 9.840 3.329 36.39 45 Jul 89-Jun 90 6629 38.87 9,582 3328 27,88 87.99 3030 6.467 1.880 3736 46 Jul 90-Jun 91 68.81 33 80 8246 2 789 2724 9431 34,77 9.378 3 314 3864 48 Jan 92-Dec 92 77.70 34 15 8.635 3099 3470 84.97 34,01 9.004 3238 3852 50 Jul 93-Jun 94 7344 3107 7004 2412 30.07 89.32 23.07 4.944 1651 3432 60 Statistical Appendix Table A.7: Poverty and distributional measures for Indian States (contd.) 8: Madhya Pradesh N Survey Rural Urban S period S R 0 Mean Head- Poverty Squared Gini Mean Head- Poverty Squared Gini u consump count Gap Poverty index consurnp count Gap Poverty index n tion index index Gap tion index index Gap d (Rs/pers/ index (Rs/pers/ index mth) mth) 13 Sep 57-May 58 56.35 62.86 24.000 11544 41.12 64.49 45.53 15.733 7.196 3487 14 Jul 58-Jun 59 59.23 56.49 15.904 8.092 36.66 6879 45.57 13.920 5.499 3645 15 Jul 59-Jun 60 5987 52.24 16.427 6.668 33.33 67.04 51.27 16.856 7.023 39.12 16 Jul 60-Aug 61 56.50 50.91 15.456 6.438 29.59 62.70 55,58 15S43 5.787 3407 17 Sep 61-Jul 62 64.18 4813 14.006 541 34.22 57.27 49.67 17 594 9019 3360 18 Feb 63-Jan 64 64.51 45.40 13.508 3.422 33 51 6073 5629 16.714 6.218 35 14 19 Jul 64-Jun 65 53.32 5020 15056 6155 30.99 59.50 53.15 16.173 6.390 33.25 20 Jul 65-Jun 66 55.18 57.23 17404 7 158 31.77 5321 58.15 19316 8.420 31 68 21 Jul 66Jun 67 45.00 67.68 24.305 11.411 29.17 55.98 58.00 19.670 8.586 3477 22 Jul 67-Jun 68 44.49 70.96 27.319 13.337 32.60 5056 64.31 22.183 9.835 32.98 23 Jul 68-Jun 69 48.78 65.55 22.975 10.661 32.45 58.97 54.94 19.715 9.356 37 54 24 Jul 69-Jun 70 49.63 63.80 22.566 10.442 32,67 58.36 5412 16,00 6,930 33.56 25 Jul 70-Jun71 50.47 62.42 20.991 9.205 31.24 61.29 51.34 16145 6,797 34.93 27 Oct 72-Sep 73 49.36 64.99 21.713 9.370 30.99 6151 5435 15.747 6.054 35 51 28 Oct 73-Jun 74 47.76 66.09 21.852 9.351 29.16 52.62 56.56 16876 6 663 27.85 32 Jul 77-Jun 78 51.56 64.71 21419 9.317 33.52 6149 59.75 23.649 11819 4463 38 Jan 83-Dec 83 5625 53.04 15.363 5.977 29.60 61.00 48.20 12.854 4.566 3033 42 Jul 86-Jun 87 58.30 54.35 14.975 5.566 31.64 6520 45.19 12.301 4.649 32.98 43 Jul 87-Jun 88 59.55 47.80 12.393 4 342 28.53 67.91 43.45 11.527 4.022 33 12 45 Jul 89-Jun 90 64.41 44.74 11.790 4353 27.39 72.78 39,27 10.168 3,501 3435 46 Jul 90-Jun91 59.52 47.93 12.834 4.662 2907 75,48 40,22 11732 4.647 37.79 48 Jan 92-Dec 92 61.70 56.09 13.945 4.766 3455 69.83 34.45 9.088 3383 3050 50 Jul 93-Jun 94 60.45 45.36 11.431 4.082 27.96 70.08 39.83 10.003 3.519 33 Statistical Appendix 61 Table A.7: Poverty and distributional measures for Indian States (contd) 9: Maharashtra N Survey Rural Urban S period S R 0 Mean Head- Poverty Squared Gini Mean Head- Poverty Squared Gini u consump count Gap Poverty index consurnp count Gap Poverty index n tion index index Gap tion index index Gap d (Rs/pers/ index (Rs/pers index mth) mth) 13 Sep 57-May 58 14 Jul 58-Jun 59 45.48 7081 24.910 11 080 30.28 70.65 38.89 10.730 4.091 31.85 15 Jul 59-Jun 60 46.97 67.97 22.686 9.784 29.39 71.49 4671 14.272 5.754 39.65 16 Jul 60-Aug 61 5026 6000 18.608 7910 28.14 67.31 44.61 15.414 6.886 35.17 17 Sep 61-Jul 62 51 72 58,36 17.583 7 097 28,02 74,19 40.63 *12,618 ,316 3773 18 Feb 63-Jan 64 53.36 5772 17 035 6.680 29.16 70.42 4$29 14.614 6.103 37.57 19 Jul 64-Jun 65 43 51 71 87 24.727 10.919 27.29 72.82 45.85 15.607 6.848 40.19 20 Jul 65-Jun 66 4466 7094 23.922 10.401 2790 70.59 46.63 15.064 6.306 38.97 21 Jul 66-Jun 67 4171 7556 27.495 12.823 28.70 66.53 46.72 15.424 6.734 36.65 22 Jul 67-Jun 68 4269 72.48 24.332 10.832 25.84 62.17 50.23 17.829 8.269 36.69 23 Jul 68-Jun 69 46.05 69.28 23.345 10.011 28.82 63.42 49.73 16.439 6.994 35.45 24 Jul 69-Jun 70 4601 6862 22.120 9.231 27.19 71.38 38.62 11.516 4700 33.84 25 Jul 70-Jun71 48.06 61.96 18.584 7541 25.14 69,79 43.23 13.284 5.40$ 35,95, 27 Oct72-Sep73 41.34 8113 30.340 9.999 31.52 66.94 45.59 15.615 7,240 37.41 28 Oct 73-Jun 74 47.58 6461 20.317 8.506 26.99 59.95 51.17 16.654 7A32 33.77 32 Jul 77-Jun 78 59 65 78 78 21 958 7.709 42.84 70.03 43.51 14,517 6.546 37.61 38 Jan 83-Dec 83 54.98 54.56 15.789 6.073 28.82 68.41 41.79 12.551 5.133 34.25 42 Jul 86-Jun87 55.71 53.92 16.027 6.392 29.98 70.21 41.21 13.094 5.575 35.05 43 Jul 87-Jun 88 59.52 52.32 14.023 5.163 31.52 72.30 38.79 11.481 4.542 34.27 45 Jul 89-Jun 90 60.10 45 53 11.652 4.149 28.25 77,99 42.72 12.785 4955 40.72 46 Jul 90-Jun 91 63.56 4305 11951 4.498 3018 71.33 38.03 11380 4,631 33.68 48 Jan 92-Dec 92 52.05 6063 18,071 7073 29.23 74.10 38.15 1.229 4.490 3523 50 Jul 93-Jun 94 60.67 47.81 13.16 5.089 30.65 77.10 36.23 10.936 4419 35.67 62 Statistical Appendix Table A.7: Poverty and distributional measures for Indian States (contd.) 10: Orissa N Survey Rural Urban S period S R 0 Mean Head- Poverty Squared Gini Mean Head- Poverty Squared Gini u consump count Gap Poverty index consunp count Gap Poverty index n tion index index Gap tion index index Gap d (Rs/pers/ index (Rs/pers/ index mth) mth) 13 Sep 57-May 58 48.06 6506 23 180 11129 32.15 5540 6407 23338 10.619 3945 14 Jul 58-Jun 59 53 10 5608 18 967 8474 30.73 6021 46.79 20.991 12646 36.95 15 Jul 59-Jun 60 5905 62.49 19494 7723 27.30 48.50 64.79 28756 16.047 3908 16 Jul 60-Aug 61 50.85 61 72 20129 8287 2978 44.18 69.12 25,752 12.081 29.55 17 Sep 61-Jul 62 58.90 4689 13909 5912 2954 6803 5081 17760 8016 4133 18 Feb 63-Jan 64 51.23 5804 17954 7468 2789 5589 54.88 18437 8 258 33 12 19 Jul 64-Jun 65 49.39 61 36 18528 7548 2683 53 12 6007 19.023 7.639 3089 20 Jul 65-Jun 66 49 96 59 98 18 280 7 806 27.57 49 69 64.84 23051 10.676 33.45 21 Jul 66-Jun 67 4724 6286 19001 7963 2493 5194 65.47 23.314 10.503 36.09 22 Jul 67-Jun 68 49.93 6340 19.980 8449 29.46 55 11 59.11 21.315 10.125 36.58 23 Jul 68-Jun 69 4461 7029 24.308 11 070 28.73 55.96 61.90 20.050 &.426 36.21 24 Jul 69-Jun 70 46,16 6620 22.902 10637 29 12 60,76 49,19 16.567 7.473 34 14 25 Jul 70-Jun 71 4667 6477 22 173 10 162 2854 56.72 54,43 17 162 7.121 3224 27 Oct 72-Sep 73 47.33 67 03 23693 10952 31 45 5845 55.53 18 133 7853 3546 28 Oct 73-Jun 74 5063 5867 17662 7 108 26.75 55.15 5999 20.203 8.773 3494 32 Jul 77-Jun 78 50.08 62 52 20.443 8 955 30.63 54.07 57.26 19,743 9,057 33.28 38 Jan 83-Dec 83 51.25 56.76 16.962 7 126 2701 5558 54.94 16.701 6.710 3059 42 Jul 86-Jun 87 59 14 4495 11.950 4462 27.35 66.78 49.81 14.793 5.789 37.70 43 Jul 87-Jun 88 5768 47.86 11699 3 840 25,58 6112 47.53 13371 5.014 3062 45 Jul 89-Jun 90 63.77 3948 8.454 2403 25,37 67.69 41.09 11.599 4.385 3220 46 Jul 90-Jun 91 69.70 27 14 5376 1532 2492 72.26 40,42 10,912 3.92S 3484 48 Jan 92-Dec 92 6823 3657 8 195 2530 2937 86.24 48.74 17120 7366 47.96 50 Jul 93-Jun 94 61 04 4028 8724 279 2457 6667 40.76 11257 4148 3069 Statistical Appendix 63 Table A.7: Poverty and distributional measures for Indian States (contd.) 11: Rajasthan N Survey Rural Urban S period S R 0 Mean Head- Poverty Squared Gini Mean Head- Poverty Squared Gini U consump count Gap Poverty index consump count Gap Poverty index n tion index index Gap tion index index Gap d (Rs/pers/ index (Rs/pers/ index mth) mtb) 13 Sep 57-May 58 65.95 50.64 17.935 8495 40.39 64.31 48.26 14.570 5.957 35.13 14 Jul 58-Jun 59 6116 49.07 14510 5734 32.50 70.74 41.11 11.564 5.941 37.92 15 Jul 59-Jun 60 72.79 40.43 10.630 4,513 37.49 56.37 46.94 11672 3.917 23.82 16 Jul 60-Aug 61 55 70 57.47 18 218 7.580 33.12 66.60 46.56 13951 5,634 33.15 17 Sep 61-Jul62 57.66 5589 19.958 9.239 3656 63.12 44.40 12.838 5.113 31.68' 18 Feb 63-Jan 64 58.03 4994 15.420 6.381 3049 70.81 43.28 12,134 4.465 35,79 19 Jul 64-Jun 65 55.10 5579 17.943 7.779 32.25 62.08 48.26 13.501 5.009 31.93 20 Jul 65-Jun 66 54.98 55 13 17,724 7.767 31.8o 57.89 54.51 15.347 5.528 30.73 21 Jul 66-Jun 67 52.06 63.13 22.066 10 191 34.88 56.11 57.95 18.241 7.504 33.68 22 Jul 67-Jun 68 51.65 6002 21 176 10073 33.16 59.82 49:15 13,612 5.184 30.49 23 Jul 68-Jun 69 5247 6724 25016 12.092 39.34 61.36 52.11 15.170 5.780 33.97 24 Jul 69-Jun 70 49.11 6887 25.280 11 829 35.51 60.01 5250 14,9&1 5397 32.98 25 Jul 70-Jun 71 4741 6528 24.459 12.006 32,77 68,57 4412 13.,033 5.277 35,77 27 Oct 72-Sep 73 5002 6342 21.949 9.891 32.19 64.44 47.47 13.601 5.164 33.97 28 Oct 73-Jun 74 5109 59.26 18.460 7.634 28.40 57.03 54.19 14.665 5,197 29.55 32 Jul 77-Jun 78 7621 53.52 14.611 5575 46.28 65.45 42.68 11.191 3.950 31.06 38 Jan 83-Dec 83 62.04 4899 15586 6.654 34.59 6963 36.94 9.606 3548 30.96 42 Jul 86-Jun 87 61.30 45 65 13.168 5 238 30.30 78.44 3331 8.069 2.734 34.22 43 Jul 87-Jun 88 57.99 50.35 14,888 5967 30.34 7379 37.68 9.154 Z992 32.56 45 Jul 89-Jun 90 65.74 4049 12.194 5.190 32.36 7951 31.08 7.450 2,411 33.08 46 Jul 90-Jun 91 64 53 38 96 12.097 5 045 23.54 77,57 29.98 7,219 2.422 30,61 48 Jan 92-Dec 92 5746 50 90 13.761 5 249 28.93 74,25 29.13 6.543 2.037 28 87 50 Jul 93-Jun 94 58.07 47.52 11.762 4.118 26.48 75,87 2938 6,549 1,948 29.36 64 Statistical Appendix Table A.7: Poverty and distributional measures for Indian States (contd.) 12: Tamil Nadu N Survey Rural Urban S period S R 0 Mean Head- Poverty Squared Gini Mean Head- Poverty Squared Gini u consump count Gap Poverty index consump count Gap Poverty index n tion index index Gap tion index index Gap d (Rs/pers/ index (Rs/pers/ index mtli) nth) 13 Sep 57-May 58 4174 7273 29469 15 138 32.12 7008 49.92 17.173 7.511 41 16 14 Jul 58-Jun 59 4676 66 15 24.866 11 987 32.02 64.80 42.62 12.263 4.969 3070 15 Jul 59-Jun 60 4427 7084 25036 11381 2900 62.38 4808 13846 5.330 32.71 16 Jul 60-Aug 61 46.40 65 12 24322 11 565 3052 6047 4256 12.633 5461 29 15 17 Sep 61-Jul 62 53 19 5720 18.501 8 155 3114 64.85 47.04 13.277 5229 3433 18 Feb 63-Jan 64 56.03 53 83 16500 6790 3087 6629 4606 12587 4 730 3449 19 Jul 64-Jun 65 48.45 6469 21.340 9346 2960 6424 50.74 15 130 6.044 3639 20 Jul 65-Jun 66 46.70 6668 22.515 9959 2889 59.75 54.53 17262 7 097 3521 21 Jul 66-Jun 67 4441 7086 24184 10666 2796 5652 53.93 16.888 7.132 3196 22 Jul 67-Jun 68 46.84 65.52 21.310 9.150 27.33 60.67 55.37 17.366 7.339 37 13 23 Jul 68-Jun 69 46.06 6764 22795 10015 28.46 71.02 3834 10.668 4.147 3328 24 Jul 69-Jun 70 4577 6973 24722 11.256 30.69 60.62 51.09 15.670 6.492 33 89 25 Jul 70-Jun 71 46.56 6636 20881 8716 2641 6051 6231 21232 9210 4183 27 Oct 72-Sep 73 5103 5927 17 730 7 153 27.61 60.78 4983 14283 5434 32.02 28 Oct 73-Jun 74 50.93 5929 17801 7 148 2744 6005 4996 14.167 5334 3120 32 Jul 77-Jun 78 5478 5807 18.196 7 560 32.41 63.30 47.69 14.752 6.230 34.48 38 Jan 83-Dec 83 56.42 54.96 17.682 7 587 32 90 6854 4453 13498 5.634 3642 42 Jul 86-Jun 87 61 88 44.78 13.139 5.313 30.46 7225 39.54 10.664 4.147 35.32 43 Jul 87-Jun 88 62.81 48.35 13417 4891 32.33 7337 39.74 10.847 3.933 3537 45 Jul 89-Jun 90 6146 4199 9901 3277 2693 74.43 3502 9030 3 150 32.91 46 Jul 90-Jun 91 61 16 4202 11573 4377 2729 7873 33.80 8624 2989 34.02 48 Jan 92-Dec 92 60.52 4665 12888 4910 2965 9326 3050 8.473 3454 41 60 50 Jul 93-Jun 94 701 3674 8592 2982 3124 8043 3132 7769 2.834 3484 Statistical Appendix 65 Table A.7: Poverty and distributional measures for Indian States (contd.) 13: Uttar Pradesh N Survey Rural Urban S period S R 0 Mean Head- Poverty Squared Gini Mean Head- Poverty Squared Gini u consump count Gap Poverty index consump count Gap Poverty index n tion index index Gap tion index index Gap d (Rs/pers/ index (Rs/pers/ index mth) mth) 13 Sep 57-May 58 53 45 5501 17.498 7 599 2967 58 80 58.43 20.971 9624 3922 14 Jul 58-Jun 59 6094 5069 15.767 6.516 34.01 55.35 6017 21.218 9.433 35.96 15 Jul 59-Jun 60 65.76 3834 10.492 3930 28.83 5677 5948 21,523 9.907 3789 16 Jul 60-Aug 61 6437 4082 11796 4622 29.56 61.80 5028 16,859 7.623 3621 17 Sep 61-Jul 62 73.44 3449 8397 2769 31.60 59,79 53.54 18258 8 156 3611 18 Feb 63-Jan 64 5838 4945 14169 5 513 2977 65.27 53.05 16872 6851 3894 19 Jul 64-Jun 65 5293 5682 17336 7208 29.24 51.79 63.28 23.186 10.821 3545 20 Jul 65-Jun 66 57.12 5077 14.520 5 594 2898 51-82 62.95 22.179 9951 3408 21 Jul 66-Jun 67 5031 5925 18.810 8046 28 10 54.14 60.77 20937 9.366 3534 22 Jul 67-Jun 68 4777 6472 21.142 9 108 28.33 55.25 6086 20.442 8.938 35.93 23 Jul 68-Jun 69 59.00 5029 14571 5828 31 16 5648 5800 20.330 9.089 35.80 24 Jul 69-Jun 70 5504 53 71 16 160 6508 2928 5552 59,12 19.735 8.540 34,88 25 Jul 70-Jun 71 61.09 45 16 11 908 4338 2930 57.51 54.28 17.474 7.410 3338 27 Oct 72-Sep 73 54.66 5 51 15246 5620 2825 5602 56.42 17 122 6770 3171 28 Oct 73-Jun 74 52 18 55 59 14462 5 125 2430 51.63 62.17 19.093 1576 2999 32 Jul 77-Jun 78 6269 45 09 11 804 4 342 3034 57 59 56.55 9395 2,627 25 96 38 Jan 83-Dec 83 61 66 44 72 12.004 4406 29.41 61.00 5003 14.592 5.593 32.32 42 Jul 86-Jun 87 7041 36.22 8970 3 120 30.65 6548 46.85 14.429 5.819 34.76 43 Jul 87-Jun 88 6443 4136 9646 3024 28.26 67.96 44.20 11.923 4.197 33.51 45 Jul 89-Jun 90 71 02 31 14 7009 2 140 2667 68.14 44,71 12.091 4,236 3376 46 Jul 90-Jun 91 62.40 3688 9.079 3255 2560 70.59 3754 9,430 3241 3216 48 Jan 92-Dec 92 61.80 4667 12694 4681 3053 66.59 46.22 12 420 4349 3376 50 Jul 93-Jun 94 63.52 41 6 10195 3462 28 12 7472 34.28 8.498 2844 3233 66 Statistical Appendix Table A.7: Poverty and distributional measures for Indian States (contd.) 14: West Bengal N Survey Rural Urban S period S R 0 Mean Head- Poverty Squared Gini Mean Head- Poverty Squared Gini u consump count Gap Poverty index consump count Gap Poverty index n tion index index Gap tion index index Gap d (s/pers/ index (Rs/pers/ index mth) mth) 13 Sep 57-May 58 53.96 52.92 15.024 5.860 26.81 74.73 3437 10.581 4.521 3296 14 Jul 58-Jun 59 5834 48.31 12.773 4629 28.06 83.60 25.41 5352 1.518 30.46 15 Jul 59-Jun 60 56.15 50.41 13590 5.027 2708 8634 26.75 5.700 1.628 33,60 16 Jul 60-Aug 61 67.80 32.36 7.614 2469 25,60 81.28 29.99 6.776 2.076 32.70 17 Sep 61-Jul 62 5863 5020 11536 3.795 27.68 84.19 27.35 6.705 2,365 33 18 18 Feb63-Jan64 5293 55.68 15.401 5754 26.37 83.14 25.69 6,735 2.443 31.47 19 Jul 64-Jut 65 50.31 5654 15,773 6.086 23.88 72.24 3849 11.091 4.231 33.32 20 Jul 65-Jun 66 46.76 64.38 20.546 8837 26.30 66.05 42.52 11.806 4A59 31.88 21 Jul 66-Jun,67 45.38 6785 21,386 9005 25.56 70.15 37.35 9.895 3.602 3119 22 Jul 67-Jun 68 40.33 7584 26.057 11.580 24.08 64.96 41.83 11.660 4.452 31.02 23 Jul 68-Jun 69 44.25 69.63 21.030 8.418 23.13 70,46 35.12 8.724 3.011 29.67 24 Jul69-Jun70 4938 60.15 17.467 6.827 25.11 77.07 32,68 8.090 2,814 32.50 25 Jrul 70-Jun 71 48.02 62.57 19.519 8.053 26,32 75.62 3236 8.773 3.304 32.92 27 Oct 72-Sep 73 5095 60.51 20.159 9.026 30.64 71 7$ 19.16 11750 4.728 34.40 28 Oct 73-Jun 74 48.78 6324 21.399 9665 30.13 68.13 3927 11.739 4.805 32.31 32 Jul 77-Jun 78 53.70 5625 17.067 7.016 29.75 70.93 3815 10.830 4.165 33.05 38 Jan 83-De 83 56.59 4921 14.786 6202 28.83 7939 32.30 8.324 3.026 34.42 42 Jul 86-Jun 87 64.21 34.10 7.723 2.634 24.27 87.22 30.20 7.510 2.620 37.64 43 Jul 87-Jun 88 65.32 34.87 7.303 2.142 25,57 82.02 32.19 7,125 2.077 34.25 45 Jul 89-Jun 90 71.71 26 19 4.614 1.134 25.23 84.76 25.93 5.493 1.518 32,58 46 Jul 90-Jun 91 6540 39.11 9.520 3.083 27.62 89,09 21.27 4878 1,399 33.28 48 Jan 92-Dec 92 68.54 28.15 5.310 1417 24.21 94.03 25.66 6037 1.939 3598 50 Jul 93-Jun 94 70.45 27.27 4.697 1 294 25.41 95.28 22.45 4.405 1.153 33.84 Statistical Appendix 67 Table A.7: Poverty and distributional measures for Indian States (contd.) 15: Punjab (including Haryana) N Survey Rural Urban S period S R 0 Mean Head- Poverty Squared Gini Mean Head- Poverty Squared Gini u consump count Gap Poverty index consump count Gap Poverty index n tion index index Gap tion index index Gap d (Rs/pers/ index (Rs/pers/ index mth) mth) 13 Sep 57-May 58 7474 3339 9687 3838 3221 71 80 3957 11309 4278 3406 14 Jul 58-Jun 59 81.82 2809 7.239 3495 3690 7662 35.21 8.088 2441 3287 15 Jul 59-Jun 60 7337 3279 8530 3079 31 37 73 55 35.51 9 136 3 172 3244 16 Jul 60-Aug 61 82.06 3200 7282 2249 30 16 83.71 26.25 7219 3210 3383 17 Sep 61-Jul 62 8345 3136 8442 3 106 3533 70.93 4157 12619 5146 3608 18 Feb 63-Jan 64 7087 33 92 8989 3 297 30.05 78.61 38.99 10.610 3873 38 39 19 Jul 64-Jun 65 72.10 36 12 9281 3384 7135 37.59 8750 2.750 20 Jul 65-Jun 66 7043 3843 10042 3858 6508 4713 13 141 4.882 21 Jul 66-Jun 67 6803 3864 9358 3 159 6545 43.14 11.107 4.035 22 Jul 67-Jun 68 6131 4422 12.189 4703 65.58 3707 11.530 4.825 23 Jul 68-Jun 69 7140 3238 8227 2943 68.39 39.15 9416 3.037 24 Jul 69-Jun 70 7230 35 57 9.724 3696 73 73 33.72 8281 2.802 25 Jul 70-Jun 71 7289 31 73 7.895 2742 80.37 2861 6585 2.178 27 0ct 72-Sep 73 81 14 2547 5.591 1765 8208 27.03 6156 2.007 28 Oct 73-Jun 74 69 13 3508 8.556 2 851 71.55 36.40 7849 2.280 32 Jul 77-Jun 78 8190 2546 5950 2010 82.28 32.11 7.762 2.792 38 Jan 83-Dec 83 8289 21 52 4.193 1239 8751 25.29 5615 1.815 42 Jul 86-Jun 87 85.39 2296 4.658 I 326 89.80 16.71 2.785 0.727 43 Jul 87-Jun 88 8587 2000 3 798 1 135 86.78 19.25 3620 0.945 45 Jul 89-Jun 90 8218 14 16 2 755 0820 9674 15.91 2680 0.620 46 Jul 90-Jun 91 82 17 1855 3.413 0943 92.57 13.44 2452 0.631 48 Jan 92-Dec 92 88.41 1814 3474 0987 8731 1602 2844 0.709 50 Jul 93-Jun 94 79.23 2523 5.172 1.608 100.34 1142 1.720 0.404  References 69 References in Text Approach Paper to the Ninth Five Year Plan (1997-2002) Planning Commission, Government of India, 1997. Anand, S and Ravallion, M. 1993. Human Development in Poor countries. On the Role of Private Incomes and Public Services. Journal of Economic Perspectives, Vol. 7, No. 1, Pg. 113-150. Bandopadhyay, D. 1986. Land Reforms in India: An Analysis. Economic and Political Weekly, June 21-28. Bandopadhyay, D. 1988, Direct Intervention Programmes For Poverty. Economic and Political Weekly, Vol. XXIII No 26, June 25, pp. A77-A88. Beck, T. 1996. The Experience of Poverty- Fighting for Respect and Resources in Village India London: Intermediate Technologies Publications. Bliss, C. and Stern, N. 1982. Palanpur The Economy ofan Indian Village Oxford: Oxford University Press. Burra, N. 1995. Born to Work- Child Labor in India. Delhi, Oxford Univ. Press. Chatterjee, M. 1996. Lessons Learned From 20 Years ofICDS, mimeo. New Delhi, World Bank. Chatterjee, M. 1997. Does ICDS Alleviate Poverty? mimeo. New Delhi, World Bank. Datt, G. 1995. Poverty in India 1991-1992- Trends and Decomposition, mimeo. Policy Research Department, the World Bank Datt, G. 1997, Poverty in India and Indian States An Update. International Food Policy Research Institute, Washington D.C. Datt, G. and Ravallion, M. 1996a. Macroeconomic Crises and Poverty Monitoring: A case Study of India. Policy Research Department, the World Bank. Datt, G. and Ravallion, M. 1996b. Why Have Some States of India Performed Better than Others in Reducing Absolute Poverty?' Policy Research Working Paper 1594, the World Bank. (Economica, forthcoming). Drbze, J.P. and Sen, A.K. 1989. Hunger and Public Action. Oxford: Clarendon Press. Dr6ze, J.P. and Sen, A.K. 1995. India- Economic Development and Social Opportunity. Oxford: Oxford University Press. Drbze, J.P. and Srinivasan, P.V. 1995. Widowhood and Poverty in Rural India- Some Inferences from Household Survey Data. Working Paper No. 33, Centre for Development Economics, Delhi School of Economics. Dr&ze, J.P. and Srinivasan, P.V. 1996. Poverty in India. Regional Estimates 1987-88. Working Paper No. 36, Centre for Development Economics, Delhi School of Economics. Drze, J.P. and Sen, A.K. (eds) 1997. Indian Development: Selected Regional Perspectives Oxford: Oxford University Press. 70 References Dr6ze, J., Lanjouw, P. and Sharma, N. (Forthcoming). Credit in a North Indian Village--An Empirical Investigation. New York: Oxford University Press. Eajan, S. Irudaya and Jayakumar, A. 1992. Impact on Noon-Meal Programme on Primary Education: An Exploratory Analysis. Economic and Political Weekly, Oct. 24-31. Expert Group, Government of India, 1993. Report of the Expert Group on Estimation of Proportion and Number of Poor. Planning Commission, Government of India. Fan, Shenggen and Hazell, P. 1996. Should the Indian Government Invest More in Less Favored Areas? Environment and Production Technology Division, International Food Policy Research Institute. Filmer, D., King, E.M., Pritchett, L. 1997. Gender Disparity in South Asia: Comparisons between and within Countries, mimeo. Policy Research Department, World Bank. Foster, J., Greer, J, and Thorbecke, E. 1984. A Class ofDecomposable Poverty Measures. Econometrica, 52: 761-765. Ghose, A.K, 1996. Employment on Organized Manufacturing in India. The Indian Journal of labor Economics, April- June. Grootaert, C. and Kanbur, R. 1995. Child Labor. A Review. Background paper for the World Development Report. World Bank, 1995. Policy Research Working paper 1454. Hentschel, J., Lanjouw, P. and Poggi, J. 1997. Combining Survey Data with Census Data to Construct Spatially Disaggregated Poverty Maps, mimeo. Policy Research Department, the World Bank. Human Rights Watch/Asia. 1996. The Small Hands ofSlavery- Bonded Labor in India. Impact Evaluation: CARE India's Title II Program. 1994. International Labor Organization, 1996. India. Economic Reforms and Labor Policies. ILO-SAAT, New Delhi, India. Jayaraman, Raji and Lanjouw, P. 1997. Rural Living Standards in India: A Perspective from Longitudinal Village Studies Jodha, N.S. 1989. Social Science Research on Rural Change in Bardhan, P. (ed) Conversations Between Economists and Anthropologists (Delhi: Oxford University Press). Kakwani, N., and Subbarao, K. 1993. Rural Poverty in India, 1973-1987, in Micheal Lipton and J. Van der Gaag (ed.) Including the Poor. World Bank, Washington D. C. Lanjouw, P and Stem, N. (Forthcoming). Population, Outside Employment and Agricultural Change. In Lanjouw and Stem (eds). A growth of Sorts: Palanpur 1957-1993. Oxford: Oxford University Press. Lanjouw, P. and Stem, N. Forthcoming. A Growth ofSorts- Palanpur 1957-1993. New York, Oxford University Press. Loh, Jacqueline. 1995. Education and Economic Growth in India: An Aggregate Production Function Appraoch. EPDP Bureau, Ministry of Human Resource Development/Department of Education, new Delhi, Processed. Measham, A. Heaver, and Richard, A. 1996. India's Family welfare Program: Moving to a Reproductive and Child health Approach. World Bank. Washington, D.C. References 71 National Council For Applied Economic Research, 1996. Human Development Profile of India: Inter-State and Inter- Group Differentials. Vols. 1-2, processed, NCAER. National Evaluation of Integrated Child Development Services. National Institute of Public Cooperation and Child Development, 1992 0zler, Berk, Datt, G. and Ravallion, M. 1996. A Database on Poverty and Growth in India, mimeo. Policy Research Department, World Bank, Washington, D.C. Prasad, Eswara K. V. 1995. Social Security for Destitute Widows in Tamil Nadu, Economic and Political Weekly, April 15. Radakrishna and Subbarao. 1997. Targeting the Public Distribution System, mimeo. World Bank Ramachandran, V. K. 1990 Wage Labour and Unfreedom in Agriculture. Oxford: Clarendon Press. Ravallion, M. and Datt, G. 1996a. India's Checkered History in the Fight Against Poverty- Are there Lessons for the Future9 Economic and Political Weekly, Vol . XXXI. * Ravallion, M. and Datt, G. 1996b How Important to India's Poor Is the Sectoral Composition of Growth? World Bank Economic Review, 10(1): 1-26. Ravallion, M. 1994 Poverty Comparisons (Harwood Press). Ravallion, M. and Datt, G. 1995. Is targeting through a work requirement efficient? some evidence from rural India, in D. van de Walle and K. Neads (eds) Public Spending and the Poor: Theory and Evidence. Baltimore: John Hopkins University Press. Ravallion, M 1996. Poverty and Growth Lessons from 40 Years of Data on India's Poor. DEC Notes Research Findings, No. 20, September. Sen, A.K. 198 1. Poverty and Famines an Essay on Entitlement and Deprivation. Oxford: Clarendon Press. Sen, Abhijit 1996 Economic Reforms, Employment and Poverty: Trends and Options. Economic and Political Weekly, Vol. XXXI, No. 35-37. Shankar, Kripa (forthcoming). A Requiem for Land Reforms, in Ravi S. Srivastava (ed.) Land Reforms in Uttar Pradesh: An Unfinished Agenda, (New Delhi: Sage) Srivastava, R. 1997. Poverty Reduction in India: A survey of the literature, mimeo. Department of Economics, Allahabad University, India. State of the World: A Worldwatch Institute Report on Progress Toward a Sustainable Society. New York, Norton, 1997 Thakur, D.S. 1985 A Survey ofthe Literature on Rural Poverty in India Margin, April. Unni, J. 1997. Non-Agricultural Employment and Rural Livelihoods. Background paper. Weiner, M. 1996. Child Labor in India Putting Compulsory Primary Education on the Political Agenda. Economic and Political Weekly, pp. 3007-3014 World Bank. 1989. India, Poverty, Employment, and Social Services Country Economic Memorandum. Washington, D.C. 72 References World Bank, 1990. Poverty World Development Report, Washington, D.C. World Bank, 1991. Gender and Poverty in India- Issues and Opportunities Concerning Women in Indian Economy. A World Bank Country Study. Washington, D.C. World Bank, 1993. Investing in Health, World Development Report. Washington, D.C. World Bank, 1995. India: Policy and Finance Strategies for Strengthening Primary Health Care Services. Washington, D.C. World Bank, 1996. Five Years of Stabilization and Reform and the Challenges Ahead. Country Economic Memorandum. Washington, D.C. World Bank, 1997. Primary Education in India. Development in Practice Series. World Bank. Washington, D.C. World Bank, 1997a. India, New Directions in Health Sector Development at the State Level, An Operational Perspective. Washington, D.C.  ―之 e) 心才 荔 ―戶 - ―化 -O& ：焉遛 ．亡斗喜＿ ｝鑼邸 l -