_____________________I6033 Human Capital Development HCD Working Papers Agricultural Growth and Poverty in Pakistan Rashid Faruqee Kevin Carey September 1996 HCDWP 71 Papers in this series are not formal publications of the World Bank. They present preliminary and unpolished results of analysis that are circulated to encourage discussion and comment; citation and the use of such a paper should take account of its provisional character. The findings, interpretations, and conclusions expressed in this paper are entirely those of the author(s) and should not be attributed in any manner to the World Bank, to its affiliated organizations, or to members of its Board of Executive Directors or the countries they represent. Agricultural Growth and Poverty in Pakistan by Rashid Faruqee Kevin Carey Abstract The paper examines the contribution of agricultural growth to poverty reduction in Pakistan. Total growth in agriculture in Pakistan has been impressive. However, growth was held back by policy distortions, and the benefits of this growth would have been more widely distributed in a less distorted policy environment. In particular, rural employment, which has proved a key conduit from growth to poverty reduction in other countries, was severely affected by policies favoring mechanization. The policy distortions also adversely affected the income-earning opportunities of smallholders. So a structural reform program in agriculture will not only enhance the sector's efficiency and growth rate, but also increase the contribution of agricultural growth to poverty reduction. Contents Introduction .......t r o d u c t i on....,.,,., 1 The Agricultural Growth and Poverty Link-age in Pakistan .3 Impact of Factor Market Distortions on Poverty 4 Was Agricultural Machinery Taxed or Subsidized? ......................................................4 Credit Policy Towards Tractors.................................................................................6 Effect of Distortions on Factor Markets and Rural Wages.7 Trend of Real Wages: Evidence of the Impact ..............................................9 Impact of Input Provision and Price Policy on Poverty.., .. ....................................... 10 Access to Publicly Provided Agricultural Inputs ......................................... 10 Output Price Policy ................................ 13 Impact of Land Access and Land Use on Poverty . 15 Policy Implications ................................... 17 Conclusion ..................................19 References...........................20........ . 20 Introduction The last thirty years have witnessed significant agricultural growth and poverty reduction in Pakistan. For instance, the headcount measures of poverty surveyed in Malik (1988) show a decline in rural poverty rates among households from 37 percent in 1963-64 to 24 percent in 1983-84. Agriculture undoubtedly played a role in poverty reduction in Pakistan. However, the question remains whether the agricultural sector contributed as much to poverty reduction as it could have. The purpose of this paper is to probe that question. The development process across countries appears to involve productivity growth in agriculture and a transfer of labor to even more rapidly growing industrial sectors, which in turn is associated with poverty reduction (World Bank, 1993). The international evidence shows a complex interaction between agricultural growth and poverty reduction across countries (Table 1). Table 1: Agricultural Growth, Transitions, and Poverty Levels Across Countries Poverty Levels Average Share of Agricultural Agriculture in Growth GDP Period Initial Final 1970-93 1970 1993 Pakistan (1968-84) 44 24 3.5 37 25 Malaysia (1973-87) 37 14 4.1 29 19 Indonesia (1970-87) 58 17 3.6 45 19 Thailand (1962-86) 59 26 4.1 26 10 Source: World Development Report 1995, World Bank (1993) and Malik et al (1994). While Table 1 shows significant agricultural growth in all countries - growth that must have contributed to poverty reduction - it is also apparent the extent of poverty reduction from a given rate of growth differs greatly across countries. Pakistan's agricultural growth record is little different from that of East Asian countries, and yet this growth was not associated with a rapid decline in poverty levels. This indicates that the pattern of growth in agriculture (as well as in other sectors) also matters in reducing poverty. The pattern of growth involves considerations such as whether output growth was accompanied by total factor productivity growth, and the relative growth in demand for different factors. These considerations play an important role in determining real wages in agriculture, and since the major asset of the poor is their labor, the trend in real wages have an important role in poverty reduction. This paper examines more closely the role of the pattern of agricultural growth in poverty reduction in Pakistan. This is a potentially difficult exercise, because migration between rural and urban areas can be a complicating factor. Imagine a world where the agriculture sector experiences slow growth in incomes and employment and there is substantial rural-urban 2 migration. In such a situation rural poverty is presumably lower than what would have been without migration. The poor performance of agriculture in that situation may be reflected in substantially higher urban poverty. Conversely, if agriculture does well in the presence of a stagnant urban or industrial sector, it may attract a substantial pool of people waiting for an employment or cultivation opportunity, this might lead to an erroneous conclusion that agriculture has made a poor contribution to rural poverty reduction.' It is, therefore, important to look at total poverty outcomes and not just rural poverty outcomes in order to assess the role of agricultural growth. What is the aggregate picture and breakdown of poverty in Pakistan? Table 2 looks at overall, and urban and rural poverty rates in Pakistan and the provinces for three selected years. As Burki (1988) has said, 'Without a great deal of deliberate effort, Pakistan appears to have reduced the number of people living in absolute poverty by quite a significant amount" Nevertheless, it is worth noting that poverty in Pakistan is still predominantly a rural, agricultural phenomenon. Poverty rates are twice as high in rural areas, and with about two-thirds of the population in rural areas, this puts 80 percent of the poor in rural areas. We focus on the opportunities offered to the poor in agriculture and the role of increased demand for labor in poverty reduction. A general characteristic of the poor is that their major asset is their labor services, while they lack the ability to invest in physical or human capital. Thus agriculture benefits the poor to the extent that it provides them with a return on their labor and enables them to accumulate physical or human capital. One clear effect of agricultural growth is to provide some combination of higher farm incomes, higher agricultural wages, higher agricultural employment, and lower agricultural prices, all of which should help reduce poverty. This suggests the need to focus on how agricultural growth has affected agricultural employment and the economic fortunes of smallholders, especially tenants. Rates of poverty are particularly high among agricultural laborers, but poor owner- cultivators are the largest group among the rural poor (Table 3). Clearly, many of those involved in the agricultural sector are having great difficulty deriving an adequate livelihood from the sector. IA further complicating factor in Pakistan's case is that migration to the Middle East and the associated remittances have made a significant contribution to poverty reduction in Pakistan. A careful assessment of agriculture's role in poverty reduction will have to factor out the impact of emigrants' remittances on rural areas. 3 Table 2: Urban-Rural Poverty Differentials in Pakistan 1985 1988 1991 Pakistan 18.3 16.6 17.2 Urban 11.1 8.7 9.8 Rural 21.1 19.6 20.6 Balochistan 27.5 9.3 7.1 Urban 17.0 4.4 4.5 Rural 28.5 10.0 7.7 NWFP 9.6 15.5 20.2 Urban 7.5 12.4 14.3 Rural 9.0 16.0 21.4 Punjab 19.0 19.9 19.0 Urban 12.8 11.9 11.4 Rural 21.3 22.6 21.9 Sindh 15.3 9.5 12.3 Urban 7.0 3.1 6.7 Rural 22.2 14.6 17.6 Source: Malik et al (1994). Sampling error in the household survey appears to have distorted the figures from Balochistan. The urban poverty rate from Sindh also seems unusually low. The Agricultural Growth and Poverty Linkage in Pakistan Compelling evidence of a link between agricultural growth and poverty reduction is provided in the Indian context by Ravillion and Datt (1995).2 The question we want to pursue in this paper is to examine whether a similar relationship holds for Pakistan. How can the role of agricultural growth in the poverty process be established? The best approach to this question would be to relate a time series of absolute and relative poverty measures to various measures of agricultural growth. The lack of a consistent series on poverty 2 Ravillion and Dant (1995) relate poverty measures to agricultural yields and rural real wages. The yield measure captures either farmners' income or employment gains; the data does not allow these effects to be distinguished. They find that poverty responds significantly to agricultural yields in the short-run, and to yields and real wages in the long-run. A one percent increase in yields is associated with a 0.28 percent reduction in headcount poverty in the short-run, and a 1.59 percent reduction in the long-run. A I percent increase in real wages is associated with a 1 percent reduction in headcount poverty. The yield and wage effects presumably capture the ability of agricultural growth to improve the fortunes of smallholders and the landless. 4 makes this exercise impossible for Pakistan. As explained before, we can therefore, look at the access of smallholders to the sources of growth in agriculture, the trend of agricultural employment, and the access to land of the poor. In particular, we will look at agricultural growth, the labor market, and net profitability in agriculture. Our central thesis is that government policies towards agriculture have distorted factor markets. In addition, small tenant farmers face a difficult economic environment that has stagnated their income-generating opportunities. The lack of opportunity for these groups underlies the high rates of rural poverty in Table 2. In particular, we examine the role of a subsidy regime that has been monopolized by large farmers, coupled with a highly distorted land market that leaves smallholders and tenants at a disadvantage. Past policies on inputs, pricing, irrigation, and services, favored large farms at the expense of small farms. Finally, a non-functioning land market has stifled the emergence of small family farms, who also tend to be the most labor- intensive amongst all classes of farm. We now deal with each of these issues in turn. Table 3: Employment Profile of Households Heads and Poverty Incidence in Pakistan, 1991 Percent of rural sanple Incidence of Poverty Percent of all rural poor Owner-cultivators 36.6 30.2 32.0 Tenants 13.6 43.8 17.3 Agr. Laborers 7.0 56.0 11.4 Other Agr. 6.4 21.0 3.9 Source: Pakistan Integrated Household Survey Impact of Factor Market Distortions on Poverty Pakistan has undertaken extensive interventions to influence the price of agricultural machinery. As is often the case, the various policies were not always consistent with each other, and it is necessary to establish whether the policies combined resulted in a subsidy or a tax on agricultural machinery relative to the non-distorted price. Second, once this has been established, we turn the effects of the policy on the agricultural labor market. Was AgriculturalMachinery Taxed or Subsidized? Pakistan appears to have pursued two policy goals relating to agricultural machinery: (1) to encourage the emergence of a domestic tractor industry, and (2) the modernization of the agriculture sector. A host of other influences have operated on the agriculture sector over this time period, leaving a wide range of influences on the price of tractors. We summarize these in the table and elaborate on each below. 5 Table 4: Incentives for Tractor Use for Pakistan Factor Effect on Demand for Tractors Protection of Domestic Industry Labor Migration + Bulk-buying by Pakistan + Credit Distortions + Land Reform + Source: Ilahi, et al (1995) Until recently, tractors could be imported as kits duty-free and assembled in Pakistan, providing protection for domestic assemblers. Domestic manufacturers of harvesters enjoyed the protection of a substantial tariff (95 percent). More recently, the tractor kits were subject to a 10 percent duty and 15 percent sales tax, while imported tractors were subject to a 35 percent duty. While trade policy towards agricultural machinery is now in a state of flux, it appears that protection for the domestic tractor industry will remain at about the average tariff rate (29 percent), while protection for the harvester industry will be somewhat higher than this. Given Pakistan's history of exchange rate overvaluation, the assemblers were further assisted by being able to import the kits at artificially low cost. The key issue is the net effect of protection policy on prices and incentives. Many countries have chosen much higher tax and tariff rates on tractors than Pakistan, possibly to correct some of the distortions we have outlined above (Table 5). Even while allowing the domestic tractor industry some protection, the markup allowed to the industry has been officially controlled. While the markup is significant, the c.i.f. price of tractors is so low in Pakistan that the overall tractor price remains low relative to other countries. The table also illustrates the fact that the concept of a "world price" for tractors is very difficult to interpret, given such wide variation in the CIF price of tractors in different countries. Pakistan appears to obtain particularly favorable prices for tractor kits through volume discounts on bulk purchases. These bulk purchases take place by competitive bidding by large international manufacturers and results in particularly favorable prices for Pakistan. As already mentioned, mark-ups to this price once the kits are imported are strictly controlled. We would therefore argue that despite the apparent protection given to the agricultural machinery industry in Pakistan, prices in Pakistan are "low" relative to an appropriate world benchmark. Table 6 gives a time series of retail prices of selected tractors for a period of 1985-86 to 1994-95. The dollar price per horsepower show a declining real price of tractors. 6 Table 5: Tractor Prices (dollars per horsepower) in Various Countries, 1985. Country Retail c.i.f. Markup and Distribution Pakistan 136 74 62 India 175 na na Brazil 176 na na Indonesia 198 119 79 Mexico 212 na na Egypt 212 145 67 Source: Binswanger et al (1987) Credit Policy Towards Tractors. Complicating the analysis of tractor pricing is the issue of credit provision. The role of the Agricultural Development Bank of Pakistan (ADBP) is particularly important. ADBP is one of the primary credit providers to the agriculture sector, and has financed most (about 70 percent) of the tractor acquisition. Distortions have occurred in two ways: loans were typically at below market rates of interest, and loans are often not repaid. In 1992-93, tractor loans carried an interest rate of 9 percent, versus a market rate of 19 percent. Given average (and presumably anticipated) inflation in Pakistan in recent years of about 10 percent per annum, this works out at a zero or negative real interest rate. In addition, a discounted price for tractors is available to those who purchase using an ADBP loan. Whereas the down payment on the loan would normally be proportional to farm size, tractor loans require a flat down payment of 10 percent. Finally, the term of the loan is extended from 8 years to 10 years. For 1992-93, Ilahi et al (1995) estimate the credit subsidy on tractors as 29 percent of the purchase price. Non-repayment of tractor loans compounds the problem. About 40 percent of ADBP loans remain unrecovered, and a large proportion of ADBP loans are for tractor purchases. While the Awami Tractor Scheme (ATS) has done away with some of the explicit subsidies, non- repayment of loans is likely to remain a significant problem. Given that non-repayment has been a problem for such a long time, and that enforcement problems are so well-known, it seems reasonable to assume that non-repayment is in fact anticipated so that the true ex ante price is indeed lowered by these distortions. The credit concessions have offset the rise in tractor prices caused by trade policy: the size of the discount due to credit policy illustrated in Ilahi et al (a 52 percent subsidy on the purchase price with partial repayment of loans) is much larger than the rates of protection on tractors (which as we argue above likely overstate Pakistan's protection relative to the 'World price'). Since the trade distortions are being reduced, while the credit distortions remain in place, the effective price of tractors may actually fall. 7 Table 6: Tractor Prices (1990/91 US dollars per horsepower) MF-240 Ford-3610 Friat 480 BYLA.MTZ-50 1985/86 201.79 205.12 222.96 175.80 1986/87 215.72 217.59 242.52 196.88 1987/88 219.42 210.89 236.50 185.18 1988/89 195.89 176.24 217.67 173.53 1989/90 188.54 147.07 217.16 175.67 1990/91 170.32 125.89 207.34 160.94 1991/92 154.78 180.20 152.46 1992/93 148.82 185.38 144.59 1993/94 121.96 138.76 123.45 1994/95 119.96 127.91 109.05 Growth Rate -5.35 -8.88 -5.51 -4.90 Source: World Bank Sources. Effect of Distortions on Factor Markets and Rural Wages Suppose it was the case that there was a net subsidy to agricultural machinery. What will be the effect on labor markets? The analysis for a two-factor open economy (Neary, 1976) provides a range of answers depending on the assumptions one makes. Consider a two-sector world where capital in one sector receives a subsidy. In a Hecksher-Ohlin world with commodity prices determined at the world level, employment in the subsidized sector must increase. The subsidized sector will also experience a rise in the capital labor ratio, as long as it is relatively labor intensive. The results do not change significantly with assumptions about factor mobility once we retain the small country assumption. However, if the subsidized sector faces a downward sloping demand curve, the price of the product will fall with expansion of the sector, reducing the value marginal product of labor and possibly employment. Unfortunately, results are less clear-cut when we allow for variable factor supplies in this framework. The analysis is further complicated by the nature of the output price regime that farmers in Pakistan faced. Given the extensive intervention in output markets, it cannot be said that farmers faced an exogenous world price or a freely functioning domestic market. Presumably one motivation for the machinery policy was to compensate for the effect of output taxes. Most important (and least easy to quantify) are the general equilibrium effects of a farm machinery subsidy in a multi-sector economy. Under reasonable conditions, subsidized mechanization of farms deprives non-farm sectors of capital resources. Suppose now that 8 subsidized mechanization leads to a expansion in output in large farms. This will generate more employment on those farms, as in the partial equilibrium analysis above. However, if the large farm sector is more capital intensive than the average non-agricultural sector, overall labor demand in the economy will decline. The small farm sector will have to absorb more labor, worsening the position of that sector. The key point is that mechanization could be labor-using within the agriculture sector while being labor-replacing for the economy as a whole. Terms of trade effects between sectors are also important. The expansion of the large farm sector could lower agricultural prices and hurt smallholders. There is a distributional element to the credit distortions that worsens the price distortion. Despite government quotas for lending to small farms (70 percent of credit disbursed is supposed to go to farms smaller than 5 hectares), smallholders have little participation in the institutional credit system (Table 5). Smallholders tend not to have passbooks (a basic requirement for a mortgage), and they rely almost exclusively on non-institutional credit sources for their credit needs.3 ADBP is the primary institutional credit provider, and its portfolio illustrates how its lending has been weighted towards larger farms. Loans larger than Rs 100,000 account for over 60 percent of total ADBP lending, and about 34 percent of ADBP credit goes to farms bigger than 50 acres. The 1984 Census of Agricultural Machinery revealed that 38 percent of the tractor purchases on farms over 50 acres had been financed by the ADBP, while only 24 percent of tractor purchases on farms less than 12.5 acres in size had been financed by the ADBP. Smaller farmers presumably had to rely on informal credit to finance these purchases, thus not benefiting from the extensive subsidies available on formal sector loans. Commercial banks claim to reach a larger proportion of smallholders than the ADBP, but the practice of proxy loans (commercial bank loans are also subsidized) makes this claim unreliable. Cooperative credit has been open to significant abuse for a long time and has not helped smallholders either. While small farms would undoubtedly fare worse than large farms even in a market driven credit system, the system of subsidies has created an incentive for rent- seeking - at which large farmers are clearly more adept. 3It should be noted that detecting a relationship between land size and another variable in the data is clouded by the fact empirical land size distributions do not allow for land quality differences. 9 Table 7: Access to Credit by Size of Farm Owned, 1985. Size Class Percent of Households Percent of Total Loans in Last Year (hectares) with Passbook Institutional Non-institutional Under 0.5 0 2 98 0.5to 1.0 1 7 93 1.0 to 2.0 3 13 87 2.0to3.0 7 24 76 3.0 to 5.0 12 31 69 5.0 to 10.0 19 46 54 10.0 to 20.0 31 61 39 20.0 to 60.0 38 64 36 over 60.0 51 58 42 All Farms 9 36 64 Source: Rural Credit Survey, 1985. The failure of tractors to generate employment is in stark contrast to the earlier experience with tubewells (World Bank, 1989). The emergence of tubewells gave rise to a significant small- scale engineering industry to produce, service, and repair tubewell machinery and equipment. By contrast, growing tractorization did not have these effects, in addition to its failure to increase yields. Large imported tractors did not lend themselves to a local small-scale maintenance industry in the same manner as tubewells. In addition, the technology associated with modem tractors presupposed literacy on the part of those working on maintaining them, which tended to exclude the rural population. Finally, the spread of tractors is dependent on rural infrastructure. In particular, electricity is a crucial input to the tractor maintenance industry, so Pakistan's well- known deficiencies in this area have constrained the employment generating effects of tractors. Thailand's experience with mechanization is quite different to that of Pakistan's. In the former case, the agricultural mechanization process was demand-driven, and did not depend on government subsidies and assistance (World Bank, 1983). In the mid 1960s, arable land expansion in Thailand was at the rate of 5 percent per year, while labor force growth was slowing. The demand for labor was increasing throughout the country, so the additional land could only be cultivated by increased mechanization. Pakistan's policies favoring capital are particularly perverse given that it is a labor-abundant country and thus its comparative advantage should lie in labor-intensive production. Taiwan and Korea responded to a labor-abundant resource endowment by pursuing a labor-intensive pattern of rural development involving land reforms, infrastructure development, rural services, and agricultural diversification. Trend of Real Wages: Evidence of the Impact Some substitution of capital for labor would have taken place in the absence of policy distortions. This is because real wages have risen since the mid 1960's. Between 1966 and 10 1991, urban real wages grew by 65 percent, while rural real wages grew by 76 percent. Rural- urban migration and migration to the Middle East reduced the supply of agricultural labor. The point to be emphasized is that the pattern of sectoral labor movements seems to be not what would one would have expected in the absence of policy distortions (recall the lack of an agricultural transition in Table 1).4 More direct evidence of the general equilibrium effects of capital subsidies would require data on wages of different classes of workers. The best we can do is compare trends in the wages of casual and permanent agricultural workers from 1984 on.5 Between 1984 and 1992, wages of casual workers increased 17 percent, while wages of permanent workers increased 29 percent. These trends are suggestive of a small pool of workers on capital intensive farms experiencing rapid wage growth, with a much larger pool of workers competing on a casualized labor market and experiencing slower wage growth. It should also be noted that Pakistan's industrial sector is also generally considered to have excessive capital-output and capital-labor ratios. The ability of the industrial sector to absorb surplus labor from the rural sector is thus also very limited, putting additional pressure on labor intensive-sectors. This is again a case where migration to the Middle East has played a crucial role in protecting Pakistan from the worst consequences of its economic policies. Impact of Input Provision and Price Policy on Poverty Access to Publicly Provided Agricultural Inputs. The access of small and marginal farmers to publicly provided inputs is an important determinant of their ability to enhance their return from agricultural activities. Most of the past sources of growth (irrigated land, mechanization, and modern inputs) were all areas where the poor had less access than the rich. In addition, price policy and taxation policy tended to penalize the poor more than the rich. Pakistan has pursued policies similar to other South Asian countries in providing heavy subsidies for irrigation, electricity, credit, and modern chemical inputs. Almost inevitably, subsidized items must be rationed, and large farmers have an inherent advantage in obtaining rationed products. This basic principle underlies our analysis of all the subsidy programs. Since water is a necessary input to agricultural production in Pakistan, it is useful to look at the distribution of irrigated land. Ahmad and Sampath (1994) estimate the magnitude of inequality in the distribution of irrigated area from three Censuses (1960, 1972, 1982) using 4 Another serious constraint on rural labor supply--although not an effect of distorted incentive policy, but an outcome of human resource development programs--relates to gender (World Bank, 1989). While many women participate in the agricultural labor force, they are generally constrained to work on their own holdings, and are not available to the agriculture sector as a whole. Since much of this employment is likely to be seasonal, these women face the problems of underemployment and low-productivity employment. These characteristics in turn are likely to be correlated with the incidence of poverty. Past employment policy is perhaps guilty of failing to address this serious under-employment problem in the rural sector, but the problem may also reflect traditional attitudes to work in rural Pakistan. 5Data from the Statistical Bulletin, Govermment of Pakistan, cited in Ilahi et al (1995). 11 district level data categorized into nine farm-size groups. A striking finding is that when one compares the distribution of total land with the distribution of total irrigated land, the level of inequality in the distribution of irrigated land is much higher. As they point out, the key difference between the distribution of irrigated land and the distribution of total land is that the distribution of irrigated land 'is determined to a great extent by the irrigation water distribution policy of the provincial governments" (p6O). The conclusion is that irrigation policy has aggravated the level of inequality in the distribution of land. Ahmad and Sampath are also able to decompose the level of inequality in irrigation into within-district and between-district inequality, and they find that within-district is the dominant factor in explaining total irrigation inequality. This means that most of the inequality is not explained by differences in the level of irrigation development across districts. It is rather irrigation distribution policy at the district level that is mostly responsible for the level of inequality - a finding that complements their first conclusion above. In addition, these findings does not take into account the many abuses that occur in the actual distribution of water at the field level. Illegal pumping from canals is widespread, and in practice, local water resources are often controlled by a small number of politically powerfiil farmers. The fundamental problem is that irrigation water is provided to farmers at far below its economic cost or its financial value to farmers. Current water charges do not meet operations and maintenance requirements on the irrigation system, and water charges of Rs 25 per acre foot in the field are well below financial marginal values of water of Rs 700 per acre foot, and private tubewell prices of Rs 100-400 per acre foot (World Bank, 1994). Water therefore must be rationed by some non-price means. Obtaining direct evidence on the access of farmers to inputs by size is difficult, but some suggestive evidence is available. A study of the factors determining whether farmers knew the correct location of a seed depot and whether they visited the depot6 found that literacy and extension contact were the key explanatory variables in both cases (Heisey et al, 1990). Once these variables were controlled for, size was an insignificant determinant of the seed access variables. Since other studies had found that farm size was generally positively correlated with seed access, the implication is that larger farmers have higher levels of extension contact (and literacy), and that this is the source of their superior access to seeds. Large farmers are more likely to use high quality seed because they are more familiar with the provision of services, not because of any inherent characteristic of larger farms. Table 8 shows how literacy and farm size interact to give large farmers much greater contact with the extension system. 6Obtaining seed from a depot is an important productivity indicator because such seed will be of higher quality than seed obtained from previous plantings or from other farmers. 12 Table 8: Relation Between Extension Contact, Education, and Farm Size, Punjab 1986 Fann Size Percentage of farmers: Less than 5 ha 5 to 10 ha More than lOha with contact in last year 13 24 61 literate 32 50 71 with secondary schooling 14 25 32 Source: Byerlee (1987). Fertilizer subsidies and state marketing of fertilizer were important components of agriculture policy in Pakistan until recently. As with any subsidized input, smallholders are likely to fare worse than large farmers in obtaining access to it, and use of fertilizer remains low amongst smaller farmers (Table 9). Coady (1995) presents evidence that access to fertilizer is strongly and significantly related to farm size in a sample of farmers from Punjab and relates this to characteristics of farmers. The problem for smallholders was partly their lack of access to credit - a problem which we have already documented above.7 Smallholders were also hampered by their lack of knowledge of best practice farming techniques, which corroborates the evidence in Table 8. Table 9: Use of Fertilizers and Insecticides By Size of Farm, 1990. Size Class Percent of Farms using Percent of Farms Using (hectares) Fertilizer Insecticides Under 0.5 33 13 0.5to 1.0 46 20 .O to 2.0 52 24 2.0 to 3.0 56 28 3.0 to 5.0 55 30 5.0 to 10.0 50 32 10.0 to 20.0 44 33 20.0 to 60.0 41 34 over 60.0 40 41 All Farms 49 25 Source: Census of Agriculture, 1990 7 Intriguingly, Coady also finds that when one looks at farmers who do have access to fertilizer, households with more land apply lower levels of fertilizer per acre. The positive effect on fertilizer use of having more assets, larger households, and better agricultural practices is not enough to overcome the effect of increasing relative risk aversion on fertilizer use. 13 Since the provision of all these inputs is weighted towards larger farmers, the poor are penalized in two ways. First, their position relative to large farmers clearly suffers. Second, as Srinivasan (1993) points out, to the extent that inputs are financed by subsidies out of general government expenditure, they crowd out other types of government expenditure that might help the poor, such as spending on rural primary education and infrastructure. Pakistan has indeed been deficient in the provision of education and infrastructure which are particularly important to the social mobility of the poor. The contrast with Indonesia's approach to rural development is instructive. In the 1970s, Indonesia used its mineral windfall to develop social and physical infrastructure, and an important component of this was the INPRES expenditure program. This was a revenue-sharing scheme between levels of government designed to fund infrastructure projects such as roads, schools, and health facilities as well as their operations and maintenance. The program funded many small-scale infrastructure projects, providing substantial employment for unskilled workers in rural areas - amounting to 1.5 million man-years of work in 1982. In general, East Asia has been characterized by strong infrastructural and service support for small-scale agriculture (World Bank, 1993). Output Price Policy The massive distortions created by output policy have reduced agricultural growth and thus have directly hurt the poor. In addition, a recent study (Ali and de Kruijk, 1994) found that the net contributors to the income transfers resulting from price policy were farmers in especially poor rural areas, such as Southern Punjab. This region faced the highest implicit tax rates under the government's marketing arrangements. Net recipients included people living in urban areas and other relatively richer zones. Producers of cotton and wheat experienced the largest income transfers. Smallholders account for a small share of total cultivated land; since the transfers are based on the size of farm (via quantity produced), more of the transfers accrue to larger farmers. An additional consideration is that smaller farmers crop mix further reduced their gains from price policy. It is difficult to make a fully convincing case for this assertion, since the crop mix is an endogenous (and perhaps poorly measured) variable. However there is some evidence of regressive transfers in the case of sugarcane. In a study of panel data from 1986-88, Adams (1995) finds that when households are placed in quintiles based on average income over the three years, households in the top quintile received 36 percent of their income from sugarcane, while households in the lowest quintile received 14 percent of their income from sugarcane. This reflected a difference in the proportion of harvested area devoted to sugarcane between the quintiles: 6.7 percent for the lowest quintile, and 14.8 percent for the highest quintile.8 In addition, the marketing arrangements for crops are surely an important determinant of returns; it is likely that large farmers can obtain better terms from the sugar mills than smallholders. In addition, since sugarcane is a water-intensive crop, farmers with superior access to water have a comparative advantage in its production, and large farmers have better access to water than smallholders. The many inefficiencies of the marketing boards typically fall on farmers, since the board is protected by its monopoly position. s Standard errors are not reported. The sample size is 680. 14 More generally, by holding down the prices of rice and wheat, price policy was hurting the poorer farmers to the extent that they were net suppliers of wheat and rice. Price policy would help the poor to the extent that they are net consumers of foodcrops, but in the case of wheat (the dominant food crop), the benefits of the wheat subsidy have been absorbed by flour millers and have not been passed on to consumers (Ryan and Khan, 1993). A possible offsetting consideration is that, by controlling the price of wheat, pricing policy may have allowed net wheat producers to capture more of the benefits of technical change in wheat production than if wheat was a freely traded commodity. However, more of this gain would accrue to the more technically advanced farmers and the bigger producers, who are more likely to be the large farmers for reasons we outlined above (in addition to the fact that large farmers by definition produce more wheat). In addition, if the pattern of government expenditures is anti-poor, then, by funding such expenditures, state marketing schemes worsen the relative position of the poor. A rare exception to extensive state marketing is provided by Thailand, which has allowed private middlemen to conduct marketing and processing operations in the agriculture sector (Myint, 1985) The result has been rapid export expansion in rice and other products. Even where farmers gain from government interventions (such as payments above the world price for crops, or subsidized irrigation systems), these interventions are rarely an effective means of reducing poverty. This is for the well-known reason that the distribution of such transfers is proportionately related to farm size, whereas poverty is inversely related to farm size. Thus the amount of subsidies that the poor receive, even if large relative to their total income, is very small. A poor farmer with exactly the same yield on a subsidized crop as a rich farmer will obtain far less of the total subsidy amount. One gauge of the effect of pricing policy on rural households is provided by trends in real per hectare profits and real household income (Tables 9 and 10). Table 9 shows the annual rate of growth between 1965 and 1987 in profits for hectare for small and large farms in irrigated and rainfed areas in Punjab.9 Table 10 shows, for the same period and a sub-period, growth in real income per household for the same groups plus the landless in both areas. Table 10: Trend Growth in Real Per Hectare Profits (constant 1980s Rs) from crop production in Punjab, 1965-87 (percent per year) Rainfed Areas Irrigated Areas Small Farmns Large Farms Small Farms Large Farms -2.7 2.7 -0.4 2.7 Source: Renkow (1991). 9While of course profit calculations are complicated by the presence of family labor, likely more important for small farms, it is not clear that such issues would lead the rate of growth of farm profitability of small farms to be understated relative to large farms. 15 Table 11: Trend Growth (percent per year) in Real Income (constant 1980s Rs) per Household for Rural Punjab, 1965-87 and 1976-87 Rainfed Areas Irrigated Areas Landless Small Large Landless Small Large Farms Farms Farms Farms 1965-87 6.6 4.4 0.8 4.6 -1.7 1.7 1976-87 5.1 14.9 10.4 -1.7 -5.3 1.0 Source: Renkow (1991). Both tables demonstrate that small irrigated farms have been squeezed relative to other classes of farm. While small rainfed farms suffered comparable declines in profit per hectare (Table 10), farms and the landless in rainfed areas experienced strong growth in household income. This is because these households have more non-farm income, notably remittances. The combination of low non-farm income and developments within agriculture resulted in small irrigated farms faring worse than any other group over the period (Table 11). This diversification by rainfed households has led to some improvement in rural income distribution in Pakistan, but this cannot be attributed to the agriculture sector. In fact, as Renkow says, 'it appears that the cumulative effect of the matrix of government agricultural policies has been to impede the ability of farm households to improve their well-being". Impact of Land Access and Land Use on Poverty Failures in the land market have interacted with credit market distortions to worsen the position of smallholders and the landless. Numerous attempts at land and tenancy reform have created an air of uncertainty, leading to a preference for self-cultivation with limited hired labor over tenancy. Land has also been used as a tax shelter by the wealthy. Small farms are unable to increase the size of their holding or invest in their land, because the supply of land for rental is limited by the reluctance of large farmers to rent, and they lack sufficient collateral to obtain credit. These constraints apply even more severely to the landless, who must choose between being agricultural laborers in perpetuity or exiting the sector completely. An indication of the difficulties facing smallholders is that the distribution of operated land area in Pakistan has worsened since 1970. The gini coefficient for operated area (calculated from Census data) was 0.52 in 1970, and 0.58 in 1991. Various government subsidies have tended to increase the profitability of large farms artificially, such as credit and irrigation subsidies, and preferential tax-treatment for agriculture. As we outlined earlier, large farmers in practice are able to obtain highly valuable water at the expense of smallholders. This pattern of rent-seeking at the expense of small farms is consistent with the experience of agriculture sectors in Southern Africa, as outlined in Deininger and Binswanger (1995). The emergence of a demand for casual labor is partly a product of policy-influenced developments in the land market. Self-cultivated capital intensive farms will only have a demand for outside labor at peak times of the year. Thus the demand for casual labor increases, while the 16 number of tenants (evicted from large farms) declines. The decline of tenancy has been particularly pronounced in Punjab, while tenancy persists on about 70 percent of cropped area in Sindh. It is important to note that some decline in tenancy and casualization of the labor force would take place even in the absence of policy distortions. The dynamics of the land market are determined by the interplay between two offsetting determinants of optimal farm size: smaller farms are more able to monitor and supervise labor supply, while larger farms are more able to overcome credit constraints. Technical change in agriculture has raised the return to investment in land, and large farms are able to reap these returns through their ability to overcome credit market imperfections. The labor supervision problem is mitigated through the use of capital rather than labor. Rural wages have risen because rural-urban mnigration and migration to the Middle East reduced rural labor supply. We would thus see the emergence of capitalized farms at the expense of tenants and hired labor even in a liberalized environment. In Pakistan's case, with a peak labor shortage compounded by the fact that many women are 'locked into" labor supply on their own farms only, the market solution is likely to be an operational size holding which can be cultivated by the family. An active land market is essential in facilitating this size adjustment, but as we have noted, the land market does not appear to be matching the supply of land with those who have labor to provide-the poor. Thailand, by contrast, seems to deal with the peak labor shortage issue in a more efficient fashion. One study (World Bank, 1983) estimated that virtually all of the seasonal increase in employment in Thailand's agriculture sector was accounted for by unpaid family workers, who enter and exit the labor force in conjunction with labor requirements in agriculture. Thus rural labor supply in Thailand is closely matched to the demand for agricultural labor-a match that is facilitated when land is operated on a family farm basis. Tax policy has also distorted the land market. With land essentially exempt from income and land taxation until recently, land was often held as a tax-shelter rather than for productive agricultural purposes. This raises the value of land above the present value of the proceeds from use of the land for agricultural production, and so renders it inaccessible to farmers with little or no equity finance. In addition, machinery deductions were allowed in the wealth tax, creating an additional incentive for large farms to purchase machinery. The fact that poorer farmers cultivate more unirrigated land relates to our earlier findings about agricultural services. Research breakthroughs, including the Green Revolution package, have been concentrated on irrigated land. In Pakistan, as in other countries, research issues on unirrigated land have been neglected. Since the poor are more dependent on rainfed land, this would appear to provide an additional tilt of policy against poor farmers. However, matters are somewhat more complicated than this. In general equilibrium, the distributional impact of technical change depends on how it is mediated through goods and factor markets, in addition to its direct impact on the beneficiaries. Net consumers of a commodity in the rainfed areas gain from research improvements that lower the price of that commodity, even if there is no breakthrough in rainfed production. However, net producers in the rainfed area lose because their production technology has not improved but they receive a lower price. Since these effects 17 operate through the price of the commodity, government price policy will inhibit some or all of the above effects from taking place. If technical change is labor using, then the rainfed areas also gain if there is migration to the irrigated areas. Migrants receive higher wages as a result of the increased demand for labor, and wages in the rainfed areas rise because the supply of labor there is reduced. Such rural-rural migration was an important factor in the Indian Punjab. Of course, the forces we have outlined above which reduced the demand for labor relative to what otherwise might have been will interfere with this process. Other events also complicate the analysis of labor market effects. For instance, income sources in rural rainfed areas tend to be more diversified than in irrigated areas, precisely because those in rainfed areas seek outside opportunities to compensate for the lack of opportunities in agriculture. In Pakistan's case, one consequence of the lack of attention to rainfed areas is that these were major participants in the Middle East migration boom, leading to very rapid income growth in the rainfed areas. Of course this is an exogenous event that surely cannot have been part of the original decision to focus research efforts on the rainfed areas. The technical change therefore has a direct impact on producer costs, and indirect impact via goods and labor markets. A general equilibrium exercise analyzing the net impact of these effects is undertaken by Renkow (1991). Some key features of households in Pakistan underlie the results. Large farmn households in rainfed areas and small and large farm households in irrigated areas are net wheat producers. Small farm households in rainfed areas, the landless, and urban consumers are net consumers. The latter group benefit from technical change in wheat production as long as it lowers the price of wheat. To the extent that government controls (or free trade) prevent the price of wheat from falling as the supply curve shifts outwards, then net producers gain from technical change. In this sense, it cannot be said that research improvements concentrated on irrigated areas are unambiguously 'anti-poor" However, since farm income is a larger proportion of total income for households on large irrigated farms than for other groups, profit-enhancing technical change will benefit this group proportionately more than other groups. In the full simulation, Renkow finds that technical change in irrigated areas does not have undesirable equity consequences, because it helps small irrigated farmers without hurting the rainfed areas. However, he points out that the rainfed areas are not hurt by a bias in research efforts because they have diversified their incomes in response to past lack of opportunity in agriculture. If their ability to diversify was weakened (e.g. by a decline in migration to the Middle East), then a redirection of research efforts towards the rainfed areas would become more important. Notably, rural-rural migration has not played an important role in diffusing benefits of technical change in agriculture to the rainfed sector. In other words, agricultural growth has played little role in reducing inter-regional income differences through the labor market. Policy Implications Pakistan is hardly unique in its pursuit of policies to promote agricultural growth without taking into account poverty alleviation through agricultural growth. As explained in the 1995 World Development Report, many countries in Asia and Latin America pursued similar policies, with similarly anti poverty consequences. In Colombia, as in Pakistan, a satisfactory rate of 18 agricultural growth did not translate into significant employment growth, as large farmers preferred to cultivate their land through mechanization and ranching, rather than rent out to tenants. Whereas Pakistan had the safety valve of Middle East migration, the effect of these policies in Colombia was significant rural violence and environmental degradation. A clear implication of the analysis is that market distortions in Pakistan have favored rich farmers over poor farmers. To enhance the poverty-reducing effect of growth (as well as to realize fully the potential of sustainable growth), these distortions should be eliminated. Pakistan is currently heading in this direction. As we have argued, a key influence on the price of machinery is subsidized credit. The continuing interest subsidy on tractor loans makes little economic sense; interest rates on tractor loans (and indeed most agricultural loans) should be moved to market rates. The other concessions available to tractor purchasers (the preferential price, term extension, and down payment reduction) should also be removed. Equally significant is the implicit subsidy provided by non-repayment of tractor loans. This subsidy approximates one-third of the total price of machinery, and is heavily weighted towards large farmers because of their access to official credit and their higher default rate conditional on having official credit. The legal framework for repayment enforcement must be improved. The entire system of targeted credit for agriculture will have to be rethought, as such a system is almost inevitably open to abuse. Reform in the irrigation sector will play an important role in enhancing the position of smallholders. A move towards a market in water will improve the access of smallholders, as allocation of water will be based more on its value in agricultural use, and less on political power in the district. As things stand now, unproductive land use is sustained by the extremely low water charges-and even at their current level, evasion of water charges by some farmers is substantial. Trade and taxation policy have also played a role in policy distortions. Price policy undercut the competitiveness of smallholders, and taxation policy led to distortions in land use that did not favor the tenants and smallholders. So trade and price liberalization will enhance opportunity for smallholders and tax reforms will generally improve efficiency of the sector and favor the poor. Most importantly, land market reform will have a strongly favorable impact on poverty. The past approach of forced redistribution has clearly been counterproductive. In addition, any land redistribution by edict can be quickly undone by the market if the problem of smallholders' limited access to capital markets is not resolved. Carter and Mesbah (1993) suggests that a land mortgage bank, which would offer concessionary finance for long-term capital investment in land, is a promising route for market-based land reform. The interest rate on such loans can be set so that smallholders will self-select as borrowers from the bank. The key insight is that since smallholders are more rationed than large farmers in their access to credit, their shadow price of credit is much higher, and so they are willing to pay a higher interest rate on loans than large farmers. The substantial borrowing by smallholders from the informal sector is an illustration of this fact. 19 Conclusion Our argument can be summarized by the following key points. First, total growth in agriculture in Pakistan has been impressive. However, growth was held back by policy distortions, and the benefits of this growth would have been more widely distributed in a less distorted policy environment. In particular, rural employment, which has proved a key conduit from growth to poverty reduction in other countries, was severely affected by policies favoring mechanization. The policy distortions in the past also damaged the income-earning opportunities of smalUholders. Structural and policy reforms in agriculture will not only increase the efficiency and growth rate of the sector, but also enhance the contribution of agricultural growth to poverty reduction in Pakistan. References Adams, Richard H. Jr., 1995. "Agricultural Income, Cash Crops, and Inequality in Rural Pakistan", Economic Development and Cultural Change. 43 (3): 467-491. Ahmad, Manzoor, and Rajan K. Sampath, 1994. "Irrigation Inequalities in Pakistan, 1960-1980: A District Level Analysis," Pakistan Development Review 33 (Spring): 53-74. Ali, Mubarak, and Lourdes E. Velasco, 1993. "Intensification Induced Resource Degradation: The Crop Production Sector in Pakistan", International Rice Research Institute, Social Sciences Division. Ali, Shaukat, and Hans de Kruijk, 1994. Incidence of Taxes and Transfers in Pakistan. Ministry of Finance of Pakistan and Erasmus University, mimeo. Barham, Bradford, Michael R. Carter, and Wayne Sigelko, 1995. "Agro-export production and peasant land access: Examining the dynamic between adoption and accumulation", Journal of Development Economics. 46: 85-107. Binswanger, Hans P., 1987. Agricultural Mechanization: Issues and Options. Washington: World Bank. Birdsall, Nancy, David Ross, and Richard Sabot, 1995. "Inequality and Growth Reconsidered: Lessons from East Asia", The World Bank Economic Review. 9 (3): 477-508. Burki, S.J., 1988. "Poverty in Pakistan: Myth or Reality?", in T.N. Srinivasan and P.K. Bardhan, eds., Rural Poverty in South Asia, New York: Columbia University Press. Byerlee, D., 1987. Maintaining the Momentum in Post Green Revolution Agriculture: A Micro- Level Perspective from Asia, International Development Paper No. 10, Michigan State University. Carter, Michael R., and Dina Mesbah, 1993. "Can Land Market Reform Mitigate the Exclusionary Aspects of Rapid Agro-Export Growth", WorldDevelopment. 21 (7): 1085- 1100. Coady, David P, 1995. "An Empirical Analysis of Fertilizer Use in Pakistan", Economica, 62, 213-234. Deininger, Klaus, and Hans P. Binswanger, 1995. "Rent Seeking and the Development of Large- Scale Agriculture in Kenya, South Africa, and Zimbabwe", Economic Development and Cultural Change. 43 (3). 21 Heisey, Paul W. (editor), 1990. Accelerating the Transfer of Wheat Breeding Gains to Farmers: A Study of the Dynamics of Varietal Replacement in Pakistan. CIMMYT Research Report No. 1, Mexico D.F., Mexico. Ilahi, Nadeem, Zareen Naqvi, Anjum Nasim, and Aasya Akhlaque, 1995. 'Policy Induced Distortions in the Demand for Agricultural and Industrial Labor in Pakistan', background paper for World Bank Poverty Assessment of Pakistan. Washington: World Bank. mimeo. Malik, Sohail J., Safiya Aftab, and Nargis Sultana, 1994. Pakistan's Economic Performance 1947 to 1993: A Descriptive Analysis, Sure Publishers: Lahore. Myint, Il-a, 1985. 'Growth Policies and Income Distribution", Development Policy Issues Series, Economics and Research Department. Washington: World Bank. Neary, J. Peter, 1978. "Capital Subsidies and Employment in an Open Economy," Oxford Economic Papers, 30 (3): 334-56. Renkow, M., 1991. Modeling the Aggregate Effects of Technological Change on Income Distribution in Pakistan's Favored and Marginal Production Environments. CIMMYT Economics Paper No. 4. Mexico, D.F.: CIMMYT. Ryan, Joe, and Shaheen Rafi Khan, 1993. 'The Impact of the Government of Pakistan's Wheat Pricing Policy in Flour Prices", The Pakistan Development Review, 32:4 Part II: 823-832. Ravallion, Martin, and Gaurav Datt, 1995. 'Growth and Poverty in Rural India" Policy Research Working Paper 1405. Washington: World Bank. Srinivasan, T.N., 1993. "Poverty Alleviation in South Asia". Working Paper, Yale University. World Bank, 1983. Thailand, Rural Growth and Employment. Washington: World Bank. World Bank, 1989. Pakistan Employment Issues and Prospects. South Asia Regional Office. World Bank, 1993. The East Asian Miracle: Economic Growth and Public Policy. Oxford University Press. World Bank, 1994. Pakistan Irrigation and Drainage: Issues and Options. Agriculture and Natural Resources Division, South Asia Region. Human Capital Development Working Paper Series Contact for Title Author Date paper HROWP40 Integrated Early Child Mary Eming Young October 1994 0. Shoffner Development: Challenges and 37023 Opportunities HROWP41 Labor Market Insurance and Deepak Lal October 1994 M. Espinosa Social Safety Nets 37599 HROWP42 Institutional Development in Alberto de Capitani October 1994 S. Howard Third World Countries: The Douglass C. North 30877 Role of the World Bank HROWP43 Public and Private Secondary Marlaine E. Lockheed November 1994 M. Verbeeck Schools in Developing Emmanuel Jimenez 34821 Countries HROWP44 Integrated Approaches to T. Paul Schultz November 1994 M. Espinosa Human Resource Development 37599 HROWP45 The Costs of Discrimination in Harry Anthony Patrinos November 1994 I. Conachy Latin America 33669 HROWP46 Physician Behavioral Nguyen X. Nguyen December 1994 M. Espinosa Response to Price Control 37599 HROWP47 Evaluation of Integrated T. Paul Schultz January 1995 M. Espinosa Human Resource Programs 37599 HROWP48 Cost-Effectiveness and Health Philip Musgrove January 1995 0. Shoffner Sector Reform 37023 HROWP49 Egypt: Recent Changes in Susan H. Cochrane February 1995 0. Shoffner Population Growth Ernest E. Massiah 37023 HROWP50 Literacy and Primary Kowsar P. Chowdhury February 1995 M. Espinosa Education 37599 HROWP51 Incentives and Provider Howard Barnum March 1995 0. Shoffner Payment Methods Joseph Kutzin 37023 Helen Saxenian HROWP52 Human Capital and Poverty Gary S. Becker March 1995 M. Espinosa Alleviation 37599 HROWP53 Technology, Development, and Carl Dahlman April 1995 M. Espinosa the Role of the World Bank 37599 HROWP54 International Migration: Sharon Stanton Russell May 1995 0. Shoffner Implications for the World 37023 Bank HROWP55 Swimming Against the Tide: Nancy Birdsall May 1995 A. Colbert Strategies for Improving Equity Robert Hecht 34479 in Health HROWP56 Child Labor: Issues, Causes Faraaz Siddiqi June 1995 I Conachy and Interventions Harry Anthony Patrinos 33669 HCOWP57 A Successful Approach to Roberto Gonzales July 1995 K. Schrader Partcipation: The World Bank's Cofino 82736 Relationship with South Africa Human Capital Development Working Paper Series Contact for Title Author Date paper HCOWP58 Protecting the Poor During K. Subbarao July 1995 K. Labrie Adjustment and Transitions Jeanine Braithwaite 31001 Jyotsna Jalan HCOWP59 Mismatch of Need, Demand Philip Musgrove August 1995 Y. Attkins and Supply of Services: 35558 Picturing Different Ways Health Systems can go Wrong HCOWP60 An Incomplete Educational Armando Montenegro August 1995 M. Bennet Reform: The Case of 80086 Colombia HCOWP61 Education with and with out the Edwin G. West September, 1995 M. Espinosa State. 37599 HCOWP62 Interactive Technology and Michael Crawford October 1995 P. Warrick Electronic Networks in Higher Thomas Eisemon 34181 Education and Research: Lauritz Holm-Nielsen Issues & Innovations HCOWP63 The Profitability of Investment George Psacharopoulos December 1995 M. Espinosa in Education: Concepts and 37599 Methods HCDWP64 Education Vouchers in Practice Edwin G. West February 1996 M. Espinosa and Principle: A World Survey 37599 HCDWP65 Is There a Case for Antonio Zabalza March 1996 M. Espinosa Government Intervention in 37599 Training? HCDWP66 Voucher Program for Alberto Calder6n Z. May 1996 M. Espinosa Secondary Schools: The 37599 Colombian Experience HCDWP67 NGO-World Bank Toshiko Hino June 1996 A. Thomas Partnerships: A Tale of Two 31151 Projects HCDWP68 The Disability-Adjusted Life Nuria Homedes July 1996 L. Arias Year (DALY): Definition, 35743 Measurement and Potential Use HCDWP69 Equitable Allocation of Ceilings Philip Musgrove August 1996 Y. Attkins on Public Investment: A 35558 General Formula and a Brazilian example in the Health Sector HCDWP70 The Economics of Language: Barry Chiswick September 1996 I Conachy The Roles of Education and 33669 Labor Market Outcomes HCDWP71 Agricultural Growth and Rashid Faruqee September 1996 C. Anbiah Poverty in Pakistan Kevin Carey 81275