## Jean Drèze and Amartya Sen

Print publication date: 1991

Print ISBN-13: 9780198286370

Published to Oxford Scholarship Online: January 2008

DOI: 10.1093/acprof:oso/9780198286370.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (oxford.universitypressscholarship.com). (c) Copyright Oxford University Press, 2021. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in OSO for personal use. date: 26 January 2021

# 10 The Elimination of Endemic Poverty in South Asia

## Some Policy Options

Chapter:
(p.347) 10 The Elimination of Endemic Poverty in South Asia*
Source:
The Political Economy of Hunger: Volume 3: Endemic Hunger
Publisher:
Oxford University Press
DOI:10.1093/acprof:oso/9780198286370.003.0011

# Abstract and Keywords

This chapter evaluates policies for combating persistent and mass poverty, with empirical illustrations from the experiences of some South Asian countries. It argues that direct action is possible and even desirable, and that different policy options are available for achieving higher growth and greater equity. The chapter scrutinizes various poverty alleviation programmes, many of which failed but are promising if effectively planned and implemented. To design an efficacious anti-poverty policy, it is essential to evaluate the whole package of several schemes together. The success of direct action programmes depends greatly on their political background. It is argued that political constraints are not only difficult to overcome but also difficult to understand.

# 10.1. Introduction

In this age of mammoth scientific achievements we have the necessary technology, raw material, and skills to grow all the food we need. We also have a transport system—or at least the capacity to build up the transport system—which will carry the food to everybody. Nevertheless we have failed to do so, and persistent mass poverty is a part of twentieth‐century life. That this appears shocking to most of us shows that we have an inherent tendency to underestimate the complexities of social and political engineering. The oft‐heard view that removing poverty is not difficult at all and our failure is merely because of vested interests or lack of willingness among powerful people reflects a misunderstanding of the word ‘difficult’. Given that vested interests are a part of the world, as also are powerful and uncaring people, it is indeed a difficult task. In designing policies, it is important to recognize this.

This work is a contribution to the problem of formulating policies to combat persistent poverty, while recognizing that there are interests and incentive structures which need to be overcome or tiptoed past. The context of almost all the analysis will be South Asia, in particular, Bangladesh, India, and Sri Lanka.

The study begins by considering the issue of basic needs provision.1 In order (p.348) to improve the standard of living of the masses, do we have to wait for the benefits of growth to trickle down or could we adopt ‘direct action’? This is the subject matter of section 10.2.

Section 10.3 is a detailed consideration of one aspect of basic needs: the removal of poverty and hunger. South Asia has experimented with myriad schemes for supporting the poorest sections and providing employment to the rural landless labourers, for example, the food‐for‐work programme and the Employment Guarantee Scheme. These are surveyed briefly and suggestions are made for making them more effective.

Even if we accept the desirability of direct action and also know what programmes and schemes to adopt, the actual implementation of such a policy may be hindered because of political constraints. The subject of political constraints is a difficult one and section 10.4 makes a few preliminary comments on it.

# 10.2. On direct action and ‘trickle‐ down’

It is an old debate as to whether, in a nation's fight against poverty and deprivation, the aim should be to undertake direct action against these or to strengthen the forces of growth and let the poor benefit from the trickle‐down effect. Not only does this debate have important policy‐consequences for South Asia, but the varied experience of regions in South Asia provides useful evidence for designing effective policy. From our point of view the most remarkable evidence comes from Sri Lanka and the State of Kerala in India. In terms of economic performance, especially growth rates and per capita incomes, these are backward regions; but their achievements in terms of basic needs and standards of living, captured mainly by health and social statistics, are outstanding.2 Not surprisingly, the role of direct action in these areas has been extensively written about and debated.3

I shall begin by commenting on the Sri Lanka debate, comments on Kerala being reserved for later. The Sri Lankan ‘paradox’ is well known. In terms of the economic criterion that receives most attention, to wit, the per capita real income, Sri Lanka is a very poor country. But its social statistics—variables which indicate the standard of living or quality of life—are very close to those of advanced industrialized countries.4 As far as per capita income goes, Sri Lanka ranked exactly on par with only one other country in 1983: Sierra Leone. Both these countries had a GNP per capita of $330 in that year. (p.349) However, in Sri Lanka life expectancy at birth was 69 years, whereas in Sierra Leone it was 38. Turning to the geographic area being studied in this paper, which also happens to be culturally comparable to Sri Lanka, we find a similar contrast. In India and Bangladesh, life expectancy in 1983 was 55 and 50 years respectively. In Pakistan, where per capita income at$390 is higher than that of Sri Lanka, life expectancy was only 50. Infant mortality in Sri Lanka was 37 (per 1,000 live births), whereas in India it was 93 and in Bangladesh 132. In terms of the percentage of children of the relevant age group attending secondary school,5 Sri Lanka with a score of 54 is ahead of India (30), Bangladesh (15), and even China (35). In fact, the only country with a comparable per capita income and with similar standard‐of‐living indicators is China.

It is possible to find countries with six or seven times Sri Lanka's income with standards of living not markedly different from it. Examples include South Korea, with a per capita GNP of $2,010, and Mexico with$2,240. In all the three standard‐of‐living indicators discussed so far,6 Mexico does equally (p.350) well or marginally worse than Sri Lanka. In short, a comparison across countries shows Sri Lanka as an ‘outlier’: In terms of most social and health indicators, Sri Lanka's actual figures are substantially superior to what we would ‘predict’ on the basis of its per capita income.7 For these reasons Sri Lanka is often held up as exemplary for other poor countries and also as proof of the fact that ‘public action’ can lead to the provision of basic needs and, therefore, to an improvement in the quality of life (see e.g. Jayawardena 1974; Fields 1980; Isenman 1980; Sen 1981 a; Caldwell 1986).

This very broad thesis has recently been subjected to a detailed econometric scrutiny and has been the source of much controversy and debate (Bhalla 1988 a, 1988b; Bhalla and Glewwe 1986; Sen 1988). Bhalla and Glewwe argued that the existing studies, in particular Fields (1980), Isenman (1980), and Sen (1981 a), were flawed because they were based on a comparison across countries at a point of time. This, they showed, led to the ‘initial conditions’ getting omitted. So they set out to examine Sri Lanka's performance over the period 1960 to 1977. As I have already argued and is clear from Table 10.1, if we look at Sri Lanka's ratio of social indicators to per capita income, it is markedly better than that of almost all other countries. Now in order to net out the effect of ‘initial conditions’ what needs to be done is this: construct a table similar to my Table 10.1, but replace column 1, which shows per capita income at a point of time, with a column showing changes in per capita income between 1960 and 1978; and similarly replace the other columns with ‘changes’ in variables from the existing ‘values at a point of time’. Does Sri Lanka continue to stand out as a country which has done exceedingly well in improving living standards compared to other countries whose change in per capita incomes is similar to that of Sri Lanka? This is the question that Bhalla (1988 a, 1988b) and Bhalla and Glewwe (1986) examine meticulously, and they come out with a clear answer: No. On this they are completely convincing; where they err is in drawing inferences from their finding.

Table 10.1  Inter‐country social statistics, c. 1983

GNP per capita, 1983 (\$US)

Life expectancy at birth, 1983 (years)

Infant mortality rate, 1983 (per 1,000 live births)

Number enrolled in secondary school as % of age group, 1982

Mexico

2,240

66

52

54

Korea, Rep. of

2,010

67

29

89

Pakistan

390

50

119

14

Sierra Leone

330

38

198

12

Sri Lanka

330

69

37

54

China

300

67

38

35

India

260

55

93

30

130

50

132

15

Source: World Bank (1985).

For instance, they treat their finding as evidence that a larger social expenditure need not lead to higher living standards. But to maintain this on the basis of a regression analysis of the kind conducted by Bhalla and Glewwe, it is clearly necessary to have comparative data on changes in social welfare expenditure in all the countries appearing in the analysis.8 Without this we cannot jump to any conclusion about the impact of social welfare expenditure based on the finding that in terms of ‘changes’ between 1960 and 1978 Sri Lanka is not an outlier. Consequently, the basis of their claim gives way when, (p.351) after conducting their econometric exercises, Bhalla and Glewwe announce that they do not have these crucial data on welfare expenditures.

As a matter of fact, the proposition that large social expenditures do not necessarily lead to better standards of living is obvious enough not to require any formal econometric study. This is because money is easy to fritter away. One region with a small expenditure can achieve what another region with a large expenditure fails to because of differences in efficiency or methods of disbursement. Thus in Kerala and West Bengal the volume of medical facilities available is comparable, but utilization is much higher in Kerala (Nag 1983); not surprisingly Kerala performs much better than West Bengal in terms of health statistics. So it is easy to concede that while a certain volume of social expenditure is necessary to improve living standards, it is not a sufficient condition.

The other important question that emerges from this debate is: can basic needs be provided directly or do we have to wait for the trickle‐down effects of growth? Bhalla and Glewwe treat their findings as implying that reliance on growth rather than direct action is essential. But to answer the above question we have simply to seek examples where income is low but living standards are high, i.e. do the kind of exercise that Isenman (1980) and Sen (1981 a) have done. This, as we have just shown, is the case with Sri Lanka, and, as I show below, is also partly the case with Kerala. Hence, whatever it is that Sri Lanka did has led to its achieving very high levels of provision of basic needs. Clearly this is all we need to show that we do not have to rely on growth to mitigate poverty and provide basic needs. To assert that something works, but direct action does not, implies a semantic misunderstanding of the term ‘direct action’.

It is possible to argue that the cause of the misunderstanding is that many analysts have labelled whatever the Sri Lanka Freedom Party (the one associated with the Bandarnaikes) did as ‘direct action’ and have taken the failure of those policies to mean a failure of direct action. I will argue later that despite the equity orientation of the Bandarnaike government its policies were often inappropriate.

Given that direct action is possible, the next question is: what is the appropriate direct action for South Asian countries? Before delving into this general subject, let us take a look at the causes of Sri Lanka's—and later I comment on Kerala—exceptional performance. The main contribution of Bhalla's work is to shift our attention to the period before the 1970s. As it turns out, there is clear evidence of governmental action. In particular, three policies which played important roles are: provision of health facilities, free and compulsory education, and free or subsidized food rations.

Sri Lanka's record in terms of health facilities is good. Richards and Gooneratne (1980) have documented some of these and I note here only a few salient features. It is well known that Sri Lanka's excellent mortality figures owe a lot to the malaria eradication programme of 1946. In one year, between (p.352) 1946 and 1947, the death rate fell from 20 to 14. As one would expect, there were sharp improvements in infant mortality and life expectancy around the same time. Clearly a malaria eradication programme is a direct welfare activity of the government. More generally, the ratio of doctors and nurses to the total population in Sri Lanka has been high relative to the South Asian experience. In the early 1970s the nurse–population ratio was four times that of India. From the mid‐1950s to the early 1970s real expenditure on health services rose steadily.9

Despite all this, for a health programme to succeed, it is essential that the people accept what the government offers. It is not sufficient to have a large family planning programme if people are not willing to listen, or a large vaccination centre if people view it as a torture cell. So for a health programme to succeed it is important that people have some education. Isenman's regression analysis confirms the crucial role of literacy in improving life expectancy and infant mortality.10 Demographers have often stressed the positive relation between literacy among women and infant mortality. Some data suggesting this relation can be found in the Survey on Infant and Child Mortality 1979, conducted by the Office of the Registrar‐General in India. This is presented in Table 10.2. The table suggests not only a relation between health and the 0–1 concept of literacy but a monotonic relation, with increasing education resulting in lower mortality.

Table 10.2  Mortality and the education of women in India, 1978

Education of mother

Infant mortality

Rural

Urban

Illiterate

145

88

Literate but below primary

101

57

Primary and above

71

47

Source: Office of the Registrar‐General of India (1981 a).

Given that Sri Lanka has had a long history of free education—dating back to its pre‐independence period—it is not surprising that its health programme has been successful. Its education programme is an example of meaningful social expenditure. As Isenman (1980: 239) notes, ‘As a result of high expenditures and high enrolment rates at all levels of education, adult literacy increased from 58% in 1946 to 78% in 1971.’

Finally, Sri Lanka has had a comprehensive system of free food rations for (p.353) about 40 years.11 This has taken the form of providing a certain amount of free rice (and, on occasions, wheat) and some additional rations at a subsidized price. What is unusual about Sri Lanka's rationing system is its comprehensive coverage of the entire population, including the urban and rural areas. This is in sharp contrast to India and Bangladesh. It will be argued later that Sri Lanka would have done better to cover a target population, namely those in poverty. Nevertheless, Sri Lanka's system of total coverage was better than having a system where only the urban sector has access to rationed foods. There has been a tendency for this to happen in some countries and its deleterious consequence was clearly evident during the Bangladesh famine of 1974. The urban sector, being covered by a rationing scheme, could ensure that a substantial part of the food would be diverted to it. The rural entitlement crisis was thus heightened by the very fact of the rationing scheme (McHenry and Bird 1977; Sen 1981 b).

Not surprisingly, Sri Lanka's daily calorie intake is much higher than that of India or Bangladesh. Coupled with the fact that poverty in Sri Lanka is less, this suggests that the calorie intake of the poor is superior. The Sri Lankan food ration and subsidy scheme has been noted by several authors as the cause of its greater equality and better nutrition (Jayawardena 1974; Sen 1981 a). The change in inequality in Sri Lanka over time is a much more controversial subject.12 In fact, the whole question of changes in economic conditions in the 1960s and 1970s and their connections with the politics of the nation is an interesting and much misunderstood subject. But before going on to that I want to briefly dwell on the other striking example in South Asia of a poor region attaining a high standard of living—the State of Kerala.

The status of Sri Lanka among nations and that of Kerala among the Indian States bear striking resemblances. Kerala is one of the poorer States in India in terms of per capita income and also per capita nutrition (see Centre for Development Studies 1975; Dholakia and Dholakia 1980; Bardhan 1984 a)—though its nutritional status is probably much better than a direct reading of the data suggests (Centre for Development Studies 1975: ch. 1). In terms of standard‐of‐living indicators, however, Kerala is markedly ahead of other States and in some ways quite close to advanced industrialized countries. In 1978 literacy in Kerala was 72 per cent as compared to the all‐India figure of 39 and Punjab's 47. Infant mortality in Kerala was 39 compared to the all‐India (p.354) average of 126 and Punjab's 103. These statistics and more are presented in Table 10.3.

Table 10.3  Inter‐state social statistics in India, 1978

Literacy (%)

Male literacy (%)

Female literacy (%)

Infant mortality (per 1,000 live births)

All India

39

50

27

126

Kerala

72

77

66

39

Punjab

47

54

39

103

33

46

19

167

West Bengal

47

56

35

78 (approx.)

Source: Office of the Registrar‐General of India (1981 a).

As in the case of Sri Lanka, Kerala demonstrates that improvement in the standard of living does not have to come via growth. Also, as we look into the factors behind Kerala's remarkable achievement, a lot of the same factors stand out here as in the case of Sri Lanka.

Kerala's educational performance is a direct consequence of progressive government policies. The State has had a long tradition of free primary schooling which has led to high enrolment rates and, more importantly, low drop‐out rates. These tendencies have been strengthened by the provision of free meals at schools to some categories of students (CDS 1975).13 Compared to other States, Kerala's performance in terms of female literacy is particularly striking, as Table 10.3 shows. This, as argued earlier, helps in the attainment of a better health status.

It seems generally accepted that health facilities are better distributed in Kerala.14 The State's progressive health policies go far back into history, to the policies of the princely State of Travancore (Panikar and Soman 1984).

A major factor behind Kerala's health statistics is its food rationing scheme. As in the rest of India, subsidized foodgrains are sold in limited quantities to individuals through ration shops. However, whereas, all over India, food rations are essentially an urban feature, in Kerala the coverage is comprehensive, including urban and rural areas. This is again a striking similarity with Sri (p.355) Lanka. In Kerala the public distribution scheme reaches 97 per cent of the population; in Sri Lanka, before 1978, 93 per cent of the population was covered. Detailed studies of Kerala (Kumar 1979; George 1979) seem to establish the important contribution of the public distribution system to the citizens' nutrition and health. This is particularly true of the poorer classes.

State‐wide data on social expenditure are difficult to get and what little is available presents problems of interpretation. Hence, it is difficult to judge whether social expenditure was high in the case of Kerala. This, however, as argued earlier, is not the crucial question. Clearly the expenditure must have been above a certain level since the provision of free education, subsidized health facilities, and rationed food entails this. Apart from this it is not clear whether Kerala spent little on social expenditure very efficiently or a lot inefficiently. The important point is that Kerala undertook direct action in government activities and this resulted in a relatively higher standard of living than one would expect from its general economic indicators.

What then is the basis of the view that direct redistributive and equity‐oriented policies cannot work? One reason must be that policies come in bundles and it is difficult for us to sort out which is the cause of success and which of failure. We have some broad stereotypical notions of different governments: South Korea and Taiwan are laissez‐faire economies, so if they succeed it is a success of non‐intervention. China is a socialist country, so its success indicates the effectiveness of intervention. Reality can be different. For instance, in the case of South Korea and Taiwan, some policies that have been followed by their governments (e.g. land reform) are more radical than those any South Asian government has tried.15 Given the scope of this essay, I shall concentrate on South Asia and illustrate this point with the example of Sri Lankan politics.

For all practical purposes, Sri Lanka has had a two‐party system: the Sri Lanka Freedom Party (SLFP) associated with the Bandarnaikes, and the United National Party (UNP) associated with Senanayake and now Jayawardena.16 The SLFP is considered the left‐of‐centre party and the UNP the right. The Sri Lankan electorate is volatile and responsive and regular changes in government have been seen, as follows: 1947–56, UNP in power; 1956–65, SLFP;17 1965–70, UNP; 1970–7, SLFP; 1977 onwards, UNP.

Economists, implicitly or explicitly, have taken 1956–65 and 1970–7 (because these were the times when the SLFP was in power) as the periods when ‘direct action’ was at its peak; failures during these periods have been taken to (p.356) be failures of direct governmental action. But politics is a more complex game. A country which espouses socialism within its boundaries may follow a foreign policy of supporting right‐wing regimes. Even in a party's domestic policy one can find unexpected mixtures which render easy labelling hazardous.

Note first that Sri Lanka's welfare programmes concerning education, health, and food were in effect during the first UNP government. Also, in 1978 it was the UNP government that made the food rationing scheme more equitable by ruling that only families whose annual income fell below Rs 3,600 would be eligible for free rations.18 With this progressive move, the total number of beneficiaries fell from 93 per cent of the population to 54 per cent.19 In retrospect it is clear that this progressive move was but a first step towards dismantling most of the public distribution schemes Sri Lanka had had. In 1979 the Jayawardena government abolished the rice rations and introduced a food stamp programme in its place. But the value of the food stamp programme was fixed in nominal terms and its real value has unfortunately declined rapidly because of inflation.20

As far as the SLFP is concerned, it is vital to distinguish between its two periods of government. During its first reign, that is, 1956–65, it stepped up social expenditure sharply—from about 5.3 per cent of GNP to 12.8.21 It passed the Paddy Lands Act (1958), to increase security of tenure, and undertook many other domestic measures.

The second period, 1970–7, was very different.22 The government's foreign policy was now more radical (e.g. it established diplomatic relations with several Communist countries), but its domestic rule was inefficient and inequitable. There was first of all considerable instability, including that stemming from the Communist insurgency of 1971, which the Prime Minister, Mrs Bandarnaike, put down ruthlessly. Unemployment was on the increase and money wages began to slip behind the price index. The real downturn came in 1973–4 with a severe failure of food crops and also as a consequence of the general international slump. Calorie intakes fell and health indicators deteriorated slightly. In addition, welfare expenditures were on the decline. Real expenditure on health programmes was probably falling, but, more (p.357) importantly, the size of the population per doctor or assistant medical practitioner was rising: from 3,800 in 1971 to 4,250 in 1976.23 As Richards and Gooneratne (1980: 152) note, ‘This coupled with drug scarcities and rising costs must indicate a recent deterioration in levels of medical care.’ Government expenditure on education as a percentage of GNP also fell, from 4.3 in 1970 to 2.7 in 1977.24 It is not surprising, therefore, that a study of change during the 1970s does not show up Sri Lanka in a particularly favourable light.

The final issue that I want to discuss in this section is the larger macroeconomic one of the alleged trade‐off between growth and equity. If we do go in for the kind of basic needs policy discussed above, do we have to sacrifice growth? A lot of traditional thinking presupposes the existence of such a trade‐ off, though several authors have challenged this presupposition (a particularly vehement and clear case was made by Myrdal 1970).25 There are two reasons why a simplistic view of a growth–equity trade‐off may be wrong. First, it is reasonable to assert that, even if there is a relation between growth and equity, it is not a ‘functional’ one but in the nature of a ‘correspondence’, whereby a set of values of growth is compatible with each level of equity. Though in this case there may be a potential trade‐off, this may not be of any immediate concern to the economy. This would be true of an economy which is functioning below its ‘possibility frontier’ in the growth–equity space. This is the view that I take here: in most LDCs there is sufficient slack for them to have more of both equity and growth.

A second and more fundamental criticism (one which is not pursued here) is to deny any relation between growth and equity in an economy. The absence of any relation between two variables may at first sight seem difficult to imagine but there are plenty of examples of this in economics. For example, it is meaningless to talk of the relation between price and supply in a traditional monopoly. If this happened to be true of growth and equity, then we would not be able to talk of a trade‐off, immediate or potential. There can be other, more philosophical problems. For instance, one could argue that the only growth‐equity combination that can occur in a country is the one that actually does occur. The question of what rates of growth are compatible with a different level of equity may then be meaningless because a different level of equity may be logically inconsistent with the other given features of the country.

One reason for the widespread belief in a trade‐off is that policies often come in bundles: the government that enhances controls (an act which often has a negative impact on growth) is the one that has a more radical poverty‐removal programme (which brings about greater equity). On the other hand a government like the one formed by Jayawardena tends to cut controls and minimize expenditure on basic needs programmes.26 It is this policy bundling which (p.358) may give us an exaggerated view of trade‐ offs, in particular the false view that an increased social spending must result in sluggishness of growth.

The amount of slack in most LDCs is large. The fact that a country takes up an activity x with zeal therefore does not mean that there will be less of activity y.27 (Of course, we may nevertheless lament that the zeal is for x and not for y.) Parents who curb a child's zeal for sports, in the belief that this will help him become a scientist, will usually be disappointed. A better strategy for the parents would be to try to develop the child's enthusiasm for science.

Given the political structure of South Asian governments, there is an inherent tendency for them to indulge in controlling activities: licensing, price‐setting, and fixing trade tariffs and quotas. These often encourage inefficiencies and curb growth, but the established domestic businesses have a vested interest in this structure of controls which makes it very difficult to dismantle it. Every time some import restriction is removed and consumers find they can buy some product cheaper and of better quality, one hears cries of dumping and ‘foul’. There is much to be gained on the growth front by lowering trade restrictions and allowing prices to play a more important role;28 and these policies need not be a curb on basic needs policies. In fact in the long run basic needs policies may actually be helped by these.

Bhagwati (1985) has argued that the problem with a large basic needs programme is not that in itself we cannot have it or it will not be beneficial; but simply that we may not be able to sustain such a programme for long. Indeed Sri Lanka was under severe fiscal strain in the 1970s. To ensure the sustenance of poverty alleviation programmes, we need some concomitant policies. We need to bolster growth so that the government has adequate resources for its programmes. What I have just been arguing is that it is not necessary to prune the poverty alleviation projects in order to have growth. Growth can be bolstered through a different set of policy instruments while continuing with the same basic needs policies. Of course, we may have to try to restrict the recipients of government support to the really needy.29 This, as we have seen in the case of Sri Lanka, is a difficult task; but it is essential for keeping the fiscal burden manageable. This subject is discussed in the next section. Thirdly, what can be implemented or sustained depends on the political climate of a nation. This is discussed in section 10.4.

In the next section I narrow the focus from basic needs policy in general to a part of it: the provision of food and the combating of poverty. This narrowing (p.359) down reflects the scope of this chapter and is not meant to suggest the unimportance of the other features of a basic needs programme.

# 10.3. Poverty alleviation programmes

South Asia has seen a plethora of schemes for alleviating poverty and acronyms have multiplied faster than one can count. Popular programmes include: food‐for‐work (FFW); Employment Guarantee Scheme (EGS); food ration and subsidy schemes; food stamps; nutrition programmes like Tamil Nadu's noon‐meal scheme; Integrated Rural Development Programme (IRDP); credit support schemes like the Grameen Bank in Bangladesh. Of these, the IRDP and Grameen Bank, while very different from each other, are also very different from the other schemes and on these I offer only a few remarks. Programmes like FFW and food rations are more directly concerned with attacking the problem of food entitlement and unemployment and these are of central interest in this section.30 I shall however begin with a few comments on programmes which provide some credit support.

IRDP is an important programme in Bangladesh and India. In India 15 million people were meant to be assisted by it during the Sixth Five Year Plan.31 The motivation for the programme arose out of a feeling that the earlier anti‐poverty policies were piecemeal and needed consolidation and that they were working mainly as subsidy schemes on which the poor were likely to get chronically dependent. The aim of IRDP was, therefore, to offer a combination of subsidies and credit to poor households so as to bolster their asset position and enable them to have a higher income and become self‐sufficient. The scheme was supposed to help the poorest sections of the rural sector. ‘Poor’ was now being defined differently from what was conventional. First, the household, instead of each individual, was being treated as a unit. Secondly, poverty was not defined in terms of landholdings, but in terms of the family's income from all sources. If this was below a certain level the family was considered eligible for support from IRDP. While this is, in principle, a better criterion for selecting the needy, it is also one which makes the problem of identifying the beneficiaries more acute.

Studies of the functioning of IRDP in India so far reveal that it has been very successful in terms of the target number of beneficiaries. But the average amount of investment that each beneficiary received was only about half of the planned investment of Rs 3,000 per family at 1979–80 prices. Also, in terms of (p.360) the objective of reaching the poorest people, the achievement of IRDP is doubtful, though no definitive study on this is available.

The targeting problem was, to a certain extent, preordained because of some inconsistencies in the planned objectives of the project. While (1) it was supposed to benefit the poorest, (2) it was also expected to assist families so as to push them above the poverty line. There is some ambiguity about objective (2). Two alternative elucidations of (2) are: (a) the number of families that cross the poverty line should be maximized. (b) A family which is assisted ought to be assisted sufficiently to cross the poverty line. Under both interpretations (2) conflicts with (1). With (b) the conflict is indirect and stems from the fact that during the process (over time) of assisting a family, it may cease to be the poorest. The conflict between (1) and the first interpretation is direct. If we want to maximize the number of families that cross the poverty line, we should choose as beneficiaries families which are close to the poverty line. This would, of course, conflict with objective (1). The incompatibility between (1) and (2) is very marked in practice. Sundaram and Tendulkar (1985 a) have shown that, given some assumptions, if the average amount of benefit per family disbursed during the first three years of the Sixth Five Year Plan had gone to the poorest 30 per cent of the rural population, then none of the IRDP beneficiary households would have crossed the poverty line, that is, objective (2) would have remained completely unsatisfied.

Though we have to wait for more detailed empirical research, there are some grounds for maintaining that the poorest people were not reached. For one, there were other objectives which militated against this happening. For instance, the so‐called ‘cluster approach’ which was adopted by IRDP meant that its assistance would go to areas where credit institutions existed and there was ‘the capacity to absorb credit to the extent envisaged’ (Sundaram and Tendulkar 1985 b: 209). Further, since a part of the IRDP package consists of assistance with credit, it would face the difficulties which credit support policies face in general in trying to reach the very poor.32 I am in sympathy with Rath's (1985) argument that, though credit and subsidy for self‐employment are important, we cannot ignore the need for massive programmes for creating public and private employment (see also Dantwala 1985 and Hirway 1985). Consequently, the waning of official interest in India for such schemes (see Rath 1985) is unfortunate. A similar sentiment is expressed by Dandekar (1986: A‐100): ‘… while the possibilities of creating self‐ employment should be explored, the main reliance will have to be on offering wage‐employment’. He rightly observes that this does not mean a perennial dependence because the ‘more thrifty, provident, and enterprising’ among the beneficiaries of an (p.361) employment scheme are likely to ‘set themselves up’ in the long run and become independent.

Before turning to employment programmes, let us take a brief look at pure credit support policies which have also been tried over a long period in South Asia. In India, the coverage provided by institutional credit has steadily increased (thereby hopefully diminishing the hold of the legendary rural money‐lender). But nevertheless the credit policy has been far from a success. The main reason is that institutional credit has failed to reach the poorest people, who have by and large remained dependent on the unorganized sector, i.e. on private money‐lenders.33 Hence it is possible that institutional credit, instead of having alleviated the condition of the poorest, has merely bolstered the position of the not‐so‐poor. The reason for this is sufficiently deep as to imply that it cannot be corrected by merely doctoring the credit programmes. The main problem of providing organized credit to the poor is the recovery of loans. The village money‐ lender has a variety of methods for ensuring repayment34 which are not open to the manager of the bank sent from the nearest city to provide rural credit. Apart from the problem of absconding from repaying, there is the genuine problem of bankruptcy. The very poor who borrow to survive may well find themselves with no liquidity at the time of repayment. The private money‐lender may then ‘bond’ them, that is, acquire a right over their labour for a long enough period to recover the loan.35 It is this knowledge which gives the money‐lender the confidence to lend to the poor. Clearly this option is not open to government‐run banks. Moreover, the rich and influential have the ability to divert to themselves the organized‐sector low‐interest credit; and even if their repayment record is poor they continue to get loans (see e.g. Bhende 1986). It is for these reasons that a credit scheme is very unlikely to be able to confer benefits on the poorest sections.

The Grameen Bank of Bangladesh however seems to provide a counter‐example. Established in 1983, it is a successor to the Grameen Bank Project which was launched in 1976 by Muhammed Yunus, an economics professor at Chittagong University. It is considered one of the most successful rural credit schemes in South Asia. The Grameen Bank is a public sector credit institution and its aim is to provide loans to the poor on reasonable terms, the idea being that of enabling the poor to become self‐sufficient. A member of any household owning less than half an acre of cultivable land can avail himself of the services of the Grameen Bank. The workers of the bank actually search for poor people who need financial support but are too ignorant or diffident actively to seek credit. Paradoxically it has done well exactly where other schemes have (p.362) floundered. Its repayment record is excellent. Studies done in the early 1980s show that only 1 or 2 per cent of all outstanding loans were overdue (Siddiqui 1985; see also Ahmad and Hossain 1985). Secondly, its targeting has been quite good. According to a 1983 survey, a vast majority of loanees owned no agricultural land (for detailed tables, see Siddiqui 1985: 175). Another striking feature of the Grameen Bank is the prominence of women among the borrowers. In 1980, 39 per cent of the borrowers were women and by late 1984 the figure had risen to 54 per cent (Hossain 1985: 12).

A part of the Grameen Bank's success lies in its organizational structure. For example, the manager of a new branch has to survey the concerned villages for two months without subordinates. This ensures first‐hand knowledge and direct involvement. The Grameen Bank is, in part, a co‐operative. It organizes credit recipients into groups and helps them build up their own savings for emergencies. The group is useful in preventing defaults by applying ‘social’ pressure on individuals to stick to schedules. This co‐operative aspect of the bank may have had other indirect benefits. It has, for instance, been found that an increasing amount of credit is being taken from the bank for co‐operative ventures like purchasing equipment for irrigation which would benefit a group or buying rice hullers or even leasing market‐places (Ahmad and Hossain 1985). According to Hossain (1985: 12–13), among the main causes of the Bank's good record are the ‘close supervision of the activities in the field by the managing director’ and the dedication of the bank workers, ‘most of whom have taken the job as providing services to the poor rather than simply as an income earning opportunity’. In the context of the present chapter the important question is: to what extent can such a credit programme be replicated elsewhere?

In the early 1980s the Grameen Bank was a small project with plans for vast expansion through the 1980s. In May 1980 the bank had 24 branches. By February 1987 this had grown to 298 and loans had been disbursed to about 250,000 households. The plan was to keep up the expansion for some more years.36 With such expansion it will of course lose the personal touch; and the robustness of its organization and its relevance for large countries, like India, will only then be fully tested. However, its success thus far is good reason for other South Asian countries to examine the viability of such a credit programme.

More direct methods of enhancing the food entitlements of the poor are the EGS or FFW and food rations or food stamps. These have had a fair measure of success in the Indian subcontinent. Food‐for‐work in Bangladesh was begun in the wake of the 1974 famine. The programme has run mainly on food received as aid—primarily from the US. In the first year of the programme, (p.363) 32,000 tons of wheat were disbursed through the programme and 8.6 million man days of work were created. In 1982–3, approximately 371,000 tons of wheat were used to create 101 million man days of work. The FFW programme is computed to have created jobs equivalent to only 2–3 per cent of the total annual unemployed labour time in Bangladesh. Also, it has been estimated that 30 per cent of the workers came from households owning more than half an acre of land (Ahmad and Hossain 1985).

In India FFW (which was subsequently renamed the National Rural Employment Programme, NREP) was begun in April 1977, with the objective of creating jobs, creating durable infrastructure, and using up the surplus grain which had accumulated with the government.37 In this respect India differed from other countries; its programme was not based on food received as aid.38 In 1978, 12 lakh tonnes of wheat were used and 286 million man days of jobs were created.39 Though the programme was running fairly successfully it got mired in political controversy and was somewhat tarnished. First of all the programme caused a certain amount of alarm because it was having the sort of impact it was intended to have: a study of the Planning Commission found that, of the twenty Districts surveyed, six had experienced a significant upward movement of wages. This was undesirable from the point of view of rich landlords and may have been a factor behind the loss of official enthusiasm for FFW. Secondly, FFW became the source of Centre–State conflicts and this resulted in a curb on foodgrain supplies to several States, including West Bengal (for a discussion, see Basu 1982).40 In terms of economic efficiency, where the programme has really floundered is in its effort at creating infrastructure. While it did create a certain amount of employment, it was usually unproductive. But indeed the sheer fact of doling out food may be of value.

The rice ration scheme or the provision of food stamps41 are methods of (p.364) doling out food and, as I have argued in the previous section, they have played an important role in raising living standards in Sri Lanka and Kerala. Ideally FFW and rice rations should be used as complementary schemes because their points of strength are very different. FFW, properly executed, can provide a self‐selection device for picking out the poor and also it has the advantage of being productive. Its main disadvantage vis‐à‐vis a rice ration scheme is that it discriminates against the old and the disabled. Since the old and disabled are easier to identify, one possibility is to have free rations restricted to them and an FFW open to all and hopefully, because of its self‐selection property, utilized only by the poor.

For countries as large as India and Bangladesh, the problem of selecting the poor is so important that reliance on FFW seems natural. If too many non‐deserving people rely on such governmental support, the burden on the exchequer may be unbearable. To get the full advantage of an FFW we need to organize it very skilfully. To bring out the salient features of FFW, I present some simple analytics.

The difference between an FFW and a food ration scheme is that they alter entitlements differently. Suppose a person earns z rupees from a day's labour;42 and that his daily non‐labour income is Rs v. Let y denote his total income. Thus

$Display mathematics$

Let p be the price of the foodgrain in the open market. In the absence of a price support scheme, his budget set is given by y0A is Fig. 10.1. Now let us consider two alternative support programmes.

Fig. 10.1. Food for work and food ration

Food ration scheme: let us suppose that a free rice ration of R units is given. Then this person's budget set becomes 0yCB. Here and below I assume that the transaction cost of reselling food is prohibitive.

Food‐for‐work: now suppose that instead of a food ration scheme food‐for‐work is started where foodgrains worth G units are given in exchange for a day's labour. Then his budget set is 0vDE.

So what these schemes do is alter the entitlements of individuals. The interesting feature of FFW is that z will vary between individuals. If for a person z is sufficiently high, then his food‐for‐work budget set becomes a subset of 0yA (see Fig. 10.1) and he would prefer not to avail himself of the opportunity of FFW. It is easy to check that the FFW budget set is subsumed in the normal budget set if

$Display mathematics$
(10.1)

Hence a person will certainly not work for FFW if condition (10.1) is true. (If 10.1 is false he may or may not work.) It is clear therefore that those who have (p.365) access to a well‐paid labour market or high productivity in their own farms (basically a high z) will not participate in FFW projects. This is the self‐selection property of FFW.

There is one problem. A person with a high non‐labour income, but low z, may nevertheless come to work for FFW and clearly we would not like to have such people. There are two mitigating factors. First, manual labour is considered demeaning for anybody who has access to other income, especially a large non‐labour income. Thus for reasons of status, those with a sufficiently high v are unlikely to come. Secondly, it is quite possible that v and z are positively correlated: it is the richer people who have access to well‐paid labour markets. This may be briefly captured by asserting

$Display mathematics$

In that case it is clear that if a person's v is large, (10.1) will be satisfied and the FFW programme will be spared the labours of such a person.

A controversial matter concerning FFW is the level of wages that ought to be paid to the labourers.43 It is a widely held view that the wages paid on FFW ought to be higher. It will, however, be argued here that both on grounds of keeping the self‐selection property sharp and also for ethical reasons the wage paid at FFW sites, i.e. G, should be kept as low as possible (in a sense made clear below).44

This seems to be a surprising recommendation if our objective is to remove (p.366) poverty. But such a feeling of surprise arises from an implicit ‘headcount’ view of poverty. This comes out clearly from Dandekar and Sathe's (1980) study of FFW and EGS in Maharashtra. They found that 90 per cent of the people working on this scheme continue to be below the poverty line despite such work. From this they went on to conclude that wages should be raised. In the case of Bangladesh, Ahmad and Hossain (1985: 80) observed that wages paid to FFW workers were substantially below the officially stipulated wage rate. ‘It has been shown that about 56 per cent of the workers did not know about the stipulated wage rate. Those who know do not bargain lest they do not get the jobs at all as there are many others who are unemployed and would be too willing to take them up on the offered terms and conditions’ (my italics).45 It is the italicized part which suggests why underpayment need not be unethical, since that will make it possible to employ a larger number of people who are needy enough to be willing to work for a low wage.

Suppose we subscribe to a headcount view of poverty46 and try to minimize this. Then, given a total stock of foodgrain X, which is to be disbursed through the FFW, we would try to heap it on people so as to ensure that the maximum number of people cross the poverty line. But clearly our more intuitive normative penchant (as opposed to one formally derived from trying to minimize the headcount index) would be to spread out X over the poorest people, even if that leaves the numbers on the two sides of the poverty line the same. Fortunately, according to some more sophisticated measures, this will register a decline in poverty.

To formalize this argument suppose X is the total amount of grain available for giving out as wages in a FFW. For simplicity I am assuming that wages are paid entirely in terms of food grain. Let L be the number of labourers supplying their labour to FFW. As usual, we assume

$Display mathematics$
(10.2)

This supply curve of labour is depicted in Fig. 10.2.

Fig. 10.2. Employment and optimal wage

Given a wage of G, the maximum number that can be employed, which may be labelled ‘potential employment’, is given by X/G. The relation between G and potential employment is depicted in Fig. 10.2. Clearly this is a rectangular hyperbola. What is being recommended here is that G should be minimized subject to L(G) > X/G. Let the solution of this be defined by G *. This is easily seen to be given by the point of intersection of the two curves in Fig. 10.2. G *, it is being argued here, is the wage that we should aim to offer.47 (p.367)

As I have argued in Basu (1981) on the basis of some Planning Commission data (Project Evaluation Organization 1979) it seems likely that the wage that has been paid in India is above G *. A similar claim seems to be possible for Bangladesh on the basis of Ahmad and Hossain's paper.

It has already been argued that G * is more equitable (in the sense of ameliorating poverty) than a higher wage. G * has another advantage. It sharpens the self‐selection property of FFW since from (10.1) it follows that, as G becomes smaller, the wealthier (in terms of labour income) will be less inclined to come for FFW jobs, thereby paving the way for the poor to take these up.

The above analysis implies that there are three ways of raising the wage: (1) by improving the opportunities open to the labourers (by, for example, having infrastructural investments in the rural sector. This would raise the supply curve in Fig. 10.2). (2) By assigning a larger food stock for distribution through FFW. This raises the ‘potential employment’ curve. (3) By simply deciding to set G above G * and maintaining an excess supply of labour. What I have argued above is that we should raise wages via methods (1) and (2); option (3) ought not to be normally used. It may be used only if we can devise some additional selection criterion whereby the poorest among the job applicants are selected. Here and elsewhere it ought to be kept in mind however that accurate targeting of benefits may itself be an expensive exercise and there may be times when it is suboptimal to perfect one's marksmanship in this respect.48

Finally, I make some brief comments on the form of wage payment. Should this be paid in cash or kind? In other words, should we have an FFW programme or a cash‐for‐work programme? I do not want to give a firm answer (p.368) here but merely point out the pros and cons. The popular argument for paying in food grains is to encourage the poor to have better nourishment even if that is not what they would do with the same amount of income. The ethical strength or weakness of this will not be discussed here. I will simply evaluate the positive argument that underlies this prescription, namely, that payment in foodgrains ensures that people will not fritter away their income on ‘useless’ consumption.

The weakness of this proposition becomes apparent as soon as we recognize that money is fungible. To make the proposition as strong as possible before attacking it let me assume that people cannot sell off foodgrains earned as wages (perhaps because of the high transaction costs of selling such small quantities). Nevertheless, as soon as we grant that most people will have some income from other sources (i.e. other than what they earn from FFW), it follows that the portion of that other income which they would have spent on food they can now spend on other things given that their food supply now comes from the FFW project. In other words, the fact of paying wages in food does not mean that food consumption will go up by that amount and it may not in fact go up at all. A little calculation shows that, given the actual facts of FFW in India (see Basu 1981, it is the case that the form of payment is unlikely to affect the volume of consumption of food. The evidence from Bangladesh suggests that payment in wheat does not affect significantly the volume of foodgrain consumption, but it does tilt its composition in favour of wheat (Osmani and Chowdhury 1983).

An important case for payment in food stems from the recognition of some political constraints, which apply especially to India. In the Indian economy, thanks to several compulsions including those emanating from the agrarian lobby, it seems likely that the food procurement policies of the government will continue for some time. If wages are paid in foodgrain at rural works programmes then the procured grain—or at least a part of it—will get earmarked for use in FFW. If, on the other hand, wages are paid in cash, the government will soon be looking for avenues for selling this grain. Once the grain gets converted into cash, it is unlikely that it will be used for the poor. Given the forces around government—and this is the central point of Griffin's paper (Griffin 1985)—it is more likely to get spent on amenities for the urban middle class. One may object here that if a government is inclined to do so, it will do so anyway. But the logic of large organizations like government does not work that way. The same ‘end’ may be infeasible by one route but by another route it may just be possible to tiptoe past vested interests.

The subject of political constraints has been mentioned several times, but in passing. It is an important issue in food policy but also a difficult one to write about; and so I reserve only a brief section for it.

# (p.369) 10.4. Optimal policies and political constraints

Keith Griffin (1985) has pointed out in a lucid essay on the state of poverty in Asia that the problem of poverty can be largely eliminated through ‘purposeful government intervention’; but ‘given the class basis of the state, a question arises about the possibility of effective action’. He argued correctly that, in most South Asian countries, governments are closely in alliance with the wealthier sections of society. These people therefore guide policy to their advantage. The government's rhetoric is meant for the poor who are not near enough to see the actual policies, and the actual policies are inclined to help the already wealthy who keep a close watch on them. More disturbing is Gunnar Myrdal's (1970) accurate observation that not only policies but even the perception of truth can deviate in favour of the dominant classes.

There is a danger in pushing this kind of an argument too far. Though such a charge cannot be levelled against Griffin or Myrdal, several radical and conservative writers have been too successful in explaining the backwardness of LDCs, in the sense that their models leave no chinks for changing the situation. If such an all‐encompassing explanation was valid, then there would be no point in making policy recommendations.

More formally, what I am claiming is that if a model of an economy is constructed in which every agent knows what is in his own interest and acts to achieve it, then ‘advice’ can have no role in affecting the ‘outcome’ of such an economy. This is the case with the Arrow–Debreu model of general equilibirium. There is no provision here for ‘advice’ to change any agent's behaviour. Let us call an agent whose actions can be altered by some advice an ‘open end’ of a model. If we want to construct a model and make recommendations, the model must be one which has open ends. In some conventional models the government is treated as an open end. If however we treat the government as either completely subservient to certain class interests or comprising of fully informed individuals with well‐defined objectives, we lose the open end and policy advice becomes redundant again. In reality there is scope, however slender, for influencing the outcomes of economies.

First note that people do not always act selfishly and our hope for more equitable policies lies in the fact that altruism, anger, envy, or the influence of writings can lead us to work against our own interest. Secondly, the rich constitute a group. Consequently, even though a particular policy may be in their own interest, they may fail to implement it because their individual rationality may be incompatible with their group rational behaviour. In short, within their own class they may interact as in the Prisoner's Dilemma.

Finally, even though governments do exhibit class bias, there are chinks in the system, which can be exploited to divert resources to the poor. A government or a Minister chooses his policy bundle subject to several constraints. Some of these are, however, unusual constraints in that they may be political in character: certain powerful lobbies could make some policies (p.370) impossible to implement. How much of a government's preference or ideology is revealed in its choice of policies depends on how restrictive these constraints are. Given the several lobbies that surround South Asian governments49 the room for manœuvre must indeed be small. Hence two Finance Ministers with very different ideologies would draft budgets which are not too different (otherwise at least one of them would cease to be Finance Minister). In brief, South Asian economies are Finance Minister neutral—or almost so. This makes it clear that if government policies are to be influenced, instead of trying to alter the government's preference, our aim should be to alter the constraints. The poorer sections have to be made more demanding and more conscious of their rights.

There is an interesting self‐referential reason why some downtrodden groups are so dormant. Note first that if all members of a group, H, consider it unreasonable to demand a larger share of the cake then, given that a government is always under severe pressure to make concessions to different groups lobbying for more, it will indeed be reasonable not to give H a larger cut. After all H is not pressing for more and others are. Hence each individual member of H will be right in supposing that it is unreasonable to ask for a larger share. If however each member of H chooses to be unreasonable and demands more, the government may find it in its own interest to concede to the group's demand. Also, the government would be able to persuade the other lobbies that it has to concede to H, because of its large ‘voice’. In other words, it would no longer be an unreasonable demand to ask for more.50 This is where the media can play a major role. They can influence our view of what is acceptable. And, as has just been argued, what is acceptable depends, at least in part, on what we consider to be acceptable.

# 10.5. Concluding remarks

The aim of this chapter was to evaluate policies for combating persistent and mass poverty in South Asia. In so doing it was necessary to go into some general analytical questions and to comment on existing debates. Though most of the empirical discussions were based on India and Sri Lanka, it is hoped that the analysis and policy suggestions which emerged will be of interest for the larger problem of poverty in the Third World.

The chapter began with the question as to whether ‘direct action’ for mitigating the poverty problem could be fruitful or would it be necessary to wait for the benefits of growth to ‘trickle down’. It was argued that direct action (p.371) was possible and desirable. To pursue this objective it was not necessary to sacrifice growth. Different policy instruments are available for achieving higher growth and greater equity. Given the enormity of this problem my comments were, perforce, in the nature of an overview. Several subthemes would have to be the subject of much more detailed research. The present chapter pursued in greater depth one particular subtheme—that of poverty alleviation programmes in rural South Asia.

Several schemes, for example, food‐for‐work and the Integrated Rural Development Programme, were commented on. However, it ought to be mentioned that, while I have analysed the schemes separately, in designing a full anti‐poverty policy it will be essential to evaluate the whole package together. This is because poverty has many dimensions51 and it will be necessary to use more than one programme to mitigate poverty,52 and the value of one particular programme may depend on what else is being implemented. Some complementarities, for example, that between food‐for‐work and food rationing, were discussed in this chapter but a more systematic evaluation of packages will have to be undertaken in future.

The last part of the chapter was devoted to the politics of anti‐poverty programmes. The brevity of section 10.4 reflects the difficulty of the subject, not its unimportance. Political constraints do not necessarily arise out of the wilful machinations of groups and lobbies but could be the inadvertent consequence of a multitude of individuals, each acting atomistically and in his own interest. This makes political constraints difficult not only to overcome but even to understand; and this subject must loom large in any agenda for research on poverty.

(p.372) References

Bibliography references:

Ahluwalia, M. S. (1974), ‘The Scope for Policy Intervention’, in Chenery et al. (1974).

———(1976), ‘Inequality, Poverty and Development’, Journal of Development Economics, 3.

Ahmad, A. K., and Hossain, M. (1985), ‘An Evaluation of Selected Policies and Programmes for the Alleviation of Rural Poverty in Bangladesh’, in Islam (1985).

Ahmed, H. (1986), ‘Rural Landless in Bangladesh: An Enquiry into the Economic Results of Grameen Bank’, mimeo (Oslo: Institute of Economics).

Bandyopadhyay, D. (1985), ‘An Evaluation of Policies and Programmes for the Alleviation of Rural Poverty in India’, in Islam (1985).

Bardhan, P. K. (1984a), Land, Labour and Rural Poverty: Essays in Development Economics (New York: Columbia University Press).

———(1984b), The Political Economy of Development in India (Oxford: Basil Blackwell).

Basu, A. M. (1986), ‘Birth Control by Assetless Workers in Kerala: The Possibility of a Poverty‐Induced Fertility Transition’, Development and Change, 17.

Basu, K. (1981), ‘Food for Work Programmes: Beyond Roads that Get Washed Away’, Economic and Political Weekly, 16.

———(1982), ‘Food for Work: Some Economic and Political Consequences’, Economic and Political Weekly, Review of Agriculture, 17.

———(1984a), The Less Developed Economy: A Critique of Contemporary Theory (Oxford: Basil Blackwell).

———(1984b), ‘Implicit Interest Rates, Usury and Isolation in Backward Agriculture’, Cambridge Journal of Economics, 8.

———(1986), ‘One Kind of Power’, Oxford Economic Papers, 38.

Bhaduri, A. (1977), ‘On the Formation of Usurious Interest Rates in Backward Agriculture’, Cambridge Journal of Economics, 1.

Bhagwati, J. N. (1985), ‘Growth and Poverty’, Occasional Paper No. 9(East Lansing, Mich.: Center for Advanced Study of International Development, Michigan State University).

Bhalla, S. S. (1988a), ‘Is Sri Lanka an Exception? A Comparative Study in Living Standards’, in Srinivasan and Bardhan (1988).

———(1988b), ‘Sri Lanka's Achievements: Fact and Fancy’, in Srinivasan and Bardhan (1988).

———and Glewwe P. (1986), ‘Growth and Equity in Developing Countries: A Reinter‐pretation of the Sri Lankan Experience’, World Bank Economic Review, 1.

Bhende, M. J. (1986), ‘Credit Markets in Rural South India’, Economic and Political Weekly, Review of Agriculture, 21.

Caldwell, J. C. (1986), ‘Routes to Low Mortality in Poor Countries’, Population and Development Review, 12.

Centre for Development Studies, Trivandrum (1975), Poverty, Unemployment and Development Policy: A Case Study of Selected Issues with Reference to Kerala (New York: United Nations ST/ESA/29).

Chenery, H., Ahluwalia, M. S., Bell, C. L. G., Duloy, J. H., and Jolly, R. (1974), Redistribution with Growth (Oxford: Oxford University Press).

(p.373) Dandekar, K. (1983), Employment Guarantee Scheme: An Employment Opportunity for Women (Bombay: Orient Longman).

———and Sathe, M. (1980), ‘Employment Guaranteee Scheme and Food‐for‐Work Program’, Economic and Political Weekly, 15.

Dandekar, V. M. (1986), ‘Agriculture, Employment and Poverty’, Economic and Political Weekly, Review of Agriculture, 21.

Dantwala, M. L. (1985), ‘“Garibi Hatao”: Strategy Options’, Economic and Political Weekly, 20.

Datta Chaudhuri, M. K. (1981), ‘Industrialization and Foreign Trade: The Development Experiences of South Korea and the Philippines’, in Lee (1981).

Dholakia, B., and Dholakia, R. (1980), ‘State Income Inequalities and Inter‐state Variations in Growth of Real Capital Stock’, Economic and Political Weekly, 15.

Edirisinghe, N. (1986), ‘The Food Stamp Program in Sri Lanka: Costs, Benefits and Policy Options’ (Washington, DC: IFPRI).

Fields, G. S. (1980), Poverty, Inequality and Development (Cambridge: Cambridge University Press).

George, P. S. (1979), ‘Public Distribution of Foodgrains in Kerala: Income Distribution Implications and Effectiveness’ (Washington, DC: IFPRI).

Griffin, K. (1985), ‘Rural Poverty in Asia: Analysis and Policy Alternatives’, in Islam (1985).

Herring, R. J., and Edwards, R. M. (1983), ‘Guaranteeing Employment to the Rural Poor: Social Functions and Class Interests in the Employment Guarantee Scheme in Western India’, World Development, 11.

Hirway, I. (1985), ‘“Garibi Hatao”: Can IRDP Do It?’ Economic and Political Weekly, 20.

Hossain, M. (1985), ‘Institutional Credit for Rural Development: An Overview of the Bangladesh Case’, Bangladesh Journal of Agricultural Economics, 8.

Isenman, P. (1980), ‘Basic Needs: The Case of Sri Lanka’, World Development, 8.

Islam, R. (ed.) (1985), Strategies for Alleviating Poverty in Rural Asia (Bangkok: ILO).

———and Lee, E. (1985), ‘Strategies for Alleviating Poverty in Rural Asia’, in Islam (1985).

Jayanntha, D. (1985), ‘Sri Lanka: The Political Framework (1947–84)’, mimeo (Washington, DC: World Bank).

Jayawardena, L. (1974), ‘Redistribution with Growth: Some Country Experience—Sri Lanka’, in Chenery et al. (1974).

Khan, A. R. (1972), The Economy of Bangladesh (London: Macmillan).

Kumar, S. K. (1979), ‘Impact of Subsidized Rice on Food Consumtpion in Kerala’ (Washington, DC: IFPRI).

Lee, E. (ed.) (1981), Export Led Industrialization and Development (Bangkok: ILO).

Lee, E. L. H. (1977), ‘Rural Poverty in Sri Lanka, 1963–73’, in International Labour Organization, Poverty and Landlessness in Rural Asia (Geneva: ILO).

Lipton, M. (1983), ‘Poverty, Undernutrition and Hunger’, World Bank Staff Working Paper No. 597 (Washington, DC: World Bank).

McHenry, D. F., and Bird K. (1977), ‘Food Bungle in Bangladesh’, Foreign Policy, 27.

Morrison, B. M., and Waxler, N. E. (1986), ‘Three Patterns of Basic Needs Distribution within Sri Lanka: 1971–73’, World Development, 14.

(p.374) Mukhopadhyay, S. (ed.) (1985a), The Poor in Asia: Productivity‐Raising Programmes and Strategies (Kuala Lumpur: Asian and Pacific Development Centre).

———(ed.) (1985b), Case Studies on Poverty Programmes in Asia (Kuala Lumpur: Asian and Pacific Development Centre).

Myrdal, G. (1970), The Challenge of World Poverty (London: Allen Lane).

Nag, M. (1983), ‘Impact of Social and Economic Development on Mortality: Comparative Study of Kerala and West Bengal’, Economic and Political Weekly, 28.

Office of the Registrar‐General of India (1981a), Survey on Infant and Child Mortality 1979 (New Delhi: Government of India).

———(1981b), Levels, Trends and Differentials in Fertility 1979 (New Delhi: Government of India).

Osmani, S.R., and Chowdhury, O.H. (1983), ‘Short Run Impacts of Food For Work Programme in Bangladesh’, Bangladesh Development Studies, 11.

Panda, M. K. (1981), ‘Productivity Aspects of Wages in Food for Work Programme’, Economic and Political Weekly, 16.

Panikar, P. G. K., and Soman, C. R. (1984), Health Status of Kerala (Trivandrum: Centre for Development Studies).

Project Evaluation Organization, Planning Commission (1979), A Quick Evaluation Study of Food for Work Programmes (New Delhi: Government of India).

Rahman, A. (1979), ‘Usury Capital and Credit Relations in Bangladesh Agriculture: Some Implications for Capital Formation and Capitalist Growth’, Bangladesh Development Studies, 7.

Rath, N. (1985), ‘“Garibi Hatao”: Can IRDP Do It?’ Economic and Political Weekly, 20.

Reutlinger, S. (1984), ‘Project Food Aid and Equitable Growth: Income Transfer Efficiency First!’ World Development, 12.

Richards, P., and Gooneratne, W. (1980), Basic Needs, Poverty and Government Policies in Sri Lanka (Geneva: International Labour Office).

Rodgers, G. B. (1976), ‘A Conceptualisation of Poverty in Rural India’, World Development, 4.

Scitovsky, T. (1985), ‘Economic Development in Taiwan and South Korea: 1965–81’, Food Research Institute Studies, 19.

Sen, A. K. (1981a), ‘Public Action and the Quality of Life in Developing Countries’, Oxford Bulletin of Economics and Statistics, 43.

———(1981b), Poverty and Famines: An Essay on Entitlement and Deprivation (Oxford: Oxford University Press).

———(1985), Commodities and Capabilities (Amsterdam: North‐Holland).

———(1988), ‘Sri Lanka's Achievements: How and When?’, in Srinivasan and Bardhan (1988).

Siddiqui, K. (1985), ‘An Evaluation of Grameen Bank Operations’, in Mukhopadhyay (1985b).

Srinivasan, T. N., and Bardhan, P. K. (ed.) (1988), Rural Poverty in South Asia (New York: Columbia University Press).

Streeten, P. (1984), ‘Basic Needs: Some Unsettled Questions’, World Development, 12.

Sundaram, K., and Tendulkar, S. D. (1985a), ‘Anti‐poverty Programmes in India: An Assessment’, in Mukhopadhyay (1985a).

(p.375) ——— ———(1985b), ‘Integrated Rural Development Programme in India: A Case Study of a Poverty Eradication Programme’, in Mukhopadhyay (1985b).

Taylor, C. L., and Hudson, M. C. (1972), World Handbook of Political and Social Indicators (New Haven, Conn., and London: Yale University Press).

World Bank (1985), World Development Report (New York: Oxford University Press). (p.376)

## Notes:

I am grateful to T. C. A. Anant, Elias Dinopoulos, Jean Drèze, Keith Griffin, Subbiah Kannappan, Sunil Sengupta, and Susan Watkins for discussions. I thank Surjit Bhalla for the unpublished papers and monographs on Sri Lanka he has sent me, though I am sure he will disapprove of the use to which these have been put. This work was done at the Institute for Advanced Study, Princeton. It would have been difficult to complete it without the excellent support provided by the Institute.

(1) The definition of ‘basic needs’ can be controversial. As Streeten (1984: 973–4) notes, ‘The ILO considers employment a basic need; Sidney Webb included leisure. High on the list, as China recognized in the six guarantees, is a decent funeral …’ In common parlance basic needs or standard of living is taken to mean the availability of basic economic necessities (food, shelter, clothing, etc.) and minimum standards of health and medical facilities (captured by demographic data on nutrition, life expectancy, etc.). I take advantage of these conventions and escape having to provide exact measures and definitions. It is arguable that ‘basic needs’ should also include indices of political and social ethos. However, the absence of commonly accepted statistics on these makes them too difficult for use in any meaningful discussion. For those who are overly enthusiastic about introducing indices of political and social climate to evaluate ‘basic needs’ policies, a reading of how such indices are computed could be sobering: see Taylor and Hudson (1972).

(2) There are other such examples outside South Asia. These include Costa Rica, Cuba, and China.

(3) See Centre for Development Studies (1975), Fields (1980), Isenman (1980), Richards and Gooneratne (1980), Sen (1981 a, 1985), Nag (1983), Bhalla and Glewwe (1986), Morrison and Waxler (1986).

(4) All the data cited in this and the next two paragraphs, as well as some additional ones, are presented in Table 10.1. All these statistics, except where specified to the contrary, refer to 1983.

(5) The reason I have chosen secondary school registration data is because this indicates completion of at least the primary level of education. Primary school registration on the other hand may not be sufficiently informative because of high drop‐out rates. For example, in UP initial school enrolment in 1965–6 was much higher than in Kerala whereas towards the end of primary school the situation was completely reversed (Centre for Development Studies, 1975: 122).

(6) One variable which is often taken to indicate living standards and is omitted here is fertility. The usual presumption is that in a poor country a drop in fertility rates reflects improved living standards. A careful scrutiny of Kerala's demographic data, however, seems to jeopardize this presumption. It has been shown (A. M. Basu 1986) that the sharper drop in fertility rates in Kerala has occurred among the landless labourers, suggesting that poverty is sometimes as likely to induce fertility declines as are improved living standards.

(7) Though I used only a few countries and a few variables to demonstrate this, there are more systematic studies which have established it (Isenman 1980; Sen 1981 a).

(8) In the absence of this a second‐best option is to confine the study to the early 1960s (instead of 1960–78) because there is some evidence that welfare expenditure was increasing in Sri Lanka in the late 1950s. We could therefore argue that the economy in the early 1960s reflected this increase (which we may suppose was sharper in Sri Lanka than in other countries).

(9) Richards and Gooneratne (1980).

(10) It also probably influences fertility. The World Development Report, 1985, notes that the percentage of married women of child‐bearing age using contraceptives in Sri Lanka in 1982 was 55. This compares favourably not only to Bangladesh's 25 but to India's 32.

(11) This has been largely dismantled in recent years, a point that is discussed later.

(12) It is commonly believed that Sri Lanka's income distribution in 1973 was more equitable than the ones in 1963 or 1953. It has in fact been shown (Fields 1980: 198) that the distribution in 1973 Lorenz‐dominates the earlier ones. These facts run into controversy once we look at consumer expenditure data. It can be shown that in terms of this the share of income accruing to the bottom one‐fifth of the population fell and the share going to the highest one‐fifth rose between 1963 and 1973 (Lee 1977). Lee has argued convincingly that because of sharp relative price changes and a consequent index‐number problem, the income data is rendered a less accurate indicator of what happened between 1963 and 1973 than the expenditure data.

(13) The Keralite's general penchant for education is also well known. This partially explains the recent mushrooming of private English‐medium schools in Kerala. As a cynical bureaucrat noted, in Kerala you have to simply put up the label ‘English Medium’, think of a good English name, like John or Mary, prefix it with a ‘Saint’, and you are in business.

(14) The percentage of population within two kilometres of medical facilities in 1978 is 64 for Kerala compared to the Indian average of 35 (Office of Registrar‐General 1981 a). Percentage of live births unattended by trained medical practitioners in 1978 in Kerala is 38% in the rural sector and 25% in the urban sector. The all‐India figures for the same are 42 and 33 (Office of Registrar‐General 1981 b).

(15) See Lee (1981) and in particular Datta Chaudhuri's (1981) essay in that collection. For a very interesting analysis contrasting South Korea and Taiwan, see Scitovsky (1985).

(16) There are several other parties of varying importance and governments have generally been coalitions, but such details are being ignored here.

(17) This is so excepting for a brief period in 1960 when, after a general election in which there were no clear victors, Dudley Senanayake of the UNP was appointed Prime Minister.

(19) From full coverage of the population, in 1972 Mrs Bandarnaike ruled that income tax payers would not get rationed food. The effect of this was negligible (Edirisinghe 1986).

(20) The full consequence of this cannot as yet be judged but, based on some preliminary evidence, Jayanntha (1985: 47) has noted in his lucid monograph on Sri Lankan politics, ‘Available evidence suggests that the nutritional status especially of infants and children under two years has deteriorated significantly.’ Regarding why this did not erode the popularity of the UNP government (judging by some by‐election results), he observes, ‘These groups though vulnerable are not articulate or politically organized. Moreover malnutrition is often not recognized by the mother … Thus increased malnutrition alone … may not be translated into overt forms of political discontent.’

(21) Bhalla (1988 a).

(22) See Jayanntha (1985) for an excellent account.

(23) Richards and Gooneratne (1980).

(24) Bhalla (1988 a).

(26) There are important exceptions to this, as just discussed in terms of Sri Lanka's experience between 1971 and 1977.

(27) Scitovsky's (1985) study shows that Taiwan is ahead of Korea in terms not only of equity and standards of living but also (albeit marginally) of growth and economic performance. What comes out of this study is that Taiwan's growth rate instead of having been curtailed by its greater equity was probably aided by it. For a detailed cross‐country study of growth and equity, see Ahluwalia (1976).

(28) I have discussed some of these issues in several short articles in the Indian Express and in a longer essay in the Statesman of 5 Nov. 1985.

(29) As Ahluwalia (1974) emphasizes, the very purpose of a support programme is to exercise selectivity.

(30) Also omitted from the purview of the present paper are more structural reforms like land redistribution. Programmes like the EGS and FFW do not usually require structural changes and, as Herring and Edwards (1983) rightly observe, they may even help ossify existing structures.

(31) For critical assessments, see Sundaram and Tendulkar (1985 a, 1985b) and Bandyopadhyay (1985). A general survey of poverty programmes in South Asia is contained in Islam and Lee (1985).

(32) In general the ‘very poor’ tend to raise problems quite distinct from the ones associated with the poor. Can the very poor be helped so as to become permanently self‐sufficient? Are they in a position to avail themselves of the benefits provided for the poor? Such questions need investigation (see Lipton 1983).

(33) The problem seems to be similar for Bangladesh (Khan 1972; Rahman 1979).

(34) The use of personalized relationship and collateral has been discussed in Bhaduri (1977) and Basu (1984 a, 1984b).

(35) This, as is well known, may be long enough to cover the lifetime of the borrower and may even spill over to his children who would then be born into bondage.

(36) See Ahmed (1986) for a discussion of the prospects of the Grameen Bank and the impact of credit availability on employment. He argues that in the long run technology upgradation will be necessary for fully realizing the benefits of a credit programme.

(37) Contrary to the official proclamation, this last ‘objective’ is not really an objective. The availability of surplus grain simply makes it easier to fulfil the real objectives of employment and income transfer and production.

(38) FFW or some variant of it has been used in several LDCs, e.g. Tunisia, Morocco, and Egypt. Tunisia is also one of the few countries where FFW began without foreign support, though within a few years food shortages developed and it had to use US wheat.

(39) India's FFW had an important precursor in the EGS which had been in operation since 1972 in Maharashtra. It was merged with the FFW programme when the latter was started. For a study of the EGS, see Dandekar (1983).

(40) The original organization structure of the FFW was as follows. The Centre supplied foodgrains to the State governments, which were responsible for setting up labour intensive projects in the rural regions. The States had considerable freedom in terms of the actual execution of projects. They could, for instance, pay wages purely in terms of foodgrain or pay partly in foodgrains and partly in cash.

(41) The food stamp programme, as implemented in Sri Lanka, is something in between a free rice ration scheme and a negative income tax. A recipient of food stamps can exchange them for a certain range of food items, including rice. Members of households which earn less than Rs 300 per month were eligible. Adults received stamps worth Rs 15 per month and children a little more (see Edirisinghe 1986).

(42) I am assuming that a day's labour is an indivisible unit. That is, a worker does not face a choice between hours of leisure each day and daily income.

(43) Dandekar and Sathe (1980); Basu (1981, 1982); Panda (1981).

(44) I argued this in Basu (1981) and the next few paragraphs draw heavily on that paper.

(45) See Bandyopadhyay (1985: 137) concerning underpayment in FFW in India.

(46) For a critique of such a view of poverty see Sen (1988: ch. 3 and Appendix C).

(47) The actual execution of this may not be as easy as it appears. In Afghanistan, wages were set so low, in an effort to maximize the spread, that the projects were perennially short of labour. In Lesotho the wage was set so high that landowners were quitting working on their own land to work at FFW sites.

(48) Reutlinger (1984) has suggested a method of computing the relative efficiency of giving aid in the form of different goods, which takes into account the delivery cost. A similar exercise should be possible whereby the efficiency of transferring income to the poor via different schemes is computed, taking into account the cost of organizing the schemes.

(49) Bardhan (1984 b) has argued that Indian policy making is a compromise of the interests of three dominant classes: the industrial bourgeoisie, wealthy farmers, and the bureaucracy.

(50) Using the idea of self‐fulfilling conjectures I have tried to show in a different context (Basu 1986) how certain unwanted political power structures could be sustained by a web of reinforcing beliefs among individuals.

(51) See Rodgers (1976) for a discussion of the concept of poverty viewed as a ‘multivariate phenomenon’.

(52) There is a semantic problem here in that we may think of a combination of programmes as yet another programme. This would render the claim that one programme can never be sufficient erroneous. The defence against this criticism is that the expression (one programme) is being used here to describe one member of the existing menu of anti‐poverty schemes.