Home

 ABSTRACT

Professor Amartya Kumar Sen was awarded the Nobel Prize for economics in 1998 in recognition of his "several key contributions to the research on fundamental problems in welfare economics", as noted in the Nobel citation. This paper examines three major areas of economics, which it calls "unfashionable economics", and in which Sen’s contributions have been particularly notable. These areas are inequality, poverty, and hunger and famine. The paper argues that Sen’s pioneering research in these areas has not only helped to resolve some theoretical and/or policy issues, it has also made a significant contribution to generating public interest in the problems that face some of "the most impoverished members of society", in the words of the Nobel citation. In addition to putting Sen’s work in the context of established thinking in the relevant areas, the paper demonstrates how his contributions have helped to improve our understanding of the issues involved, and how such advances have influenced policy-making. To make the paper accessible to the interested non-specialist, the paper uses a style of exposition that is less technical even where the issues involved are of a highly technical nature.

JEL Classification: A11, A13, I32.

In October 1998, the Royal Swedish Academy of Sciences awarded the Nobel Memorial Prize in Economics to Professor Amartya Kumar Sen in recognition of his "several key contributions to the research on fundamental problems in welfare economics"(Nobel Citation, quoted by the editors of the Scandinavian Journal of Economics 101(2), 1999). The citation goes on to record that

The Nobel Economics Prize, established in 1968, was first awarded in the following year. To date, it has been awarded to 44 economists twenty of whom have been single (non-shared) winners. As is to be expected, the range and the depth of contributions of the winners of this coveted award are varied and outstanding. But, even so, the contributions Sen has made are distinctive in several ways. For example, he has challenged what may be termed the orthodoxy of mainstream economics wherein the "rational individual" is supposed to be engaged in maximising his or her personal utility (satisfaction) subject to a budget constraint. And social welfare is perceived as the sum total of individual utilities. There is no place in this line of utilitarianism for socially-based sentiments such as ‘sympathy’ or ‘commitment’ which often form part of an individual’s motivation to action even in an economic context (Sen 1977).

An aspect of utilitarianism is that it judges the merit of a particular state of affairs exclusively in terms of individual utilities in the state of affairs in question. Sen calls this welfarism, and he rejects it as too narrow. Pursuit of rationality in this orthodox sense, Sen points out, implies that the utility-maximising individual has but

If such rational agents existed, they should qualify as ‘rational fools’ in Sen’s telling terminology. Social welfare should be based on more than just individual utilities defined in the narrow sense of the individual’s command over goods and services, contends Sen. He went on to develop the notion of capabaility which

This notion is conceptually much wider than welfarism, and has significant policy implications.

Another side of the traditional notion of rationality involves the acceptance of the efficiency of the free market. The analytical elegance of the competitive model, and the existence of Walrasian equilibrium form the base of Pareto-efficient allocations which help maximise output and minimise costs. Even within traditional economic thinking, the efficiency of the free market is constrained by the existence of externalities or absence of some of the features of perfect competition. But Sen has gone beyond these to raise fundamental questions about some of the assumptions underlying this model. One of these is whether it is correct to make the (implicit) assumption that all agents in these markets are able to survive even when they are not engaged in the market-based exchanges. In cases where this ‘survival’ assumption is violated, the efficiency of the competitive market is no longer guaranteed, and appropriate forms of government intervention may be necessary to ensure complete survival. This issue has been discussed illuminatingly by Coles and Hammond (1995) who, after an extensive review of the survival aspect of Walrasian equilibrium and Pareto efficiency, observe

Sen has also pointed out how individual actions are often socially determined, and how individuals acting in isolation may often produce a suboptimal result which can be improved upon by collective action (Sen 1967, 1984). His work on ‘isolation paradox’ is another example of the individual-collective dichotomy (such as the Keynesian paradox of thrift) which the market may not resolve in an efficient manner (Bagchi 1998, Patnaik 1998), but either an appropriate government intervention or cooperative action by individuals can.

If these ideas constitute a rebuttal of some widely held views in mainstream economic theory, Sen’s quest for analytic rigour and policy-guidance has led him to raise searching questions and to challenge accepted ideas in areas that affect human well-being in major ways. This paper looks at three such areas: poverty, inequality and famine. The choice of these areas is prompted by our belief that they help to establish the importance of applying economic theory to not only understand some pressing problems in societies, but also find possible solutions to them. We have called this ‘unfashionable economics’ because mainstream economics, arguably, had not addressed these issues with such rigour - indeed passion - as Sen has done.

In section 2 below we examine Sen’s work on inequality which then leads on to section 3 where the related area of poverty is examined. Section 4 concentrates on Sen’s work on the availability of food and famine. A general overview of Sen’s influence as an economist and a social scientist is presented in section 5. The paper ends on the concluding observations of Section 6.

A word about the style of our presentation in this article: an outstanding theoretical economist that Sen is, he also has the extraordinary gift of communicating the most abstract and technical of economic ideas to the uninitiated in an effective, even interesting, way. In the same spirit, but, of course, with considerably less ability, we have endeavoured in this article to be largely non-technical even though many of the issues we are addressing here are of a highly technical nature.

The concept of inequality has a positive and a normative dimension. Statisticians and economists have long struggled to capture (a) how a given total income (wealth) is actually distributed over a given population, and (b) how a redistribution of that income (wealth) is to be judged from a welfare point of view. The former is an issue which is largely in the domain of positive economics, while the latter involves normative judgements.

If the cumulative percentage of a population is graphed from the poorest to the richest along the horizontal axis, and the cumulative percentage of income received by the bottom x percent of the population up the vertical axis, one obtains the Lorenz curve which, typically, has a convex-from-below shape. Had income been distributed absolutely equally so that each percentile of the population received the same percentage share of the total income, the Lorenz curve would coincide with the diagonal between the population and income axes. The farther away the actual Lorenz curve is from this line, the greater is the inequality in the distribution of income. This notion is captured in the Gini coefficient of income distribution which is measured as twice the area between the diagonal and the Lorenz curve. It implies that the farther away a Lorenz curve is from the diagonal, the higher the value of the Gini coefficient.

The Gini coefficient is a widely used positive measure of inequality; its strength deriving in part from the fact that it is independent of the mean income and the population size. It also satisfies the Pigou-Dalton condition which requires that any transfer from a rich to a less rich income recipient must reduce the Gini coefficient (by shifting the Lorenz curve upwards between the relevant centiles).

While the positive measures of inequality are helpful in indicating the nature of a given distributional arrangement, it is the social welfare implications of these arrangements that are perhaps more interesting and useful from a policy point of view. The Pigou-Dalton condition, for example, imposes a rather normative requirement on a ‘good’ positive measure of inequality, viz. the Gini coefficient. The basic idea underlying the Pigou-Dalton condition is rooted in the notion that it is the social welfare potential of one distributional arrangement over another that a measurement ought to be capable of capturing. An intuitively appealing way to capture the message of a utility-based measure of inequality such as the one implied in the Pigou-Dalton condition is to compare an actual distribution with the equal distribution, and view inequality as being the distance between the two.

This notion of inequality was revived more recently by Kolm (1966) and Atkinson (1970). Atkinson demonstrated that, with identical mean incomes and population sizes of two groups, if the Lorenz curves of their income distributions are non-intersecting, then the group with the Lorenz curve nearer to the line of absolute equality will enjoy a higher social welfare. Thus, Atkinson views lower inequality as higher welfare, and the ranking of distributional states is provided by Lorenz dominance, i.e the positions of the Lorenz curves vis-à-vis one another. This result is based on the assumptions of additively separable individual utility functions which are increasing, concave, and identical for all individuals.

2. Sen’s Generalisation of Atkinson’s Inequality Measure

While accepting the usefulness of the welfare-based normative measures of inequality, Sen has criticised Atkinson’s measure as being restrictive because of its requirement of additive separability. Dasgupta, Sen and Starrett (1973) demonstrated that, to satisfy the Pigou-Dalton condition, the social welfare functions need only to be non-decreasing, s-concave functions. This is a weaker requirement than concavity or even quasi concavity.

To explore further the nature of Sen’s critique of Atkinson’s measure, it would be helpful to explain the latter first. Atkinson’s inequality measure can be defined as:

(1)

where m_E is called the equally distributed equivalent level of income. This is the average level of income, which, if allocated to every individual, would yield the same level of welfare as the existing distribution with the average income . The welfare function implicitly

underlying m_E is:

(2)

assuming that the preference of all individuals are alike, and that the utility function of individual i, U_i, is concave in i’s income, x_i, but not necessarily strictly concave.

The specification of the welfare function is obviously based on the value judgement of the person specifying it. Atkinson restricts the class of welfare functions by using some concepts from the literature on risk aversion. He conceives the utility function as:

(3)

where ensures the concavity of the function , and it represents the weight attached to the inequality by society. If is zero, society is indifferent between distributions, and as the value of e increases, society’s concern for the poor increases.

In critiquing this approach, Sen considers two situations involving the distribution of a given total income, 10, between two persons. Let these be (0,10) and (5, 5). If a utility function is such that it is proportional to income, i.e. the marginal utility of income is constant, then both distributions will yield the same Atkinson index of inequality. Sen’s criticism of Atkinson’s approach is thus directed at the latter’s implied utilitarian form. Sen puts this succinctly in the following quote:

Sen generalises the Atkinson measure by considering social welfare as an increasing

function of individual incomes (thus avoiding the problem of an individualistic social welfare function). Such a function is further assumed to be symmetric and quasi-concave, and it is of the form:

(4) W = W(x₁,x₂ …., x_n)

Sen defined a generalised equally distributed equivalent income as

If this is called , with (4) above being the assumed social welfare function, then for every distribution of income. Therefore the measure of inequality can be written as

5. S = 1-

A and S would be completely equivalent if an additive utility function is assumed.

S is a normative measure like the Atkinson measure but it is based on a less restrictive welfare function, as argued above.

Sen’s contribution to the inequality issue is confined not merely to the refinement in the measurement of inequality, but also to a number of applications of the idea of inequality in important social policy areas.

The use of generalised Lorenz curves, and the interpretation of situations involving intersecting Lorenz curves are among the extensions that have emerged out of Sen’s theoretical work in the area of inequality measurement (see Foster and Sen 1997). These developments are immensely significant in making inter temporal and/or inter spatial comparisons of income distribution.

 

3. Sen on Poverty

From the question of inequality and its measurement to that of poverty, its satisfactory measurement and policy implications would appear to be a natural sequence. The lower ranges of the distribution of a given total income amongst populations are usually linked to poverty in many studies and in policy contexts.

However, Sen’s classic work on inequality, based on his Radcliffe Lectures, and published in 1973, did not address the issue of poverty. His pioneering contribution to the measurement of poverty was given its first formal treatment in his 1976 paper in the Econometrica (Sen 1976b). He initiated an axiomatic approach to measuring poverty taking into account the proportion of a population below a defined minimum, and the degree of inequality amongst the poor. The major themes of Sen’s significant contribution in the field of poverty-related research are explored below.

 

3.1. Absolute and Relative Poverty

The concepts of absolute and relative poverty are often invoked in the literature on poverty. But because of the ambiguous nature of these concepts, they often generate controversy and debate. The absolute concept may suggest greater precision and objectivity, while the idea of relativity may convey a sense of subjectivity and arbitrariness. The choice between these concepts would depend partly on the purpose for which the concept is to be used, and partly on philosophical and moral considerations. The use of the relative measure can help establish that poverty can exist even in societies with a high mean income, for example. This concept is gone into later in this section; first we look briefly at the notion of absolute poverty. An absolute measure may be independent of a society’s mean income.

Sen invoked the notion of deprivation to capture the essence of an absolute measure of poverty (Sen 1980). One well-used criterion of deprivation is malnutrition: if a person is unable to satisfy a minimum nutritional standard, there can be said to be poverty in an absolute sense. There are of course difficulties in defining and measuring malnutrition as the ‘irreducible core of absolute deprivation’, but Sen has argued in favour of using it as a measure of absolute poverty (Sen 1981). He accepts that what constitutes ‘deprivation’ may vary from society to society, but observes that these variations are matters of objective study.

"We could of course debate about the exact ways in which normative judgements should take note of such social variations, but the primary exercise of diagnosing deprivation cannot be sensitive to the way various types of hardships are viewed in the society in question" (Sen 1992, p. 108).

Sen has sought to separate the issue of identifying deprivation, and that of devising policies to deal with it- the latter being a derivative of the former.

 

3.2 Income Poverty and Capability

While ‘low income’ may be an indicator of poverty, a target income is not an end in itself, but only a means to achieving what Sen has called ‘human functioning’ (Sen 1987). To emphasise the connections between poverty and human functioning, Sen has developed the concept of ‘capabilities’. To have capability in this sense is to be able to ‘function’ in a chosen way, or to have options in a ‘capabilities set’. In interpreting this concept, Atkinson (1995) writes, quoting Sen, that

"in the chain

Commodities—>Characteristics—>Capabilities—>Utility,

it is ‘the third category- that of the capability to function- that comes closest to the notion of standard of living’,… and that, if the argument is accepted, it provides a basis for ‘sorting out… the absolute-relative disputation in the conceptualisation of poverty. At the risk of over-simplification, I would like to say that poverty is an absolute notion in the space of capabilities but very often it will take a relative form in the space of commodities and characteristics’" (p. 17).

This method of viewing absolute and relative deprivation has not only helped to resolve the conceptual ambiguity inherent in the concepts themselves, it has also influenced empirical implementation of the ideas to broaden the definition of human development. The Human Development Index, which has come to be used by the United Nations Development Program (UNDP), uses life expectancy at birth, literacy, and real income per head. This concept has clearly been influenced by Sen’s thinking (see Anand and Ravallion 1993).

Another empirical application of the capabilities approach has been in the area of intra-household distribution of commodity-bundles. In a number of studies (Sen 1983, Sen and Sengupta 1983, Kynch and Sen 1983, Sen 1984, Drèze and Sen 1989), this issue of deprivation within families has been addressed. Sen has argued that, while the ‘entitlements’ of households determine what commodity bundles a household can acquire, the consumption of these commodities within the households depends on the distributional arrangements practised by households. In reference to the frequently-observed disparities in consumption patterns within households, he comments:

"if the focus is on the deprivation of personal capabilities, then inequalities within the family have an importance of its own, no matter what view we take of the sustainability of the notion of individual welfare in cultures in which the family plays a dominant role. There is no escape from the grave tragedy of the disproportionate undernourishment of children (or sharper undernourishment of the female children in distress situations,..), or the unusual morbidity of women…. The problem has to be distinguished from the deprivation of goods as such since the ‘capability approach’ is concerned with what goods can do to human beings. The undernourishment and morbidity- not to mention mortality- provide a better focus than food-intake itself…." (Sen 1984, p. 363).

 

The discriminatory treatment of the female members of households has a direct bearing on their mortality. By comparing expected male-female ratios in the absence of discrimination against the female with the actual ratios, Drèze and Sen (1989) have estimated the number of "missing women" in selected countries and regions of the world to be around 100 million in 1986. The dimensions of this problem and its policy implications are clearly immense. Sen himself and his associates have addressed these aspects of the problem. Their work has inspired other researchers to study the issues of societal inequalities as well.

 

3.3. Measurement of Poverty

Social scientists and policy makers have long paid attention to issues related to poverty in societies. Attempts have been made to measure poverty in a manner that is intuitively acceptable, and useful from a policy point of view. A familiar concept used to measure what proportion of a given population is ‘poor’ is the head-count ratio (H). The use of this method of measuring poverty requires the establishment of a ‘poverty line’ which indicates what level of income or consumption is needed by a person or a group of persons (a household) to satisfy a defined (minimum) level of needs. Once this line is set, those whose income (or consumption) falls below it would be identified as poor. A related concept is that of the ‘income-gap’ (I) which is the gap between the actual incomes of those below the poverty line and the poverty line income itself. If the income gaps of all those who are below the poverty line are aggregated, one obtains an estimate of the total income needed to raise everybody to the defined minimum level.

Apart from the fact that setting the poverty line itself is an arbitrary exercise, the H- measure is also insensitive to any redistribution of incomes within the poor group. If income is transferred from a poor person to someone richer, H will either remain unchanged or go down, but it will never go up. Sen criticised the head-count measure on this ground. He pointed out that, given the nature of this measure,

"… any government faces the temptation to concentrate on the richest among the poor, since that is the way that the number of poor- and the head-count ratio- can be most easily reduced. Recasting the empirical measurement of poverty in a distribution sensitive way has the effect of making comparatively better use of income data (despite the overall limitation of that income-based informational base)’ (Sen 1992, pp. 105-106).

Sen proceeded to remedy these shortcomings and develop a measure that avoids them.

He introduces two ‘axioms of legitimacy’ for this purpose. Let us first formally define

H and I:

6. 

(7)

where z is the poverty line income, x_i the income of individual i, q the number of people below the poverty line, and n is the total population.

The proposed axioms are:

Monotonicity Axiom: other things the same, a reduction in the income of any person below the poverty line must increase measured poverty; and

Transfer Axiom: a pure transfer of income from a person below the poverty line to anyone who is richer must increase measured poverty.

This second axiom was made somewhat weaker in Sen 1981 which requires that a pure transfer to a poor person below the poverty line from a richer person, without making either cross the poverty line, must reduce measured poverty.

Clearly, H does not satisfy either of the axioms, while I satisfies the monotonicity but not the transfer axiom.

Sen’s dissatisfaction with the H measure, as noted above, stemmed from its inability to capture the severity of poverty amongst the poor. It is intuitively acceptable that a poor person’s sense of deprivation depends not just on the person being identified as poor in some objective sense, but also on the person’s relative position amongst the poor. Therefore, "the lower a person is in the welfare scale, the greater his sense of poverty" (Sen 1976b, p. 227). In axiomatic terms, this proposition leads Sen to develop weights that are based on a poor person’s income rank among the poor- the axiom of Ranked Relative Deprivation (R). A "simple case of such a relationship is to make the weight on any person’s income gap equal the rank value r(i)". Poverty is then measured as a weighted sum of the poverty gaps of the poor:

(8) P = A (z,q,n)g_iv_i (z,x)

Where g_i=(z–x_i) is the poverty gap of person i, v_i(.,.) is the weight on it (in the sense of axiom R above, r(i)= v_i(z,x), see Sen 1981, p. 36), and A(z,q,n) is a normalising parameter involving the total population n, the number of poor people q, and the poverty line income z.

The weights are made equidistanced which, Sen points out,

The insensitivity of both H and I measures to the distribution of income among the poor makes them unsatisfactory in situations where income is unequally distributed among the poor. However, as Sen observes, in situations where income is equally distributed among the poor, a combination of H and I together would suffice to measure poverty in a satisfactory way. He proposes the axiom of Normalised Absolute Deprivation (Axiom A) which states that, if every person below the poverty line has the same income, then poverty can be measured by:

(9) P = HI

Using the two axioms R and A, Sen presents a more general poverty measure:

(10) P = H{I + (1-I)G_P}

Where G_P is the Gini coefficient of income distribution among the poor. It is clear that when all the poor persons have the same income, G_P equals zero, and equation (10) reduces to equation (9).

The poverty measure P is thus a function of the number of poor (H), the aggregate poverty gap (I), and the income inequality among the poor (G_P). The relative deprivation among the poor is ranked in the manner discussed above, and included in P via the Gini coefficient. The measure P is also insensitive to changes in the incomes of those above the poverty line.

In another practical application of the measure, Sen has used this analytical approach (Sen 1976a) to correct real national income of large economies for distributional inequalities. If e is the per capita income of such an economy, G the Gini coefficient of income distribution, then e (1-G) would measure its per capita income, adjusted for inequality in the distribution of income.

The development of the poverty measure P has rescued the analytical literature on poverty from its almost-universal reliance on H, the head-count ratio. Sen’s own pioneering work and those based on it have led to a much better understanding of the nature and severity of the problem of poverty. A particularly widely-used measure in this genre for example is the Foster-Greer-Thorbecke measure (Foster, Greer and Thorbecke 1984, Subramanian 1995).

Despite these conceptual and analytical refinements, acceptance of Sen’s methodology amongst compilers of official statistics remains rather limited. As Atkinson observes:

The policy significance of Sen’s thinking in this area cannot however be diminished by the apparent reluctance of policy-makers to accept new ideas.

4.1 Food, Hunger and Conventional Economic Theory

So begins the introduction to Drèze and Sen’s classic work Hunger and Public Action (1989), a work that extends Sen’s earlier research which was brought together in his critically acclaimed book Poverty and Famines: An Essay on Entitlement and Deprivation (1981).

Although food is among the barest of human necessities, and the search for it has been a major factor behind human migrations in history, it would be fair to say that conventional economic theory has not paid a great deal of attention to studying the factors and forces behind severe food shortages and the undernourishment and large scale morbidity and mortality that usually result from it. True, governments and policy makers have often investigated the causes of specific instances of food shortages and famines, and even taken note of some of the findings in their policymaking. But a systematic study of the phenomenon as such, principally from an economic theoretic angle, but also bringing in the tools of other social sciences, where such tools are helpful, can be said to have begun with Sen’s work in the 1970s and 1980s. Due largely to the contributions of Sen and his associates, a new area of investigation can be said to have opened up in social sciences dealing with not just food shortage and hunger but also with some legal, ethical and moral issues relating to how societies divide up their resources and real incomes. In what follows we examine briefly the issues in the debate that Sen has addressed. Our discussion will also touch on some of the empirical dimensions of the research of Sen and his associates in this area.

2. Food, Famine and Entitlements

Not having enough food to eat is a state of starvation; an extreme state of starvation affecting a population is famine. In the popular perception, a famine occurs when there is a shortage of food- a supply failure, in the language of economics. Such a situation may come about because of a harvest failure, for example. The connection between food shortage and famine is so strong in peoples’ thinking about famines that to say otherwise is to risk incredulity. So, when Sen’s research, first published in the 1970s, found that the Great Bengal Famine of 1943, which claimed the lives of some two to three million people, took place in a year in which the supply of the two main food grains- rice and wheat- was "only about 5 percent lower than the average of the preceding five years", it was bound to be controversial. And it was, for more reasons than one. His findings also showed that the supply of food grains in 1943 was 13 percent higher than in 1941 when there was no famine. Sen went on to examine several other recent famines in Asia and Africa, and, again, his findings confirmed that these famines were not caused by food shortage. So, he rejected food availability decline (FAD) as a necessary factor causing famine, although it can of course be a sufficient factor behind a famine.

So, what does cause a famine? Sen advanced his now-famous entitlement approach to explain endemic hunger and famine that afflict large populations from time to time. The basic message of this approach is that, for various reasons, certain societies or certain sections thereof suffer a decline in their ability to command food, i.e. their entitlement to food. The entitlement approach is an analytical device defining certain relationships in societies which link production, consumption and the rights of people to the goods and services available in societies at any point in time. Sen uses three basic concepts to define the relationships involved: the endowment set, the entitlement set and the entitlement-mapping or E-mapping, for short. The concepts can be applied to define the relationships for an individual or for a group of individuals such as a household (family) or a social class.

The endowment set refers to the ‘resources’ which are legitimately at the command of the agent in question. Such resources may be some personal attributes of the agents such as their labour power, knowledge and skill, or some objects possessed by them such as property and other productive assets like animals or equipment. The legitimacy of the resources may be conferred legally and/or by accepted customs and norms of a society.

The entitlement set refers to the set of goods and services a person’s endowment set can enable the person to lawfully obtain. Agents endowed with the resources can engage in production, exchange or transfer to obtain the goods and services they wish to enjoy. All possible sets of goods and services belong to the agents’ entitlement set, as long as the actions involved in transforming resources to entitlement are in keeping with the law of the land.

How the endowment set at the command of a person gets translated into an entitlement set is captured in the notion of E-mapping. Such translations can involve production by the person using resources out of his/her endowment set, or may involve trading a resource or an asset for a usable good or service. The set of all possible bundles of goods and services that the person can command, given an endowment set, is the person’s ‘exchange entitlement’. In addition to the production and exchange-based transfers, a person can also have access to ‘non-entitlement transfers’ such as charity or some other form of transfer.

When a person’s entitlement set is incapable of guranteeing him/her adequate food, without non-entitlement transfers, the person faces starvation. Such a situation is caused by what Sen calls an ‘entitlement failure’. When a large number of people face such an entitlement failure, there is famine. Thus, a person faces starvation- or a class of people faces a famine situation- when there is a reduction in their exchange entitlement which makes it impossible for them to acquire adequate food. This may be caused by a decline in food supply which pushes food prices up and thereby reduces their capacity to purchase the required quantity of food. This is a situation where there is food shortage, but the proximate reason for starvation or famine is not food shortage per se, but the reduction in exchange entitlement brought about by increased food prices, according to this line of thinking. Sen calls it a failure of exchange entitlement (FEE), which may be related to FAD, but it does not have to be.

Given the three concepts outlined above, FEE can be caused by a decline in a person’s endowment set and/or by adverse changes to his/her entitlement mappings. It is not of course enough just to point out that famines are caused by entitlement failures, one has to investigate what caused entitlement to fail if such failures are to be avoided by appropriate policy measures. Sen has studied four recent famines closely to verify how FAD and FEE as explanations of famines stack up. These studies are taken up for discussion briefly below. We note in passing first that the entitlement approach as proposed by Sen has been the subject of intense criticism, even some ridicule (see Osmani 1995 for an excellent assessment of the concept and the criticisms).

The four famines Sen studied are: (i) the Bengal Famine of 1943; (ii) the Ethiopian Famine of 1972-74; (iii) the Famine in the Sahel, 1973; and (iv) the Bangladesh Famine of 1974. As the detailed chronology of events leading to each of these famines, their consequences and lessons are readily available (see Sen 1981, for example), this study no more than highlights some of the salient aspects of these catastrophes to assess how far their occurrence relates to FAD and FEE. For reasons of brevity, we combine (i) and (iv) as one set, and (ii) and (iii) as the other. The first set may be termed the Bengali famines, and the second the African famines.

The province of Bengal comprised the state of West Bengal in India and the sovereign state of Bangladesh. When British India was partitioned into India and Pakistan in 1947, the eastern part of Bengal became East Pakistan. After a brief civil war in 1971, East Pakistan seceded from Pakistan to become Bangladesh. Like some other parts of the Indian subcontinent, Bengal has suffered famines from time to time in recent centuries. The Famine of 1943 was particularly severe in its impact, and has been the subject of numerous reports and enquiries, both official and private. Sen identifies three phases of the famine: from the beginning of 1942 to March 1943; from March to November 1943; and from November 1943 through most of 1944. His findings also reveal that deaths due to starvation peaked in the second phase, while the death rate as a whole was at its highest in the third phase when epidemics were claiming many lives.

Not surprisingly perhaps, the conventional explanation of the famine runs in terms mainly of a supply failure (FAD) which in turn was caused by a number of factors as the following quote notes:

Sen found this assertion unconvincing. His study reports that the food supply situation in 1943 was only marginally worse than the average of the previous five years. It was in fact 13 percent higher than in 1941 when there was no famine, as noted earlier. His own explanation runs in terms of the failure of exchange entitlements of many rural inhabitants. The reason for this was the decline in rural wages relative to the price of rice, the main staple diet. The rise in the price of rice had a demand and a supply side to it. On the demand side were factors such as increased war-time spending- mostly out of newly printed money- by the government, and its spin-offs within the private sector which took advantage of the increased demand for certain types of commercial activities. This latter saw an uneven expansion of incomes and purchasing powers of certain groups of people in particular occupations and locations within the province. On the supply side were factors such speculative withdrawal and hoarding of rice by farmers and dealers, and panic purchase by those who had the money to do so.

While these and other factors combined to raise the price of rice, rural wage rates of agricultural and other labourers did not keep pace as disruptions to normal economic

activities added to the swelling of the number of people seeking (unskilled) work, put downward pressure on wage rates. This led to a drastic reduction in the entitlement set of a large proportion of the rural population who faced starvation- and death- on a massive scale. Sen points out that peasant-farmers and share-croppers were among the least affected groups. This is because they had in their endowment sets assets such as land and/or a right to share the crop they grew. This helped them avoid severe reduction in their food entitlements unlike the landless agricultural labourers, artisans, and other non-agricultural labourers who had to rely on a money wage the E-mapping of which had fallen.

An interesting aspect of the entitlement relations, drawn attention to by Sen, concerns the legal systems protecting ownership and property rights in societies. The apparently baffling instances of starving people dying in front of well-stocked food shops in Bengal is ‘explained’ by the fact that these people lacked legal entitlement to the food. The ownership rights of the shop-owners were protected by the law of the land. This raises the question as to how far law should have a moral dimension in order for it to command willing acceptance in a society.

The 1974 famine in Bangladesh was preceded by devastating floods which affected mainly the northern districts of the country, and washed away much of the seasonal rice crop. The price of rice rose sharply, and the incidence of starvation rose with it. Famine was officially declared by the government in September, and relief operations in the form of providing cooked food to destitutes were put in place under government sponsorship. By November, as another rice crop was harvested, rice prices started to come down, and the provision of free cooked food was stopped. Mortality estimates vary, but some estimates put it at over a million when deaths due to diseases caused by malnutrition in the immediate aftermath of the famine are taken into account.

In seeking to relate this famine to the output of rice, Sen found, again,

In terms of the availability of all foodgrains, both total and per capita, 1974 was again a local peak, observed Sen. Even in the districts most affected by the flood, food availability was higher in 1974 than in 1973.

Thus, FAD does not appear to be the factor causing the famine as there had not been a decline in the availability of foodgrains, total or per capita, in 1974 in Bangladesh as a whole, despite the flood. Indeed, in the three districts which bore the brunt of the famine- Sylhet, Rangpur and Mymensingh- per capita availability of food was higher.

The explanation, as Sen went on to show, lay in the decline in the entitlements of certain groups of people. A relevant observation in this context is that the largest group of destitutes who sought help at the food distribution centres were labourers, closely followed by farmers. Many of these destitutes owned no land at all, some had a small amount of land. The only asset in the endowment set of these people therefore would have been their labour power. As in the Bengal famine of 1943, the rural wage rate in Bangladesh in 1974 failed to keep pace with the increased price of rice. Coupled with reduced employment opportunities brought about by the flood, food entitlements of a large number of people declined drastically, causing them to starve. It is thus the FEE approach that can explain the famine.

The Ethiopian famine of 1972-74 affected some of the northern provinces more severely than the rest of the country. In particular, Wollo and Tigrai provinces were badly hit as was Harerghe, but to a lesser extent. The major factor contributing to the famine was the failure of rainfall in the north of the country in mid-1972 and early 1973. There were signs of distress and movement of people out of the drought-stricken areas by late December 1972, but recognition that a major catastrophe might be building up was slow as were the necessary relief operations which did not begin in a significant way until the latter part of 1973.

As this famine followed a severe drought which was most likely to have affected the size of the food crops, it is natural to expect that its explanations would be in terms of FAD. Sen has examined the available data on output of food crops in different parts of Ethiopia over several years around the years of the famine. His findings show that there was "very little evidence of a dramatic decline in food availability in Ethiopia coinciding with the famine" (Sen 1981, p. 92). Thus, once again, FAD was not a satisfactory explanation for the famine for the whole of Ethiopia. However, there was a shortage of food in the province of Wollo. The famine there, Sen points out, can be explained in terms of an entitlement failure on the part of the Wollo farmers who suffered a reduction in the output of- and therefore in their consumption of- food. This would be a direct entitlement failure resulting from a reduction in the endowment set of the (subsistence) farmers. Consumers of food other than the farmers also suffered because their earnings would be cut reducing their command over food, a trade entitlement failure.

The purchasing power of the Wollo farmers and cultivators would also have been reduced by other factors such as a fall in the market price of land and of livestock, and by loss of livestock as a result of the drought. All of these would have contributed to a reduction in the incomes and the entitlements of the farmers and others, both directly and through trading, in line with the entitlement approach.

Moving on to the drought and famine in the Sahel region which, in Sen’s study, includes Mauritania, Senegal, Mali, Upper Volta, Niger and Chad. These West African countries are semi-arid in climatic terms, but areas within these countries are in the arid zone. The drought that caused famine in this geographic region of West Africa existed over the period 1968-73. Food shortage started to be experienced in early in the 1970s, and, by the spring of 1972, the UN World Food Programme recognised the seriousness of the drought and the food shortage situation, and recommended that emergency food aid be made available. Incidences of starvation and starvation-related mortality increased around this time relative to the situation over the preceding four years. The famine situation peaked in 1973, and the number of famine deaths rose to an estimated 100,000.

Unlike in the other famines discussed above, Sen starts his analysis of the Sahelian famine by accepting that a sharp decline in the average food availability per head had taken place in the Sahel region as a whole during the drought of 1969-73. This would seem to lend support to the FAD approach to explaining famines. As Sen puts it:

But, as Sen then goes on to examine the available data in greater detail, he finds the FAD analysis again deficient. For example, looking at food availability figures on a country-by-country basis, he finds that there was no systematic connection between the severity of the famine and the availability of food in particular countries. The countries that experienced the sharpest decline in food availability did not necessarily experience the most severe famine, nor did the countries that had smaller decline in food availability avoid the famine altogether.

The emphasis of the FAD approach to aggregate supply of food makes it overlook distributional changes in incomes and spending powers of different groups brought about by a natural disaster such as a drought or a famine. Therefore, while the approach may help forewarn that a famine situation is developing in a certain geographic area, it cannot predict which groups are likely to be the worst affected amongst the general population of the area, argues Sen. The FEE approach on the other hand can help predict how any existing entitlement arrangements might change when food shortages begin to occur as a result of a natural calamity. In the Sahel famine, for example, destitution and mortality among the nomadic population was higher than the share of the nomads in the total population. The reason for this disproportionality was the reduction in the incomes of the nomads who suffered a significant loss of animals as well as a sharp fall in the livestock prices which combined to cut their entitlements, as Sen documents. It is factors such as these which the entitlement approach with its emphasis on information at the micro level helps to highlight.

As mentioned earlier, Sen’s analytical tools and empirical results have been subjected to copious criticisms. But the fact remains that his work in the area of hunger and famine has helped draw attention to some basic issues of human survival and well-being. Indeed, he has succeeded in offering a distinctly non-Malthusian explanation of endemic malnutrition and widespread hunger that still persist in a world that has, paradoxically, grown more opulent over time. Hunger and famine are not the result of population growth outstripping food supply growth, as Malthusians would argue; they arise more often because society’s distributional arrangements sometimes fail to protect the basic entitlements of some of its members.

If Sen’s work on famine and endemic hunger were confined to just developing and applying the notion of entitlement, it would have advanced our understanding of these problems in a major way, as it has done. But he moved on beyond measurement and evaluation to address policy issues with examples of numerous detailed case studies. In the preface to that remarkable treatise, co-authored with Jean Drèze, entitled Hunger and Public Action, the authors state that "the primary focus of the book is on action, rather than on measurement"(Drèze and Sen 1989, preface).

In addition to assessing the extent of hunger and malnutrition in the contemporary global society, the authors investigate the causes of such deprivation. The failure of entitlements naturally figures prominently in the explanation. In an extensive review of the practices of different economies- rich and poor, free market or planned, with single-party or pluralistic systems of government- they identify the strengths and weaknesses of public policies and how such policies relate to the state of hunger and malnutrition and the general well-being of the people. They have coined the phrases ‘growth-mediated security’ and ‘support-led security’ to capture the essence of the approaches used to raise the standard of leaving of people. This is how they explain the two approaches:

We use a few selected examples from the vast collection of country-studies to illustrate the important points Drèze and Sen make. They draw pointed attention to the very different achievements of India and China in regard to reducing hunger and undernourishment, and expanding life expectancy over the period since the early 1950s to the late 1970s. Both of these large, poor countries had achieved similar (modest) economic growth rates in this period. Their agricultural expansion was also slow. And yet, China’s system of public provision of food through rationing, and extensive network of health services promoted peoples’ entitlements to food and other basic necessities, including medical care, much more effectively than was the case in India where such public provision was more selective and patchy. Over this period, China managed to improve the average life expectancy at birth of its population from the low 40s to the high 60s, without any dramatic improvement in its general level of opulence, while India’s achievement was much more modest. China’s record is an endorsement of the support-led security idea, according to Drèze and Sen.

An interesting twist to this comparison of the world’s two largest countries concerns India’s relative success in avoiding major famines since its independence in 1947, while China’s disastrous Great Leap Forward experiment of the late 1950s generated the famine that cost some 30 million lives. The major contributing factor in this case, the authors argue, is India’s democratic system of government, and its free and inquisitive press media which, by publicising and critically assessing food supply situations, impel public authorities into action- an advantage China’s single-party rule and state-controlled media lack. This insight has led Sen to observe that

And yet, with all the advantages of India’s relatively free political institutions, its record of avoiding regular hunger and non-extreme deprivation is much worse than China’s, Drèze and Sen point out. Because such (quiet) regular tragedies do not make headline news, their eradication through public action is much more difficult to achieve. In the contemporary world, many a democratic nation, with greater affluence than India’s, experience creeping poverty and deprivation amongst sections of their populations. But, without active public interest in the eradication of such entitlement failures, they are allowed to persist. Socialist China’s ‘political commitment’ to providing the basic needs of the people may well have served them well in this regard.

Even within India, the success of the state of Kerala in raising the average life expectancy at birth to 66 years, against India’s 52 years is the result of the state’s sustained programme of public support in medical care. Kerala’s high literacy rate is also a factor, according to the authors, in making people better informed and careful about health issues.

The authors have also examined the records of a group of other countries which used mainly the strategy of growth-mediated security to improve the living conditions of their peoples in a wide sense. Countries such as Hong Kong, Singapore and South Korea are among the countries that achieved rapid economic growth since the 1960s. This enabled them to raise the general standard of living of their people, and to improve their basic entitlements to food, health care and basic education. Drèze and Sen point out that in these countries too, the state has played an active role in promoting social security which in turn has helped their economic development.

Examples abound in this book of advocacy for public action. Countries as diverse as Sri Lanka, Chile, Kuwait, Costa Rica, Cuba and Jamaica- and a good many others- have featured in this social-scientific study. The message is for active public involvement- and commitment- to fight poverty, hunger, ill health and ignorance. And the message applies as much to developed, richer, nations as it does to the poorer ones.

The selected themes of Sen’s wide-ranging contributions detailed above show how influential Sen’s scholarly work has been in areas where conventional economics has either not ventured at all, or, at best, confined itself to esoteric theorising. Sen’s reputation as an economic theorist is founded on his earlier work in areas covering economic growth and development, capital theory, social choice and welfare economics in general, project evaluation and cost-benefit analysis, to mention just a few major ones ( The Scandinavian Journal of Economics 1999 has categorised Sen’s publications into no fewer than sixteen different areas, not all of which are in ‘economics’).

When Sen was working on his doctoral research at Trinity College, Cambridge in the late 1950s, the debate raged as to whether labour-abundant developing economies should encourage the use of labour-intensive methods of production, as common sense would suggest. Sen pointed out that a developing country needs to maximise its savings rate to finance investments which would help raise its income and consumption streams over a longer time horizon. This might require a labour-rich economy to use production techniques which are relatively capital-using as that might generate a larger surplus and therefore larger funds for investment. In conceptual terms, this requires assigning a higher ‘shadow’ wage rate to labour than the market rate. The usefulness of shadow prices in the context of investment decisions is now well understood, thanks largely to Sen’s early work in this area (Sen 1960). The contributions of Little and Mirrlees (1969) and of Dasgupta, Marglin and Sen (1972) have made these ideas more readily accessible to practitioners in the field of project appraisals especially in the context of developing economies.

Another significant contribution of Sen’s in the field of development economics relates to the notion, put forward by Lewis in the 1950s (Lewis 1954), that most developing economies have a ‘surplus’ of labour. These economies are characterised by a dualistic structure which makes it possible for the ‘modern’ industrial sector to grow by drawing labour away from the ‘traditional’ agricultural sector without causing output in the latter sector to decline. The traditional sector’s requirement of labour for any given level of output therefore is less than its actual employment of labour- a phenomenon referred to as ‘disguised unemployment’.

While casual observation in many a developing country may lend support to such a possibility, it is hard to justify the rationale behind employing labour at a positive wage rate when the marginal product of labour is nil. Sen (1966) pointed out that many traditional activities in developing societies are based on family members working together and sharing the resulting income. When some labourers are removed, others put in more effort to make up for their loss so that total output does not decline. In other words, the objectives of such family (or cooperative) enterprises are different from those of capitalist, profit-maximising, ones. This distinction between labour input and labourers employed is a subtle but important one in the context of explaining ‘rational behaviour’ in peasant (and/or cooperative) economies.

A related theme involving the relative efficiency of farm sizes in traditional agriculture was addressed in Sen (1962) which suggested an inverse relationship between farm size and the productivity of land, and offered explanations for this relationship. Significant subsequent research in this area largely corroborates such a relationship (see Ray 1998 pp. 354-61 for a good summary), thus confirming Sen’s pioneering role in yet another area.

Sen’s reputation as an outstanding theoretical economist derives from his contributions in a diverse range of issues, as indicated above. Perhaps the area in which his intellectual influence has been the most pervasive is social choice theory. Although this is outside the scope of this article, it is worth noting that Sen’s contributions here extend well beyond what one would regard as the domain of economics. As Basu (1998) observes about Sen’s research in social choice theory

Sen, the accomplished economic theorist, is also a "leading light" amongst philosophers, "having published extensively in philosophical journals", as Arrow (1999, p. 163) notes.

And yet, for all for all his insightful contributions to numerous areas of the social sciences and philosophy, his influence may have been even greater when he takes on the activist role. At a time when the world seemed to have moved away from the idea of a benevolent role for the government in economic decisions, Sen emphasised the importance of public action in dealing with some entrenched problems facing humanity, as discussed above. As Drèze and Sen (1989) concede

They ask

But Sen and his research collaborators have emphasised the very positive role that public participation can play "in both collaborative and adversarial ways vis-à-vis governmental policy" (op. cit. p. 259). Because their advocacy for public action has been based not just on sound theoretical arguments but also on extensive empirical research, they have had a useful ‘hearing’ in policy-making circles, as noted above. By going against the tide in contemporary economic thinking, Sen has helped keep the debate about the appropriate role of public action alive at a time when there was a danger that the advocates of free-market policies would hold an unchallenged sway.

In this paper we have discussed three major areas of economics within the wider discipline of social sciences where the 1998 Economics Nobel Laureate Amartya Sen has made pioneering contributions- the areas being poverty, inequality and famine. In addition to highlighting the distinctive nature of Sen’s contributions in these areas, we have also pointed out how Sen has used economic theoretic insights to advocate policies to deal with some endemic problems societies face from time to time. Economists are often perceived as technocrats who are interested in resolving technical problems which may have no more than an indirect relevance to the problems of the ‘real world’. It is to Sen’s credit that he has helped "to legitimize the investigation of certain topics- such as famine- that had previously been regarded by many as outside the profession’s concerns" (Atkinson, 1999, p. 189). He has challenged many a current orthodoxy in economic thinking so as to raise doubts about the legitimacy of ‘received wisdom’ in areas that are fundamental to particular ideological persuasions. In doing all this, Sen has drawn attention to the fact that economics as a social science has a ‘moral’ dimension as well as an ‘engineering’ one. This is a tradition that comes down from the classical economists such as Adam Smith and Karl Marx for example, but is often overlooked in the standard textbook treatment of economic problems. Sen has restored for economics the role that makes it socially useful, and, by so doing, has been, in the words of another Economics Nobel Prize winner Robert Solow, "the keeper of the conscience of our trade".

Drèze, J. and A. K. Sen (1989): Hunger and Public Action. Oxford: Clarendon Press

Raj, D. (1998): Development Economics, Princeton: Princeton University Press.

Sen, A. K. (1960): Choice of Techniques. Oxford: Basil Blackwell.

Sen, A. K. (1973): On Economic Inequality. Oxford: Clarendon Press and New York: Norton

Sen, A. K. (1976a): "Real National Income", The Review of Economic Studies, 43, 19-39.

Sen, A. K. (1980): "Deprivation as Choice", Oxford Economic Papers, 32, 353-69.

Sen, A. K. (1984b): Resources, Values and Development, Oxford: Basil Blackwell.

Sen, A. K. (1987): On Ethics and Economics. Oxford: Blackwell.

Sen, A. K. (1992): Inequality Reexamined. Oxford: Clarendon Press