77069 ON "HUNGER AND PUBLIC ACTION" A Review Article on the Book by Jean Dreze and Amartya Sen Martin Ravallion Recent writings on poverty and hunger agree broadly on objectives and means but diverge significantly in emphasis. Views differ on the importance and func- tion of economic growth and on how much weight to give to direct public sup- port. These differences can matter in policy choices. Hunger and Public Action, by Jean Dreze and Amartya Sen, is an important contribution to the literature on antibunger policy. This article critically examines the issues raised by the book, its differences with other recent writings, and the implications for both policy and future research on policy. D reze and Sen's book is a good starting point for a discussion of antihunger policy. The first book in a seriesi reporting results of research sponsored by the World Institute for Development Economics Research, Hunger and Public Action is an outstanding contribution to the lit- erature. It is also accessible, written with clarity of thought and economy of technique. But it is more than a book about specific policies; it is also an ap- plied study in aspects of the theory of policy. It prescribes not only things to do but also a way to think about what we should do. New concepts are advo- cated for that task. This article's evaluation of the arguments of Hunger and Public Action aims to assess implications both for policy and for the way we think about policy. Hunger and Public Action is a scholarly book, but it is one with a clear and potentially influential prescriptive message. I found myself in sympathy with that message. And I believe it is one around which a consensus could now be formed in the development comrmunity, particularly in the light of not The World Bank Research Observer, vol. 7, no. 1 (January 1992), pp. 1-16 C 1992 The International Bank for Rteconstruction and Development/THE WORLD BANK dissimilar views expressed in other recent works on this topic, notably the United Nations Development Programme's Human Development Report 1990 and the World Bank's World Development Report 1990 (UNDP 1990; World Bank 1990). I did, however, feel uncomfortable with certain aspects of Hunger and Public Action's message and with some of the data and methods used to support it. The issues in question are not easy to resolve but are nonetheless important. This article will also try to identify points where further research on antihunger policy seems called for. Concepts Dreze and Sen define hunger to embrace all kinds of social as well as bio- logical disadvantages associated with inadequate food intakes-a definition much broader than the concept frequently found in research and policy. For example, a person who adapts to low food energy intakes by reducing activity may not consider himself or herself hungry; Dreze and Sen would, nonetheless, deem this a food-related deprivation within the scope of their study-in short, a problem of hunger. I believe their broader perspective is warranted. Antihunger policies should surely help people who adapt to low food energy intake by diminishing their level of activity. Yet both research and practice sometimes (at least implicitly) dictate otherwise. One example is the common use of child anthropometric measures (such as weight-for-age) to indicate nutritional need. Children can maintain seemingly satisfactory physical growth rates at low levels of food en- ergy intake by not playing (see, for example, Beaton 1983); that is surely a serious food-related deprivation for any child. The concept of entitlement is used throughout Hunger and Public Action, following Sen's book Poverty and Famines (1981a). A person's entitlements over some time period are the various consumption bundles that are legally attain- able from that person's initial endowments, given prevailing prices. Sen (1981a) made an enormous contribution in exposing the limitations of the view that famines are caused by a decline in aggregate food availability. Sen argued that this view has little analytical force in understanding some major famines of this century. Many of these were not associated with an obvious decline in avail- ability of food. And even for those that were, the arithmetic of aggregate food statistics does not explain why similar shortfalls at other times or places did not also result in famines. Nor does it help us understand the incidence within a population of famine mortality or the extensive forms of nonfatal suffering during and after a famine. Instead, Sen advocates study of the entitlements of famine victims. Of course, people may avoid or suffer starvation without a change in their entitle- ments; they may avoid it by illegal means or suffer it by choice. But, because it emphasizes how individual budgets are determined, the concept of entitle- 2 The World Bank Research Observer, vol. 7, no. 1 (January 1992) ments is useful in understanding how specific famines happen and how they may be averted. This perspective has clear antecedents in the literature on fam- ines in India (see, for example, Bhatia 1967); indeed, recognition of the impor- tance of entitlements goes back centuries in the subcontinent and elsewhere. Sen has contributed greatly in redirecting attention in analysis and policy to the role entitlements play in causing and relieving modern famines. A theme of Hunger and Public Action is that the expansion of entitlements should not, however, be viewed as the final objective of human well-being. Rather it is the capability for doing valued things (called functionings) that ul- timately matters: "Formally, a person's capability is a set of functioning bun- dles, representing the various 'beings and doings' that a person can achieve with his or her economic, social, and personal characteristics" (Hunger and Public Action, p. 12, n. 18; also see Sen 1985). The authors argue that the goal of public action should be to expand human capabilities rather than entitle- ments per se. The capability central to the topic of Hunger and Public Action is that of being alive and healthy. Being adequately nourished is seen as essential for this capability. Here again the authors take a broad view, emphasizing the variabil- ity over time and between people in nutrient needs for good health as well as that access to health care, basic education, clean drinking water, and adequate sanitation are important complements to nutrition. Does the attention paid to capabilities represent a significant departure from mainstream approaches to the welfare analysis of public policies? Dreze and Sen emphasize the inadequacies of regarding consumption of commodities as an end in itself (this is also a prominent theme of the Human Development Report). But these are not adequate grounds, in terms of theory at least, for their departure from mainstream thinking. Current welfare economics does not insist that income, or comrnand over consumption, is the final objective of per- sonal well-being. Nor does it insist that income is the sole determinant of in- dividual well-being, or that personal characteristics do not matter, at given consumption levels. Whether we emphasize capabilities or stay with main- stream theory, we can conceive of well-being as dependent on the same basic ingredients: goods consumed, the time spent in various activities, and personal characteristics. It is in the practice of policy analysis that the main difference becomes clear. Consideration of capabilities forces us to take a broad view of the range of factors relevant to individual well-being. The more eclectic nature of main- stream analyses readily allows certain special assumptions to be made in ap- plications-often about those determinants of individual welfare that are difficult to observe. A common assumption of this sort is that income or con- sumption is an adequate indicator of individual well-being-in other words, that: individuals are identical in all other respects. I doubt if this assumption would ever be easily accepted when applying the capabilities approach. Martin Ravallion 3 This may or may not matter. In some circumstances it can be innocuous to ignore non-income differences across individuals. But in other circumstances it could make a big difference to conclusions about policies. Social sector policies, for example, have sometimes been judged solely according to what they do for productivity and, hence, incomes. Even if one confined attention to that most basic of capabilities-to be healthy-one would not take such a narrow point of view. Dreze and Sen also differ from some other writers in the importance they attach to the public provision of the goods needed for expanding human capa- bilities. A consensus has been emerging in recent literature that the develop- ment of human resources must be given a higher priority than in the past; see, for example, World Bank (1989, 1990) and UNDP (1990). There is less sign of consensus on the further question: What is the relative importance of private versus public provision as instruments for achieving human resource develop- ment? Many of the things important in developing and maintaining human re- sources are essentially private goods (in that the benefits are exclusive to the individual consumer), and so there is a possibility for efficient provision through markets. Human resource development will then require that people want and can afford these goods. The case for direct public provision then depends on the adequacy of policy instruments influencing incomes and their distribution, as well as on the existence of relevant market failures. Some well-known arguments against public provision of private goods do not arise in the framework of Dreze and Sen. An example is the concern about the welfare losses that arise from imposing a common level of public pro- vision when tastes vary. Within the Dr&ze-Sen approach, the case for public provision rests heavily on the strength of the relationship between capabilities and incomes. In the case of nutritional capabilities, if the link between incomes and hunger is weak relative to other factors, then Dreze and Sen are justified in putting less emphasis on generating incomes in reducing hunger (except in- sofar as growth facilitates the financing of public support, a function to which the book does attach importance). Recent debates about the relationship between income poverty and hunger have bearing on this issue. A great deal of effort has gone into investigating how responsive nutrient intakes-particularly but not exclusively food energy intakes-are to changes in income. Recent econometric estimates using house- hold-level data have suggested that food energy intakes are relatively unrespon- sive to income changes, even for the poor. (For a good survey, see Behrman 1990.) Although I agree that some estimates have overstated the response of individual intakes to changes in income, it does not follow that the capability of being adequately nourished for good health is also unresponsive to income changes. Whether it is or not will depend on how changes in intakes are re- flected in health status, which will depend in turn on how far nutrients fall short of needs, as well as on other factors, such as the initial state of health. These interactions can strengthen the link between income poverty and hunger. 4 The World BanikResearch Observer, vol. 7, no. 1 (January 1992) For example, measures of undernutrition in Indonesia have been found to re- spond quite strongly to income gains, even though caloric intakes at the indi- vidual level are relatively unresponsive (Ravallion 1990). In this instance, a capabilities perspective suggests that incomes may matter more than was otherwise thought, not less. One familiar concept-food security-is conspicuous in its absence from Hunger and Public Action. Earlier inconsistencies in the usage of this term have largely vanished; food security is now widely defined as "access by all people at all times to enough food for an active and healthy life" (World Bank 1986, p. 1). Dreze and Sen are careful about words, and I am sure they have their reasons for avoiding this term. I expect they feel it gives undue emphasis to command over goods, and just one bundle of goods, food. Nonetheless, much of Hunger and Public Action is unquestionably about attaining food security, and the book clearly has much in common with other recent writings on food security, including the Bank's policy paper on the topic (World Bank 1986). Data and Methods The methods of empirical analysis in Hunger and Public Action, like the concepts, are distinctive-although this time not so much for their sophistica- tion as for their lack of it. Dreze and Sen seem particularly shy of econometric methods. Possibly they were concerned to make the book accessible to readers unfamiliar with these methods. But surely accessibility is determined more by the way results are presented than by the methods used to obtain them. None- theless, as someone who uses econometric methods routinely, I was struck by how far the authors seem to be able to take us with little more than the inves- tigative tools of 100 years ago. One example should suffice. Hunger and Public Action presents some dev- astatingly simple estimates of the number of women "missing" in the world owing to the gender bias in mortality. There are 1,050 females to every 1,000 males in Europe and 1,022 to every 1,000 in Sub-Saharan Africa. The cor- responding figures are as low as 941, 931, and 905 per 1,000 in China, India, and Pakistan; these numbers represent 44 million, 37 million, and 5 million missing women, respectively, when judged against the African ratio of females to males. Dreze and Sen attribute this discrepancy to higher rates of female mortality, which are taken to reflect sex bias in access to food and health care. They do not discuss other possible explanations. For example, the same factors that lead to such biases would presumably also yield some underreporting of females in household surveys. Nonetheless, these unsophisticated figures hint at an alarming problem of gender bias in parts of the developing world. But this sort of empirical analysis can take us only so far. A great deal of data that bear on this and other questions raised in Hunger and Public Action have not been exploited in the book. I suspect that at least part of the reason Martin Ravallion 5 is that the more sophisticated methods needed to learn from such data have been shunned. For example, hardly a scrap of household-level data on living standards has been used, despite the fact that many of the countries mentioned in the book now have such data available, and often for more than one point in time. And these are data with quite direct bearing on assessments of capa- bilities (as well as on more conventional measures of welfare). An understand- ing of the circumstances of the poor at the household or individual level can greatly inform antihunger policies (Lipton 1988; World Bank 1990). In discussing progress in reducing chronic hunger in part 3 of Hunger and Public Action, Dreze and Sen rely almost exclusively on a few social indicators, such as aggregate child mortality and average life expectancy. I found this sur- prising. These indicators are useful up to a point, but they can be rather unin- formative about the quality of the lives that people live-their capabilities, if you wish. One could hardly contend that they are good indicators of the at- tainment of the most basic capabilities, beyond being alive. In making compar- isons over time, they may also be deceptively sensitive to very small improvements in primary health care, particularly when the mortality rate is high to begin with. Preventing the common infectious and parasitic diseases such as malaria is undeniable progress, but there is a lot more to the elimina- tion of chronic hunger and related deprivations. Real progress-or lack of it- in reducing the persistent poverty underlying chronic hunger may show up little in these aggregate social indicators. Famines A large part of Hunger and Public Action is devoted to policies for prevent- ing famine, building on Jean Dreze's recent work (particularly Dreze 1990a; 1990b), as well as Amartya Sen's well-known earlier writings (particularly Sen 1981a). The key policy issue is how to avoid the contractions in entitlements suffered by vulnerable groups. This is not just a matter of avoiding transient distress, although that is ample motivation. Longer-term chronic hunger is both a cause and an effect of suffering during a famine. Protecting the entitlements of the poor at such times can play an important role in development (both in the narrow sense of economic growth and in the richer sense of expanded hu- man capabilities). Dreze and Sen further expose the fallaciousness of classifying public actions in this context into mutually exclusive categories of relief and development. Markets and Famines The first step in formulating sound policies is to understand the workings of existing institutions for allocating resources, of which markets are generally the most important. The crucial response of markets to impending famine, and the 6 The World Bank Research Observer, vol. 7, no. i (January 1992) role of public action, is given adequate attention in Hunger and Public Action, but some of the issues bear elaboration, and some points are not particularly well coveredin the book. The main policyissues in question concern both the spatial and intertemporalperformance of food-grainmarkets. There can be no presumnption that spatial food movementsthrough private trade will alleviatea famine;the food will probably moveaccordingto market prices which need not be higher in the worst-off regions (sufficientto cover transport costs). Following Sen (1981a), Hunger and Public Action points to the fact that there havebeen examplesof food export from famine-affected re- gions. The term "slump famine" has been coined for this phenomenon, to indicate that famine often arises from contractionsin purchasingpower in af- fected regions. There is a risk of some confusionhere. The export of food during a famine does not mean that external trade has not helped stabilizedomestic consump- tion. The crucial question is whether food exports declined, or imports in- creased, during the famine. This begs two further questions: Did food-grain exports respond to changes in relative prices? Did the relative prices facing traders reflectdomestic scarcities?Hunger and Public Action does not address these questions. The evidenceon at least one supposed caseof slump famine- the severefamines in British India around the turn of this century-does not support the view that trade destabilized consumption (see Ravallion 1987a). Equally, there can be no presumption that government restrictions on food movements will improve matters, and there is evidence that they have not (Ravallion 1987b). A better strategy, well advocated in Hunger and Public Action, is to raise food purchasingpower of the poor in affected regions. Nor, according to popular opinion, are markets very good at the intertem- poral allocation of food during famines; panic buying and excessive hoarding often exacerbate current scarcities. Some experts endorse this opinion: a well- known handbook on famine and disaster relief, for instance, recommends var- ious actions against food hoarding (see Masefield 1967). The very existence of private food stocks during famine has understandably given rise to anxiety, not least among those threatened with starvation. But markets may have a part to play in encouraging storagc-a potentially important means of stabilizing consumption over time, which could improve the prospects for survival of the poor, as well as the profits of speculative stock- holders. How well markets will work in this regard will depend on how com- petitive they are and on how shrewdly traders have anticipated future scarcity. These are difficult empirical questions. In the major famine in Bangladesh in 1974, there is evidence of very harmful destabilizing speculation. Errors in price forecasting appear to have been positively correlated with readily avail- able information on damage to the future harvest (Ravallion 1987b). Traders therefore stored more rice against these anticipated production losses than they would have if markets had been functioning efficiently. Rice prices rose to record levels in the months between the flooding and the arrival of the next Martin Ravallion 7 (depleted) harvest, and prices fell sharply in the week or two preceding the ar- rival of that harvest. The severe contraction in the food entitlements of the poor led to a sharp increase in mortality, which also peaked before the decline in aggregate food availability (Ravallion 1987b). The evidence on the Bangladesh famine also calls into question the perfor- mance of nonmarket food institutions. Rice prices in Bangladesh are influenced by changes in the government's food-grain stock, itself determined by previous imports (including foreign aid) and internal procurement efforts. Thus traders' expectations of the effect of preharvest crop damage on future prices require an assumption about the government's response to shared information. The most plausible conclusion is that the stockholders' overoptimistic price expec- tations or anticipations of future rationing during the 1974 famine were pre- mised on a belief that the government would be unable to respond suitably to the reported damage to the future crop (Ravallion 1987b). A ban on grain hoarding is not the most effective government response. It is unlikely to work (stocks are easy to hide) and may even make matters worse by fueling excessive price expectations, premised on a lack of confidence in the government's credibility. Rather, public action should be geared to support- ing public confidence in future food availability and the stability of prices, through open market operations and food distribution policies, backed up by adequate stocks or stabilizing external food trade or aid policies. By such means, Bangladesh has done a better job of avoiding famine in the 1980s (Osmani 1987). Food or Cash? Direct food delivery to famine-affected areas is the most common policy in- strument for famine relief. This method has two distinct functions: raising ag- gregate food availability and raising the food entitlements of vulnerable groups. Dreze and Sen argue that there is a potential gain from separating these two functions. Selling food aid or imports in domestic markets and using the pro- ceeds to reach the needy may be more effective than directly delivering food. Cash transfers can also help even if aggregate food availability cannot be in- creased; for example, the authors point to the success of the state government of Maharashtra, India, in avoiding famine in 1972-73 using a rural public em- ployment scheme, even when food supply could not be rapidly increased. The development community remains divided on the merits of monetizing food aid. (For discussion of the various views, see Reutlinger 1984; Berg 1987; and Singer, Wood, and Jennings 1987.) Dreze and Sen are persuasive in arguing that whether cash transfers or food relief will be more helpful to those in need must be judged according to the specific setting, including the performance of food-grain markets. Under competitive conditions, the case for cash relief is strong either if food prices in the affected region are below world prices at the border (since recip- 8 The World Bank Research Observer, vol. 7, no. I (January 1992) ients can then purchase more food locally with the cash than the aid agency can obtain with the same money on world markets) or if private traders can deliver the food more efficiently than the aid agency. The case is weaker if local markets perform poorly, although even then the change in policy can be pref- erable, provided that traders can still deliver food at lower cost than the aid agency (Coate 1989). A common criticism of cash relief is that it will lead to higher food prices. Dreze and Sen argue that this outcome is fine provided that the higher food prices are the result of an effectively targeted cash transfer policy; those in need will still be better off. One could quarrel with the theoretical generality of the authors' argument here; the conditions we are discussing by no means rule out the possibility that the recipient's initial gains are substantially eroded by sub- sequent changes in relative prices (Ravallion 1987b). But a reader becoming mired in that debate would risk missing the main point of Hunger and Public Action, which is to stimulate a more creative ap- proach to the formulation of policy and careful consideration of the relative merits of the options. As elsewhere in the book, Dreze and Sen demonstrate that clearer thinking in distinguishing ends from means can enhance the pros- pects for effective public action against hunger. Reaching the Vulnerable Whether in the form of cash or food, if transfers are to prevent or relieve famine they must be genuinely redistributive. The benefits of a well-targeted transfer scheme are plain enough, although the costs to the poor may not be insignificant. Without the necessary administrative capabilities, targeting is not easy. For example, as Dreze and Sen point out, the underlying intrahousehold allocation of food will often adjust to thwart attempts to reach certain indi- viduals through supplementary feeding programs. The goal must be to protect the household entitlements of vulnerable individuals. Among the alternative targeting mechanisms, Dr&ze and Sen argue for a combination of employment provision to help those willing to work at low wages and unconditional relief to those who are obviously unemployable. Di- rect provision of employment, as an element of a comprehensive safety net for the poor, also gets favorable reviews in other recent literature (see, for example, ILO 1988; World Bank 1990, chap. 6; and UNDP 1990). Rural public works schemes have proved repeatedly in South Asia and recently in parts of Africa and Latin America that they are effective in reaching the able-bodied rural poor in times of need. (Ravallion 1991 surveys the theory and evidence for South Asia.) The reason relief work schemes seem to be well targeted is that they impose a cost (forgone income, transport cost, disutility of work) on would-be partic- ipants, and this cost tends to be lowest for those in greatest need. Critics have argued that this is a deadweight loss, but that ignores the fact that alternative Martin Rauallion 9 policy instruments-such as perfectly targeted lump-sum transfers-that allow costless redistribution in this setting are simply not available (Ravallion 1991). In any case, the cost of participation for the poor is unlikely to be high at times when famine threatens, so that the net transfer is likely to be a reasonably high proportion of wage receipts. An aim of this type of public action is to avoid the onset of famine by pro- tecting the entitlements of the vulnerable; they need not already be starving, but they can see the impending risk. An effective system of famine relief should be ready to go into action as soon as it is needed-usually well before external food aid can arrive. And it should ideally fall out of use when it is not needed. Public administrative capabilities will constrain the policy options in practice, although much can still be done. Hunger and Public Action discusses a number of success stories in famine prevention through such domestic public action, such as Maharashtra in 1972-73, and Botswana in 1982-87. Saving lives is the overriding goal, but is not the only reason to favor effec- tive protection of entitlements through public action. Although I have not seen any estimates, it is surely plausible that even the narrowly defined current fi- nancial costs of an effective system of early response are well below the costs of belated famine relief and rehabilitation. A timely domestic policy response can also avoid the potentially disastrous consequences to the poor of their more desperate later responses, such as sale of assets and migration. Although indi- vidually rational, these responses can readily add up to a higher collective risk of death, through exposure to disease, and leave those lucky enough to survive this time facing longer-term destitution because of a diminished asset base or more debts. In such cases, the provision of an effective safety net for the poor complements the longer-term alleviation of poverty through economic de- velopment. Chronic Hunger It is one thing to prevent famine and quite another to eliminate chronic hun- ger. Most of Hunger and Public Action's discussion of chronic hunger concen- trates on earlier experiences in using direct policy intervention in raising nutritional and related capabilities over the longer term. In characterizing these experiences, a distinction is made between growth-mediated security and support-led security; the former relies on growth to finance public support, while the latter does not. The central question addressed in the book is whether direct intervention can work. Dreze and Sen examine in detail the outcome of some relatively successful efforts at longer-term public support in developing countries, as indicated by rates of improvement in social indicators, particularly under-five mortality. Ex- panding on earlier work by Sen (particularly 1981b), the authors show how countries such as Chile, China, Costa Rica, Cuba, and Sri Lanka have used 10 The World Bank Research Observer, vol. 7, no. 1 (January 1992) direct intervention in public provision to achieve the social indicators typical of richer countries. For example, Sri Lanka's life expectancy of about seventy years is higher than that of many countries with far higher average incomes- such as Brazil, where life expectancy is sixty-five years for a country with roughly four times Sri Lanka's gross national product per capita. Hunger and Public Action gives a lot of attention to the interesting compar- ison of the performance of China and India in public support over the past forty years or so (see also Sen 1989). The relative success of China in enhancing longevity (with, the authors argue, similar rates of economic growth), is con- trasted with the greater success of India in avoiding famine since the 1950s. Dreze and Sen argue that China was able to raise longevity through active in- tervention in health care and food distribution. But, whereas India has success- fully avoided famine through effective public intervention over this period, China experienced what was arguably the worst famine in recorded history in 1958-61; the estimates of people who died in that famine vary from 17 million to 30 million. India's freer press and more open political environment is cited as an important impetus for its more successful efforts in avoiding famine, which is far more newsworthy than chronic hunger. Drawing on these experiences, Dreze and Sen are encouraging about the scope for direct support-led security. Similar views are found in FAO (1987), Berg (1987), UNDP (1990), WHO (1990), Brown University Faculty (1990), and a series of public declarations in recent years.2 There is a subtle but potentially important difference of emphasis between these writings and the World Devel- opment Report 1990 (World Bank 1990). The World Development Report clearly sees direct public support as a less important instrument for alleviating poverty in the longer term than achieving the right sort of economic growth. Nonetheless, there appears to be wide agreement that growth is not sufficient for eliminating chronic poverty and hunger in its various dimensions, and that a combination of growth-mediated and support-led security is needed. The real difference lies in the position taken on what should be the proper balance between the two. On that issue, one question that many readers of this book will ask is: Can support-led security be achieved without sacrificing long-term expansion in those capabilities that do require private consumption goods? Economic growth in some of the economies that have pursued support-led security over long periods, such as Sri Lanka, has been less than impressive. One gets the impression that Dreze and Sen do not see any necessary tradeoff-that support- led security need impinge little on future prospects for growth-mediated secu- rity. I believe that quite a credible argument to that effect can be made, but it will not be found in Hunger and Public Action. Whether a significant growth cost is incurred in practice will depend on various conditions, notably the struc- ture of the economy (including any existing macroeconomic imbalances) and the resource costs of intervention (which, as Dreze and Sen do point out, but with little hard evidence, may be quite modest in poor countries). The method Martin Ravallion 11 of financing will be crucial on both counts; financing direct support by cutting other public spending would plausibly damage future growth-mediated security less than financing by tax increases (particularly on export earnings, as in Sri Lanka for many years) or borrowing. By at least one assessment, the cost to growth of direct support is likely to be small provided that (and it is an important proviso) there are no significant macroeconomic imbalances (Bourguignon 1991). The possibilities for combining support-led with growth-mediated strategies, and their timing, merit further research. Hunger and Public Action throws little light on the prospects for a development strategy that combines support-led se- curity at an early stage with growth-mediated security in the longer term. It would be interesting to know if any of the countries identified by Dreze and Sen as examples of growth-mediated security would have been examples of support-led security at some earlier time when they were much poorer. A number of related questions are left begging about the role of public action in promoting equitable growth in private incomes-not as an end in itself, I hasten to add, but as an instrument for higher ends ("increasing utility" or "expanding capabilities," as you prefer). Little att.ention is given to the identi- fied examples of growth-mediated security (Hong Kong, the Republic of Korea, and Singapore). I would expect that, with the sustained periods of equitable growth experienced by these countries, capabilities requiring private consump- tion goods have expanded greatly for the poor, as have those requiring public provision. There are numerous questions about how public action can help in this process. To give just one example, scant regard is given in Hunger and Public Action to the policy issues that arise in ensuring that the poor share fully in the potential benefits from technical progress and commercialization in ag- riculture (Binswanger and von Braun 1991). The enthusiasm of Hunger and Public Action for public provision of social services-notably health and education-is shared with other recent writings, including the Human Development Report 1990 and the World Development Report 1990. Most of these writings have left me feeling uneasy about the way the attitude toward the public provision of certain private goods, notably those used in health care and education, differs from that toward income transfer and safety net policies. It is not always clear that the same standards are being applied in evaluating the two types of public spending. For example, the (difficult) policy choice between targeting and uniformity that Hunger and Public Action covers quite adequately in the context of famine relief policy also arises often in the social sectors but is hardly mentioned by the book in that context. A similar observation can be made about the World Development Report 1990: the chapter that deals with transfers and safety nets gives far more attention to cost-effectiveness and targeting than does the chapter on social services, although the problem would appear to be no less acute in the latter context. There is a lot we do not know about how to deliver social services to the poor in developing countries. Analogous to Dreze and Sen's study of countries 12 The World Bank Research Observer, vol. 7, no. 1 (January 1992) with good social indicators relative to their incomes, it would also be of interest to look at those with good social indicators relative to their levels of public expenditure on social services. There can be no presumption that simply throw- ing more money to the social sectors will much expand the capabilities of the poor; indeed, in many countries a disproportionate share seems to have gone to raising the capabilities of those whose capabilities are already relatively ample (Jimenez 1987, chap. 5; World Bank 1990, chap. 5). The same factors that act to restrict the extent of income redistribution in favor of the poor pre- sumably will also constrain the possibilities for targeting social services to the poor. The nature and cost of an effective system of public support for the poor will depend crucially on those constraints. The extent to which public action is constrained by adminisitrative capabilities and political realities, and the de- tailed issues of policy design and implementation subject to those constraints, are of the utmost importance to many of the countries that could benefit from switching to a strategy of support-led security. Conclusions Dreze and Sen argue convincingly that there is a positive role for public ac- tion in both famine relief and the longer-term alleviation of poverty and hunger, and that it is a role that poor countries with limited domestic resources can go a long way toward fulfilling. They need not wait to be rich countries. In arguing the case for such public action, the authors put less emphasis on raising incomes (except insofar as that mediates public support) than one finds in other recent writings on this topic. Dreze and Sen are entirely justified in their criticism of the use of average output as the sole indicator of progress in development. A high average income can allow great scope for widespread well-being, but it does not guarantee it. However, one should be careful not to lose sight of the fact that mean in- come is one of the relevant parameters of development; the others are the dis- tribution of income and the provision of crucial nonmarket goods, notably social services. There are some notoriously difficult problems in finding the right balance between growth and equity, and between the provision of goods by private and public means. The position one takes will depend in part on one's views about human well-being and social welfare at quite an abstract level. It will also depend on country-specific circumstances, including past rates of growth, levels of incorne inequality, and the existing provision of social ser- vices. Focusing on the expansion of basic human capabilities as the object of development can help clarify some of these issues, but I am unconvinced that this perspective calls seriously into question the instrumental importance of at- taining growth with equity in most developing countries. It is not enough, then, to know that poor countries need not wait to be rich countries before providing direct support; the question is also whether they will Martin Ravallion 13 have to wait any longer. It cannot be presumed that they will, though there is no way of being sure without knowing quite a lot about economic circum- stances and policy details, including methods of financing and their implica- tions for growth. Dreze and Sen concentrate almost exclusively on the success stories of direct support. This teaches the valuable lesson that direct support can work, but there are also lessons to be learned from the cases in which at- tempts at direct support have failed in the short term or have been costly to the poor in the longer term. This is an important area for future research. A number of questions also remain unanswered about the implications of identifying capability expansion as the objective of public action. For example, since the approach eschews welfare assessments based on utility, a policy choice that has the greatest impact on capabilities, such as health, might be judged inferior to some other choice by all affected individuals. It is risky to rely solely on information on utility; for example, Sen (1970, chap. 6) has shown how utility-based assessments of policy choices may contradict other valued princi- ples, such as liberty. That is certainly no less true in discussing matters related to health. But these arguments fall well short of convincing me that one should ignore information on utility altogether. It remains unclear what is the best way to incorporate that information in evaluating policies, and what tradeoff against other objectives is to be accepted. Hunger and Public Action does offer authoritative encouragement on the po- tential for effective public action against hunger. Famines can be avoided, and we need not wait for an increase in domestic aggregate food availability to do so (although this is not to deny that such an increase will almost certainly help). Chronic hunger and the threat of destitution can be greatly alleviated, and we need not wait for growth to do the job (although this is not to deny that the right sort of growth is a powerful weapon, and that public action has a role in promoting that growth). I believe that a reasonably broad consensus now exists on these propositions, but great challenges remain in both the anal- ysis and the implementation of effective public action against poverty and hun- ger in specific countries. There is still much we do not know about how to go about making a success of public support. This should be a high priority for future research on development policy. Notes Martin Ravallion is a principal economist in the Population and Human Resources Depart- ment of the World Bank. He is grateful to Jock Anderson, Jere Behrman, James Boyce, Gaurav Datt, Michael Lipton, Amartya Sen, Dominique van de Walle, Michael Walton, and the Observ- er's referees for their comments on this article. 1. Other volumes so far include Dreze and Sen (1990). 2. Among these statements are the Cairo Declaration of the World Food Council, the Bellagio Declaration from a conference organized by Brown University's World Hunger Program, and the Bangkok Affirmation by the Task Force for Child Survival; these are reprinted in the August 1990 special issue of Food Policy. Also see the introduction to that issue by Kasperson and Kates. 14 The World Bank Research Observer, vol. 7, no. I (January1992) References The word "processed" describes informally reproduced works that may not be commonly available through libraries. Beaton, George H. 1983. "Energy in Human Nutrition: Perspectives and Problems." Nutrition Reviews 41: 325-40. Also in Gittinger, Leslie, and Hoisington 1987. Behrman, Jere R. 1990. "Nutrient Intake Demand Relations: Incomes, Prices, Schooling." University of Pennsylvania, Department of Economics, Philadelphia. Processed. Berg, Alan. 1987. Malnutrition: What Can Be Done? Baltimore: Johns Hopkins University Press. Bhatia, B. M. 1967. Famines in India: A Study of Some Aspects of the Economic History of India (1860-1965). London: Asia Publishing House. Binswanger, Hans P., and Joachim von Braun. 1991. "Technological Change and Commercial- ization in Agriculture: The Effect on the Poor." World Bank Research Observer 6, no. 1 (January): 57-80. Bourguignon, Francois. 1991. "Optimal Poverty Reduction, Adjustment, and Growth." World Bank Economic Review 5, no. 2 (May): 315-38. Brown University Faculty. 1990. "Overcoming Hunger: Promising Programmes and Policies." Food Policy 15: 286-98. Coate, Stephen. 1989. "Cash versus Direct Food Relief." Journal of Development Economics 30: 199-224. Dreze, Jean. 1990a. "Famine Prevention in India." In Dreze and Sen 1990. - 1990b. "Famine Prevention in Africa." In Dreze and Sen 1990. Dreze, Jean, and Amartya Sen. 1989. Hunger and Public Action. Oxford: Oxford University Press. , eds. 1990. The Political Economy of Hunger. 3 vols. Oxford: Oxford University Press. FAO (Food and Agriculture Organization of the United Nations). 1987. The Fifth World Food Survey. Rome. Gittinger, J. Price, Joanne Leslie, and Caroline Hoisington, eds. 1987. Food Policy: Integrating Supply, Distribution, and Consumption. Baltimore: Johns Hopkins University Press. ILO (International Labour Office). 1988. The Challenge of Employment: Rural Labour, Poverty, and the ILO. Geneva. Jimenez, Emmanuel. 1987. Pricing Policy in the Social Sectors: Cost Recovery for Education and Health in Developing Countries. Baltimore: Johns Hopkins University Press. Kasperson, Jeanne X., and Robert W. Kates. 1990. "Introduction: Halving Hunger." Food Policy 15: 274-76. Lipton, Michael. 1988. The Poor and the Poorest: Some Interim Findings. World Bank Discussion Paper 25. Washington, D.C. Masefield, G. B. 1967. Food and Nutrition Procedures in Times of Disaster. Rome: Food and Agriculture Organization of the United Nations. Osmani, S. R. 1987. The Food Problems of Bangladesh. Working Paper 29. Helsinki: World Institute for Development Economics Research. Also in Dreze and Sen 1990. Ravallion, Martin. 1987a. "Trade and Stabilization: Another Look at British India's Controver- sial Foodgrain Exports." Explorations in Economic History 24: 354-70. 1987b. Markets and Famines. Oxford: Oxford University Press. 1990. "Income Effects on Undernutrition." Economic Development and Cultural Change 38: 489-516. 991. "Reaching the Rural Poor through Public Employment: Arguments, Evidence, and 1. Lessons from South Asia." World Bank Research Observer 6, no. 2 (July): 153-77. Reutlinger, Shlomo. 1984. "Project Food Aid and Equitable Growth: Income-Transfer Efficiency First." World Development 12: 901-11. Martin Ravallion 15 Sen, Amartya. 1970. Collective Choice and Social Welfare. Amsterdam: North-Holland. - . 1981a. Poverty and Famines: An Essay on Entitlement and Deprivation. Oxford: Oxford University Press. . 1981b. "Public Action and the Quality of Life in Developing Countries." Oxford Bulletin of Economics and Statistics 43: 287-319. 1985. Commodities and Capabilities. Amsterdam: North-Holland. 1989. "Food and Freedom." World Development 17: 769-81. Singer, Hans, John Wood, and Tony Jennings. 1987. Food Aid: The Challenge and the Oppor- tunity. Oxford: Oxford University Press. UNDP (United Nations Development Programme). 1990. Human Development Report 1990. New York: Oxford University Press. World Bank. 1986. Poverty and Hunger: Issues and Options for Food Security in Developing Countries. Washington, D.C. . 1989. Sub-Saharan Africa: From Crisis to Sustainable Growth. Washington, D.C. 1990. World Development Report 1990: Poverty. New York: Oxford University Press. WHO (World Health Organization). 1990. Priority for the Poorest. Geneva. 16 vol.7, no. 1 (January1992) The WorldBankResearchObserver,