Amartya Sen, who won the Nobel Prize for Economics in 1998, has spent most of his career worrying about how to understand, measure, and promote “human development,” in the sense of the term that refers to the general condition of the people and not the achievements of a fortunate few. He is probably best known for his work with Jean Drèze on famines and how to prevent them, but his work covers a very broad range. He has worked in social choice theory, a technical discipline that examines the precise relationship between judgments about what is good for society and judgments about what is good for its individual members.

Sen has held a joint chair in economics and philosophy at Harvard. He has written extensively, often in these pages, on poverty and inequality, as well as on gender differences in nutrition and life expectancy in the developing world. His most recent book, which is also the one most obviously aimed at an audience outside professional economics, begins by pointing out that

we live in a world of unprecedented opulence, of a kind that would have been hard even to imagine a century or two ago…. And yet we also live in a world with remarkable deprivation, destitution and oppression.

It would be wrong to suppose that our collective achievements can be measured just by the opulence, without taking account of the continuing deprivation, much of it persisting within rich countries as well as poor ones. Overcoming this deprivation “is a central part of the exercise of development,” yet to do so we cannot just react to problems as they come along, relying on an instinct for recognizing development when we see it. We need a coherent conception of what human development is in order to be able to pursue it as a goal. This book is therefore an attempt to develop such a conception.

According to the particular conception Sen espouses, development is “a process of expanding the real freedoms that people enjoy.” He contrasts this with “narrower views…such as identifying development with the growth of gross national product, or with the rise in personal incomes, or with industrialization, or with technological advance, or with social modernization.” It might sound odd to call “social modernization” a “narrow” view of development, but to see what Sen means, and why he thinks a view of development based on freedom is both broader and better than its rivals, it is important to understand the background from which his thought has evolved.

The common thread running through the varied body of Sen’s work is the idea that public policy has for too long been informed by excessively simplistic pictures of human society, and specifically of human beings and their needs and aspirations. This is a claim that has to be interpreted with some care. It would be trite to assert that human beings and human societies are more complex than you would ever imagine from looking at the GNP statistics. Of course they are, and the GNP statistics would be quite useless for anything unless they abstracted from a vast amount of this complexity.

So Sen’s target is not simplification per se: the complexities of human society have to be simplified in order for public policy to assign any priorities at all. Doing so necessarily involves excluding relatively unimportant information about society (and therefore about society’s individual members) in order to concentrate on the relatively important. Economic policy in particular has been motivated by the search for single numerical measures that allegedly capture most of what is important for a given problem (GNP, say, or the retail price index) and that allow you to say precisely how well a given policy project is doing, and even to rank countries by how well they have achieved the goals set by that project. The really interesting question is not whether public policy has had to simplify, but whether it has simplified too much.

The message of much of Sen’s previous work can be summed up by the statement that single aggregate numbers simplify too much. The average national income per person ignores its distribution among people. The average income of a household ignores its distribution among household members. Income itself, which sums up the importance for people’s lives of marketed goods and services that can be assigned a price, thereby ignores other things people care about, such as health or political autonomy (health care has a price—and how—but health itself does not). Rising incomes, Sen points out, may mean little if people live in much greater fear of violence and expropriation. Likewise, statistics reporting the total availability of food in a given country or region do not tell you how that food is distributed among people. Shifts in prices or in purchasing power may make some people fall below the threshold of starvation even if the total amount of food in a region does not change. In each case you need to know more than the overall averages tell you; something important has been left out.


But what precisely is it that these single aggregate numbers leave out? Part of the answer is easy: averages of any quantity leave out the distribution among people, and that distribution is important, because we worry much more about conditions at the bottom end of the scale than about those at the middle or the top. But part of the answer is not easy. Incomes, life expectancies, and other such measures are only imperfect indicators of what is sometimes called “the quality of life.” What precisely these measures leave out is the question Sen’s account of freedom is supposed to answer. The quality of a person’s life, he writes, is to be judged according to “the capabilities that a person has, that is, the substantive freedoms he or she enjoys to lead the kind of life he or she has reason to value.” Measures of economic welfare can be useful indicators of whether people enjoy a good quality of life, but they are far from perfect indicators, and when they conflict with other evidence of the quality of life it is sometimes the other evidence that is to be preferred. High incomes are associated on average with better health, for example, but it is the health that counts and not the incomes per se. Incomes, Sen writes, are “only instrumentally significant.”

What difference does it make to talk about “capabilities”? Here’s an example. Arguments have raged for years about whether poverty is best understood as a relative notion or an absolute one: Are you poor when you have less than a certain real income or when your real income falls below some proportion of the average? Sen wants us to think of poverty as a condition that occurs when people fall below some absolute level of the quality of life. Whether someone is counted as poor should not depend on the statistical accident of the incomes of those who happen to be included in the calculation of the average. But the absolute level should be defined by the “capabilities” people have rather than their incomes. People care about clothes, say, partly for simple functional reasons like protection from the cold but also for reasons of self-respect and respect within their communities. You may be poor if your income is insufficient to enable you to achieve self-respect, which sounds like an absolute criterion. But the amount of income it takes to achieve self-respect may be higher in Beverly Hills than in Bangladesh, thereby explaining why the relative concept of poverty is not wholly mistaken after all. In this case, Sen would emphasize the capability to achieve self-respect.

So capabilities are an important explanatory concept because they are, at least in principle, the underlying valuable goals of development of which measures such as GNP are only indirect indicators. But what have capabilities to do with freedoms? To this reviewer at least the answer is not always luminously clear, even on multiple readings of Sen’s book. At times Sen talks of capabilities and freedoms interchangeably to describe the various activities that are necessary for a full and active life: living long and healthily, enjoying the esteem of family and friends, communicating with others and learning about the world we live in, participating in decisions about our own future and the government of our community. At other times Sen worries that we might confuse the genuinely deprived with those who have voluntarily chosen to forgo some of the activities of a full life: ascetics, those on a diet or fasting for religious reasons, mountaineers losing their lives in the quest for adventure. So he takes capabilities to mean the possibility of enjoying a full life whether this possibility (“freedom”) is exercised or not. At yet other times Sen appears concerned to avoid making capabilities sound like an arbitrary list of things he happens to approve of; so his conception of the enhancement of the range of human opportunities (also “freedom”) appears as what defines the list of worthwhile capabilities: the ability to take part in sport, say, might count as a capability but the ability to take part in military conquest would not. Whether Sen’s appeal to freedom strengthens his overall argument is a point to which I shall return.

The book develops its argument in three main stages. The first, which occupies most of the book, uses large numbers of examples to appeal to our intuitions that simple indices such as GNP or even life expectancy do indeed leave out dimensions of human development of foremost importance. Judging countries solely by their average performance on some index—be it GNP or infant mortality or literacy—can lead to the shocking neglect of groups and regions within those countries. Thus “even entire [Indian] states such as Uttar Pradesh (which has a population as large as that of Brazil or Russia) do not do much better than the worst-off among the sub-Saharan countries in terms of these basic indicators of living quality [adult literacy and infant mortality].” So distribution within countries matters as well as the averages, especially since the averages are arbitrarily affected by the boundaries of nation-states.


Furthermore, the hope that GNP might be a good shorthand for overall development progress because it is correlated with other important indicators may lead us to overlook some major issues of concern. Thus “in the United States, African Americans as a group have…a lower chance of reaching advanced ages than do people born in the immensely poorer economies of China or the Indian state of Kerala.” And even for individuals, the presence of handicaps and different social circumstances means that incomes are not just an unreliable but a systematically misleading way of comparing the welfare of different people. Someone who is handicapped may have fewer capabilities than another person in every aspect of life, even with a substantially higher income. What matters, then, are not the incomes but the human capabilities that such incomes can support—and different people in different circumstances can convert income into human capabilities with strikingly different degrees of effectiveness.

Most of all, respect for people’s human and civil rights is no less central a part of their development because there do not exist markets in which the satisfaction of such rights can be bought. On the contrary, human rights matter because the exercise of almost all other capabilities depends on one’s not being subject to arrest, imprisonment, torture, and execution; civil rights matter because they can promote the capability of being involved in making decisions about one’s community. Countries differ greatly in their respect for such rights, and Sen does not want to see these differences overlooked just because GNP does not reflect them. He also joins those who have argued that democracy is no barrier to economic development, but is positively instrumental in bringing development about. His own work with Jean Drèze on famines illustrates this well, since it is often the most authoritarian countries that have suffered the worst famines. A flourishing civil society and a free press are often the best assurance against the occurrence of crises affecting minorities whose plight is ignored by the aggregates and averages of conventional policy analysis.


Many of Sen’s examples would be persuasive even in the absence of any overall theory of development. Whatever definition of development we might reasonably agree on, it is clear that millions of people in Uttar Pradesh are excluded from it. However, other examples are more controversial. Some people will question an approach that makes human and civil rights a central part of the project of international development. Objections to this have come in part from such blatantly interested parties as the political leaders of authoritarian countries and are not usually shared by the movements that oppose them. But less partisan observers have also worried that concern for human and civil rights may be somehow less objective than other aspects of development, and specifically may reflect a “Western” or similarly ethnocentric bias. The second stage of Sen’s argument is therefore devoted to showing that human and civil rights are no less grounded than other capabilities in values that are found throughout human societies. Partly this follows just from the fact that respect for these rights is an essential precondition for enjoying other capabilities. You cannot enjoy a happy family life while in jail or being tortured, or a full intellectual development when your reading matter is restricted to government propaganda. Partly Sen’s argument follows from noting important points in common in the evaluative language of different societies. In the process, he deftly and wittily disposes of the view that “Asian values” do not include respect for freedom, citing many historical examples from China and India to show that argument and dissent are deeply embedded in their cultures.

Still, it is one thing to argue that freedom has an objective basis as a component of human development—that it is one of the capabilities that, for example, make participation in public decisions possible. It is quite another to argue that development can be defined as the expansion of freedom, that all the capabilities people may acquire are to be understood as exemplifying freedom, which is Sen’s third stage of argument. Or rather, it is what Sen asserts, at various stages through the book, and also implies through his frequent use of the word “freedom” when discussing various aspects of development such as education and infant mortality. As I have already indicated, I do not find these assertions and implications either clear or convincing.

Nor is it evident what would be gained by agreeing that freedom is the touchstone by which to evaluate claims about the nature of development. For instance, the wish to see freedom as the fundamental value underlying every other even leads Sen at one point to talk about mortality as a denial of “the freedom to survive.” Well, yes, one can call it that, but is it really illuminating to suggest that what matters about being dead is the lack of freedom that goes with it? Being dead is also bad for the health and has a significant statistical association with dropping out of college, but personally I think it’s the deadness that would bother me. This may sound like a debating point, but it is fundamentally an argument about what constitutes a good explanation. Explanations proceed by taking something that is not well understood (human development, say) and demonstrating its connections to something else that is rather better understood. Freedom doesn’t really fulfill this function, since, in many respects, it is at least as obscure and bitterly contested as the concept of development itself.

Someone, for instance, might accept Sen’s view that the enhancement of freedom is the proper definition of development but also believe that one of the most important freedoms is the self-determination of ethnic groups. Should it therefore follow that ethnic self-determination is part of the definition of human development? Sen himself would certainly not think so. And the most fruitful way to begin resolving the dispute would surely not be to discuss whether or not self-determination is an intrinsic freedom, but rather to look at the kind of society that would result if the proposed right to self-determination were to be honored in a particular case. We can always say that the society we like best is the one that most advances freedom, but a claim of that kind sounds remarkably like the claim in Molière that opium sends people to sleep because of its dormitive faculty.

If a general appeal to freedom is not a good way of resolving issues about the nature of human development, are there better ways? I think there are. Those that seem pertinent and persuasive depend to a large extent, I would suggest, on which actual or imagined audience we think of ourselves as addressing. Audiences made up of cultural or intellectual historians tend to concentrate their assessment of human development disproportionately among the more complex products of human communication and symbolism, and be less interested in the conditions (even the intellectual conditions) of the population at large. In contrast, assessments of our economic and social development might be seen as an interim report card addressed to our political leaders and our institutions of collective action, and therefore as requiring a more general and democratic scope. But they also limit the complexity of the criteria by which development is to be assessed, because of the nature of that audience.

One of the central problems of political thought has been that of finding ways to align the interests and incentives of political leaders with those of the rest of society. An important insight drawn from the study of incentives in human organizations is that the broader and less focused the goals politicians are required to pursue, the harder it is to hold them to account for their success or failure in doing so. So one way of thinking of the nature of development may be to see it as those aspects of human aspiration to which we can realistically require our political leaders to devote themselves. In order to provide an effective set of incentives, these aspects will necessarily be narrower and more focused than the whole range of goals to which human beings might reasonably aspire.

The point is worth stressing because Sen sometimes writes, and those who cite his work approvingly often write, as though the development of social indicators such as GNP statistics were motivated by a philistine insensitivity to what makes human beings tick. Yet I know of no one who really believes that gross national product, or a rise in personal incomes, or any of the narrower goals against which Sen argues can be wholly identified with development. Every source of cultural inspiration from Diogenes to Dickens to Dynasty reinforces the message that money isn’t everything, and no sane person genuinely believes otherwise. That does not prevent such intermediate goals from being essential yardsticks by which the performance of our political leaders can be more reliably judged than if we make them responsible for human happiness, freedom, and fulfillment in all its infinite variety.

No country is so misgoverned that there is nothing whatever to be said in its favor, and no country is so blessed with enlightened government that nothing could be done to improve matters. So unless the domain of political responsibility is somehow narrowed the arguments about relative freedom and fulfillment may go on forever, and it may be quite impossible to hold politicians and administrators to account. And conversely, if those who govern us are philistines, they may be philistines by design.

This is not to say that the narrower the domain of political responsibility, the better. Sen’s work has undoubtedly done much to make policymakers take seriously the importance of personal autonomy and human variety in the process of economic and social development. Unfortunately his recent book, though evidently intended by the publishers to reach a wide audience, has many defects as an introduction to this body of work. Both writing and editing bear signs of haste, and sentences like “If person i has more of a significant functioning than person j, and at least as much of all such functionings, then i clearly has a higher valued functioning vector than j has” suggest the author may not have accurately assessed the needs of the average intelligent reader. The editors could have wielded the blue pencil less timidly to valuable effect.

The book also too often misrepresents itself as plowing a lonely furrow against received opinion, the latter being characterized by vague allusions without references (contrasting oddly with the extensive references to Sen’s own work and that which supports his views). A particularly flagrant example occurs in a chapter on “Markets, State and Social Opportunity,” where Sen writes that “the virtues of the market mechanism are now standardly assumed to be so pervasive that qualifications seem unimportant.” “Standardly” by whom? The world is a large place, and while you can find many people who believe this (try talk radio), it is a serious misrepresentation of the work of Sen’s professional colleagues. “Market failure” is a central component of every course on general microeconomics taught at every reputable university in the world. In another chapter Sen talks of “the belief that has been so dominant in many policy circles that ‘human development’…is really a kind of luxury that only richer countries can afford.” You’d never guess from this that the importance of health and education for development and alleviating poverty is now a central tenet of policy discussion in all the international institutions and among all aid donors, from Oxfam to the World Bank. Sen has certainly been influential in changing received opinion, and he has no need to represent that opinion as being more philistine than it really is.

To the extent that “policy circles” still do pursue narrower goals than Sen would like them to, it is important to understand some of the reasons that may lead them to do so. Nobody can seriously dispute the richness and variety of human development about which Sen writes so persuasively—but nobody of sense would want to. If narrower goals of development have been pursued it is not in many cases through simple stupidity but partly because of the difficulty of granting politicians and administrators more grandiose mandates and then holding them reliably to account. The more multifarious the goals they claim to be pursuing, the harder it may be to ensure they pursue any of them successfully at all.

Indeed Development as Freedom is curiously silent about the difficulty of devising mandates that work. In the preface Sen writes, “I have, throughout my life, avoided giving advice to the ‘authorities.’ Indeed, I have never counseled any government, preferring to place my suggestions and critiques—for what they are worth—in the public domain.” But if Sen’s suggestions are good ones it might seem a pity not to have counseled the authorities directly. The authorities tend not to be particularly avid readers of the American Economic Review, or even, regrettably, of The New York Review of Books. But in any case the point seems a rather facile one, since the book arises out of lectures given to the World Bank, which wields more power in the world than many governments, and it is praised on the cover by Kofi Annan, who has far less power but must count as one of “the authorities” if anyone does.

In recent years the range of issues with which the World Bank concerns itself has broadened, to include the environment, health care, gender, political decentralization, and local accountability. These are all admirable goals, though they lie some way outside what management consultants would call the “core competencies” of the bank. There are good arguments for broadening the mandate, but the arguments for continuing to restrict it are not stupid ones either. The International Monetary Fund, the bank’s sister organization, has recently come under justified criticism for its confused and vacillating policies toward Russia, in which economic orthodoxy—for example, a requirement that borrowers have plausible prospects of repaying loans—has been adapted to the political concerns of the US, its dominant shareholder. Regardless of the merits or otherwise of American political pressures, and regardless of the organization’s other alleged deficiencies, the clarity and effectiveness of IMF policy have been damaged by the growing fuzziness of its mandate.

Amartya Sen has done much, in this book and in his career, to argue for a humanist conception of public policy with a broad range of aspirations. His arguments are powerful ones, but they are not the end of the story. Important questions remain about how to pursue the project of international development in a way that is simple enough for even the most plausible and slippery politician to be held to account.

This Issue

March 29, 2001