J Econ Inequal (2006) 4: 233–244 DOI 10.1007/s10888-006-9023-y

Review essay, “The 2006 world development report: Equity and development” John E. Roemer

Received: 1 March 2006 / Accepted: 2 March 2006 / Published online: 15 June 2006 © Springer Science + Business Media B.V. 2006

The 2006 World Development Report, issued by the World Bank and entitled “Equity and Development,” marks a major step forward in the discussion of poverty and development, and it is a significant event that this step has been taken by so august an institution as the Bank. For many years, economic development has been viewed as co-extensive with the growth rate of GDP per capita; that view has been increasingly challenged, as contemporary political philosophy and welfare economics have been brought to bear on the analysis of development. Notably, the UNDP initiated, some years ago, an annual report in which a ‘human development index (HDI)’ was calculated for a large panel of countries, an index in which GDP per capita is averaged with measures of education and health. The political philosophy which motivated the HDI was Amartya Sen’s idea of functionings and capability [14, 15]. In contrast, the political philosophy upon which the 2006 WDR draws has, at its root, the idea of equal opportunity, as it was first outlined by Ronald Dworkin [6] in his theory of resource egalitarianism, amended and modified by Richard Arneson [1] and G.A. Cohen [4], and reduced to an economic algorithm by John Roemer [11, 12]. The 2006 WDR is by far the most thorough application of these modern ideas in political philosophy, and welfare economics, to evaluating economic development. Because I believe that economic development must be conceived of in terms of the development of human welfare (I do not here wish to take on the more subtle question whether the welfare of all sentient beings should be made to count, as Singer [16], for example, advocates), and because the deep questions of how to conceive of human welfare are most seriously addressed by philosophers, I think the Report is moving in just the right direction. In this review, I will first give a brief presentation of the theory of equal opportunity; then a very brief outline of what “Equity and development” covers; and then spend the majority of my space on some critical reflections on the Report. That

J. E. Roemer (B) Elizabeth S. and A. Varick Professor of Political Science and Economics, Yale University, New Haven, CT, USA e-mail: [email protected]

234

J Econ Inequal (2006) 4: 233–244

allocation of my space does not accurately reflect my evaluation of the pluses and minuses of the Report: I am its great fan, but as a reviewer, my responsibility is to be a critic, and push the boundaries farther, if I can.

1. The theory of equal opportunity We have a population, and we are concerned with evaluating, from an ethical viewpoint, the distribution of some kind of advantage in it: That advantage could be income, educational achievement, health status, wage-earning capacity, or even welfare, if we could construct a measure of it. Advantage is an outcome, thought to be of value, by the agency or society which is carrying out the analysis. Let me specialize to the case where advantage is something objective – like income or educational achievement – as this will avoid knotty problems of individual and idiosyncratic conceptions of welfare. It is a useful abstraction to think of the advantage that an individual ends up acquiring or achieving as a function of his or her circumstances and effort, where circumstances are environmental factors, either social or biological, and effort comprises the set of actions that we view to be under the control of the individual. In addition to circumstances and effort, the advantage level a person achieves will be a function of the policy that the agency or country adopts. Thus, to summarize, we may write the level of advantage acquired by the person as a function u(c, e, ϕ) where c is a vector of circumstances, e is a vector of efforts or choices by the person, and ϕ is the policy, chosen from set of feasible policies 8. I assume that the function u is universal across the population, although that could be modified. For purposes of analysis, it is usually necessary to partition the population in question into a fairly small set of types or groups, where a type is a set of individuals all of whom have the same circumstances. For example, we may choose the single circumstance of parental educational level, and then partition the population into three or four types, defined by the level of their parents’ education. Suppose, for specificity, that u is wage-earning capacity; we are concerned how individuals from different family backgrounds acquire that capacity. The policies, in this case, might be different choices of how to distribute educational finance monies in the society – in poor societies, one might be particularly concerned with the allocation of educational finance between primary and more advanced schooling, where increasing the investment in primary schooling might help most the most disadvantaged type. Associated with any policy there will ensue a distribution of efforts in each type. If the types are denoted 1,2,...,T, and if we simplify by assuming that effort can be summarized by a number, then we may denote the distribution function of effort in type t if the policy is ϕ by Fϕt – a cumulative distribution function on the non-negative real numbers. The philosophical idea behind equality of opportunity, as it emerged from the literature I mentioned above, is that individuals should not be held responsible for their circumstances, but it is morally all right to hold them responsible for their effort. Effort, however, is a complex thing, and the question arises as to how to measure it. The approach I have taken is as follows. First, effort should be measured relative to other individuals in one’s own type. Thus, suppose we think of effort, in the wageearning capacity example, as the number of years of schooling the individual chooses to acquire. There is surely personal choice involved in this: But that choice is also influenced by circumstances. Thus the distribution of effort is itself a circumstance, it

J Econ Inequal (2006) 4: 233–244

235

is a characteristic of the type. In measuring an individual’s effort, we should attempt to sterilize out that aspect of effort which is attributable to circumstance. A simple way of doing so is to identify an individual’s degree of effort with the quantile which he or she occupies on the distribution of advantage of his or her type. Thus, if A and B occupy quantiles 0.5 and 0.7, respectively, on the distribution of wage-earning capacity of the type to which they both belong, then we say B has applied higher effort. This requires only that it be the case that greater effort leads to a higher level of the advantage in question, holding circumstances fixed. (This assumption is not innocuous – it rules out ‘luck’ as a cause of advantage.) Perhaps more contentiously, we say that if A and C both occupy the 0.5 quantile on the distribution of wageearning capacities of the different types to which they belong, then we assert they have exerted the same degree of effort. We now say: Equality of opportunity for the acquisition of advantage of the kind u measures has been achieved if, at every level of effort, the levels of advantage across types are the same. In other words, given the discussion above, if the cumulative distribution functions of advantage across types are identical. More generally we say that a policy ϕ1 equalizes opportunities for advantage more than policy ϕ2 if the distribution functions of advantage across types are ‘closer’ under ϕ1 than under ϕ2 . I will not discuss how we measure the distance between advantage distribution functions (for that, see Roemer [12, 13]); there are a number of possibilities. In other words, equality of opportunity is a state in which the only differences in advantage are due to effort, and not to circumstance. It is often difficult to measure the distribution functions of a particular kind of advantage, especially for developing countries, where the data may be poor. The 2006 WDR adopts an acceptable compromise: It looks only at the means of these distributions. Thus, it asserts that equality of opportunity for a kind of advantage will have been achieved if the means of the advantage distributions, across types, are equal. Now, to be more precise, we are not simply concerned with equalizing distribution functions, or equalizing their means, but rather with equalizing them at the highest possible level. This leads immediately to the concept of maximin: That is, an opportunity egalitarian should seek to maximize the mean advantage level of that type with the lowest such mean. Thus, if the mean level of advantage in type t under policy ϕ is denoted µt (ϕ) then the optimization problem is: max min µt (ϕ) ϕ∈8 16t 6T

(1)

Let us call the value of this program µ∗ ; it is obviously a function of the set of policies we have identified as feasible, and the set of circumstances chosen, which determine the types. In the example I have given, I chose the level of parental education as the unique circumstance – in fact there may be many other circumstances, such as natural talent, race, caste, sex, region of the country, urban vs. rural, and so on. So, much of the inequality within types that we observe, if parental education is taken to be the unique circumstance, will in fact also be due to (other) circumstances. In other words, because in practice we always choose a small set of circumstances, and then identify the residual difference in advantage within types as due to effort, we are underestimating the degree of inequality of opportunity. In the limit, if we thought of each individual as a type of his own, as we would if we took the view that no choice a person makes is under his own control, but all choices are

236

J Econ Inequal (2006) 4: 233–244

due to circumstances, then we would attribute all the inequality in a society to circumstances. The equal-opportunity objective would then reduce to equality of outcome. In practice, however, we are always concerned with measuring inequality of opportunity due to a small and discrete set of circumstances, and so much of the inequality that we measure will be attributed to inequality of effort, and hence will not be morally disturbing.

2. Outline of “Equity and Development” The definition of equity chosen by the authors of the WDR is a state in which “individuals have equal opportunities to pursue of the life of their choosing and be spared from extreme deprivation in outcomes (p. 2).” Thus they append to a concern with equal opportunity, one with avoiding extreme poverty. So, if opportunities have been equalized, but some individuals, due to imprudence (very low effort) end up in extreme deprivation, they should be rescued – out of compassion or charity. For practical terms, the authors often identify extreme deprivation as having an income of less than $1 per day. The report consists of three parts: First, ‘Inequity within and across countries’, second, ‘Why does equity matter?’, and third, ‘Leveling the economic and political playing fields’. The first part presents a gold mine of data, summarized in charts and graphs, of inequality of opportunity within and across countries; the second part argues that inequity matters not only for its own sake, but also because it slows development, and the third part is concerned with policies to reduce inequity, and, given the claim of part 2, to thereby improve the rate of development. The circumstances that are chosen to create the partition of a population in a country into types include parental education level, rural versus urban, parental economic status, sex, caste, and gender of household head. The kinds of advantage measured are income, educational achievement, infant mortality rates, rates of child immunization, and stunting level of children. When global inequality is discussed, as opposed to within-country inequality, then the main circumstance is country of residence. In this case, kinds of advantage are life expectancy, years of schooling, and income. The policies upon which the report concentrates in the third part have to do with early childhood development and education, improving the health of the poor, guaranteeing better ‘justice systems’, chiefly through strengthening the rule of law, microeconomic reforms to improve the efficiency of labor, credit, and land markets, macroeconomic policy, and improvements in global markets.

3. What is development? Throughout the report, there is an emphasis on the claim that equity is not only a good in itself, but it is good for development. I will cite several examples from the report’s overview: Greater equity is thus doubly good for poverty reduction: Through potential beneficial effects on aggregate long-run development and through greater opportunities for poorer groups within any society (p. 2)

J Econ Inequal (2006) 4: 233–244

237

If the opportunities faced by children like N. are so much more limited than those faced by children like P. or S., and if this hurts development progress in the aggregate, then public action has a legitimate role in seeking to broaden opportunities....(p. 3) Third, the dichotomy between policies for growth and policies specifically aimed at equity is false (p. 10) In Part 2, one of the major emphases is that more equitable societies develop better institutions which in turn make the societies more prosperous; the justification here of equity is the prosperity (presumably measured by income per capita) that it generates. The concept of economic development that is implicit in these and many other examples from the report is one of GDP per capita. GDP per capita may be of interest for two reasons: As a measure of the society’s endowments (capital, human capital, technology, and the other resources that go into producing wealth), and as a welfare measure. Presumably as economists or philosophers, we are interested in GDP per capita as a welfare measure. How can it be so justified? The obvious way is as follows. We take the concept of welfare for a country to be the total (or per capita) utility of its members, and we measure the utility of an individual by his or her income. Thus,
let v xi be the utility of individual i as a function of his or her income xi , and choose v(x) = x. We then immediately get total income of a country (or per capita income, assuming that population is fixed in the short run) as the measure of social welfare. Notice there are two requirements for this justification of GDP per capita: First, that our conception of justice is total welfare (utilitarianism), and second, that welfare is equal to income. Let us modify only the second premise; it is surely more reasonable to suppose that there is decreasing marginal utility in income, as more urgent needs are met first. Suppose, for example, we chose v(x) = log x to reflect this. Then the utilitarian ethic would require us to X Y max log xi (ϕ) ∼ = max xi (ϕ). ϕ∈8

ϕ∈8

Here, x (ϕ) is the income of individual i under the policy ϕ. That is, the utilitarian would not be concerned with maximizing income per capita but rather with maximizing the product of incomes. Obviously, under this second construal of utilitarianism, there would be a much greater concern with preventing very small incomes then under the first construal. Thus, to use GDP per capita, as a welfare measure, requires not only embracing a utilitarian ethic, but also taking a particular view of the relationship between welfare and income that denies the notion of basic needs and urgency. Of course, the adoption of a utilitarian ethic is inconsistent with the report’s stated ethic, which is ‘equity,’ as defined above. So why does one have to justify policies that improve equity by arguing that they improve GDP per capita? I see no reason to do so; this is a logical inconsistency. Perhaps what the authors of the Report mean to be saying is something like this: “Even if you, reader, disagree with our view that equity is the right ethical posture to adopt, and you remain utilitarians of the type that endorse maximizing per capita GDP, then you should nevertheless support our equity-improving policies, because i

238

J Econ Inequal (2006) 4: 233–244

they are the path to maximizing GDP per capita.” If that is the writing between the lines, then the Report’s message is weakened. The work of the philosophers upon which the Report explicitly rests (Dworkin, Arneson, Cohen, Roemer) emerged out of John Rawls’s [9] magisterial attack on utilitarianism: While these writers disagree about the details of how equal opportunity should be defined, they agree on the rejection of utilitarianism (as does, as well, Sen). So the repeated attempt of the Report’s authors to justify their concern with equity by a bottom-line endorsement of utilitarianism is inconsistent, and missing the main lesson in the evolution of political philosophy in the last 40 years. Furthermore, the claim that improving equity is the best way to maximize ‘prosperity’ or GDP per capita is surely false. The easiest way to see this is to note that, except in singular situations, one cannot simultaneously maximize two objective functions. The maximization of the program P in (1)
will almost never lead to the same policy choice as the maximization of v xi (ϕ) . To make this point slightly more formally, consider a very small country with just two individuals, each of whom comprises one type. Suppose that total income is maximized at point P in income space (see Figure 1a), and also opportunities for income are maximinned at P. In this simple case, the indifference curve of the opportunity – maximizing welfare function (1) is the rectangular Leontief curve LL, and the indifference curve of the utilitarian social welfare function is the straight line UU. The simultaneous double maximization means that the point P belongs to the income-possibilities set, but there are no points of the incomes-possibility set lying in the shaded region illustrated in Figure 1b. But this can only be the case if the income-possibilities set has a kink, a non-differentiable boundary, at P. This is the singularity referred to above. In other words, the claim that maximizing equity will maximize prosperity or GDP per capita can hold in a world with non-differentiable income-possibilities sets. This is a perhaps pedantic way of saying that the growth-with-equity school, which maintains that growth in GDP per capita is maximized by maximizing equity is almost surely making an opportunistic argument. Perhaps a correct and honest statement of the view would be “Maximizing equity is important as an end in itself, and probably the sacrifice in GDP per capita that would be endured by doing so is quite small.” I am willing to believe that this may be the case, but I am not willing to believe the stronger

x2

x2 U

U

L

P

L

P

L

L

U

U

x1

a

x1

b

Figure 1 Simultaneous maximization of maximin and utilitarian social choice functions.

J Econ Inequal (2006) 4: 233–244

239

case that the Report attempts to make, of possible simultaneous maximization of the two objectives. There is, moreover, always the possibility that maximizing equity will require some substantial sacrifice in GDP per capita. The Report implicitly argues that this is not the case, because various market imperfections (e.g., in credit markets) prevent poor, talented individuals from realizing their potential; and if that potential were realized by improving their opportunities, then everyone would benefit – or more circumspectly, GDP per capita would grow. This may indeed be the case in many developing countries, but it is not necessarily and always the case. To defend my claim, I cite results of Betts and Roemer [3], in which we ask what policy of educational finance would equalize opportunities for achieving wage earning capacity among children in the United States. We study this question for various different typologies. Typology 1 takes parental education as the unique circumstance; we partition the population of young male workers into four types, defined by the level of their mother’s education. We take the set of feasible policies to be the possible distributions of the US educational budget, where different amounts of investment can be targeted on each of the four types of student. We compute that the equal-opportunity policy would indeed increase total earnings by about 2.6%. This is a case where equity is good for growth. But we then study Typology 2, where again there are four types, but this time both parental education and race are taken as circumstances. Now, the equal opportunity policy reduces total wages by 2%. In other words, equity is bad for growth. In this case, the policy maker is forced to choose: Is she pro-equity or -growth – is she an opportunity egalitarian or a utilitarian? I do not think that one can deny that such cases are likely to occur fairly often – at least in the short run. If one understands the simple mathematical principle that two different objective functions can rarely be maximized by the same policy, then there will be in theory cases where maximizing one of them reduces the value of the other, and given the complex world we live in, these cases will occur. I do not mean to heap all the criticism on the growth-with-equity school; the same peccadillo is committed by the neo-liberal school of development, which tries to argue that reducing poverty is best accomplished by maximizing entrepreneurial opportunities for the wealthy. Again, these are two different goals, and they will not in general be maximized with the same policy. Neo-liberalism, however, is, thankfully, not the philosophy of the Report upon which I am here commenting.

4. What is efficiency? As I have noted, the Report does not make explicit its definition of development, but from its statements, we may deduce that it accepts the classical definition of development as measured by GDP per capita, or the rate of development as measured by the rate of growth of GDP per capita. There is a similar ambiguous treatment of the term efficiency in the Report. Efficiency is not defined by the authors. The standard meaning of efficiency in economic theory is Pareto efficiency (PE): A policy (in our case) is PE if it is not possible to find another policy that increases the welfare (or the advantage, or the opportunity for advantage) of all. But it seems that often the Report uses efficiency in the utilitarian sense: That is, a policy is efficient if it is impossible to find another

240

J Econ Inequal (2006) 4: 233–244

policy that engenders a larger GDP per capita (or a faster rate of growth of such). For instance: there may be various short-run policy level trade-offs between equity and efficiency (p. 3) Greater equity implies more efficient economic functioning (p. 3) This inequality is objectionable on both intrinsic and instrumental grounds. It contributes to economic inefficiency....(p. 9) Related to this is the view that inefficient institutions are often perpetrated by the powerful: ...market imperfections may arise not by accident but because they distribute income and power in particular ways. In this view, there will be social conflict over the institutions of society and incentives for people who control power to shape institutions in ways that benefit them (p. 8). I prefer the following triple nomenclature: Technological efficiency, Pareto efficiency, and social efficiency. An allocation is technologically efficient if, with the same bundle of inputs, it is impossible to achieve a vector of outputs that dominates, component-wise, the first vector of outputs. Pareto efficiency was defined above. Social efficiency means that the allocation (or policy) maximizes the social objective function – in the case of this Report, equity. If one writes ‘efficiency’ without a modifier, the default meaning, for economists, is Pareto efficiency. The most sensible interpretation of some of the quotations I have given is that efficiency means ‘social efficiency in the utilitarian sense,’ namely, maximizing GDP per capita or its growth rate. That is inconsistent with the concept of social efficiency which the Report has advocated. As I have been at pains to say in the last section, this Report should consistently advocate equity as the concept of development, and hence social efficiency as the maximization of equity – formally, of the program stated in (1). Consider the last quote reproduced above. ‘Market imperfections’ are always identified by economists as a source of Pareto inefficiency. But in this quotation, the authors are saying that market imperfections may often be maintained because they benefit the wealthy and powerful. But if market imperfections are ipso facto Pareto inefficient, then why cannot these powerful rich people find a set of institutions that Pareto-dominates the present ones, which, by definition, would make them better off? In other words, either the economic theory is wrong, or the wealthy and powerful are not really that powerful – not powerful enough to implement the Pareto-dominating allocation that could be achieved by reducing market imperfections. Market imperfections that exist may well arise by accident, not by the conscious manipulation of the rich and powerful. The last quotation smacks of functionalism – the error of asserting that because an institution benefits some powerful social class, that class must have brought it about. The best insurance against functionalist errors is to always search for mechanisms by which the benefiting group actually does bring the benefiting institution into being.1 As far as I can see, the Report does not do this.

1 The strongest advocate of this view is Elster [7].

J Econ Inequal (2006) 4: 233–244

241

My own view is that imperfect markets may well be part of a Pareto efficient set of institutions, in the sense that, given what is politically feasible, eliminating those imperfections would harm some individuals – namely, the rich and powerful. But if this is the case, then the authors of the Report should not claim that reducing imperfections is in everyone’s interest, or would be a Pareto improvement. Doing so might well improve equity, and as well might improve the rate of growth of GDP, but it also might not be a Pareto improvement, in the sense that the wealthy would suffer. On a more historical note, let me also express some skepticism about the claim that improving equity will always improve growth. The Report argues (p. 9) that the difference between the successful development path of North America and the unsuccessful one of South America was due to there being labor scarcity in the North and labor abundance in the South. The argument is that, in the South, unskilled labor was prevalent, and wealth could be created by mining and plantation agriculture, without technological innovation and increasing labor productivity. In North America (with the exception of the US South) this was not the case; capitalists had to compete for scarce labor, and introduce technical innovation to save on labor. Workers in the North won the franchise, and other liberties, because their scarcity forced the wealthy and powerful to grant them concessions. Thus, equitable institutions had ‘positive consequences for long-run economic development.’ I am not a historian, and cannot judge this claim from my own research. But there was a long debate in the journal Past and Present, in which the historian Robert Brenner argued quite the opposite: That the reason France’s industrial revolution came late, and Britain’s early, was that the French yeomen peasantry succeeded in winning a large degree of independence, and were able to live as subsistence farmers without innovating or become proletarianized, whereas in Britain, weaker farmers were driven off the land, proletarianized, and capitalist competition among both large agriculturalists and manufacturers led to rapid economic development. Here, then, we have two class-oriented tales, with precisely opposite conclusions.2

5. How much inequality is due to inequity? One of the important analyses of the Report is presented in chapter 2, where the authors attempt to measure the proportion of inequality that is due to inequity, in various countries. For instance, Figure 2.11 takes the circumstance to be education of the household head, and reports the fraction of income inequality which is explained by type variation, as opposed to ‘within type.’ In other words, this figure reports an attribution of total income inequality to inequality between distributions of income, by education of household head, and the inequality within such distributions. Two different ways of making the attribution are computed. In Brazil, over 30% of income inequality is due to such inter-type inequality while in Belgium, less than 5% is so explained. From the figure, it appears that the degree of inequality due to inequity is fairly low in the high-income economies (Luxembourg is the highest, with about 15%) and East Asia (for the most part), and it is high in Latin America and sub-Saharan Africa. This would suggest that as countries become wealthier, equity improves.

2 The exchanges in this debate are available in Aston and Philpin [2].

242

J Econ Inequal (2006) 4: 233–244

The caveat here is that there is only one circumstance: The education of the household head. That is not an optimal choice of circumstance, because the choices that determine income overlap with the choices that determine the education of the household head. In other words, if low effort determines low education of the household head, it may also determine low income of that family. It would be better to look at the earnings of the children of parents typed by their level of education, but I am sure those data are available for only a handful of countries at present. Because one should, ideally, take into account many other circumstances, one can confidently conjecture that the proportions of inequality due to inequity reported in this chapter are only lower bounds for the true proportions: The authors make this point. Moreover, one cannot be confident that the proportion of inequality due to inequity truly falls with increasing wealth of the country, because it may be the case that the salient circumstances change with wealth. For instance, natural talent may be the circumstance that accounts increasingly for high incomes, as markets become more efficient. And is this not a circumstance that should count with respect to our conception of inequity? Is it morally right that those with high natural talent should have higher consumption than those with less? Perhaps we would consider it a great achievement if the only circumstance which made a difference with regard to income were natural talent; but as wealth increases, is it not right that nations should cast the net of circumstances more broadly? When we eliminate inequality of opportunity due to typologies based upon caste, race, parental education, and parental social connections, should we not start including certain genetic dispositions as grounds for social compensation? We already do so in advanced democracies, to some extent: Consider the Americans with Disabilities Act (ADA), which requires employers to equalize opportunities for functioning at work for disabled workers. American businesses and other institutions have had to spend many millions of dollars, for example, to install wheelchair ramps to conform to the ADA.

6. Global inequality The chapter on global inequality takes, as the circumstance, the country a person is born in. Many have argued that birthplace is morally arbitrary, and this leads to what is called the cosmopolitan egalitarian view, that advantage (say, income) should be equalized across countries. It also may imply a liberal immigration policy. The Report argues, correctly in my view, that global inequalities are unjust, and it is important to take steps to rectify them. Here, I wish only to comment that the Report is a bit ahead of contemporary political philosophy: By this I mean that there are some political philosophers who advocate egalitarian measures within nations, but not across them. Two names to mention are John Rawls and Thomas Nagel. Rawls [10], in his writing on global inequality, did not advocate applying the difference principle (roughly speaking, equality of opportunity). And recently, Nagel [8], who would agree with this Report’s stance towards within-nation inequality, would not extend that ethic to inequality among nations. Nagel argues that justice is a concept that applies only to collections of citizens who belong to a state, and because there is no supra-national state, the concept of justice does not apply globally. Citizens of rich states, he maintains, have no obligations to citizens of poor ones beyond the prevention of extreme deprivation, and whatever charity might dictate.

J Econ Inequal (2006) 4: 233–244

243

I believe that Nagel is wrong; for a response to his view, see Cohen and Sabel [5]. My intent here is only to point out that the philosophical authority for maintaining that country of birth is properly taken to be a circumstance compensable at the bar of justice is still in contention, although I, personally, think that in a century or two, it is global inequality that will seem to be the most morally egregious.

7. Conclusion The authors of “Equity and development” have made an important contribution both to the literature on development and welfare economics: This book is the most thorough application to date, of which I am aware, of important ideas in recent political philosophy to a massive world problem. Not only will it change the terms of the debate in development circles (at least, so I hope), but it shows the importance of political philosophy for welfare economics. Economists have, far too frequently, lauded themselves on their discipline’s being ‘value free’, and eschewed any justice talk in favor of efficiency talk. This is, I believe, a mistake, and the publication in question makes a strong case against that narrow view. In critique, I say the Report has not gone far enough. It correctly takes inequality of opportunity as the appropriate concept of inequity, but then goes on to argue that there is no contradiction between that view and utilitarianism of a certain narrow type, namely, the view that justice recommends the maximization of per capita income. I have belabored this point above; but see also page 78 of the Report, where the authors backtrack and say they are not necessarily recommending the equalization of opportunities, but only that opportunity be the argument of the social objective function. In the highlighted box on this page, they say they take no stand as between equalizing opportunities across types, or maximizing the sum of opportunities across types, or maximinning opportunities across types. I have argued that only the last of these corresponds to the notion of equity advanced elsewhere in the Report. Note that if one takes ‘mean income of the type’ as the measure of its opportunity, then maximizing the sum of opportunities across types, if they are (reasonably) weighted by their population fractions, becomes: Maximize GDP per capita. I can only conjecture that the language on page 78 was a concession to a minority view on the editorial committee of the Report. It is inconsistent with the rest of the Report. I believe that the conception of development which is consonant with the Report’s definition of equity is the value µ∗ of the program in (1). The rate of economic development should be taken to be the rate at which the mean advantage level of the worst-off type grows over time. I agree that maximizing this objective will probably, in most cases, lead to an increase in GDP per capita, although I disagree that it will maximize the rate of that growth. I look forward to a future number of the WDR that carries out the computation, across countries, of this new definition of economic development. References 1. Arneson, R.: Equality and equal opportunity for welfare. Philos. Stud. 56, 77–93 (1989) 2. Aston, T.H., Philpin, C.H.E. (eds.): The Brenner Debate: Agrarian Class Structure and Economic Development in Pre-Industrial Europe. Cambridge University Press, Cambridge (1985)

244

J Econ Inequal (2006) 4: 233–244

3. Betts, J., Roemer, J.: Equalizing opportunity for racial and socioeconomic groups in the United States through educational finance reform. In: Peterson, P. (ed.) Schooling and Equal Opportunity. MIT, Cambridge, MA (2006) 4. Cohen, G.A.: On the currency of egalitarian justice. Ethics 99, 906–944 (1989) 5. Cohen, J., Sabel, C.: Extra rempublicam nulla justicia? Philos. Public Aff. (forthcoming) 6. Dworkin, R.: What is equality? Part 1. Equality of welfare. Philos. Public Aff. 10, 185–246 (1981) (and What is equality? Part 2. Equality of resources. Philos. Public Aff. 10, 283–345) 7. Elster, J.: Nuts and Bolts for the Social Sciences. Cambridge University Press, Cambridge (1989) 8. Nagel, T.: The problem of global justice. Philos. Public Aff. 33, 113–180 (2005) 9. Rawls, J.: Theory of Justice. Harvard University Press, (1971) 10. Rawls, J.: The Law of Peoples. Harvard University Press, MA (1999) 11. Roemer, J.:A pragmatic theory of responsibility for the egalitarian planners. Philos. Public Aff. 22, 146–166 (1993) 12. Roemer, J.: Equality of Opportunity. Harvard University Press, MA (1998) 13. Roemer, J.: Equality of opportunity. In: Blume, L., Durlauf, S. (eds.) The Palgrave Dictionary of Economics, 3rd edn. (in press) 14. Sen, A.: Equality of what?. In: McMurrin, S. (ed.) The Tanner Lectures on Human Values. University of Utah (1980) 15. Sen, A.: Commodities and Capabilities. North-Holland (1985) 16. Singer, P.: Animal Liberation: A New Ethics for our Treatment of Animals. Random House, New York (1975)