Version classiqueVersion mobile

Arguing about justice

Axel Gosseries
Philippe Vanderborght

The ideological roots of inequality and what is to be done

John E. Roemer


La philosophie anti-étatiste, dont Robert Nozick est le représentant le plus exemplaire, a fourni le fondement idéologique des inégalités contemporaines. La théorie économique y a également contribué, en faisant jouer un rôle moins central a la théorie de l’équilibre général pour se focaliser sur la théorie du contrat, la fonction principale des marchés étant alors de fournir des incitants, plutôt que de coordonner l’activité économique. Cet accent placé sur les incitants a suscité le pessimisme quant a la faisabilité d’une fiscalité redistributive. Je défends ici l’idée selon laquelle ce recentrage est déplacé. La polarisation extrême des revenus qui caractérise l’économie américaine ne répond en rien a une nécessité d’efficience économique. Au contraire, elle entrave cette dernière.

Note de l’auteur

Note portant sur l'auteur1

Texte intégral

Three arguments for inequality

1To address properly the growing inequality in the advanced democracies, particularly in the United States but also increasingly in the UK and continental Europe, I believe it is necessary to review the main arguments for inequality, and to inquire into their legitimacy. There are, today, two main arguments for inequality: first, an ethical one, that individuals deserve to benefit from what nature and nurture endows them with, and second, an instrumental one, that inequality is good for everyone. The first argument is presented in its most compelling form by the philosopher Robert Nozick, who in his 1974 book, Anarchy, State and Utopia, advanced the idea that a person has a right to own himself and his powers, and to benefit by virtue of any good luck that may befall him, such as the luck of being born into a rich family, or in a rich nation. Any voluntary exchanges that take place between persons are legitimate, and in this way, it is not hard to imagine that a highly unequal distribution of income and wealth can be built up fairly rapidly from these unequal endowments. Nozick is the first to admit that actual capitalist economies are not characterized by historical sequences of legitimate, voluntary exchanges: there is much coercion, corruption, and theft in the history of all societies. But Nozick’s point is that one can imagine a capitalism with a clean history, in which vastly unequal endowments of wealth are built up entirely from exchanges between highly talented, well educated people and simple, unskilled ones, and this unequal result is ethically acceptable if one accepts the premise that one has a right to benefit by virtue of one’s endowments – biological, familial, and social – or so he claims.

  • 1 G.A. Cohen (1997) has argued that Rawls is inconsistent in calling just allocations in which some (...)

2The second major argument for inequality is the instrumental one: that only by allowing highly talented persons to keep a large fraction of the wealth that they help in creating will that creativity flourish, which redounds to the benefit of all, through what is informally called the trickledown process. In a word, material incentives are necessary to engender the creativity in that small fraction of humanity who have the potential for it, and state interventions, primarily through income taxation, which reduce those material rewards, will kill the goose that lays the golden eggs. Even among left-liberal political philosophers, the incentive argument for inequality is acknowledged: John Rawls describes as just inequality that is necessary, for incentive reasons, to render the worst-off class as well off as possible1.

3A third argument for inequality, which is currently most prevalent in the United States, is one of futility: even if the degree of inequality that comes with laissez-faire is not socially necessary in the sense that the incentive argument claims, attempts by the state to reduce it will come to naught, because the government is grossly incompetent, inefficient, or corrupt. Thus, better to let the rich keep their wealth and invest it profitably, than to hope that the state can manage it more fruitfully. Incredible as it may seem, it is now becoming increasingly popular in certain circles of economists in the United States to say that the productivity of government investment is zero.

How current economic theory justifies inequality

4I wish to discuss how economic theory influences our views about inequality. I will dismiss macro-economic theory quite curtly. Its two pillars, rational expectations and the efficient markets hypothesis, have both been quite effectively demolished by the current financial and economic crisis. Rational expectations is simply a premise of mathematical convenience. Suppose that agents are making decisions under uncertainty, which is, of course, ubiquitous in the real world. Suppose, for example, each individual is deciding how much education to take, under an assumption about the probability distribution of wages associated with each level of education he/she might acquire. If one is interested in the long-run properties of such a process, one needs to make some assumption about the relationship between the prior distributions of wages conditional upon educational levels that people hold, and the actual distribution of wages which occurs after people are educated and acquire jobs. The simplest – and indeed, focal— assumption that one can make is that these distributions are the same: that is, people take, as their prior, the distribution that is in fact realized. This is the rational expectations hypothesis. Although mathematically convenient as a way of closing the model, the assumption is impossible to justify as empirically accurate. The reason the assumption is convenient is that there is no simple way to close the model without it.

5The efficient markets hypothesis says that the prices in stock markets accurately summarize all the information that is available about the expectations of the future values of firms. A consequence of this hypothesis is that bubbles cannot exist in stock markets – or , if housing is efficiently priced, in housing markets.

6The financial crisis and concomitant economic crisis which we have recently experienced are not possible under these two views. Diehard advocates of these views argue that there was no housing bubble, and there is currently no involuntary unemployment: those who are unemployed are rationally investing in the search for new jobs. Somewhat less militant advocates of these views do not reject the claim that a crisis occurred, but rather construct Ptolemaic arguments purporting to demonstrate that nothing could have been done to prevent it, and that, in particular, any aggressive intervention by the state would have then or will now only make things worse.

7My own take is that the hold these views have on macro-economists demonstrates that economics is not yet a science. For a theory to be a science, it must be falsifiable. I do not see how any real event could falsify the main tenets of current macro-economic theory, if the current crisis could not. History suggests that individuals who have long-held views or methods do not change: we can only hope that a new generation of macro-economists will emerge, whose members have the courage and intellectual power to reconstruct the theory.

8There has also been an important transformation in micro-economic theory during the past thirty years, and although less noted, it may be just as important as the rational-expectations revolution in macroeconomics, for our views on inequality. Economists have long realized that markets perform two functions: they coordinate economic activity, and they provide incentives for the development of skills and innovations. It is not easy to give a definition which distinguishes precisely between these two functions, but there is no question that a conceptual distinction exists.

9Let me give two examples. An important problem facing Soviet planners was called the transportation problem. There are a number of plants in the nation which produce steel, with various fixed capacities, and a number of locations in the nation which demand steel – in various quantities. There is a cost of shipping steel from any supplier to any demander – these costs vary. What is the optimal plan? How much should each steel producer ship to each market, so that all demands are satisfied, and transportation costs are minimized? This problem led the Soviet mathematician L. Kantorovich to co-invent linear programming in the 1940s, for which he and Tjalling Koopmans, in the United States, jointly received the Nobel prize in economics.

10Competitive prices can solve this problem without linear programming. If competitive prices for transport and prices of steel at the various producers characterize this market, then each demander orders steel to minimize its costs, the steel market clears, and indeed, the market will have found the mathematical solution to the cost-minimization problem. This is an example of prices performing a coordinating function.

11Consider, now, the problem that I alluded to earlier. A young person is deciding on how much education to take, and wishes to maximize a utility function in which both future income and the nature of the occupation are important. Prices (which is to say expected wages) guide him/her to make the decision. We should think of this as an instance of where the market is performing an incentive function – it is inducing the individual to make a choice on which skill to develop based on his self-interest, and if the market is well functioning (competitive), the aggregation of all such choices will be Pareto efficient.

12Now neither of these examples is pure. Indeed, in the transportation problem, equilibrium prices are realized as the result of managers of firms having the incentive, for some reason or other, to minimize costs. In the education example, the market is coordinating the allocation of workers to occupations in a way which is, under some assumptions, Pareto efficient. And, indeed, in any example, there will be both a coordination and an incentive interpretation of the market’s role. I maintain, nevertheless, that there is a distinction between these two roles, and that each example illustrates different mixes of the coordination and incentive functions of the market.

13In the last thirty or forty years, the economic theorist’s view of the market has changed, from being an institution which performs primarily a coordination function to one that is primarily harnessing incentives. Indeed, the old definition of micro-economics was the study of how to allocate scarce resources to competing needs. (One will find this statement of the task of economic science in almost any textbook published before 1970.) This is entirely a coordination view. The elegant theorem at the pinnacle of this view is that equilibrium prices in a competitive economy with no externalities, no public goods and a complete set of markets will induce an allocation of resources which is Pareto efficient. That kind of efficiency is the mark – perhaps the definition of -- successful coordination. The technique that provided the intellectual justification of the coordination view was general equilibrium theory: it reached its apex in the 1950s and ’60s.

14It may surprise you to hear that the phrase ’principal-agent problem’ was only introduced into economics in 1973, in an article by Steven Ross (1973). In the principal-agent problem, coordination is not the primary issue – rather, a principal must design a contract to extract optimal performance from an agent whose behavior he cannot perfectly observe. This is par excellence an incentive problem.

15The technique which developed to analyze incentive problems was game theory – more specifically, parts of game theory called contract theory and mechanism-design theory. The problems that are analyzed typically involve small numbers of individuals, where markets do not exist, and the individuals must set prices or write contracts to ’get the incentives right.’ But the idea that incentives are the real guiding force which markets must address extends to much larger groups of actors – even to entire economies. Some economists have even gone so far as to say that prices have no coordinating function at all: rather, they say we should look at all economic exchanges as episodes of bargaining between individuals; what determines the price of the exchange is the relative power of the two bargainers – in the sense of how scarce the commodity they are each supplying is – and prices simply ratify these fundamentals. Under this view, markets do not coordinate at all, they simply provide the venue for bargaining.

16What is the import of the change in emphasis of the market’s role for the problem of inequality? I believe it is to reinforce the importance of material incentives: more radically, to justify the view that the very high salaries and incomes, for those at the top, that have come to characterize capitalist economies in the last thirty years, are necessary for efficiency.

17To see this, I propose a simple thought experiment. Suppose that there were an economy where individuals cared only about the kind of occupation in which they engaged, and the amount of education they required to qualify for the occupation. Assume that there is a variety of preferences in the population over occupations and education. Wages are not important to people, as long as they suffice to purchase a certain minimum of consumption. There is one firm, which produces a single consumption good that everyone needs, and the firm can produce this good using many different possible combinations of workers employed in the various occupations. The firm wishes to maximize its profits, which will equal the revenues from selling the output to consumer-workers, minus the wage costs of the workers. A vector of occupational wages, and a price for the output good, comprise a competitive equilibrium for this economy if, at those wages and prices, all the occupational labor markets clear and indeed the demands for workers in the various occupations put forth by the firm, maximize its possible profits at those wages and price.

  • 2 For a more careful discussion of this model, see Roemer (2010).

18Now the peculiar thing about this economy is that workers do not care about wages: they make decisions about what occupation to engage in solely by looking at the attributes of the occupations and the educational requirements to enter them. Each worker chooses the occupation and concomitant educational level which maximizes his utility over the available choices. Equilibrium wages are ones which will direct the profit-maximizing firm to demand exactly the supplies of occupations which are on offer by the population2.

19I propose that in this economy, the market is providing primarily a coordination function. Granted, the firm has, by hypothesis, an incentive to maximize profits: but in some sense the main function of the market is to coordinate the fixed occupational choices of workers with the firm’s demands for workers in the various occupations. Some workers will, indeed, earn more than others at the equilibrium, because their wages are higher: but, by hypothesis, as long as all wages are sufficiently high with respect to the price of the output, these wages form no part of workers’ decisions.

20Suppose one now imposed a purely redistributive income tax on workers. Then the occupational choices of workers will be unchanged. Indeed, the equilibrium wages and price of output will not change. All workers will choose exactly the occupations they chose before, and the firm will produce the same amount of output as before, with the same profits: all that will happen is a redistribution of the consumption good among worker-consumers. This is a case where redistributive taxation has no effects on occupational and productive choices. There are no efficiency consequences to taxation.

21Consider, now, the polar opposite model, where workers care only about income and the education they must endure to qualify for different occupations. They do not take into consideration the characteristics of occupations at all. A wage vector will still allocate workers to different occupations, but this time the nature of the occupation is unimportant for workers – the wage makes it possible for them to optimize their trade-off between income and time-consuming education. Here, wages are providing an important incentive function, and workers will choose different occupations only because they have different degrees of antipathy towards education, or to time lost from earning income while going to school. If we introduce an income tax, occupational choices will change: indeed, there may be considerable real effects on occupational structure if the tax is sufficiently large; the efficiency consequences of taxation are no longer costless.

22The punch line I am proposing is this: to the extent that the market is primarily a device for coordination, taxation can redistribute income without massive efficiency costs. But if the market is primarily a device for harnessing incentives, the efficiency costs of redistribution may be high. I believe that the transformation in micro-economic theory that I have described – one that views the market primarily as harnessing incentives, as opposed to coordinating -- provides an important intellectual foundation for the inadvisability of interfering with market rewards via a re-distributionsist state policy.

Ideological strategies for egalitarian policy

23There are three key ideas that we must address, to provide a sound intellectual basis for advocacy of a more equal society – I think that this intellectual task is of utmost importance. Ideas really do trickle down, even if money does not! We must argue:

24First, that children should not be disadvantaged by the bad luck of being born into a poor or poorly educated family. I believe that there is considerable support for this idea in the advanced democracies – but often the consequences of this view have not been thoroughly understood by citizens, or implemented in policy. Consider education, for example. In the United States, the average amount of state (federal, state, municipal) funding for the education of advantaged children is far greater than for disadvantaged children. A moderate liberal goal is to equalize funding for all children. But the correct policy is to invest more state funds in disadvantaged children than in advantaged ones, to compensate them for their family’s lack of resources. To the extent that disadvantaged children are of a minority race or a minority nationality, it is more difficult to win citizens to this implementation of the requirements of equal opportunity, due to racism and anti-immigrant sentiment. Intellectually, winning this battle means confronting Nozick’s libertarianism, and in particular the tenet that a person has a right to benefit from his or her good luck in the birth lottery (Betts & Roemer 2007).

25A more radical view, that I also hold, is that we must extend the compensation to include genetic bad luck. There is no reason, from the ethical viewpoint, that individuals who are less mentally endowed should have poorer material conditions than the better endowed. To the extent that this occurs, it can only be excused by the necessity of incentive payments to wage earners and the infeasibility of redistribution. There is no ethical basis for it, for such a basis can only exist by advocating the right to benefit from good luck.

26Secondly, we must understand the true division of labor in the market’s coordination and incentive roles. I think this is, intellectually, the most difficult problem and probably the most politically important one. What I here propose is a question for research, rather than an answer. As I said earlier, I believe that the degree to which redistribution is feasible with small efficiency costs is closely related to how important material incentives are for productive activity. Although economic theory has shifted on this question during the last generation, it is far from obvious that the shift is empirically justified.

27My current view is that material incentives are important for the great majority of workers, in their role in directing educational and occupational choices, but that they are far less important for the high fliers at the top of the income and wealth distribution, and indeed for innovation. I do not mean that the high fliers and innovators do not like their huge incomes, but rather that these incomes are not socially necessary, in the sense that we would lose innovation and good corporate management without them.

28Many people think that the high incomes paid to corporate management, and in particular to those in the financial sector, are due to monopolistic elements – back-scratching among interlocking directorates of firms made possible by shareholder apathy, for example. I do not endorse this view; I believe it is entirely possible that the high salaries and bonuses these people receive are competitive, in the sense that each is receiving his expected marginal product as evaluated by the board that hires him. Even if this is not so, and those monopolistic elements do exist, I will make my argument under the competitive assumption, because it makes the argument stronger. In other words, I claim that even if extremely high CEO salaries are competitive, they are inefficient. Their existence is a market failure.

29The theoretical argument for the possibility that these salaries are competitive is that, if you are the pilot of a very large ship in Arctic waters, a very small change you make in the angle of the rudder can significantly affect the probability of the ship’s impact with an iceberg. If you are the CEO of a large corporation, with $10 billion in annual revenues, a very small difference in the quality of your decisions may make a 1% difference in annual revenues, which is to say a $100 million difference. If the hiring board is convinced that you are just slightly better than the next candidate, you can probably bargain for $25 million or so of your $100 million marginal product, or maybe more.

30But the payment of these high salaries is inefficient because, even if the expected marginal product of the incumbent CEO is $100 million greater than the next best candidate, the social costs of creating a class with such huge wealth are large. First, the institution creates a class which has great power to influence politics. Members of this class, if private campaign financing is legal, will make large contributions to political parties to maintain their privileges. Secondly, as we have seen in the financial sector, paying such huge salaries on the basis of expected marginal product can induce behavior that is far too risky from the social viewpoint. This negative externality is not internalized by the firm, because, inter alia, of the implicit insurance provided by the state against failure in the case of a bad outcome on these risky gambles. This is why the US government’s allowing Lehman Bros. to go bankrupt in September 2008 so scared the financial sector, and why the government effectively reversed itself a few days later by rescuing the insurance firm AIG.

31My final argument is based on the claim that the most important reward for the very high fliers is not money income, but power – power over others, and the respect of their peers. Therefore, if these high salaries were taxed at very high rates, we would not see a significant change in productive behavior. (One might have to make the taxation international, to prevent migration.) What other occupation could these high fliers choose which would give them the same power and respect of their peers? Will they, alternatively, be satisfied to retire and play golf or read poetry?

32I have argued that these high incomes are inefficient, because of risk-taking externalities that they induce, that they are unnecessary for incentive provision, and that they create a class with disproportionate political power. Finally, there is the very important negative externality of the creation of a social ethos which worships wealth. Ordinary people lose self-esteem, and do what they can to ape the consumption behavior of the very wealthy.

33In sum, the positive social value of the institution of extremely high salaries that the leaders of the corporate world, and in particular, of the financial sector, receive, is a big lie. It may well be a competitive outcome, but it is a market failure which could be corrected by regulation or legislation. In the US, at least, there are no visible prospects for such legislation, and I maintain that this is due, in no small part, to the hold of the incentive argument for inequality on the voting population.

34I believe, in contrast, that incentive pay for the working population in general is important. The experience of China since 1979 is an example: I do not attribute the Chinese growth rate of over 8% per annum in this period entirely to the lack of coordination which existed prior to 1979 due to the absence of markets. On the other hand, I do not think that material incentives, especially in the wealthier countries, are as important as many economists insist. We see a very large difference in relative material rewards to skilled labor across the advanced democracies – at its most extreme, between the Nordic countries and the United States – without concomitant efficiency differences. Once basic needs are met, I believe that people put substantial weight on the nature of their work.

35Many people could increase their income by changing occupations, but we do not, because we think our work is important, or we find it interesting. So although material incentives are important for ordinary workers, they are probably the least important for those workers who pay significant taxes in market economies – namely, those whose incomes are high, and have interesting careers. If my conjecture is correct, there would be no major impediments to substantial redistribution, on account of incentive problems.

What can we do?

  • 3 For a summary of the state of this research, see Salverda, Nolan & Smeeding (2009)

36Most social scientists who study inequality have focused upon measuring its degree, and studying the micro mechanisms through which it is produced3. Micro mechanisms include a substantial variety of processes: how tax policy contributes to the distribution of wealth, how the intergenerational transmission of income and wealth occur, what role education plays or could play in modifying the distribution of income, how neighborhoods and social networks affect behavior and choices which have an impact on distribution.

37Of course we must understand these processes. But in a sense, I think understanding them is of secondary importance, with regard to the massively important phenomenon of increasing inequality in the advanced democracies. I believe that democracy works in the advanced countries, in the sense that policies can only survive if they are approved by the majority (or a substantial fraction) of the electorate. Now the electorate may have poor information, or may have learned an ideology which does not serve its true interests, or may subscribe to a social ethos which prizes a kind of de jure individual freedom over a secure livelihood, but in the end, the electorate must acquiesce to social policy in our advanced democracies. This is surely one great accomplishment of the political revolutions that began in 1776 and 1789.

38Today, I think the most important problem for the social science of inequality is understanding how electorates have come to acquiesce to policies which increase inequality, and to try and reverse this acquiescence. To an extent, revealing the logic of the micro mechanisms is one strategy for doing so. A second strategy is the philosophical one, of arguing that justice requires equality – the last 40 years in political philosophy have seen a lively debate about exactly what kind of, and how much of that kind of equality, is required. A third strategy is showing how the inequality-increasing policies that electorates have agreed to have been aggressively advocated and proselytized by the wealthy, in order to maintain their material privileges. This work is mainly done by historians and some political scientists. A fourth strategy is critiquing the social ethos of individualism and greed which has grown immensely since 1980: the ethos to replace it is one of solidarity. This strategy overlaps with the third one, since (I believe) the nurturing of the individualistic ethos has been a conscious strategy of the wealthy, and perhaps its most powerful weapon. Finally, I believe, as I have said, that we must do the research to challenge the view that interfering with the incentives the market provides necessarily reduces economic welfare.

39We face a situation where we must move from describing the phenomenon we study, to figuring out how to bring evidence to bear against the ideology which sanctions it. That ideology has created a monster, capable of inflicting truly disastrous consequences on civilization. Because theory and ideas are our metier, we have a vital role to play.



BETTS, J. & ROEMER, J. (2007), 'Equalizing opportunity for racial and socioeconomic groups in the United States through educational finance reform', in P. Peterson (ed.), Schools and the equal opportunity problem, Cambridge, MA: MIT Press, 209-238.

COHEN, G. (1997), 'Where the action is: On the site of distributive justice', Philosophy & Public Affairs, 26: 3-30.

ROEMER, J. (2010), 'Incentive provision and coordination as market functions', Journal of Labour Market Research, 43: 6-16.

ROSS, S. (1973), 'The economic theory of agency: The principal’s problem', American Economical Review, 63: 134-139.

SALVERDA, W., NOLAN, B. & SMEEDING, T.M. (2009), The Oxford Handbook of Economic Inequality, Oxford: Oxford University Press.


1 G.A. Cohen (1997) has argued that Rawls is inconsistent in calling just allocations in which some are much better off than others, when this differential is caused by personal, selfish decisions.

2 For a more careful discussion of this model, see Roemer (2010).

3 For a summary of the state of this research, see Salverda, Nolan & Smeeding (2009)

Notes de fin

1 This paper is based on a lecture given at a meeting (February 2011) of the GINI Project, a European, multi-disciplinary social-science effort to study and understand the roots of economic inequality in contemporary advanced societies.

© Presses universitaires de Louvain, 2011

Licence OpenEdition Books

Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search