THREE SOURCES OF ECONOMIC INEQUALITY

Abstract

There are three distinct forces that conspire to produce a great deal of economic misery. We can refer to them, for convenience, as misfortune, unfairness, and improvidence. Political philosophers have often shown an interest in one or another of these, but seldom all three. Furthermore, those who do acknowledge all three have often felt driven to collapse them into one root cause of inequality. My goal in this essay will be to argue that the three are independent of one another, but more importantly, that they each require distinct remedies. This is important for egalitarians because it defeats any attempt to develop a “one-size-fits-all” policy aimed at creating a more equal society. This analysis helps us to understand several of the tensions that arise in our attempts to combat inequality that are often obscured by an overemphasis on the power of redistributive taxation as well as generalized inattention to the way that successful welfare states achieve meaningful progress in promoting greater equality.

I. Introduction

My best friend died at the age of forty-three from a ruptured brain aneurysm. None of us was surprised. At his fortieth birthday party, we had all congratulated him quite earnestly for having made it that far. He lived what could be referred to, euphemistically, as “a hard life.” He was what some people call a schlimazel—that is, a chronically unlucky person. I got a call from him one evening. He was at the police station, needing me to come down and drive him home. Earlier in the day he had been riding his motorcycle and got sideswiped. The bike was totalled, but he managed to walk away. Later on, though, his leg was causing him pain, so he got in his van and drove himself to the hospital. While he was getting examined, thieves broke the side window of his van in the hospital parking lot and took everything in it. As he was driving home, he got pulled over by the police. Having noticed the broken side window, they naturally suspected him of having stolen the van. Unfortunately, while cleaning out the glove compartment the thieves had taken the registration, so he was unable to prove his ownership. The police impounded the van and threw him in jail. This sort of thing happened to my friend a lot. Bad luck seemed to follow him wherever he went.

He was also a drug addict. He spent several years off and on shooting up heroin. That is what eventually killed him. He managed to avoid HIV, but he did contract Hepatitis C, which was slowly destroying his liver. The brain aneurysm hastened his demise. Addicts often suffer from this, which is the result of impurities and bacteria in the drugs they inject that weaken their arteries. When he was off the heroin, he would drink too much (and, as I found out later, do cocaine). And smoking, always smoking. Every morning he would go out to the corner store and buy a pack of filterless Camel cigarettes, which he would finish off by the end of the evening. He never bought a carton, always by the pack. Then, he would go for coffee at his local haunt. Sometime past lunch, he would devise a plan for dinner and pass by the neighborhood shop to buy food. Being with him was like watching money burn. If he had just gotten in his van, driven to the grocery store, and made a few purchases in bulk, he could easily have halved his monthly food bill.

The odd thing is that, despite living in fairly abject conditions, he was never out of money for long. Although he had dropped out of high school, he later learned Linux and became a talented systems administrator. He had difficulty holding down a regular job, but he was able to line up a lot of freelance work, all of which paid quite well. In a good year he would make twice as much as I did (with my boring-but-stable professorial salary). Obviously, he had no savings; he would occasionally ask me for an emergency loan or a place to sleep. The point is that, among his many problems—and they truly were legion—an inability to earn market income was the least of his concerns.

The larger, philosophical point that I am building up to is that social and economic inequality is complex. My friend suffered from a lot of problems that are not uncommon among working-class men. Although his financial situation was something of a catastrophe—no assets, no savings—he was not exactly poor. In fact, had he been able to solve any of his other problems, he could easily have been quite well off. There was often a great deal of money flowing through his household; it was just not being effectively translated into a better quality of life. The reasons for this are not difficult to ascertain. There are three distinct forces that conspire to produce a great deal of economic misery. We can refer to them, for convenience, as misfortune, unfairness, and improvidence. Political philosophers have often shown an interest in one or another of these, but seldom all three. Furthermore, those who do acknowledge all three have often felt driven to collapse them into one root cause of inequality. My objective in this essay will be to argue that the three are independent of one another, but more importantly, that each calls for a different set of institutional arrangements to redress its effects. This is important for egalitarians because it explains why the mere redistribution of wealth is generally insufficient to remedy economic inequality, and thus why successful welfare states rely upon a more complex mix of policies in order to achieve this goal. ¹

II. Sources of Inequality

Social inequality is complex and multifactorial. Some important aspects of it, such as status inequality, are difficult even to measure. As a result, there has been an understandable tendency among egalitarians to focus on more tractable forms of inequality, such as inequality of income or wealth. Not only is economic inequality easier to measure, but it is also more amenable to policy intervention since most states have the organizational capacity to take money away from some people and give it to others. Yet even these more limited forms of inequality remain complex. While it is relatively easy to measure economic inequality, identifying the sources of this inequality—its etiology—is far more difficult. Among empirically oriented economists, for instance, there are few who would be so bold as to claim that they are able to offer a complete and decisive explanation of interindustry wage differentials or the gender gap in income, to pick just two prominent examples. ²

Normative theorists, by contrast, have gravitated toward relatively simple models of economic inequality. There are legitimate reasons for wanting to stylize somewhat in order to identify the normative issues more clearly. But there is also, unfortunately, a temptation to allow one’s normative views about the forms of inequality that one considers to be justifiable or unjustifiable to influence one’s empirical understanding of how the various forms of inequality come about. Most obviously, those who are keen to redress inequality may be more likely to attribute it to forces that are outside the control of those who are worse off, whereas those who are less averse to inequality may be inclined to attribute it to the effects of individual choice. These issues are difficult to adjudicate, however, without first achieving some clarity about the empirical question of how economic inequality does in fact come about and different ways in which choice and preference can affect it.

It is widely understood that even if one were to establish perfect equality, according to any designated conception, things would not remain equal for very long. ³ There is, however, a persistent failure to acknowledge the diverse forces that can be expected to drive the allocation away from equality. Most obviously, if one were to establish perfect economic equality, but do nothing about inequality in various other dimensions, such as social status or race relations, the latter would soon disrupt the former, since there are many ways in which individuals are able to parlay noneconomic advantages into economic ones. ⁴ Yet apart from these mechanisms that transmit inequality, there are also a variety of sui generis forces at work in the economic realm that would, over time, generate inequality. It is the latter that I would like to discuss in this section. As a further simplification, I will focus only on systemic sources of economic inequality that affect the pattern of distribution across the population as a whole.

A. Misfortune

The first source of inequality is simply good or bad luck, which is so obvious that it may seem barely worth mentioning. Where I come from, in Western Canada, the purest example of this is a summer hailstorm, a rare and essentially unpredictable event that may destroy one farmer’s entire crop, while leaving the neighbors unaffected. Events such as these have, throughout human history, left their victims wondering which god they have offended or how. (Many insurance contracts still make reference to “acts of god.”) This is what we call misfortune and it is an important source of economic inequality.

It has sometimes been thought, not least by egalitarians, that a useful distinction can be drawn between the inequality that results from acts of god of this sort and that which results from choices made by mortals. ⁵ This turns out to be rather difficult to do, since the gods seldom act entirely alone when it comes to the production of misfortune. Even the farmer, in the example above, might have chosen to locate the farm elsewhere or to plant a crop less sensitive to destruction by hail. ⁶ Even within the domain of individual choice, luck can play a significant role in generating inequality. One might compare the person who bets everything on a single, ill-considered gamble, but then happens to win against the person who makes a series of small, conservative choices, but against all odds, happens to lose them all. ⁷ More generally, two individuals who make a set of mathematically identical choices are more likely than not to get outcomes of different value. (Consider that when a fair coin is flipped ten times, the probability of getting exactly five heads and five tails is less than 25 percent.) Thus, there is likely to be considerable dispersal of outcome in human affairs.

Not all of this bad luck, it should be noted, will be neutralized by the “large numbers” effect. It is true that a person who is unlucky in some areas or on some days, will tend to be lucky in others. This is why the term schlimazel is a mildly humorous designation (especially with the suggestion that it parallels the schlemiel) since in principle there should be no such thing. ⁸ And yet, random sequences are a great deal more “streaky” than untutored intuition leads us to believe, which is what underlies our impression that some people are cursed to suffer a great deal more bad luck than others.

B. Unfairness

While bad luck is essentially exogenous to the prevailing set of social arrangements, the way that the economic institutions of society are organized can also, through their normal operations, be expected to give rise to inequality. The extent to which the market mechanism—or a system of decentralized production and allocation coordinated through the price mechanism—produces inequality has been the subject of considerable debate over the past two centuries. Many have claimed, somewhat counterintuitively, that although the outcomes produced by the market are highly unequal, the market is not actually the source of any of these inequalities. On the contrary, the market is essentially transparent to inequality; it merely takes some background allocation of resources, however unequal, and transforms it into a more efficient outcome that preserves the same inequalities. ⁹ Theoretical support for this claim is based on the so-called Second Fundamental Theorem (SFT) of welfare economics, which shows that for any desired allocation of economic outputs, it is possible to design a perfectly competitive market that, based on some initial allocation of resources, generates precisely that outcome as an equilibrium. This suggests, in turn, that in order to achieve any particular outcome, including one that satisfies equality according to some specified conception, there is no need to interfere with the operations of the market; one need only rearrange the initial allocation of assets in such a way as to produce that outcome. ¹⁰ This means that, to the extent that the outcomes of the market are unequal, it is not due to the operations of the market, but rather to the allocation of the initial inputs.

I describe this claim as counterintuitive because, when contemplating the vast fortunes amassed by a very small number of individuals, it seems rather unlikely that these are entirely due to a superabundance of natural talent or some other aspect of their initial endowment. ¹¹ Many have suspected that there is some mechanism internal to the operations of market institutions that biases outcomes in the direction of greater inequality. Karl Marx was, of course, the most prominent and early proponent of such a claim, arguing that the misery of the working classes, along with the relative affluence of the bourgeoisie, could be explained by the way that the labor contract permits the appropriation of surplus value by the employer. ¹² On this view, capitalism entrenches a specific form of exploitation that is different from that which is found under other institutional arrangements.

This analysis became less attractive to theorists over time, in part because of the declining fortunes of the labor theory of value, but also due to the realization that the concept of exploitation does not offer a robust account of inequality. Indeed, as John Roemer has shown, when the notion of exploitation is carefully specified, it is possible to construct a scenario in which the poor exploit the rich. ¹³ Thus the central conclusion of his important article “Should Marxists Be Interested in Exploitation?” is that, no, they should not. ¹⁴ His suggestion is that those who want to analyze ways in which the market produces inequality should focus directly on that, keeping equality as the central normative concept and not get sidetracked by Marxian analysis. ¹⁵

Marx was committed to discovering the “secret” of capital, but if one sets aside that ambition, in order to observe some of the more superficial qualities of markets, it is not difficult to find two features that produce inequality. Markets institutionalize a system of cooperation that employs both a division of labor and the realization of scale economies to produce enormous benefits. The division of these benefits—what David Gauthier refers to as the “cooperative surplus” ¹⁶ of the interaction—is determined by the set of factor payments that are in turn determined by the price mechanism. This is why, under previous modes of production, questions of “just price” (or more specific questions, such as the price of bread or the wage level) were considered essential issues of social justice. The price essentially dictated how the burdens and benefits of the cooperative scheme were to be divided up between the parties involved. In a market economy, however, the competitive determination of prices is intended to maximize efficiency in a way that is essentially orthogonal to questions of distributive justice or of equality and inequality. This has required abandoning the preoccupation with justice in pricing. Thus, the first source of unfairness in distribution is due to the fact that economic exchange divides up the cooperative surplus that it generates in accordance with principles that are arbitrary from the standpoint of distributive justice.

This point tends to be obscured by ways of thinking about the market inspired by the SFT because the latter provides only a static analysis. Even more egregiously, however, the SFT provides a model of an economy in equilibrium. An actual market economy, however, is never in equilibrium. Unlike a planned economy, in which the set of calculated prices is supposed to be in equilibrium (at least in principle), the market relies upon the dynamic adjustment of individual agents to out-of-equilibrium prices to move prices in the appropriate direction over time. This serves as a second source of unfairness because it means that economic actors will be able to appropriate rents (that is, payments to providers that exceed what would be necessary to keep some factor in its current employment). These are often enormous, but there is considerable reluctance to control them, precisely because the incentives they provide are what generate the competitive dynamic that, in turn, moves prices toward market-clearing levels.

C. Improvidence

While people are often short-sighted in their decision-making, this is not necessarily of interest from the standpoint of justice since it is difficult to specify a priori what the appropriate attitude toward balancing present and future satisfaction should be. While there is something to be said for temperance, there is also a great deal to be said for living in the moment—or at least not unduly postponing gratification. One might therefore be tempted to treat this entire domain as one of reasonable disagreement, save for the fact that there is also a “warp” in the way that we evaluate future satisfaction that leads us often to act in ways that undermine our own plans. ¹⁷ The most familiar instances of this occur when we experience temptation, which causes us to deviate from some prior resolution only to subsequently regret that choice. Because vulnerability to such reversals of preference is appreciably worse with some people than with others, it winds up becoming an important source of inequality.

A simple illustration of this warp is to consider the choice between receiving a $100 cashier’s check that can be deposited right away or a check for $200 that is post-dated by three years. Most people would take the $100, but given a choice between a check for $100 that can be cashed in six years or a check for $200 that can be cashed in nine years, they would opt for the $200. ¹⁸ These two choices are not only inconsistent, but dynamically unstable, since the latter can be expected to generate a reversal of preference in six years, when the opportunity to cash the first check right away presents itself. There is some debate about the psychological underpinnings of this phenomenon—for example, whether it is a consequence of a pure time preference, ¹⁹ the occurrence of a “hot” psychological state, ²⁰ or a glitch in our system of goal construal ²¹—but it can be modeled straightforwardly as a “hyperbolic” discounting function. Rather than representing the agent’s temporal attitude as a consistent discount rate, applied in exponential fashion (like an interest rate), it can instead by treated as a function that imposes a significant discount rate on near-term delay, combined with a flattening of the curve in the long term. Such a model accounts for the way that, when a choice is far removed from us in time, we resolve to choose the “larger later” good over the “smaller sooner” one, but then as time passes these preferences reverse, leading us to select the “smaller sooner” good. This corresponds to the ordinary notion of succumbing to temptation. The model offers a similarly parsimonious account of procrastination, which is the inverse phenomenon of choosing the “larger later” bad over a “smaller sooner” one. ²²

The hyperbolic-discounting model also provides a morally neutral way of representing the phenomenon traditionally referred to as “improvidence,” in the sense of a general failure to plan appropriately for the future and to be resolute in the execution of that plan. It is not difficult to study discount rates across a population and to see that the magnitude of the warp differs from one person to another, which in turn makes some more likely to exhibit improvidence than others. ²³ This can have a direct effect on achieved welfare levels, especially if one agrees with contemporary theorists who believe that this warp provides the psychological underpinnings of addiction. ²⁴ (What was striking about my friend, for example, was how the pattern of improvident choices that he made was so consistent across domains.)

Until the development of effective birth control, the major self-control issue involved the number of children that individuals had and the age at which they had them. As life expectancy increased over the course of the twentieth century (and birth rates declined in wealthy countries), other issues, such as the rate of savings, began to eclipse this. ²⁵ Wealth is much more unequally distributed than income, in part because lower income is associated with not just lower savings, but a lower rate of savings. This amplifies inequality because savings represent more than just deferred consumption. Because of the productiveness of capital, wealth is a source of income, and so those who save acquire the ability to save more over time, a trend that is then exacerbated through inheritance. Another major issue is educational attainment, which requires the ability to defer satisfaction for many years. Improvidence leads individuals to refrain from making appropriate investments in their own human capital, which in turn limits their income in future years.

III. Reductive Analyses

These three sources of economic inequality are distinct from one another and, superficially at least, it is not difficult to see how they conspire to produce a great deal of human misery. They are also interconnected in complex ways. For example, living under conditions of material scarcity tends to exacerbate hyperbolic discounting, so while improvidence is an obvious source of poverty, poverty from other sources may also produce increased improvidence, which will in turn deepen poverty, and so on. ²⁶ There is a complex skein of causal relations among the various sources of inequality. Many theorists, however, rather than merely tracing out the connections between these different sources of inequality, have tried to reduce them all to just one, which is then taken to be the root cause of all inequality. The motivation for this is usually not difficult to discern; it is typically driven by normative commitments. Marx was only interested in forms of inequality generated by the market mechanism; he had little concern for either bad luck or improvidence. More recently, however, egalitarian philosophers have been struck by a desire to reduce all forms of inequality to bad luck. ²⁷ Those who are more tolerant of inequality, by contrast, have for a long time sought to reduce it all to improvidence. ²⁸ The underlying normative intuitions are in both cases apparent, since those who experience a run of bad luck are clearly more sympathetic victims than those who suffer from their own poor choices. Unfortunately, these efforts at reduction impede our capacity to confront the sources of inequality in their full complexity, which in turn makes it difficult to develop appropriate remedies or even to see the attractions of certain remedies that are available.

My own view is that these three sources of inequality are not only irreducible to one another, but they are also best addressed through different policy interventions. As a result, the effort to tie them all back to one cause has a perverse consequence. While doing so simplifies normative argumentation, it does so only by positing relations of causal dependence that are largely spurious. This in turn encourages policy interventions that are destined to fail because they are based on wishful thinking about the etiology of the inequalities that they seek to redress.

The problem with reductionist accounts can be seen most clearly by considering the apologetic version, which seeks to justify the plight of the poor by claiming that it is all a consequence of improvidence. This conviction is what underlies a great deal of the rhetoric among conservatives about the importance of “personal responsibility.” For example, James Q. Wilson was fond of citing the statistic that one need only do three things in the United States to radically diminish the risk of poverty: “finish high school, marry before having a child, and marry after the age of 20. Only 8 percent of the families who do this are poor; 79 percent of those who fail to do this are poor.” ²⁹ (One could easily add “do not become addicted to drugs” to the list in order to make the percentages even more favorable to the argument.) On this view, criminality, teenage pregnancy, single parenting, dropping out of school, and substance abuse are all seen as impulse-control problems, which impair the ability of individuals to participate in the workforce or limit them to low-skilled, poorly remunerated occupations. ³⁰

The upshot of this analysis is twofold. First, it suggests that those who are the losers in the general distribution of goods in society bear primary moral responsibility for that outcome. The second element of the analysis then consists in the claim that the best way to discourage this sort of improvidence is to allow individuals to bear its consequences. Improvidence serves as a remedy unto itself; no social or state intervention is required. Indeed, any attempt to reduce the inequality that results should be discouraged because of the moral hazard effects it would provoke. That is, by failing to make individuals bear the full burden of their own improvidence, the state is encouraging them to act this way. ³¹ The optimistic version of this argument asserts that improvidence is caused by the indiscriminate charity of others. Charity, on this view, creates the demand for its own services. But since not all misfortune can fully be compensated, everyone would in fact be better off if certain tender souls stopped helping others because then everyone would start taking responsibility for their own lives. The more hard-hearted view recognizes that there would probably still be a great deal of improvident conduct, even if individuals were forced to bear the full consequences of their decisions, but in this case the suffering is not bad because it is deserved.

This reductive argument allows one to dismiss any apparent unfairness of the market as merely an indirect consequence of improvidence, such as failure to develop one’s human capital, failure to accumulate savings, etc. As for bad luck, much of it can also be traced back to improvidence, such as failure to exercise appropriate caution, failure to save for a rainy day, etc. Finally, even in cases in which there is no direct connection to improvident conduct, the moral-hazard argument can be appealed to as an argument against doing anything to improve outcomes for the losers, on the grounds that doing so will deter them or others from taking proper precautions in the future. For example, even if no one could have anticipated the hailstorm, compensating farmers who have suffered losses will just encourage them to keep planting crops that are vulnerable to hail. So even when bad luck is not a direct consequence of improvidence, still nothing should be done about it because that would only bring more bad luck. ³²

Egalitarians have responded to this argument in various ways, but by far the most common has involved adopting a position that directly mirrors it. Instead of arguing that the primary factor driving economic inequality is improvidence, these egalitarians have argued that the primary factor driving it is bad luck. In its “luck-egalitarian” formulation, this involves accepting the basic premise that inequality is justifiable in cases in which the losers (and the winners) can be held responsible for the difference in outcome, but then claiming that when it is not acceptable to hold individuals responsible, differences in outcome ought to be eliminated. ³³ Some critics have expressed puzzlement over how this normative framework could avoid generating conservative policy implications. ³⁴ And yet the widespread perception among proponents of luck egalitarianism is that their view implies some type of recognizably left-wing politics. ³⁵ This makes very little sense, unless one takes luck egalitarians to be committed also to a set of empirical premises that leads them to believe that the poor are seldom responsible for the outcomes they receive. ³⁶ In other words, they take their view to be egalitarian, as opposed to anti-egalitarian, because they think that as a matter of fact most inequality can be traced back to some form of bad luck.

With respect to the inequality generated by the market, the primary luck-egalitarian strategy involves accepting the basic SFT framework, which suggests that the unequal outcomes produced by the market are a consequence of an unequal allocation of initial endowments. ³⁷ On this view, inequalities that result from wage differentials are seen as a consequence of unequal endowment of “natural talent.” ³⁸ These initial endowments are then criticized on the grounds that they are the product of luck, such that individuals cannot be held responsible for them. Some people get born lucky in the genetic lottery, which gives them various natural talents that they are then able to parlay into economic advantages. This analysis allows egalitarians to characterize low wages as the result of misfortune.

With respect to the effects of improvidence, the strategy is similar. It takes the form of a “deresponsibilization” argument, aimed at showing that certain self-defeating choices made by individuals are not choices at all, but rather are determined by background social conditions. ³⁹ The standard argument involves taking some prima facie improvident action, such as bearing a child before graduating from high school, then showing how it is the product of larger social forces or that it is subject to various “social determinants.” One might point to empirical studies showing a correlation between family income and teenage pregnancy or to evidence that children raised in single-parent households are more likely to become pregnant at a younger age. It is then suggested that individuals should not be expected to bear the cost of these choices because they have no control over these background conditions. Again, this argument comes in more and less optimistic variants. The optimistic version holds that, absent these background inequalities, no one would make such improvident choices, and so egalitarian redistribution, far from generating moral hazard, will solve the problem of improvidence. The less optimistic version expects that improvident choice will persist even under conditions of greater equality, but that individuals should nevertheless not be expected to shoulder the cost and that the moral hazard effect of redistribution should be tolerated.

I believe that one can recognize in these two grand reductions the core of an extremely familiar debate between the Left and the Right, which has been dragging on unproductively for decades. It is not difficult to see why so many people are drawn to one or another of these views, depending on their background sympathies. Each provides an analysis that radically simplifies the normative challenge presented by the phenomenon of economic inequality. In so doing, however, each adopts an essentially procrustean approach toward the empirical basis of this inequality, which in turn makes it more difficult to develop policies aimed at countering it.

In the case of egalitarians, the focus on bad luck has led to excessive emphasis on the distribution and redistribution of endowments as the key to redressing inequality. One can find the most extreme manifestation of this among those who recommend lump-sum redistribution of wealth as an all-purpose solution to the problems of inequality. In The Stakeholder Society, for instance, Bruce Ackerman and Anne Alstott argue that large monetary payments should be made to every American upon reaching the age of maturity (at the time of writing they suggested $80,000). ⁴⁰ Although willing to accept the moderately paternalistic constraint that the funds be disbursed in four annual installments of $20,000, their primary response to the concern over improvident use of the money is to recommend the introduction of education programs for high school students, such as “How to Manage Your Stake” classes, which they expect to become as popular as driver’s education. ⁴¹ Unfortunately, the same psychological trait that makes certain people unlikely to make wise use of large lump-sum cash payments also has a tendency to make them inattentive in, or absent from, formal educational settings. This is symptomatic of a general feature of the view, which is that it fails to take seriously the problem of improvidence, preferring instead to engage in wishful thinking about the potential for material redistribution to abolish its underlying causes.

The same focus on endowments has led many egalitarians—most prominently G. A. Cohen—to worry a great deal about differential endowments of talent, as though these were a major source of inequality in labor-market returns. ⁴² As a result, this analysis ignores several more important forces that drive inequality in the distribution of reward in a market economy. Natural talent does not translate directly into higher wages; compensation depends on how many other people possess the same talent or what close substitutes are available. Workers in highly competitive markets, who can easily be replaced, are going to earn lower wages than workers in less competitive markets, regardless of how objectively “talented” they are. This is why international trade can have major effects on wages in particular occupations: it increases the competitiveness of labor in particular sectors by expanding the supply of available workers. Similarly, skill-biased technological change will affect wages by increasing and decreasing demand for certain categories of labor. These two forces, which have been central to the stagnation of blue-collar wages in the past three decades in the United States, are both instances of arbitrariness in the division of the cooperative surplus that arises endogenously in market institutions; neither has anything to do with background endowments. As a result, they have largely been ignored by luck-egalitarian philosophers.

IV. Promoting Equality

With the decline of the socialist project in its more assertive variants—which demanded fundamental changes in the ownership of productive assets—egalitarians have now by and large accepted the market as the central institutional mechanism for organizing the division of labor in society. The result has been an enormous emphasis on redistribution by the state as the primary mechanism for promoting greater economic equality. ⁴³ This has encouraged a view according to which a major function of the welfare state is to take money away from the rich (primarily through a taxation scheme that is progressive with respect to income) and give it to the poor (through entitlement programs that offer either cash payments or in-kind provision of goods). This is often referred to as the “Robin Hood” model of the welfare state. The temptation to endorse this model is, of course, enhanced by the reductive-egalitarian view, which regards income disparities as a downstream consequence of a morally arbitrary distribution of endowments. If this were the problem, then it would seem only natural that the solution should involve redistributing these endowments either directly or, in cases in which they are non-fungible, indirectly (for example, through a tax on earned income).

The problem with this approach is twofold. First, it fails to take into consideration the extent to which different sources of inequality can resist or undermine a purely redistributive remedy. Thomas Piketty, for instance, puts enormous emphasis on wealth inequality, but says surprisingly little over the course of several long books about the fact that the distribution of wealth is mediated by the savings decisions of individuals and that, when it comes to these savings decisions, improvidence is a significant challenge to the egalitarian project. ⁴⁴ It is certainly not an accident that, for most of the population, what wealth they do possess is sequestered in the form of owner-occupied housing, locked-in retirement savings, or pension funds. These vehicles all serve as commitment mechanisms, imposing significant financial cost or inconvenience on those wanting to use their savings to finance current consumption. The effectiveness of these institutional arrangements in this regard has, however, been eroded in recent years through financial innovation, which makes it much easier to drain the value of or to borrow against these assets. ⁴⁵ This is where the crucial policy issues now lie, since there are very limited returns to engaging in redistribution of wealth in a society in which a significant fraction of the population is willing to carry credit card debt at interest rates often exceeding 20 percent.

Second, the focus on redistribution generally overstates the amount of pure redistribution that is undertaken by the state, which in turn leads to a relative neglect of other strategies that are employed by a successful welfare state to combat inequality. On the “input” side, the progressive income tax largely conforms to the Robin Hood picture, although even here it should be noted that most European welfare states rely on a very high value-added tax (VAT), with rates between 19 percent and 25 percent, which amounts to a proportional tax on income with a complete exemption for savings. Thus, the tax regimes in many of the most egalitarian societies are not nearly as progressive as the selective focus on income taxes might lead one to conclude. ⁴⁶ It is on the “output” side, however, with respect to the provision of government programs and services, that the redistributive model encounters the most significant difficulties. If one takes a single-year snapshot of state activity, it can look as though a great deal of redistribution is taking place, but if one looks at expenditures over the lifetime of a citizen, much of this apparent redistribution turns out to be life-cycle smoothing. For example, John Hills has calculated that approximately three-quarters of state benefits received by individuals in the United Kingdom are “self-financed” over their lifetime, and thus not redistributed between individuals at all. ⁴⁷

When thinking about how to redress inequality, it is important therefore to recognize the significant limitations of pure redistribution and to examine more carefully the way that successful welfare states have managed to achieve greater reductions in inequality. In this section, I will focus on a set of institutional arrangements that are aimed at countering the specific sources of inequality identified above. These have all featured centrally in the development of the welfare state (although they are often obscured by the Robin Hood model): social insurance, active labor-market policies, and the “scaffolding” of individual choice. ⁴⁸ All of these features of the welfare state are, it should be noted, responses to broad-based economic inequality (of the sort measured by a Gini coefficient ⁴⁹); the question of what to do about the increasing fraction of national income going to the top 1 percent of earners is a separate issue.

A. Social insurance

Although the welfare state does a great deal to diminish economic inequality, the nature of the transfers that it undertakes are widely misunderstood. It is common to compare the distribution of market income to the distribution of income after “taxes and transfers” in order to gauge the amount of redistribution going on. This is what generates the impression of a Robin Hood logic. In reality, very little of what the welfare state does on the expenditure side follows this pattern. Most of these transfers, especially those belonging to what is often referred to as the “social safety net,” are better understood as a set of insurance programs administered by the state and financed through taxation. This is particularly obvious with disability and unemployment insurance as well as cases in which health care is financed through a public insurance system, but it is equally true of pensions. Although state pensions are financed through fixed contributions, they typically make payments from the age of retirement until death, which makes them equivalent to life annuities, a type of insurance product designed to protect individuals from the risk of outliving their savings. ⁵⁰

The central characteristic of insurance systems is that they offer individuals indemnification against specific losses: health insurance transfers money from the healthy to the sick, disability insurance transfers money from those who work to those who are forced to stop working, pensions transfer money from those who die young to those who live a long time, and so on. These transfers promote equality in the local sense that they limit the inequality that can develop with respect to the particular loss that is indemnified. This may or may not promote equality in the global sense since the transfers need not be progressive with respect to income or wealth. For example, the rich tend to live longer than the poor, and so pensions have a modestly anti-egalitarian bias with respect to wealth. They are egalitarian only to the extent that they eliminate the effects of a particular form of bad luck. Of course, when the state provides insurance, it is able to charge whatever premiums it likes (either directly, in the form of payroll deductions, or indirectly, by financing the program out of general taxation). It is therefore possible to finance these programs in a way that will be progressive with respect to income (just as one can finance provision of public roads and parks in a way that is progressive with respect to income). However, the transfers that are undertaken on the public-facing side of the program do not follow an egalitarian-redistributive logic; instead, they follow what Francois Éwald refers to as “une logique assurantielle” (an insurance logic). ⁵¹

Marx’s reading of early capitalism has become so influential that we are inclined to view the living conditions of the working class in nineteenth-century Britain as a consequence of injustice in the distribution of the economic product. And yet, if one looks at the emergence of friendly and fraternal societies in the late-eighteenth century and their growth over the course of the nineteenth, a different interpretation suggests itself. On this view, industrialization and proletarianization had the effect of dissolving many aspects of traditional village and parish life that had protected individuals from risk. Decreased demand for labor under feudalism, for example, was typically diffused by agrarian institutions over many individuals, producing generalized underemployment. The result was a great deal of poverty in the countryside, but very little unemployment. The growth of factory labor, by contrast, eliminated these arrangements, while initially offering no risk-pooling arrangements in their place. A great deal of the misery that we associate with the factory slum—sickness, injury, disability, unemployment, and the lack of support for orphans—was the product of a sudden individualization of exposure to risk, not distributive injustice per se. While it is trivially true that with enough money no one needs insurance, there is no feasible redistributive scheme that obviates the need for insurance in the core categories of health, life, and retirement. Thus, the solution to the miserable conditions of the nineteenth century arose from the development of contractual insurance, initially in the private and nonprofit sectors. Only later did these functions become public and provide the foundations for the growth of the welfare state. ⁵²

If one examines the major successes of social-democratic parties in the twentieth century, those are based largely in the popularity of social insurance systems. Unlike purely redistributive programs, which have an essentially zero-sum structure, social insurance produces ex ante welfare gains for all participants. The result is that these schemes do not automatically give rise to a constituency committed to their abolition. This is why the Conservative Party in the United Kingdom finds itself championing the National Health Service and the Republican Party in the United States now presents itself as committed to protecting Social Security. Government turns out to be an efficient provider of insurance because it benefits from the largest risk pool, suffers from no adverse selection, and has low overhead costs due to its administration of the tax system. As a result, even when it engages in cross-subsidization between individuals in the plan, the state is nevertheless able to provide basic insurance on terms that offer some benefit to almost every citizen. This has the effect of promoting equality by minimizing the consequences of certain forms of bad luck, while generating benefits that outweigh the deadweight losses occasioned by the financing of these programs through taxation.

B. Labor-market policies

St. Augustine believed that the “earthly city” was incapable of achieving true justice, but that it could produce an outcome that bore an external resemblance to it, namely, peace. Similarly, when it comes to dealing with the market, the state has largely abandoned the goal of producing outcomes that correspond to anyone’s intuitive conception of justice, such as ensuring that the wicked never prosper and the honest receive their just reward. Most of its contribution consists in preventing overt conflict from erupting. This is apparent in the case of labor and employment arbitration, which takes an entire category of decision-making out of the hands of the courts, with tribunal decisions typically protected from judicial review by privative clauses, in order to apply a set of more pragmatic principles to the resolution of these disputes. Arbitrators do not work from first principles; instead, they rely primarily on precedent and interfirm comparisons, tacitly acknowledging that there is no way to determine what constitutes just compensation in the context of a market economy.

To the extent that the welfare state seeks to diminish wage inequality, it does so largely by attempting to influence the competitiveness of different segments of the labor market. The major reason that wages are not more closely tied to intrinsic features of the work performed (such as its onerousness or its value to society), is that they are affected by how many people are willing and able to do that particular type of work. Wages in more competitive sectors are lower than wages in less competitive sectors. In cases where labor-market segmentation is based on irrational factors, such as gender stereotypes, the state can work to diminish wage differences by breaking down the barriers that prevent movement of workers from one class of employment to another. In other cases, where segmentation is based on possession of certain skills, the state can promote movement of workers through training programs and other active labor-market policies. Another often-overlooked set of policies promotes geographical mobility either through portability of social insurance benefits or direct tax incentives. Both of the latter two policies are particularly important at transitioning workers out of industries that are declining due to trade or technological disruption.

With respect to the average wage rate, the state is able to promote full-employment policies and manage a certain measure of inflation. Modern welfare states are also all involved, to varying degrees, in implementing industrial strategy and trade policy, funding research and development, and subsidizing higher education—all aimed at promoting high-value-added work. Finally, states can promote more egalitarian wage policies through support for unionization, most obviously by enforcing “closed-shop” arrangements and restrictions on the hiring of replacement workers. Unions tend to have a leveling effect with respect to the wages of members, although they often increase interindustry wage differentials, and so their impact on overall inequality can be difficult to assess. Generally speaking, they are likely to have the most positive impact when rates of unionization are very high, as is the case in several northern European countries.

C. Scaffolded choice

Perhaps the most delicate area of welfare state activity involves policy areas in which there is a need to discourage or prevent individuals from making improvident choices. Despite the familiar criticism of the “nanny state,” modern welfare states are in fact hesitant to intervene in many areas out of a desire to avoid violating the liberal prohibition on state paternalism. The perception of state paternalism arises often from a misunderstanding of social-insurance schemes. ⁵³ The fact that they offer indemnity against specific risks is sometimes misinterpreted as a paternalistic constraint on income redistribution. ⁵⁴ Similarly, the fact that citizens are not allowed to opt out is seen as paternalistic rather than being a response to adverse selection. Thus, it is important to be careful when it comes to distinguishing genuine paternalism from arrangements that exhibit what might be referred to as “byproduct paternalism,” that is, where the fact that individuals are forced to act in their own interest is accidental rather than essential to the justification of the arrangement. Generally speaking, there has been a decline in the willingness of Western states to intervene coercively in many areas of individual choice: truant officers no longer drag students physically back to school, psychiatric institutions largely refrain from involuntary commitment of patients, addicts in recovery programs remain free to leave, and so on. As a result, certain highly visible forms of inequality, such as homelessness, remain persistent because in many jurisdictions they are a consequence not just of poor housing policy, but also drug addiction and mental health problems, where the state refrains from heavy-handed interventions.

At the same time, a certain number of arrangements are clearly paternalistic. The payment of social assistance, unemployment benefits, and child support as a monthly check instead of as a yearly lump-sum payment integrated into the income-tax system, has a transparently paternalistic rationale. Many other policies are semi-paternalistic. A great deal of life-cycle-smoothing transfer payments, for example, would be unnecessary if individuals could be counted on to exercise greater self-control in their savings decisions. Both paid parental leave and subsidized daycare would be unnecessary if people set aside some money before having a child or could borrow to cover these expenses. As it is, parents wind up paying for these programs in the form of higher taxes over the course of their working lives, and so these policies are only redistributive between those who have children and those who do not. Since there is no good reason to penalize those who refrain from having children, the redistributive aspect of the policy is unmotivated by any plausible conception of justice, and so the justification for the policy must lie primarily in its paternalistic quality. ⁵⁵ Similarly, universal primary education, tax benefits for retirement savings, policies that promote home ownership, grants or subsidized loans for higher education, and consumer protection laws, are all policies that have a significant paternalistic dimension.

There is, however, an important difference between these interventions and the classic forms of hard paternalism criticized by liberal theorists. John Stuart Mill’s objection to state paternalism was that it most often involved the imposition of other people’s values on those who do not share them, under the guise of knowing where their “true interest” lay. ⁵⁶ In the case of dynamically unstable preferences, however, such as one finds with hyperbolic discounting, it is much less tendentious to say that state power is being used to promote the individual’s own interests because the individual can be expected to endorse the policy both before and after the time at which its coercive effects are felt. ⁵⁷ Recent policy thinking has also focused on modes of intervention that can more easily discharge the justificatory burden, such as asymmetric paternalism (where the cost of avoiding the intervention is low relative to the benefits that flow to the genuinely improvident) or nudge paternalism (where the intervention does not change the rational structure of the individual’s choice, and so is only costly for those who act irrationally). ⁵⁸ These policies can be thought of as “scaffolding” individual choice. Rather than imposing an objective on individuals, they are generally aligned with the individual’s own goals. What they do instead is buttress the individual’s will, often just by providing a temporal restructuring of incentives, for example, bringing certain costs or benefits forward in time, so that they will not be as sharply discounted.

D. The upshot

If one thinks of taxation of income followed by cash transfers to individuals as “redistribution,” then all of the interventions described above can be described as “predistribution.” Rather than allowing inequalities to arise and then attempting to undo them through the tax system, these policies involve preventing inequalities in pretax income from arising in the first place. ⁵⁹ This has been the subject of increased attention, in part due to the realization that European welfare states such as France are more egalitarian than the United States, not so much because they do more redistribution, but because they do a great deal more to prevent inequality. The two approaches are, of course, not incompatible with one another, but it is important to see why predistributive policies have come to play the large role that they have. Although economic analysis has tended to focus on labor-market policies in this regard, it is not difficult to see how social insurance and choice scaffolding achieve similar outcomes with respect to other sources of inequality. ⁶⁰

Two major points can be made about the politics of predistribution: the first unequivocally positive and the second cautionary. A major problem with purely redistributive policies is that they have a zero-sum logic, which means that they tend to become politically acrimonious. ⁶¹ Attempting to persuade people that they have no entitlement to their income and so should greet its appropriation by the state with equanimity—as Liam Murphy and Thomas Nagel advise ⁶²—has so far proven more popular among philosophers than with politicians or the broader public. An important feature of the predistributive policies described above is that they all have a positive-sum character. State intervention works by producing economic or welfare benefits, which are then allocated in a way that favors those who suffer some disadvantage or loss. Socialized medicine resolves a major market failure in the health-insurance sector, state pensions provide life annuities for millions of workers who are unable to purchase them privately, job-training policies increase overall worker productivity, parental subsidies produce nontrivial benefits for children, and so on. Apart from making these programs more philosophically defensible, this positive-sum character also has the effect of making them more politically resilient. Because they do not automatically generate a constituency with an economic interest in seeing them abolished, they are more likely to survive the inevitable alternation of left-wing and right-wing political parties in power.

The major problem with these more localized, predistributional programs is that they are not necessarily compatible with one another; for example, an intervention that reduces inequality from one source may well exacerbate it from some other. One has the sense that a significant attraction of the Robin Hood model of the welfare state is that it relieves the theorist of the need to worry about a great deal of institutional detail, a simplification that is further enhanced by ignoring everything but the income tax and setting aside any worries about tax incidence. The commitment to equality then translates into a single, consistent, straightforward intervention. The attempt to reduce pretax inequality, by contrast, can easily generate policy dilemmas, which a mere commitment to equality is powerless to resolve. For example, unemployment insurance can interfere with active labor-market policies aimed at shifting workers out of particular sectors. Since it can be difficult to tell the difference between a cyclical downturn and a permanent decline, it may be practically impossible to prevent these policies from working at cross-purposes. Similarly, one consideration that speaks against raising minimum wages is that, since a large fraction of minimum-wage workers are teenagers, a high minimum wage can encourage students to drop out of school. The wage provides, in effect, a misleading impression of how much low-skilled labor is valued in the economy. The desire to combat improvidence in educational attainment can therefore conflict with the goal of improving living conditions among low-skilled or unskilled workers.

This analysis allows us to see why certain programs, such as socialized medicine and public pensions, form the backbone of a successful welfare state. The ideal public program is one that generates a significant cooperative surplus by resolving a market failure, but where the benefits can be allocated in a way that diminishes inequality from some particular source without having perverse effects that increase inequality from other sources. This is why cross-subsidization across participants in a social-insurance system is so much more successful than redistribution between taxpayers as a mechanism to promote greater economic equality. None of this is incompatible with the basic role that political philosophers have seen for the welfare state in promoting greater equality. It merely shows that the task is more complicated than it might at first seem: first, because of the potential tension between different policies aimed at redressing inequality from different sources, and second, because of the internal connection that it reveals between the efficiency-promoting and the equality-promoting activities of the state.

V. Conclusion

In his celebrated article “On the Currency of Egalitarian Justice,” Cohen makes the offhand remark that some people achieve low levels of welfare “because they are negligent or feckless in a morally culpable way: they buy their food at Fortnum’s because they cannot be bothered to walk up to the Berwick Street market.” ⁶³ This passage has always reminded me of my friend’s shopping habits, along with the fact that “feckless” consumption choices are not confined to the upper classes. It also impressed upon me the extent to which Cohen, from his perch at All Souls College, failed to appreciate the magnitude and extent of human suffering that can be attributed to such faults of character. More generally, it shows how little the major prescriptions of the egalitarian Left or the anti-egalitarian Right would have helped my friend during his life, since there was a clear sense in which the major problems that he suffered were a consequence of his own choices. (Even the epic run of bad luck that landed him in jail began with a motorcycle accident, an event that never happens to those who are more cautious in their choice of transportation modality.) Many people would have been happy to abandon him to suffer the consequences of his decisions. Other more sympathetic souls might have been willing to overlook the self-inflicted character of his injuries, yet their efforts to help him would have taken the form of a cash transfer, which would have done very little to improve his life.

It is important to observe, however, that my friend did derive significant benefits from several major welfare-state programs as currently constituted. He benefited most obviously from the Canadian public health-care system, not just for his chronic medical problems, but also for psychiatric counselling and periodic addiction treatment. He also benefited from unemployment insurance, job training, subsidized housing, and as a parent, Quebec’s public childcare system. Although his life did not go that well, it could have gone a great deal worse had it not been for these affordances. From a policy perspective, these programs also had a significant effect on reducing the contribution that he made to economic inequality. My primary objective in this essay has been to urge philosophers who are sympathetic to this objective to develop a normative analysis that expresses these goals, but in a way that sticks closer to the actual mechanisms through which existing state programs achieve their effects. Recognizing that there are three different sources of economic inequality, each irreducible to the other, is an important step in this direction.