About this Issue

Here’s a tempting argument: Income inequality is rising, and its rise means something is wrong with our society. To fix that problem, we should adopt more redistributive policies that will correct for income inequality.

In this month’s issue of Cato Unbound, our very own Will Wilkinson challenges this tempting argument from start to finish. Income inequality is less important than we think, it has changed less than we think, and redistributing income probably won’t fix any of the underlying problems that might cause income inequality.

These questions sit at the intersection of sociology, economics, and ethics, and so we’ve invited an expert in each. Respectively, they are Lane Kenworthy of the University of Arizona, John V.C. Nye of George Mason University, and Elizabeth Anderson of the University of Michigan.


Lead Essay

Economic Inequality and the Mirage of Injustice

Prior to the election of Barack Obama, income inequality was a hot topic. Much of the credit for this should go to powerhouse New York Times columnist and economics Nobel Laureate Paul Krugman. His alarmed and pugnacious book about American income inequality, The Conscience of a Liberal, marked a high point of public interest in the subject, which has died down considerably since Obama took office and the recession took hold. Krugman argues that the widening income gap had chiefly political rather than structural causes and demands a political, largely redistributive response. Perhaps the presence of a Democratic president and Congressional majorities in Washington have invited the sense among Krugman-style egalitarian liberals that the matter is now in good hands and will be addressed in due time. But now, in a season of double-digit unemployment and rising poverty rates, it might pay to consider whether income inequality is really among our most urgent problems. Even governments have a limited attention span. More careful allocation of this and other scarce resources would be most welcome.

In this essay,  I argue for following three points.

  1. The level of real economic inequality is lower than popular treatments of the issue have led many of us to think.
  2. The level of economic inequality is an unreliable indicator of a society’s justice or injustice.
  3. Inequality distracts us from real injustices that are given too little attention.

In the following I draw from my paper, “Thinking Clearly about Economic Inequality, published by Cato in July, though there are some fresh references and most of it is new.

Getting the Story Straight

The first step in getting a story straight is to identify the main character. Most discussions of economic inequality focus on the trend of nominal income inequality. The familiar story about nominal income inequality is that since the 1970s income inequality has been rising steadily and is now at levels not seen since the Gilded Age. I think this story gets on the wrong track at the start.

As far as I can tell, when most people are worried about economic inequality, they’re usually worried about inequalities in real standards of living — in the real material conditions of life. Income plays a crucial role in this story, but it’s a supporting role. An individual’s or household’s standard of living is determined rather more directly by the level of consumption than by the level of income. Current income does finance current consumption, but so do credit and savings. If we can directly inspect measures of consumption, and we can, then we probably should. When we do, we find that the familiar story of income inequality is rather misleading and that trends in consumption inequality are considerably less dramatic. Different datasets and analytical methods produce somewhat different results, but most stories of consumption inequality are stories of stability or a relatively mild rise.

Daniel Slesnick says:

The widely reported U-turn in inequality in the United States is an artifact of inappropriate use of family income as a measure of welfare. When well-being is defined to be a function of per equivalent consumption, inequality either decreased over the sample period [the 1990s] or remained unchanged.

Dirk Krueger and Fabrizio Perri say:

[T]he increase in income inequality in the U.S. has been much more pronounced than the corresponding increase in consumption inequality. … Consumption inequality … has remained substantially stable.

And in the most impressively wide-ranging and technically sophisticated evaluation of inequality data I’ve seen to date, Jonathan Heatchote, Fabrizio Perri, and Giovanni Violente say:

First, consistent with basic economic theory, consumption inequality is substantially lower than income inequality. Second, the rise in consumption inequality is much smaller than the rise in disposable income inequality.

Even if we let income hog the spotlight in our story about inequality, a good deal of evidence suggests that income inequality has been overestimated. In a new paper weaving together several strands of recent evidence, Robert Gordon reports that improved use of income datasets “shows that there was no increase of inequality after 1993 in the bottom 99 percent of the population, and can be entirely explained by the behavior of income in the top 1 percent.”

What happened in the top one percent? The bulk of the evidence seems to point in the direction of changes in executive compensation. The debate over the details of these changes is too complicated to adequately address here, though I’ll touch on it below. For now let me just highlight the finding that there has been no increase in income inequality within the bottom 99 percent of the distribution for more than a decade.

A number of other recent studies indicate that measured income inequality has been overstated due to inadequacies in traditional methods for constructing price indices and estimating real income. For example, in the latest iteration of a much-discussed paper Christian Broda and John Romalis find that

the relative prices of low-quality products that are consumed disproportionately by low-income consumers have been falling over this period. This fact implies that measured against the prices of products that poorer consumers actually buy, their “real” incomes have been rising steadily. As a consequence, we find that around half of the increase in conventional inequality measures during 1994 – 2005 is the result  of using the same price index for non-durable goods across different income groups.

Many popular narratives about inequality are grounded on the alleged fact that wages and incomes at the middle and bottom of the distribution have been stagnant for decades. It appears that this, too, may be an artifact of insufficiently sophisticated methods for constructing the price indices used to calculate rates of inflation. Using an updated price index, Christian Broda, Ephraim Leibtag, and David Weinstein [pdf] find that

the real wages at the 10th percentile increased by 30 percent from 1979 to 2005. In other words, the real wages of low earners have not remained stagnant, as suggested by conventional measures, but actually have been rising on average by around 1 percent per year.

No doubt  there are sensible methodological objections to all these studies. But taken together they are impressive. Suppose we accept their upshot. Suppose economic inequality has actually increased very little lately. Suppose that the American lower and middle classes have over the last few decades enjoyed gains from economic growth equal to those enjoyed by the bulk of the upper class. Would you conclude that Krugman raised a false alarm? I hope you would. Would you say the United States’ socioeconomic system is really pretty fair after all? Would it change your opinions about what does or does not need to be done?

Maybe it shouldn’t.

Inequality Is Not an Indicator of Injustice

A good deal of popular liberal political and economic commentary treats a country’s level of economic inequality as a rough indicator of the justice or injustice of its institutions. This kind of pop liberal egalitarianism would suggest that the results laid out above should put our minds at ease. But I think the inequality-as-justice-barometer view is based on a mistake and is bound to confuse and distract us. The reason not to get up in arms about the United States’ level of income inequality is not that it’s really pretty decent, considering. The reason not to worry about it is that the level of income inequality is a fact astonishingly devoid of morally relevant information and at best shares a common cause with real problems.

The level of inequality within a country can be influenced by a variety of factors. Some factors, like the age composition of the population, are plainly irrelevant to questions about the justice or praiseworthiness of a country’s institutions. Some good things tend to make inequality rise. Some bad things tend to make inequality fall. The same level of inequality can have better or worse underlying causes. For example, the United States and Ghana have approximately the same level of income inequality, as measured by the Gini coefficient, but on any sensible account of justice or goodness, American institutions are a lot better. Neither the level nor the trend of inequality conveys the sort of thing we need to know in order to responsibly pass judgment on a society’s institutions or its social, political, and economic order.

If income inequality in the United States is symptomatic of injustice, the problem is unlikely to be the level of inequality as such, but the institutional mechanisms or social norms — such as predation by political elites or the systematic exclusion of ethnic minorities from economic opportunities — that tend to generate income inequality. If you believe that American income inequality does reflect injustice in the structure of America’s institutions, then it is important to identify precisely where and how the system is unjust instead of simply fixating on the fact that there is inequality. If the level of inequality is a knock-on effect of a more fundamental injustice, we should focus our attention on the original site of wrongdoing. The fire is the problem, not the people who look around when they hear the alarm.

Let me illustrate my point by examining two of the most popular inequality-influencing mechanisms: “skill-biased technical change” and executive compensation.

So, some innovations in technology increase the productivity of some groups of workers more than of other groups of workers. A worker’s wage reflects her productivity (among other things). And new technology can make some workers more productive while doing little or nothing for others. As it happens, a great deal of new technology over the past several decades has been “skill-biased,” meaning that it has enhanced the productivity of highly skilled workers more than less-skilled workers. This is a large part of the economics profession’s most popular theory of rising wage and income inequality.

So what are we to make of this? Taken by itself, the fact of “skill-biased technical change” seems neutral as a matter of justice or social morality. The actual distribution of skills may reflect injustice in, say,  the education system. (I think it does.) But universal access to indisputably satisfactory schooling will by no means erase all the inequalities in skill that skill-biased technical change parlays into wage inequality. The mere possibility of this kind of morally neutral process widening the income gap should leave us skeptical of barometer-of-justice approaches to inequality. Of course, skill-biased technical change is not just a possibility.

The demand for skilled labor relative to supply is another piece of the “skill-biased technical change” story. Wages reflect productivity, but they also reflect supply and demand. The American educational system, the story goes, has not been producing skilled workers at a rate sufficient to keep up with the demand for them. This has contributed to an increasing wage premium for skilled workers, which has boosted inequality.

Is the low supply of skilled labor relative to demand also a neutral inequality-influencing factor? As with many factors that influence the level of inequality, I don’t think it is morally neutral, but I also don’t  think its effect on nation-level inequality has much to do with it.

Unlike the fact that much new technology does not affect productivity uniformly, a shortage of skilled workers can be more or less directly alleviated by government policy. If injustice in the public school system is partly responsible for poor skills and flagging rates of growth in college admission, we can’t treat the low supply of skilled workers as a neutral fact. It reflects a real problem. Additionally, the United States could easily admit many more skilled foreign workers, but does not. In the context of rising demand for skilled workers, a not-very-generous cap on visas for skilled workers from abroad acts as a subsidy to American skilled workers, and this puts upward pressure on American income inequality.

Now, I would contend that the effect of the quota on visas for skilled foreign workers on the level of American income inequality is neither good nor bad. I think the real questions of justice here pertain to access to adequate education, the legitimacy of an implicit government-granted subsidy, and the openness of the United States’ labor markets to foreign workers. Maybe I’m wrong that these are the real issues. The point is that locating the real problems, whatever they are, is incredibly important. Fixing on income inequality and moving straightaway to thoughts of redistribution helps us do the wrong thing.

Let’s glance at CEO pay as a final example. Changes in executive compensation are almost certainly behind the dramatically disproportionate income gains in the top one percent of the distribution. After all the scholarly tweaks to the income datasets and the Consumer Price Index are done, that’s where the action remains. So why did those changes occur? I really don’t know why, but the story matters.

If the story boils down to a change in the tax code and a shift in social norms about compensation, then we’ll need to dig deeper. Was the change in the tax code wrong? Ill-motivated?  A solution to some other problem with unforeseen consequences? Did the old norms of remunerative moderation collapse under the weight of Ayn Rand novels? Or what? Let’s say the story is that an old-boys network of scheming CEOs sitting on one another’s boards of directors conspired to give each other raise after massive raise with no regard for the interests of their firms’ shareholders and creditors. That’s a story of theft on a grand scale. But theft is an injustice whether or not it moves the Gini coefficient. Any rise in national income inequality due this or, say, Bernie Madoff’s boundless cupidity would be an epiphenomenon of injustice, the echo of a gunshot.

Keeping Our Eye on Injustice

There’s ample of reason to suspect that the gap in standards of living has widened a whole lot less than Paul Krugman and most everybody thought. My point is, So what? That fact doesn’t tell us what we need to know.

What’s really interesting is why inequality has been overestimated. If, for example, large systemic economic forces — expanded trade with China, Wal-Mart’s downward pressure on prices — helped make poorer Americans a good deal richer than we had thought they were…. Well, that’s terrific. And informative. This story suggests who might get hurt if the United States gets in a trade war in China, or if Wal-Mart unionizes and shifts some gains from trade away from consumers and toward its employees. Getting the story straight is the first step in making a better story.

Economic inequality is too easy. It makes us lazy in our quest to sniff out injustice, which, after all, is not so hard to find. There is overwhelming reason to believe that in the United States the deck really is stacked against some people. As a consequence, many millions are doing much less well than they might be.

Legions of inner-city kids consigned to abysmal public schools are systematically denied a fair chance to develop the capacities need to participate fully in our institutions, or to enjoy their potentially ample rewards. The United States imprisons a larger share of its citizens than any country on Earth, literally disenfranchising hundreds of thousands of men and women (though they are mostly men) and leaving hundreds of thousands more dispirited and damaged. Undocumented immigrant workers increasingly constitute a permanent economic underclass explicitly denied many of the basic legal protections of citizens, inviting both government and private abuse. And, at the level of culture, patterns of private discrimination continue to constitute for millions a web of real, seemingly inescapable barriers to opportunity and achievement, and help to generate self-reproducing patterns of diminished expectations and wasted potential. We should focus  our attention and energy to the task of rectifying these vicious injustices.

Maybe fixing all this would decrease the variance in national incomes. But the idea that fixing all this somehow requires “fixing” the pattern of incomes is an excellent way to avoid the real problems and fix nothing.

Will Wilkinson is a research fellow at the Cato Institute and editor of Cato Unbound

Response Essays

Is Consumption the Grail for Inequality Skeptics?

According to the best available data, in 1979 the average after-tax income of the top 1% of American households was approximately $340,000. By 2006, at a similar point in the business cycle, it had risen to $1,200,000 [1]. The average for the bottom 60% of households rose much more modestly, from $29,000 in 1979 to $35,000 in 2006. Over this period the share of total income going to the top 1% of households jumped from 7% to 16%, while the share for the bottom three quintiles fell from 36% to 28%.[2] This is a substantial rise in income inequality. Very few social scientists deny its existence. The debate among them focuses on its characteristics, timing, magnitude, and causes.[3]

But there is far less consensus about whether, and if so to what extent, we should worry about this development. Inequality skeptics have offered a number of reasons for downplaying its significance: inequality is the product of free choices, what really matters is equality of opportunity, measures of income inequality ignore upward mobility, higher inequality boosts economic growth, focusing on a single country fetishizes national borders, and others.

Will Wilkinson emphasizes the distinction between income and consumption:

As far as I can tell, when most people are worried about economic inequality, they’re usually worried about inequalities in real standards of living — in the real material conditions of life… An individual’s or household’s standard of living is determined rather more directly by the level of consumption than by the level of income… Different datasets and analytical methods produce somewhat different results, but most stories of consumption inequality are stories of stability or a relatively mild rise.

Wilkinson makes other arguments in his piece, and additional ones in his earlier Cato essay. But I want to focus on this point. I don’t find it compelling.

One reason is that existing analyses of consumption inequality suffer from a problem similar to that which, until recently, hindered the study of income inequality: limited data on those at the top. Consumption data come from the Consumer Expenditure Survey (CEX). Like its income counterpart, the Current Population Survey (CPS), the CEX is not designed to effectively capture developments at the top end of the distribution. Since this is where a good bit of the rise in income inequality is centered, researchers may have underestimated the degree to which consumption inequality has increased.

Suppose, though, that the very rich have been consuming relatively little of their additional income. Should we then conclude that the economic inequality we care about hasn’t risen much? No. The fact that income isn’t spent doesn’t render it irrelevant. If my income were to balloon to more than a million dollars, my household might not increase its consumption by much. But it’s not as though the additional income would thereby disappear. I could cut back on teaching and devote more of my time to research, or take an unpaid sabbatical. My wife could quit her job and spend more time with our children or do more volunteer work. Or we could invest the money, which might produce considerable additional income in future years.[4] This could help ensure that, among other things, we’d be able to afford to send our kids to expensive private colleges. Or we could retire early. Or simply accumulate assets and pass them on when we die. None of these uses would show up as consumption in the survey data (except the college payments, though that would come some years down the road). But they surely would enhance our well-being.

The point is that income adds value even if it is not spent right away. Consider, as a potentially helpful analogy, political freedom. On virtually every scoring or classification of political liberties, the United States receives the highest possible score.[5] Yet many Americans make little use of these liberties. Only about half of those eligible actually vote in presidential elections, and the share is much smaller for off-year national elections and smaller still for local elections. Many who do vote have limited knowledge of key issues, and they tend to be heavily myopic in their thinking.[6] Few Americans participate in politics in other ways, such as active involvement in a party or other political organization, campaigning for a candidate, engaging in organized political discussions, or giving money to a political party or candidate. Only a small portion of Americans, in other words, make much use of the political freedoms they enjoy. If we care about political liberty only insofar as people make use of it, perhaps we should judge our country to have far less political freedom than is commonly thought.

I don’t think that’s the right interpretation, though. It matters a great deal that Americans have the right to vote, to assemble, to join a political organization, and to demonstrate, even if many don’t exercise these liberties. Similarly, it matters if the incomes of those at the top triple or quadruple, even if they use only a limited part of the additional income to boost their consumption.

There also is the issue of how increased consumption among households at the low end and in the middle has been financed. Consumption smoothing over the life course — borrowing when young, repaying later when income is greater — is one thing. But if growth in consumption inequality among the bottom 99% has been suppressed by poor and middle-income Americans borrowing too heavily, this ought to be of concern.

I am not suggesting we shouldn’t care about the distinction between income and consumption. It matters that huge income increases at the top helped propel a housing bubble that raised the price of expensive homes, especially in and around the cities where a disproportionate share of the top 1% live. It matters that Wal-Mart and imports from China have reduced the prices of many consumer goods for low-income Americans. It matters that many people now have access to communication via e-mail and cell phones, information via the Internet, and entertainment via cable TV and iPods. And it helps when government services enhance material well-being — by reducing violent crime or expanding access to health insurance, for instance. Here’s how I put this latter point in a recent comment:

Imagine an America in which high-quality public services raise the consumption floor to a high level: most citizens can put their kids in high-quality child care followed by good public schooling and affordable access to a good college; they have access to good health care throughout life; they can get to or near work on clean and efficient public transportation or roads with limited congestion; they enjoy clean and safe neighborhoods, parks, roads, museums, libraries, and other public spaces; they have low-cost access to information, communication, and entertainment via reliable high-speed broadband; they have four weeks of paid vacation each year, an additional week or so of paid sickness leave, and a year of paid family leave to care for a child or other needy relative. Even if the degree of income inequality were no less than today and we still had CEOs, financiers, and entertainers raking in tens or hundreds of millions of dollars in a single year, that society would be markedly less unequal than our current one.

I agree that we should pay attention to consumption in assessing changes in economic inequality. But we need better data on consumption at the top. And even if consumption inequality has increased only modestly, that by no means renders the large rise in income inequality moot [7].



[1] These figures are in 2006 dollars. The top 1% includes about 1.1 million households.

[2] These data are from the Congressional Budget Office. They are estimates created by merging IRS tax records with Current Population Survey data on incomes. 2006 is the most recent year for which these data are available. For additional discussion, a much longer time series on incomes at the top, and cross-country comparison, see the recent paper by Anthony Atkinson, Thomas Piketty, and Emmanuel Saez (free version here).

[3] Wilkinson mentions a recent paper by Robert Gordon as questioning this conclusion. But as is clear from the passage Wilkinson cites, while Gordon raises questions about the timing and nature of the rise in inequality among the bottom 99% of households, he does not challenge the consensus view that with the top 1% included income inequality has risen sharply.

[4] This is a key part of the reason why the distribution of wealth is much more unequal than the distribution of income.

[5] See, among others, the Freedom House scores.

[6] See Larry Bartels’ Unequal Democracy.

[7] For my views on why we should be concerned about rising income inequality, see my “Rising inequality: what’s the problem?” and chapter 2 of my book Jobs with Equality.

Why Things Will Feel Worse As They Get Better: The Downside of Growing Consumption Equality

There is little doubt in the economics profession that popular views of income inequality are misleading. Furthermore there is good evidence that the trends that have been observed exaggerate the rise in true inequality. Will aptly summarizes some of the more important recent findings in the area. He could have gone further by noting that inequality seems misleadingly worse when looking at households vs. individuals (people are more able/willing to live alone today, thereby lowering the average) or that high immigration may complicate the story (low-wage immigrants might plausibly raise natives’ income yet still increase dispersion and lower the measured averages) or that a focus on pretax earnings cannot — by definition — show the benefits of progressive tax policy designed for redistribution.

Nonetheless, there is still evidence that income inequality has risen over the last 30-40 years and there is certainly the perception that incomes have grown more unevenly. That the numbers show that the change is smaller than we had first thought doesn’t seem to affect how largely the issue looms in public policy debates. This suggests that much of the debate on inequality is not just about inequality but also beliefs about fairness and about which social outcomes are to be encouraged or critiqued. Figuring out the source of our disagreements is made worse when we look to evidence that is misinterpreted or is poorly tied to the policies we wish to advocate.

While Will tries to get us to focus on the justice or injustice of various social institutions as the “true” issue we should be concerned about, I want to make a different observation that flows from theoretical and empirical considerations about economic inequality.

The desire for greater equality is partly an outgrowth of the worldwide spread of democracy and democratizing tendencies over the last two centuries. Tocqueville was one of the first to notice how important the sense of equality was to democratic America, quipping that Americans “would rather be equal in slavery than unequal in freedom.” But ironically, the very economic growth that has fostered greater material equality of consumption in the industrial era, and the expectations of continued improvements in both absolute welfare and relative equality that have accompanied this growth are likely to make the perception of inequality worse no matter what the data can be made to demonstrate.

And this is because of the problem of positional or status goods. To understand the core idea, it is best to think of the different classes of goods and services that humans consume. Assume for the sake of argument that many, perhaps most goods are theoretically reproducible with improved technology and greater efficiency. Food, clothing, cars, equipment, toys all fall in this category. Expensive as an MRI machine or luxury sports car are today, we can easily imagine worlds in which those goods (or their close substitutes) become cheap enough so that almost anyone could afford them. In contrast there are some goods — let us say, the most desirable housing spot, or the best view of a lake — which cannot be readily multiplied simply by changing technology and improving economy. Then by definition, any changes that improve global availability of the former category of goods, even to the point of absolute equality, will mean that the remaining category of goods that cannot be easily reproduced will loom larger in people’s consciousness. When salt and spices were expensive, access to these were part of the measures by which people differentiated themselves materially. Today, salt is so trivially cheap that it’s not worth considering in a comparative budget. As more items and services — whether TV sets or MP3 players or computers or even triple bypasses — become widely available, then by definition, the goods and services that remain objects of envy and inequality will be those that are most insensitive to improved efficiencies. Hence our sense of perceived inequality might actually increase as material inequality decreases.

Some goods may literally be fixed — there are only so many Picassos in the world today. Others are fixed because they are purely about relative position — there can by definition only be ten top 10 universities or ten best places to live. Some of these — such as desirable locations in a city — might vary with the number of desirable cities and the quality of public access. But the latter will vary in ways that are only tangentially related to improved productivity. And increased wealth will not make access easier. Indeed, to the extent that greater income allows more people to bid for these positional goods, we will be frustrated to find that our expectations of which goods we can consume will not match our reality if everyone gets wealthier at similar rates. Moreover, goods dependent on human labor must perforce become relatively more expensive as people get richer, because to say that people are more prosperous is to say that human labor is more expensive.

The upshot of all this is that the better off we are, the more we spend and focus attention on consuming items which are less amenable to economic and technological fixes.

Housing is an interesting case because it is a curious mixture of a physical commodity reproducible with better technology and a positional good not amenable to productivity gains. Consider what people mean when they say that a house is expensive. They are not usually speaking about the physical structure per se. A new 3-bedroom house costs vastly more in a nice area of San Francisco or New York than an identical house would in rural Idaho or Tennessee. Thus, housing is expensive because certain locations are more desirable. Some of this might simply be snob appeal. Some locations may have more status. But much of that cost is driven by what other people want to do. Where people wish to live is often where jobs are available. Your preferred neighbors are also those with desirable social traits and appropriate income status. Is the local school system good? That has more to do with the quality of the students who attend a particular public school than the physical structure of facilities of the school. The latter are not unimportant, and good schools in good areas are more likely to have more resources as well. But affluent children from good households tend to make better students than poor and ill-prepared children from low income families suffering from various dysfunctions. Good facilities tied to poorly motivated students from uninvolved families located in run-down areas will not appeal to someone paying large sums for a nice house.

These are differences in consumption. But the problem of non-material inequalities will be reflected in payment schemes as well. Growing inequality in high-end compensation has been partly driven by the laudable goal of limiting positional or other non-material perks to the firm’s leaders. One of the perks of management that has slowly been devalued in the more democratic and open marketplace has been security of tenure and freedom from competition. In earlier times or in less open economies, access to control of important firms and the political bureaucracy was more restricted than it is in the major developed nations today. The ability to limit competition was also tied to greater deference from subordinates and the capacity to use company funds for personal expenditures. The economist John Hicks once noted that the best of all monopoly profits is a quiet life. As monopolies have been eroded and secure positions have been threatened, firms have had to offer larger and larger financial sums to motivate their best individuals, who find it not only much easier to change jobs than before but even to change residency or citizenship. The more socially constrained the access to jobs and position, the less unequal the official monetary payments need to be. The more one insists on social equality — or at least its pretense — in the workplace, the more the elite will ask to be compensated in raw financial terms. Jobs which are more secure and grant tenure will tend to pay less — all else being equal — than those which are volatile and reward risk-taking.

There are some good unintended consequences of this positional struggle. Because positional and other fixed goods become harder to buy as people get wealthier, those who are particularly status-obsessed will have to work more to buy these goods, which ultimately benefits us all. Working longer hours to buy the choicest piece of real estate or to be the first to get the latest new toys serves as a voluntary progressive tax on income. Especially in the latter case: those who slaved away to buy flat screen TVs when they were rare and cost $10,000 made it easier for most of us to eventually own similar sets at a tenth the price.

Conversely, attempts to limit positional competition only push status seeking to different dimensions, which are often less accessible to all. Consider that in countries like the Soviet Union, where the role of money and markets was officially repressed, competition became fiercer for government connections and “socially approved” professions, like chess master or ballet dancer, since that was one way to obtain both fame and highly restricted material goods like special imports.

Perhaps the real issue that will be taken up by the other commenters will be to discuss why certain inequalities which are also unevenly distributed — such as looks, intelligence, ability, or personality — do not invite as much social envy or opprobrium as disparities in income or wealth resulting from hard work or shrewd dealing. These differences are probably as large or larger than the measured inequalities in dollar income, yet go unmeasured and often excite no commentary in discussions of inequality. Sometimes differences in income are simply byproducts of these differing endowments. But it is rare that policies are directly and specifically designed to tax those whom nature has favored unfairly rather than those who have chosen to convert their advantages into cold hard cash.

But given my limited space I am content to repeat my main point. Whatever one’s views on material inequality, it will certainly be the case that more we succeed in promoting wider access through growth, the more we will be focused on acquiring and differentiating ourselves on margins that are hard to change or difficult to equalize. Which will inevitably lead to more frustration as we mistake the material manifestation of those inequalities for the underlying differences themselves.

What Should Egalitarians Want?

I agree with Will Wilkinson that, considered in isolation, the pattern of income distribution in a society doesn’t tell us much about whether that society is just or unjust. Nevertheless, as an egalitarian, I think inequality in wealth and income, when considered in its causal context matters a lot. Economic inequality is objectionable when it makes people, especially the less well-off, less well-off than they otherwise would be.It is also objectionable when it reflects government favoritism toward the better-off.In the United States, economic inequality is objectionable for both of these reasons.

There are several ways in which economic inequality undermines people’s well-being.(I don’t claim that mere awareness that others are better off makes people worse off:I don’t credit envy with a genuine grievance.)For lack of space I’ll set aside the intriguing evidence that economic inequality makes people ill, reduces life expectancy, increases stress and violence, and undermines social trust.Here I’ll argue that economic inequality (1) makes the less well-off vulnerable to domination, (2) makes them vulnerable to stigmatization, and (3) causes governments to favor the better-off at the expense of the worse-off.

Consider the problem of domination.People’s bargaining power is tied to their relative exit costs.This often enables people with superior wealth to dominate others.Women are more likely to submit to domestic abuse and control if the cost of leaving their partners is a drastic fall in material well-being.Low-wage workers are often unable to bargain for predictable work hours because they can’t afford to quit.This empowers employers to wreak havoc with workers’ lives, as arbitrary changes to their work schedules leave them scrambling for (often unavailable or very costly) child care, or forced to withdraw from community college courses, turning their hard-earned tuition payments into a deadweight loss and crushing their dreams for a better life.Such despotic control over workers’ time can be humiliating.At one Nabisco plant, female workers were forbidden to take time off the production line to use the restroom, and were told to wear diapers.

Economic inequality can also lead to stigmatization.Adam Smith famously observed that, as the general level of consumption increases, so does the level of consumption needed by each individual to be able to appear in public without shame:

A linen shirt, for example, is, strictly speaking, not a necessary of life. The Greeks and Romans lived, I suppose, very comfortably, though they had no linen. But in the present times, through the greater part of Europe, a creditable day-labourer would be ashamed to appear in public without a linen shirt, the want of which would be supposed to denote that disgraceful degree of poverty, which, it is presumed, no body can well fall into without extreme bad conduct.

The consumption of the better-off thereby raises the cost of living for the worse off.The demands of respectable appearance can be crushing in regions where spiteful competition inflames a culture of conspicuous consumption.In the 2008 documentary Kids + Money, Lauren Greenfield depicts the competitive consumption among teenagers of all classes in Los Angeles, sometimes to the financial ruin of parents who are less well-off.One single mother, desperate to see her daughter wear socially accepted clothes, couldn’t pay her utility bills because she spent her income on a pricey t-shirt for her daughter.

Finally, economic inequality causes governments to make favorites of the better-off and neglect the interests of the worse-off.A recent study by Larry Bartels finds that the preferences of low-income constituents have no influence at all on senators’ votes, while the preferences of middle-income constituents have middling influence, and the views of high-income constituents had great influence. (His findings apply to Democratic as well as Republican senators.)

How does catering to the preferences of the well-off harm the worse-off?In some cases, the state grants or extends monopoly privileges to the better off.Consider the extension of copyright terms, dramatic consolidation of broadcasting ownership, and huge giveaways of public spectrum to existing broadcasters due to IP and communications laws passed in the 1990s.In other cases, the state effectively exempts the well-off from laws that are supposed to protect the worse-off.After discovering that Wal-Mart had violated child labor laws dozens of times, the Labor Department under President Bush agreed to give Wal-Mart 15 days notice before inspecting its stores for further violations, giving it time to hide its lawbreaking from authorities.This is hardly an exception.Many employers steal from their low-wage workers with impunity.One recent study found that a quarter of low-wage workers were paid less than the minimum wage, most were denied meal breaks to which they were legally entitled, many were forced to work off the clock, not paid for overtime, suffered the theft of their tips, and illegally fined, among numerous other labor-law violations.In still other cases, the state grants the better-off legal loopholes or vast subsidies at the expense of the worse-off.Payday loans are largely exempt from usury laws.For years, the federal government paid a 9.5% interest rate to student loan companies, even though prevailing interest rates were much lower than this, and the government guaranteed the loans, obviating any rationale for a risk premium.Even after Congress had enacted a law to end this absurd subsidy, the Bush administration failed to act on it.That’s billions of dollars to the wealthy at taxpayer expense.

Yet to focus on government subsidies and favors to wealthy rent-seekers offers a narrow view of the ways the better off use their superior wealth to skew government policy toward their interests, at the expense of the less well-off.As economic inequality increases, the better off perceive fewer and fewer shared interests with the less well-off.Because they buy many critical goods — health insurance, education, security services, transportation, recreation facilities — individually from the private sector, or pool the provision of these goods within private gated communities or municipalities governed by zoning regulations designed to exclude the less well-off, they tend to oppose public provision of these goods to the wider population.At the same time, the ways they provide these goods for themselves raises the costs to the less well-off of obtaining them.When the better-off vote down public transportation because they own their own cars, the less well-off must buy cars too.When the better-off spend more on schooling, so too must the less well-off, to give their children a fighting chance to compete for better opportunities.Since municipal zoning regulations create class-segregated cities, the less well-off must tax themselves more heavily to provide inferior public services to what their better off counterparts in other towns pay.

Wilkinson’s evidence on consumption inequality fails to account for these facts, because he looks only at individuals’ consumption of private goods, not their access to publicly provided goods.Yet for the less well-off, consumption of the sorts of consumer goods that the market has — to its credit — so successfully delivered to all classes pales in importance to public and publicly provided goods.To live in a low-crime, orderly, unpolluted neighborhood, free of run-down and abandoned property, graffiti-marred buildings, open drug dealing, prostitution, and gangs; to have access to public parks where one’s children can safely play, to well-maintained sidewalks and roads, to schools that offer an education good enough to qualify one for more than menial, dead-end jobs: how many cell phones and athletic shoes is that worth?Enjoying warranted confidence that the police will, if one is innocent of any crime, treat one with the respect one is due, rather than as a criminal suspect, and extend one the protections of the law rather than police harassment, accusation, and prosecutorial abuse:here there can be no question of comparison to consumer goods. But access to these public goods depends on one’s relative wealth.As wealth inequality increases, such goods are less and less within reach of the less well-off.

Wilkinson says, “There is overwhelming reason to believe that in the United States the deck really is stacked against some people. As a consequence, many millions are doing much less well than they might be.”I agree entirely with this judgment.I also agree with Wilkinson that to deal with this, we must get the causal story right, and attack the root causes of injustice, rather than just the epiphenomena.I suspect that we also agree that right causal story would identify state action (or inaction) as critical to stacking the deck.But I suggest that economic inequality is also among the root causes of the stacked deck, because it motivates the affluent to skew state policy in their interests, at the expense of the less well-off.As wealth inequality increases, the rich lose interest in state action to promote security, prosperity, and opportunity to all, and instead use their vast wealth advantage to buy access to government officials via campaign contributions, and to out-lobby more diffuse and less well-funded public interests.They thereby acquire agenda-setting and veto powers over public policies, which they use to limit provision of public goods to the less advantaged, and to obtain special privileges for themselves — monopolies, subsidies to favored industries, and effective exemption from laws that are supposed to protect the less well-off from exploitation and domination by the better-off.Economic inequality thereby harms the worse off, and reduces them to second-class citizens.

The Conversation

Is Redistribution the Wrong Cure?

In my initial comment I addressed Will Wilkinson’s contention that consumption inequality is much more important than income inequality. Here I turn to his central argument, which is in the second half of his essay.

Will argues that regardless of the degree to which inequality has risen, focusing on it is a distraction from the mechanisms that contribute to that rise, and some (not all) of those mechanisms are the real things we should object to and try to fix. He seems to be saying that the standard liberal or progressive proceeds as follows: observe rising income inequality –> advocate an increase in redistribution to reverse that rise. Will is suggesting that instead we should inquire into the causes of the rise in inequality. If a particular cause is not morally objectionable — such as, in his view, skill-biased technological change or an increase in immigration — no policy response is called for. If the cause is morally objectionable — such as unequal access to good schooling or racial discrimination — we should try to address the cause rather than turn to redistribution as the cure.

I have a different view, though it isn’t diametrically opposed to Will’s. First, I agree that we should look for the particular injustices that underlie a rise in income inequality. Indeed, we should do that whether or not inequality is increasing.

Second, I also agree that some of the causes of rising inequality are more objectionable and deserving of policy remedy than others. For instance, though the research isn’t conclusive, I strongly suspect that increases in imports from low-wage countries and in immigration into the United States  have contributed to the rise in income inequality over the past several decades. But unlike some American progressives, I believe that on balance both are desirable developments, so I don’t favor heightened trade or immigration restrictions as a policy response. The same holds for technological change.

I do favor government action to help the Americans most vulnerable to these processes to better adapt and adjust. Potentially helpful policies include more aggressive early education, improved K-12 schooling, portable health insurance and pensions, more generous unemployment compensation, wage insurance, retraining, job placement assistance, a minimum wage pegged to inflation, and a beefed-up EITC for those without children. Some of these involve “redistribution” in the usual understanding of the term, while some don’t.

There are other injustices that I, like Will, think we should attempt to address head-on, irrespective of their impact on income inequality. I agree with him that discrimination and inadequate schools are among these.

Suppose, optimistically, that we succeed in reducing these barriers to opportunity and yet income inequality continues to rise. My sense is that Will would say no policy response — in particular, no increase in redistribution — is merited. In this view, everyone starts out with roughly equal opportunity. Differences in outcomes, such as earnings and income, are a product mainly of effort and hence are morally neutral.

I disagree. Hard work, drive, and diligence are heavily influenced by luck. To a considerable extent, they are a result of things over which we have no control: our genes, development in utero, birth order, our parents’ traits, childhood nutrition and health, early social experiences with peers, and stumbling into an occupation that suits our likes and abilities, among others.

Now, the fact that luck plays a significant role in outcomes by no means implies that we should redistribute until outcomes are perfectly equal, or even close to equal. How much to redistribute requires taking a variety of things into account, including administrative costs and incentives for work, investment, and innovation. But luck’s influence means that redistribution is a justifiable remedy. Skyrocketing CEO pay does not have to be the product of improper collusion, and wage stagnation need not owe to racial discrimination, for us to consider an expansion of taxes and/or transfers an appropriate response. While we sniff out and rectify the injustices on which Will commendably wants us to focus, it is perfectly reasonable to also respond to heightened income inequality by increasing redistribution.

Kenworthy on Consumption and the Value of Savings

I’m immensely grateful to professors Kenworthy, Nye, and Anderson for such thoughtful, challenging replies to my initial contribution. I could spend days responding in detail to each essay — and I probably will! I’d intended to write a relatively short omnibus post touching on one or two points in each of the reaction essays, but in the interest of getting the informal blog discussion under way, I’m going go ahead and post my response to each essay separately as soon as I finish it.

First up: Lane Kenworthy!

I have argued that consumption is a better proxy for material welfare than income and that the consumption gap has widened much less than the income gap, if it has widened at all. So, I conclude, inequality in material welfare has not increased much. Kenworthy argues that I’m overlooking the value of accumulated wealth. Money in the bank, he says, widens our horizon of opportunity and these opportunities are valuable to us even if we do not take advantage of them. I agree entirely. But I still believe that, among standard economic measures, consumption is the best indicator of  individual’s or household’s real standard of living.

So what accounts for our disagreement? I’m not certain, but there seem to be two questions at issue. I think the fundamental question is: What are we trying to measure? The second question is: What’s the effect of savings on whatever it is we’re trying to measure?

I’ve been speaking variously, and probably somewhat carelessly, about “material welfare,” “real standard of living,” and “economic well-being.” Whether or not these terms mean precisely the same thing, I think people are generally talking about something along these lines when they’re talking about distinctively economic inequality. Whatever economic well-being is, it’s narrower and less comprehensive than overall well-being. Overall well-being might include such things as happiness, health, a sense of purpose, the development and exercise of certain human capacities, and more. I take it that specifically economic well-being has something to do with that part of overall well-being directly affected by the economic resources at our disposal, or at the disposal of our household. What we consider a sensible measure of economic well-being, then, is likely to turn on what we count as an economic resource and the weight we put on the various effects of these resources on our well-being.

So, for example, money in the bank, or in a mutual fund, is indisputably an economic resource. And I’m inclined to agree with Kenworthy that the peace of mind and sense of open opportunity that affords should be included in an accounting of economic well-being. But suppose I’ve invested a great deal in gaining skills greatly in demand in the labor market. Is my “human capital” an economic resource? Suppose I’ve got a stellar credit rating and so I could get access to a big line of credit with which I could do all sorts of things. My sense is that both human capital and reputational capital should count in much same savings does. I am, as a matter of fact, still in debt from graduate school. I certainly don’t worry about it, since I’m confident that I’ll come out way ahead in the end. But my “coming out ahead” is largely a matter of future consumption — a matter of what I expect to do with the money I’ll make. And it seems to me that the expectation of future consumption is a large part of the value of savings, or of the value of access to credit. So, while current consumption may be a bit misleading as a measure of an individual’s or household’s current economic well-being, aggregate measures that encompass people at all points of the life-cycle seem to me pretty likely to capture much of the value of savings, access to credit, and so on.

That said, I think nominal consumption remains a pretty rough measure of economic well-being. For example, the quality of public goods in a place certainly contributes to the standard of living of the people who live there. The quality of public goods and services here Iowa City is excellent — much better than in Washington, D.C. Public schools in Iowa are remarkably good. (Iowa consistently has one of the highest high-school graduation rates and one of the highest average standardized test scores in the United States, which is really remarkable when you think about it.) Basic cost-of-living issues aside, it seems pretty clear that both income and consumption measures badly underestimate the level of real consumption enjoyed most Iowans — especially low-income Iowans. Meanwhile, low-income Washingtonians likely have it rather worse than consumption surveys would indicate. It might be interesting to talk a bit how social scientists could correct for this sort of thing in their estimates of economic inequality.

What Does Income Inequality Tell Us?

Lane Kenworthy gets to the heart of the matter. He states that even after we correct for discrimination, basic health and welfare needs, et cetera, if inequality continues to rise, it should be cause for redistribution. He says, “luck’s influence means that redistribution is a justifiable remedy.” I disagree. Moreover, I believe that we do not and almost never have enough evidence (certainly not at the national level) to make judgments about whether luck was the primary cause of a given inequality. What philosophical principle tells us that it’s okay for the top ten percent to make five times more than the poorest ten percent (as an example) while earning 100 times the bottom ten percent is not? And how much do you need to know about what generated that disparity to decide?

There is no way to infer useful information about the nature of just or unjust compensation from the raw factoid of an aggregate income distribution for a nation, state, or even small community. That some income inequalities may be associated with grave injustices does not tell us which measured inequalities are relevant.

Consider the recent housing crisis. Say that two families start out in the early 90s with similar incomes and jobs. But family A saves and invests prudently and conservatively, limiting family spending, sometimes in unfashionable ways (e.g. children are subject to the mockery of school fashionistas or colleagues deride parents for not eating out enough), and chooses to rent accommodations till 2009 when the crisis makes it possible to buy a home cheaply and with low interest. They now have reasonably high net worth and are quite well off.

Family B does not save much, borrows and spends on disposable consumer goods, and buys a house in 2005 with no money down and a subprime variable rate. The market crashes. They have difficulty making the monthly payments which have been adjusted upwards and they cannot refinance. In 2009 our measures tell us that family A is worth 10 times more than family B and the latter is in danger of losing their house to foreclosure. The raw fact of income or (in this case) wealth inequality will tell us nothing of how A and B got to the current situation nor guide us in deciding whether it is ethical to tax A to fund housing assistance for B.

And what about the non-financial inequalities I asked about? Consider that a boy from the ghetto who is bright enough to be singled out for a scholarship to MIT or Stanford may have no monetary assets at age 21 but might be — in a lifetime sense — effectively “wealthier” than his out-of-work uncle who is bad with numbers, likes to drink and gamble, but just won $1,000,000 in a lottery. Should someone talented and athletic enough to be drafted into the NBA or NFL be taxed for their good fortune even before they’ve earned a dime?

Or what if two equally talented people with equal opportunities take very different paths in life? Person A takes a satisfactory but low paying job with a lot of intellectual opportunity and a good support network. He is happy and healthy, but is living on a lower middle class income. Person B does not merely work hard, he takes financial risks, and even ruins his health by neglecting to eat and sleep properly and does not cultivate healthy social relationships with friends or peers. However, B manages to “hit the jackpot” and successfully turns a few thousand dollars into several million. I still do not understand the philosophical claims that would justify taking from B to give to A. Why should people who are equally endowed by nature and luck be treated differently because some chose to increase their financial wealth while others chose to focus on their social or spiritual lives? And this of course leaves out the inconvenient fact that we never have this much information about the people at the high end of the distribution relative to those at the lower end. It is equally hard to separate the deserving from the undeserving rich as it is to separate the deserving from the undeserving poor.

There are many, many nations where the bosses of top companies do not earn as large a premium relative to their workers as do CEOs in the United States. But that is more than compensated by secure positions, relatively large, often untaxed perks, and political and social clout that a U.S. billionaire might envy. Would we consider it to be an improvement if CEO salaries were capped in exchange for a greater “private-public” partnership in which officials of top companies joined with selected government regulators to engage in protectionist industrial policy that limited competition and promoted job “stability”?

Absent very specific contextual information, why should one ever care about raw income inequality? How do you know that measures to address inequality don’t create more injustice? What philosophical principles help me decide why monetary inequality is more deserving of attention than ability, beauty, social connections, or blind luck?

And who will take the blame if redistribution is mistakenly applied so it leads to a harmful loss of productivity and reduces economic growth (which hurts the poorest most)? The onus of showing the redistribution is harmless should be on those who choose to make the transfer.

Nye on Positional Goods and the Paradox of Growth

Before I reply to John Nye’s fascinating essay, I want to mention that I probably wouldn’t be writing about inequality if not for John. My views on the subject have been greatly shaped by a conversation we’ve been having on and off for almost ten years now. And a lot of what little I know about economics, I learned from John. He’s a good friend and a great teacher.

In his reply essay, John points out a striking paradox of modern economic growth: decreasing real economic inequality creates the illusion of its opposite. John points out that there are goods — positional goods — which are by definition in exceedingly short supply. As it turns out, people tend to care a great deal about their relative position in hierarchies of social status, and we tend to try to establish and signal status through consumption. As access to manufacturable goods becomes increasingly widespread, the status-signaling value of these goods declines relative to essentially scarce positional goods. Meanwhile, as incomes rise, a rising portion of the population finds itself with money to spend on status games. The result is a bidding war for positional goods that forces their prices to astronomical levels and ever further from an average household’s reach. Since most of us are tuned in to the local status-signaling channel, goods that broadcast strong status signals captivate us while other goods don’t. As broad prosperity advances, the background of attention fills with a swelling inventory of stuff plain folks have in common with rich folks while increasingly inaccessible positional goods loom ever larger in the mind’s eye.

I’m sold. I think John’s story brilliantly accounts for the data. Our society has become more economically egalitarian, in one profoundly importance sense at least, but it sure doesn’t seem like it.

One question I’ve been asking myself is how it is that John’s story can sound so similar to a story Robert Frank tells (in Status Luxury Fever, for example), while the lesson each draws from his story seems so different. I think much of the difference is explained by the fact that Frank sees little but waste (too many people competing to be rock stars!) and emotional harm (your BMW makes me feel worse about my Honda!) in positional competition, while John understands status-seeking more along the lines of Adam Smith and David Hume — as an engine of the efforts that drive growth and raise standards of living. If this is not quite John’s view, it is mine.

I’d say, along with Adam Smith, that our tendency to compare ourselves to others, and to want badly to compare favorably, is a source of frustration, dissatisfaction, and even humiliation (as Elizabeth Anderson rightly points out in her reply essay). But when our institutions and norms focus the energy of status-seeking into the right kinds of activities, it tends to leave almost all of us economically better off even if it leaves the most status-obsessed feeling wretched. And here’s John’s paradox from another angle: our hunger for inequality can, given a suitable institutional and cultural setting, act as an equalizing force.

This is counterintuitive, to say the least. And I think the differences between the views of people like John and me and the views of those more like Robert Frank (Lane? Elizabeth?) raise some really hard questions about how best to approach positional or status inequality when thinking about economic inequality. To repeat a theme from my response to Lane, I think it depends on what we have in mind when we’re thinking about “material welfare” or “real standard of living” or what have you. People evidently care a good deal about signaling to one another. Signals of various kinds are part of what many people are trying to buy with their money. If the price of signaling a certain level of social status rises sharply relative to income, as John says it has, should we take that into account when trying to determine how well off people are economically? Maybe we should. But how much weight should give it? I honestly don’t know, but I have some thoughts about it.

Status is multidimensional. Money isn’t the only status game in town. Many academics could make a lot more money doing something else, but there’s a certain respect and admiration extended to professors of economics or philosophy or sociology that is hard to come by as a financial analyst or a software engineer. I don’t want to say that well-regarded academics are economically better off than they might at first appear, due to their relatively high social status. I certainly don’t want to say that a professor making $75,000 at Harvard is economically better off than a professor making $75,000 at Northeastern, since Harvard is so much more prestigious.

Now, if the price of beef rises while the price of chicken stays the same, that doesn’t much affect the cost of living as long as people like chicken about as well as they like beef. Suppose the price of signaling the next position up in a certain domain of status competition rises rapidly relative to income growth. Can we just switch to another, more affordable, dimension of status competition in the way we can just switch from beef to chicken? I say yes. Do people do this? Yes. But what about people who just don’t like chicken? What if that’s most people? What if most people could learn to like chicken just as well as beef, but they never try? Do we say their money buys them less now that the price of beef has gone up? Or do we say they’re wasting money?

John asked why we give so little consideration to inequalities in natural endowments, unless they’ve been translated into economic inequalities. It’s a very good question. In the context of a conversation about economic inequality, the answer might seem obvious, but it isn’t. Some people are a lot funnier than other people, and being funny has a lot of advantages. But a quick wit and good comic timing can’t be bought. So humor inequality isn’t an economic inequality, right?

However, it used to be that a perfect set of teeth, which also has its advantages, couldn’t be bought. Now they can be bought. What if perfect teeth becomes the norm? Does that mean smile inequality became an economic inequality? If naturally perfect teeth weren’t an economic asset before, are they now? My total bafflement suggests to me that there’s something wrong with this question, but I’m not sure what it is.

Anderson on Economic Inequality, Domination, and Stigmatization

In her challenging reply essay, Elizabeth Anderson argues that economic inequality

(1) makes the less well-off vulnerable to domination, (2) makes them vulnerable to stigmatization, and (3) causes governments to favor the better-off at the expense of the worse-off.

In this post I’ll address the first two propositions. I’ll save the third, and I in my opinion the most compelling, for a separate post.


Anderson says that “People’s bargaining power is tied to their relative exit costs. This often enables people with superior wealth to dominate others.” I agree. However, it seems to me that most cases in which Jane Rich is able to dominate Joe Poor are cases in which Poor finds himself with no tolerable, feasible alternatives to submitting to Rich’s demands. Suppose Poor has open to him several perfectly decent, readily accessible courses of action that would put him out the reach of Rich’s domination. It seems that the problem is solved. Now suppose the wealth gap between Poor and Rich widens. It seems that the problem is still solved.

If, to use one of Anderson’s examples, the problem is that a low-wage worker, Joe Poor, is in too weak a bargaining position to negotiate a regular working schedule from his boss, Jane Rich, then what’s the solution? Suppose the government raises Jane Rich’s income tax rate 20 percent and then spends the revenue on bombs and health care for retirees. The disposable income gap between Poor and Rich will have narrowed considerably. Has it thereby become easier for Poor to negotiate regular hours from his Rich? It seems not. He still lacks decent alternatives, which suggests economic inequality is not the problem.

It may be that some government policy could both help Poor and reduce economic inequality. And it may be that some government policy could help Poor and increase economic inequality. For example, suppose Poor’s town reduces the tax rate on businesses, and this attracts new businesses that employ workers like Poor. The market for low-wage workers becomes a bit more competitive, and so Rich, worried about losing relatively experienced workers, starts giving Poor regular hours. Suppose Poor, now happier about his job, becomes a bit more productive and this makes Rich a little richer. I think that’s a good result!

Social Comparison and Stigma

Anderson argues that “as the general level of consumption increases, so does the level of consumption needed by each individual to be able to appear in public without shame.” This may be true, but I think it can be misleading. I think the question at hand is whether “the level of consumption needed by each individual to be able to appear in public without shame” has become easier or harder for the least well-off to attain. Anderson twice offers the example of clothes. The evidence suggests that it is easier than it has ever been for low-income households to buy affordable, quality, relatively fashionable clothes. School togs from Wal-Mart or Forever 21 may not have the cachet of designer gear. But if the status-conscious cool kids have to look close or check out the label to detect a bargain-basement designer knockoff, the kids with thrifty parents are doing just fine. However, some families can’t get to a good discount retailer and have no Internet connection. Even then, most cities have a Salvation Army, Goodwill, or St. Vincent’s within reach. The wealthier the society, the better the castoffs.

One of my worries about putting too much weight on the effects of social comparison is rooted in the fact that our habits of comparison and our reactions to our comparisons are to a fair degree up to us — or at least up to our parents and those who constitute our developmental milieu. I wouldn’t go as far as a Buddhist monk and say that our patterns of attention and reaction are almost entirely within control, but we do have some control over them, and we therefore bear some responsibility.

The story Anderson relates about the single mother who “couldn’t pay her utility bills because she spent her income on a pricey t-shirt for her daughter” strikes me not as a story about the toxic consequence of economic inequality, but a story about a mother with mixed up priorities. The mother did not have to be “desperate to see her daughter wear socially accepted clothes.” She was not compelled to care so deeply about the opinion of those among her daughters peers who would have found affordable clothes unacceptable. She did not have to leave unchallenged her daughter’s anxiety about fitting in. This is not to say that the mother bears all the responsibility. She didn’t get to pick her parents either. And there’s a good chance she would have chosen to pay her utility bill had she lived in a community in which missing payments is considered even less socially unacceptable than a child in perfectly decent but slightly dated clothes.

Finally, not to sound like a broken record, but it’s not clear what countrywide economic inequality has to do with this sad tale. The poorest households go broke trying to emulate the consumption patterns of slightly less poor households. Lopping off the top decile of the income distribution would leave this scenario entirely unchanged.

How Much Should We Redistribute? And How?

In an earlier post I argued that “luck’s influence means that redistribution is a justifiable remedy.” John Nye says he disagrees. This is a fairly straightforward difference in philosophical views, and I doubt further discussion of it would persuade anyone to change sides.

But suppose one does think luck’s impact justifies some redistribution. John still objects, because “we do not and almost never have enough evidence (certainly not at the national level) to make judgments about whether luck was the primary cause of a given inequality.” That’s true. Policy makers would need to know intricate details of each person’s life history. They can’t. Nor would we want them to. And even with this information there would be uncertainty about how much influence luck has had in any particular case.

So what should we, as a society, do? John seems to think the best course is to forego redistribution in order to avoid mistakenly rewarding someone who isn’t faring well but is in that condition mainly due to laziness, a short time horizon, or a preference for leisure, and to avoid mistakenly penalizing someone whose income is large because of prudence, hard work, or risk-taking. He offers a few hypothetical cases to illustrate. Undoubtedly such cases exist. But as best I can tell, he isn’t suggesting they are the norm. With any government action there are mistakes, but that in and of itself doesn’t strike me as a compelling reason to favor nonaction. Moreover, in my view these types of characteristics, preferences, and behaviors are, to an extent not often appreciated, themselves heavily influenced by luck.

As a practical matter, we, like all other affluent countries and most not-so-rich countries, have moved beyond the question of whether or not to redistribute. We do redistribute. The interesting questions have to do with how much to redistribute and how to do it.

The average inflation-adjusted pretax income of the top 1% of American households jumped from about $500,000 in 1979 to about $1,700,000 in 2006. For households in this group, the effective tax rate (taxes paid as a share of pretax income) for federal taxes in 2006 was about 31%. That was down from 37% in 1979 and 34% in the mid-to-late 1990s.[1] I favor pushing that effective tax rate up to around 35% (there are various ways to do this) and using some or all of the additional tax revenue to increase the Earned Income Tax Credit.

Would this improve the material well-being of Americans at the low end of the income distribution? Enhance their freedom and dignity? Would it boost employment? Stifle innovation? Impede economic growth? Reduce opposition to immigration? Increase social cohesion? These are, to my mind, the kinds of questions that ought to be at the forefront of debates about inequality and redistribution.


[1] These income and tax rate data are from the CBO. The federal taxes included in the CBO’s calculation of the effective tax rate are individual and corporate income taxes, the Social Security and Medicare payroll tax, and excise taxes.

How Much Equality Is Enough? And Who Decides?

I do not understand the argument Elizabeth Anderson is making.[1] In fact I agree with the Hayek quote she cites, but I don’t see its import. Let us assume you have the most generous welfare state in the world: How is that relevant to the measurement of income/wealth inequality?

To be concrete: Sweden already has a famously generous welfare state, yet in 2000 the Wallenberg family controlled roughly 40% of the value of the Swedish stock market (In contrast, Bill Gates’s 50 billion is less than one percent of the more than 10 trillion dollar U.S. market). Is that still unjust inequality? Are we better off taxing such families or opening up the market to further competition?

What measure of inequality do Kenworthy and Anderson wish to observe that would lead them to say, “Ok. No more needs to be done”?

And I agree with Lane Kenworthy that absolute equality is undesirable, so what measured inequality would lead him to say, “We have gone too far?” Should we trust Congress to decide this? Or a select committee perhaps? Does any country do a good job of dealing with measured inequality (as opposed to merely enhancing the welfare state)?

Even if I were to concede that such a redistribution might make sense, what tools do we have for deciding which measures tell us what to do?


[1] Anderson also says I stack the deck to only include stories where people get what they deserve. But I specifically cited a case where someone received $1M in a lottery and someone else got a scholarship but had no financial wealth. That has nothing to do with just deserts. Why should the former be taxed for the latter?

Brief Response to John Nye on How Much Equality Is Enough

John Nye asks: “What measure of inequality do Kenworthy and Anderson wish to observe that would lead them to say ‘Ok. No more needs to be done’?”

My response: There is no such precise measure or amount, and in my view it would be silly to offer one. My judgment about whether I would favor reducing the existing level of income inequality is based on how much it has increased and why, what seems feasible (taking into account, among other things, prior levels of inequality and levels achieved by other countries), and what the benefits and costs of reducing the existing level might be.

I think this type of question, like the one I addressed in my previous post, distracts us from genuinely interesting issues related to inequality and redistribution (including the one raised by John in his initial comment, which I hope to say something about in a future post).

Thinking Carefully about Monetary and Non-Monetary Inequality

There is a large disconnect between the specific problems and remedies that are being discussed here and the macro measures of income inequality that generated this debate. Measured inequality is often unaffected by these policies.

Consider Elizabeth Anderson’s desire to increase welfare and government transfers to the poor. No matter how generous these are made, they cannot have an effect on income equality since, by definition, these welfare payments are excluded from income measures which are usually derived from tax data. So if welfare payments are good, it is not because they address the measured income inequality.

It is well known that attempts to measure the incomes of the lower tail are severely biased by ignoring transfers from the state and from friends and relatives as well as the problem of unreported earnings from the underground economy. The latter issue includes anything from unreported tips and cash payments to illegal businesses and organized crime. For that reason, economists like to look at measures of consumption, which as Will has already noted show a much lesser degree of inequality than measures of income. This is consistent with the intuition that the lower end incomes are simply not measured properly or go unreported.

At the other end of the scale, it is also well known that reported income measures do not capture well the wealth of the very rich. That is why Lane Kenworthy’s suggestion of raising the top marginal income tax rate a few points is so odd. Research on worldwide trends suggests that the wealthiest members of society have been shifting income to equity because it makes it easier to shield their earnings from the taxman. And there are well known limits on how much and how easily you can tax equity and corporations without strong negative consequences. Moreover, the wealthiest are more able and willing to use legal and accounting techniques to hide their assets from the state. Hence any tax increase will almost certainly hit the upper middle class and high wage earning members of the population while leaving the true plutocrats relatively unscathed. It is easier to tax successful doctors and managers than the Warren Buffetts and Bill Gateses of this world. Indeed, in much of Europe there are lower corporate taxes than there are in the United States, and capital gains taxation is quite varied the world over.

At the same time, the fact that measured inequality ignores differences in the cost of living in different parts of the country means that a tax on high nominal income will punish dual wage earners in expensive areas relative to those who might have higher real living standards but lower costs. Do you really believe that big-city couples with high incomes should pay higher taxes than rural or suburban families with similar consumption but lower reported earnings? And of course, any attempt to adjust taxes to take this into account would result in nightmares of bureaucratic compliance and rent-seeking, not to mention induce distortions in production and the labor market. Moreover, it is not clear that the changes Lane Kenworthy proposes would do much to affect the overall income inequality measures. Would this justify continuing to increase taxation till some particular index flipped substantially?

For these and many, many other reasons (e.g. differences in age profiles will affect measured inequality, as will treatment of singles vs. households) most countries focus on simply enhancing welfare to the poor or balancing different public policy concerns, and they do not directly address the measured inequality gaps. Explicitly redistributive taxes to curtail the incomes of the rich (as opposed to the entirely different issue of how much of a safety net to provide) often have quite negative consequences. To take as a very crude example: In the 1960s, most countries in the West had high marginal tax rates, often in the 90-95% range. (The Beatles’ song “Taxman” had a line, “It’s nineteen for you and one for me,” reflecting their own struggles with taxation.) These have all been abandoned because they had huge negative consequences, such as forcing emigration or damaging productivity, and were often evaded by the wealthiest as capital markets became more fluid. Yet even with lower modern tax rates, many notable celebrities in France, for example, have changed residence to Switzerland to preserve income.

Moreover, I asked about non-material inequalities because I still believe that the focus on monetary resources as the only assets deserving of redistribution is misguided. Many of the greatest inequalities are not monetary and often the financial differences we see are only the residual fallout of these other differences.

Consider that the dominance of bosses and corporations is not just about wealth but power and social control. And the latter, especially when supported or created by legal or social inequities, are often substitutes for income. In much of the world, interlocking directorates and cozy corporate arrangements favoring the well-connected and politically influential are much stronger than in the United States and are often more important than cash payments. These don’t show up in measured inequality. Sometimes this could be as simple as generous allowances for discretionary spending that go untaxed, or special perks like memberships to exclusive golf clubs where the top deals are made (I believe these memberships can run to millions of dollars in value in some countries). Other times it is in anti-competitive measures that limit takeovers or shareholder democracy. In Northern Europe, cartels and price-fixing were either legal or much less severely constrained than in the United States for most of the last century.

I don’t think any income comparison of the United States to many other nations, nor to historical cases like the communist bloc or aristocratic Europe, even begins to capture the ways in which other nations had or have social arrangements that sustain inequality outside the cash nexus. Even in the United States today, much of the worst lobbying comes from different groups seeking a variety of special tax exemptions and industry favors. Increasing the complexity and scope of taxation will only intensify these efforts.

This is why I focused my original remarks on positional goods and inequalities due to individual or social endowment, in the hope of getting us all to Think Outside the Box. For reasons I still don’t understand, many people have an aversion to inequalities that are monetized but are otherwise oblivious to more persistent differences.

Even in the U.S., much of the income inequality of the last 30 years has been driven by the greater returns to education in the workplace. The meritocracy that has been promoted since World War II, however imperfectly, was partly a response to the problem that bright people from certain backgrounds were disadvantaged unfairly. To take only one prominent example, Jews and other minorities were often excluded or disadvantaged in admission to university, especially the Ivies (cf. Jerome Karabel’s The Chosen). Opening up university access and emphasizing academic merit has brought more talent to the fore and contributed to the diversity and success of American science and industry and higher education itself, but it has also led to a growing wedge in the compensation gap between the skilled and unskilled.

This is not just driven by inequalities in talent, training, personality, and intelligence but is exacerbated by differential trends in the family. Assortative mating has meant that highly educated couples in stable relationships produce offspring more likely to do well in school and the workforce than the poor, who are more and more composed of single parents with low human capital. Thus a child of bright, educated parents is often blessed with advantages over the child of a broken home in a bad neighborhood that cannot easily be fixed. It is readily observed that poor immigrant children from intact families do better even in the same schools and the same neighborhoods than their peers who are raised without fathers. Once more, a single-minded focus on financial inequality will miss this deeper and more intractable inequality.

Thus, most of the concerns about injustice or inequality I see in Kenworthy and Anderson do not directly correspond to the measured inequality stats. Do not be surprised if standard attempts to change the distribution are ineffective or do not assuage people’s concerns about their well-being, while nonetheless having negative unintended consequences. Let the focus be on unequal access, or minimal standards of living, or health care, or employment opportunity, not on coefficients of distribution.

What Does Inequality or Redistribution Have to Do with Strengthening the Hand of Low-Wage Workers?

Elizabeth Anderson asks “how likely is it?” that competitive labor markets will provide low-wage workers with a range of decent options that help ensure the value of their exit rights. I believe the answer is “very likely,” and I think I could mount a strong empirical case.

However, at this point, I simply want to flag that the question is indeed an empirical one. The thrust of my initial argument was that we need to take more seriously the questions of which policies and institutions actually help the poor, and to recognize that what helps may have nothing to do with inequality. Anderson seems to concede my main point, that income inequality is orthogonal to the issue of low-wage workers’ bargaining power and the value of their exit rights. Her fall-back position seems to be that whatever secures the worth of these rights will also reduce income inequality. If true, this would suggest that, while inequality is not quite the real issue, it’s somehow inextricably related to the real issue. This is possible, but I think there’s little reason to believe it. Anderson writes:

A generous welfare state (or rather, social democracy) is one way to provide poor with reasonable alternatives, but, as I’ve noted, a rational funding mechanism for this will reduce economic inequality. Alternatively, Poor’s bargaining power can be increased by raising Rich’s cost of exit — say, by unionizing the workforce of Rich’s company, or the workforce as a whole. This will also tend to reduce inequality between Poor and Rich.

I’m afraid all this just begs the relevant questions. A “generous welfare state” could apply to an indefinite number of schemes, any one of which might make the problem either better or worse. And, because the issues are completely logically unrelated, whether a well-functioning redistributive scheme also reduces income inequality is an open question. It is certainly possible that a better-designed set of programs could do more good for less money. And large increases in the generosity of redistributive programs aimed at improving the conditions of the poor could be financed through changes in the overall level and composition of government spending, leaving tax rates and inequality entirely unaffected. (Just imagine what could be done with a third of Navy’s budget!) The effects of unionization, again, depend on the details. Unions indeed improve the bargaining power of some workers. But this often comes at the expense of other workers who may then find it harder get work and face higher prices as consumers. And a higher unemployment rate and higher consumer prices tend to increase income inequality, not decrease it.

If it is wanted, I can bring forth evidence that poorly designed welfare programs have increased economic inequality in the United States by negatively affecting labor-market participation among low-skill workers. And I’d be happy to point to evidence on the typical effects of unionization. Indeed, I’d rather have this kind of evidence-based debate. What really helps? What really works? My argument has been that simply assuming that inequality is the problem does not help. I have not argued against redistribution. What I am arguing is that the demand for more generous progressive redistribution is often little more than a reflex, and that the design, not the generosity, of our redistributive policies is generally the pertinent issue.

I’d like to apologize to Prof. Anderson for dithering on my promised reply to her argument about the effects of inequality on democracy. I’ll come through tomorrow.

One More Round on Conspicuous Consumption

Wilkinson’s latest post argues that the poor who squander their resources on competition over appearances should take more personal responsibility for their spending decisions. In response, I’d like to point out that there is a difference between a policy perspective and a perspective of individual choice. I thought we were talking policy here, and not giving advice to individuals. As modest Midwesterners, I’ll bet Wilkinson and I would give the same advice to people of all classes engaged in competitive consumption: don’t do it — it’s a waste of money and an exercise in empty vanity. Still, from a policy perspective it’s worth noting that collective action problems can (a) make it difficult for individuals to make the best choices and (b) impose high costs on people. Competitive consumption is a kind of collective action problem. Many people find it hard to resist when they see how their peers mete out costs and benefits in this game. And even if, in central regions of this country, this game is not played vigorously, in many other regions — notably, Southern California, Florida, and the urban-suburban-exurban Northeast, it is played with alacrity, and with cruel results. In the end, it’s a negative-sum game, which, as Robert Frank argues, offers opportunities for public policy to intervene with positive outcomes all around. My point about squeezing the differences between the top and bottom of the income distribution is that this can change the dynamics of the recognition game, by making it more difficult for all classes to distinguish themselves through consumption alone, and making it look more foolish as well, since everyone is more vividly aware of what is being sacrificed in trying to keep up appearances. The quest for distinction is not extinguished, but at least it can be channeled into more worthwhile outlets.

Brief correction on income inequality data

John Nye says measured income inequality would not be reduced by the types of redistributive policies that I (and Elizabeth Anderson) have been discussing. He seems to think neither government transfers nor taxes are included in the data used to measure inequality.

It’s true that the Gini coefficients published by the Census Bureau in its annual report on income inequality are based on a measure of income that does not include the Earned Income Tax Credit or noncash transfers such as food stamps and does not subtract tax payments. But each year the Congressional Budget Office (CBO) publishes income data that address these limitations. The information about income inequality trends that I mentioned in a prior post in this exchange is based on the CBO data.

One other small point: John says I recommend raising the top marginal income tax rate, and that this might not do any good because the rich find ways to shelter their income to avoid it being subject to that top marginal rate. Actually I said I favor increasing the effective tax rate on top incomes. That can be done by increasing the marginal rate, or by reducing exemptions and deductions, or by improving enforcement of existing rules, or some combination of these.

Real and Perceived Inequality

John Nye suggests that innovation and productivity improvements tend, paradoxically, to reduce the actual degree of material inequality but to increase perceived material inequality.

Goods such as food, clothing, housing, cars, televisions, and many others come to be produced ever less expensively. As this happens, more people can afford to purchase them. This reduces material inequality. A rich American today may have four houses and seven cars. Yet the difference in material well-being between that person and someone who lives in a modest apartment and has one car is smaller than the difference between their counterparts of a century ago, when the rich person had two homes and one car and the poor person had no car and lived in a hovel with no heating or indoor plumbing.

As this process of material equalization proceeds, according to John, “the goods and services that remain objects of envy and inequality will be those that are most insensitive to improved efficiencies” — positional or status goods such as rare paintings, a spot in a top university, the most desirable housing locations in a city, the corner office, and so on. As a result, “the more we succeed in promoting wider access through growth, the more we will be focused on acquiring and differentiating ourselves on margins that are hard to change or difficult to equalize. Which will inevitably lead to more frustration as we mistake the material manifestation of those inequalities for the underlying differences themselves.”

I find this an interesting and plausible hypothesis about how things might play out at some point in the future. John, though, appears to view it as a useful account of what’s already occurred. Will Wilkinson concurs: “I think John’s story brilliantly accounts for the data,” he says. “Our society has become more economically egalitarian, in one profoundly important sense at least, but it sure doesn’t seem like it.”

I’m skeptical. Consider developments over the past century. The level of income inequality today probably is similar to what it was in 1910 — more inequality between the top 1% and everyone else, less inequality within the bottom 99%. But as I suggested earlier, inequality of material well-being is surely lower today. Do we nevertheless perceive that material inequality is greater today, as John hypothesizes? I have no idea, and I don’t know of any historical data on perceptions that would give us even a semi-educated guess.

What about the past three decades? Income inequality has risen sharply. Consumption inequality may have risen less sharply, but hardly anyone (including Will) contends that it has decreased. Has inequality of material well-being declined? Almost certainly not. If it doesn’t seem as though our society has become more economically egalitarian over the past generation, that’s probably because it hasn’t, rather than because Americans have shifted their focus to status goods.

More Problems with Measured Inequality

I am grateful that Lane Kenworthy has addressed the issues I brought up in my original post because I think they are quite different than the normal discussions of inequality.  He concedes my points, but doesn’t think they are relevant to the last century or at least the last thirty years.

But if he starts to go down that path, he will see that most of our current measures do not take these issues into account at all.

Kenworthy argues that gross measured income inequality is roughly the same as it was 100 years ago.  Yet he also concedes that “inequality of material well-being is surely lower today.”  How can my point therefore be irrelevant?  The gross measured inequalities are utterly misleading.  A world in which food of many kinds is vastly cheaper, better prepared, and less likely to contain harmful bacteria; in which clothing of a high quality is widely available cheaply; and in which people have access to basic medical care such as antibiotics, pain-killers, and basic surgery often unavailable to their ancestors, in addition to electronics like phones, washing machines, radio and television is one in which the true gap between rich and the poor is almost certainly lower (cf. the work of Stanley Lebergott).  Moreover, a simple perusal of the life expectancy statistics will show that lifespans and infant mortality have improved for all groups in the last century but that they have improved more for the worst off than the richest (cf. Robert Fogel).  Can there be any doubt that the ratio of a poor person’s expected lifespan to that of a rich one’s in the United States (all other things held constant) is higher today than it was 30-50 years ago? This is a major reduction in inequality not captured by any income measures whatsoever.

Even standard measures of consumption inequality overestimate inequality gaps because they don’t correct for quality and the utility of the goods that are consumed. [A good may fall so much in price that it is a trivial share of the budget but it is a large share in utility terms.]  That this is true can be noted from the fact that even the so-called absolute measures of poverty do not take fixed, quality-adjusted consumption bundles into account.

Moreover, it requires at the minimum correcting for different prices faced by different classes of people.  The likelihood that this issue is important can be seen by the work of Robert Gordon and others who have documented that when a different cost of living index is calculated for the rich and the poor, the inequality measures of the last few decades seriously flatten out.  Moreover, all of these crude calculations don’t take the very differential regional cost differences into account, nor the cost differences even within the same city or state.  Nor do they combine the information from the consumption measures which indicate that a large fraction of the poor consume far more than their income (even when official transfers are taken into account).  Which is another way of saying that the real income measures for the left tail are simply wrong.  I am confident that more detailed analysis on these lines would flatten measures even further, especially considering the point (noted by Elizabeth Anderson) that we all have diminishing marginal utility of income.

I would also argue against the inference that Anderson and others draw from Robert Frank that reducing positional races necessarily has no cost.  This ignores the incentive effects to work.  There are positive externalities to the rest of the population of the rich wishing to obtain more positional goods.  If everyone with $10 million decided that they had enough and simply stopped working, those of us with much lesser wealth would be worse off.  Of course, there are ways in which positional races can be harmful to all as well, but now that is an empirical question as to how gross remedies like taxation affect the incentives to work and to distort the system (like lobbying for favors).

Finally, I think that the concern with gross measures of inequality is itself positional and partially motivated by the relative performance of the United States when compared to other nations.  If Europe were not perceived as less unequal in incomes and more generous in welfare than the United States, would we observe the same pressures to reduce income inequality as we do?  Especially given the absolute progress that has been made in improving the lot of the poor.

Why Oppose an Increase in Redistribution?

Will Wilkinson and John Nye, in their contributions to this exchange, have expressed explicit or implicit support for government action to reduce discrimination, enhance access to good schooling and health care, improve opportunity for employment, and boost the living standards of the poor. Good!

What both Will and John oppose is an increase in government action, via redistributive tax and/or transfer policies, to reduce income inequality. Why? Neither John nor Will has offered either of the two most commonly voiced objections. One is that taking from the rich and giving to the poor infringes too heavily on freedom or ownership rights. I disagree with that notion, and in any event it’s irrelevant as a practical matter. We do redistribute and will surely continue to do so; the question is whether we should redistribute more when income inequality goes up. The other common objection is that increasing redistribution would reduce economic growth or employment or worsen the government’s fiscal deficit. My research and my reading of others’ suggest to me that a moderate increase in redistribution likely would have little or no adverse impact on growth, employment, or our budget deficit.

Why, then, do Will and John oppose an increase in redistribution? Will says focusing on income inequality and redistribution distracts us from doing other valuable things, such as reducing discrimination and unequal access to good schools. Okay, that’s possible. But policy makers might manage to keep their eye on the latter even as they enhance redistribution. John says an increase in redistribution isn’t needed, because true economic inequality actually has declined. It surely has declined over the past century as a whole, but not in recent decades. Finally, John says an increase in redistribution won’t reduce measured income inequality because the data don’t include transfers or taxes. That’s simply mistaken.

John and Will have emphasized that in their view rising income inequality is not the problem. I agree. I think rising income inequality is neither the only problem nor the most important problem our society faces. But that doesn’t mean it’s not a problem worth worrying about and trying to alleviate.

I don’t believe an increase in income inequality automatically requires an increase in redistribution. But the rise in income inequality in America over the past several decades has been large, a core feature of it has been massive incomes for those at the very top, luck has played a significant role in determining who has and has not benefited, taxing away a larger portion of high incomes would do little if any harm to those households or to the economy, and increasing (certain types of) transfers and services to those in the bottom half or quarter or tenth would improve their lives. For these reasons, I think an increase in redistribution would be a good thing to do.

In an earlier post I said,

The average inflation-adjusted pretax income of the top 1% of American households jumped from about $500,000 in 1979 to about $1,700,000 in 2006. For households in this group, the effective tax rate (taxes paid as a share of pretax income) for federal taxes in 2006 was about 31%. That was down from 37% in 1979 and 34% in the mid-to-late 1990s. I favor pushing that effective tax rate up to around 35% (there are various ways to do this) and using some or all of the additional tax revenue to increase the Earned Income Tax Credit.

Perhaps Will and/or John would support this if we label it “helping the poor” instead of “increasing redistribution” and if we say the motivation for it is unequal opportunity rather than rising income inequality. If so, that’s fine with me.

How Does Inequality Matter?

There are many ways to measure economic inequality — for example, in terms of income, wealth, or consumption, either pre- or post- tax-and-transfer, or with reference to the use of material resources that are not personally owned, such as access to public goods provided by local governments, and command over the economic resources of corporations. One can criticize this or that measure of economic inequality, as John Nye and Will Wilkinson do. One can question whether a particular measure, such as the Gini coefficient of pre- tax-and-transfer income, has intrinsic moral significance in isolation from other factors (I agree with them that it doesn’t). One can question whether a particular measure tracks something of causal relevance to things that matter. What I don’t think is credible is to suppose that no measures of economic inequality have any causal relevance to the forms of social inequality I identified as objectionable from an egalitarian point of view — relations of domination, stigmatization, and state favoritism toward those in economically superior positions.

Obviously this is an empirical question, as Wilkinson properly reminds us. But even he doesn’t exactly deny that in the world today laws are, in effect, bought and sold in a “bidding war” that is won by “interests with more resources.” Surely this is why, as he concedes, “a good deal of the United States’ legal and regulatory regime has been rigged to lock in the advantages of well-financed corporate and other interests.” Instead, he criticizes a particularly crude version of the claim that economic inequality causes state favoritism toward the economically better off. The crude version says that “the rich” constitute a unified class that always votes and lobbies in its own economic self-interest, and that command over economic resources is the only factor that generates political influence. I don’t endorse that view. Rather, I think there are many and varied causal paths from particular subgroups that command superior economic resources (as measured by various criteria) to particular state policies that favor their interests over less economically advantaged groups.

Consider the fact that the extreme economic concentration of the banking sector means that there are institutions that are “too big to fail,” which are able to gamble recklessly with the fortunes of the entire global economy. No matter how left-leaning Obama’s economic advisors are (and they are far from socialist), this economic concentration sets profound constraints on feasible state policies. The result, plain for all to see, is a situation in which a particular subset of “the rich” — the executive officers of our leading private financial institutions — rake in the upside of their gambles with the world’s fortunes, and fob off the costs on everyone else via government bailouts, with virtually no accountability for their actions. Notice that in this case, the proper measure of economic inequality will not focus on the economic resources personally owned by this subset of the rich, but rather on the economic resources they command through their executive positions at the leading banks. In this case, economic inequality is not measured by the Gini coefficient on pre-tax income, but by the concentration of the banking sector.

Nye and Wilkinson seem to suppose that egalitarians are committed to redistributive taxation, directly aimed at reducing income inequality as measured by the Gini coefficient, as their primary policy tool. This is not the case. Reducing concentration in the banking sector by breaking up banks too big to fail is a way of reducing economic inequality, as measured by resources at the command of (rather than owned by) a handful of rich people, relative to everyone else. Promoting labor unions is a way of reducing economic inequality by placing a concentrated chunk of critical economic resources — labor — at the command of workers acting collectively, and thereby reducing the unequal bargaining power of employers and workers. Eliminating class-exclusionary zoning regulations is a way of reducing economic inequality, as measured by access to public goods produced by local governments. And so forth.

There is plenty of room to argue about the costs and benefits of different policies that aim to eliminate domination, stigmatization, and state favoritism by reducing economic inequality as measured in various ways. And I have no quarrel with attempts to block the conversion of economic power into relations of domination, stigmatization, and state favoritism that work by other means besides reducing economic inequality, as measured in one or another way. Yet I think it’s naïve to deny that economic inequality, as measured in different ways, generates (by diverse causal paths) objectionable relations of social inequality. And I think it’s equally naïve to take policies against various sorts of economic inequality off the table, on the supposition that alternative policies that preserve economic inequality but attempt to block its conversion into unequal social relations of domination, stigmatization, and state favoritism will always suffice. It’s an empirical question whether such strategies will suffice, and an empirical question which strategies work most effectively at least cost.

October 2009: Best of the Blogs

Here’s a roundup of some notable blog posts in response to this month’s issue. Each post is longer than what I’ve quoted here — which tries only to give you an idea of where the author is headed. All are worth reading.

Rortybomb writes:

What good is money for? Well, in a liberal society, it’s good for two things: more things and more autonomy. The things part is down — as Will is quick to point, goods at the lower end of the consumption scale are cheaper and better working over the past few decades. So even if many consumers have not seen their incomes rise, they feel richer since the goods they are getting are cheaper. True dat!

What I’m more interested in is the autonomy end of it. What about the ability to leave an abusive partner or job without worrying about health care? Travel, spend more time with your family, feel a sense of financial security, etc? Here people less worried about income inequality would say that consumers are in better position to take advantage of this as well since they are richer.

So what they have a cheaper basket of certain goods. Now though autonomy isn’t commodified and sold on a market, the items that we associate with it, insurance, education, perhaps housing, have all seen skyrocketing prices, an effect that blocks out the first effect. This leaves consumers noticeably poorer than they would be otherwise. You’d need to construct a separate index and see the effects played out; I hope researchers are able to do that.

Modeled Behavior writes:

Economists breathed a collective sigh of relief when we saw data showing that GDP really does make people happier.

However, what if all that’s really going on is that the probability of very bad events is going down. That is, the richer you are the less likely your child is to starve to death, the less likely you are to find yourself with no home, the less likely you are to be trapped in an abusive family. These bad things go down and happiness goes up.

In other words Crusoe is happy when his coconut stockpile increases not because he actually going to eat 50,000 coconuts but because implicitly the probability of any series of bad events leaving him with no coconuts or only rotten coconuts goes down.

Let me be clear about what I think we’re saying here. It’s not simply that there is some reservation level of consumption and as people get further from it they are better off. It is that there are a whole range of bad things that could happen to you and higher income gives you the ability to self-insure against more and more of those things.

This can go pretty far. The flat screen and DVD collection insures you against boredom. The night [on] the town insures you against an unhappy mate. Lots of what we think of as happiness from consumption might simply be freedom from worry. In this sense income is not merely an input to consumption, it is a substitute for consumption. We should demand income in and of itself.

Peter Twieg of The Distributed Republic writes,

Elizabeth Anderson uses the phrase “spiteful competition” no less than four times. Does anyone have any idea what exactly this is supposed to mean? I could offer some guesses, but let’s see what she apparently believes about this concept:

- Conspicuous consumption is a form of spiteful competition

- Not all forms of status competition can be classified as spiteful competition

I’m at something of a loss here. Pure status competition in general is usually seen as a zero or negative-sum activity, and if displacing someone else on the social ladder isn’t spiteful, what is? Does it have to be accompanied by an upturned nose and catty banter in order to qualify? This strikes me as an important source of confusion in Anderson’s overall argument: What delineates “good” status competition from “bad” status competition? What forms of inequalities give rise to good versus bad forms of status competition?

It’s unclear whether Anderson actually believes that reducing income inequality would actually reduce status competition instead of just causing it to be expressed along some other dimensions of identity. Observe the trendy prevalence of food and environmental snobbery among certain parts of the American population – can anyone honestly say that these causes have not become broad cultural movements used in order to create “spiteful” hierarchies of social enlightenment? Anderson appeals to the lack of conspicuous consumption in Scandanavian countries, but does she believe that the fact that these countries have much greater status stratification by job title is completely coincidental? Or is this okay because, for whatever reasons, this sort of stratification isn’t done “spitefully”?