clock menu more-arrow no yes

Everything you need to know about income inequality

What is income inequality?

Broadly speaking, income inequality refers to the fact that different people earn different amounts of money. The wider those earnings are dispersed, the more unequal they are. But that intuitive concept of dispersal can be defined in several different ways. Indeed, income itself is a somewhat ambiguous idea that can be defined in different ways.

All that said, in the contemporary United States income inequality has been increasing for several decades by essentially any measure. Similar trends are observed in most other rich countries. At the same time, on a global scale inequality is probably declining.

Inequality has become a hot topic in American politics in recent years. The Occupy Wall Street movement's slogans about the 99 percent versus the 1 percent have stirred an influential current in politics. In a December 2013 speech Barack Obama called inequality the "defining challenge of our time." Conservative politicians overwhelmingly disagree, and focus groups seem to indicate that the public prefers talk about "opportunity" to explicit discussion of inequality. Still, the trend toward more inequality shows no sign of halting and left-of-center figures are likely to keep it on the policy agenda.

How do you measure inequality?

Inequality can be defined or measured in a number of different ways.

One traditional approach was to compare the income of a relatively broad swath of affluent people — the top ten percent of the income distribution (the top decile) or the top twenty percent (the top quintile) — to the national median or average. One big advantage of this approach is that the relevant data is readily available from the Census Bureau and other survey-based sources. A major downside is that this method doesn't tell you much of anything about the earnings of the very highest earners — people in the top 1 percent, for example.

A newer line of research pioneered by Emmanuel Saez, Thomas Piketty, and their collaborators at the World Top Incomes Database has been to use tax records to focus on the incomes of the very top of the distribution. That lets you understand the top 1 percent, the top 0.1 percent, and even the top 0.001 percent. This work has been the basis of much subsequent discussion about the 99 percent versus the 1 percent but the even finer slices are interesting, too.

It is also at times interesting to look at the gap between the poor (say the bottom 10 percent) and the median household. Metrics that define poverty in relative terms tend to, in effect, look at this kind of inequality. So discussions of the living standards of the poor are normally framed in terms of poverty rather than inequality.

Last but by no means least, there is a widely used summary method of calculating inequality that is known as the gini coefficient. A gini coefficient of 0 corresponds to precise equality while a gini coefficient of 1 corresponds to a state of total inequality.

What is a gini coefficient?

The gini coefficient is the most widely used single-summary number for judging the level of inequality in a particular country or region. Unfortunately, it does not lend itself to a particularly obvious intuitive interpretation.

The way it works is to start with a Lorenz curve:


The horizontal axis on this chart represents cumulative shares of the population. The vertical axis is cumulative shares of income. Trace the curve to the 50 percent point, in other words, and you'll see what share of national income the bottom 50 percent of the income distribution earns. Trace the curve to the 75 percent point and you'll see what share the bottom three quarters earn.

In a perfectly equal society, the bottom X percent would earn X percent of the income for any value of X and the Lorenz curve would be a perfect 45 degree angle. In unequal societies, the bottom X percent always earn less than X percent of the income. Except that the bottom 100 percent by definition must earn 100 percent of the income. So an unequal society is always a society whose Lorenz curve swoops below the 45 degree line and then converges with it at the very end.

The 45 degree angle divides a square in half. The Lorenz curve divides that half into two sections — A and B. The gini coefficient equals A / (A+B).

The downside to using a gini coefficient to describe a society is that (as you can see above) it's an incredibly abstract idea that's difficult to verbally describe. The advantage to using a gini coefficent is that in principle it summarizes all the information about the distribution of income and thus facilitates easy comparisons.

Is inequality bad?

There is considerable disagreement about this point.

One prominent argument in political philosophy, found in John Rawls' book A Theory of Justice, holds that inequalities in economic and social status can be justified to the extent that they serve the interests of the least-fortunate class in society but not otherwise. In other words, if society of impoverished subsistence farmers finds that the only way to broadly raise living standards is to industrialize and create a small class of rich factory owners, that's okay. Or if paying doctors substantially more than the average person's salary is the way to induce people to master the craft of medicine and cure the sick, that's okay too. Exactly how much inequality this approach (dubbed "the difference principle" by Rawls) would countenance in practice is difficult to say.

A different criticism of inequality appeals to an idea known as the declining marginal utility of money. In other words the exact same amount of money means different things to the real living standards of different people. A modest financial loss could mean foregone food for a poor person, foregone preventive medical care for someone further up the economic food chain, a foregone vacation for someone more prosperous than that, a somewhat-less-fancy vacation for someone even more prosperous than that, and it would be completely imperceptible to a genuinely rich person. In that view, inequality-reducing redistribution could increase overall human well-being.

Still on both of these views what's bad about inequality is that it's a sign of potentially foregone opportunity to raise the absolute living standards of the less-fortunate. There are also various efforts to demonstrate that inequality per se is a cause of problems.

Many of these ideas are summed up in The Spirit Level by Kate Pickett and Richard Wilkinson, which purports to show that inequality as such drives a number of social ills including low life expectancy, obesity, and poor educational outcomes. Another line of research claims to show that inequality is associated with a lack of social mobility. Last, there is a long tradition of argument that massive levels of economic inequality subvert democratic politics by concentrating excessive political influence in the hands of economic elites.

Counterposed to all of this is a long and broadly libertarian line of argumentation that there's nothing wrong with some people becoming extraordinarily rich if they happen to provide products or services that are broadly in demand.

How economically unequal is the United States?

By international standards, the United States is a very unequal country. Here's our gini coefficient compared to the other major rich economies:


Just three countries — Chile, Mexico, and Turkey — considered developed have higher inequality than the United States. Compared to our traditional peer group of European countries, Canada, Australia, and Japan, we are the most unequal.

The contemporary United States is also unequal by historical terms. Here's a look at the top ten percent's share of income over time:


One nuance to international comparisons is that the United States is considerably larger than other rich countries. If you look at inequality across the entire European Union rather than within a single European country, the EU as a whole is less equal than almost any particular European state. Conversely, if you consider the United States separately the vast majority of states are more equal than the country as a whole.

Is inequality rising because the rich are getting richer or because the poor are getting poorer?

Mostly because the rich are getting richer. A vast swath of American households have been in financial distress since the economic crisis of 2008, but taking the long view incomes have tended to rise across the board. They've simply risen dramatically faster for the highest-income households than for the rest of the population:

As you can see above, the incomes of the very rich are also considerably more unstable than the incomes of the bottom 99 percent. That's because the richer you are the more important the ups and downs of the stock market are to your economic life. Consequently, it's not unusual for a short-term period to occur during which the incomes of the very rich fall quite dramatically. It's important to distinguish this from the longer-term trend.

Why is inequality rising?

There is some disagreement about this and also some nuance as to exactly which aspect of inequality we're talking about and what kind. Some of the most prominent accounts:

  • Skill-biased technological change. Technological improvements raise incomes, but they do so unevenly. Rewards disproportionately go to highly educated workers. On this view, rising inequality reflects slowing educational progress and better schooling is the key solution. This has been the traditionally dominant view in economics, but it doesn't explain the specific rise of the top 1 percent very well.
  • Immigration. Importing low-skilled workers to do low-paid jobs tends to raise measured income inequality. David Card estimates immigration to be the cause of about 5 percent of the total rise in inequality.
  • Decline of labor unions. Labor unions reduce inequality both by raising wages at the low end and constraining them at the high end. Bruce Western and Jake Rosenfeld estimate that the decline of labor unions as a force in the American economy is responsible for 20 to 33 percent of the overall rise in inequality.
  • Trade. Growing international trade with lower-wage countries such as China seems to have reduced the wage share of overall national income, boosting the incomes of people with large stock holdings.
  • Superstar effects. Because the world is bigger and richer in 2014 than it was in 1964, being a "star" performer — the most popular athlete, author, or singer — is more lucrative than it used to be.
  • Executives and Wall Street. Increased incomes for CEOs and financial sector professionals account for 58 percent of the top 1 percent of the income distribution and 67 percent of the top 0.1 percent. So the specific dynamics of compensation in those areas are wielding a big influence.
  • Minimum wage. David Autor, Alan Manning, and Christopher Smith find that about a third (or possibly as much as half) of the growth in inequality between the median and the bottom ten percent is due to the declining real value of the minimum wage. This is different from the issues about the top end pulling away that normally dominate political discussions of inequality, but it's a noteworthy finding nonetheless.
  • The fundamental nature of capitalism. Thomas Piketty has made waves lately with his new book Capital in the 21st Century, which argues that very high levels of inequality are the natural state of market economies. In his view, it's the economic equality of the mid-twentieth century that needs explaining, not today's inequality.

How does inequality relate to opportunity and upward mobility?

One common political response to rising inequality has been to argue that what really matters is economic opportunity: do people have a chance to rise to the top? Alan Krueger, the then-chair of the White House Council on Economic Advisors challenged this notion in 2012 with what he called "The Great Gatsby Curve." It shows a strong correlation between national income inequality and intergenerational mobility:

The_great_gatsby_curveOn the other hand, Raj Chetty, Nathaniel Hendren, Patrick Kline, Emmanuel Saez, and Nicholas Turner have published research looking at trends over time and finding that there's been no decline in mobility in the United States despite the increase in inequality.

UC Davis economic historian Gregory Clark has challenged the link in an even more profound way by arguing that properly measured, social mobility is extremely low everywhere and always has been, even in places like Sweden.

Sweden hasn't had a feudal economy or political system in quite some time, but even in the early 21st century people with aristocratic surnames earn substantially higher incomes than people with the common peasant surname Andersson. On this view, inequality may not hamper upward mobility but upward mobility is also so scarce that it's hardly an antidote to worries about inequality.

What can the government do to reduce inequality?

The most straightforward thing the government could do to reduce income inequality would be to tax the rich more heavily and give additional money to the poor. The United States has some of the lowest taxes of any rich country so there's certainly room to do more.

Laws that were friendlier to labor union organizing and to the activities of existing labor unions would also likely reduce inequality. So would altering the mix of immigrants to the United States to include a higher proportion of skilled workers. So would improvements in educational attainment. But these kinds of measures speak more to the gap between the top 10 or 20 percent and the rest, not to the gap between the top 1 or 0.1 percent and the average.

For incomes at the very top, there is taxation and then there is the idea of directly targeting the sources of high-income individuals' earnings. That could mean stricter regulation of the financial sector to reduce earnings on Wall Street, or weaker protections for the holders of copyrights to reduce the earnings of superstars.

How does income inequality relate to wealth inequality?

Income is the flow of money that you receive. For most people it's your wages or salary. For some people, there might be dividends, interest payments, or rent thrown in. Wealth is how many financial assets you have stockpiled. The equity in your home, your bank account, your retirement account, or other stocks and bonds you may have accumulated.

As this chart from Emanuel Saez and Gabriel Zucman shows, wealth is distributed very unequally. Much more unequally than income.

There are two main reasons for this.

One is that the bottom quarter or so of the population has zero or negative net wealth, due to student loans, underwater mortgages, credit card debt, auto loans, or other debt instruments. Nobody has negative income and very few people have zero income when government benefits are factored in.

The other is that wealth leads to income. A billionaire who owns tons of stock is going to earn substantial dividends from his stock holdings. Some of that income will be saved, and turn into further wealth. This tends to put the wealth of the wealthiest on an upward trajectory unless something like a war or a depression or a massive political intervention interferes.

What's happening to inequality elsewhere in the world?

Broadly speaking, income inequality appears to be on the rise in almost all developed countries. That was the conclusion of a recent report on the subject by the Organization Economics Cooperation and Development (OECD).

InequalityoecdThe absolute level of inequality continues to vary greatly from country to country. But the trend is broadly similar. So explanations of why inequality is on the rise that lean on idiosyncratic aspects of American politics tend to fall short.

What's happening to global income inequality?

Even as inequality rises in most rich countries there is evidence that global inequality is declining somewhat from a very high level. World Bank economist Branko Milanovic made the following comparison of the global gini coefficient to the gini coefficients of a few specific countries:

Global_inequality The driving force here is rapid economic growth in China and a few other developing countries. The new "global middle class" that's emerging is poor compared to the middle class in the United States or other rich countries, but it's large enough and its income is growing fast enough that global inequality is falling by this measure.

On the other hand, the richest of the world's rich — the billionaires — are getting richer at a rapid pace according to Forbes' annual tally:

Billionaires The number of billionaires is growing, but the wealth under their control is growing even faster. In other words, what you make of global trends depends on what you care about.

Why do some economists say the increase in inequality has been overstated?

While income inequality has been a growing subject of public discussion and most authorities take it for granted at this point that incomes in the United States have grown very unequal, there is some dispute about this. Richard Burkhauser, a Cornell University economist, and Scott Winship, a policy analyst at the Manhattan Institute, have been the leading proponents of the view that the new conventional wisdom overstates the increase in inequality.

The dispute hinged on a variety of conceptual issues, but also on the existence of different sources of data about income. Let's start with the data, and then get into the conceptual problems.

Census vs IRS

The biggest, broadest difference between inequality measures you will see is that some economists (following the work of Thomas Piketty and Emanuel Saez) look at tax return data from the Internal Revenue Service while others rely on the Current Population Survey (CPS) data from the Census Bureau.

The big advantage of the CPS is that it lets you find out information about non-cash compensation — mostly health insurance — and government benefits programs, both of which are important to middle- and working-class economic welfare.

The big disadvantage of the CPS is that because it's based on broad statistical counting, it doesn't give you insight into the top 1 percent, top 0.1 percent, or top 0.01 percent of the population. And since a very large share of the top 5 percent's overall income is in the hands of a very small elite (the top 0.1 percent, say) the CPS ends up undercounting the overall high-end cash income.

Household size

The average size of the American household has shrunk over time, meaning that income-per-household-member for the middle class has grown faster than income-per-household. Some people feel that "income stagnation" claims should be adjusted accordingly (which gives you a better sense of average living standards) while others do not (which gives you a better sense of the state of the labor market).

Taxes and transfers

Since 1979, the tax burden on poor and middle class families has shrunk. Social welfare expenditures (primarily health care programs like Medicaid and the Affordable Care Act) have risen, and a larger share of the population is receiving Social Security and Medicare benefits. All this means that if consider middle class incomes after taxes and transfers, you get a considerably rosier picture than if you consider pre-tax income. For similar reasons, IRS data will tend to suggest that retirees are very poor by neglecting the considerable value of government programs that lift the elderly out of poverty.

Somewhat ironically, the people who insist on including tax and transfer data in their inequality metrics tend to be associated with the political right while the people who argue in favor of progressive taxation and a generous welfare state tend to want to leave this stuff out of their inequality metrics.

Capital gains

Another conceptual issue relates to the treatment of capital gains, in other words profits earned on investments. Including IRS data on capital gains adds a lot to the incomes at the high end, because very rich people own the bulk of stock. At the same time, since profits from the sale of owner occupied housing are not normally taxed, middle class households' main form of capital gains income generally doesn't show up in this dataset.

Another issue with IRS capital gains data is that investment profits are taxed when they are realized, i.e. you pay tax on stock market profits when you sell shares. Combine that with the stock market's tendency to fluctuate up and down and this creates a very unstable income series. That, in turn, means that estimates of high-end income gains over a given period can be highly sensitive to your use of start and end dates. The realization issue also makes a difference for middle class incomes. Middle class workers tend to hold stock in tax-advantaged accounts like 401(k)s and IRAs. The value of these portfolios builds up, tax free, over time. The gains are then realized (and taxed) after retirement, when labor income has dropped to zero.

Health insurance

Many people receive deeply discounted health insurance as part of their compensation package. As the cost of health care has risen over the decades, so has the value of these benefits. Putting a precise number on their value is conceptually challenging, but it is clear that $0 is not the right number. Rich people tend to have better insurance plans than middle class people, but only to a moderate extent. Including the value of health insurance benefits makes middle class income growth look more robust and the growth in inequality smaller.

The bottom line

For all of these issues, there is merit to both ways of looking at the issue. It's important to ask yourself what, specifically, you are interested in and for authors to be clear about what data they are referring to when making polemical points. The IRS income data is the only way to measure the growth of high-end incomes, which makes it indispensable for understanding the economic elite. But it's a rather poor guide to middle class living standards.

The easy answer would be to say to look at the IRS data for a glance at the elite, and turn to different Census-based measures to understand the fate of the middle class. The problem is that one of the main things people would like to understand is how soaring incomes at the top have impacted living standards in the middle. There's simply no entirely satisfactory way of doing this. Census-based measures will miss the soaring elite incomes, while IRS-based measures will leave out health insurance, treat home sales and retirement accounts oddly, and most of all miss the impact of government programs.

Where can I learn more about inequality?

The World Top Incomes Database is the best source of raw data about income and wealth inequality around. The researchers involved in the project also regularly publish papers based on the data that are often enlightening. One of them, Thomas Piketty, has recently published a book, Capital in the 21st Century, that offers one of the most thorough treatments of the subject.

Branko Milanovich's 2011 book The Haves and the Have Nots: A Brief and Idiosyncratic History of Global Inequality is a useful discussion of the generally neglected worldwide perspective on inequality.


The tricky trans politics of FX’s Y: The Last Man


Murders are spiking. Police should be part of the solution.

The Goods

$5 jewelry and an MLM conference gone wrong

View all stories in The Latest