3

TPM’s Deep, Deep Dive into the Economics of Inequality

By Jared Bernstein


Introduction: Lots of data, all pointing toward more inequality

nderstanding the growth of economic inequality in the United States is both simple and complex. The simple part, which is also the most important part, is that by virtually all the relevant metrics, types of income, and data sets, inequality has significantly increased since the 1970s. The complex part is that to understand this increasing trend in some depth, one needs to undertake some degree of immersion in all those different metrics, data sets, and income types. That means we’ll need to look at the growth of inequality by wages, incomes, and wealth; we’ll have to understand the strengths and limitations of the various data sets.

Though not all readers will share the same tolerance for digging around in the data weeds, the exposition can be relatively quick and painless, with lots of nice pictures. When you’re trying to understand the evolution of a big, important phenomenon like inequality, think of yourself as a navigator getting a fix on her position by looking at various points in the firmament, which in this branch of economics means surveys and tax records on incomes, wages, and wealth. Since no one survey is definitive, we must fix our position by looking at all of them.

Fortunately, at least in an analytic sense, when it comes to inequality, all the surveys point in the same direction: toward greater economic distance between people and households in their economic outcomes. Though there may be some economists and policy makers who deny that inequality has increased—you can always find (or pay) someone to take the other side of any position in economics—as a very active participant in this debate, I can tell you that we don’t hear much from them. It’s easier to find a denier of global warming than of rising inequality.

Of course, there is a robust debate as to the causes of inequality, whether it matters, and what, if anything, should be done about it. In fact, as we’ve largely agreed on the trends, these, in my view, are the compelling questions regarding inequality in America.

But one must eat one’s spinach before dessert, so let’s turn first to a rigorous look at the evidence and then discuss what it all means.

The evidence: The growth of income, wage, and wealth inequality

Starting in the mid-1940s, the Census Bureau began to collect data on family incomes at various percentiles on the income scale. As Figure 1 shows, inflation-adjusted incomes at the low- and middle end of the scale rose at the same pace as high incomes between 1949 and 1979 (actual income levels are indexed to 100 in 1979 so I can plot them on the same scale), when the real income of each group roughly doubled. After that point, however, incomes diverged. Families at the 95th percentile rose by almost 50 percent between 1979 and 2014, but families at the median saw only a 12 percent gain and income for the low-income families at the 20th percentile actually fell slightly during that 35-year time period.


Census Income Data (Figure 1)

Census Income Data (Figure 1)

Having looked at the previous figure, you might conclude that the case is closed. Incomes used to grow together at the different percentiles, but that’s no longer the case. Ergo, since the late 1970s, inequality has increased. That’s true, but it’s not the whole story, in part because the Census family income data are incomplete, missing some important income sources.

For one thing, the Census leaves out the value of some government transfers, like the cash value of nutritional support. After all, if the government gives you $50 to spend on food, that frees up that same amount for you to spend elsewhere or save (technically, such a transfer is “fungible”). The Census data are also pretax, and since taxes—at least at the federal level—are progressive, we’d expect the post-tax income distribution to be less unequal than the pretax distribution. Consider, for example, the Earned Income Credit (EIC). We now redistribute around $60 billion through the federal tax code in the form of wage subsidies to low-income workers. This transfer reduces income inequality.

There are also omissions in the Census data that push the other way. Asset values—like those of stocks, bonds, and home equity that are disproportionately held by those at the top of the income scale—are not in the Census data (it includes dividend payouts, but not equity values; it excludes capital gains). Adding these into the mix should give rise to higher inequality, especially over periods when asset values sharply appreciate.

As you might imagine, adding in these forms of income and wealth create serious measurement challenges. How should we value medical benefits like Medicaid, Medicare, and Obamacare premium subsidies, especially given that the costs of these benefits are uniquely inflated in the American health care system relative to that of other advanced economies? At the low end of the income scale, we don’t want to make a poor family on Medicaid look better off just because American doctors have really high salaries.

At the high end, asset values spin off income when they are sold—capital gains—which should be included. But as you’ll see in a moment, capital gains jump around noisily based on market volatility, such that you can get a very different read on the extent of inequality depending on whether you’re taking its measure in an up or down market. Surely the structural forces that have been driving inequality for decades do not flip from year to year. And why stop with realized gains? To do so creates a bias against the rising or falling wealth of asset holders who happen not to sell their assets.

The Congressional Budget Office (CBO) generates an important household income data series that tries to tackle some of these gnarly questions. They add in the value of government transfers, both cash and non-cash (the data in Figure 1 include only cash benefits), they simulate the impact of the federal tax system (but not the state or local), they include employer payments for health insurance premiums, realized capital gains (but not unrealized gains), and more.


"It’s easier to find a denier of global warming than of rising inequality."


You may be thinking “aha!”—here, by dint of filling in all the omissions from the Census data, we have what we need to provide the final word on income inequality.

But, no. While the CBO series has many advantages, particularly regarding the actual purchasing power of household incomes at various parts of the income scale, it also has some biases. Probably the most important is the extent to which low incomes are biased up in this series due to assigning the market value of health care benefits. The bias arises out of the fact that health care costs in America are much higher than those of other countries on a per capita basis. Moreover, the cost differential is not due to quality but is generally seen as arising from unique inefficiencies within our system: high drug costs, physicians’ salaries, hospital stays, and so on. To add these costs at market value to poor households with Medicaid, for example, has the effect of making them look better off, not because they are but because some drug or X-ray costs much more here than everywhere else.

Of course, government-provided health insurance is worth something, and surely a non-trivial amount, to those who have it. But we really have no way of knowing how much. CBO’s method is a high upper bound. In fact, they also show the results of using a more fungible approach, which cuts the growth of income among low-income households by almost half.

At any rate, with that caveat, the inclusion of more income sources shows faster real income growth among households at all income levels than does the Census data. This is an important finding in the sense that many of these households, both low- and middle-income, are considerably better off after taxes and transfers than the Census data reveal. But more importantly from our perspective, the growth of inequality since the late 1970s, when the CBO data begin, mirrors that shown in Figure 1 above.

As Figure 2 below shows, the bottom 99 percent of Americans saw real gains in income under this expanded definition between 1979 and 2011. Gains in the top 1 percent, however, far outstripped the rest, as their incomes rose by 200 percent, a gain five times that of the middle class.


CBO Income Data, 1979-2011 (Figure 2)

CBO Income Data, 1979-2011 (Figure 2)

Another odd characteristic of the CBO data is the big swings in the income of the top 1% toward the end of the series, making the series look like a jagged mountain range. This pattern is due to large losses and subsequent gains in the value of financial securities held disproportionately by the wealthiest households. The pattern implies considerable volatility in income inequality, such that measuring trends between different endpoints can provide very different answers about the extent to which inequality has grown. But this is a result of financial bubbles bursting and leading to sharp stock market swings, not the fundamental, structural forces driving inequality’s growth, a topic I turn to below.

As noted, the CBO data also net out the impact of federal taxes. Figure 3 shows the share of income going to each income group before and after taxes (note that CBO’s before-tax incomes include government transfers, which, as noted above, are significantly equalizing). While the federal tax system does move some economic resources from the top of the scale down to the middle and bottom, the shape of the income distribution remains largely unchanged.


CBO Income Shares, 2011 (Figure 3)

"CBO Income Shares, 2011 (Figure 3)

In fact, the growth of inequality is also pretty much the same under either concept, as shown in Figure 4, which plots the changes in shares of income by income group, before and after federal taxes. If we look at pre-tax, post-transfer income, the poor and middle class have lost ground relative to the wealthy since 1979. For example, 1.9 percent of pre-tax income shifted from the middle to the top. If we then shift to post-tax income, these same groups have still lost ground - not quite as much, but not enough to alter the general picture.

These facts have two implications. First, it’s hard to argue that the tax system is responsible for higher inequality as income shares grew more unequal through “market outcomes,” i.e., before factoring in the impact of taxes. Second, since the growth in inequality is largely a pre-tax phenomenon (though as I’ll note later, the two sides are related), it is not likely to be reversed through tax policy alone.


Change in CBO Income Shares, 1979-2011 (Figure 4)

Change in CBO Income Shares, 1979-2011 (Figure 4)

The growth of market-based (or pre-tax) income inequality is particularly evident in a widely cited data source, that of Thomas Piketty and Emmanuel Saez (PS), based on tax records going back to the early 1900s. Along with the long time series which provides critical historical insights, the PS data have the advantage of showing the extent of income concentration at the very top of the scale: not just for the top 1%, but also for the top 0.5%, 0.1%, and 0.01%. The key finding in that analysis is that the further you go up this narrow part of the income scale, the more you observe income concentration. Figure 5 shows that in 2014, the top 1 percent of households help over 21 percent of all income, and you can see the breakdowns within the one percent in the stacked bar. The 16,500 families in the top 0.01 percent, for example, held 5 percent of all income in 2014 and their average income was over $29 million.


Breakdown of Top 1 Percent Income Shares (Figure 5)

Breakdown of Top 1 Percent Income Shares (Figure 5)

The historical series (Figure 6) shows that the share of income accruing to the top 1 percent has in recent years returned to levels not seen since the late 1920s. The 21.4 percent held by the top 1 percent in 2014 was almost three times the share held by the richest 1 percent during the 1960s.


Top 1 Percent Income Shares (Figure 6)

Top 1 Percent Income Shares (Figure 6)

The two main shortcomings of the PS data are the exclusion of non-market income sources and the fact that, since they derive their data from tax records, non-filers are absent. PS try to adjust for this second problem by adding imputed non-filers and their income back into the data. Still, while the market incomes of the rich and super-rich are well-represented in these data, some important components, including many government benefits, are missing.

Sticking with market incomes, since most working-age households depend on their paychecks (as opposed to their stock portfolios), the dispersion of wages is a fundamental source of income inequality. The Economic Policy Institute has long measured this aspect of the problem using two metrics: wages by decile (and gender) and median compensation relative to productivity growth, an increasingly widely cited measure of inequality’s growth.

Figures 7 and 8 plot wage growth by wage decile and gender for low-, middle-, and high-wage (corresponding to the 10th; 50th, or median; and 95th percentile) workers. Once again, real wage values are indexed to 100 in 1979 so the trends in different wage levels can be plotted on the same graphs. For both genders, the now-familiar pattern of fanning out is evident as high-wage earners pull away from the pack. But there are interesting gender differences. The trajectories for both middle- and low-wage men have been similarly stagnant at best; middle-wage women experienced some degree of real wage growth in the 1980s and more in the 1990s, but they haven’t gained much ground since then.


Male Wages by Decile (From EPI) (Figure 7)

Male Wages by Decile (From EPI) (Figure 7)


Female Wages by Decile (From EPI) (Figure 8)

Female Wages by Decile (From EPI) (Figure 8)

For part of the time period shown, these divergent wage trends were driven by growing differentials in wages by education level, aka “education wage premiums.” Economists argued that unmet skill demands by employers were driving the pay of college-educated workers up relative to those with less-than-college educations. Today, that premium remains high, but it hasn’t grown much for over a decade, and even some college graduates have experienced real wage erosion.

Figure 9 has arguably become the most important inequality graph, especially in this political season, as it is frequently referenced by various campaigns as a summary picture of the problem. It plots the real compensation growth—wages plus benefits—of a middle-wage worker against productivity growth, or output per hour. Students of economics are typically taught that it is through growing productivity that we improve our living standards. If we’re more efficiently turning inputs into the goods and services we want and need, we can increase our consumption without more hours of work, or increase our leisure (work less) and consume similar amounts.


Productivity and Pay Began to Diverge in the 1970s (Figure 9)

Productivity and Pay Began to Diverge in the 1970s (Figure 9)

But focus for a moment on the “we” in that last sentence. If the benefits of productivity growth are flowing to all income and wage groups, as you can see they were in the 1950s and 60s (you also see this for the income groups in Figure 1), then “we” is all of us. But as the forces of inequality took hold, they created a wedge between productivity growth and the pay of middle- and low-wage workers. In this scenario, “we” becomes a much narrower group of beneficiaries.

One reason this figure is so resonant is that workers who are no longer benefitting from productivity growth as much as earlier cohorts are still contributing something to it. Imagine the economy as a bakery owned by a corporate entity where workers and owners are paid in slices of what they produce. The bakers have consistently generated more bread and cakes and pies per unit of flour and sugar, and for decades as both they and their corporate parent benefitted from increased output per hour, both sides enjoyed bigger slices. But as inequality took hold, the bakers kept creating more pies and cakes but the size of their slices stopped growing. The owners of the bakery, on the other hand, increasing cut bigger slices for themselves.

What changed such that the benefits of growth become more narrowly shared? I’ll get to that fundamental question in a moment, but first, one last dimension of inequality, one that’s becoming increasingly important.

When Billie Holiday sang, “them that got shall get,” she might well have been thinking about the concentration of wealth. As you might imagine, wealth is far more concentrated then income. Wealth is typically measured in this research as “net worth”—assets minus liabilities—and one of the surveys of net worth is run by the Federal Reserve. In their most recent survey, from 2013, they found that when they rank families by income, the top 10 percent holds about 48 percent of all income. But when they rank families by net worth, the top 10 percent holds 75 percent of all wealth. For the top 3 percent, the comparable shares are 31 percent for income and 54 percent for wealth.

What about trends? The figure below shows the share of wealth going to the wealthiest 1 percent of families since the early 1900s. The pattern is familiar, much like the one for income shown in Figure 6. But compare the levels of the y-axes of the two graphs and you’ll see that the top 1 percent’s wealth share is roughly about double that of “their” income share (the quotes on “their” are meant to emphasize that these aren’t the same people—some of the top 1 percent by income will be in the top 1 percent by wealth but not all of them).


Top 1 Percent Wealth Share (Figure 10)

Top 1 Percent Wealth Share (Figure 10)

A final figure in this section (figure 11) combines a number of the concepts that we’ve discussed, and introduces the inequality of debt, in this case student debt, which is an increasingly evident dimension of the problem.


Ratio of Mean Education Debt to Mean Income (for families with education debt) by New Worth Group (Figure 11)

Ratio of Mean Education Debt to Mean Income (for families with education debt) by New Worth Group (Figure 11)

The figure shows the ratio of families’ mean education debt to mean income in different sections of the net worth distribution in 1995 and 2013. Families in the bottom half of that distribution owed 26 cents for every dollar they made in income in 1995, but by 2013, that ratio had more than doubled – they held 58 cents in college debt for every dollar of income. The ratio also more than doubled for the next 45 percent of Americans; only in the top 5 percent did it remain unchanged. Such differential patterns of the burden of debt associated with higher education have important implications for the inequality of opportunity, a point I return to below.

Like I said, there’s a lot to wrap your head around here in terms of different definitions, surveys, time periods, and results. But there’s simply no denying that they all point in the same direction. For some set of reasons, the distribution of economic resources has been growing increasingly unequal since the 1970s. What’s in that set, why does it matter, and what can be done about it?

What’s driving the growth of inequality?

The most common explanation for the trends shown above—a phrase uttered by politicians, policy makers, and economists—is “technology and globalization.” But this explanation is opaque at best; at worst, it is both misleading and too benign-sounding.

“Technology” captures the idea that employers’ demands for specific skill sets that complement technology in the workplace—think IT—are pushing up the wages of highly educated workers relative to everyone else. There’s obviously something to this explanation. Those with college degrees consistently make more than those without, and for years, that differential was rising.

However, demand for tech skills was only one reason for the rise in the pay of more highly educated workers, and one thing I can assure you: if we’re going to understand the factors driving inequality, we’re going to need most of the blackboard. This phenomenon can’t possibly be described by one or two developments.

For example, the college wage premium (the college wage compared to the high-school wage) grew quickly in the 1980s, but at the same time, the real value of the minimum wage fell by 30 percent. Union coverage fell by 8 percentage points (from 27 percent to 19 percent), a rate of decline that was more than twice that of the 1990s or 2000s. Between 1979 and 1989, the labor market was at full employment (meaning the number of job seekers tightly matched the number of jobs on offer) only 30 percent of the time, a situation that is particularly disadvantageous, and thus dis-equalizing, for lower wage workers.

In other words, employers’ skill demands were and are but one factor in play. While this may sound obvious, I cannot overemphasize its importance. One reason it has been so hard to craft and implement policies to fight back against rising inequality is because of the misdiagnosis which solely emphasizes the role of unmet skill demands. This emphasis leads to “more education” as the fix-all response. Of course, it is a valid and much needed response, especially for those facing barriers to quality education (starting with pre-school). But it cannot be seen as the sole solution.

"To lower today’s levels of inequality, policy would need to accomplish two broad goals: raise the bargaining power of the American worker and lower the political clout of the wealthy elite."

Globalization is also very much in the mix of explanations, though here, too, refinement is necessary. At one level, increased trade with the rest of the world increases the implicit supply of labor, and thus, absent a commensurate increase in demand, puts downward pressure on wage growth. But the key question regarding inequality is: whose wage growth is negatively affected by globalization? The fact that much of our trade—and of great importance, imbalanced trade—is with countries with larger shares of workers with low levels of education implies the pressure will be on non-college educated workers here. This is probably as good a time as any to point out that such workers—those without at least a four-year college degree—still form a majority of our workforce (about two thirds).

This may sound like an unavoidable tradeoff: for many good reasons, we want more globalization. It increases supply chains, lowers prices, and gives those in developing countries a chance to raise their own living standards. But the problem for those American workers on the wrong side of the inequality divide is not globalization per se. It’s imbalanced trade, meaning trade deficits, which is one reason why labor demand hasn’t kept up with the implicit increase in trade-induced labor supply. It’s also one reason for that absence-of-full-employment statistic noted above regarding the 1980s. In other words, imbalanced trade—the persistence of economically large trade deficits—is a significant driver of inequality.

In fact, labor market conditions were dis-equalizing not just in the 1980s, but since then as well; we’ve also been at full employment only 30 percent of the time over the full period when inequality has increased (since 1979). Also, pre-1979, over the period when incomes grew together in Figure 1, the trade deficit as a share of GDP was about zero; since then it has averaged close to -3 percent of GDP.

In other words, the causes behind inequality are multi-faceted, as you’d expect given the complexities of both global economies and inequality itself. But not unlike the data section above, where complex definitions can be usefully reduced to a single upward trend, so can the many factors driving inequality be reduced to a useful concept. At least from a labor market perspective, the unifying factor behind rising inequality is diminished bargaining power.

Here, then, is a list of the factors most experts cite as driving inequality upward, each of which reduces bargaining power, clout, voice, and political representation of the many households facing the costs of higher inequality:

  • technology, or unmet skill demands putting upward pressure on the wages of more highly educated workers;

  • globalization, or persistently imbalanced trade in industries where blue-collar, production workers are disproportionately employed;

  • the loss of manufacturing employment due to those trade imbalances in those industries hurt by foreign competition;

  • the absence of full employment, i.e., the fact that since 1979, the labor market has been at full employment only 30 percent of the time;

  • the decline in unions;

  • the erosion of labor standards, including the minimum wage (i.e., the loss of its real value) and overtime rules, and increased misclassification (classifying regular workers as self-employed and thus denying them various protections);

  • “financialization,” or the increase in occupations within industries (finance, real estate) where workers claim exorbitant “rents” (profits that go well beyond the actual economic value of their work).

  • regressive tax policy, including tax cuts to high marginal income tax rates, “preferential” rates on non-labor income, and tax avoidance associated with foreign earnings by US multinationals;

  • the interaction between high levels of wealth concentration and money in politics, leading to the legislative success of inequality-inducing policies (e.g., deregulation, regressive tax cuts) and the failure of policies aimed at reducing inequality (e.g., increases in the federal minimum wage).


Why does inequality matter?

While all economies, capitalist and otherwise, will always have some degree of inequality, our historically elevated levels of economic inequality (which, as the last bullet above suggests, correlates with political inequality) is problematic. But that needs more than an assertion. What’s wrong with high inequality?

First, consider the inequality wedge between productivity and compensation of the typical workers in Figure 9. Their living standards, and that of their families and children, are negatively affected by their increasingly diminished share of the benefits of productivity growth, though as the CBO data show, some of that difference is made up by transfer payments and tax credits (though for working-age families, these benefits tend to reach the poor more so than the middle class).

Next, as Ben Spielberg and I show in a recent review of an important and growing literature, our current high levels of inequality are reducing opportunities over the course of the life cycle for many kids in disadvantaged families. Some of the best-known work in this area is by Raj Chetty and colleagues, showing that places with high inequality are negatively correlated with the mobility of children who grow up there.

Spielberg and I argue that there are three destructive roles that inequality plays in the diminished provision of opportunity. First, it contributes to residential segregation by income, invoking the Chetty et al. findings just noted. Second, it leads to unequal access to quality education, with lifelong repercussions. Third, “inequality directly undermines equality of opportunity, likely through a variety of mechanisms. As the gap between the rich and poor widens, lower-income families have less ability relative to their rich counterparts to invest in enrichment goods for their children. Children from families with less income have relatively less extensive and privileged social networks and, compared to their rich peers, are more likely to experience the type of ‘toxic’ stress that can hamper brain development and long term academic, health, and economic outcomes.”

In addition, high levels of inequality can drag down economic growth in at least two ways. First, there’s the consumption channel. Our economy is uniquely driven by consumer spending—almost 70 percent of US GDP is consumer spending, compared to 55 percent in Europe and 35 percent in China. If inequality channels increasing shares of growth to the wealthy, who tend to spend a smaller percentage of their marginal dollar than do low- and middle-income households, we’d expect rising inequality to slow macroeconomic growth.

The Reconnection Agenda by Jared Bernsten Art by by Anthony Martinez and Kate Bernstein

The Reconnection Agenda by Jared Bernsten Art by by Anthony Martinez and Kate Bernstein

It is surprisingly hard to see this effect in the data over the long post-1979 period of rising inequality, but that may result from a second reason why inequality could restrain growth (also, income redistribution of the type we see in the CBO data should offset the inequality-induced marginal spending problem). As inequality contributes to the middle- and low-income wage stagnation shown in the wage figures above, these households may turn to credit to maintain their living standards. At the same time, wealth concentration means cheap loanable funds and greater demand for financial services. These factors drive up the debt-to-income ratio for lower-income populations and increase the size of the financial sector. Especially in the absence of regulation, this chain of events can lead to a bubble that must eventually bust, with a significant and negative impact on growth, a pattern that’s become all too familiar in the American economy in recent years.

A final reason for taking action against rising inequality was alluded to above, when I noted the absence of adequate financial market oversight. Political scientists have unearthed a toxic interaction between concentrated wealth and the unique extent to which money influences American politics. Comparisons of Americans’ expressed policy preferences with politicians’ voting records and eventual policy outcomes find that government is largely unresponsive to the opinions of low-income citizens yet highly responsive to those of wealthy constituents. While this dynamic has surely long been operative in American politics, increasing wealth concentration appears to be making this divide even more pronounced.

Interestingly, in the current election cycle, establishment Republican candidates have themselves arguably been hurt by these dynamics. To their credit, these candidates acknowledge the inequality trends documented above—it would, at this point, be hard not to—but so far, their prescriptions, as predicted by the political science findings just noted, have been those desired by the wealthy: tax cuts that lavish billions on the wealthiest households, deregulation of industry, and attacks on social insurance programs that serve as somewhat of a bulwark for the poor and middle class against the impact of market-driven inequalities. Beset by the negative impacts of inequality, many Republican constituents are rejecting these establishment candidates in favor of outsiders with more populist messages.

What to do about it?

My recent book, the Reconnection Agenda (it’s free online!) was written with the sole purpose of answering this question of what should we as a society do about rising inequality. I mention that not just to drum up interest for more of my scribblings but because the anti-inequality policy agenda is a deep one; it requires more thought and explanation than I have space for here. But as you’ve shown great stick-to-it-tiveness to get this far, allow me to reward you with a brief, thematic discussion of the solutions. Read the book for details.

To lower today’s levels of inequality, policy would need to accomplish two broad goals: raise the bargaining power of the American worker and lower the political clout of the wealthy elite. Most people, when they hear “bargaining power,” think “unions,” and more collective bargaining would unquestionably help. But to reach many more workers who have been hurt by the dynamics portrayed in Figure 9, we also need to get to and stay at full employment in the labor market.

That invokes the needs for active fiscal and monetary policy working together to generate full employment. Lowering the trade deficit is also essential, which means more attention to both dollar (exchange rate) and manufacturing policy. The importance of improved labor standards was also noted above. Direct job creation will likely be needed to assist the least advantaged, who are at risk of being left behind even when the macroeconomy is firing on all cylinders. Reducing the power of the elites calls for better oversight of financial markets (by breaking the bubble/bust syndrome, this would also help the labor market stay at full employment when we get there), less money in politics, and less privileging of unearned incomes in the tax code.

What did I leave out?

Though I tried to be exhaustive, this overview left out, or gave too short shrift to, certain aspects of the current inequality debate. While I touched on the relationship between economic inequality and economic mobility—one’s lifetime movements up and down the income scale relative to one’s cohort—there is much more to be said here. The causal channel, suggested above, where high levels of inequality erect barriers to mobility, is a key issue in social science today, yielding rich analyses. I mentioned Raj Chetty’s well-known work is this space, but sociologist Robert Putnam’s book “Our Kids” is another important and deep dive into these connections.

I did not speak of consumption inequality, though some claim it has gone up less than the other types of inequality discussed in this paper. In part, I left it out because the consumption data are simply not nearly as reliable as the data featured here. Some of what I consider to be the best work—research where the analysts try to adjust for the shortcomings in the data—find that consumption inequality has gone up about as much as income inequality (I did note above the very important observation about the growing inequality of family consumption of “enrichment goods” for their children), but there is some disagreement about this conclusion.

I also left out one important newer piece of evidence: the shift in “factor incomes” from compensation to profits. Roughly speaking, the income generated in economies can be divided into income from work, which is basically wage income (including benefits); capital income, or profitability from the ownership of capital assets; and a residual income category, not counted in what follows (e.g., self-employed income, since it’s hard to know how to divide this into the other two, much larger, shares). For decades, the wage and profit shares were relatively constant, breaking down into roughly 2/3 compensation and 1/3 profits. Over the past decade or so, the compensation share has fallen, meaning the profit share has risen, by something on the order of 4 to 7 percentage points. Since wages are more equally distributed than profits, this type of shift automatically increases income inequality.

Finally, I referenced in passing the point that tax policy doesn’t just affect the inequality of after-tax income. When the tax code privileges the wealthy, as does ours with its many carve-outs for asset-based income and overseas profits, this tends to affect the pre-tax distribution as well, as the wealthy are able to structure their pre-tax income to take advantage of the tax code. A deeper dive into this part of the research shows that American inequality continues to be exacerbated by such incentives in the tax code.

In sum, there’s no legitimate rock you could look under to refute the higher inequality story. The forces generating greater dispersion of economic outcomes are firmly embedded in our economy. They are powerful and as I’ve stressed, they are causing considerable harm to many different aspects of our economy, our lives, and our futures. Importantly, this is a view now widely held by policy makers, including those aspiring to lead the country. Whether they can or will do something about is yet to be seen.

TPM illustration includes graphics from Albund and Wongwean / Shutterstock.com