Aparna Mathur, aei.org
image (not from article) from
The full picture of income and resources shows inequality is not as bad as the doomsayers claim.
Many believe that the debate around income inequality – that inequality is increasing at a rapid pace and that the share of incomes at the top has become more concentrated over time – has been settled. However, a new study by economists Bricker et al. cautions against jumping hastily to these conclusions.
In their paper, the authors find that earlier studies, relying largely on tax returns data to measure income inequality, have overstated the pace and share of income gains at the top. Addressing measurement and conceptual problems with earlier studies, their paper finds that the income share of the top 1 percent is lower today and is growing less rapidly than the estimates reported by economists Thomas Piketty and Emmanuel Saez.
Specifically, the share of income received by the top 1 percent is reported at 23 percent in 2012 by Piketty and Saez, while Bricker et al. estimate it to be at 18 percent in 2012. Where do these differences in results come from?
The Piketty and Saez study relies on tax return data to estimate income inequality. Tax return data is good for estimating top incomes since most types of income that top earners earn, such as capital and wage income, are required to be reported in tax returns. However, tax return data does a poor job of capturing incomes for households at the bottom of the income distribution.
Since a large share of low-income households don’t owe federal taxes and do not file tax returns, these households are largely missing in tax return data. Piketty and Saez do attempt to impute incomes for these non-filers by assuming that their incomes are 20 percent of the average income of filers. The problem remains that we do not directly observe these incomes, and more importantly, as argued below, the types of incomes earned by non-filers are very different from those of filers. Therefore, top income shares may be overstated in tax data because we are simply capturing fewer people, and therefore less income, at the bottom.
The larger problem with the use of tax return data is that low-income households often earn incomes that are not required to be reported in tax returns or are simply not reported well in these data. These include types of in-kind or cash transfer programs such as Social Security benefits, food stamps (SNAP), Medicaid, Medicare and other government provided health benefits as well as employer-provided health insurance. Over the last several decades, there has been a significant expansion in these types of transfer programs with the explicit purpose of providing income support or boosting living standards for households with poor wage and salary incomes.
As these incomes don’t show up in the tax data, studies based on this tax data suffers from biases which understate incomes at the bottom, and thereby show much higher income concentrations at the top.
The Bricker et al. paper finds that accounting for these differences in the concept of income and using household survey data to better capture people at the bottom of the income spectrum causes the share of income at the top to lower by 21 percent or 5 percentage points.
The finding that accounting for these additional sources of cash and in-kind transfers significantly boosts incomes at the bottom lends support to other studies that have found that consumption inequality is not as high as the Piketty-Saez income inequality measures suggest. In other words, people’s standards of living have continued to improve over the last several decades despite much of the income gains going to the top.
In a study Kevin Hassett and I co-authored, we find that material standards of living for households have improved significantly across the income distribution, and the share of consumption expenditures for the top and the bottom 20 percent has remained fairly stable. Combining these findings with the Bricker et al. paper, it seems likely that the missing piece of the puzzle is the role that these types of transfer programs have played in boosting incomes at the bottom of the distribution and thereby allowing families to improve their consumption and well-being over time.
All of this suggests that we are still trying to determine the disparities along the income distribution. There have been large increases in programs to help low income households, and until we capture those in the data, our understanding of income inequality is incomplete.
But in my opinion, the larger point is that a focus on inequality per se is misguided. As the Bricker et al. study suggests, income inequality can narrow simply because we are redistributing income to lower income households through an expansion of transfer programs.
But is that the narrowing of the distribution that we care about? While these programs are helpful and important, we can only make real improvements in the lives of people if we can improve opportunities and outcomes through access to good education, skill building, and good jobs that lead to higher incomes and move people up the income ladder. Seeking parity in those indicators would be a much better measure of success than a narrowing of the income distribution through even more transfer programs.