Politicians, researchers, and the media have given a good deal of attention recently to widening income inequality. Yet very few have paid attention to how—and how well—we measure income. Different measures of income show very different results on whether and how much inequality has risen. Without clarity, even honest and non-ideological public and private efforts to address inequality will fall short of their mark—and, in some cases, exacerbate inequality further.
How we typically measure income
Income measures tell different stories about opportunity and can be useful for different purposes. Some studies on income inequality measure income before government transfers and taxes—for instance, studies that compare workers’ earnings over time. Studies of the distribution of wealth or capital income, too, typically exclude any entitlement to government benefits. These “market” measures capture how much individuals have gained or lost in their returns from work and saving.
More comprehensive measures examine income after transfers and taxes. Transfers include Social Security, SNAP (formerly food stamps), cash welfare, and the earned income tax credit. Taxes include income and Social Security taxes. These measures best capture individuals’ net income and what living standards they can maintain, but not their financial independence or how much they are sharing directly in the rewards of the market.
Within both sets of measures (pre-tax, pre-transfer and post-tax, post-transfer), most studies still exclude a great deal. Health care often fails to be counted, even though increases in real health costs and benefits now take up about one-third of all per-capita income growth. Most of these health benefits come from government health plans (like Medicare and Medicaid) or employer-provided health insurance. Households pay directly for only a minor share of health costs, so they often don’t think about improved health care as a source of income growth.
How improving our definition of income—and using alternative measures—sharpens our view of income inequality
Starting with a more comprehensive measure of income, and then breaking out components, can improve our understanding of income inequality and its sources. As an example, let’s use some recent Congressional Budget Office (CBO) estimates, which provide perhaps the most comprehensive measure of household market incomes (consisting of labor, business, capital, and retirement income), then add the value of government transfers and subtract the value of federal taxes.
Between 1979 and 2011, the average market incomes—that is, incomes before taxes and transfers—of the richest 20 percent of the population (or the top income quintile) grew from 20 times to 30 times the incomes of the poorest 20 percent (the bottom income quintile).
When examining income after taxes and transfers, however, the relationship between the top and bottom income quintiles is more stable. It starts at a ratio of about 5.5:1 in 1979 and increases very slightly to 6:1 by 2011.
We find somewhat similar trends when comparing the fourth quintile with the second quintile. After taxes and transfers, income ratios don’t change much over this period, though the market-based measures of income show increased inequality. Of course, the very top 1 percent still has gained significantly; in 2011 it had nearly 33 times the income of the bottom 20 percent, compared with about 19 times in 1979.
What’s still missing
Even the CBO measures, however, are far from comprehensive. They exclude many benefits that are harder to measure or distribute, such as the returns from homeownership and the value of public goods like highways, parks, and fire protection. As Steve Rose points out, even more elusive are the broadly shared gains in living standards brought by improved technology and new inventions.
Why getting it right matters
Though defining income may seem like merely a technical exercise, it has huge consequences. Inequality has always been a political football: all sides tend to quote statistics that support their policy stances while ignoring the statistics refuting them. Yet if we want good policy, we’ve got to be open to how well any particular policy might improve income equality by one measure or make it worse by another, often at the same time. This is not a new story: bad or misleading information almost inevitably leads to bad policy.