Inequality Exists. Ask the Middle Class.
Income inequality creates social tensions in two directions. From the point of view of the majority, it isn't just a question of envy, although that's real; it's also the distortion in prices of important assets like houses in good school districts.
And from the point of view of the extremely affluent, it becomes all too easy to disengage from the rest of America and to see one's wealth as divine reward for special genius, rather than for many other factors, including blind luck. That sense of difference in turn leads to the absurd sort of narcissism and insensitivity to others.
Perhaps recognizing how poorly this message plays, some economists have shifted to arguing that, because the leading studies overstate the problem, widening income inequality doesn't really exist.
An example is a widely cited study by Cornell University's Richard Burkhauser and two others. His paper, "A 'Second Opinion' on the Economic Health of the American Middle Class," concludes that Piketty and Saez overstate income inequality and that the middle class actually showed strong income growth from 1979 to 2007.
Growth in the median income of tax units (single or joint taxpayers) in the Piketty and Saez studies was only 3.2 percent over the period. But when using households as the basis, Burkhauser's preferred method, median income grew about 37 percent. Presto -- a healthy and prosperous middle class.
I take issue with that conclusion. A more complete inquiry into the data leads to the opposite conclusion -- the middle class has fallen behind over the last generation in important respects.
The Burkhauser paper has received a great deal of attention because its conclusions are highly convenient for market triumphalists, not to mention the affluent who fear their income will be siphoned off for redistribution to the 99 percent.
How can studies on the same topic come to such different conclusions? The answer comes down to the definition of income and whose income is being measured.
Because Piketty and Saez rely on actual federal income-tax returns, they have by far the best detail on all the components of market income (wages, self-employment income, income from investments) in each year, particularly at the highest levels.
But tax-return data don't include government transfers, such as food stamps and Medicaid, or in-kind benefits such as housing subsidies. So these forms of "income" aren't included in the Piketty and Saez data.
Burkhauser, on the other hand, relies on the March Current Population Survey (CPS), a detailed, face-to-face annual Census Bureau survey of about 60,000 households.
CPS data include government benefits but not capital gains (profits from selling investments, a business or a home). If you want to say something useful about how the most affluent Americans are doing relative to the rest of us, you need to include capital gains. But the Census Bureau doesn't collect information on capital gains, so the Burkhauser paper simply ignores them.
Both datasets have drawbacks. But when it comes to measuring top incomes, I prefer tax-return data. High-income Americans are likely to be more forthcoming about their exact income on tax returns (where noncompliance can lead to large civil or even criminal penalties) than a Census Bureau interview.
Another complaint with the Piketty and Saez results is that they ignore important forms of what might be called "invisible" compensation such as employer-provided health insurance. The street value of that policy -- what your employer pays -- might be in the neighborhood of $12,000 a year for a family plan.
Piketty and Saez ignore such invisible income because it isn't reported to the IRS. Burkhauser, by contrast, imputes the missing data. (That is, he does his best to estimate what the missing number might be.) This tends to bring up middle-class income relative to the affluent. Burkhauser argues that cash wages may have stagnated in real terms, but that's because employees are getting paid more in invisible currency.
There is logic to this, but there are good reasons to suspect that the value to employees of employer-provided healthcare isn't the same as the amounts employers pay. Most employees have no choice but to take the health-care program given them, and get lower wages as a result. They can't decide how to allocate their wages as they see fit.
Rapidly increasing health-care costs are in large measure the result of overconsumption of health care (and indifference to its cost). By including the imputed amount of health-care premiums at face value, Burkhauser confuses income and welfare.
Where does all this leave us? Even with its limitations, the Burkhauser data still show a substantial increase in inequality from 1979 to 2007. Congressional Budget Office data show that the market incomes of all households grew about $6 trillion, to $10.6 trillion in 2009 from $4.6 trillion in 1979, in constant 2009 dollars. The middle quintile of households -- the middle class -- captured about 10 percent of this. The top 1 percent captured more than 28 percent, or nearly three times as much.
The same data show that in 1979 the middle quintile of households earned 77 percent more market income than did the top 1 percent -- just as one would expect in a market economy. In 2007, however, the top 1 percent earned 61 percent more in market income than did the entire middle quintile. This is an extraordinary reversal in less than 30 years.
Burkhauser tells us that median household market income -- adjusting for household size and the value of employer-sponsored health care -- grew all of 19 percent from 1979 to 2007, with a third of that coming in the precrash bubble. In addition, Burkhauser's results are boosted because more adults are earning income in each household, beyond the husband and wife filing a joint return. There are also fewer mouths for those income-earning adults to feed, probably because they've decided to have fewer children.
Some of these shifts capture welfare-enhancing moves, such as gay couples cohabiting now that society is more tolerant of that choice. But when a mother-in-law moves in or an adult son won't move out, the shifts reflect deterioration in income security. Burkhauser is insensitive to the welfare implications of these developments.
At any rate, the very modest gains in median real household market incomes from 1979 to 2007 were mostly the result of more women working -- and getting paid more. Male workers in the middle saw no gains for their labor. Does this not explain a great deal about political and social unrest in America today?
But as the CBO data show, almost half the growth in median real disposable income was the result of lower taxes or larger government transfers, not real economic gains. Median disposable income grew about 35 percent from 1979 to 2007, while median household market incomes grew 19 percent.
A moment's reflection will reveal that one cannot trumpet the advances in available resources of the middle class when those advances are driven to such a large extent by government tax and transfer policies, unless you also are willing to entertain ever-increasing government transfers -- and steeper taxes on the highest-income Americans to fund them.
But market triumphalists use the work of Burkhauser and others to argue the opposite: Everything is fine, and there is no need to raise taxes.
The buoyant middle class that inequality deniers purport to have discovered is a creature of a decades-long fiscal policy of borrowing against the future by delivering more transfer benefits than our current level of tax collections can support.
(This is the second of two excerpts from "We Are Better Than This: How Government Should Spend Our Money," which will be published Oct. 1 by Oxford University Press.)
This column does not necessarily reflect the opinion of Bloomberg View's editorial board or Bloomberg LP, its owners and investors.
To contact the author on this story:
Edward D Kleinbard at firstname.lastname@example.org
To contact the editor on this story:
Paula Dwyer at email@example.com