Democrats are still stinging from last month’s epic defeat at the polls. Some of them attribute that unhappy outcome to ideological temperance. The party would do better, some argue, if their candidates took a more strident tone, denouncing the rich, preaching that Washington governs in the service of Wall Street, and calling for more aggressive redistribution of income and wealth to dampen rising inequality.
But a new study published in the Southern Economic Journal pulls at a critical thread of that narrative. It found that, at least according to one measure of income, the gap between rich and poor actually narrowed between 1989 and 2007.
The study’s authors – Philip Armour and Richard V. Burkhauser of Cornell University and Jeff Larrimore of the U.S. Federal Reserve Board – showed that measuring inequality is a tricky business, one that is highly dependent on how researchers define income. The table below summarizes some of their findings.
|Three Measures of Inequality
Average Inflation-Adjusted Growth in Income, 1989-2007
|What does it measure?||Cash market income||Post-tax and transfer income, adjusted for household size||Column (2) + accrued capital gains (includes housing)|
|Who used the measure?||Piketty||Numerous researchers||Armour, et al.|
|Top 5 percent||9.9||12.7||-2.7|
The first column presents the sort of data that Senator Warren most typically cites. It is based on the methodology used by world-renowned French economist Thomas Piketty. It measures “cash market income” – essentially money that households receive through labor or investment. By this measure, which looks at pre-tax earnings, things darkened between 1989 and 2007 for households in the lowest quintile (i.e., the lowest 20 percent of earners). Their average, inflation-adjusted earnings dropped by nearly one-third (32.9 percent). And while those in the middle quintile barely kept their heads above water (a 2.3 percent real income increase), those at the top did just fine, increasing their income by 12.9 percent. The rich got richer, the poor got poorer and the middle class stagnated.
This picture, however, is not entirely accurate. Because it measures only pre-tax income, the Picketty methodology overlooks the redistributive effects of a progressive tax regime and the transfer benefits these taxes (and government borrowing) finance: Social Security, Medicare, Medicaid, and more than 100 welfare programs. By ignoring the taxes the rich pay and the benefits that poor and middle income households receive, the Piketty method omits some rather critical data.
It is also fails to take into account social changes. People are having fewer children and living arrangements are much different than in previous generations. Two working adults living together and sharing expenses are counted as two separate households in the data sets Piketty uses, but are in fact a single unit. An accurate methodology must account for such changes in household composition.
The second column, which relies to a large extent on Census Bureau data, adjusts both for household composition and the effects of taxes and transfer benefits. This changes the picture dramatically for the lowest quintile, whose median income, after accounting for taxes, benefits and household composition, grew by 9.9 percent instead of shrinking by 32.9 percent. The remaining quintiles also fared better under this definition of income. Still, inequality grew, with the average earnings of the top quintile increasing by 15.1 percent.
The third column represents the methodology employed by the authors to compute income inequality. In addition to the Census Bureau data, they look at the Federal Reserve’s Survey of Consumer Finances, which provides data on household assets held in both taxable (e.g., bank accounts, mutual funds) and non-taxable (e.g., IRAs, 401(k)s) accounts. They also factor in data from the House Price Index published by the Federal Housing Finance Agency. They use these data to impute “accrued capital gains” to households.
Column (3) shows the rather startling results of including these additional bits of information. The big winners since 1989 have been people in the bottom quintile, whose average incomes rose by 13.7 percent between 1989 and 2007. The biggest losers were those Senator Warren loves to hate: people in the top quintile, whose incomes grew by less than 1 percent over that period, and the top 5 percent, who experienced a 2.7 percent reduction in their net income.
The authors’ methodology is not without controversy. The Survey of Consumer Finances samples only 5,000 households every three years. It is thus much less comprehensive than either IRS or Census Bureau data. Moreover, the survey’s methodology changed so much that the authors chose not to use data collected before 1989.
Critics also object to the idea of imputing income based on accrued capital gains. The fact that your house is appreciating in value or your IRA did well this year doesn’t necessarily give you more income. On the other hand, measuring only taxable, realized capital gains also has limitations. The government, for example, doesn’t tax capital gains on the sale of a residence unless that gain exceeds $500,000. The average price of a home owned by a family in the middle quintile rose by nearly $66,000 between 1989 and 2007. A homeowner who realized that amount paid no taxes on the sale of the home. In measuring household income, that $66,000 arguably should be accounted for. The authors account for it by imputing that capital gain to homeowners as average housing prices change over time.
Similarly, households in the bottom quintile had an average of nearly $43,000 in total equity investment assets in 2007, according to the Federal Reserve data. While imputing them accrued capital gains may seem an imperfect way of valuing these savings, it is arguably better than ignoring them entirely.
The Congressional Budget Office, in an analysis published last month, provides a fourth way to compare income across quintiles. Like the authors, they take into account taxes and transfer payments. Unlike the authors, they treat only taxable capital gains as income, ignoring many assets that households own. The chart below (click to enlarge) shows the Gini Index between 1979 and 2011 that CBO computed based on its methodology. The Gini Index is a commonly used measure of income inequality that varies between 0 and 1. A society in which everyone’s income was equal would get a score of 0; one in which one person gets all the income and the rest get nothing would get a score of 1.
So has inequality increased in the U.S.? As former Senator Phil Gramm and economist Mike Solon noted in a recent Wall Street Journal piece, the answer to that question depends very much on which year you choose to begin your analysis. If you choose 1979 as the base year, the CBO data would show a 22 percent increase in the Gini coefficient (.358 to .436). But if you were to select 1985, the change would be just 6 percent (.410 to .436).
Neither represents a complete picture. The Gini coefficient has experienced three big run-ups since 1979: during the central bank’s 1980s war against inflation, the dotcom bubble of the mid to late 1990s and the housing bubble that burst in 2007. All three spikes in the Gini index coincided with ambitious central bank interventions. In the aftermath, the index trended downward. The most abrupt decline occurred when housing and equity prices plummeted after 2007. The Gini index fell by nearly nine percent between 2007 and 2009. Since then, it has drifted upward.
Where it has headed since 2011 is difficult to say. The Federal Reserve has pursued extraordinary monetary policy since 2009 and there has been a concomitant rise in equity prices. Both would suggest that the Gini index would have tracked higher over the past three years.
Economic growth, however, has been tepid and government has been fiscally hyperactive. Obamacare’s new taxes on the rich took effect in 2013, and Congress also voted to raise the top rate in January of that year. Those taxes will adversely affect high income households. The new health law will extend an estimated $1.8 trillion in income-related health insurance subsidies between 2014 and 2023, most of which will benefit those in the bottom quintile. Middle income households are taking a hit, since a good portion of Obamacare’s new spending is offset by reducing the value of Medicare benefits to seniors. The new law’s thicket of regulations also has eroded employer-sponsored health benefits that most middle income households enjoy. That erosion will continue at least until 2018, when the government will impose a 40 percent excise tax on coverage that it deems to be overly generous. How this all will play out is anyone’s guess.
Inequality, in short, is very difficult to predict and quantify. It is highly dependent on what economists measure and how they measure it. No perfect index exists and various methodologies produce wildly different results. It is a subject whose nature, causes and remedies demand more study.
Instead, it will fuel more demagoguery. The Left will promote its distorted narrative to incite racial and “class” resentments and direct that animosity against Republicans. They also will use it to advance tired and ineffective policy bromides, like increasing the minimum wage, rewarding students for not making timely payments on their college loans, and confiscating more of the wealth and income of successful individuals and businesses.
Aggressive redistributionist schemes, which rely heavily on the political system to allocate the fruits of private sector output, have failed to arrest inequality in Europe, as Piketty’s own data attest. But that won’t stop his disciples from joining him in calling on government to double down on those ill-conceived policies.
The economics of envy, as Gramm and Solon noted, don’t work. Democrats are fervently hoping that the politics of envy will.