Jason Furman, chairman of the Council of Economic Advisers, set out to explain “middle-class economics” in the Wall Street Journal, March 11, in an earlier Vox blog and in a presentation to National Association of Business Economists (NABE), as well as the first chapter of the Economic Report of The President.
The intent is to make the recent economy look healthier (massaging 2.3-2.4 percent growth for 2013-14 into 2.7 percent), and to claim that “subpar” 2010-14 income gains for the middle class (generously defined as the bottom 90 percent) are not due to a subpar recovery but to something that has gone on ever since 1973. His Wall Street Journal article complains of “the decades-long trend of slower income growth for the middle class.”
Furman says, “Congressional Budget Office data (with a minor extrapolation) show, median U.S. incomes are up 17 percent since 1973.” Actually, CBO data start with 1979 and end with 2011, so it takes more than minor extrapolation to extend that back to 1973 or forward to 2013. CBO estimates show real after-tax median income rising from $45,400 in 1983 to $68,000 in 2008 (in 2011 dollars), but not yet back to the 2008 level by 2011. Making up a number for 1973 can't undo stagnation after 2008.
He continues: “But from 1948-73, median incomes rose 110 percent, according to broadly comparable Census estimates.” Yet the two series aren’t remotely comparable. Unlike pre-tax “money income” from the Census Bureau, the CBO subtracts federal taxes (middle-income tax rates were nearly cut in half since 1981) and includes rapidly increased health and other in-kind benefits from employers and government (Medicaid, SNAP, CHIP and housing allowances).
Furman repeatedly sets up 1973 as the ideal, with productivity, incomes, and fairness all supposedly downhill after that. The reason this old trick is still so popular is that 1973 was the last year Nixon’s price controls appeared to keep the consumer price index artificially low – creating a brief artificial spurt in measured real wages and productivity. When price controls exploded in a wave of shortages, average weekly earnings (in 1982-84 dollars) dropped from $341.36 in 1973 to $314.77 in 1975, $308.74 in 1979, and $290.80 in 1980. A falling dollar and rising tax rates stimulated demand and discouraged supply, giving us two nasty episodes of “stagflation.” Amazingly, those trying to blame current discontents on the distant past continue to hold up 1973 or 1979 as ideals -- idolizing the economics of Richard Nixon and Jimmy Carter.
Furman writes, “In 1973 the bottom 90 percent of households received 68 percent of the nation’s income, a figure that has fallen to 53 percent today.” But because this is no measure of the nation’s income, Furman has no idea what the bottom 90 percent share has been. Instead of using any official measure of personal or household income, Furman is citing an untenable private estimate of income reported on tax returns --with income frequently redefined by changing tax laws. “From 1944 on,” Piketty and Saez explain (Table A0), “total income is defined as total Adjusted Gross Income less realized capital gains in AGI, taxable Social Security and Unemployment Insurance benefits, and adding back all adjustments to gross income. Income of non-filers is imputed as 20 percent of average income (50 percent in 1944-1945).”
As I pointed out in a recent article and blog post, the data Furman cites report that the average real income of the bottom 90 percent was higher in 1968 than it was in 2013. Claiming to actually believe such preposterous data is a mark of unlimited gullibility or deception.
This graph illustrates a few points made in my recent Wall Street Journal article. First of all, the Piketty & Saez mean average of bottom 90% incomes per tax unit is not a credible proxy for median household income, particularly since the big reductions in middle-class taxes from 1981 to 2003.
Second, the red bars claiming bottom 90% incomes in the past six years have been no higher than they were in 1980 (Sen. Warren) or even 1968 (see the graph) is literally unbelievable. If that were true then all other income statistics -- including GDP -- would have to be completely false.
The Piketty & Saez estimates before 1944 describe total income as Personal Income less 20% (because not all income is reported). Postwar data use a modified version of Adjusted Gross Income as a proxy for personal income, with no transfer payments or health benefits, and that measure has become less and less credible over time. This makes the estimates of bottom 90% incomes simply worthless, as well as related claims that the top 1% "captured" all the cyclical gains (and losses!). If total income were calculated the same way it was in 1928, the the top 1% share would drop from 17.5% to 13.3%. Grossly underestimating total income by greater and greater amounts created an artificial increase in top 1-10% shares of such increasingly understated income.
As the blue line in the graph shows, many measures of income in 2012 or 2013 were not yet back to the peak levels of 2007 or 2000. But that definitely includes real incomes of the top 1%, which were 20.6% lower in 2012-2013 than they were in 2007.