I enjoy reading Eduardo Porter's columns in the NYT and usually learn a lot from them, but I think he has made a mistake in arguing that white workers have fared worse in the recession and recovery than African Americans. In his Wednesday piece, he included a chart showing employment among whites had actually fallen by roughly 700,000 since November of 2007, while it had risen by more than 2 million for both African Americans and Asian-Americans and by twice this amount for Hispanics. This seemed to suggest a radically worse labor market experience for whites than for other groups.

After several people asked me about this chart, I checked the change in the working age (ages 16–64) population for each group. The number of whites in this age grouping had fallen by more than 2 million between 2010 and 2015, while the number of African Americans had risen by 1.4 million and the number of Asians had risen by 1.7 million. I'm sure the number of Hispanics grew even more, but these data are on a different table. In other words, the differences in the number of people working by demographic group reflected less their differences in labor market outcomes than their different population trends.

Porter has a piece today responding to this issue, the headline of which tells readers, "across age groups, whites fared worse in employment rates." The problem is that the data really don't support this claim. Porter notes that the percentage of prime age (ages 25–54) white workers who were employed fell by 2.0 percentage points from November 2007 to November 2016. By comparison, "[l]ast month 74.5 percent of prime-aged blacks held a job, 0.3 percentage points more than in November 2007."

While that would seem to suggest that African Americans had a sharply different experience from whites, this conclusion rests on comparing single months in a very erratic data set. If we take the first 11 months of 2016, we find that the EPOP for prime age blacks was 73.2 percent. This is down by 1.7 percentage points from the 74.9 percent average for the first 11 months of 2007. This is only slightly better than the performance of the EPOP for whites, and of course starts from a lower level.

The same problem arises with the comparison of the 55 to 64 age group.

"The employment rate of whites from the ages of 55 to 64 declined slightly — to 63.6 percent from 64.1 percent. By contrast, the employment rate of blacks, Hispanics and Asians increased."

Taking the first 11 months of both years, the EPOP for African Americans declined by 0.7 percentage points, from 52.7 percent to 52.0 percent. This is actually slightly larger than the decline for white workers.

I am leaving Hispanics out, both because I don't have immediate access to their EPOPs by age group and because their EPOPs are likely to follow a different pattern due to immigration flows. Immigrants who have difficulty getting a job are likely to leave and others will not come if they face dismal job prospects. Asian-Americans actually saw a sharper decline in prime-age EPOPs than whites, as Porter notes.

To be clear, I think Porter is right in seeing support for Trump as being to a substantial extent a response to bad economic prospects. But the economic prospects of working-class whites in the last decade were not notably worse than the prospects of working-class blacks. The difference in their voting patterns can be explained by the fact that working-class whites who feel they are being left behind by the economy can see hope in a white nationalist like Trump, working-class blacks, Hispanics, and Asian-Americans cannot.