18 November 2016

How Common Are Very High Income African-Americans?

Klopfer, in an artful bit of statistical jujitsu manages to make some reliable inferences about high income African-Americans in the United States, which shows that the 1% is whiter than a moderate sophisticated intuition would expect. Full text of the paper does not appear to be available at this time, so the raw numbers which would be interesting in and of themselves, are not available.
I show that the Decennial Census and American Community Survey are valid sources of statistics on the top of the wage distribution, and replicate well-known findings from the administrative data commonly used to study income inequality. These surveys contain rich demographics that are not available in the administrative data, and I use them to provide the first race-specific estimates of top income earners in the United States. 
While the number of black Americans among the top 10, 5, and 1 percent of wage earners in the United States has grown sevenfold since the 1960s, the wages of blacks relative to whites haven't grown since the Civil Rights era. 
These facts are mechanically linked by rising wage inequality in the economy more broadly, and consequently the majority of progress by blacks into the top echelons of wage earners has occurred within industries, occupations, and levels of education, with comparatively little reallocation across these categories. 
Further, wages of blacks at the pth percentile of the black wage distribution, relative to those of whites at the pth percentile of the white wage distribution, are 70% from the 50th to the 95th percentile, but black relative wages fall to 50% by the 99th percentile of the respective wage distributions. I show that this relative wage slump occurs because top white wages follow the Pareto distribution and top black wages do not. Models that generate Pareto wages cannot explain the black distribution, but models of discrimination or human capital investment may.

No comments: