One of the main reasons that I and others have given for leaning on the Fed to keep interest rates down is that low unemployment disproportionately benefits those at the bottom. While we can and should try to help the disadvantaged through increased education, training, child care and other programs necessary to give them a foothold in the labor market, the easiest thing is allow them to get jobs.
When the Fed raises rates it is deliberately slowing the economy and thereby reducing the number of jobs available. The people who are then denied jobs are disproportionately the most disadvantaged groups, such as blacks, Hispanics, and less educated workers. These workers are hurt not only because fewer have jobs, but also because the bargaining position of those employed weakens when there is higher unemployment. In this telling of the story, wage gains for those at the bottom should be strongest during periods of low unemployment, as we have been seeing in the last few years.
For this reason, the latest data on median wages for black workers is somewhat surprising. The Bureau of Labor Statistics Usual Weekly Earnings series showed real median weekly earnings for full-time black workers in the first quarter of 2018 were up just 0.6 percent from the first quarter of 2017. Furthermore, taking the last three years together, it showed real weekly earnings for blacks were up by a meager 1.1, trailing the 2.8 percent rise in real earnings for the median white worker. The racial gap seems to be increasing even in this period of relatively low unemployment.
While that is a discouraging story, the picture is not as bad as it may first appear for a couple of reasons. First, the quarterly data are erratic, they jump around due to noise in the data. If we take a four-quarter moving average ending with the most recent quarter, this average has risen by 3.1 percent for African Americans over the last three years. That's better, but it still trails the 4.7 percent increase in the real median earnings for whites by this measure.
But there is another aspect of the labor market that needs to be considered. As noted above, tighter labor markets have disproportionately allowed blacks to get jobs. The number of blacks employed in the first quarter of 2018 is 10.6 percent higher than in the first quarter of 2015. It's 15.1 percent higher than the number for the first quarter of 2014. By comparison, employment of whites in the first quarter of 2018 was up by 2.9 percent compared with the first quarter of 2015 and 4.2 percent compared to the first quarter of 2014.
This matters for the median wage calculation, since the newly employed workers are likely to be concentrated at the bottom end of the wage distribution. These are likely people with less education and experience than most of the people who had previously been working. That means that the median black worker in 2018 likely had less education and experience than the median black worker in 2015.
This would have the effect of pulling down the wage at the median. If we wanted to do a true apples-to-apples comparison, we would be comparing the median worker in 2015 with someone higher up on the distribution in 2018. If we take the extreme case and assume that all the new entrants to the labor force between 2015 and 2018 were below the median, then the 2015 median black worker would be at roughly the 55th percentile of the wage distribution for black workers in 2018.
The usual weekly earnings for a black worker at the 75th percentile of the distribution was 56 percent higher than the median earnings in the first quarter of 2018. If we assume that the gap is linear (i.e. moving up five percentage points in the wage distribution closes the gap between the 75th percentile and 50th percentile by one fifth) it means that in this extreme case, we would have to add more than 11 percent to the median wage in 2018 to make an apples-to-apples comparison to the data for 2015. By that measure, black workers are indeed gaining ground in this tighter labor market. (This story is consistent with the real wage gains we see over the last three years for blacks lower down in the wage distribution, 9.6 percent for the 10th percentile, and 7.1 percent for the 25th percentile. These cutoffs would be less affected by the large increase in employment since many of the new workers would be above these cutoffs.)
We can do more careful analysis with the microdata, where we see what is happening to pay for African Americans when we control for age, education, gender, and other factors, but the picture in this summary data is not as negative as it may first appear. We know for certain that there has been a huge increase in employment over the last four years for black workers. Those at the bottom clearly have seen substantial wage gains over this period. It seems likely that those at the middle did as well.