The impact of the COVID-19 shutdowns showed up very clearly in the March data as the Bureau of Labor Statistics reported a loss of 701,000 jobs, a decline almost as high as the peaks hit in the housing crash. The unemployment rate jumped 0.9 percentage points to 4.4 percent, while the employment-to-population ratio (EPOP) fell by 1.1 percentage points to 60.0 percent.
The size of the impact on the March data is somewhat surprising since the reference period for this report is the pay period that includes March 12. This was well before most states had initiated shutdowns; although it does seem likely that high turnover sectors, like retail and restaurants, would have put off hiring replacements for workers who had left, given what was known at the time.
Retail lost 46,200 workers in March, while restaurants lost 417,400, 3.4 percent of total employment. There is little doubt that the hit in April will be at least two orders of magnitude higher for retail and an order of magnitude higher for restaurants, as these sectors are at the center of the shutdown.
While these sectors were hit hard by the shutdowns, job losses were seen pretty much across the board. The temporary help sector lost 49,500 jobs, the construction sector lost 29,000 jobs, and the manufacturing sector lost 18,000 jobs. Mining shed 7,000 jobs, mostly in the category of support activities. This decline precedes the sharp plunge in energy prices in the last two weeks. Job losses in the sector will almost certainly be far higher in April.
One surprise, and a very real cause for concern, is a loss of 42,500 jobs in the health care sector. This was mostly in physicians’ and dentists’ offices, presumably the result of non-emergency visits and procedures being canceled.
While the March report shows in a big way the effect of the virus, it still gives us some indication of the pre-crisis direction of the economy. This was already showing signs of weakness, although not any obvious recession signals. Wage growth, in particular, had slowed, from a peak of 3.5 percent year-over-year hit last February, July, and August, to just 3.0 percent in February of this year. This weakening of wage growth is surprising given an unemployment rate well under 4.0 percent.
There was a small pickup to 3.1 percent in March, but this was likely a compositional effect, as we disproportionately lost workers in the low-paying restaurant and retail sectors. There is also a within industry compositional effect, as the lowest-paid workers are the ones most likely to be laid off first. This will almost certainly lead to a jump in reported wage growth in April.
Virtually all the data in the household survey is bad, which is not surprising, given the overall numbers. Women were hit harder than men, with prime-age (25 to 54) EPOP falling by 1.1 percentage points for women, compared to 0.6 percentage points for men. This gap will grow next month, as we see the tidal wave of layoffs in retail and restaurants disproportionately eliminate jobs held by women. Involuntary part-time employment jumped by 1.4 million, a 33.5 percent jump. Voluntary part-time fell by 1.6 million, as jobs offering part-time employment in retail and restaurants disappeared. Not surprisingly, the share of unemployment attributable to voluntary quits plunged from 13.4 percent to 10.5 percent.
The March employment data is likely worse than most had expected, but this is simply because the impact of the shutdowns is showing up somewhat earlier than many of us had anticipated. We know with almost certainty that the April data will be far worse, in fact, the worst we have ever seen. It is important to put these terrible numbers in context. We are keeping people from working to slow the spread of the coronavirus, so we should not be upset by data showing that people are not working.
The issue for this period of shutdown is ensuring that people can be kept whole as much as possible, that they can pay for their rent, food, and other necessities. Then, once the virus is contained, we have to make sure that we can get people back to work quickly.