Gordon's Guesstimate: May Jobs

Authored by Gordon Gray, AAF’s Director of Fiscal Policy

According to last month’s Employment Situation report from the Bureau of Labor Statistics (BLS), 20.5 million Americans lost their job between March and April, surpassing any previous monthly decline by an order of magnitude. When the BLS reports its estimates for the month of May, it will detail a shallower deterioration than what was observed in April, but to levels not seen since the Great Depression. That’s the reality of putting the U.S. economy into quarantine. As the United States emerges from this hibernation, it is important to evaluate the performance accordingly.

Directionally, the intermonth changes to labor market indicators – net job losses, change in unemployment, others – will show that the United States will have lost millions of jobs and saw an uptick by previously historic margins in the unemployment rate. But it will also show a smaller deterioration than the previous month. The levels of joblessness and related labor market collapses could exceed those of the Great Depression. But what made the Great Depression so “great” and depressing was the persistence of those levels for years. To be sure, as the Congressional Budget Office recently forecast, unemployment will remain well above recent experience, but the economy should be on the road to recovery in the third quarter of this year.

But assessing the incremental progress has some challenges. First, the economic upheaval has introduced complications in the collection and interpretation of data. In an excellent and informative report, BLS identified a number of these – including the issue of firm births and deaths identified last month, lower survey response rates, and the impracticality of in-person interviews.

But one issue could swing the unemployment rate by as much as 5 percentage points. When workers were surveyed last month, millions of workers were misclassified and left out of the official unemployed rolls in the report. It is unclear how many were misclassified, but BLS has contemplated a scenario in which as many as 7.5 million workers were not counted as unemployed in the household survey. Their inclusion alone would add 5 percentage points to the level of unemployment. What is also clear is that revisions will be substantial. Additional follow-up surveys will reflect higher response rates and greater confidence in the estimate – given the low response rates and other challenges, revisions of substantial magnitudes in crosscutting directions should be expected.

If this morning’s report reflects a reclassification of those workers and a substantial share of the 12.4 million workers who filed for UI benefits between the April and May reference weeks, the unemployment rate could be north of the near 25 percent estimated in the Great Depression. This guesstimator is not quite so pessimistic and expects that May will show a decline in employment of 6.5 million, while U-3 will tick up to 22 percent. Hourly earnings should shed somewhat the compositional effect that perversely drove that measure up substantially in last month’s print, though not necessarily entirely – for this morning’s report I’m assuming earnings stay flat.