Monday, June 29, 2020

Gordon's Guesstimate: April Jobs


Gordon's Guesstimate: April Jobs

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

The highest monthly job loss on record is over 1.9 million, which occurred in September of 1945, marking the demobilization from World War II. The largest single-month increase in the unemployment rate occurred in October of 1949, when the unemployment rate jumped 1.3 percentage points from 6.6 percent to 7.9 percent. When the employment report is released at 8:30 this morning, there is reason to suspect that these historical “superlatives” will be exceeded by an order of magnitude.

But we already know that. The grim predictions above are safe assumptions, resting on higher frequency UI claims data, which ordinarily cannot be relied upon to paint a telling picture of the labor market. Unemployment claims data typically only pick up layoffs and are blind to other critical  flows in the labor market, most notably hiring. Before the pandemic, these flows typically netted to tens of thousands to low-hundreds of thousands in a given month. But the scope and scale of the necessary shuttering of the economy is such that those net positive hires will be swamped by gross job losses. Since the survey that informed last month’s job numbers, 26.5 million new unemployment claims were filed – and last month’s numbers were still near historic losses at the time.

This is not to say that hiring has completely ceased. Some research suggests that despite significant net job losses, there remains as many as 3 new hires for every 10 layoffs. Even if that ratio holds, it still suggests a net employment loss of 18.5 million – or nearly 10 times the highest single-month loss previously recorded.

Divining the unemployment rate is more of a challenge and largely hinges on the disposition of the unemployed. For context, assuming that unemployment jumps 26.5 million and that, implausibly, there is no decline in the labor force, unemployment would exceed 20 percent. The labor force is the sum of the employed and the unemployed; the unemployment rate is the percentage of the labor force that is unemployed. But to be considered unemployed, one must be looking for work. Policymakers correctly expanded the eligibility for unemployment benefits beyond those laid-off. Accordingly, some of these workers receiving benefits may not show up as unemployed, but rather as having left the labor force. Today’s report will show some of both. Before last month, the single-largest monthly decline in the labor force was just over 1 million and occurred in October of 2013. That record will be smashed as well.

Confounding the measurement of the devastation is the nature of measurement – survey responses. These surveys are highly robust, and the results are statistically cleansed. But as the Bureau Labor Statistics (BLS) notes, “If there is a substantial departure from historical patterns of employment changes in net business births and deaths, as occurred from 2008 into 2009 during the 2009 benchmark, the model's contribution to error reduction can erode.” BLS has stated that they will update the model to account for the recent and aberrant labor market movements. But from a practical perspective, employers and individuals are experiencing a disruption that is by some lights an order of magnitude worse than anything we’ve ever seen. This is to say that today’s first real look into the labor market devastation caused by COVID-19 will evolve, and its accuracy should be considered accordingly.

For a guesstimate, the magnitudes of these losses are so large that estimating these indicators to any degree of certainty would not be credible. Nevertheless, I’m estimating that the BLS will report that in April 23 million Americans lost their jobs, and 15.5 percent of the labor force was unemployed. For earnings, the composition of the labor market – concentrated job losses among lower-wage earners – may push up the earnings print. This guesstimator is assuming an 11-cent gain for a 3.5 percent year-over-year gain.

No comments:

Post a Comment