The May jobs report issued by the Bureau of Labor Statistics on Friday contained quite a surprise. In the midst of the pandemic, the unemployment rate actually decreased from April’s 14.7 percent to 13.3 percent. That sounds like good news.
Two factors could explain the unexpected improvement. The first is the easing of stay-at-home orders in several states, most notably Texas and Florida. The second is the growing use of the Paycheck Protection Program, which provides loans to companies that pledge to retain most or all of their workers.
The good news looks even better when one looks at the details. One might have expected employment to rebound noticeably in areas directly clobbered by the crisis, such as leisure and hospitality, retail and construction. These are the sectors that should benefit from easing restrictions and the PPP. But the uptick in employment went beyond that: Manufacturing employment, along with education and health, ticked upwards as well.
At the same time, the data come with a greater-than-usual level of uncertainty. In its press release, the Bureau of Labor Statistics provide a whole page of explanation of the unusual obstacles they faced in collecting data.
Because of the pandemic, the BLS was not able to conduct its usual face-to-face interviews, and it had to operate with somewhat limited staff in its four regional data collection centers. Due to this, the survey response rate was “slightly lower” than usual for establishments (businesses) and 15 percentage points lower (67 percent versus 82 percent) for households.
Perhaps as a result, the BLS states clearly that they have detected numerous cases of miscoded data. In particular, workers still employed but absent from work due to business closures should be classified as unemployed, but frequently are reported as employed. The BLS even gives a specific estimate, stating that the unemployment rate could be 3 percentage points higher if these coding errors were all detected and corrected.
As an economist, I spend a good deal of time thinking about how accurate data are and how well they represent the economic reality we need to understand. In this case, my main observation would simply be: The data are shaky at best. It is never wise to make strong inferences based on a single data point. These unemployment data are an unusually compelling illustration of the general rule.
Whether the real, true unemployment rate should be 13.3 percent or 16.3 percent, the unemployment rate is higher than at any time since the Great Depression.
Looking forward, I would caution against interpreting these numbers as signs that the COVID recession will be short-lived. The major emergency income-support programs last only until July. And state governments, obliged to keep balanced budgets by state constitution in most states, are currently scrambling to cut costs. Paradoxically, in the midst of the pandemic, some hospitals have already started to lay off staff in light of the funding crisis, as are schools.
The stimulus bills passed thus far have propped up the markets for state and local government bonds. This has allowed many localities to raise cash through bond sales. However, with balanced-budget clauses hard-wired, this borrowing can only raise short-term cash; it cannot eliminate the yawning deficits that have opened up.
Another major concern going forward is that large numbers of small businesses will be unable to weather the economic storm. Many businesses have either been rejected by the Paycheck Protection Program or did not apply. For those that cannot get PPP loans and do not have access to other credit or very large financial reserves, failure will be likely. Costs such as rent, utilities and mortgage payments will pile up during the months of complete and partial shutdown, while many businesses will find revenues greatly depressed. Small business failures will reverberate throughout communities, greatly hampering recovery.
Additionally, without measures to reassure workers that they will be safe when they return to work, it is hard to imagine the return going rapidly and smoothly. Economists such as former World Bank Chief Economist Paul Romer have been urging massive funding for virus testing and contact tracing; without this, workers face the nasty dilemma of refusing to go back to an unsafe job and ending up fired for cause, ineligible for unemployment benefits or working in an unsafe environment.
Proposals to insulate businesses from liability for health hazards at work, advanced by Senate Republicans, lower the risks for businesses, but would lay all of the risk on workers. I find it hard to imagine that workers will accept work under these circumstances, even with unemployment so high. This dilemma needs to be resolved at the national level.
Economic recovery certainly hinges on advances in the detection, treatment and prevention of the virus. However, this month’s data give us some hope that pro-active policies can cushion the economic blows. It also raises concerns about what would happen to many working families if the extraordinary programs created by Congress are not extended when they expire in late July.
Substantial support for state and local governments, along with extension of some of the safety-net programs already funded by Congress, will be needed to avoid another round of layoffs. Aggressive expansion of testing and contact tracing may well be the next key step to allowing people to work without putting their health at extreme risk. The risks of new waves of the virus, and returns to stricter limits on economic and social activity, will remain significant in any case.
In short, the good news from the jobs report is just a first step on the long and winding road to economic recovery.
Evan Kraft is the economist-in-residence for the economics department at American University. He served as director of the research department and adviser to the governor of the Croatian National Bank.