In May, the unemployment rate declined to 13.3% from 14.7% in April. That improvement was welcome news, but aside from April, the 13.3% unemployment rate in May was higher than anything we’ve seen since the Great Depression.

And, the unemployment rate is not reflecting all coronavirus-related job losses. In May, 21 million workers were counted as officially unemployed. But there were another 4.9 million workers who were out of work because of the virus but who were being misclassified—they had been furloughed and should have been counted as unemployed and on temporary layoff, but were instead counted as “employed but not at work.” If those workers had been correctly counted as being on furlough instead of employed, the unemployment rate would have been 16.4% in May instead of 13.3%. (This misclassification has gotten a lot of attention, with some suggesting that the Bureau of Labor Statistics (BLS) is cooking the books. That is not the case; BLS was extremely transparent and consistent with their treatment of the data.)

Another group of workers who are left out of the official unemployment rate are those who are out of work as a result of the virus but are not actively seeking work. As is always the case, for a jobless worker who is not on furlough to be counted as unemployed, they must be actively seeking work. Today, that remains impossible for many. I estimate that in May, there were 6.6 million workers who were out of work as a result of the virus but who were being counted as dropping out of the labor force because they weren’t actively seeking work. I arrive at that figure by assuming that the February labor force participation rate (63.4%) is what the May labor force participation rate (LFPR) would have been if the pandemic had not occurred. (Note, this is likely somewhat understating what the LFPR would have been in May without the pandemic since the LFPR had been rising slightly in the year prior to the virus—from 63.1% in February 2019 to 63.4% in February 2020.) Then I multiply the February 2020 LFPR by the May population (ages 16+) to estimate what the size of the labor force would have been in May if the coronavirus hadn’t happened, yielding 164.8 million. I then subtract from that the actual May labor force, 158.2 million. The result, 6.6 million, is the number of workers who would have been in the labor force in May if the pandemic hadn’t happened but, instead, are out of work and not actively seeking work.

Putting that all together—the 21.0 million officially unemployed, the 4.9 million unemployed who are being misclassified as “employed, not at work,” and the 6.6 million who have dropped out of the labor force as a result of the virus—that’s 32.5 million workers who are officially unemployed or are otherwise out of work as a result of the virus. If they had all shown up as unemployed, the unemployment rate would have been 19.7% in May instead of 13.3%. That’s an improvement from the same figure in April, which was 23.5%, but it is still mind-bogglingly high—nearly one in five workers are out of work.

The following chart provides these two unemployment rates—the official rate and the rate that takes into account the officially unemployed and others who are out of work as a result of the virus—by gender, race/ethnicity, education, and age.

The rates are incredibly high across the board, but job losses have been particularly stark among Black and brown workers. Historically higher unemployment rates and lower liquid savings make job losses even more devastating for Black workers and their families. Unemployment rates in May were also extremely high among women, and especially Hispanic, Asian, and Black women. Furthermore, the unemployment rate is higher for workers with lower levels of educational attainment, though even among those with advanced degrees, nearly one in 10 are either unemployed or out of work as a result of the virus.

The figure also shows that young people have been hit especially hard in this pandemic recession, with more than a third of 16–23-year-olds either officially unemployed or otherwise out of work as a result of the virus. Those graduating right now are in a particularly difficult situation, as they are experiencing extremely limited job opportunities but do not qualify for unemployment insurance, even under the expansive definitions of the CARES Act. Congress should institute a Jobseeker’s Allowance to address this issue.

Some are saying that because rehiring took place in May and the unemployment rate improved, perhaps more aid from Congress isn’t needed because workers will just return to their old jobs. This logic could not be more misguided. Of the 32.5 million workers who are either officially unemployed or otherwise out of work because of the virus, only 14.8 million workers (or 45.6%) can reasonably expect to be called back to a prior job, which means 17.6 million (or 54.4%) cannot. That calculation is the result of the following steps:

1. The survey the unemployment data are from asks unemployed workers who lost a job if they expect to return to work. Of the 21.0 million officially unemployed in May, 15.3 million expect to be called back to their old job, which means 5.6 million of the officially unemployed do not expect to be recalled to a former job (numbers don’t sum to the total due to rounding). Further, I assume that all of the 4.9 million unemployed who are being misclassified as “employed, not at work” expect to be called back to their old job, and that none of the 6.6 million who have dropped out of the labor force as a result of the virus expect to be called back. That means that of the 32.5 million workers who are either officially unemployed or otherwise out of work because of the virus, 20.2 million expect to be called back, and 12.2 million do not.
2. However, it is likely that many of those who expect to be called back to their jobs will find that those furloughs have turned into permanent layoffs. First, BLS adopted an unusually broad definition of temporary layoff with respect to the coronavirus. Typically, to be classified as unemployed on temporary layoff, a person has to either have been given a date to return to work by their employer or expect to be recalled to their job within six months. However, if an unemployed person expects to be recalled to their job but, because of the coronavirus, they are not sure whether they will be recalled within six months, their response was counted as a “yes” (they expect to be called back within six months), rather than “don’t know.” BLS reports that “this unusual step was taken as part of an attempt to classify people who were effectively laid off due to pandemic-related closures among the unemployed on temporary layoff.” (See Question 8 here.)Further, it is likely that many workers who were explicitly told they would be called back will ultimately find that their temporary layoff was actually a permanent layoff. Businesses may have told the workers they were laying off that they were going to call them back even if they were unsure they would be able to, because it was an easier message to deliver. Or, businesses may have felt confident that they were going to call back their workers at the time of the job separation but underestimated the financial strain they would face and will end up either going out of businesses or only needing to call back a fraction of the workers they furloughed.

   Historically, only 71.7% of workers on temporary furlough return to their jobs. On the other hand, 13.4% of laid-off workers who expect that their layoff is permanent are actually recalled. (These shares are found in Table II and the surrounding discussion in a 1990 paper by Larry Katz and Bruce Meyer, “Unemployment Insurance, Recall Expectations, and Unemployment Outcomes.”) It is unclear the extent to which this historic experience will be applicable in the current crisis, but these estimates nevertheless provide a rough sense of what we might expect. BLS data show that 2.3 million unemployed workers report being permanently laid off (unemployed workers who were not on temporary layoff but, for example, were not laid off because they are new entrants or reentrants to the labor market, or they voluntarily quit their job, are not included in this figure). If 13.4% of the 2.3 million unemployed workers who report being permanently laid off are in fact recalled, that is 308,000 workers who thought they were permanently laid off who will ultimately be recalled. Subtracting 308,000 from the 12.2 million calculated in step 1 that do not expect to be called back to a prior job yields 11.9 million workers who have essentially zero chance of being called back.

3. If, as described in the prior step, only 71.7% of workers who do expect to be called back are called back, that means 28.3% will not be called back even though they expect to. That in turn would imply that of the 20.2 million workers who are out of work but expect to be called back to their prior job, 5.7 million will not be called back, and 14.5 million will. Adding to that 14.5 million the 308,000 workers calculated in the previous step who thought they were permanently laid off who are likely to get recalled, that is 14.8 million workers who can reasonably expect to be called back.
4. Putting this all together means that of the 164.8 million workers who are either in the labor force or who have dropped out of the labor force as a result of the virus, 11.9 million workers, or 7.2%, are out of work with no hope of being called back to a prior job; 5.7 million workers, or 3.5%, expect to get called back to work but likely will not; and 14.8 million workers, or 9.0%, may reasonably expect to be called back. In other words, even if all workers who can reasonably expect to be called back to their prior jobs were called back, the share of the workforce out of work would still be 10.7% (7.2% plus 3.5%), higher than the highest unemployment rate of the Great Recession.

The following chart provides, overall and by demographic, the breakdown from step 4 (namely, the share of the workforce that is out of work with no chance of being called back to a prior job, the share of the workforce that is out of work and expects to get called back to their prior job but likely will not, and the share of the workforce that is out of work and can reasonably expect to be called back). The sum of the first two segments in each bar is the share of the workforce that is out of work and has no reasonable expectation of being called back to work. Overall, that figure is 10.7%, and is extremely high across the board, but it is particularly high among Black and brown workers, women, and especially Hispanic, Asian, and Black women. It is also particularly high among young workers and among workers with low levels of education.

It’s worth noting that the prospect of even those who can reasonably expect to be called back to a prior job actually getting called back will require Congress to act. If Congress provides enough aid to state and local governments, individuals, and businesses so that confidence and demand are high as the economy reopens, and puts in place public health measures and child care measures that make a successful reopening possible, then those furloughed workers will likely be needed by their former employers to meet demand and will be called back. But if not, many won’t, and we will face sustained, extremely high unemployment.

Congress needs to extend the unemployment provisions in the CARES Act, including extending the across-the-board $600 increase in weekly benefits. Opponents of the policy say that the extra benefits disincentivize people from returning to work. However, cutting off the extra money so that people must exist on our very stingy regular state benefits will not work to incentivize people to return to work when safe jobs aren’t available. Instead, what we will see is millions who can’t find work being forced to dramatically cut back their spending. That will not only cause an enormous amount of human suffering, but it will also hamstring the recovery. We estimate that allowing the $600 to expire at the end of July will cost us 5.1 million jobs over the next year.

We must also provide a huge amount of aid to state and local governments. Without it, the recovery from this recession will be delayed for years. We estimate that state and local governments face budget shortfalls on the order of $1 trillion by the end of 2021 and if the federal government doesn’t fill that in, it will cost 5.3 million jobs—in both the public and private sector—by the end of 2021.

And we can’t turn off federal relief too early. The expiration of relief provisions should be tied to economic conditions. Assigning arbitrary end dates to provisions to sustain the economy makes no sense when we could easily have provisions phase out as the unemployment rate or the employment-to-population ratio are restored to near pre-virus levels. Using automatic stabilizers would not be any more expensive than the cumulative cost of multiple extensions—but it would prevent destructive lapses in critical programs while Congress negotiates extensions, and it would alleviate corrosive uncertainty by giving businesses, states, localities, and households confidence around budgeting and planning. And if projections are wrong and jobs roar back until we are at near-pre-recession levels much more quickly than expected? That would be great news, and with automatic stabilizers, aid would simply turn off because it would no longer be needed.