How did the government misclassify millions of workers in measuring employment since the beginning of the pandemic?

To measure the unemployment rates, each month the U.S. government (and specifically the Census Bureau) surveys a representative sample of households and asks them questions about their employment status: whether they are employed, not employed and looking for a job (i.e., unemployed), not employed and not looking, temporarily absent from work, retired, and so on. During the unusual circumstances of the last months, Census interviewers have classified a significant portion of interviewed people whose employers had shut down because of the pandemic as “temporarily absent from work” for “other reasons”—and therefore employed—when they should have probably been counted as “temporarily laid off”—and therefore unemployed. This has likely caused the unemployment rate figures reported by the Bureau of Labor Statistics (the government agency which discloses current employment statistics each month) to be lower than they should have been. 

Was the improvement in May jobs numbers an indication that the employment market is starting to improve in a meaningful way?

Even correcting for the mistake, the unemployment rate was lower in May than it was in April (the reported figure is 13.3%; the corrected one 16.3%). In this respect we can say that the labor market has shown some slight improvements in the last month. It should be kept in mind, however, that this is the highest rate the U.S. has seen since the Great Depression. Only last week, another 1.5 million workers filed for unemployment insurance and another 700,000 for the new Pandemic Unemployment Assistance program. It is hard to predict what will happen but it is also hard to envision a speedy recovery: both the Federal Reserve and the OECD forecast unemployment rates as high as 9% throughout the rest of the year (as a comparison, unemployment peaked at 10% right after the Great Recession of 2007-2009).  

Does the unemployment number give a good sense of the experience of workers under ordinary circumstances?

“The unemployment rate doesn’t count people who are ‘discouraged’ and have stopped searching. These people might end up representing a non-trivial share of the population if the economy is bad for a long period of time.”

The unemployment rate measures the share of the workforce who does not have a job at a given point in time but is looking for one. Under ordinary circumstances it is a pretty good indicator of workers’ experience in the labor market. For example, it is usually low when the economy is good and there are plenty of jobs. When this is the case, not only is there room for most people who want a job to find one but people are also more likely to be able to choose a job that is a good “match” for them. Conversely, the unemployment rate is high when the economy is sagging and jobs are scarce. One caveat about the unemployment rate is that it doesn’t count people who are “discouraged” and have stopped searching. These people might end up representing a non-trivial share of the population if the economy is bad for a long period of time.