May's Employment Situation Summary report (also known as the "jobs report) from the U.S. Bureau of Labor Statistics (BLS) caused quite a stir.
The BLS reported employment rose by 2.5 million and the unemployment rate fell to 13.3% from 15.7% in May.
The stir was caused because other economic data showed U.S. employment fell in May and the consensus among economists was that May's unemployment rate would be significantly higher.
These stark differences led many to wonder if the numbers were intentionally tilted by the Trump administration to make the economy look better.
We've long worked with the BLS and can say with 99.999% confidence that the numbers weren't tampered with. The highly professional career civil servants at the BLS wouldn't allow this to happen. And if someone even tried, it would become public knowledge very quickly.
So we can safely dismiss statistical subterfuge. But that doesn't mean the BLS data is perfect.
The Washington Post's A ‘misclassification error’ made the May unemployment rate look better than it is. Here’s what happened. explains one issue with the May report. Key quote:
When the U.S. government’s official jobs report for May came out on Friday, it included a note at the bottom saying there had been a major “error” indicating that the unemployment rate likely should be higher than the widely reported 13.3 percent rate. ... The special note said that if this “misclassification error” had not occurred, the “overall unemployment rate would have been about 3 percentage points higher than reported ...”
There are other issues with how the BLS collects employment and unemployment data that lead to confusion about the rate of unemployment.
Most of these are covered in the New York Times article The Economic Pain That the Unemployment Rate Leaves Out. Key quote:
“The unemployment rate itself is significantly going to understate the drop in economic activity,” said Stephanie Aaronson, the director of economic studies at the Brookings Institution. “It also is underestimating the amount of distress that Americans are facing right now because it understates how many people have lost jobs.”
One of the major reasons the unemployment rate is understated (beyond the misclassification error listed above) is the BLS only consider those who who don't have a job but are actively looking for work as being unemployed.
But because of the COVID-19 shutdowns and health concerns, there are many who would like to have a job that are choosing not to look at this time. This means in the eyes of the BLS, they aren't unemployed. Instead, they are considered out of the workforce.
The New York Times estimates that adding this group of unemployed people back in would increase the unemployment rate by 3.7%.
So we have a 3% misclassification error and a 3.7% omission error. Adding them back means the unemployment rate is around 20%.
We think it's even a bit higher - in the 21%-23% range - when you adjust for other issues that lead to the undercounting of unemployment by the BLS.
None of this is meant as a criticism of the BLS or how they do their work. Measuring employment and unemployment on a monthly basis is challenging even in the best of times.
But the COVID-19 related shutdown of the economy means the BLS is faced with sampling, data collection and reporting issues they've never faced - or even anticipated facing. So problems are to be expected.
Despite these issues, the BLS data shows the extent of the economic problem we're facing.
The chart below (click to enlarge) is from the Calculated Risk blog. It uses BLS to data to show job losses in percentage terms for all the post World War II recessions.
As the chart clearly shows, even with May's numbers we have a long way to go to recover from the current downturn.