The monthly US unemployment rate has been 4.0% or lower since March 2018.

The idea that the US could sustain a level of unemployment this low was unexpected by mainstream economic forecasters.

For example, back in December 2012 when the unemployment rate was 7.9%, the Federal Reserve announced: "In particular, the Committee decided to keep the target range for the federal funds rate at 0 to 1/4 percent and currently anticipates that this exceptionally low range for the federal funds rate will be appropriate at least as long as the unemployment rate remains above 6-1/2 percent ..." The implication was that when the unemployment rate fell to 6.5%, that would be near the lowest level and it would be about time for the Fed to think about raising interest rates. But the unemployment rate fell below 6.5% by April 2014, and the Fed wasn't ready to start raising rates. It wasn't until December 2015 when the unemployment rate had fallen to 5.0% that the Fed started nudging interest rates upward.

Again, the implication in late 2015 and into 2016 was that when the unemployment rate hit 5%, it had fallen about as far as it was going to go. Here are a couple of illustrative figures from the 2020 Economic Report of the President, produced by the White House Council of Economic Advisers. The first figure shows the predictions from late 2016, from the Fed and the Congressional Budget Office, about that the decline in the unemployment rate was going to level out. But the unemployment rate just kept falling.

This figure shows the Fed's prediction in late 2016 for the rise in the total number of jobs in the US economy. The prediction was that the rise in jobs would level out, but the total number of jobs just kept rising.

Of course, Trumpophiles will credit this change to Trump administration policies, while the Trumpophobic will emphasize that the momentum toward lower unemployment rates and more jobs seems to have continued more-or-less uninterrupted from the pre-Trump years. Here, I want to sidestep the question of credit, and focus more tightly on what is actually going on behind these numbers.

An unemployed person can leave unemployment in two ways: either by getting a job, or by leavint the labor force so that they are no longer looking for a job. Similarly, person can enter unemployment in two way: either by losing a job and looking for a new one, or by re-entering the labor force and deciding to start looking for a job but not finding one right away. So, is the unemployment rate low because lower numbers of people are transitioning in, or because higher number are transitioning out, or some of both? Marianna Kudlyak and Mitchell G. Ochse look at how the patterns of these transitions have evolved over time in "Why Is Unemployment Currently So Low?" (Economic Letter, Federal Reserve Bank of San Francisco, March 2, 2020).

Here's one of their figures showing transitions out of unemployment (using averages over three-month periods). By historical standards, these transitions from unemployment to jobs, or unemployment to out of the labor force, do not look very different. Indeed, the rate at which the unemployed are shifting to employment is lower now than it was before the Great Depression or back in the 1990s. In other words, the very low unemployment rate doesn't seem to be occurring because the unemployed are leaving unemployment at higher rates than in the past. 

This figure shows transitions into unemployment. The red line shows that transitions from out-of-the-labor force to unemployment were high right after the Great Recession, but are now at low levels. The blue line shows, similarly, that transitions from employment to unemployment were high during the Great Recession, but are now at low levels.

What are the underlying reasons why transitions into unemployment may be lower now than in the past? Kudlyak and Ochse don't analyze this issue explicitly, but they highlight two possibilities: an aging workforce and better job matches. The intuition is that an aging workforce (along with a lower rate of start-up firms) means that people are more likely to remain in a current job, and only to move if they have the new job already lined up. The argument for better job matches is that employers have developed better tools for evaluating whether employees are a good fit--in terms of skills, personality, cognitive ability, and more--and so those who get hired into a job will on average stay longer. In other words, when the Fed and the CBO were incorrect in their forecasts of where the unemployment rate was headed, these factors were the ones they had not not taken into account.

A version of this article first appeared on Conversable Economist.