Blog Name

US unemployment remains worse than it seems as millions still out of the labor force

Jason Furman (PIIE) and Wilson Powell III (Harvard Kennedy School)



The labor market improved in February 2021 as employers added 379,000 jobs, leaving the economy at 11.9 million jobs below its pre-pandemic trend. At the same time the unemployment rate fell to 6.2 percent. Throughout the pandemic the official unemployment rate has been kept down by a misclassification error and the unusually large withdrawal of millions of people from the workforce. Our estimate of the realistic unemployment rate for February was 8.2 percent, a concept that adjusts for the unusually large decline in participation as well as changes in demography and is designed to be historically comparable to the official unemployment rate, which reached a maximum of 10.0 percent in the 2007–09 financial crisis. Another concept, the fixed participation rate unemployment rate, cited by Federal Reserve Chair Jay Powell and Treasury Secretary Janet Yellen, was 9.5 percent; the comparable concept peaked at 11.8 percent in the financial crisis.

The realistic unemployment rate fell slightly in February but remains elevated

The headline unemployment rate was 6.2 percent in February, down slightly from 6.3 percent in January. This concept works well in normal times but has had some deficiencies in the context of the pandemic. In response to these unique circumstances, we have been publishing monthly updates since June tracking what we call the “realistic unemployment rate” as shown in figure 1.

Figure 1 Alternative measures of unemployment

The realistic unemployment rate adds 757,000 workers who reported being “not at work for other reasons” as unemployed and also adds 2.6 million workers to the labor force to reflect the fact that the decline in labor force participation has been unusually large, even conditional on the overall economic weakness and adjusted for changing demographics (note that this is only a portion of the 4.2 million workers who have left the labor force since February 2020). This measure was 8.2 percent in February and is intended to be historically comparable to the official unemployment rate. The February value represents a continuation of January’s improvement, but at a slower pace. After falling rapidly over the summer, improvement has been more modest since the fall.

Recently many, including Yellen and Powell, have cited what we call a “fixed participation rate unemployment rate.” This measure fell from 9.7 percent in January to 9.5 percent in February. This concept is more straightforward than the “realistic unemployment rate” in that it simply assumes that all of the workers who have left the labor force since February 2020 should be treated as unemployed. An advantage of this approach is that it measures how much further the economy has to go until employment is back to where it was prior to the crisis. It has the disadvantage of not being comparable to official unemployment rates during previous recessions because, by construction, it is higher than the official unemployment rate. In the last recession the “fixed participation rate unemployment rate” peaked at 11.8. (Another downside of the “fixed participation rate unemployment rate” is that it does not adjust for demographic changes, as an aging population would be expected to reduce labor force participation, though this is a relatively small issue over shorter periods like a year.)

Since last summer we have also been tracking another concept, the “full recall unemployment rate,” that adjusts for workers on temporary layoff as well as the unusually large decline in participation. This measure is increasingly less relevant as the number of workers on temporary layoff has fallen, but we include it below for completeness. The full details on the calculation of the realistic unemployment rate and the fixed participation unemployment rate are available here.

Alternative measures of unemployment
Feb. 2020 to Feb. 2021 (p.p.)
  Unemployment rate 3.5 6.7 6.3 6.2 2.7
  Labor force participation rate 63.3 61.5 61.4 61.4 -1.9
  Unemployment rate 3.5 8.6 8.3 8.2 4.7
  Labor force participation rate 63.3 62.3 62.4 62.4 -1.0
Fixed participation rate          
  Unemployment rate 3.5 10.0 9.7 9.5 6.1
  Labor force participation rate 63.3 63.3 63.3 63.3 0.0
Full recall          
  Unemployment rate 3.5 7.0 6.9 7.0 3.6
  Labor force participation rate 63.3 62.6 62.6 62.5 -0.8
Note: p.p. denotes percentage points. Change based on unrounded numbers.
Sources: Bureau of Labor Statistics via Macrobond; authors' calculations.

As the unemployment rate remains elevated, the employment-population ratio is 3.5 percentage points below its pre-pandemic value

Last month, the employment-population ratio, the share of the civilian adult population that is working, increased slightly. Overall, the employment-population ratio has fallen from 61.1 percent in February 2020 to 57.6 percent last month. The change in the employment-population ratio accounts for both the change in unemployment and the change in labor force participation and thus fully captures the change in employment. Our alternative measures of unemployment capture the abnormally large increase in the number of people who have stopped looking for work, but, like the fixed participation unemployment rate, the employment-population ratio accounts for the entire decrease in labor force nonparticipation.

Figure 2 Employment-population ratio

Overall employment is 7.7 percent below its pre-crisis trend with increased hours per worker making up a small portion of that decline

The job market bottomed out in April 2020 but has made gains since then. The initial recovery was very rapid as workers were recalled to their previous jobs, but job growth has slowed in the last six months. As a result, today total nonfarm employment is 7.7 percent below its pre-pandemic trend, corresponding to 11.9 million jobs.

At the same time average weekly hours for private workers have risen from 34.4 hours in the year before the pandemic to 34.6 hours in February—a 0.7 percent increase. This is very unusual as average weekly hours generally decline in recessions. In part the increase in hours is a compositional effect as employers have shed lower-hour jobs, but it also appears as if hours have risen for many workers.

As a result, aggregate weekly hours have fallen 7.2 percent from its pre-pandemic trend. This is larger than what was likely about a 4 percent reduction in output relative to trend in February, likely because the job losses have been disproportionately in lower productivity sectors.

Figure 3 shortfall from trend


Understanding the data can help inform projections of the trajectory of labor market recovery. We saw an initial “partial bounce back” in the labor market in the late spring and summer of 2020, as the unemployment rate fell quickly at first as businesses reopened, but the pace of recovery has slowed throughout the fall, and temporarily reversed in December.

The future prospects of the labor market will depend on the trajectory of the virus, the future policy response, and how many people without jobs can quickly connect with their old jobs instead of undertaking the time-consuming process of finding a new job, or even a job in a new industry. Virus cases, hospitalizations, and deaths continue to fall while vaccination is picking up pace. The COVID-19 relief included in the December Consolidated Appropriations Act has contributed to rapid increases in consumption so far this year.

If the virus is brought under control and the US Congress passes something like the $1.9 trillion American Rescue Plan is passed, the economy would have sufficient aggregate demand to support rapid growth in 2021. This demand would be satisfied by putting millions of workers back to work and bringing millions more back into the labor force.

Data Disclosure

The data underlying this analysis are available here [zip].

More From