The United States added 266,000 jobs in April while the unemployment rate rose slightly to 6.1 percent with the realistic unemployment rate, which adjusts for misclassification and the unusual decline in labor force participation, falling to 7.6 percent as more people entered the labor market. At the same time hourly earnings rose 0.8 percent for production and non-supervisory workers, faster than any pre-pandemic monthly increase since the early 1980s, and an even higher 0.9 percent when adjusted for changing industry composition.
The economy was still 10 million jobs short of its pre-pandemic trend in April with the employment rate down 3.2 percentage points since February 2020. Even though the pace of hiring was disappointing in April, the labor market has still been behaving as if there was relatively little or even no slack left: Openings were at record levels, quits were near record levels in March, composition-adjusted wages were growing at the same pace they did in the relatively tight 2019 labor market with the largest wage gains for the lowest-wage workers, hourly earnings growing rapidly in April, and average weekly hours remain very high.
Normally, a labor market with this little apparent slack would only get tighter as the unemployment rate falls. This is because the supply of labor is usually relatively stable, so a falling unemployment rate is indicative of stronger labor demand. The current situation is different and perhaps unprecedented because there is good reason to expect further large increases in both labor supply and labor demand. To the degree supply and demand are mismatched there are risks of an incomplete jobs recovery and/or a substantial increase in price inflation, at least in the near term, with possible longer-run repercussions. But if labor supply and demand are well matched, a rapid return to high levels of employment with faster real wage gains could be the most likely scenario.
With so many conflicting signals as the labor market changes rapidly with demand and supply returning to different degrees in different sectors, it is hard to make a confident assessment from the data. Moreover, the data we have is often outdated by the time it is reported. After all, the data from today's employment release reflect the labor market of mid-April. Given the speed with which the labor market is changing, quite a lot could have happened since then. Continuing to process and synthesize the conflicting signals is essential.
The labor market has made rapid progress but remains well short of where it was
The US economy lost 22 million jobs in March and April of 2020. It has since gained back 14 million jobs. This still leaves jobs 8 million short of their level in February 2020, and 10 million short of their pre-pandemic trend, as shown in figure 1. That is the clearest sign that the labor market still has a substantial amount of room for growth.
The official unemployment rate has fallen from its high of 14.8 percent in April 2020 to 6.1 percent last month, as shown in figure 2. But 2.0 million more people than usual have left the labor force; if they were counted as unemployed, along with workers that were misclassified (as employed but not at work), the unemployment rate would be 7.6 percent. Both of these rates are well above the 3.5 percent level prior to the COVID-19 recession. Another concept of the unemployment rate cited by Federal Reserve Chair Jerome Powell and Treasury Secretary Janet Yellen, which assumes the participation rate was fixed at its February 2020 rate, was 8.9 percent in April (this concept is not comparable to the historic unemployment rate). (The recent history of these alternative measures can be found in the appendix table at the bottom of this post, and full details of the calculations are available here.)
All of these indicators are strong evidence that the labor market has a ways to go before it is healed. The question is what form this adjustment will take and what the risks are.
The labor market has been hot in important respects
The labor market has not looked like what one would normally expect in a labor market with 10 million missing jobs and an elevated unemployment rate. Judged contemporaneously based on movements in prices, wages, and labor market flows, the labor market in April appeared to have about the same amount of slack (defined as demand falling short of supply) as it had in the 2017 to 2019 period. The economy was near full employment then, with Chair Powell assessing it was somewhat below full employment, while the Congressional Budget Office (CBO) thought it was above it.
In a normal recession and recovery, the constraint on employment is lower demand for workers, but in recent months the demand side of the labor market appears to be recovering more quickly than the supply side, as businesses are increasingly posting job openings while employees have been relatively slow to return to the labor force.
One piece of evidence corroborating the slow return of individuals to the workforce is the behavior of the labor force participation rate. As shown in figure 3, even with recent increases, it has only just gotten back to its peak last summer.
Some of the decline in labor force participation since the pandemic may reflect people who do not want to re-enter the workforce because of the lack of availability of jobs. But the fact that it has remained low in the face of increasingly rapid job creation and record job openings suggests that there is more to the story. Two interrelated possibilities:
- The most important factor is likely that people continue to have concerns about workplace safety after high levels of infection and death in some frontline occupations. In all, 4 million people are reluctant to return to work because of virus fears according to the Census Pulse survey.
- Increased unemployment insurance has resulted in unemployment benefits being more valuable than wages for roughly half of the unemployed. This does not appear to have been a major factor in the labor market in 2020, but labor market conditions are completely different now. There is no clear evidence yet, but historical experience, theory, and commonsense all suggest it is one factor in labor supply—including giving people who feel unsafe returning to work more options. (Note, arguments that unemployment insurance benefits increase the bargaining power of the unemployed are equivalent to saying they reduce labor supply at any given wage, a reduction in the labor supply curve.)
Childcare does not appear to be a factor holding back labor supply, as employment rates for parents of young children have not fallen meaningfully faster than for similar people without young children.
The most striking indicator of the hot labor market was the rate of job openings, which hit a record level in February according to the Bureau of Labor Statistics (BLS) data, which goes back to December 2000, as shown in figure 4. Job postings data from the employment website Indeed indicate job openings increased a further 18 percent by the end of April. Workers' confidence in the labor market is demonstrated by the near record quits rate in February showing that workers were not concerned about finding other jobs.
While job openings are plentiful there are also many more unemployed people than usual looking for jobs. The ratio of unemployed to job openings was about 1.1 for the officially unemployed and 1.4 for the realistically unemployed in April, roughly comparable to the state of the labor market in 2017 as shown in figure 5—but still with more slack than the labor market had immediately before the pandemic.
Wages are another way to assess the balance of supply and demand in labor. Wage growth was extremely strong in April with wages up 0.7 percent for the month, among the largest monthly increases in wages. Wage numbers from month to month can be volatile and are also affected by changes in the composition of workers (i.e., wages artificially rose when low-wage workers disproportionately lost jobs in 2020 and have artificially fallen as some of those lower-wage workers have returned to work). As a result it is better to look over a longer period at a composition-adjusted wage series, like the Atlanta Fed's wage tracker or the employment cost index, which both compare wage changes for the same people over time.
The most striking fact about labor markets over the last year has been that wage growth has held up at its pre-pandemic pace as shown in figure 6. The employment cost index, which also adjusts for composition, tells a similar story. Like wage growth pre-pandemic, the strongest wage growth in the Atlanta tracker has been for the bottom quartiles of workers (higher state and local minimum wages have contributed to this fact). Overall, if the only data we had about the economy came from wages, one would be hard pressed to know that labor markets were any looser than they were in 2019, despite employment remaining 10 million below trend. Moreover, these data are from March, and wages are inertial, so it is likely that the pace of wage increases will pick up in the coming months.
Finally, hours also tell a similar story, as average weekly hours for all private workers tied its record of 35.0 in April (data goes back to 2006) and average weekly hours for production and nonsupervisory workers were the highest they have been since 1998. Hours, like the Bureau of Labor Statistics' headline wage series, are affected by the composition of workers, so they should be interpreted with some care.
Labor supply and labor demand are both likely to increase in coming months; the question is how much and how quickly
The labor market recently has not had a very high amount of slack (defined as labor supply well in excess of labor demand) but still had employment levels very far short of where the economy could and should be. Looking forward there are good reasons to expect large increases in both demand for labor and supply of labor.
Both labor supply and demand will be boosted by the continuing reopening of the economy as more people are able to get vaccinated, and COVID-19 cases, hospitalizations, and deaths continue to fall. This will hopefully help close much of the remaining jobs gap.
Labor demand will get an additional boost as the fiscal stimulus continues to spend out and the monetary stimulus continues to be applied. These measures very plausibly will lead demand to exceed pre-pandemic projections, possibly as soon as the next few months.
Labor supply will also get an additional boost as expanded unemployment insurance expires (assuming Congress allows it to do so). Nevertheless, there are questions about whether labor supply might still end up at least slightly below its pre-pandemic levels. More than one-third of the decline in the labor force participation rate from the first quarter of 2020 to the first quarter of 2021 was due to the withdrawal of workers age 55 and up, some of whom may retire, either voluntarily or involuntarily, earlier than planned. Some people have made alternative plans or delayed schooling and will be leaving the labor force. There may be scarring among people who have been out of the work for over a year and have a hard time finding new jobs. Finally, in some cases, workers' reservation wages may have increased in response to the greater risks in some workplaces and the better outside option afforded by higher unemployment insurance benefits.
Three downside scenarios: Overheating, job-short recovery, and the return of the virus
One downside scenario is overheating if labor demand increases much more than labor supply. In this scenario, demand for consumption and investment returns much more quickly than the economy's capacity to produce it. The US economy experienced this in the first quarter when domestic final demand increased at a 9.8 percent annual rate. Nearly a percentage point of that increase in domestic demand was met from foreign producers, relieving some pressure on the US economy and exporting some of our relative strength to other economies that are, in many cases, struggling more than we are. But 2.6 percentage points of domestic final demand was met by running down inventories in the United States. Inventory-to-sales ratios were near historic lows in many categories, and they cannot be run down forever.
If domestic final demand returns to or even exceeds its pre-pandemic trend sometime in the second half of 2021, fulfilling that demand would require either the 10 million workers short of trend to get jobs extremely quickly, average hours to increase a lot, or productivity to jump up. Absent those changes, the result would be a substantial increase in prices—with prices rising more than a percentage point above their pre-pandemic trend, increases well in excess of base period effects.
The consequences of the overheating downside scenario depend on the magnitude of inflation, how it affects inflation expectations, and how the Federal Reserve responds to it. A burst of inflation could set back real wage gains for workers if prices rise more than wages, something that happened in the inflationary episodes in the 1960s and 1970s when price increases preceded nominal wage increases. Another risk is higher inflation expectations. Although they are well anchored around the Federal Reserve's 2 percent goal now and are likely to remain there, expectations are very correlated with actual inflation, so when actual inflation rises, expectations could too (this situation might not be terrible, and in fact could be an opportunity to raise the average inflation target or establish nominal GDP or gross labor income targeting). The worst scenario is a boom-bust cycle where either because of the Federal Reserve's policy, raising rates too high or too fast in an attempt to curtail inflation, financial market turbulence, or the dynamics of demand, the economy falls into a recession.
A second downside scenario is an incomplete jobs recovery. In this scenario demand returns to normal and supply is able to match it with a combination of longer hours and greater productivity, for example a shift to more successful businesses or more intense work. In this scenario real wages rise as labor demand exceeds labor supply at the current wage, leading wages to go up along with productivity. The Federal Reserve's Summary of Economic Projections and many other forecasts implicitly assume something like this scenario with output at or above pre-pandemic projections by Q4 of 2021 but the unemployment rate still substantially elevated.
The third downside scenario is that the virus itself takes a turn for the worse. Evidence to date indicates that existing vaccines generally remain highly effective against known variants of the SARS-CoV-2 virus. However, if a vaccine-resistant variant were to develop, the US and global economy could end up back in their 2020 situation.
The Goldilocks scenario: A hot summer that cools off to just right as we enter fall
The most likely outcome may be the Goldilocks scenario. In this scenario both demand and supply return. Patches of mismatch in timing and sectors would lead to noticeable shortages and price and wage increases in some areas, especially over the spring and summer as high demand is temporarily unable to fully be satisfied by available labor. However, these mismatches work themselves out with only transitory increases in the level of prices and no persistent changes in inflation or inflation expectations. Employment returns relatively quickly in the fall as unemployment insurance supplements end and the risk of COVID-19 from employment is greatly reduced, and the job match process has time to work out.
Which of these four stylized scenarios—or many other possibilities—will unfold will be hard to tell as we look through a foggy rearview mirror with imperfect data, all while hurtling forward through an unprecedented type of economic recovery. Nevertheless, there is no choice but to try to continue to understand the data as best as possible as it comes out.
|Alternative measures of unemployment
Feb. 2020 to Apr. 2021 (p.p.)
|Labor force participation rate
|Labor force participation rate
|Fixed participation rate
|Labor force participation rate
|Labor force participation rate
|Note: p.p. denotes percentage points. Change based on unrounded numbers.
Sources: Bureau of Labor Statistics via Macrobond; authors' calculations.
The data underlying this analysis are available here.