By Louise Sheiner, Gian Maria Milesi-Ferretti

Most economic forecasters expect very strong GDP growth in 2021 as the $1.9-trillion American Rescue Plan works its way through the economy and more people are vaccinated. The consensus is that real GDP will increase about 6% between the fourth quarter of 2020 and the fourth quarter of 2021. If this growth materializes, output by the end of 2021 will be about back to the level it would have been had the pandemic not occurred (using the Congressional Budget Office’s pre-pandemic 2020 economic projection to gauge that level).

This will not be another jobless recovery. If the GDP forecasts prove accurate, we estimate that monthly payroll employment gains over the next 10 months will average between 700,000 and 1 million per month, a lot faster than many forecasters anticipate.

The pandemic recession was unlike any previous recession. The U.S. lost almost 9½ million jobs between February 2020 and February 2021, many more than one would have anticipated based on past episodes of declining GDP. The lost jobs were disproportionately low-wage, low-productivity jobs. The post-COVID recovery will be different as well. Jobs are likely to return at a faster pace than ever seen before on a sustained basis.

What factors will determine the pace of employment gains?

At one extreme, the economy could return to its pre-pandemic employment trajectory, reversing all the employment losses that have occurred during the pandemic. In that case, the level of employment at the end of 2021 should be about equal to CBO’s January 2020 projection of employment in 2021 Q4, which would require adding 11 million jobs over the remainder of the year. (This represents the 9½ million jobs lost since the start of the pandemic plus the normal growth of employment.) That would mean adding 1.15 million jobs per month, on average, over the next 10 months.

A more likely scenario is that the post-pandemic economy will differ from the pre-pandemic economy in some ways. Some jobs lost during the pandemic may be lost for good because firms have learned new ways to operate. Some jobs may take longer to return—some health-related restrictions may remain in place this year, and it may take time for firms to reopen and rehire even once it is safe to do so. So not all the increase in GDP is likely to come from the reopening of suppressed sectors of the economy. Instead, some of it will reflect a more broad-based increase in demand stemming from the large amount of fiscal stimulus. This increase in demand is likely to be associated with a more “usual” relationship between increased GDP and employment.

To gauge what the employment gains would be for a more normal boost to GDP from fiscal stimulus, one can use a variant of Okun’s law (devised by the late Brookings economist Arthur Okun) that relates employment changes to GDP changes. According to Ball, Leigh, and Loungani (2013), a 1 percentage point increase in real GDP growth is usually associated with a 0.54 percent increase in employment growth. Assuming real GDP increases 6% between the fourth quarter of 2020 and the fourth quarter of 2021, this would imply an average monthly job increase of 460,000 jobs per month over the next ten months—which would mean less than half of the jobs lost as a result of the pandemic would be recovered by the end of the year.

Why are these two estimates so different?

Employment losses caused by the COVID-19 recession have been unusually concentrated in low-wage jobs, as shown in the figure below (taken from the Fed’s semi-annual report to Congress). Because low-wage jobs are generally low-productivity jobs (in a national accounting sense), the job losses have been unusually large relative to the GDP losses. As these jobs come back, there will likewise be an unusually large increase in employment relative to the gains in GDP. In other words, because the workers who lost jobs produced less than the average worker, more of them would need to be hired to make up the lost output than would be suggested by Okun’s law, which is based on the historical experience, where the workers who lose jobs have wages closer to the average.

Click figure to enlarge.

Which jobs will or won’t come back?

The figure below shows an estimate of the share of missing jobs by industry. We compare actual employment growth by industry to a counterfactual of what employment growth would have been in the absence of the pandemic; the counterfactual assumes that growth rate of employment by industry would have been similar to that observed in 2019.^[1]

Employment losses in leisure and hospitality account for 33% of all missing jobs, followed by professional business services (8.8%), health services (7.5%), and local education (6.4%).  Diving into the details of the decline in professional and business services, we find that the declines are concentrated in services related to travel and facilities and business support.

Click figure to enlarge.

To get a sense of which jobs might not come back, consider which sectors of our economy are likely to take longer to recover and which sectors may never fully recover. There is a lot of uncertainty.

Take leisure and hospitality. With other countries not vaccinating as quickly as the U.S., it is possible that activities related to tourism by foreigners (and services provided in the U.S. to Americans who want to travel abroad) might remain depressed in 2021. According to the Bureau of Economic Analysis, foreigners account for 14% of spending on domestic tourism. But following a year of little travel, increased demand by Americans could make up for this.

In other industries, the uneven pace of vaccinations around the world means COVID is likely to continue to affect some supply chains. This could slow the pace at which employment in some industries hard hit by the pandemic recover.

It could take time for day care, camps, and schools to become fully operational. Vaccines for children have not yet been approved, and it is possible that schools won’t be back to normal for a while. If so, some of the employment losses caused by a decline in labor force participation of women due to childcare responsibilities may not recover right away. In addition, some of the large job losses in local education and childcare (under social assistance, in the chart above) could take time to reverse.

In other sectors, the pandemic will have permanent effects that will alter the composition of employment and GDP. Working from home will be much more common, which means fewer commuters in central business districts, resulting in a decline in demand for some restaurants, transit services, professional services (office management and general office services), and commercial construction (office buildings).

Another long-lasting effect of the pandemic maybe a reduction in business travel. Businesses account for about one-quarter of total tourism dollars, so a transition to fewer in-person meetings and conferences could slow the recovery of employment in travel and leisure and hospitality services. On the other hand, even if demand declines in the long run, the urge to make up for trips missed in 2020 may mean that business travel is robust in the second half of 2021.

Finally, in some sectors that have experienced large losses during the pandemic, total recovery seems likely. Other than possible delays in reopening schools, employment by state and local governments is likely to fully recover given the very generous federal assistance provided over the past year. Similarly, demand for healthcare in coming months is likely to be at least as high as it would have been pre-pandemic, given all the delays of elective procedures during the pandemic.

The bottom line is that most—though not all—of the job losses that have occurred since the start of the pandemic will likely be refilled over the coming year.

If 80% of the GDP growth comes from the unwinding of the effects of the pandemic, and 20% reflects a broader-based increase in demand, then employment gains will average about 1 million per month over the next ten months. If only 50% of the GDP growth is accounted for by the unwinding of the pandemic and the other 50% by an increase in broad-based demand—which seems somewhat pessimistic about the ending of the pandemic—then employment gains would still average about 800,000 jobs per month.

While those numbers seem extremely large, keep in mind that sectors most affected by the pandemic are the most employment intensive. If, for instance, employment in the hard-hit leisure and hospitality sector recovers to 90% of its pre-pandemic trend by the last quarter of 2021, this alone would add some 240,000 jobs a month between March and December.

What role does productivity play?

In the scenario where the recovery fully reverses pandemic-related job losses, we are implicitly assuming that productivity—output per hour of work—reverts to its pre-pandemic trajectory. Of course, it is possible that the pandemic and its recovery will affect productivity, but the direction of the effect, if any, is unclear.

Fernald et al (2021) suggest that productivity growth is likely to be somewhat lower following the pandemic, reflecting the fact that much of the capital investment during the year was simply “duplicating” idle office equipment and that there may be costly adjustments to learning to work from home. It is also possible that the need for social distancing and for additional safety measures (new HVAC systems, for example) isn’t fully gone by the end of the year, which may hold down productivity. A lower level of productivity than anticipated pre-pandemic would imply an even higher pace of job creation than calculated above.

On the other hand, the pandemic could lead to higher productivity. For example, substituting Zoom for a lot of business travel and commuting will save countless hours of travel and could boost productivity. Similarly, the pandemic may have induced more automation with an associated increase in productivity. It’s also possible that a surge in demand in some sectors will increase utilization of resources: every restaurant table, theater seat, and plane seat could be full, physicians may schedule more surgeries to make up for those delayed, airlines may schedule more flights, etc. If this occurs, it may lead to temporarily longer work hours and higher productivity. These effects would reduce the number of payroll jobs created for any given level of GDP growth, making the lower end of our estimated range of employment gains more plausible.

What about the unemployment rate and labor force participation?

The unemployment rate—which measures those looking for and available for work—will surely decline as the economy improves, but the degree of the decline depends on the labor force participation rate. Those who aren’t looking for work don’t count as unemployed in the headline unemployment rate.

Labor force participation fell from 63.3% in February 2020 to just 61.5% at the end of 2020. This decline was much larger and much faster than typically occurs in recessions. The question is how quickly labor force participation will rebound.

In the “unwinding of the effects of the pandemic“ scenario, and again using CBO’s pre-pandemic January 2020 projections for Q4 2021 as a gauge, labor force participation would increase sharply throughout the remainder of the year, ending the year at 62.8%. (Because of population aging, LFP has been trending down over time.) Similarly, the unemployment rate would fall sharply over the remainder of the year, ending the year at 3.6%.

The more broad-based increase in demand scenario is associated with fewer new jobs—and a less than complete labor market recovery—which would leave the labor force participation below and the unemployment rate above their pre-pandemic trajectories. Ball, Leigh, and Loungani (2013) estimate that a 1 percentage point increase in real GDP growth is associated with a 0.4 percentage point reduction in the unemployment rate. Using this relationship, we would expect the unemployment rate to decline by 2.4 percentage points during 2021, ending the year at 4.3%. This would imply that the labor force participation rate would increase by almost 1 percentage point, ending the year at about 62.4%.

Using these estimates, if 80% of the projected GDP growth comes from the unwinding of the effects of the pandemic, and 20% reflects a broader-based increase in demand, the unemployment rate at the end of 2021 would be 3.7% and the labor force participation would be 62.7%. If the shares are 50/50, the unemployment rate would end the year at about 4% and the labor force participation rate at 62.6%.

Click figure to enlarge.

Are these huge payroll gains guaranteed?

Of course not. For one, this exercise assumes that the consensus GDP forecasts are correct. If social distancing remains in place for longer than anticipated, if there is another wave of lockdowns in the fall, if many people opt for not getting a vaccine or if they take time to believe that vaccinations are making them safe, the recovery would likely be slower, lowering both GDP growth and especially employment growth, since the high-employment contact-intensive sectors would be the most affected. And, of course, even if the GDP forecast is correct, there is a wide range of possible employment outcomes depending on the extent to which the GDP growth stems from the unwinding of the effects of the pandemic.

How much growth in payrolls are other forecasters projecting?

The figure below shows the distribution of forecasts from the Wall Street Journal’s March 2021 Economic Survey, where each dot represents a particular forecaster’s projection of GDP and employment. These forecasts are broadly consistent with Okun’s law. Indeed, a regression of the percent change in payroll growth on the GDP growth predicted by Wall Street forecasters yields a coefficient of 0.56—just a tad above the Okun’s law coefficient of 0.54 we used for our “broad-based” employment scenario. The employment gain associated with a projection of 6% GDP growth using the line of best fit in the graph is just 506,000 per month—well below the numbers we discussed above. These forecasters implicitly expect the 2021 recovery to resemble that from previous recessions even though the recent pandemic-induced recession was quite different from recessions past.

Click figure to enlarge.

^{[1]In particular, because CBO expected overall job gains in 2020 to be about 75% of those in 2019, we assumed in the counterfactual that the employment increase by industry in 2020 was 75% of the 2019 employment increase by industry.}

Read More