There are still many unanswered questions about the health and economic consequences of COVID-19. How many people have gotten sick? How much longer will the health crisis last?
And now, with concerns growing over COVID-19 survivors still suffering long after an illness, there are fresh questions: How many so-called “long haulers” are there, and how is the condition affecting their ability to work, their employers and the overall U.S. workforce?
Gopi Shah Goda, a senior fellow at the Stanford Institute for Economic Policy Research (SIEPR), has some answers. Applying rigorous science to the question, Goda and her coauthor, Evan Soltas, a PhD student at MIT, estimate in a new study that roughly 500,000 U.S. adults are out of the labor force due to prior COVID-19 illness.
This information — and the research behind it — can shed light on the longer-term costs of COVID-19 illnesses, which may influence strategies that policymakers and businesses can take to soften the blow. For example, social safety net programs like disability insurance might need to be changed. Employers looking to reduce turnover and protect workers might readjust sick leave or take further steps to try to stem COVID-19 illnesses at the office.
Reliable estimates on the pandemic’s workforce effects have been hard to come by — resulting in a mishmash of back-of-the-envelope calculations. Goda and Soltas have harnessed the best data to answer this question so far.
Along with estimating COVID-19’s impact on half a million sidelined workers, Goda and Soltas also show that workers with a COVID-19 illness earn on average 18 percent less over the subsequent year. This result includes people who move into lower-paying and/or part-time jobs.
All told, the researchers estimate that, over the pandemic, the consequences of COVID-19 illnesses have cost U.S. workers about $62 billion in wages per year. That is about half of comparable estimates of the productivity costs of cancer or diabetes.
Goda and Soltas say their study appears to be the first empirical analysis of the pandemic’s direct impacts on labor supply. Other widely reported estimates rely on subjective data that don’t adequately represent all U.S. workers or don’t compare those with prior COVID-19 illness to similar workers who did not become sick with COVID-19. These analytical shortcomings have led prior research to overstate COVID-19’s impacts on workers.
“While smaller than prior estimates, our estimate of 500,000 fewer workers who are not able to fully engage in the economy is a substantial number,” says Goda. “It gives policymakers and employers clarity on the true scope of the problem as they think about COVID-19 responses more generally.”
Finding answers to a pressing question
Goda and Soltas set out to examine COVID-19’s effects after working together at the White House Council of Economic Advisers (CEA). Goda was on sabbatical from Stanford beginning in the summer of 2021 to serve as the council’s senior health economist for a year (she returned to SIEPR on Sept. 1).
While at the CEA, Goda and Soltas knew that reports of COVID-19 survivors with either ongoing symptoms or permanent impairment to their physical health were increasing. A top policy concern from severe cases like long COVID-19 is the economic impact — including on the country’s already tight labor market — if people are unable to work.
Meaningful insights into these effects are hard to uncover. For one thing, Goda says, “there isn’t a consistent definition of what long COVID-19 entails, so even estimating its prevalence is difficult, let alone its consequences.” And relying on questionnaires of COVID-19 survivors can yield misleading estimates because people’s memories of why, for example, they quit working are not always complete or objective.
To overcome these and other hurdles, Goda and Soltas relied upon a large, representative household survey, the U.S. Current Population Survey (CPS), which is the top source of monthly labor force statistics and is jointly produced by the Census Bureau and Bureau of Labor Statistics. Using CPS data, they track the earnings and job status of all working adults aged 16 and older from the start of the pandemic through this summer.
Goda and Soltas find that, in a typical week during the pandemic, 10 out of every 1,000 workers were absent for a week or more for reasons related to their own illness or injury. That’s a significant increase from before the pandemic, when the absence rate was about 6 per 1,000 workers. The increase in health-related absences has fallen hardest on frontline workers, whose jobs put them at the highest risk of COVID-19 exposure.
The researchers then use multiple government data sources to link the increase in week-long work absences to COVID-19 case rates where these workers live. Their analysis suggests that the employees were out sick because they had come down with more severe cases of the virus. Next, they track these workers for up to 14 months after they first missed work, which is the maximum amount of time that CPS data track workers.
Identifying risks by demographic
Goda and Soltas find that workers who miss a week or more of work are 7 percentage points less likely to have a job a year after their infection than similar workers who did not take time off from work because of an illness.
To better understand what this means, consider two groups of 100 people. The first group of 100 falls ill with COVID-19 and everyone misses a full week of work as a result. The second group of 100 is not exposed at all over the course of a year. “Our results suggests that, after that year, there would be about 7 more people not working in the group of 100 ill people than in the control group of 100 people,” Soltas says.
Goda and Soltas show, too, that older workers are most severely affected by COVID-19 illness. And while COVID-19 has hit some demographics harder than others, the researchers find the likelihood that someone will drop out of the labor market because of long COVID-19 or other serious, virus-related complications has little to do with their race, ethnicity or education.
“Age aside, the effects we find are surprisingly similar across demographics,” Goda says. In other research, Goda has analyzed the pandemic’s economic toll on older and disabled workers.
Goda and Soltas caution that, while they have good reason to think that these job exits are due to severe cases of COVID-19, the CPS data only tracks health-related absences broadly — and not COVID-19 illnesses specifically. “Based on how excess absences track infection rates, there is solid evidence that these excess absences are due to COVID-19,” Goda says.
What’s more, the researchers can’t say for certain what happens for workers after the 14-month cutoff in their data. “But even at 14 months later, we still see substantial effects on people who say at that time that they don’t intend to search for a job for at least another 12 months,” Soltas says.
Goda and Soltas say their study provides new ways to estimate the impacts of COVID-19 illnesses and could lead to new insights into the long-term health and labor market consequences of the pandemic.