How Many People Can Work from Home?

In order to be able to understand, for example, the short-run hit that the economy is going to take during this period when we can’t go to work, the most basic thing you might want to know is: How many people, in fact, can work from home? How many people won’t be able to work at all if they can’t leave their home to go to a workplace?

To answer this question, [Chicago Booth’s] Jonathan Dingel and I looked at a number of survey questions administered by a group called O*Net that tries to describe various characteristics of given occupations. Some of the questions would be, for example: Do you operate heavy machinery? Or do you have to run during a typical day to conduct your job?

Based on these survey responses, we can essentially classify any given occupation as an occupation that could be done from home versus an occupation where you really couldn’t do it from home. A police officer or a firefighter, for instance, would be a very easy example where you couldn’t really do that from home. Another example that you couldn’t really do from home, that’s perhaps better because it would not be exempted from these stay-at-home orders, would be certain kinds of commercial drivers or certain kinds of workers that need to be physically active. A classic example would be for restaurants, a waiter can’t really do their job from home.

By contrast, for financial professionals, for a lot of service jobs, it’s not such a big deal to do your work from home. So Jonathan and I estimated that in the United States as a whole, 37 percent of all jobs could be done from home, could be done entirely from home. And it’s worth noting that that’s a much larger percentage of jobs than those that are currently done entirely from home.

That leaves, of course, a very large set of workers that can’t. I mean, that’s very scary and problematic and suggests there’s going to be a really severe economic hit from strong social distancing measures put in place to combat the health effects, of course, and to protect the public health.

Now these 37 percent aren’t the same as their complement. Those jobs that can be done from home typically are higher-paid than those jobs that can’t be done from home. So as much as 37 percent of the jobs can be done from home. They account for a larger set, 46 percent of all wages earned in the US economy. So about 46 percent of wages is done by jobs that can be performed at home.

Does that mean that the economy is going to perform under social distancing at a level equaling 46 percent relative to what it was before social distancing? Of course not. There’s a lot of elements that make it really difficult to map from our number to a guess or a forecast of GDP.

One thing that’s important to note is that there’s been a big demand shock, or there’s likely a very big demand shock associated with this crisis. Furthermore, just looking at the supply side, for those jobs that are done at home, we have not evaluated whether when they’re done at home, they’re done with equal efficacy, with equal productivity as when they were done at work.

And finally, it might be that various jobs can be done at home, but they’re done less productively because when they’re done at work, they experience positive interactions, positive feedback, essentially, from interacting with other jobs that can’t be done at home.

For all these reasons, we think our number is helpful, but more work would have to be done to translate it into an assessment of what the actual hit to the economy will be from strong forms of social distancing.

There’s a lot of heterogeneity across cities and across industries in the share of jobs that could be done at home. For instance, in some cities like San Francisco, a large share of the workers could, in fact, continue their work even under very strong forms of social distancing, where in other places like Las Vegas or Bakersfield, California, it’s going to be much more difficult.

This is important to know, just because it does imply that the effects of this crisis are going to hit very unevenly across households and regions, cities and industries. But also we hope that being able to flesh this out, being able to provide these statistics, might be useful as policy makers try to target aid to those regions and occupations and industries and people that most need it.

I should also mention that the heterogeneity that we described not only exists within the United States across cities, but Jonathan and I are now in the early stages of looking at the equivalent statistic across various countries. And it’s very early, but our work so far suggests that poorer countries, more emerging markets that aren’t at the level of development of, let’s say, the United States, typically have an even lower share of their jobs that can be performed at home.

Thinking globally about the impact of the COVID-19 crisis, our analysis does suggest that it might be an even harder challenge to keep workers going and doing what they used to do for emerging markets and poorer countries than in the rich world, where it’s, of course, still a very large challenge.