Judging from the media coverage of the work from home (WFH) phenomenon, you’d think it’s become near universal. It’s not. In July, only about one in eight workers were teleworking—the Bureau of Labor Statistics’ (BLS) preferred term—and those are heavily concentrated in a few sectors and occupations, and among the highly credentialed.

According to BLS stats, in July 2021, just 13% of workers are doing so remotely because of the pandemic, down from 35% in May 2020, the first month the numbers were collected. (See graph below.) And that initial 35% number was inflated by the fact that so many workers who couldn’t work from home, like those in retail and hospitality, had been laid off. Since that peak, the share has declined in twelve of the subsequent fourteen months.

And, as the paired bar graphs below show, WFH is heavily concentrated in a few demographics. Asians account for 15% of teleworkers, well over twice their share of the employed, no doubt reflecting their strong presence in tech; Hispanics/Latinos, just 11%, well under their share of employment. By education, workers with less high school diplomas or less barely register—but those with advanced degrees are heavily over-represented.

If we define the heavily teleworking segments as those with an intensity (share of teleworkers divided by share of employment) of over 150%, just eight occupations out of twenty-two (computer and mathematical, business and financial operations; legal; life, physical, and social science; architecture and engineering; arts, design, entertainment, sports, and media; management; and community and social services) accounted for 69% of teleworkers in July, over twice their share of employment (30%). The concentration by industry is even sharper: just five sectors out of twenty-three (finance and insurance, professional and technical services, information, public administration, and utilities), pass the 150% intensity test. Together, they account for 70% of teleworkers, but just 21% of employment. (These aren’t graphed because they’d be too big for a little blog post.)

Telework was somewhat more democratic in the early days of the pandemic. In May 2020, 46% of those working from home had less than a bachelor’s degree, somewhat less than their 55% share of total employment. In July, that share was down to 15%. In May 2020, 57% of those in managerial and professional occupations were working from home; in July 25% were. Over the same period, the share of those teleworking outside the professional and managerial occupations went from 17% to 5%.

On the left, you often hear its said that we should just pay people to stay home. Of course, if people can’t find work, they need generous support. It’s cruel to cut off expanded unemployment insurance benefits.

But most jobs just can’t be done remotely. Those who can work remotely have to be served by the majority of lower-status, lower-paid workers who make things and move them around. If they were paid not to work, money would quickly become worthless because there’d be nothing to buy with it—no food, no electricity, no appliances, no medical equipment, nothing.

Telework is common among high-paid, high-visibility fields, but it’s just not the “new normal” for most workers. And it’s hard to see how it could be otherwise.