The pandemic was historically bad for working-class women

By just about every measure other than mortality (which is pretty important!), the Covid-19 pandemic has been harder on women than on men. The majority-female services sector lost jobs at at a faster rate than the majority-male goods-producing sector — which has never happened before in a recession. More women than men have dropped out of the labor force. Even women who held onto their jobs generally had to shoulder an outsized share of the new child-care and schooling burdens.

Much of the media coverage of this "she-cession" has focused on the tales of professional-class women who have struggled to balance pandemic-imposed home responsibilities with work and in some cases completely dropped out of the workforce. Those struggles are real. But one place where they haven't really showed up is in the jobs data.

The February 2021 employment numbers from the Bureau of Labor Statistics provided a unique opportunity to gauge who was hardest hit by the pandemic recession. February 2020 was the last month of normal life before Covid-19 changed everything, and some labor-market indicators are reported on a not-seasonally adjusted basis that makes year-over-year comparisons the only really meaningful ones.

The data show a jobs recession that was hardest on women without college degrees and Black and Hispanic women, and was pretty tough on men without college degrees and Hispanic men too. College-educated women, White women and Asian women, on the other hand, experienced smaller jobs losses that were roughly equal to those of their male peers.

The best way to measure these losses is with the employment-population ratio, or Epop. 1 It's a more complete measure of the state of the labor market than the unemployment rate because, unlike the unemployment rate, it factors in those who have left the labor force and given up looking for jobs. Here's how the Epops of men and women 25 and older of varying education levels changed from February 2020 to February 2021.

A tough year for high school grads, especially women

The US employment-population ratio fell across the board, but the sharpest drop was for women with only high school degrees*

The three education groups in the middle are the largest, although given how many women are in graduate school (an estimated 1.8 million in the US in the 2019-2020 academic year, versus 1.2 million men) it won't be much longer before women with advanced degrees outnumber those with only high-school degrees in the workforce. The no-high-school-degree category is relatively tiny, especially for women, so one shouldn't make too much of its seemingly against-the-trend Epop trajectory over the past year.

The three education groups in the middle are the largest, although given how many women are in graduate school (an estimated 1.8 million in the US in the 2019-2020 academic year, versus 1.2 million men) it won't be much longer before women with advanced degrees outnumber those with only high-school degrees in the workforce. The no-high-school-degree category is relatively tiny, especially for women, so one shouldn't make too much of its seemingly against-the-trend Epop trajectory over the past year.

The education profile of the American workforce

The hard-hit high-school-degree-only group remains one of the biggest*

One explanation for the big job losses among those with only high school degrees is the ability to work from home. Other surveys have shown this to be closely related to education level — the more education you have, the more likely it is that you can do some or all of your work remotely. Especially in the early days of the pandemic, those who couldn't work remotely were much more likely to lose their jobs.

At the same time, women with high school degrees were more likely than their male counterparts to have new responsibilities at home imposed by the pandemic (such as taking care of kids whose schools had halted in-person instruction) that were incompatible with non-remote jobs. Even if they didn't get laid off because of the pandemic — even if they were essential workers — many had to quit.

These gender disparities in who is expected to take care of the kids exist across the educational spectrum, of course, but women with bachelor's degrees and higher have been much less likely to leave the labor force, presumably because more of them could work from home (and afford child-care help when they couldn't). The stresses faced by women in professional and managerial jobs do seem to show up in other metrics: Their share of scientific papers published and venture capital funding received plummeted during the pandemic, for example. But for the most part they didn't lose their jobs.

The job losses also fell very unevenly by race and ethnicity. Hispanics have had the toughest time overall, but the real standouts, in a bad way, have been both Black and Hispanic women. Here the Epops are for what's known as "prime-age," ages 25 through 54 (the data by educational attainment aren't available in that form):

A tough year for black and Hispanic workers, especially women

Prime-age employment-population ratios fell the most for Hispanic women, followed by Black women*

Ability to work from home would seem to be a factor here as well: In one BLS survey from before the pandemic, only 16% of Hispanics could work from home versus 31% of non-Hispanics, and 20% of Blacks versus 30% of Whites and 37% of Asians. This is in turn probably linked to education: Asian adults are the most likely to have college degrees, followed by Whites, Blacks and then Hispanics.

Sorting out causes and effects here can be tough. There actually are some statistics available on Epops sorted by gender, education, race and ethnicity but the numbers of people involved can be so small that the numbers are noisy and inconclusive. For example, Black women with just bachelor's degrees experienced a smaller Epop decline than their White peers over the 12 months ending in February, according to the BLS data, but those with advanced degrees saw a bigger decline. With Hispanic women it was the other way around.

Another interesting statistic of uncertain meaning is that married women (and men) seem to have weathered the Covid jobs crisis better than those who are on their own. Is that because marriage provides a support structure that helps couples get through difficult times, or just that Whites, Asians and those with college educations are more likely to be married than everybody else? Or both?

It helps to be married

Single men and women faced sharper employment-population ratio declines than married ones*

Finally, the industry one works in mattered a lot as well. In general, heavily female in-person service sectors have been hit much harder by the pandemic than the heavily male manufacturing, construction and transportation. These data are from the establishment survey of employers rather than the household survey from which the employment-population ratio is determined, so they're expressed simply in numbers of jobs rather than Epops.

Majority-women sectors were hit the hardest

Payroll employment fell the most in leisure and hospitality, government, and education and health services

The establishment data are not broken down by education or race, but there's ample evidence that lower-paid, lower-status jobs were hit hardest. Average pay, which usually falls in a recession, leapt last spring as low-wage jobs disappeared, with April's gain in real average hourly earnings the biggest ever recorded. More than half the job losses in the health-care sector were at nursing and residential-care facilities, where 80% of employees are women and average weekly earnings of non-supervisory employees just $618.

This past year has been pretty awful for almost all of us, but has placed unique demands on women. In the job market, these have weighed heaviest on the women with the fewest resources. Even as aid pours out of Washington and the economy gains strength, there's a good chance a lot of them still need help. 

Disclaimer: This article first appeared on bloomberg.com, and is published by special syndication arrangement.