Multiple Jobholding Among Women In the Pandemic Economy

To follow up on last week’s post showing how women are still more likely to be working multiple jobs than men (both among the employed and among their total populations), here’s a reminder that not all women are the same. Let’s look at multiple jobholders as a share of employed, across race and age categories.

Black women are substantially more likely to hold multiple jobs than any other race categories of women. Notably, while multiple jobholding fell for all groups in the spring when unemployment peaked (and number of overall jobs in the economy cratered), multiple jobholding as share of employed has already (as of September) returned to a “normal” level for Black women but not for other women.
By age categories, the youngest of working women (ages 20-24) have always been the most likely to hold multiple jobs, because they are most likely to have to piece together multiple part-time jobs (often in the leisure/hospitality sector) to make a living. These women were most likely to lose at least one of their multiple jobs at the start of the pandemic, and regain work as businesses reopened in the summer.

There are many factors that could explain the differences by race, probably most significantly that Black women are more likely to be sole earners in their households (as well as single parents) yet also more likely to earn lower hourly wages. The different trends by age reflect that multiple part-time jobs are often the closest a young adult (even a college-educated one) can come to a full-time job–and that the human-service-intensive jobs many young women work in were the ones that disappeared the most at the start of the pandemic and have not and will not likely fully come back even when the public health crisis eventually wanes. The Pandemic Recession — or “She-cession”– has not just been hard on women because of the severity of the lowest depths of job loss experienced, but because it’s really “jerked around” the women who were already the most economically vulnerable.

Women Still Working a Lot of Jobs During the Pandemic

Last summer –as in all the way back in 2019–I wrote this post about how multiple jobholding had become more common in the labor force overall and how this seemed to be a “new normal” for the (very strong 2019) economy–something no longer limited to recessionary times when people might be forced to piece together several smaller jobs when they couldn’t get a full-time job. I underscored my finding that multiple jobholding was becoming especially more prevalent among working women, and I hypothesized on some reasons why, including: (i) women need multiple jobs to add up to the pay of one job (because women typically earn less than men); (ii) women choose to work multiple jobs to add up to the hours of one full-time job–yet with the greater flexibility/control over work schedule that multiple part-time jobs allow; and (iii) women often choose to “work” not to maximize earnings but for personal fulfillment, which often calls for a variety of work whether paid or underpaid or unpaid, rather than just one job.

In 2019 all these reasons were already, in the slow-but-steady recovery from the Great Recession, becoming part of the “new normal” of an economy and labor market increasingly dominated by women. Last week I learned of this new Census analysis based on a new measure of multiple jobholding:

We create a measure of multiple jobholding from the U.S. Census Bureau’s Longitudinal Employer-Household Dynamics data. This new series shows that 7.8 percent of persons in the U.S. are multiple jobholders, this percentage is pro-cyclical, and has been trending upward during the past twenty years. The data also show that earnings from secondary jobs are, on average, 27.8 percent of a multiple jobholder’s total quarterly earnings. Multiple jobholding occurs at all levels of earnings, with both higher- and lower-earnings multiple jobholders earning more than 25 percent of their total earnings from multiple jobs. These new statistics tell us that multiple jobholding is more important in the U.S. economy than we knew.

“A New Measure of Multiple Jobholding in the U.S. Economy,” by Keith A. Bailey and James R. Spletzer, Working Paper #CES-20-26, September 2020.

So the Census researchers conclude that multiple jobholding is pro-cyclical (no longer just a “make ends meet in hard economic times” phenomenon), has been rising in prevalence over the past 20 years (not even just since the Great Recession), and “is more important in the U.S. economy than we knew.” They might as well have said that the role of women in the workforce, and how women choose to work, is more important than we knew.

This prompted me to go back to my summer 2019 blog post and update the data through September 2020 (the latest monthly employment statistics from household survey data) to see what the past year looks like in terms of multiple jobholding and its prevalence among workers overall and men vs. women. Here are some charts that look at multiple jobholding as share of people who are employed, in the workforce (“labor force” meaning employed or unemployed/looking for work), and relative to total population–and then comparing each of the three shares for men vs. women. All charts show BLS unadjusted monthly household employment survey data for populations age 16+, from January 2018 through September 2020:

Shares of total (men+women) population measures: Overall, multiple jobholding dropped dramatically from February to April 2020 at the start of the Pandemic Recession as the number of jobs in the overall economy plummeted, and as of September is still below what had been the “new normal.”
As share of total employed, comparing men (blue) vs. women (orange). Note that multiple jobholding in the recovery (thus far) from the Pandemic Recession first rose for both men and women, but since July has decreased among working men while increasing among working women.
As share of those in the labor force, we see basically the same story as among those employed; since July, women’s multiple jobholding has increased while men’s has decreased.
And even as a share of their total population, women are more likely to be multiple jobholders than are men.

In summary, we see that multiple jobholding has come down and then back up during the pandemic, along with jobs and employment in general. Women still juggle a lot of jobs, whether paid work in the labor market or unpaid work at home. Even as workers have lost some of their multiple jobs and have yet to fully regain them, we see that multiple jobholding is a significant phenomenon in the U.S. economy and one that is not necessarily a bad thing if it has made it easier for workers–especially working women–to better tailor their work opportunities to their personal circumstances and preferences. Women have always chosen multiple part-time jobs so that work fits into the rest of their lives better, even pre-pandemic. With the pandemic placing only more burdens and constraints on women’s time (with kids and elderly parents to care for), multiple jobholding will become an even more important way for women to stay connected to the workforce. But multiple part-time jobs have never added up to the level of economic reward one can get from one full-time job–whether it be in the form of wages or benefits (such as subsidized health insurance, childcare assistance, and paid leave). Women are too easily relegated to “secondary earner” status (read: “you should be the one to stay at home now”) which makes it too common for them to disengage from market work when family circumstances change. The way women juggle their different jobs at home and at work will require a lot more attention and focus from economic policymakers if we want our economy to not just “survive” this pandemic recession but to actually “thrive” over the longer term.

Update on Pandemic Recession vs. Great Recession Unemployment by Race and Gender (incl. Asian Women)

10/22/20: Belated UPDATE through September data to charts below, originally shared on Twitter here. (I’ll make future updates here as well in order to keep in one place where some might look thanks to the new Center for American Progress report by Mike Madowitz and Diana Boesch published today.)

Today the BLS released their monthly “Employment Situation” report –and the report’s underlying data (which you can access here), including numbers on Asian women separated from Asian men which are not included in the report. Here’s a small, mostly quantitative update to my original post (with Mina Kim) from last month. First, the bottom-line/main-story bar chart:

… and below, the updated table with detailed unemployment stats by race and gender:

One thing I decided to adjust in the bar chart and add as note to the table is that April 2020 was when the unemployment rate peaked for the workforce overall as well as for most of the race-gender categories–except that the peak for Asians (both men and women) and for Black women (and Black overall but not Black men) happened in May. So the bar chart compares the absolute change in unemployment rate during the (aftermath of the) Great Recession compared with the change from this February to either April or May–whichever was the worst point for each race-gender category.

Some findings worth highlighting (or repeating):

  • Asian women fared the best in the Great Recession in that their absolute change in unemployment rate from the start of the recession to peak unemployment was the smallest of all race-gender categories;
  • From the start of the Pandemic Recession (Feb. 2020) to either peak unemployment in April or May, or to the latest data for September, White men have fared the best (their September unemployment rate is just 2.8 percentage points higher than in February);
  • Hispanic women fared the worst in the Pandemic Recession through this spring (April)–but up through September, Asian women have seen the largest increase in unemployment (+6.7 percentage points);
  • While Asian unemployment peaked in May at 16.6% for women and 13.2% for men, both had failed to recover as much as for other race-gender categories through August. Unemployment rates in August were 11.5% for Asian women and 10.0% for Asian men, still higher than they ever got even in the depths of the Great Recession. By September, Asian male unemployment (at 8.0%) was lower than at its peak in the Great Recession, yet Asian female unemployment (at 9.7%) was still higher than at any time in the last recession–and in fact, higher than ever measured in the BLS data.

As I hinted at in my first post on this subject, there are many different possible explanations for why this Pandemic Recession has been so hard on women (it’s the “She-cession”) and Asian women in particular. The intersection of Asian and female provides a uniquely-focused lens into what’s so different about this recession and how it’s affecting all workers through the roles and demands that Asian women just happen to disproportionately represent. To be continued!

More Pictures on the Pandemic “She-cession” by Race

Here are a bunch of charts that illustrate why the Pandemic Recession is also sometimes referred to as the “She-cession.” These are based on BLS monthly data through August (released last Friday, Sept. 4). I’m also re-posting the bar charts I put up on Twitter and here last Friday, so the latest figures are in one place. (See original post on “Seeing Asian Women…” here and the table with all the unemployment data through August by race-gender categories here.)

First, compare male minus female unemployment during the Great Recession to during the Pandemic Recession (thus far):

During the Great Recession (Dec. 2007-2009/10 recovery period), male unemployment relative to female unemployment grew dramatically–and in general/overall male unemployment stayed higher than female even years after.
During the Pandemic Recession (since Feb. 2020), female unemployment increased dramatically and surpassed male unemployment in all racial categories, and only among Black workers did male unemployment again exceed female unemployment this summer.

The differences in unemployment across racial categories have been stark. It’s not just a recession with disproportionate impact on women, but one with disproportionate impact on people of color. Zooming in a bit more recently, here’s what the unemployment rates by race since January 2019 look like:

Unemployment rates by race categories, men and women combined. Note that Asian unemployment started the very lowest in February 2020 and is the second highest in August 2020.

Going back to looking at the employment outcomes among women by race over the past year and a half (since January 2019), here are charts showing the unemployment rates, the employment-to-population ratios, and the labor force participation rates:

Female unemployment rates, by race, January 2019 through August 2020. Note that like for race categories overall, Asian females had the lowest unemployment rate in February 2020, but in August 2020 have the second highest–as Hispanic female unemployment has shown the best recovery since its peak in April.

Female employment-to-population ratios by race, January 2019 through August 2020. Hispanic women had the lowest e/pop at worst part of the Pandemic Recession and have traded places with Asian women a couple times since. Black women had had the highest e/pop going into the pandemic but in August, white women had surpassed black women.
Female labor force participation rates by race. Asian female participation was relatively low pre-pandemic, but is second highest this summer–showing that many Asian women consider themselves still “in the labor force” even if currently unemployed.

And below is the headline bar chart I previously posted comparing absolute changes in unemployment rates for women by race in the Great Recession vs. during this Pandemic Recession (through August):

The Pandemic Recession is a “She-cession” and has had disproportionate impact on women of color, particularly on Asian women–who face a still higher unemployment rate in August (at 11.5%) than at their worst point in the Great Recession (8.4%).

And looking at Asians alone (both men and women), the “double whammy” of being both Asian and female in this Pandemic Recession is underscored:

Asians (especially the women) had very low increases in unemployment during the Great Recession relative to other race categories. This time is totally different.

Seeing Asian Women to Better Understand the Pandemic ‘She-cession’

By Mina Kim and Diane Lim (with assistance from Taylor D. Jones), August 2020

This pandemic recession has been like nothing we’ve ever seen in our lifetime. The U.S. has experienced the largest contraction since the Great Depression. Every week, we eagerly await each data release for any sort of good news, only to be met with confirmation after confirmation of the calamitous impacts of the coronavirus pandemic. Economists were quick to point out early in the pandemic that increases in unemployment have been more severe for women than men and more severe for people of color than for White people. But in all the analyses, Asian women have been missing. Given the goal of national representativeness in most surveys, the “invisibleness” of Asian women is understandable. We represent a very small percentage of the population. We are also probably not as responsive to surveys given how 1st and 2nd generation Asian American women are taught to blend in and not stand out. (Both of us reflect on how this “blend in” behavior was reinforced in us very early on in our childhoods, by our teachers, classmates, and even our immigrant parents.) As Asian female economists, we had personal incentive to take a good look at the data—we just wanted to find ourselves there! What we found is that not only have Asian women experienced the largest increase in unemployment during the pandemic recession, but that the case of Asian working women probably best shines a spotlight on how women’s employment depends as much or more on what’s going on at home (and the demands for our time there) than on what’s going on with the market demand for our paid work. What follows is a story of what has happened to women in the labor market during this Pandemic Recession and how this recession is different from the last, made clearer by our ability to dive deeper into details across race categories.

NOTE: For this analysis we used the BLS “One-Screen Data Search” tool here. While the monthly employment reports of the Bureau of Labor Statistics (BLS) present breakdowns of the unemployment measure by race and by race-gender categories, they only show data for the Asian category as a whole because of the valid concern that the Asian sub-sample sizes are too small to be statistically comparable with other race-gender categories. But the BLS does publish and make publicly available the data on Asian women separated from Asian men on its website. All data are monthly and unadjusted (not seasonally adjusted), for the populations age 16 and over. The “Asian” category includes only people who identify themselves as (only) Asian on the Census survey, not those who identify as two or more races; it also does not include Pacific Islanders. (A helpful BLS report which explains their categorization of people by race and ethnicity is here.)

  1. The “Pandemic Recession” is a “She-cession” (Figure 1). C. Nicole Mason, CEO of the Institute for Women’s Policy Research, first dubbed the pandemic recession the “She-cession” back in early May in this New York Times story, when it had become apparent that more women than men had become unemployed since February. At the time, the leading explanation was that women make up a disproportionately large share of the workforce in the industries that have been most adversely impacted from the “stay-at-home” orders: the leisure/hospitality sector (think restaurants, bars, sports venues, concert halls, movie theatres), other people-intensive personal services businesses (think beauty salons/barbers, gyms, yoga studios), and non-emergency medical care offices (think general/family practices, dentists, eye doctors).

Now there is recognition of the role childcare (and in-person school), or the lack thereof, plays in affecting the ability of women to work (see Alon, Doepke, Olmstead-Rumsey and Tertilt; Mongey, Pilossoph, and Weinberg; Dingel, Patterson, and Vavra; Modestino, Ladge and Lincoln; and Heggeness). Compared with a typical economic recession that impacts male employment more than female, the Alon et al. paper stresses that the “She-cession” has more severe short-term/cyclical impact on economic activity due to diminished “within-household insurance”—a reduced ability for parents to trade off/coordinate roles at home vs. work—and will also lead to a longer-term (persistent) widening of the gender wage gap even after the recession is over.

Figure 1.
  1. This recession is very different from the last one (Figure 2). The “Great Recession” (left panel) impacted male-dominated jobs in the economy more than the female-dominated ones. Before the start of that recession, goods-producing jobs (particularly in the durable goods/”heavy” manufacturing sectors) had already begun a longer-term loss due to both offshoring and automation. As the economy began its long, slow recovery from the recession (officially in “recovery” by June 2009), unemployment rates continued to rise, peaking for men in early 2010 at a much higher rate than for women. Overall female unemployment peaked later (in the middle of 2010) but remained below male unemployment until the middle of 2011. Many of the men who had been laid off from their manufacturing jobs during the Great Recession never came back to those jobs after the economy recovered. These were disproportionately middle-aged, White men who came out on the other side of the recession saying “well, I guess I’m retired now.” (These were also the White men disproportionately represented in the “deaths of despair” that Anne Case and Angus Deaton revealed.)

This time (right panel) the economic impact was felt suddenly due to a non-market event and has fallen disproportionately on the employment of women. After several years of very low unemployment where women and men switched ranking but stayed very close to each other, female unemployment very quickly overtook male unemployment at the onset of the pandemic and has retained its “lead” thus far.

  1. Pandemic unemployment has been dominated by the laying off and rehiring of largely part-time workers (Figures 3 and 4). The part-time share of total employed fell as unemployment peaked (in April) and rose as unemployment has declined (since May). The timing is consistent with job losses concentrated in the leisure/hospitality, personal services, and retail sectors, where workers are disproportionately young, part-time, and low-wage (no benefits). (This is also why government assistance has been critical for households to be able to maintain basic consumption and prevent a more severe “reverse multiplier” effect on the economy more broadly.)
Figure 3.

While the trends in unemployment of part-time workers impact women more than men across all race categories, Black and Asian women more often work full-time jobs than do White or Hispanic women.

Figure 4.

If pandemic unemployment were mainly driven by occupational and industry factors (layoffs from part-time jobs in the leisure/hospitality, retail sales, and personal services sectors), we might see Hispanic and White women suffering the greatest unemployment, and Asian and Black women, less so. In fact, it’s not that simple…

  1. The “She-cession” story differs across race (Figures 5 and 6). It appears to have come and gone among Black Americans but has persisted for other race categories. By June, Black male unemployment was back above Black female unemployment, while in all other categories female unemployment has exceeded male unemployment since April. In the prior recession, men overall faced more severe unemployment than women, and it was especially severe for Black men. In this recession, Black men have not had as large an increase in unemployment as many of the women, but only because they started so high. Before the pandemic, in February, the Black male unemployment rate (at 7.3%) was the highest of all the race-gender categories. As of July, the Black male unemployment rate (at 15.6%) was still the highest of all the categories.
Figure 5.
Figure 6.
  1. The “She-cession” experience has had biggest impact on Asian women (Figures 7 and 8). Among all women, Asian women had the lowest unemployment rate in February, yet their unemployment rate has been above average and in line with the unemployment rate of other women of color during the pandemic. In terms of the change in unemployment rates during this pandemic recession so far, Asian women have experienced the single worst change in unemployment through July—not just among women but across all the race-gender categories. (Hispanic women suffered the largest increase in unemployment from February to April; besides figures below, see details in the Appendix Table at the end.)
Figure 7.
Figure 8. (Note: Worst point in the Great Recession is going to differ for each race category. See the Appendix Table for the exact dates.)

Caveats, Take-Aways, and “To Be Continued”s:

  • We’ve provided a one-step-more-granular analysis here, simply by including Asian females in our look at employment outcomes by race and gender. But this still doesn’t include everyone! “Asian” doesn’t include those who identify as only part-Asian (of multiple races), nor the “Pacific Islanders, American Indians, Alaska Natives, Native Hawaiians, and Other Pacific Islanders” who also tend to get left out of the “Asian” category in analyses by race—due to even smaller sample sizes than that of Asian females.
  • Yet, this race-gender intersectional analysis is still meaningful in that we do see clear differences in employment outcomes across both race and gender and all the race-gender categories. The analysis “works” because as broad as these race-gender categories are, and as heterogenous as the characteristics of the populations within each category, there are still enough characteristics that are common (or prevalent) within each category yet distinct enough from those in other categories. We suspect, for example, that the “Asian Female” category contains a sizable group of women who have relatively low educational attainment and work relatively low-skilled and low-paid personal services jobs (think nail salon), which would partially offset whatever we see that is driven by the high-skilled, higher-paid Asian women working in professional service jobs. Yet the average characteristics and outcomes within the Asian Female category still put the group as a whole in higher-skilled occupations and higher-paid jobs compared with women in the other race categories.
  • And those distinct characteristics and outcomes are not merely random differences. Race is a social construct, and that’s why it is useful in studying social and economic behaviors, influenced by social and cultural norms. Asian women look “special” in this analysis of unemployment during the Pandemic Recession because their roles in the labor market and at home and the resulting tradeoffs they face in making decisions about work vs. home are likely different enough from other women. For example, Asian women are less likely to experience the gender wage gap, probably because of their high educational attainment and occupational choices—and that’s a plus for holding onto jobs during this pandemic. (Distinct from other women, Asian women are more likely to work high-paid professional service jobs than low-paid leisure/hospitality or retail jobs.) But Asian women are also more likely to be in a traditional family structure, married with children, and married to a man who has even higher market-earning power. (We hope to do a follow-on piece on the family structure and marriage/”assortative mating” factors later.) So, as highly productive as an Asian woman might be in her job, she is likely to be the household’s “secondary earner” and also the household’s “primary caregiver.” This is why a deeper study of what’s going on with Asian women in this pandemic recession would shine a light on the struggles of all working mothers, who all face demands on their time at home with the kids. Those demands (regardless of the financial opportunity costs) too easily tip the scales in favor of staying at home and cutting back on or quitting their paid work.
  • For Asian women, the cultural factors influencing our labor market decisions are perhaps stronger than for other women. (See this paper on “Babies, Work, or Both?” by Mary Brinton and Eunsil Oh.) Many working-age Asian women are immigrants or children of immigrants. Diane’s parents immigrated to the U.S. from Korea (dad) and the Philippines (mom, of Chinese ethnicity) in the 1950s, both to pursue PhDs in Chemistry. Diane’s mom finished her PhD in 1962, just months after Diane was born, and spent nearly a decade staying at home with Diane and her younger sister Janice and was never able to make up for that lost first decade of her career. Mina’s parents immigrated from Korea in the 1970s to pursue a PhD in Physics (dad) and a Masters in Nursing (mom). Mina’s mom put her career on hold to take care of her and her two sisters when family could not help. Yet both moms constantly coached their daughters to “do it all”: to both stay ever-active in the workforce and take primary responsibility for the proper care of the kids on the home front.
  • This “first look” at Asian women’s employment outcomes has underscored the importance of considering how social and cultural expectations about the roles of women at home vs. at work influence female employment outcomes, especially in this pandemic with kids learning from home and daycare and elder care facilities closed or downsized (or viewed as undesirable/unsafe by families). It also suggests that the increased demands placed on women’s time on the home front does not just force women who earn relatively low wages (below the cost of child care) to give up their paid job, but is likely causing many higher-earning women (who can afford to pay for child care even as it has become scarcer) to cut back on or even drop out of the labor force, even when the financial and economic opportunity cost of doing so is high.
  • Further investigation is needed, going still more granular and down to more household-level data, to better understand what truly drives women’s choices and outcomes concerning work. These are questions fundamental to the study of Economics and requires research methods that have largely remained outside the Economics profession. Further research should: (i) start with interviews and focus groups; (ii) use these findings to develop surveys; and (iii) field the surveys to intentionally oversample the population subgroups that best represent and emphasize the decisions we are trying to better understand.

Appendix Table: detailed unemployment rates and changes, comparing previous to current recession