Occupation Gender Segregation and Birth Rates in the U.S.

–

When the going gets tough, the weak noodle around with data.

I don’t know if I’ve seen anyone directly compare the gender composition of women’s occupations with their birth rate in the U.S. before. If you’ve seen this, please let me know. If you would like to look into it more, feel free to look at the Stata data file and code for this exercise and use any of it that you find useful.

This uses the 2021-2023 American Community Survey data, which asks people what their occupation is (including their most recent occupation if they are not currently employed), and whether they had a birth in the previous year (for women). And a bunch of other demographic variables.

I start with a sample of all adults age 15-49, and calculate the percent female within each of 530 detailed occupations. Then I calculate the proportion of women in each occupation who had a birth in the previous year. This is the basic result:

There is clear a positive relationship between the percentage of women in an occupation and their likelihood of having a birth. Among large occupations, those with highest birthrates — more than 9% having a baby in the past year — are female dominated: Telemarketers (66% female), Optometrists (63%), Nurse practitioners (89%), Physician assistants (71%), Personal appearance workers (95%).

Near the bottom are some jobs with very few women, like Carpenters (3% female, 3% with a birth) and Maintenance workers (5% female, 3% with a birth). But some female dominated jobs have very low birth rates, like restaurant hosts (82% female, 2.5% with a birth). And smaller occupations scattered around. So there are other factors at play, obviously.

To help with the influence of other variables, I ran a logistic regression model with controls for age, marital status, number of times married, education level, race/ethnicity, and nativity. Here is the effect of occupation gender composition with those variables controlled:

Don’t like ads? Become a supporter and enjoy The Good Men Project ad free

This shows a difference of about 1 percentage point, with women in the most male-dominated occupations having children at a rate of just under 5% per year, and those in the most female-dominated at just under 6% per year. (The other variables perform as expected, with highest birth rates for women ages 25-29, those with less education, those who are married (especially for the second time), Black and Hispanic women over White and Asian/PI women, and foreign-born women.)

One caveat is that some women may have moved into their current occupation after the birth in question; I can’t tell from the data. So if a carpenter quit or lost her long-time job last year, then had a baby and took a job as a telemarketer, I would erroneously tell you that a telemarketer had a baby. The other problem is I can’t use things like spouse/partner characteristics, earnings, or work hours, because all those things likely changed in the last year for women who had babies, so it would be misleading; I just leave those variables out. There are better datasets for things like this, they just have smaller and take longer to come out.

I didn’t look for literature on this very hard (which might have ruined my fun), but I did find an old article about the Netherlands showing the same pattern. There is a lot of research on education levels, fertility intentions, marriage, and labor force participation, but I didn’t see much about occupations.

Anyways, if this is a pure gender segregation effect (which is a stretch based on this evidence) it would be interesting if one result of gender integration was a decline in birth rates.

ADDENDUM

I opened the post by saying “I don’t know if I’ve seen anyone directly compare the gender composition of women’s occupations with their birth rate in the U.S. before.” I say things like this to communicate that I am not claiming my idea to be original. It’s just something that popped into my head yesterday – which makes it highly unlikely to be original. For a blog post, the obligation to document the intellectual lineage of an idea is, in my opinion, reduced. So instead of litigating the originality, I just disclaim precedence.

Note, however, I did not say that I am not aware of the idea that female-dominated occupations are somehow more “family-friendly” and better for mothers. This has been an explanation for occupational gender segregation for more than 40 years, at least since Polacheck proposed the idea that women choose occupations that don’t penalize their time out of the labor force, thereby causing gender segregation. One of Paula England’s first very influential publications (1982!) was a debunking of that claim, when she wrote, “there is no evidence that plans for intermittent employment make women’s choice of traditionally female occupations economically rational.”

Since then, there have been thousands of papers on how motherhood affects wages and career outcomes, and how childbearing can affect occupational choice, and how “family friendly” policies reduce work-family conflict, and why flexible professional jobs with more linear returns to hours worked — basically, pharmacists, in the work of Goldin and colleagues — attract women. After all this, though, it’s still not a safe assumption that women are concentrated in family-friendly occupations.

I have added a label for pharmacists to the figure above: 64% female, 7.4 births per hundred women, along with optometrists (63%, 10.8). Note Goldin is basically only concerned with college graduates, but if someone were more interested in the theory generally they might like to argue that telemarketing jobs are super mother-friendly (66% female, 11.7 births per hundred). Flexible hours! Work from home! Do we want more jobs like this?

Which brings me to several critiques I heard on social media, responding to the clickbait line I used to promote the post: “What if gender integration lowers birth rates.” One person wrote, “In our current environment, this gives fodder to the claim that women should stay out of predominantly male occupations or should not work at all. Without addressing these issues, you’re helping enemies of women’s liberation.” Another suggested privately that, “a simple correlation between gender segregation and fertility could be seen as justification for further segregation.”

I appreciate the concern, but it’s not the job of feminists to boost birth rates, or to promote occupational gender segregation. Maybe I should have reverse-coded my comment to say, “What if gender segregation raises birth rates.” That means exactly the same thing, but would it have raised concerns?

Of course, gender integration DOES lower birth rates – and gender segregation raises them – if you take any kind of historical perspective. In modern economies, women entered new occupations, had more rewarding, more independent careers, and had fewer children. This. Is. Good. Don’t let the pronatalists convince you otherwise. If they want more children to labor in support of their AI machines, they can just open the borders. (But what they really want is White supremacy and eugenics.)

Finally, if we have learned anything in the last month, it includes this: the forces of reaction do not need, and do not want, social science research to justify their evil ends. And they don’t read. So I’m not too worried that my blog post is going to increase gender segregation.

—

Previously Published on familyinequality with Creative Commons License

***