The Sorting of Women into Geographically-Flexible Jobs

Alan Benson

“Grad students divorce and break up at much higher-than-average rates. It's a stressful experience, full
stop. And it doesn't necessarily get easier for your partner when you're done. One of the ironies of
academia is that it's hard to be a two-career family when one member might have to move three times in
fifteen years.”

With these words, a veteran of my program greeted new graduate students. Having spent the previous year
touring doctoral programs with other admitted students, I was familiar with what was sometimes called the
“two-body problem” and the sacrifices career-oriented couples make.

Of course, the two-body problem is not unique to academia. For the fresh college-graduate, the mid-twenties
can an awkward time-- not knowing where you'll be in two years, and if you're dating, there's a good chance
they won't know where they'll be either. And for the most ambitious young couples, work lives align as
often as often as the planets. My friend Tara had just switched from a career in physics to a career
teaching high school science, which she believed would be more-feasible given that her fiancé was a naval
architect. An old dorm-mate in apparel design had trouble finding a job in DC after her husband got a job
as a lobbyist, and ended up dropping out of the labor force.

As a first year-graduate student, the experiences of my fellow PhD students and friends bore little
resemblance to the standard models of job search I studied. Job search models, such as those recognized by
this year’s Nobel Prize, treat workers as “atomistic” agents who make job decisions and earn money alone.
What was missing, I felt, were the broader implications of a labor market where men and women choose to
between jobs requiring frequent and calculated moves and those that can be performed anywhere.

Here’s an example of what I mean.

Geographic Concentration of Physicists

This map, which I created from 2000 US Census data, shows the concentration of astronomers and physicists
in US metropolitan areas. Upon showing this to Tara, she instantly recognized what the map represents,
pointing out Oak Ridge National Laboratory in Tennessee, Pacific Northwest National Laboratory in
Washington, and NASA’s Space Flight Center in Alabama. Her dream job, where she spent her summers, was in
Los Alamos, New Mexico (not many marine engineers there). A particle physicist, she noted, might bounce
from one lab and one postdoc to the next, completing graduate school as a late-twenty-something and perhaps
landing a permanent position in his or her (but usually his) mid-thirties.

Geographic Concentration of Elementary Teachers

Here’s a map of teachers. “Yeah, teachers are everywhere,” she shrugged.

And lastly, here her husband’s fellow marine engineers and naval architects.

Geographic Concentration of Marine Engineers

Comparing this graph to the first, she noted few few cities a naval architect might rub shoulders with
a physicist.

And so, I mused over this project's first big question.

Big Question One: If you could look across all occuaptions, would you find that men segregate into
geographically-clustered occupations that involve frequent relocations, while women segregate into
geographically-flexible occupations robust to spousal moves?

Innervated by curiousity, I went to my favorite pastime—-crunching labor market data. For all 474 Census
occupations, I calculated how many workers would need to move so that the occupation would have the same
number of workers per capita in every US. city. For physicists, that’s 45%. For marine engineers, that’s
63%. For primary school teachers, who are already widely-dispersed, that’s a mere 6%. I called this number
the occupation’s clustering index and graphed it against the female share.

Sex-Based Occupational Clustering

I wasn’t sure what to expect, but found that, regardless of the reason, occupations dominated by men do
tend to be more geographically-clustered than those dominated by women. Examining within dual-career
couples, I found that husbands with occupations that are more-clustered than his wife’s outnumber the
reverse two-to-one.

You may be wondering-- if dual job searches are such a burden, why would anyone, including men, enter
geographically-clustered occupations? It turns out that men in clustered occupations tend to have much
higher earnings than dispersed ones-—about 20-30% for highly-educated early and mid-career men. Young,
unmarried, men and women who enter geographically-clustered occupations look about the same: both relocate
for work more-often, and earn higher salaries than their peers in geographically- flexible occupations.

However, by mid-career, the story changes. Upon marriage, young men in clustered occupations continue to
relocate for work at nearly the same rate as those who are unmarried. However, women’s propensity to move
for work plummets. Similarly, while the earnings of men in clustered occupations continue to thrive,
earnings for women tend to drop below those of their dispersed counterparts. The geographical distribution
of a job, it seemed, identified high-paying jobs that penalized women and that women avoided.

Big Question Two: Do mobility, earnings, marriage, and divorce patterns implicate the dual body problem as
the root cause for occupational segregation and the pay gap?

As is common when a young researcher becomes fixated on an idea, I began to see it everywhere: in couples
actively balancing two careers, in students planning careers and children around graduations, in lunch
conversations with my fellow (unmarried) friends over whether the careers of dates are compatible, and even
in undergraduates who are concerned that their career ambitions were in conflict with their ultimate life
goals. One of the nice things about practicing medicine, rather than doing medical research,” someone doing
his residency explained to me, “is that you can do it anywhere. So what are you researching?”

And so, I went back to number-crunching. Beside impaired earnings, women entering geographically-clustered
occupations also appear to make life sacrifices, tending to marry two or three years later than those
in dispersed occupations, and tending to have higher incidence of divorce. The effect of occupational
choice men’s marriage timing, if anything, is the reverse. Of course, this doesn't account for stay-at-home
parents who focus on domestic work and child-rearing-- the ultimate "dispersed" occupation, and one where
women outnumber men eight-to-one.

Big Question Three: Are young men and women's family and career expectations self-fulfilling?

In her TED Talk and highly-circulated article in The Atlantic, Hanna Rosin cites recent labor market
trends as indications of the coming “End of Men:” “For every two men who get a college degree, three women
will do the same. Women […] became the majority of the American workforce, and they are starting to
dominate lots of professions: doctors, lawyers, bankers, accountants.” I couldn’t help but notice that her
first two examples were the two most-dispersed occupations where the majority of workers have graduate
degrees, and indeed are occupations women have made great strides. The majority of financial managers,
accountants, and auditors recently became majority-female, and they too rank among the most
highly-dispersed occupations. Indeed, women’s greatest strides have occurred in those occupations that
are most dispersed. Even as our society and labor market evolves, perhaps the family dynamics that
underlie occupational sex segregation stay the same.

Throughout my account, I've shared anecdotes of how the dual job search shaped men and women’s lives. But
to the extent these anecdotes are common, cannot they also shape our expectations? To the extent men might
expect to marry someone who will follow him and women might prospect marrying someone whose career choices
were made under this pretense, it would seem the rationale behind educated young men and womens' chosen
career trajectories are aggregately self-fulfilling.

Interested in learning more? Read the full research paper or contact me at alanmb AT mit DOT edu.

Back to my homepage .