A Theory of Dual Job Search and Sex‐Based Occupational Clustering

• Author: Alan Benson
• Date: 01 July 2015
• DOI: http://doi.org/10.1111/irel.12095
• Published date: 01 July 2015

A Theory of Dual Job Search and Sex-Based

Occupational Clustering*

ALAN BENSON

This paper theorizes and provides evidence for the segregation of men into clus-

tered occupations and women into dispersed occupations in advance of marriage

and in anticipation of future colocation problems. Using the Decennial Census,

and controlling for occupational characteristics, I find evidence of this general pat-

tern of segregation, and also find that the minority of the highly educated men

and women who depart from this equilibrium experience delayed marriage, higher

divorce, and lower earnings. Results are consistent with the theory that marriage

and mobility expectations foment a self-fulfilling pattern of occupational segrega-

tion with individual departures deterred by earnings and marriage penalties.

Introduction

In 2010, 67 percent of U.S. families with at least one labor force participant

featured dual-earner couples, up from 41 percent in 1980.

1

Given the frequency

with which young, career-oriented workers relocate for work or schooling and

the impact these decisions may have on the career trajectories of the partner, it is

not surprising that researchers’attention to the colocation problem has also

grown. Mincer (1978) and Sandell (1977) introduced a neoclassical approach to

household relocation decisions, noting that decisions that optimize the family’s

net career outcomes may impair the careers of tied stayers or tied movers. Frank

(1978) hypothesized that the co-location problem will concentrate women in

*The author’saffiliation is Carlson School of Management, University of Minnesota, Minneapolis, Minne-

sota. Email: bensona@umn.edu.

JEL: J24, J16, J31, J12, J61.

This paper is based upon Chapter 2 of the author’s doctoral dissertation at the MIT Sloan School of Man-

agement. The author is indebted to David Autor for comments and feedback throughout the project, and also

thanks Dan Fehder, Stephanie Hurder, Thomas Kochan, Ferran Mane, Matt Marx, Kathy McGinn, Paul

Osterman, Al Roth, participants in the MIT Economics Labor Lunch, MIT Economic Sociology Working

Group, the 2011 LBS Transatlantic Doctoral Conference, the 2011 IWAEE conference, the 2011 MOOD

conference, the 2011 People and Organizations conference at Wharton, and colleagues at the Institute for

Work & Employment Research for comments. Replication materials are available upon request. The usual

disclaimer applies.

1

Figures based on author’s calculation. Includes full-time, part-time, and unemployed labor force partici-

pants where both heads are ages 18–65.

INDUSTRIAL RELATIONS, Vol. 54, No. 3 (July 2015). ©2015 Regents of the University of California

Published by Wiley Periodicals, Inc., 350 Main Street, Malden, MA 02148, USA, and 9600 Garsington

Road, Oxford, OX4 2DQ, UK.

367

large cities. Costa and Kahn (2000) attributed the increasing concentration of col-

lege-educated “power couples”in large metropolitan areas in part to their ability

to sustain two simultaneous careers. Empirical studies have challenged the sym-

metry of these models due to evidence that household relocation decisions favor

husbands’careers. Compton and Pollak (2007) found that only the education of

the husband predicts migration to large metropolitan areas, and the higher inci-

dence of power couples in large metropolitan areas is due to higher rates of

power couple formation. Others find that household migration usually advances

husbands’careers to the detriment of their wives (Boyle et al. 2001; Clark and

Huang 2006; Jacobsen and Levin 2000; Long 1974; McKinnish 2008).

However, by emphasizing why and for whom household relocation deci-

sions are made, research on couples’job search neglects the broader chal-

lenges and opportunities presented by variation in the geographic flexibility of

jobs and potential endogeneity in occupational choice. Newly trained physi-

cists, computer scientists, naval architects, and nuclear engineers are typically

geographically constrained and require calculated moves early in their careers.

In contrast, administrative assistants, dentists, schoolteachers, nurses, and gen-

eral managers are ubiquitous and typically enjoy relative geographic flexibility.

This study asks how the co-location problem might differentially affect men

and women. If men sort into occupations that benefit from calculated moves,

does the colocation problem discourage women from doing the same?

In the first section I introduce a model whereby early specific career invest-

ments, couples’desire to colocate, and variation in the geographic flexibility

of jobs result in a coordination problem that prompts men and women to sort

into geographically clustered and dispersed occupations in advance of mar-

riage, a phenomenon I refer to as “sex-based occupational clustering.”

2

In the

model, occupations are clustered or dispersed for reasons exogenous to their

sexual composition, and highly clustered occupations compensate for the di-

samenity of geographic constraints but penalize couples with two such careers

due to the colocation problem. The model is consistent with a number of well-

known features of inequality, including the tendency for women to be tied

movers when a family relocates for work and the segregation of women into

lower-paying jobs. It also yields specific testable predictions regarding the

marriage and earnings penalties that deter desegregation, and implies that these

2

The model and timing are similar to others in which the household division of labor and child-rearing

expectations affect men’s and women’sex ante training investments (see, for example, Becker 1991; Eche-

varria and Merlo 1999; Engineer and Welling 1999; and Hadfield 1999). However, although these models’

ex ante investments follow from specialization in household production (and the inability to efficiently con-

tract domestic work), ex ante investments in the model proposed here follow from variation in the geo-

graphic constraint and flexibility across occupations.

368 / ALAN BENSON

predictions should be greatest for those with a high degree of formal

education.

I construct an occupational clustering index in the second section. Using the

2000 Decennial Census, for each occupation, I use the index to calculate the

share of workers in that occupation that would need to relocate to equalize its

workers per capita in every metropolitan statistical area (MSA). I find that

workers in clustered occupations tend to have higher rates of relocation for

work among never-married and single-earning men and women.

In the third section I provide evidence that occupational clustering is an axis

of occupational segregation: occupations dominated by women tend to be geo-

graphically dispersed, and occupations dominated by men tend to be geographi-

cally clustered. This correlation holds after controlling for several occupational

features including hours worked, mean age, physical strength requirements, the

extent to which the occupation involves assisting and caring for others, and the

required level of math. To provide intuition for the magnitude of this number,

prime-age dual-earner couples that feature a husband in an occupation that is

more clustered than his wife’s outnumber the reverse about two to one.

In the following section, I provide additional evidence that occupational

clustering is an axis of earnings and marriage inequality. Consistent with the

model’s predictions, women who enter geographically clustered occupations

experience lower earnings, later marriage, and higher divorce rates compared

with those entering dispersed occupations. Men, in contrast, enjoy earnings

premia and lower divorce rates in clustered occupations. I find that the results

for occupational segregation, earnings inequality, and marriage delay and

divorce are primarily driven by individuals in occupations dominated by work-

ers with a bachelor’s degree or higher. This suggests the model’s results are

most consistent where the costs of switching occupations (for example, follow-

ing geographic displacement) are highest.

The study aims to introduce a framework for thinking about the colocation

problem in a life-course setting and to examine whether the geographic distri-

bution of jobs appears to be an empirically (if not theoretically) useful axis

from which to analyze inequality in occupations, earnings, and marriage.

A Model of Occupational Clustering and Sex Segregation

In this section I demonstrate how a simple model that incorporates standards

of occupation choice, marriage search, and job search in a life-course setting

can yield sufficient conditions for equilibrium sex-based occupational cluster-

ing. The model considers risk-neutral workers who initially differ only in sex

and who are interested in low marriage search costs and high family earnings.

A Theory of Dual Job Search / 369