Decomposing gender wage gaps: a family economics perspective

• Published date: 01 January 2024
• Author: Dorothée Averkamp,Christian Bredemeier,Falko Juessen
• Date: 01 January 2024
• DOI: http://doi.org/10.1111/sjoe.12542

Scand. J. of Economics 126(1), 3–37, 2024

DOI: 10.1111/sjoe.12542

Decomposing gender wage gaps: a family

economics perspective

Doroth´

ee Averkamp

University of Wuppertal, DE-42119 Wuppertal, Germany

averkamp@wiwi.uni-wuppertal.de

Christian Bredemeier†

University of Wuppertal, DE-42119 Wuppertal, Germany

bredemeier@wiwi.uni-wuppertal.de

Falko Juessen†

University of Wuppertal, DE-42119 Wuppertal, Germany

juessen@wiwi.uni-wuppertal.de

Abstract

We propose a simple way to embed family-economics arguments for pay differences between

genders into standard decomposition techniques. To account appropriately for the role of the

family in the determination of wages, one has to compare men and women with similar own

characteristics – and with similar partners. In US survey data, we find that our extended

decomposition explains considerably more of the wage gap than a standard approach, in line

with our theory that highlights the role of career prioritization in dual-earner couples.

Keywords: Gender wage gap; wage-gap decomposition; dual-earner households;

discrimination

JEL classification:J31; J16; J12; J71; J24

1. Introduction

The gender wage gap decreased substantially in the second half of the 20th

century, but a persistent gap remains (see, e.g., Olivetti and Petrongolo, 2016).

On average, women in the United States continue to earn close to 20 percent

less per hour than men. As shown by, for example, Blau and Kahn (2017),

a considerable part of the wage gap can be related to observable gender

differences in individual characteristics, such as work experience, occupation,

and industry. In turn, the closure of the gender wage gap can be explained

to a substantial extent by women catching up in terms of human capital (i.e.,

education and experience). However, an open question remains as to why the

†Also affiliated with IZA.

c

2023 The Authors. The Scandinavian Journal of Economics published by John Wiley & Sons Ltd on behalf of F¨

oreningen

f¨

or utgivande av the SJE.

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution

and reproduction in any medium, provided the original work is properly cited.

4Decomposing gender wage gaps: a family economics perspective

gender wage gap is (still) so large or, put differently, why a man with the

characteristics of the average woman earns, according to the estimates of Blau

and Kahn (2017), about 7–9 percent more than the average woman does.

There are two approaches in the literature that seek to explain remaining

gender gaps. The first approach, reviewed by Bertrand (2011), Azmat and

Petrongolo (2014), and Blau and Kahn (2017), argues that gender differences

in personality traits or gender norms can lead to self-selection of women into

lower-paid jobs and less steep career paths.1Several studies have documented

that a part of the wage gap can be attributed to such factors, but their

quantitative role seems to be limited (Mueller and Plug, 2006;Leetal.,2011;

Nyhus and Pons, 2012; Heinz et al., 2016; Chen et al., 2017; Flinn et al., 2018;

Jung et al., 2018; Reuben et al., 2015; Roussille, 2022).

The second approach emphasizes the role of the family for the gender

wage gap. An important dimension is women’s relative temporal inflexibility

due to their dominant role in childcare and non-market work in many families

(e.g., Goldin, 2014;Alm

˚

as et al., 2023;Cort

´

es and Pan, 2023). At the same

time, many husbands are their families’ primary breadwinners and see their

careers prioritized in many decisions of the family such as migration decisions

(see, e.g., Mincer, 1978; Compton and Pollak, 2007; Foged, 2016; Braun

et al., 2021), the choice of employers (see, e.g., Bredemeier, 2019; Petrongolo

and Ronchi, 2020), and job-search investments (Flabbi and Mabli, 2018).

In this paper, we connect this family-based approach to the literature on

decompositions of the gender wage gap. We propose a simple way to embed

family-economics explanations for the wage gap into standard decomposition

techniques. The key common aspect of the family-based explanations is that

important family decisions induce a trade-off between spouses’ careers, and

that the family often has an incentive to prioritize the career of the spouse

with the higher earnings potential. For the individual worker, this means

that realized wages depend not only on their own characteristics but also

on whom they are married to. Two workers with identical characteristics

but different partners are treated differently by their respective families and

will thus experience different career trajectories. For decompositions of the

wage gap, whose purpose it is to compare observationally identical men and

women, family economics implies that one should compare men and women

with similar own characteristics and similar partners, in order to account

appropriately for the role of the family for wages.

1There is a significant empirical literature, mostly experimental, on gender differences in

non-cognitive abilities, personality traits, and preferences, including the willingness to compete

(Gneezy et al., 2009; Flory et al., 2015; Buser and Yuan, 2019), negotiation styles (Babcock and

Laschever, 2003; Exley et al., 2020), promotion-seeking (Bosquet et al., 2019), the willingness

to take on non-promotable tasks (Babcock et al., 2017), risk aversion (Croson and Gneezy, 2009;

Dohmen and Falk, 2011), and self-promotion (Exley and Kessler, 2022).

c

2023 The Authors. The Scandinavian Journal of Economics published by John Wiley & Sons Ltd on behalf of F¨

oreningen

f¨

or utgivande av the SJE.

D. Averkamp, C. Bredemeier, and F. Juessen 5

For the Oaxaca–Blinder decomposition approach, which remains the most

frequently applied decomposition approach of the gender wage gap, this means

that the wage equation should include the characteristics of the individual’s

partner. For instance, a worker’s own education should be included in the

equation for the worker’s wage – but also the education of the worker’s

partner to account for the effect of the partner’s education on the family’s

investment into the workers’s career. In the decomposition, one would then

capture the extent to which women’s relative wages are compressed by

their husbands’ characteristics through career-prioritizing decisions of the

family. The implication to include partner characteristics is not limited to

the Oaxaca–Blinder approach but applies to all approaches that seek to

assign a part of the wage gap to differences in observable characteristics. For

example, matching-based approaches (e.g., Meara et al., 2020; Strittmatter

and Wunsch, 2021) should include partner characteristics in the matching

process, independent of the specifics of this process.

To make our point explicit, we set up a model of dual-earner couples

deciding upon investments into spouses’ careers. The model has two

investment margins, one of which includes a trade-off between spouses’

careers and the other allows for potential positive spill-over effects of

investments into one partner’s career on the career of the other partner.

While there are many interpretations for the first channel, we frame it as a

joint location choice where couples have to compromise between locations

promoting the husband’s career and locations promoting the wife’s career.

For a couple, it is rational to prioritize the career of the spouse with the higher

earnings potential and to choose to live closer to the place that optimally

promotes the career of the spouse with the higher earnings potential. As

a consequence, the realized wage of a worker depends positively on the

individual’s own earnings potential and – through the mediator distance to

optimal location – negatively on the earnings potential of the individual’s

partner. The second investment choice, which we call the spill-over channel,

induces a positive relation between one’s own wage and a partner’s earnings

potential as a high potential of the partner may induce the family to invest

heavily into the partner’s career, from which one’s own career benefits as well.

We use the model to show that a decomposition that ignores partner

characteristics misestimates the fraction of the wage gap that is due to

observable characteristics (i.e., the explained part of the gap). Whether the

explained gap is overestimated or underestimated depends on whether, on

average, the career-prioritization or the spill-over effect is the dominant

channel from partner characteristics to wages. With positive assortative

mating along observables, the explained wage gap is underestimated when the

career-prioritization channel is dominant. Reversely, the spill-over channel

being dominant would imply that the standard decomposition overstates

the explained wage gap. We then show that extending the decomposition

c

2023 The Authors. The Scandinavian Journal of Economics published by John Wiley & Sons Ltd on behalf of F¨

oreningen

f¨

or utgivande av the SJE.