Productivity or Gender? The Impact of Domestic Tasks Across the Wage Distribution
| Author | Jennifer L. Hook,Lynn Prince Cooke |
| Published date | 01 June 2018 |
| Date | 01 June 2018 |
| DOI | http://doi.org/10.1111/jomf.12467 |
L P C University of Bath
J L. H University of Southern California∗
Productivity or Gender? The Impact of Domestic
Tasks Across the Wage Distribution
This article assesses the wage impact of domes-
tic tasks across women’s and men’s wage dis-
tributions given the cross-distribution variation
in unpaid work time. The productivity–volume
versus the gender–class normative argument
developed here suggests competing hypotheses.
Analyses of pooled 2010–2015 waves of the
American Time Use Survey using unconditional
quantile regression revealed that an increase
in the lesser time women at the top of the wage
distribution spent doing routine housework pre-
dicted a smaller wage penalty than at the bottom
of women’s wage distribution. Conversely, men
at the top of their wage distribution spent the
least time doing routine tasks, but incurred
the largest penalty for an increase in that time.
Increases in nonroutinehousework or child-care
time did not negatively affect the wage distribu-
tions of women or most men. Results supported
the volume–productivity argument for routine
housework among women, but a gender–class
normative argument for men.
Researchers usually take one of two causal
perspectives when exploring the relationship
between wages and domestic time. Some test
Department of Social and Policy Sciences, University of
Bath, Bath BA2 7AYUK (l.p.cooke@bath.co.uk).
∗Department of Sociology, Universityof Southern
California, Los Angeles, CA 90089-2539.
Key Words: housework, income and earnings, unconditional
quantile regression.
Becker’s (1985) argument that time in unpaid
tasks undermines employment productivity and
predicts a wage penalty (Bryan & Sevilla-Sanz,
2011; Coverman, 1983; Hersch & Stratton,
1997; Noonan, 2001). Others assess Gupta’s
(2006, 2007) autonomy perspective asserting
that higher wages allow individuals to opt or buy
out of routine housework (Craig, Perales, Vidal,
& Baxter, 2016; Gupta, Evertsson, Grunow,
Nermo, & Sayer, 2010; Hook, 2017). Carlson
and Lynch (2017) showed that both perspectives
are simultaneously valid for routine housework
among U.S. married individuals, although there
are gender differences in effects. Wives’ wages
and housework time are reciprocally related,
whereas the causal direction for husbands is
from housework to wages only (Carlson &
Lynch, 2017). This article also brings together
the two causal perspectives, but to theorize and
assess how, given the wage-related variation in
housework and child-care time, the magnitude
of the associated wage penalty varies across
U.S. women’s and men’s wage distributions.
One reason why the impact of domestic tasks
might vary across the wage distribution is that
productivity is affected only after the volume of
housework exceeds some threshold (Hersch &
Stratton, 1997). For example, women have been
argued to incur larger wage penalties for core
housework time because they spend appreciably
more time doing these tasks than men (Hersch,
2009; Hersch & Stratton, 1997, 2002; Noonan,
2001). Findings from the autonomy literature
Journal of Marriage and Family 80 (June 2018): 721–736 721
DOI:10.1111/jomf.12467
722 Journal of Marriage and Family
suggest a novel way to test the volume argu-
ment among women and among men that we
undertake here. As higher wages predict women
spend less time in core housework (Gupta, 2006;
Hook, 2017; Killewald & Gough, 2010), a vol-
ume argument is supported if the wage penalty
for housework decreases across women’s wage
distribution. Conversely, the minimal wage vari-
ation in men’s housework time should lead to
little variation in the housework penalty across
men’s wage distribution.
An alternative theoretical argument devel-
oped here is that gender differences in the
cross-distribution wage effects of domestic tasks
reect and reinforce material hierarchies of the
gender–class stratication system (Acker, 1990;
Ridgeway & Correll, 2004; Risman, 2004).
Gendered divisions of labor reinforce widely
held cultural beliefs about paid work as inherent
to masculinity and unpaid work as inherently
feminine (Ferree, 2010; Sullivan, 2013; West
& Zimmerman, 1987). Relatedly, organizations
continue to value career-committed, inherently
masculine ideal workers unencumbered by
family demands (Acker, 1990; Bianchi, Robin-
son, & Milkie, 2006; Williams, Blair-Loy, &
Berdahl, 2013). Employer expectations of work
devotion are particularly high for executives,
with executive men still assumed to have a
stay-at-home wife to handle family demands
(Blair-Loy, 2003; Williams, 2000). So in con-
trast to the volume–productivity argument, the
work devotion expectation among high-wage
workers suggests that wage penalties for time in
unpaid domestic work should be larger at the top
of the wage distribution irrespective of howlittle
time high-wage women or men might spend
doing these tasks. The gender–class normative
pressures on elite men in particular suggest
that wage penalties for domestic time will be
greatest at the top of men’s wage distribution.
Estimating “average” wage effects of domes-
tic tasks as done in all tests of the productivity
argument to date provides no insight into how
effects vary across women’sand men’s wage dis-
tributions. Instead, we select a two-step uncon-
ditional quantile regression (UQR) procedure
developed by Firpo, Fortin, and Lemieux (2009).
As discussed more in the Method section, the
advantage of UQR for our analysis is that it
provides ordinary least squares (OLS) estimates
of the effect of an identically sized increase in
individuals’ existing domestic time at different
quantiles (percentiles) of the unconditional wage
distribution, controlling for other covariatesused
to predict wages. Hence we pool 2010–2015
cross-sectional waves of the American TimeUse
Survey (ATUS) and use UQR to estimate the
wage impact of unpaid domestic time across
women’s and men’s wage distributions. Time
diary data are ideal for this task as they con-
tain measures to distinguish core from nonrou-
tine housework as well as child care. The data
also allow us to focus analyses on the effects of
unpaid activities on employment days when they
have the greatest likelihood of affecting produc-
tivity (Hersch, 2009; Hook, 2017). These excel-
lent data and our innovative analytical approach
reveal between- as well as within-gender varia-
tion in the wage effects of domestic tasks that
reinforce gender–class systems of stratication
only partially related to productivity.
T U T–W R
Impact of Wages on Domestic Time
Researchers became interested in the impact of
earnings on domestic tasks with the accruing
evidence that women’sgreater relative resources
do not predict egalitarian divisions of couples’
housework as indicated by bargaining models
(Bittman, England, Folbre, Sayer, & Matheson,
2003; Brines, 1994; Greenstein, 2000). Gupta
(2006) instead proposed an “autonomy” model,
arguing a woman’s absolute income is a more
important determinant of her domestic time, in
part because money provides the possibility of
purchasing market substitutes (see also Cohen,
1998; Craig et al., 2016). Analyzing the National
Survey of Families and Households (NSFH),
Gupta (2006) illustrated that both married and
single women’s time in routine housework
decreased as their earnings increased. Limit-
ing analyses to partnered women, he further
found that the predicted impact of wives’
absolute earnings was signicantly larger than
the impact of their relative share of house-
hold earnings (Gupta, 2007). Gupta (2006)
concluded that these effects highlight a class
dimension of housework time among women,
with high-wage women able to purchase greater
domestic time equality with men than lower
waged women.
Nevertheless, his analyses also revealed that
even high-wage women could not eliminate
daily housework entirely (Gupta, 2006, p. 988).
Later analyses with different data sets conrmed
To continue reading
Request your trial