What women want (their men to do): Housework and Satisfaction in Australian Households

ABSTRACT

The time allocated to household chores is substantial, with the burden falling disproportionately upon women. Social norms about how much housework men and women should do are likely to influence couples’ housework allocation decisions and satisfaction. Using Australian data spanning 2001–14, this study employs a two-stage estimation procedure to examine how deviations from housework norms relate to couples’ satisfaction. The study finds that satisfaction is negatively affected by predicted housework time and that women’s satisfaction, but not men’s, is robustly affected by their partners’ residual housework time. When he exceeds housework norms, she is happier with housework allocations, but less happy in broader dimensions. The study suggests several reasons for the results, including that housework is more salient in women’s lives than in men’s, that housework generally is not a preferred activity, and that some degree of gender-norm conformity in regard to housework can positively affect women’s life satisfaction.

KEYWORDS:

• Satisfaction
• social norms
• housework
• Australia

JEL Codes:

• D13
• I31
• Z13

ACKNOWLEDGMENTS

This paper uses unit record data from the Household, Income and Labour Dynamics in Australia (HILDA) Survey. The HILDA Project was initiated and is funded by the Australian Government Department of Social Services (DSS) and is managed by the Melbourne Institute of Applied Economic and Social Research (Melbourne Institute). The findings and views reported in this paper, however, are those of the authors and should not be attributed to either DSS or the Melbourne Institute. We thank Deborah Cobb-Clark, Paul Frijters, Joyce Jacobsen, Charlene Kalenkoski, Terra McKinnish, Edward Millner, Paco Perales Perez, and seminar participants at Monash University, the University of New South Wales, Virginia Commonwealth University, Deakin University, and the 2016 ANU-hosted Labour Econometrics Workshop for helpful comments on earlier drafts of this paper. We are also greatly indebted to James Stratton for outstanding research assistance. All errors remain ours.

SUPPLEMENTAL DATA

Supplemental data for this article can be accessed at http://dx.doi.org/10.1080/13545701.2019.1609692.

Notes

1 For a review of the broader literature in sociology regarding household labor – its measurement, division within the household, and associations with economic and psychological outcomes – see Beth Shelton and Daphne John (1996).

2 If, as beautifully articulated by Candace West and Don H. Zimmerman, “the ‘doing’ of gender is undertaken by women and men whose competence as members of society is hostage to its production” (1987, 126), then by implication men who “do” more female-ness (for example, by allocating more time to housework) risk being perceived as being less competent members of society – even, presumably, by their partners.

3 Bootstrapping in this case is necessary to correct the standard errors of the coefficients in the second-stage models because four variables used in those models (namely, his and her predicted and residual housework times) are in fact values that we predict in the first stage. We draw repeated random samples from our data with replacement, and generate successive estimates of the predicted and residual housework values from stage one using each of those samples, that we then plug into our stage-two models. The empirical distribution of standard errors for the second-stage coefficient estimates that we recover from repeating this procedure 100 times forms the basis for our bootstrapped standard errors.

4 Persons younger than age 20, men older than age 64, women older than age 61, and 20–23-year-olds enrolled full time in higher education are excluded. The different age restrictions by gender approximately reflect the different ages at which men and women are eligible to receive pensions in Australia. Fewer than 5 percent of couples are excluded due to inconsistent or incomplete information regarding the relationship.

5 Non-response by the household or individual, or failure to complete the self-reported questionnaire on which housework time is reported, together account for 60 percent of the lost couple-year observations and 27 percent of the lost couples. Singleton reports account for 12 percent of the lost couple-year observations and 60 percent of the lost couples. Couples who jointly report no time spent on housework or individually report more than 70 hours on housework are also dropped, as we judge these reports not to be credible. This criterion excludes only 421 outliers (less than 1 percent of couple-year observations).

6 This question is answered to the nearest minute in all HILDA waves except the first; in 2001, it is answered to the nearest hour. In our models, any difference in average measured quantity of housework caused by this change in granularity across reporting years is captured by year dummies.

7 At both the individual and household levels, hours spent on housework are positively associated with hours spent on these other forms of unpaid labor.

8 Analysis of satisfaction-with-housework is consequently restricted to 22,322 observations (4,240 couples). These individuals have characteristics very similar to those of the full sample, though they do tend to be a bit younger and more educated.

9 Christopher Ambrey, Jennifer Ulichny, and Christopher Fleming (2017) report evidence generated using the same dataset that the satisfaction of both Australian men and Australian women has fallen over time (attributing this fall to a decline in social connectedness). We also find a decline over time for both genders in our broader satisfaction measures.

10 The incidence of individuals responding that they spend no time on housework is sufficiently uncommon (less than 0.4 percent for women and less than 7 percent for men) to make nonlinear estimation unnecessary.

11 These satisfaction measures are recorded on an ordinal scale. Ada Ferrer-i-Carbonell and Paul Frijters (2004) note that fixed effects models of satisfaction generally produce similar results to ordinal latent-response models (for example, logit or probit).

12 For women, the within-couple variation in predicted housework driving these results is attributable primarily to changes in household composition. Approximately 70 percent of the within-couple variance in women’s housework is explained by changes in the number of children ages 0–4 – that is, new births. For men this factor accounts for only 12 percent, while changes in his or her disability status explain 22 percent. That onset of a disability would decrease satisfaction, at least in the short run, is in line with prior results from the literature on life satisfaction (Pagán-Rodríguez 2010). Andrew E. Clark, Ed Diener, Yannis Georgellis, and Richard E. Lucas (2008) and Paul Frijters, David W. Johnston, and Michael A. Shields (2011) find that births are associated with increases in life satisfaction in the short run, but are also anticipated, such that the actual birth may precipitate a decrease in life satisfaction because satisfaction is temporarily higher directly prior to a birth.

13 The share of household housework time supplied by women in two households may be the same (say 70 percent) when the hours spent are quite different (say 7 hours in a household reporting 10 hours of housework, versus 21 hours in a household reporting 30 hours of housework). Spending 21 hours on housework constitutes a much greater burden than spending 7 hours on housework, which may have important implications for satisfaction. This is why in our baseline first-stage results we predict reported hours spent, rather than share.

14 This story is consistent with the findings in Gigi Foster and Leslie S. Stratton (2018) that document different gendered norms by education when it comes to housework responsibilities.

Additional information

Notes on contributors

Gigi Foster

Gigi Foster is Professor with the School of Economics at the University of New South Wales, having received her BA from Yale University, majoring in ethics, politics, and economics, and her PhD in economics from the University of Maryland. She works in many economic subfields, including education, social influence, behavioral economics, lab experiments, and time use.

Leslie S. Stratton

Leslie S. Stratton received her PhD in Economics from the Massachusetts Institute of Technology in 1989 and is currently Professor at Virginia Commonwealth University in the US. She is an applied econometrician working primarily in the labor economics field. Her research has centered on the time allocation decisions of individuals and households. As such, she has explored intrahousehold time allocation decisions, marital wage differentials, labor supply decisions, and postsecondary enrollment patterns.