Who Uses Paid Domestic Labor in Australia? Choice and Constraint in Hiring Household Help

Abstract

This paper investigates why some Australian households use paid help with domestic labor while others do not. Consistent with earlier studies, the analysis examines hypotheses relating to resources, objective demand, and gender attitudes. Additionally, this paper examines the impact of attitudes toward using paid domestic labor, an area that has not been investigated previously. The findings suggest that resources and objective demand provide the parameters within which employing household help is made possible or necessary, but beliefs about the appropriateness of this strategy also play a role in determining whether Australian households use paid domestic labor. The paper concludes that understanding whether Australian households pay for domestic help is dependent not just on the level of resources and objective demand but also on whether individuals view paid domestic help as an appropriate strategy for undertaking domestic work.

Keywords:

• Paid domestic labor
• housework
• attitudes
• gender
• JEL Codes: J12, J16, J22

Notes

¹ One exception is Oropesa (1993) who examined the impact of attitudes on paying for domestic labor in the United States. The attitude items included how uncomfortable the respondent would feel if their house was not clean, as well as the respondent's degree of attachment to cooking and meal-related activities (I like to cook; the kitchen is my favorite room; I feel guilty serving convenience foods; and meal preparation shouldn't take long). The results show no effect on housecleaning-service purchases, although there is some evidence that those who cite the kitchen as their favorite room are less likely to eat out. It is, however, not clear exactly what the service-related attitudes measure. As Oropesa notes, feeling uncomfortable if one's house is not clean may actually mitigate against hiring a cleaner since this may involve leaving the house untidy until the designated cleaning day, as well as negotiating over standards. Moreover, it was unclear whether the items in Oropesa's work suggested positive or negative views toward hiring paid domestic help.

² This paper focuses specifically on heterosexual couples as there were insufficient numbers of same-sex households for meaningful analyses.

³ There were seven respondents with missing data on this question who were dropped from the sample.

⁴ In earlier analyses, we examined a measure of equivalised household income, that is, income adjusted for household size. But the results were very similar to those for the unequivalised measure. Since we include a variable for the number of children in the household, we decided to present the unequivalised result. We also investigated disaggregating husbands' and wives' incomes. Other studies have shown that women's earnings may be more important than men's in the allocation of household labor (Sanjiv Gupta 2006), while others have found greater effects of women's incomes than men's on outsourcing expenditure (Oropesa 1993; Cohen 1998). But our results show only mixed support for these earlier findings. Although we find evidence that both men's and women's incomes are important in predicting domestic outsourcing in the earlier models, in the full models with all variables men's income is associated with using paid domestic labor while women's income is not (results not shown). Thus counter to recent studies, our research indicates that men's incomes, not women's, are most important for determining the use of paid domestic help. This result requires further investigation in future studies.

⁵ Intuitively, one might expect that the same interactions should be found for women and men because we have measures of both partners' characteristics and relevant household factors. However, we do not have data on partners' attitudes; thus, in the male equation, we control for men's attitudes, and in the female equation we control for women's attitudes. The models are therefore not symmetrical by gender, and it is logically possible to find different interactions for men and women. The asymmetry in the models is another reason to conduct separate analyses by gender.

⁶ To derive this graph, we assumed men of average age, working forty hours per week, with vocational qualifications, average gender attitudes, other service beliefs that were opposed to paid domestic labor, two children under 18 living at home, no children under 5, partners with good health working twenty hours per week, and vocational qualifications. We then varied household income in Australian dollars from $0 to $206,000 (the 99^th percentile) and derived the predicted probabilities, assuming men agreed or disagreed with the statement “You really should be doing it yourself.” Choosing other values for the covariates changes somewhat the levels of the probabilities and the shape of the graph but not the starkness of the interaction between belief and income.

⁷ In this equation, we assume the same values on the covariates as we did for men, except that we allow women's hours of paid employment to vary between zero and seventy (the 99^th percentile), and we assume that their partners work average hours. We also assume an average household income.

⁸ The predicted probabilities for women are much lower than for men, because we have assumed an average household income in the right panel of the figure (approximately AU$34,000) and allowed income to vary freely in the left panel of the figure. Household income is strongly associated with paying for domestic help for women and men.