Housework and earnings: intrahousehold evidence from Latin America

ABSTRACT

This paper analyzes the intrahousehold allocation of housework and paid work in five Latin American countries. Prior work has consistently shown that income plays a major role in the region’s large gender gaps in the distribution of unpaid work at the aggregate level. However, the extent to which earnings shape intrahousehold decisions regarding the allocation of unpaid work remains unexplored. Using harmonized time-use surveys for Chile, Colombia, Mexico, Peru, and Uruguay, we analyze the relationship between earnings and housework drawing on the framework of the dependency, gender deviance neutralization, and autonomy. We find that in Latin America, increases in women's absolute earnings are related to decreases in the hours women devote to housework. At the same time, the allocation of men’s time into housework does not seem to be related to their own or their partners’ earnings. Against our expectations, differences in contextual gender inequality across countries does not seem to be relevant. These findings help us assess how well existing theories, formulated to account for phenomena of the developed world, apply to more unequal contexts that have higher levels of gender inequality and where a high proportion of women are excluded from paid work.

KEYWORDS:

• Unpaid family work
• time use
• housework
• intrahousehold gender gaps
• Latin America

JEL:

• C81
• D13
• C83

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 The index is calculated based on seven indicators included in the World Value Surveys (WWS). For each indicator a variable takes the value of 1 when an individual has a bias and 0 when the individual does not, and results are then aggregated in a composite index which reflects the percentage of people with biases.

2 Traditional methods to deal with joint endogeneity (instrumental variables, two-stage least squares, or panel data) have been scarcely used. Among the exceptions, Connelly and Kimmel (2007) instrument spouses' wages through standard first-stage procedures to reduce endogeneity problems, whereas Killewald and Gough (2010) exploit the panel structure of their data. Bloemen and Stancanelli (2014) also attempt to control for endogeneity—in their analysis of the effect of partners’ wages on partners’ allocation of time—by estimating a model comprising ten simultaneous equations that uses job characteristics to identify wages. Carlson and Lynch (2017) estimate a structural equations model and two-stage least-stage regressions to deal with possible reciprocal causality between personal earnings and housework.

3 A separate body of research, not summarized here, has suggested reverse causality; time spent on housework has a negative effect on both men’s and women’s earnings.

4 The sum of housework and paid work in Figure 1 does not add to total work, as we are not considering time devoted to unpaid care.

5 For Colombia and Mexico, homes with domestic service are those where a domestic worker lives with the family, so it captures only a part of outsourcing. In the case of Peru and Uruguay, it includes both the cases where a domestic worker lives in the household and also the situation where household chores are done with the help of an external worker. In Chile, the survey item identifies households where housework was done by domestic workers during the last week.