Gender Differences in Childcare: Time Allocation in Five European Countries

Abstract

This article analyses the intrahousehold allocation of time in households headed by heterosexual couples to show gender differences in childcare in Denmark, France, Germany, Italy, and Spain. Using data for the five sample countries from the European Community Household Panel (ECHP; 1994–2001) and the framework of a general efficiency approach, each parent's hours spent on childcare are regressed against individual and household characteristics. Empirical results show a clear inequality in childcare between fathers and mothers, with this disparity being more evident in Mediterranean countries. Panel data estimates reveal that, in general, caring tasks are mainly influenced by the presence of young children in the household, by the total nonlabor income, and by the ratio of mothers' nonlabor income to family's nonlabor income, with this latter variable exhibiting different behavior across genders and across countries.

Keywords:

• Childcare
• gender differences
• intrahousehold allocation
• time use
• JEL Codes: D13, J22, C33

Acknowledgments

This paper was partially written while José Alberto Molina was Visiting Fellow at the Department of Economics of the University of Warwick (UK), to which he would like to express his thanks for the hospitality and facilities provided. Thanks are also due to the editors of Feminist Economics, as well as the three anonymous referees. Finally, the authors are grateful for the financial support provided by the Spanish government (Grant ECO2008-01297). The usual disclaimer applies.

Notes

¹ Mark Aguiar and Erik Hurst (2007) use alternative measures of leisure to find a different behavior in the US across educational groups, which result in an aggregate increase of leisure time for both men and women over time.

² This result is also found for the US by Kalenkoski, Ribar, and Stratton (2007) and for Switzerland by Alfonso Souza-Poza, Hans Schmid, and Rolf Widmer (2001). However, Korenman, Liao, and O'Neill (2005) do not find influence of own wages on time devoted to childcare in the US.

³ Consequently, we cannot develop a full structural model containing all possible uses of time, as exists in other studies which use Time Use Surveys (Korenman, Liao, and O'Neill 2005; Connelly and Kimmel 2007; Kato and Matsumoto 2007; Kimmel and Connelly 2007).

⁴ Finland is also included in the ECHP, but only since 1996. Sweden has not been considered in our study, since information about hours devoted to childcare is not available. Although the case of the UK has also been excluded, previous evidence for this country exists (see, for example, Kalenkoski, Ribar, and Stratton [2005]).

⁵ For more information on this database, see Bettio and Plantenga (2004).

⁶ See Bernard Fortin and Guy Lacroix (1997); Pierre-Andre Chiappori, Bernard Fortin, and Guy Lacroix (2002); Bittman et al. (2003); Leora Friedberg and Anthony Webb (2006), among many others. One measure seldom used is the difference in age between husband and wife (see Michael Myck, Olivier Bargain, Miriam Beblo, Denis Beninger, Richard Blundell, Raquel Carrasco, Maria-Concetta Chiuri, François Laisney, Valérie Lechene, Ernesto Longobardi, Nicolas Moreau, Javier Ruiz-Castillo, and Frederic Vermeulen [2006]), with this being inconsequential in our study. We thank one anonymous referee for calling our attention to this point in a previous version of this paper.

⁷ Adults in the survey are considered those individuals 16 years or older. There is no other information about the number of children by age groups (under age 6, etc.).

⁸ As shown in the studies for the US by Connelly and Kimmel (2007) and Kimmel and Connelly (2007), childcare exhibits a behavior quite close to that of paid work, in the sense that first, the number of hours devoted to each activity have increased over time (especially in women), and second, time devoted to paid work and childcare reacts positively to changes in wages.

⁹ See also Chris van Klaveren, Bernard M. S. van Praag, and Henriette Maassen van den Brink (2006) and Martin Browning and Metter Gortz (2006).

¹⁰ Álvarez and Miles (2003) opted to eliminate the time spent in paid work from the analysis to avoid endogeneity in the estimation of the determinants of time spent in housework. We would prefer to deal with such endogeneity by applying instruments.

¹¹ We thank anonymous referees for their suggestions regarding the treatment of endogeneity.

¹² Despite problems of endogeneity being somewhat intractable to a simple resolution, our data allows us to instrument variables, with the results, presented in Table A2 in the Appendix, not varying substantially from those obtained without the instrumented variables.

¹³ By contrast, Connelly and Kimmel (2007) find that a higher value of the ratio increases the share of childcare by mothers in the US.

¹⁴ With respect to evidence in other EU countries not included in this paper, Kalenkoski, Ribar, and Stratton (2005), for the UK, using an exogenous indicator variable for the receipt of nonlabor income, find no influence of family income on childcare.

¹⁵ Note that OLS estimates are generally found to be statistically significant at the 5 percent level. A remarkable result is that more paid hours worked by mothers implies less time in childcare by them, which is statistically significant at the 5 percent level in all countries when endogeneity is not instrumented; this result is observed only in Germany when endogeneity is dealt with.