The effects of schooling on parental time in education production* Jeff DeSimone, University of South Florida and NBER Angela Dills, Clemson University November 2005 Abstract This study uses recently released American Time Use Survey (ATUS) data to examine whether parents with greater educational attainment allocate more time towards the educational activities of their children. This question is inherently empirical because theoretical predictions about the direction of the relationship are ambiguous: net of correlations with preferences, opportunity costs, and resource availability, human capital and efficiency effects of schooling act in opposite directions. Regression results show that after controlling for a wide array of demographic and geographic characteristics including a variety of family structure and budget constraint measures, an additional year of schooling is associated with increases in time spent with children of 12 percent on reading, four percent on homework and two percent overall. However, the nature of the relationship differs across activity categories. For reading, virtually the entire relationship occurs on the extensive margin (i.e. spending any time with children) rather than the intensive margin (i.e. time spent conditional on allocating any time). The schooling gradient is large and significant for parents of both pre-school and school age children, married parents regardless of work status, both mothers and fathers, and both the school year and summer, and is larger at higher levels of schooling. For homework, effects are equally divided between the extensive and intensive margins but are marginally insignificant in each case, and are strongest among parents of school-age children, married parents who work, single parents, and mothers, and during school year weekdays. For overall time, the impact is primarily on the intensive margin, but is significant for both margins. The variation in importance across demographic groups is similar to that for homework time, but the relationship is driven by time spent during weekends, holidays and summer. Our failure to reject the hypothesis that schooling coefficients are equivalent for degree and non-degree years suggests that they at least partially reflect a human capital effect rather than unobserved heterogeneity with determinants of graduating from various levels. * We are grateful to Jason Fletcher, Lawrence Getzler, Mark Long, Daniel Player, and Duc-Le To for detailed suggestions and to other participants in a session at the 2005 APPAM meetings for helpful comments.1 1. Introduction This paper examines the amount of time parents spend in educational production at home using recently released data from the American Time Use Survey (ATUS). In particular, we investigate the association between years of schooling and parental time spent reading with their children, helping them with homework and in any activity with them. It is well-established that family background is an important predictor of student achievement. Hoxby (2001), for example, finds that family variables account for 90–93 percent of the variation across individuals in income, educational attainment, and 12th grade math scores. More educated parents clearly raise more educated children. This relationship appears to be both genetic and causal. There is weak, but growing, evidence that increasing parental education causally improves educational outcomes of their children. Behrman and Rosenzweig (2002) difference out monozygotic twins’ genetics and find that children of better educated mothers attain less schooling. However, Antonovics and Goldberger (2003) show that these results are sensitive to data coding. Moreover, Plug (2004) finds no significant effect of mother’s schooling on that of her adopted child, although children adopted by higher educated fathers attain more schooling. Other studies use compulsory schooling laws to estimate causal effects of parental education on children’s education, obtaining mixed results. Chevalier (2004) finds a positive effect of maternal educational attainment, but no effect of paternal attainment, on child schooling. Oreopoulous et al. (2003) estimate that increasing the education level of either parent reduces the likelihood that a child is held back in school. Using Norwegian data, Black et al. (2005) find that increasing maternal schooling raises child schooling.2 Little is known, however, about the mechanisms through which additional parental schooling improves the educational outcomes of their children. This paper investigates one such mechanism: we consider whether more educated parents spend more time in educational production. Because economic theory does not unambiguously predict the direction of this relationship, our question is inherently an empirical one. There are at least three reasons to expect a positive relationship between parental schooling and time in home education production. First, additional schooling might transmit an understanding that the marginal benefit of additional educational time with a child is larger than would otherwise have been assumed. Second, schooling could directly increase the marginal benefit of additional educational time at any time allocation amount. Third, higher educational attainment almost certainly reflects stronger preferences for educational activity that will also be manifested in spending more time in child educational production. An important distinction between these explanations is that the first two reflect causal impacts of schooling on time in educational production, while the third is symptomatic of a spurious correlation between schooling and child educational time through preferences regarding educational activity. Results from some of the previous related literature are consistent with these theoretical arguments. In early time-use studies focusing on maternal time use in two-parent families, Hill and Russell (1974 and 1980) find that mothers with greater educational attainment spend more time with their children. Leibowitz (1977) suggests that more highly educated mothers participate in activities, such as reading, that improve child verbal abilities. Datcher-Loury (1988) estimates that schooling increases the effectiveness of maternal child care time in increasing child educational attainment.3 Evidence on whether education increases productivity in home production is mixed.