My dissertation emphasizes the role played by siblings in the determination of individual outcomes. In the first chapter of my dissertation, I study the direct effect of siblings on adolescent
outcomes, while the remainder of my dissertation considers siblings in the context of family fixed effects methods of estimation.
Longitudinal data shows that the likelihood of a child smoking more than triples if the child has an older sibling who also smoked. In my first essay, “Endogenous Sibling Effects in Adolescent Substance Use”, I unpack this striking correlation, asking whether it is causal or a result of correlated
unobservables such as parental investments that are endogenous to the child’s behavior. In particular, I examine whether smoking or drinking by an older sibling influences the corresponding
behavior of a younger sibling. To control for endogeneity in the older child’s outcome, I use differences in smoking and drinking by gender and age among older siblings as instrumental variables.
While previous studies have established gender differences in teen substance use, I find that these differences vary across age. For example, gender differences in drinking are small at younger ages,
but males drink consistently more than females from mid to late adolescence. The instruments are plausibly exogenous of unobservables such as parental substance use, and will control for parental investments under the assumption that the older sibling’s age and gender do not directly affect resources invested in the younger child. I empirically investigate this assumption using data on measures of parental investments, and find no evidence of a correlation between the instruments and younger sibling investments. The results point toward significant and positive sibling effects for smoking as well as drinking. These findings indicate the presence of opportunities for resource constrained parents to invest efficiently in favor of their firstborn, to reduce “bad” behavior. Positive sibling effects imply that curtailing the older child’s behavior in this way can have spillover effects on the younger sibling, leading to greater payoffs to parents in overall child quality. Within family
social multipliers may also serve to amplify the effects of public policies aimed at curtailing smoking and drinking.
For almost two decades, a vast literature has concerned itself with the association between family structure and child outcomes. These studies have typically found that individuals who grow up
in traditional families (with two biological parents) are better off across several indicators such as educational attainment, health and fertility as compared to their counterparts from other types of family structures like single mother or blended families (with a biological mother but one in which the father may be step for all, or step for some and biological for others). In my second essay, “Family Structure, Parental Investments and Child Well-Being”, I adopt a parental investment perspective and propose that differences in parental investments across varied family structures may provide one explanation for differences in child outcomes by family structure. There are several interesting findings. First, children in single mother and blended families receive lower investments as compared to traditional families, with the gap in investments being larger between single
mother and traditional families. Second, among single mothers, the group driving lower investments appears to be never married single mothers. Third, I find that joint biological children of
both parents in blended families do not differ significantly in the amount of investments they receive relative to their counterparts in traditional families, but non-biological children of fathers are significantly disadvantaged in investment levels. This provides indirect evidence for a biological preference motive. Investigating this further solely within blended families, I find direct evidence in favor of a biological preference motive wherein biological children receive higher investments than their half siblings.
Parental investments in children are usually motivated by models of intra-household allocation which suggest that parents, in making investment decisions, have information about their children’s endowments and respond to them. One factor is that parents motivated by efficiency concerns invest more heavily in better endowed children, presumably due to greater marginal returns from the investment for them. An alternative factor is that rather than reinforcing endowment differences, equity concerns may motivate parents to make compensatory investments in their children by investing relatively more in the less endowed children. There have been several empirical studies that have tried to test these implications, but they have almost entirely assumed that all families will reinforce (compensate) child endowments to the same extent. However, we might expect to
see heterogeneity in the degree of unequal treatment. In my last essay, “Family Structure and Intra-Household Resource Allocation”, I explore one such source of heterogeneity namely, family structure. For example, single mothers, experiencing greater resource constraints might have more incentives to make efficient investments in their children as compared to traditional families with two biological parents present. Studying how investment allocation varies across family structure can be viewed as an important link to understand why individuals from varied types of family structure have such different outcomes on a variety of indicators.