Racial inequality in educational and occupational attainment has been shown to be related to racial inequality in test scores and cognitive skills. Most research and policy attention has been given to the ability of schools to equalize test scores. I argue that a major reason why researchers have been unable to explain why schools have not closed the gap is because by the time children begin school it may be too late. Cognitive skills develop from infancy and as such, it should be unsurprising that by the time children are five years old the differences across groups are firmly established. Thus, this research attempts to uncover where the racial test-score gap begins by examining infants. I perform a series of analyses using ordinary least squares regression (OLS) and structural equation modeling (SEM) using the first wave of the Early Childhood Longitudinal Survey--Birth cohort (ECLS-B). I utilize the mother's race, rather than the child's race, in the analyses because looking at the mother's race makes the most logical sense since the mother's race is more likely than the child's to determine household income, marital status, mother's education, parenting styles, and so on. I demonstrate that there is little to no raw gap in cognitive skills between the infants of White and Black mothers in the United States. However, through SEM I find that when one controls for social, human, and financial capital, and for differences in health and type of childcare, the infants of African American mothers would actually do better than the infants of White mothers because of their precocious motor development. I find no support for genetics and childcare and only limited support for financial and human capital as mediators of the gap. However, there is support for family social capital and low birth weight as key mediators of the small Black-White test score gap in infancy.

Pesquisas recentes vêm encontrando que alunos negros têm pior desempenho escolar que alunos brancos em testes cognitivos padronizados. A segregação racial é freqüentemente apontada na literatura internacional como uma das principais responsáveis por essa diferença. Nessa dissertação, foi analisado o efeito da segregação racial escolar no diferencial de proficiência escolar entre alunos brancos e negros do ensino fundamental brasileiro. Nos modelos estimados, mesmo após a utilização de diversos controles, foi encontrada evidência de que onde há maior segregação, os negros tem pior desempenho relativamente aos brancos. / Recent research have found that black students have worse schoolar performance than white students in standardized cognitive tests. Racial segregation is frequently pointed out as one of the main factors behind this scenario in international literature. We have studied the effects of racial seggregation on the black-white grade gap in Brazil. In the estimated models, even after the inclusion of several control variables, we have found that where the racial seggregation is higher, the grade differential is higher against black students.

Pesquisas recentes vêm encontrando que alunos negros têm pior desempenho escolar que alunos brancos em testes cognitivos padronizados. A segregação racial é freqüentemente apontada na literatura internacional como uma das principais responsáveis por essa diferença. Nessa dissertação, foi analisado o efeito da segregação racial escolar no diferencial de proficiência escolar entre alunos brancos e negros do ensino fundamental brasileiro. Nos modelos estimados, mesmo após a utilização de diversos controles, foi encontrada evidência de que onde há maior segregação, os negros tem pior desempenho relativamente aos brancos. / Recent research have found that black students have worse schoolar performance than white students in standardized cognitive tests. Racial segregation is frequently pointed out as one of the main factors behind this scenario in international literature. We have studied the effects of racial seggregation on the black-white grade gap in Brazil. In the estimated models, even after the inclusion of several control variables, we have found that where the racial seggregation is higher, the grade differential is higher against black students.

This dissertation consists of three essays studying the effects of collective household decisions on early childhood development from both empirical and theoretical perspectives. The first chapter outlines the dissertation, by presenting the motivations, methods, conclusions, and policy implications for the entire dissertation. Chapter two examines early childhood development using a collective model with children's cognitive production. We jointly estimate the home input demand with children's cognitive production functions based on a simultaneous equations model. Biases are considered that are caused by the non-random selection of time inputs and possible correlations across inputs and outcomes functions. A direct measure of time inputs relying on children's time diaries from the Child Development Supplement of the Panel Study of Income Dynamics (PSID-CDS) has been constructed. We thereby relax the assumption that there is no difference between parental time spent on children and leisure. Our results show that parental time inputs, especially the active time interacting with children's daily activities, have substantial effects on both children's math and reading test scores. The time inputs vary across parents' age, race, and eduction levels. In chapter three, we conduct a standard Blinder-Oaxaca decomposition to evaluate the role of home inputs in the black-white test score gaps based on the empirical model presented in chapter two. Aside from the finding that children's ability accounts for a large proportion of the differences, we find that home inputs can also explain a significant portion of the gap. When the maternal time is equalized at the average levels of white children, the racial differences in children's reading and math test scores can be closed by approximately 30%-50%. The last chapter extends a collective model with household production to the general equilibrium framework. We concentrate on the impacts of a global bargaining power shift within the household on children's cognitive achievement, especially on those who live with single mothers. The model shows that a global bargaining power change in favor of the female may not necessarily be beneficial to the children living with their single mothers. An increase of female's market equilibrium wage rate as a result of reduced labor supply by married women may induce single mothers to work longer hours, spend less time with children, and compensate them with more monetary investment compared with the case when the equilibrium wage rate stays constant. / Ph. D.