Abstract

Adolescence is a critical development period characterized by biological, cognitive and social-emotional changes that profoundly affect the development of successful lives. In this period, youth is also susceptible to surrounding environments. This thesis explores: (i) the effects of cognitive and noncognitive (or social-emotional) skills in adolescence on education and earnings in adulthood, (ii) the development of the various dimensions of human capital during adolescence and (iii) the consequences that victimization among adolescents has on health and psychosocial outcomes. The thesis uses high-quality longitudinal data from the Young Lives survey in Vietnam that follows children into adulthood. Young Lives provides a rich data set on diversified aspects of children, their families and communities. The diversity of children in terms of various attributes and experiences allows for analyzing causal relations and the dynamics of child development over time in a low-resource setting.

Chapter 1 examines the effects of cognitive and noncognitive skills on college completion decisions and subsequent earnings. I explicitly embed a model of endogenous education decisions and subsequent earnings into a latent factor model. This approach allows for the identification of latent competencies to capture multiple skill dimensions more accurately and correct for measurement errors in observed measures of skills. It also allows for the isolation of the effects of these skills on earnings into components explained by schooling and productivity. Furthermore, this approach solves the endogeneity and reverse causality problems of skills, schooling and earnings by excluding education variables from earnings equations, introducing latent skills and using panel data with skills and outcomes observed at different times. The findings indicate that both cognitive and noncognitive skills in adolescence are associated with college completion and better earnings in early adulthood. Both types of skills are important in directly determining earnings and indirectly determining earnings through their influence on schooling.

In Chapter 2, I analyze the process by which current levels of cognitive skills, noncognitive skills and health depend on past cognitive and noncognitive abilities, past health, parental cognitive and noncognitive abilities and parental investments. I estimate a dynamic production function model with endogenous parental investments to examine dynamic complementarities and interactions among different inputs and factors in forming child human capital. I use a maximum likelihood approach to estimate the joint distribution of latent factors, which are proxied by observable measures and dynamic CES production functions of human capital. My results show strong effects of parental investments on child cognitive skills, noncognitive skills and health and indicate that parental investments are driven by parental skills and resources. I find evidence that there are dynamic complementarities among the inputs in human capital production, implying that returns to investments are higher for children with better initial conditions. I also find evidence of high levels of self-productivity and the existence of cross-productivity from noncognitive skills and health to cognitive skills and from cognitive and noncognitive skills to health.

Chapter 3 uses a structural model combined with an instrumental variable strategy to deal with the endogeneity and measurement error issues of bullying to study the consequences of peer victimization on a range of health and psychosocial indicators. The findings indicate that peer victimization has strong effects on subjective well-being, alcohol consumption and emotional and mental distress of children. These results are consistent with evidence from both developed and developing countries that bullying has substantial consequences on health risks and psychosocial outcomes. I do not find evidence of associations between bullying victimization and self-rated health.