This dissertation is composed of three essays. Chapter 1 studies the role of the composition of intermediate inputs in shaping the dynamics of skill premium in developing countries following trade liberalization. The sharp increase in the skill premium in developing countries following trade liberalization is an empirical regularity that is at odds with the predictions of standard Heckscher-Ohlin trade theory. In this paper, I argue that most of the rise in the skill premium is accounted for by complementarity between high skilled labor and foreign intermediate inputs. To quantitatively assess this effect, I build a model featuring two types of labor, high skilled and low skilled, which interact with foreign and domestic inputs at varying degrees of substitution. In this environment, increased access to foreign intermediate inputs raises the relative demand for skilled labor, and consequently, the skill premium. Using firm-level microdata from Ghana, I provide evidence for this mechanism and estimate the degree of skill-complementarity. I find that with differences in substitution elasticities, changes in the observed shares of foreign intermediate inputs account for 75% of the rise in the skill premium observed in the data. Chapter 2 studies the determinants of the negative educational gradient in divorce. The data suggests that couples with higher levels of education face a lower risk of divorce. This observation is puzzling in the context of standard models of household formation, which are based on the trade-off between gains from joint-consumption and match quality. In these models, the gain from joint consumption is lower for college- educated people due to decreasing returns in consumption, but the gain from match quality is higher, leading to a higher divorce rate. To resolve this puzzling observation, I provide evidence that divorce reduces the chances of success for children. Moreover, this reduction is much more pronounced for children with college-educated parents than for those with non-college educated parents. This additional cost for college-educated parents, in the form of a lower likelihood that their offspring complete their college degree, provides a stabilization effect. I build a model of household formation to quantitatively assess to what extent this larger reduction can account for the divorce gap observed in the data. I estimate that these higher costs impute to parents of higher levels of education account for almost two thirds of the divorce gap. This finding suggests that the benefit of marriage conferred on college-educated parents in rearing children is the primary factor in shaping the differences in divorce. Chapter 3 explores the causes behind an increasing gap between college graduates and high school graduates in entrepreneurship entry. Up until the 1980s, both college graduates and high school graduates had the same propensity of becoming entrepreneurs. The fraction of these two groups who were entrepreneurs was 13%. However, after the 1980s, there was a substantial gap between the number of college and high school entrepreneurs. In 2010, the fraction of college graduates who were entrepreneurs increased to 20%, while the same fraction stayed at 14% for high school graduates. I build a model of educational choice and occupational choice to unravel the underlying economic forces in shaping this increasing gap. The findings of this paper suggest that an increasing complementarity between managerial ability and education is the main driver of the gap in entrepreneurship entry.