The institution of marriage plays some role in determining one’s risk of exposure to HIV. Since the transmission of HIV in the population is mainly through sexual activity, avoiding infection depends on risk-avoiding behavior. Consistently, empirical results show that excess mortality is concentrated in not-yet married adults aged 20-39 among both men and women. Therefore, the choice of when and who to marry appears to be related to risk of exposure. The objective of this paper is to determine the effect that schooling has on HIV/AIDS excess mortality, using panel data from South Africa. This paper tests the hypothesis that schooling affects when and who one marries and thus impacts the risk of mortality from HIV/AIDS. The effect could be negative or positive. On the one hand, since educated agents have incentives to secure returns to their human capital in the future, more education implies earlier marriage, given that the marriage institution effectively decreases the HIV-related mortality risk. On the other hand, education increases the opportunity costs of marriage especially for women, who need to increase their time spent in the household. Thus, schooling may increase mortality risks due to the increased risk of HIV infection. The empirical analysis of this research uses an identification strategy for mortality shock based on the following observation: before 1980, fertility decisions and mortality incidences in South Africa were not related to the coming AIDS epidemic in the mid 1990s. This condition enables us to argue that cohort-specific factors among those who were ages 20 or above in the mid 1990s (when the first round of our data was collected) are not correlated with the AIDS epidemic. Therefore, after controlling the non- AIDS mortality rate, the cohort-specific mortality changes in the period of 1998-2004 can be mainly attributed to the AIDS epidemic, and are treated as exogenous shocks to the adult population. In the analysis to estimate the cohort-specific schooling effects in the mortality equation, we also control for household-level fixed unobservables and age (cohort)-specific unobservables in order to base our statistical inferences on within-household and within-age group variations in mortality incidence and schooling. Results show that schooling increases excess mortality among women, but not among men. This gender difference is consistent with their marriage behavior. The probability of marriage decreases among educated women if the direct cost of marriage increases (that is, if Lobola—a gift from the groom to the bride’s family—payment is the norm in the community). For men the probability of marriage is lower than for women. In contrast to women, educated men are more likely to get married when the direct cost of marriage increases. In sum, schooling increases the opportunity cost of marriage for women, which delays marriage and increases their mortality risks in high HIV-prevalence societies, but has the opposite effect on men. viii Our analysis demonstrated the need to integrate our understandings of the marriage market, the labor market, schooling investments, and youth behavior to identify the determinants of AIDS-related excess mortality. The marriage institution potentially protects the youth from excess mortality, but the interactions between marriage and labor markets complicate the role of schooling in determining excess mortality. This finding is also policy relevant as we need to pay special attention to the differentiated mortality risks between women and men to effectively reduce gender-specific AIDS-related excess mortality.