Identifying Marginal Returns to Education Through Social Networks

Abstract

This thesis explores the role of peers in influencing the decision of individuals to attend college and the resulting labour market outcomes. It proposes a model, combining the econometric methods of networks and treatment effects, to estimate the marginal treatment effect of education when peers have influence on the wage outcome and probability of treatment for an individual. Using Monte Carlo simulations, the effect of networks on the treatment effects model is investigated. We further explore the model by varying the network structure and conducting sensitivity analyses, considering the impact of networks on policy. Our results suggest that networks initially have a significant positive impact on the returns to education and the effects of policy. However, this effect is reduced once homophily in characteristics is introduced.