Structural Estimation of Pairwise Stable Networks : An Application to Social Networks in Rural India

This paper studies what can be learned from pairwise stable networks. I specify the marginal utility of a pair forming a link as a function of individual-level unobserved heterogeneity as well as pair level observed characteristics. I derive a tetrad-level utility inequality from pairwise stability, which substantially simplifies the complex model of strategic network formation. The individual-level unobserved heterogeneity is canceled out within the tetrad-level inequality, and utility parameters are point identified. I propose a maximum score type estimator that is consistent when the number of networks goes to infinity, and its finite sample performance is investigated with Monte Carlo experiments. I apply the estimation method to social networks in rural India and find evidence of homophily in caste, gender, and education