Junior Recruitment Seminar presented by Andrei Zeleneev, Econometrics

Abstract:

Homophily based on observables is widespread in networks. Therefore, homophily based onunobservables (fixed effects) is also likely to be an important determinant of the interactionoutcomes. Failing to properly account for latent homophily (and other complex forms ofunobserved heterogeneity, in general) can result in inconsistent estimators and misleadingpolicy implications. To address this concern, I consider a network model with nonparametricunobserved heterogeneity, leaving the role of the fixed effects and the nature of their interactionunspecified. I argue that the outcomes of the interactions can be used to identify agents withthe same values of the fixed effects. The variation in the observed characteristics of suchagents allows me to identify the effects of the covariates, while controlling for the impact ofthe fixed effects. Building on these ideas, I construct several estimators of the parameters ofinterest and characterize their large sample properties. The suggested approach is not specificto the network context and applies to general two-way models with nonparametric unobservedheterogeneity, including large panels. A Monte-Carlo experiment illustrates the usefulness ofthe suggested approaches and supports the large sample theory findings.

Paper available here.

Paper available here.