Identification of Peer Effects using Panel Data

TL;DR

This paper introduces new methods for identifying peer effects in panel data models with complex network structures, leveraging exogenous mobility to handle unobserved heterogeneity.

Contribution

It provides novel identification results for peer effects in panel data with unobserved heterogeneity and general network structures, using exogenous mobility assumptions.

Findings

Positive peer effects on surgeon behavior in hospitals

Method applicable to various network structures

Handles correlated unobserved heterogeneity

Abstract

We provide new identification results for panel data models with peer effects operating through unobserved individual heterogeneity. The results apply for general network structures governing peer interactions and allow for correlated effects. Identification hinges on a conditional mean restriction requiring exogenous mobility of individuals between groups over time. We apply our method to surgeon-hospital-year data to study take-up of keyhole surgery for cancer, finding a positive effect of the average individual heterogeneity of other surgeons practicing in the same hospital