Network Formation with Unobserved Homophily: Identification and Evidence from Bangladesh

Abstract

We consider a network formation model that features degree heterogeneity and homophily in both observed and unobserved node characteristics. To identify the model parameters, we employ a two-step approach, where the first step is to construct tetrad-level inequalities in which degree heterogeneity is cancelled out. In the second step, the observed and unobserved homophily parameters are identified by showing that the conditional density of the network-formation decision given the observed and unobserved homophily is recovered from observed variables. We propose an easy-to-implement parametric estimation method that yields a consistent estimator under many-node asymptotics. We show that the estimator has an asymptotic normal distribution, and demonstrate its finite-sample performance. Using a unique two-period network dataset from Bangladesh, we empirically investigate students’ homophilic preferences. We utilize the differences in the eigenvector centrality between students and between their parents in the previous year, respectively as a proxy and an instrument, to control for homophily in unobserved personality. Our results suggest that observed homophily estimates could be substantially biased if unobserved homophily is not properly taken into account.

About the presenter’s visit

Dr Jun Sung Kim will be visiting the School of Economics on 6 June 2019.  While here he will be using room 520A Colin Clark building. If you would like to meet with Dr Kim or have lunch or dinner with him please contact Dr Fu Ouyang who will be his host while at The University of Queensland.  Dr Ouyang can be contacted on f.ouyang@uq.edu.au.