“The Company We Keep:” Endogenous Network Formation and Peer Effects in Churn
(A2020-59621)

AUTHORS

Yulia Nevskaya, Washington University in St. Louis; Yijun (Sally) Chen, Washington University in St. Louis

KEYWORDS

peer effects; customer churn; endogeneity

ABSTRACT

The paper addresses the endogeneity in social ties in empirical measurement of causal peer effects in customer churn from observational data. Peer effects in churn are of paramount importance for products consumed in a socially connected manner. Equipped with the knowledge of causality, managers can craft effective retention campaigns. To tackle the endogeneity, we directly model network formation. After accounting for the choice of peers, we model the interdependence of agents' churn decisions and estimate the causal peer effect in churn. Modeling network formation first allows to recover the latent individual-specific parameters that might affect both the selection of peers and churn. These parameters correct for tie endogeneity in the peer effects model. We use data from a popular online game and find a strong causal peer effect in churn decisions. Our post-estimation analysis informs a firm’s policy of inducing formation of peer groups less prone to peer effects in churn.