Welcome to CIREQ’s new website !

On the Estimation of Social Effects with Observational Network Data and Random Assignment (with T.J. Chan, J. Estrada, K.P. Huynh, C.T. Lam and L. Sánchez-Aragón)

CIREQ-McGill Seminar 2021-2022
joint with the Department of Economics, McGill University

Organizer Saraswata Chaudhuri (McGill University)

* Virtual Seminar. Please contact the organizer for Zoom login information.


Résumé: This paper proposes a new method to identify and estimate the parameters of a linear model of peer effects in situations where an initial randomization of peers induces the observed network of interest. We argue that the initially randomized peers do not generate social effects. However,  after the randomization, agents can endogenously form relevant connections that can create peer influences.  We introduce a moment condition that aggregates local heterogeneous identifying information for all the individuals in the population. We show that it is possible to identify parameters of interest by using the exogenous variation in the randomized groups. We characterize the root-n rate of convergence for peer and contextual effects. Assuming ψ-dependence in the network space, the asymptotic variance-covariance matrix is shown to depend on different proportions relating to the number of nodes for which it is possible to find distance-p paths in the randomized network.

Upcoming Events