Abstract
It is now widely recognised that social interactions play a crucial role in economic decision-making and outcomes. Taking such interactions into account is of first-order importance to identify the true mechanisms underlying microeconomic behaviour and to assess individual and social welfare. Almost all empirical work that studies these interactions, however, proceeds by invoking strong parametric assumptions in the econometric model. This does not only obfuscate the empirical content of the model; it also leads to biased estimates, and therefore to biased conclusions. In response to these issues, this project makes a double contribution. First, I will develop a flexible model that significantly generalises the widely-applied linear-in-means model of social interactions, in which the outcome of an individual arbitrarily depends on the characteristics and outcomes of her connections. I will subsequently show that this model can be identified and estimated with nonparametric IV techniques using the average characteristics of connections-of-connections as an instrument, extending previous results from a linear to a nonlinear context. Second, I will investigate the channels through which social interactions can arise in structural models of demand. Drawing from the literature on differential demand and revealed preferences, I will then characterise what are the testable implications of these models. An experiment will empirically validate various specifications of the model.
Researcher(s)
Research team(s)
Project type(s)