Binary Outcomes and Linear Interactions

Working Paper: CEPR ID: DP15505

Authors: Vincent Boucher; Yann Bramoull

Abstract: Heckman and MaCurdy (1985) first showed that binary outcomes are compatible with linear econometricmodels of interactions. This key insight was unduly discarded by the literature on the econometrics of games.We consider general models of linear interactions in binary outcomes that nest linear models of peer effects innetworks and linear models of entry games. We characterize when these models are well defined. Errors musthave a specific discrete structure. We then analyze the models’ game-theoretic microfoundations. Undercomplete information and linear utilities, we characterize the preference shocks under which the linear modelof interactions forms a Nash equilibrium of the game. Under incomplete information and independence, weshow that the linear model of interactions forms a Bayes-Nash equilibrium if and only if preference shocksare iid and uniformly distributed. We also obtain conditions for uniqueness. Finally, we propose two simpleconsistent estimators. We revisit the empirical analyses of teenage smoking and peer effects of Lee, Li, andLin (2014) and of entry into airline markets of Ciliberto and Tamer (2009). Our reanalyses showcase themain interests of the linear framework and suggest that the estimations in these two studies suffer fromendogeneity problems.

Keywords: binary outcomes; linear probability model; peer effects; econometrics of games

JEL Codes: No JEL codes provided


Causal Claims Network Graph

Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.


Causal Claims

CauseEffect
binary outcomes (C25)compatibility with linear econometric models of interactions (C51)
preference shocks are iid and uniformly distributed (D11)linear model forms a Bayes-Nash equilibrium (C73)
linear model forms a Bayes-Nash equilibrium (C73)existence of equilibria (C62)
shock dispersion increases (D39)unique Nash equilibrium (C72)
econometric methods used (C51)validity of results obtained (C52)

Back to index