Identification and Estimation of Demand Models with Endogenous Product Entry and Exit

Working Paper: CEPR ID: DP18398

Authors: Victor Aguirregabiria; Alessandro Iaria; Senay Sokullu

Abstract: This paper deals with the endogeneity of firms’ entry and exit decisions in demand estimation. Product entry decisions lack a single crossing property in terms of demand unobservables, which causes the inconsistency of conventional methods dealing with selection. We present a novel and straightforward two-step approach to estimate demand while addressing endogenous product entry. In the first step, our method estimates a finite mixture model of product entry accommodating latent market types. In the second step, it estimates demand controlling for the propensity scores of all latent market types. We apply this approach to data from the airline industry.

Keywords: Demand for differentiated product; Endogenous product availability; Selection bias; Market entry and exit; Multiple equilibria; Identification; Estimation; Demand for airlines

JEL Codes: C14; C34; C35; C57; D22; L13; L93


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
endogenous product entry (D40)demand estimation (C51)
failure to account for endogenous product entry (D43)biases in demand parameter estimation (C51)
neglecting latent market types (J46)attenuation biases in demand own-price elasticities (D12)
correlation between market characteristics and entry decisions (L10)demand estimation (C51)

Back to index