Bayesian Estimation of a Dynamic Model of Two-Sided Markets: Application to the U.S. Video Game Industry

This paper develops and estimates a structural model of two-sided markets with durable platform intermediaries and affiliated products. It models buyers’ purchase decisions of platforms and affiliated products and sellers’ decisions of price setting and entry, accounting for the dynamic interaction between the two distinct groups of platform participants. To estimate the proposed model, this paper develops a Bayesian Markov chain Monte Carlo estimation approach that incorporates nonparametric approximation and interpolation methods. The proposed model and estimation method are applied to the 32/64-bit generation of the U.S. video game industry. The results of counterfactual experiments show that the dynamic behavior of platform participants has significant impacts on platform adoption and the affiliated product market, and that a failed platform could have survived if it had priced the two sides properly in a dynamic two-sided market environment.

This paper was accepted by Matthew Shum, marketing.