Abstract: We study competition between Airbnb and hotel accommodations in Paris in 2017 to assess the welfare implications of Airbnb’s presence on hotels and travelers. The existing literature on the subject exclusively uses across city variation in Airbnb diffusion. Consequently, it does not take into account that the location of an accommodation within a city might be an important dimension of product differentiation. We combine granular data from three different sources to compile a dataset, which allows to model heterogeneity in consumer preferences across geographical locations. Preliminary results based on a nested logit model, which segments consumer preferences along the type (hotel/ Airbnb) and geographical location (districts of Paris), suggest geographical differentiation is an important dimension of product differentiation. We also benchmark the nested logit results against the Inverse Product Differentiation Logit (IPDL) Model, which imposes less restrictive assumptions regarding the nesting structure. Identification is based on plausibly exogenous variation in the number of offered Airbnb listings within a district. Additionally, we propose an instrument exploiting geographical heterogeneity in demand shocks.
Joint with Kevin Tran.