Research

Abstract: This paper studies how retailers use assortment to respond to local market structure when prices are set at a national level. Since the pricing channel is unavailable for firms under uniform pricing, retailers can instead vary assortment to adjust to demand and market shocks. Extensive store-level data and a structural equilibrium model of store competition allow to disentangle the market power effect on assortment choice from other market forces such as local preferences and logistics. The counterfactual analysis points out the retailers' incentives to exercise local market power when consumers have limited choice by providing fewer and more expensive positions.

With Anders Munk-Nielsen and Morten Sæthre

Abstract: We study the entry decisions of the three retail pharmacy chains in Norway over the period from 2004 to 2012. Following deregulation of entry, the market grew rapidly, doubling the number of pharmacies.  We document that repeated entry by an already present incumbent chain occurs with non-trivial frequency and set out to investigate whether preemptive motives play a key role. We propose and estimate a highly flexible spatial demand model with overlapping sets of consumers across space. While the estimates imply substantial demand heterogeneity, we reject the hypothesis that the repeated incumbent entries can be explained by market segmentation by store format differentiation. Instead, we propose that private information about local market conditions may play a role. Indeed, we find that an incumbent chain is significantly more likely to respond to local market heterogeneity than competing chains. 

With Sergey Kokovin, Shamil Sharapudinov, Alexandr Tarasov, and Philip Ushchev 

Abstract: Our novel approach to modeling monopolistic competition with heterogeneous firms and consumers involves spatial product differentiation. Space can be interpreted either as a geographical space or as s space of characteristics of a differentiated good. In addition to price setting, each firm also chooses its optimal location in this space. We formulate conditions for positive sorting: more productive firms serve larger market segments and face tougher competition; and for the existence and uniqueness of the equilibrium. To quantify the role of the sorting mechanism, we calibrate the model using cross-sectional data on the haircut market in Bergen, Norway, and perform a counterfactual analysis. We find that inequality in the distribution of the gains among consumers caused by positive market shocks can be substantial: the gains of consumers from more populated locations are 3-4 times higher.