Research

The fees that platforms charge to consumers and merchants may be inefficient due to market power, network externalities, and business-stealing externalities. Using a structural model of platform competition estimated on data covering all major US food delivery platforms, I quantify distortions in platform fees. Consumer fees are nearly optimal due to offsetting market power and offline business stealing distortions. Restaurant commissions, by contrast, are nearly twice their socially optimal levels, primarily because platforms do not fully account for consumer benefits from increased restaurant variety on platforms. I also consider whether platform competition corrects inefficiency in platform fees.

We estimate the effects of offline stores on online spending using data on US store locations and on internet browsing and transactions. The direct effect of a multi-channel retailer’s offline stores on its online sales depends on two opposing forces: (i) a negative cannibalization effect and (ii) a positive cross-channel complementarity, which may reflect offline customer service for items purchased online or brand awareness effects of offline stores. The direct effects of an offline store on rival retailers’ online sales similarly depend on (i) a negative business-stealing effect and (ii) a positive showrooming effect, which arises when offline stores bolster consumer interest in a product category. Our empirical approach captures consumer heterogeneity through variables characterizing the categories of websites that consumers visit, which proxy for unobserved tastes, as well as the demographic profiles of consumers' neighbourhoods. Our estimates for 2007–2008 imply that a multichannel retailer's online sales fall by 1.1–3.8% on average when a rival adds an offline store. These sales increase by 7.1–32.3% when the retailer adds an offline store of its own, suggesting that cross-channel complementarities exceed cannibalization effects. Additionally, our estimated effects of offline stores on rivals' online sales vary across retailers and categories. Notably, offline stores often boost Amazon's sales, suggesting the relevance of showrooming effects.

This paper characterizes the identifiability of demand models with network externalities. Such models are generally not identifiable with market-level data, although micro data linking consumers' decisions and characteristics permit identification under plausible conditions. Identification relies on instrumental variables reflecting across-market variation in the distribution of consumer characteristics or in the characteristics of products on offer. Guided by the identification analysis, I empirically evaluate the role of network externalities in the US dating websites industry. The results suggest  mergers that make  websites interoperable may benefit consumers, in part because reduced price competition may lower prices by leading the merged firm to internalize cross-site network spillovers.

This article develops techniques for the empirical analysis of repeated sequential search over unordered alternatives using data on consumer search processes. I use these techniques to assess why consumers conduct little search in e-commerce and often pay significantly above the minimum available price for a product. Search costs could explain these facts, as could pre-search seller differentiation: consumers with low search costs may not visit stores they dislike based on information known before search. I find that seller differentiation is primarily responsible for limited consideration and market power.