|
Recent years have witnessed a surge in the significance of digital platforms and online retailers. The analysis of digital firms and the industries in which they compete is useful for both furthering general economic knowledge and for informing policy. First, digital firms’ activities incidentally generate digitized data that facilitate the empirical analysis of economic concepts applying to settings traditional and digital alike. Examples of such concepts include network externalities, pass-through of tariffs in vertical relationships, and consumer search. In addition, digital firms differ from traditional firms in ways that matter for the effects of regulation; this makes an understanding of digital firms crucial for intelligent policymaking in contemporary industries. To illustrate, the relative values of the fees that digital platforms charge on different sides of a market depend on industry features distinctive to platform markets, including network externalities and capacity for multihoming. The effects of platform fee regulations depend on these distinctive features. Additionally, the extent of search frictions and of seller differentiation in e-commerce may differ from those in brick-and-mortar retail. These differences have implications for the efficacy of interventions intended to remedy retailer market power — e.g., if search frictions are low at baseline in e-commerce, then informational interventions may do little to intensify price competition among online retailers. In recognition of the value of understanding digital firms, I empirically characterize digital platform and e-commerce markets in this dissertation. This analysis includes assessments of (i) the welfare impacts of digital firms and (ii) the effects of policies targeting these firms. To study digital firms, I collect data from varied sources and develop empirical models of firm competition and consumer behaviour.In the first chapter, I evaluate caps on the commissions that food delivery platforms charge to restaurants. Commission caps may entice restaurants to join platforms, thereby benefitting consumers who value variety in platforms’ restaurant listings. A reduction in platform commissions may also lead restaurants to lower their prices, further benefitting consumers. But commission caps may lead platforms to raise their consumer fees, thereby reducing consumer ordering on platforms and consequently platforms’ value to restaurants. The net effects of caps on restaurant and consumer welfare are thus uncertain. To estimate caps’ effects, I assemble data on restaurant orders, platform adoption, and platform fees. An initial analysis of the data suggests that caps raise platforms’ consumer fees, reduce ordering on platforms, and lead restaurants to join platforms. To analyze these effects, I develop a model of platform pricing, restaurant pricing, platform adoption by restaurants, and consumer ordering. Counterfactual simulations using the estimated model imply that commission caps benefit restaurants at the expense of consumers and platforms. I estimate a total welfare reduction of caps equal to 6.2% of participant surplus from platforms.In the second chapter, I empirically evaluate the contributions of search frictions and presearch seller differentiation to limited consumer search and markups in e-commerce. The internet facilitates consumer learning and allows firms sell products without physical stores. These conditions seem capable of inducing high consumer awareness and cut-throat price competition. In practice, though, consumers exhibit limited consideration in online markets and often pay significantly above the minimum available price for a product. High search costs could explain these facts, as could pre-search seller differentiation: consumers with little aversion to search may not visit a store they believe they are unlikely to purchase from based on information known prior to search. I assess these explanations for limited consideration and market power using a model of sequential consumer search and retailer price competition. I estimate this model on data describing browsing and transactions in contact lens e-commerce. I find that pre-search seller differentiation, not search costs, is primarily responsible for limited consideration and market power in contact lens e-commerce.In the third chapter, I characterize the identifiability of demand models with network externalities. Guided by my identification analysis, I empirically evaluate how network externalities shape the effects of consolidation among US dating websites. Network externalities often arise in differentiated products markets, and especially in platform markets. I show that demand models with network externalities are generally not identified with market-level data alone. This result reflects the impossibility of independently varying product characteristics and market shares at the market level. However, an extension of results in Berry and Haile (2023) establishes that demand models with network externalities are identified under reasonable conditions with microdata linking consumers’ decisions and characteristics. I estimate demand for dating websites using online browsing microdata. The estimates imply that a site’s user values a 10% increase in site usership at USD6.34/month. I find that welfare losses from price rises outweigh gains from network externalities associated with monopolization. Additionally, I find that—due to platform differentiation—a firm earns higher profits from joint ownership of two large dating sites when it does not integrate these sites.
|
