Value Creation and Competition in the Grocery Delivery Market

Abstract

This dissertation contains three chapters, each of which is pertinent to the topic of how value is created to consumers and platform competition in the same-day grocery delivery market. All chapters make use of tools from empirical Industrial Organization. All data describing choices made by consumers used for both empirical evidence and demand estimation presented in chapters 1 and 2, respectively, pertain to the Nielsen Company (US), LLC and marketing databases provided by the Kilts Center for Marketing Data Center at The University of Chicago Booth School of Business. In the first chapter, I use the roll-out of two major same-day delivery services in several metro areas in the United Stated to study the impact of these new alternatives on consumers' retailer choice. To do so, I construct a new dataset with the timing of entry decisions of two grocery delivery platforms combine this geographic entry information with scanner data on consumer purchases to evaluate how store choices change once these new services are introduced. To measure the importance of user switching costs, in the second chapter, I estimate a demand model where consumersincur costs to update their delivery platforms choices over time. I extend Katz (2007)'s store choice model to a dynamic setting where, in addition to choosing bundles of products and retailers, consumers also pay a sunk cost to subscribe to memberships that augment their choice set of online retail alternatives. In addition to the revealed preference relations used in Katz (2007) which identify utility parameters, I estimate costs associated with subscriptions (fees and switching costs) using a second set of moments. I construct these moments using revealed preference conditions which compare the utility of maintaining the consumer's subscription choice to the utility of switching. To estimate switching costs, I use constraints that impose rational switching behavior identifying bounds on differential continuation values between subscriptions. I present evidence that switching costs are substantial: fewer than 50% of customers switch to a competitor in the face of savings of up to $40 per purchase. Using the model, I find that switching costs significantly affect consumer platform use. In the absence of switching costs, consumers would alternate between platforms from one purchase to the next ten times more often. By itself, this suggests a potential harm from the major firm's acquisition as lock-in would allow the combined firm to exercise market power in the future. In the third chapter, I model firm decisions as a dynamic entry game in which consumers' transition across platforms, predicted by the estimated demand model, governs the law of motion of firm revenues. Firms then compete in continuous time across independent markets in a similar fashion to Arcidiacono et al. (2016). I use this empirical framework to conduct a retrospective analysis of this recent acquisition. I show that an important aspect of the welfare impact of Big Tech's acquisition of the national grocery chain was Grocer Partner's strategic entry response. Big Tech's main rival could have responded to the merger by either conceding or entering markets more rapidly. When met with the competitive threat presented by the merger, Grocer Partner's own intent to build a loyal customer base increases this firm's incentive to chase a first mover advantage by entering new geographical markets earlier. Moreover, because Grocer Partner's entry costs are low, this firm is able to pursue this accelerated entry strategy giving rise to fierce competition for new markets. I find that the acquisition significantly increased both firms' speed of entry cross new markets, giving consumers earlier access to the services and generating important welfare gains in the short run. Specifically, had the acquisition not happened, both firms would have entered new markets over two years later, on average. The combined costs associated with the two firms' earlier entry due to the acquisition amount to a loss of $624 M in producer surplus. However, consumer benefits across markets that were served earlier due to this merger are larger, representing a total welfare gain of $846 M. Additionally, the fact that this merger allowed the large online retailer to enter multiple markets earlier provides an explanation for the premium paid for the acquisition. Moreover, until this merger occurred, this retail chain was Grocer Partner's largest affiliated retailer, giving it access to approximately 23 million consumers. This supports the fact that Grocer Partner anticipated how the acquisition would affect its ability to serve certain markets and reacted through earlier entry. I perform a second counterfactual that simulates a potential horizontal merger between Big Tech and Grocer Partner resulting in a monopoly. I find that, due to the lack of significant competitive threat, the monopolist would not have an incentive to serve markets early. This shows the role of competition in the timing of entry of these services. I also show that consumer losses due to delayed entry by the monopolist are larger than cost savings from this merger. In both analysis, the focus is on entry timing and firms do not choose prices in the model. For this reason, this paper is limited in its ability to capture possible future harm to consumers through prices. However, I use the demand model to show how consumers' substitution patterns as response to price changes under switching costs shed light onto issue. I find evidence that competition is important to keep prices low, especially if the firm' business model relies on economies of scale. This paper contributes to the literature on the role of consumer inertia in competition by measuring the importance of switching costs for entry strategies in a nascent market and highlighting the implications of this mechanism for consumer welfare. There is a large body of literature relating switching costs to price competition. There is also a theoretical literature relating switching costs to other dimensions of firm strategic behavior, including entry decisions: Klemperer (1988), Farrell and Shapiro (1988), Klemperer (1995), Farrell and Klemperer (2007), Klemperer (1987) and Schmidt (2010). Furthermore, switching costs are deemed theoretically important for preserving advantages to early movers: Lieberman and Montgomery (1988), Shapiro and Varian (2000), Amit and Zott (2001). However, the implications of switching costs for entry decisions have been studied less extensively empirically and measurement of first-mover advantages is sparse (Gómez and Maícas (2011)). This paper also relates to the literature measuring the importance of entry timing to firm decisions. In my setting, the source of early entry incentives is explicitly present in the demand model. I model the mechanism driving consumers' inertia and its relationship with firms' strategic behavior. There is a vast theoretical work on this topic since the early technology diffusion literature (Reinganum (1981a)), (Reinganum (1981b)) and (Fudenberg and Tirole (1985)). The empirical literature on this issue is much sparser due to the difficulty to single-out the motive driving timing from other sources of strategic behavior.