Abstract 

Platforms often use fee discrimination within their marketplace (e.g., Amazon, eBay, and Airbnb specify a variety of merchant fees). To better understand the impact of marketplace fee discrimination, we develop a model that allows us to determine platform equilibrium fee, category, and retail entry decisions that depend on the extent of fee discrimination available to the platform. Isolating the effects of fee discrimination (by not allowing the platform to enter into commerce), we find that greater fee discrimination allows the platform to serve more markets in its marketplace but also worsens double marginalization in the high surplus markets. However, if the platform enters into retail, then the platform reduces its fees and generates greater retail competition. These effects mitigate distortions from fee discrimination and improve welfare. In terms of policy, we show that (1) banning fee discrimination and platform entry is detrimental to welfare, (2) a vertical merger within a retail market mitigates double marginalization but is often worse than an equilibrium with platform entry into retail, and (3) taxing the platform in retail (not merchants) levels the retail playing field and can generate a Pareto improvement upon a policy that bans platform retail entry. 

Paper available here.

Abstract 

The online grocery market has seen massive entry over the last five years by firms ranging from traditional brick-and-mortar chains to online platforms. This paper documents the welfare impact of a major online retailer’s acquisition of a national grocery chain. The largest delivery services offer subscriptions, and, in the data, consumers rarely switch between them. I find that switching costs significantly affect consumer platform choice, suggesting potential for future exercise of market power. To study this acquisition’s impact, I model competition between two large platforms. The first is the major online retailer engaging in the merger. The second, the rival, is an independent platform with massive coverage. The platforms compete in a dynamic entry game, and two opposing forces influence entry timing. On the one hand, firms chase a first-mover advantage resulting from consumer lock-in. On the other hand, I find that entry costs fall over time for both firms, leading to significant costs of early entry. I estimate that, before the acquisition, the major online retailer had very large entry costs. The acquisition reduced this cost, posing a competitive threat to the rival. Consumer lock-in then contributed to raising the stakes of early entry for both firms, accelerating entry timing by more than two years across new markets. Further, I find that a merger between the two platforms, resulting in a monopoly, would delay entry significantly. These results show that strategic competition in entry timing plays an important role in mergers’ welfare effect. 

Paper available here.

Abstract 

In a Bayesian Persuasion framework with heterogeneous priors, I identify a novel strategy which is optimal for a monopolistic sender if the society is sufficiently polarized in terms of priors. Persuasion can still arise with competition, if two necessary conditions are met: a) biased senders; b) limited attention by receivers. I examine a sequential game where first each receiver decides which sender to pay attention to, and then senders persuade as monopolist in their own segments. In this setting, Echo Chambers might arise endogenously and such allocation of attention is Pareto dominated by a monopoly. With any positive fraction of single-homing receivers, full revelation is not an equilibrium. Conversely, when the fraction of single-homing is sufficiently large, the optimal strategies under monopoly constitute an equilibrium even in presence of multi-homing receivers. 

Paper available here.

Abstract 

Understanding the implications of digital platform strategies on user welfare has become a crucial issue for policy makers and economists alike. In payments, distributed ledger technologies such as the blockchain have introduced new models of platform governance in the form of vertically disintegrated platforms (VDP). In this paper, we develop a framework to study the implications of vertical disintegration on user welfare. A VDP mediates between users and processors that enable interactions between users. The VDP controls the price structure but the price level is set in a competitive equilibrium between processors; proceeds from operating the platform service are split between processors and the VDP. We find that the welfare ordering between traditional (integrated) and disintegrated platforms crucially depends on the platform cost structure and the regulatory conditions in the market for processors. When the cost of integrating processors depends on the transaction volume, the VDP can produce higher welfare. In contrast, when this cost is fixed, an unregulated VDP is less desirable than the integrated platform. However, regulatory limits on the VDP power over processors can make user welfare under a VDP dominate. Within our framework, we analyze recent applications in the payment industry, including central bank digital currencies, and show how different design choices affect user welfare.

Paper available here. 

Abstract 

This article uses an original dataset on apps targeted at very young children to explore the types and scope of data that is collected about children when they use online mobile applications. US legislation provides a stringent privacy regulation to protect children’s personal data and digital platforms have implemented a self-certification program to encourage developers opt-in to support the regulation. We use an original dataset of mobile apps published in the US that target very young children to explore the types and scope of the data collected while children use mobile apps. We measure the effect of developer’s size in collecting sensitive data and evaluate the effectiveness of the platform policy in protecting children. 

Abstract 

Review systems are ubiquitous in online marketplaces. I construct a game-theoretic model to study the impacts of review systems on market outcomes. I document how the emergence of review systems changes firms' branding decisions and lowers the value of brand names. Review systems improve social welfare by incentivizing firms to invest in quality, but inefficiencies persist partly due to over-investments. 

Abstract 

Many modern innovations depend on interconnectivity, which require technology standards as a common language to successfully link up. This paper develops and estimates a structural model to understand how competition between firms affects their incentives to cooperate by supplying technologies to a common standardization process. I study these incentives empirically by focusing on the standardization of the mobile telecommunications technologies. In the model, firms face two decisions. They decide whether to join a group to develop a component of the system and, in that case, how much effort to exert. When making these choices, firms consider 1) how their effort increases the common value, 2) how much of this common value they can privately appropriate through their patents, and 3) their capacity to profit from the technology in the downstream part of the market. In this setting, patents have an ambiguous effect on the development of a common innovation. On the one hand, they alleviate the free-rider problem and induce firms to exert more effort. On the other hand, they bias firms' participation towards groups with less competition over patented technologies even where their effort may be less valuable. To study the net effect of these forces in equilibrium, I estimate the model using a novel dataset on 3G and 4G technologies. I also show that the enforcement of royalty-free clauses reduces firm participation and effort, ultimately delaying the completion of the initial releases of 4G by almost 1 year.

The paper is available here.

Abstract 

We analyse the implications of allowing exclusive dealing between a marquee product seller and two competing platforms that are asymmetric in their locked-in consumer base and compete for ‘single-homing’ new consumers. On the seller side, the platforms choose the number of fringe products, which generate uncertain value, and compete for the marquee product, which is better than an average fringe product, by offering a menu of exclusive dealing and non-exclusive dealing contracts. We show that an exclusive dealing (ED) equilibrium where the big platform has the marquee always exists. ED equilibrium is unique if the marquee gives competitive advantage to the big platform, that is, when new consumer market is covered and the small platform’s quality would be lower without the marquee. Otherwise, there exists also a nonexclusive equilibrium where both platforms have the marquee. ED harms welfare by lowering variety and the expected quality on the small platform. If the platforms are allowed to price discriminate between their loyal and new consumers, in the ED equilibrium the big platform offers more fringe products and the small platform offers fewer fringe products. Price discrimination lowers fees for new consumers and increases fees for locked-in consumers. 

Abstract 

Roam-Like-At-Home rules (RLAH) passed on the 15h of June 2017 stipulate the end of roaming charges for EU residents that travel to countries within the European Economic Area (EEA). In this paper, we use micro-level data from a telecom operator to estimate the welfare effects of mobile internet for travelers. Decomposing the surplus gains by consumer types based on pre-regulation mobile data usage, we highlight important heterogeneity. Users with higher pre-regulation usage experience higher absolute gains, which are mainly driven by the revaluation of existing mobile data usage, and less by additional usage. Conversely, users with lower pre-regulation usage experience higher relative gains, which is mostly driven by additional usage. This suggests that the regulation was at least partially effective in closing a digital divide between locals and visitors. We further show that consumers use mobile data for both information consumption and information provision, and also look into the indirect welfare effects of RLAH for content providers. 

Abstract 

A firm may gain competitive advantage over its rivals through access to market information. Yet evidence suggests only large firms invest in technology that facilitates information provision, potentially contributing to increase their leverage over smaller competitors. This paper aims to empirically investigate how getting access to Big Data information would affect small and medium firms’ performance and decision-making. To do so, we evaluate the impact of an unprecedented Big Data information service diffused at zero cost by a large European bank among its small and medium-size business customers. Upon adoption, the bank provided monthly reports with rich information about each firm’s clientele portfolio and that of its competitors coming from the analysis of Big Data credit card transactions. Using first-differences we find adoption is associated with a 4.5% increase in establishment revenue, whereas IV estimation results show that adoption increases revenue by 9% for those establishments whose adoption decision is most strongly affected by the instrument. The main mechanism behind this result appears to be the information technology prompting establishments to target existing, yet unexploited, business opportunities. Consistent with this mechanism, we find that adopting establishments increase their sales to underserved customer segments. Not only they increase their number of customers, their new customers also come from underrepresented geographic areas and gender-age groups in their customer portfolio prior to adoption. Our evidence is also consistent with establishments improving their resource allocation efficiency upon technology adoption. These findings suggest that small and medium enterprises obtain substantial returns from information access, and therefore, high adoption costs and lack of awareness are likely to be key barriers preventing these firms from investments in Big Data technology.

Abstract 

This paper analyzes consumers' privacy choice concerning their data and firms' ensuing pricing strategies. Consumers decide whether to reveal private information in form of cookies. We incorporate this endogenous decision into a duopoly model with behavior-based pricing. We find unique pure-strategy equilibria for two disclosure rules. Under revelation to both firms, consumers disclose their information and firms price discriminate. Under revelation to only one firm, consumers hide their information and firms price uniformly. In our laboratory experiment we confirm these pricing strategies. Consumers, however, have a higher than predicted disclosure rate which is driven by learning effects and privacy concern.

The paper is available here.

Abstract 

Platforms greatly facilitate transactions between buyers and sellers. This allows platforms to gather detailed information on transactions and tailor their strategies when deciding to introduce their own products that compete with independent traders. Concerns have been raised that such an information advantage of the platform can hurt traders. However, we show that the usage of more detailed and individualised information by the platform can actually benefit traders by relaxing competition between them. This occurs as traders have more incentives to raise their prices in order to hide the popularity of their products and prevent entry of the platform in their product categories. The competition relaxing effect is particularly strong when traders are close substitutes and face little demand uncertainty within their category. In addition, we show that the interests of platforms and traders are more aligned and both can benefit from more individualised information usage when proportional fees are used, but consumers are hurt in this case. 

Abstract 

We demonstrate how privacy law interacts with competition and trade policy in the context of the European General Data Protection Regulation (GDPR). We follow more than 110,000 websites for 18 months to show that websites reduced their connections to web technology providers after GDPR became effective, especially regarding requests involving personal data. This also holds for websites catering to non-EU audiences and therefore not bound by GDPR. We further document an increase in market concentration in web technology services after the introduction of GDPR. While most firms lose market share, the leading firm, Google, significantly increases market share. 

Paper available here. 

Abstract 

Digital technologies  allow sellers to improve their ability to better target the state-contingent needs of buyers by designing the information structure. One of the critical questions is to understand how to model  such information in a setting where the seller's quality is unknown to the buyer. Our paper investigates under which conditions a high-quality seller can separate in a context in which the seller can either acquire exogenous signals (information acquisition) or design the information structure (information design). The cost of endogenously reducing uncertainty allows for studying which of the two informative setups make the separation more likely. When this cost is low, the separation cannot emerge under information design. When this cost is high, the parameter region in which separation can occur is larger under information design. 

Abstract 

This paper analyzes the impact of intermediary concentration on the allocation of revenues in online platforms. We study sponsored search documenting how advertisers increasingly bid through a handful of specialized intermediaries. This enhances automated bidding and data pooling, but lessens competition whenever the intermediary represents competing advertisers. Using data on nearly 40 million Google keyword-auctions, we first apply machine learning algorithms to cluster keywords into thematic groups serving as relevant markets. Using an instrumental variable strategy, we estimate a decline in the platform’s revenues of approximately 11 percent due to the average rise in concentration associated with intermediary M&A activity.

Abstract 

This article studies the entry and pricing decisions of sellers in a market with a reputation system. We provide a model of dynamic oligopoly with heterogeneity in marginal and opportunity costs and individual reputation as a state variable. We show that new sellers are generally less likely to reenter the platform than incumbents and sellers who have a lower chance of entering in subsequent periods set on average higher prices. The mechanism behind these findings is selection on marginal costs. We apply our model to a dataset on sellers on a large ridesharing marketplace. We showcase a negative correlation of tenure on the platform, measured by the number of reviews, and prices set by drivers. However, after accounting for drivers' unobserved characteristics, which we interpret as marginal costs, we find a positive relationship. We provide, further, evidence of selection on unobservables by studying reentry decisions. Finally, we calibrate our dynamic model to uncover the distribution of opportunity costs. 

Abstract 

The aim of this paper is to provide new evidence on the impact of ultra-fast broadband investment on local growth, and in particular on local new business establishment. We leverage on a unique dataset collecting municipality-level ultra-fast broadband deployment data for the period 2013-2018 matched with granular information on the number of both new and closing businesses for 15 industrial sectors.  We exploit the staggered roll-out of fiber-based ultra-fast broadband networks starting from 2015. We deal with the endogeneity of broadband diffusion by exploiting the distance between each municipality and the closest central office equipped with "last mile" device. Our preliminary results show that the introduction of ultra-fast broadband technology reduces both new firm establishment and firm exit.

Abstract 

We study choice of content substitutability by competing media platforms when consumers can multihome. Two platforms compete for consumers and advertisers. Advertisers want to reach eyeballs. A consumer multihomes if the second platform provides sufficient added value. We show that providing similar content makes multihoming less likely. Competition for consumers will then be fiercer. But if more consumers singlehome, competition for advertisers will be less fierce as platforms have a monopoly on the eyeballs of their singlehoming consumers.

Our model has implications for mergers among media firms. If platforms primarily make profits from consumer payments, they prefer to merge with firms that provide similar content, enhancing their market power on the consumer market. If they primarily profit from providing eyeballs, however, they prefer to merge with firms that offer complements, to build ecosystems that effectively lock in consumers. Doing so enhances their market power on the market for eyeballs.

Abstract

Following the development of autonomous vehicles (AVs) and GPS systems, fleets will gain prominence over private vehicles. We analyze the welfare effects of the transition from a fully decentralized regime, in which all travelers are atomistic and do not internalize the congestion externality, to a centralized regime, where all travelers are supplied by a fleet of AVs controlled by a monopolist. In our model, heterogeneous individuals differing in the disutility from congestion may travel on one of two lanes, which may endogenously differ in the level of congestion, or they may not travel. We show that the monopolist sorts travelers across the two lanes differently than the decentralized regime. Moreover, depending on the severity of congestion costs, it may also exclude some travelers. We find that centralization is always welfare detrimental when the monopolist does not ration travel. If instead rationing occurs, centralization may be welfare beneficial, provided that congestion costs are sufficiently high. We then analyze how to restore first best with road taxes. While congestion charges are optimal under decentralization, taxes differ markedly in a centralized regime, where restoring first best may require subsidizing the monopolist. 

Download the paper here.

Abstract

We provide a unified framework to compare the merits of various strategies that platforms can adopt to facilitate the interaction among users and generate value from these interactions. Using a micro-foundation of transactions among users (typically, buyers and sellers), we analyse different tools that a platform can use to manage participation and transactions. On top of charging membership and/or transactions fees, we allow the platform to control the price of the traded products, or to use data analytics to help sellers practice differential pricing. Our (current) results are the following: a platform setting membership fees (i) attracts fewer sellers and more buyers than a free platform; (ii) does not achieve larger profits if it can control the price of the traded products, (iii) or when it shares information about buyers with sellers  

Abstract

This article studies competition between different types of ad-funded platforms attracting consumers with free services. Consumers often find advertisements a nuisance on such platforms. We study how under a competitive setting platforms balance the tension between attracting consumers and rent extraction from the advertising side. We propose a flexible yet simple model that studies competition between standard platforms and social media platforms (with same-side network effects). We find that an increase in either positive same-side network effects or an increase in consumer disutility from advertisements leads to a reduction in the number of ads on that platform. When competing platforms merge, consumer side network effects do not impact prices and the number of ads is higher. In a setting where consumers present a negative (congestion) externality on each other, competition fails to protect consumer welfare and behaves erratically. Finally, we present a few extensions and discuss some policy implications.   

Abstract

In many sectors, platforms or matchmakers are competing with vertically integrated incumbents. Uber competing with taxi companies and AirBnB competing with hotels are the two preeminent examples of this type of competition. In this paper, we model this process of competition between a platform and vertically integrated incumbents. A striking feature is that competition takes place at two levels. The platform competes for the market to affiliate users to its service. The platform also competes in the market with the incumbents for service provision to the affiliated customers. We construct a dynamic model that integrates these two levels of competition. We distinguish two stages in the competitive process: the mature market where there is only competition on the market and a transition phase where there is both competition in and for the market.

Abstract

Marketplaces are increasingly using algorithms to make some sellers more prominent than others. In this paper, we study the incentives of a platform to conceive prominence as  default option for a share of the consumer population to vertically differentiated sellers. We find that if a platform follows a "consumer-centric" algorithmic rule, prominence is granted to the high-quality seller. When the platform offers quality-upgrade programs in exchange for revenue-sharing, the platform designs a scheme such that only the low-quality seller joins the program. We find that this practice squeezes sellers' profits and can have ambiguous effects on consumer welfare. 

We will start our reading group on AI and Algorithms. In this first session, we will discuss the paper by Emilio Calvano, Giacomo Calzolari, Vincenzo Denicolò, and Sergio Pastorello on "Artificial intelligence, algorithmic pricing and collusion". 

Abstract

Pricing algorithms are increasingly replacing human decision making in real marketplaces. To inform the competition policy debate on possible consequences, we run experiments with pricing algorithms powered by Artificial Intelligence in controlled environments (computer simulations).

In particular, we study the interaction among a number of Q-learning algorithms in the context of a workhorse oligopoly model of price competition with Logit demand and constant marginal costs. We show that the algorithms consistently learn to charge supra-competitive prices, without communicating with each other. The high prices are sustained by classical collusive strategies with a finite punishment phase followed by a gradual return to cooperation. This finding is robust to asymmetries in cost or demand and to changes in the number of players. 

Professor Axel Gautier launched the 2020 LED meetings by giving an overview on the economic challenges and legal issues related to the usage of AI and Algorithms in the digital economy.