Periodic Reporting for period 3 - DIPVAR (Digital Platforms: Pricing, Variety and Quality Provision)
Periodo di rendicontazione: 2023-01-01 al 2023-08-31
With the rapid development of internet economies, digital platforms have become crucial for the world economy. The first goal of this project is to develop tractable and applicable models of markets with digital platforms capturing their unique features. These models should account for the presence of both participation and transaction externalities between different user groups (like buyers and sellers). Positive participation externality from buyers to sellers means more buyers participating makes the platform more attractive for sellers. Similar definition applies for participation externality from sellers to buyers, and for transaction externality from buyers to sellers. The models capture the participation demand on both sides responding to prices while allowing distinct participation and transaction decisions by consumers, and platforms making revenue by charging for transactions between the two sides. They consider quality choices of the platform by allowing one side (sellers) to differ in the value they offer to the other side (buyers), considering pivotal sellers of superior/marquee products, allowing for platform competition and dynamic interactions. The second goal is to investigate the implications of platforms’ business practices (contract types and conditions they sign with their sellers, investment and pricing decisions, etc.) on variety and quality provision to buyers, small rivals, and potential platform entry. The first goal is the key to achieving the second since once we develop tractable and applicable models of digital markets, we can address the key questions of this project as well as a long list of things we ought to be including, which might be a lot to put into this summary. The results of this project will hopefully outline conditions under which platforms’ privately optimal practices deviate from the socially optimal actions, the direction of these distortions, and potential policy tools to mitigate them. This project will thereby provide a tractable framework to address antitrust concerns related to digital platforms’ practices. More generally, it will outline when and which digital platforms’ privately optimal practices could become anticompetitive and harm consumers or when they could be beneficial to consumers by leading to more efficient provision of quality and/or variety. The findings have the potential to generate novel insights into the business models of digital platforms and to highlight potential distortions in these multi-billion markets. The project’s results will provide practical policy recommendations for competition policy and thereby improve consumer welfare in the EU as well as in other countries using these rules.
We provide a canonical and tractable model of an online trade platform controlling the number of differentiated products it hosts and product prices by charging a percentage fee on third-party sellers and choosing whether to sell its own products (hybrid mode), like Amazon marketplace. This model captures interactions between a large number of small sellers and a sizeable range of differentiated platform products competing against them. It also captures two-sided network effects (more buyers attract more sellers and vice versa). Using the model, we illustrate that the hybrid platform steers consumers towards its own products by charging higher seller fees than pure marketplace (insidious steering), so hybrid mode raises final prices and lowers variety on the platform. Banning hybrid mode benefits consumers if and only if the ban leads to pure marketplace.
We also derive general conditions under which the hybrid mode benefits or harms consumers. This benefits consumers when products on the platform are undifferentiated, for example, at the Amazon Standard Identification Number (ASIN) level, by lowering prices in a suite of models: a gatekeeper platform facing a competitive fringe of sellers, when fringe sellers also have their own channels substitutable to the platform; when the gatekeeper platform with fringe sellers competes against a big seller with market power on a differentiated alternative channel; and when the gatekeeper platform hosts only a big seller with market power. This might harm consumers when a big firm sells both on the platform and its alternative channel or when products on the platform are differentiated.
We develop a model of competition for a superior input between two firms asymmetric in their demand potential. The superior input reduces the cost of quality and can either be possessed exclusively by one firm or non-exclusively by both. We show that exclusive dealing is a unique equilibrium when the input changes the competitiveness of either firm (affecting its equilibrium quality). Otherwise, there also exists a non-exclusive equilibrium. We illustrate that exclusive dealing harms consumers but might increase welfare and provide implications for digital platforms.
We study news quality provision on internet by modelling the editorial standard, a cutoff for how confident a news firm must be to post an article. We illustrate conditions under which the firm posts news later (with more information) or earlier (with less information) on the internet compared to offline publishing. We also build a theory model studying the effect of the frequency of communication on the information provided between two strategic players: a sender of information and a receiver.
We develop a novel methodology to study data collection by e-commerce platforms based on the marginal value of information. This enables us to study how platforms collect data and related distortions compared to efficient data collection.
We also derive general conditions under which the hybrid mode benefits or harms consumers. This benefits consumers when products on the platform are undifferentiated, for example, at the Amazon Standard Identification Number (ASIN) level, by lowering prices in a suite of models: a gatekeeper platform facing a competitive fringe of sellers, when fringe sellers also have their own channels substitutable to the platform; when the gatekeeper platform with fringe sellers competes against a big seller with market power on a differentiated alternative channel; and when the gatekeeper platform hosts only a big seller with market power. This might harm consumers when a big firm sells both on the platform and its alternative channel or when products on the platform are differentiated.
We develop a model of competition for a superior input between two firms asymmetric in their demand potential. The superior input reduces the cost of quality and can either be possessed exclusively by one firm or non-exclusively by both. We show that exclusive dealing is a unique equilibrium when the input changes the competitiveness of either firm (affecting its equilibrium quality). Otherwise, there also exists a non-exclusive equilibrium. We illustrate that exclusive dealing harms consumers but might increase welfare and provide implications for digital platforms.
We study news quality provision on internet by modelling the editorial standard, a cutoff for how confident a news firm must be to post an article. We illustrate conditions under which the firm posts news later (with more information) or earlier (with less information) on the internet compared to offline publishing. We also build a theory model studying the effect of the frequency of communication on the information provided between two strategic players: a sender of information and a receiver.
We develop a novel methodology to study data collection by e-commerce platforms based on the marginal value of information. This enables us to study how platforms collect data and related distortions compared to efficient data collection.
The project developed the first tractable yet rich online trade platform model to study their business practices’ implications. It documented a new theory of harm arising from hybrid business model. This has significant policy implications for current policy proposals in the EU and the US of dominant hybrid platforms, like Amazon, Google Play, and Apple App Store. It provides the first model that can be used to empirically estimate the economic effects of different policy interventions that are under consideration. So it will serve to improve consumer welfare in digital markets significantly.
Using the theory model, we expect to highlight fundamental differences and similarities between private labels of retailers and hybrid business models of platforms. We aim to study different demand specifications and different forms of platform contracts with sellers to do a robustness analysis. Next, we aim to link variety provision of the monopoly platform to variety provision of multi-product monopolists. We expect to provide the properties for distributions of consumer tastes for variety that would generate different types of variety distortions arising from the firm’s choices.
We plan to extend the model to platform competition and different qualities of third-party products, and analyze the incentives of competing platforms in choosing which products to allow on the platform (curation).
We extend the scope of markets the project contributes to by considering quality provision in media markets by modelling the editorial standard choice of a publisher (how confident the firm should be to post an article) in a dynamic environment where the firm faces uncertainty about the likelihood that the story is correct. We also study the impact of targeting algorithms of a social media platform on news quality provided by a publisher. To do that we model a publisher that chooses the quality of a fact and opinion article, how to prioritize these articles on its website, and analyze how the social media platform targeting algorithms affect these decisions. This analysis will provide implications for regulation of social media platforms to improve content quality.
We next consider environments where a monopoly platform can collect information on buyers’ demand and so has more precise information on demand than third-party sellers to study strategic information acquisition and disclosure of a monopoly platform. We expect to illustrate the welfare impact of the platform’s demand information disclosure to sellers. We also expect to address the question of how much costly information a firm would acquire about its customers to determine the price to show to a given customer.
The model of asymmetric platform competition for the exclusive product is the first to study endogenous quality choices of platforms when exclusivity provisions affect equilibrium quality provision of asymmetric platforms. Building on this, we plan to analyze the direction of innovation by a firm as a response to dynamic platform competition for exclusivity of this innovation. We expect the results of this work to highlight potential distortions in direction and amount of innovation due to the acquisition of the innovator or exclusive provisions by dominant digital platforms.
Using the theory model, we expect to highlight fundamental differences and similarities between private labels of retailers and hybrid business models of platforms. We aim to study different demand specifications and different forms of platform contracts with sellers to do a robustness analysis. Next, we aim to link variety provision of the monopoly platform to variety provision of multi-product monopolists. We expect to provide the properties for distributions of consumer tastes for variety that would generate different types of variety distortions arising from the firm’s choices.
We plan to extend the model to platform competition and different qualities of third-party products, and analyze the incentives of competing platforms in choosing which products to allow on the platform (curation).
We extend the scope of markets the project contributes to by considering quality provision in media markets by modelling the editorial standard choice of a publisher (how confident the firm should be to post an article) in a dynamic environment where the firm faces uncertainty about the likelihood that the story is correct. We also study the impact of targeting algorithms of a social media platform on news quality provided by a publisher. To do that we model a publisher that chooses the quality of a fact and opinion article, how to prioritize these articles on its website, and analyze how the social media platform targeting algorithms affect these decisions. This analysis will provide implications for regulation of social media platforms to improve content quality.
We next consider environments where a monopoly platform can collect information on buyers’ demand and so has more precise information on demand than third-party sellers to study strategic information acquisition and disclosure of a monopoly platform. We expect to illustrate the welfare impact of the platform’s demand information disclosure to sellers. We also expect to address the question of how much costly information a firm would acquire about its customers to determine the price to show to a given customer.
The model of asymmetric platform competition for the exclusive product is the first to study endogenous quality choices of platforms when exclusivity provisions affect equilibrium quality provision of asymmetric platforms. Building on this, we plan to analyze the direction of innovation by a firm as a response to dynamic platform competition for exclusivity of this innovation. We expect the results of this work to highlight potential distortions in direction and amount of innovation due to the acquisition of the innovator or exclusive provisions by dominant digital platforms.