DIPVAR - Digital platforms: pricing, variety, and quality provision

Digital platforms—such as online marketplaces, app stores, and other digital ecosystems—now play a central role in the economy by organizing interactions between buyers and sellers, or users and developers. The project’s first objective was to develop tractable and empirically applicable models that capture key features of these markets: separate participation and usage decisions; cross-group externalities; transaction-based monetization, often through ad valorem commissions; heterogeneity in seller quality and consumer valuations; and platform practices such as hybrid business models, seller fees, curation and certification, exclusivity, and innovation acquisition.The second objective was to use these models to study how platform conduct affects prices, variety, quality, market structure, and welfare, and to derive implications for competition policy and regulation. A central conclusion of the project is that conduct that is privately profitable for a dominant platform—such as self-preferencing through fee setting, excessive commissions, under-enforcement of quality standards, exclusive contracts, and acquisition strategies that shape innovation—can generate systematic distortions in market outcomes. At the same time, interventions aimed at one harmful practice may trigger adjustments in other platform decisions, with unintended effects on quality, variety, and consumer surplus. The project therefore provides tools for evaluating policies currently debated or implemented in the EU and elsewhere, including rules on self-preferencing, payment routing, liability, exclusivity, and startup acquisitions.

Work Package 1 develops a tractable framework for online trade platforms and related digital ecosystems, with a focus on pricing, variety, and quality. A flagship result is Anderson and Bedre Defolie (2024), which studies a dominant platform that charges third-party sellers an ad valorem fee and may also sell its own products (hybrid mode). The paper identifies a new form of self-preferencing, “insidious steering”: a hybrid platform raises seller fees, which increases rivals’ costs and prices, reduces third-party entry and product variety, and can lower consumer welfare. Banning the hybrid mode may therefore benefit consumers, but not always: if the platform would become a pure reseller rather than remain a marketplace, the welfare effect may reverse. Anderson and Bedre Defolie (2022) establish more general conditions under which hybrid platforms lower or raise prices and help or harm consumers. The key policy implications of this work have been summarised in the La Fonte Blog piece, "Hybrid marketplaces: Acting as the referee and a player”. These theoretical findings are complemented by empirical work using Amazon Italy data across five product categories. Bedre Defolie and Sokullu (2026) structurally estimate a model in which Amazon chooses whether to enter as a retailer and sets fulfillment fees, third-party sellers choose whether to enter, demand for fulfillment services, and set prices; and consumers decide whether to buy on Amazon or not. The estimates show that the effects of hybrid model differ substantially across categories, but counterfactual simulations indicate sizeable consumer-surplus gains from banning Amazon’s own retail presence, especially for more recent products. The project also extended the analysis to app stores and other two-sided platforms. Anderson and Bedre Defolie (2025) study a platform that charges sellers an ad valorem commission and buyers a membership fee, in an environment where sellers differ in quality and buyers differ in willingness to pay and transaction intensity, and each match has a different value proportional to the seller quality and the buyer value. The model endogenizes cross-side network effects through equilibrium participation on both sides. It shows that the profit-maximizing platform fees  lead to too little participation by buyers and sellers. The paper identifies a simple statistic—the ratio of seller revenue to consumer surplus—that governs the strength of the distortions and the effects of regulation. For policy, it demonstrates how commission caps or third‑party payment mandates stimulate app entry but also induce the platform to raise buyer fees, so high-value users may benefit while low-value users lose. Consumer surplus effects depend on the type of distribution of buyers and sellers. The project also developed new theory on platform governance of quality and seller entry through curation and certification: Bedre Defolie, Johansen, and Madio (2026) identify conditions under which a revenue‑maximizing platform may not remove harmful sellers to strengthen incentives for effort provision by normal sellers: investing in costly effort increases the likelihood of being certified and so decreases the likelihood of being pooled with low-quality and harmful sellers. The social efficiency might also keep some harmful sellers to provide effort-provision incentives, but the platform over‑tolerates harmful sellers relative to a social planner.The project also examined how information intermediaries shape markets through the design of information policy and market composition, including targeting, disclosure, and certification. Kyriazis (2026) studies how an intermediary’s choice of which consumers are exposed to a monopolist’s offers affects the monopolist’s screening strategy. It shows that a platform that places more weight on consumer surplus than on profits prefers to have some intermediate-valuation consumers as well as the highest-value segment. Related work with Lou shows that an intermediary’s information policy can induce a monopolist to offer a broader product menu, increasing consumer surplus but not necessarily efficiency. These papers identify a new channel through which transparency, targeting, and information design affect market outcomes. Work Package 2 studies exclusivity, acquisitions, innovation, and the dynamics of competition when firms can also compete through quality. Bedre Defolie and Biglaiser (2023), extended in later work, analyze competition for exclusive ownership of an input that lowers the cost of quality provision. Firms first compete for exclusivity and then choose quality and price. The analysis shows that exclusivity arises in equilibrium and disadvantages the excluded firm, which lowers both quality and price. The exclusive firm raises quality and price; when competition is weak, the price increase can exceed the quality gain, so all consumers are worse off. The results also imply that banning exclusivity only for dominant firms can backfire by shifting exclusivity to weaker firms and reducing welfare.Bedre Defolie, Biglaiser, and Jullien (2025) study how a startup endogenously chooses the direction/fit of innovation to potential acquirers to intensify bidding competition, making “decreasing dominance” (a laggard becoming the leader) more likely while simultaneously lowering total and consumer welfare due to distorted innovation direction. Bedre-Defolie and Nitsche (2020) provide an overview of factors that facilitate or mitigate tipping in multi‑sided platform markets. It provides policy‑relevant screening questions and clarifies trade‑offs between short‑run welfare gains from network effects and long‑run losses from reduced innovation/quality after tipping.The project’s findings were disseminated through peer-reviewed publications, working papers, seminars, conferences, and policy-oriented contributions. Several outputs speak directly to ongoing regulatory debates, including self-preferencing rules for gatekeeper platforms, commission regulation and payment routing in app stores, platform liability for harmful products, limits on exclusivity, and the competitive effects of startup acquisitions.

Overall, the project delivered tractable but rich theoretical and empirical frameworks for platform markets that jointly determine prices, entry, variety, and quality—dimensions that standard two-sided market models often could not analyze together. The hybrid-platform framework provides a theory of harm operating through endogenous commissions and entry; the empirical Amazon project measures these effects in practice; the curation framework explains why platforms may under-enforce quality; the app-platform pricing model identifies a sufficient statistic governing distortions and regulatory incidence; the intermediary-design papers show how targeting and information policies change firms’ product menus and market outcomes; the exclusivity work shows how exclusive ownership of quality-enhancing technologies can harm consumers; and the startup-acquisition model explains how acquisition incentives shape the direction of innovation itself. Together, these contributions expand the set of policy-relevant questions that can be studied with models closely tied to real platform instruments and practices.