Algorithmic Collusion or Competition: the Role of Platforms' Recommender Systems

TL;DR

The paper addresses how platform recommender systems alter competition in algorithmic pricing by embedding a recommender module into a repeated-game framework of sellers and consumers. It develops a structural sequential-search demand model with heterogeneous preferences and position effects, calibrated via real data, and analyzes two platform objectives—revenue maximization and consumer utility maximization—using Q-learning pricing and numerically integrated recommender policies. The key findings show revenue-focused recommenders amplify collusion and prices, while utility-focused ones reduce collusion, with a counterintuitive rise in prices under more displays at high horizontal differentiation; these effects have important regulatory and design implications. The work advances the economics of AI by studying AI–AI interactions, informs platform design and antitrust considerations, and suggests a three-tower framework that accounts for consumer, seller, and platform responses in dynamic online markets.

Abstract

Recent scholarly work has extensively examined the phenomenon of algorithmic collusion driven by AI-enabled pricing algorithms. However, online platforms commonly deploy recommender systems that influence how consumers discover and purchase products, thereby shaping the reward structures faced by pricing algorithms and ultimately affecting competition dynamics and equilibrium outcomes. To address this gap in the literature and elucidate the role of recommender systems, we propose a novel repeated game framework that integrates several key components. We first develop a structural search model to characterize consumers' decision-making processes in response to varying recommendation sets. This model incorporates both observable and unobservable heterogeneity in utility and search cost functions, and is estimated using real-world data. Building on the resulting consumer model, we formulate personalized recommendation algorithms designed to maximize either platform revenue or consumer utility. We further introduce pricing algorithms for sellers and integrate all these elements to facilitate comprehensive numerical experiments. Our experimental findings reveal that a revenue-maximizing recommender system intensifies algorithmic collusion, whereas a utility-maximizing recommender system encourages more competitive pricing behavior among sellers. Intriguingly, and contrary to conventional insights from the industrial organization and choice modeling literature, increasing the size of recommendation sets under a utility-maximizing regime does not consistently enhance consumer utility. Moreover, the degree of horizontal differentiation moderates this phenomenon in unexpected ways. The "more is less" effect does not arise at low levels of differentiation, but becomes increasingly pronounced as horizontal differentiation increases.

Figures (4)

• Figure 1: Learning Dynamics
• Figure A1: Consumer Search and Purchase Process
• Figure E1: Learning Dynamics ($\mu = 0.25$, $\omega = 5 \times 10^{-8}$)
• Figure F1: Learning Dynamics (MAB)