Foundations for Antitrust and Policy on Digital Platforms

Introduction

Internet markets tend to concentrate in the hands of a very few large platforms. These platforms have been accused of abusing their monopoly power vis-a-vis their users (consumers, sellers, advertisers) and maintaining the latter through hostile behavior towards potential competitors.

Concerns of Monopoly Power

They are said to harm users by:

• Self-preferencing
• Data harvesting
• Creating 'monopoly positions'
• Extracting resulting rents with high fees

All the while avoiding competition by acquiring, copying, and otherwise disadvantaging potential competitors. This proposal addresses these concerns in four parts.

Part 1: Dual Role of Online Marketplaces

Part 1 focuses on the dual role of online marketplaces, whereby the platform both runs the marketplace and acts as a seller on it. I aim to understand how such hybrid marketplaces conduct themselves toward consumers and third-party sellers.

The model will be used to evaluate recent policy proposals and suggest ways to avoid significant unintended consequences.

Part 2: Steering Consumers to Sellers

Part 2 studies how platforms steer consumers to sellers. As most platforms let sellers set prices and collect fees on revenues, a platform's own algorithm and its choice to augment/replace it with a position auction may be consequently driven by revenue maximization.

I plan to show that steering systems may drastically alter pricing, leading to mediated competition.

Part 3: Nature of Recommendation Algorithms

Part 3 explores the nature of recommendation algorithms, particularly the interplay between consumer search and algorithm effectiveness.

I demonstrate that algorithms may be self-fulfilling and self-defeating, which determines their effectiveness and significantly alters the resulting allocations and their efficiency.

Part 4: Acquisitions of Upstart Platforms

Part 4 explains the circumstances in which an incumbent platform may acquire an upstart platform. These depend on:

1. The overlap of existing user bases
2. Increasing returns to data
3. Monopoly power over advertisers

Acquisitions may be used in situations of both no and substantial overlap in user bases, with mixed welfare consequences.