Algorithmic Steering and Advertising on Platforms

The paper presented in this ZEW Research Seminar analyzes steering/self-preferencing by a retail platform, using both economic theory and simulations of AI algorithms. When the platform does not sell advertising slots to merchants (which guarantee the merchant will be displayed to consumers), the platform has incentives to exclude merchants when its commission rate is in a lower range, but not when its commission rate is higher. Facing the threat of not being displayed, the merchant’s optimal response is to raise its price on the platform, so as to reduce the substitution threat it poses to the platform’s own product. Thus, even when the equilibrium outcome is for the merchant to be displayed, the self-preferencing possibility raises overall prices and harms competition. The authors then introduce an advertising option, and show that this can lead to lower prices while still ensuring the merchant is displayed to consumers. However, the merchant need not gain from the introduction of advertising. Various extensions are considered.