First-party selling and self-preferencing

In this paper, I analyze the welfare effect of a vertically integrated gatekeeper platform selling its own first-party product, i.e., first-party selling, as well as the platform's incentive to favor the first-party product in the product recommendations it makes, i.e., self-preferencing. I find that, irrespective of self-preferencing, both consumer welfare and platform revenue are higher under first-party selling because first-party selling mitigates double marginalization. Additionally, third-party product prices are lower in expected terms under first-party selling, either because the platform reduces the commission fee (with self-preferencing) or downstream competition is fiercer (without self-preferencing). Finally, I show that both consumers and the platform are better off if the platform commits not to engage in self-preferencing. • A platform can sell its own (1P) product in the marketplace (1P selling). • The platform may recommend the 1P product favorably to consumers (self-preferencing). • 1P selling reduces double marginalization and increases welfare. • The platform's revenue and consumer welfare are higher without self-preferencing. • If the platform has commitment power, it does not engage in self-preferencing.