Research

Abstract

Search platforms, such as Google and Bing, increasingly answer queries with artificial intelligence (AI) search features that scrape and synthesize content from online publishers. In this paper, I develop a theoretical model to examine how the introduction of AI search features affects content investment decisions, platform design choices, and welfare. Although improving AI content quality increases the direct benefit each user receives on the search platform, it also diverts users away from each publisher’s site. This weakens the publishers' incentives to invest in high quality content indirectly reducing the quality of the AI content. Therefore, it is optimal for the search platform to degrade the quality of its AI search feature to temper this feedback effect. In face of this tradeoff, I show that AI search features improve user surplus, but may reduce social welfare if publishers sufficiently value user clicks. Applying my framework to several proposed policies, I find that copyright restrictions can implement the socially optimal outcome; platform competition reduces social welfare; and opt-out options, licensing deals, and pay-per-crawl payments have ambiguous effects on welfare.

Abstract

Artificial intelligence (AI) synthesizes existing knowledge into refined knowledge, which the human brain then recombines to generate new ideas. By lowering the cognitive burden of information processing, AI can accelerate discovery, but it may also reduce knowledge spillovers by filtering out information that later proves valuable. We show that research productivity varies nonmonotonically with AI efficiency: AI boosts innovation when knowledge is very scarce or very abundant, yet may create a mid-knowledge level AI trap where faster AI progress slows down innovation and lowers productivity. In our endogenous growth model, faster AI raises long-run growth, but its effect on R&D labor share is ambiguous. 

Abstract

Many online retail platforms, such as Amazon and JD.com, have recently begun providing sellers with proprietary consumer data. In this paper, we investigate how the sharing of such data can incentivize seller collusion through personalized pricing. We find that the effect of data sharing on collusion sustainability and profitability depends on the mode operated by the platform. When a platform acts as both a host and seller, sharing data leads to more collusive outcomes. However, when a platform only intermediates purchases, sharing data hinders collusion. Our results suggest that imposing a ban on data sharing may be ineffective and harmful to consumer welfare. 

Abstract

Platforms use fee concessions to recruit sellers and costly support to build their capability. We develop a two-period model of a monopoly platform in which heterogeneous sellers choose when to enter, only early entrants acquire capability, and the platform cannot commit to future fees. Future entrant fees determine the value of waiting and hence the concession required for early entry. As the return to early capability rises, the unique limited-commitment equilibrium passes through four regimes: positive late entry, a zero-late-entry plateau, renewed early recruitment, and seller saturation. The plateau arises because recruiting an additional early seller lowers the waiting concession paid to the entire early cohort, a saving that disappears when late entry closes. At that boundary, the marginal early seller would earn no rent by waiting, yet the platform still subsidizes early entry to expand buyer-side revenue. With active late entry, greater capability front-loads seller entry yet expands period-2 participation on both sides. When the platform and sellers discount the future at the same rate, full commitment eliminates late entry but can reduce early recruitment. Whenever allocations differ across commitment regimes, full commitment raises platform profit while reducing period-2 participation on both sides and lowering social welfare, even though support intensity is unchanged. Thus, resolving the platform’s time-inconsistency problem can benefit the platform at the expense of market development.

Abstract

Platforms often enforce pricing policies onto sellers such as the minimum margin agreement (MMA). MMAs require that platforms receive a guaranteed profit-margin on seller goods. If the margin is not met then any difference between the actual margin is taken from the seller. I develop a model where a seller sells through a direct channel and serves as a supplier for a separate platform. Under MMAs, a platform can threaten the seller to flood the market demand by pricing low and capturing a guaranteed profit. I show that MMAs potentially lead to an increase in platform facilitated purchases and inflate both direct and intermediated prices. Furthermore, I find that MMAs may be more prevalent when the seller faces competition and platforms can steer buyers towards specific goods.

Abstract

Bernheim and Whinston (1990) famously show that contact across multiple markets may facilitate collusion given that firms or markets are not identical. As pricing decisions are increasingly made by algorithms, antitrust authorities are concerned that algorithms may autonomously collude on supracompetitive prices. In this paper, we test how multimarket contact affects the behavior of pricing facilitated by reinforcement learning algorithms via Q-learning.