Incentives for Exploration at Competitive Equilibrium
Vijay Kamble, Assistant Professor of Information and Decision Sciences
University of Illinois Chicago
Abstract: Online marketplaces face an exploration problem: the qualities of new supply units are unknown and must be discovered through customer feedback so that higher-quality supply gets prioritized for matching. However, customers are generally myopic and unwilling to participate in exploratory matches, leading to the well-known concern of incentivizing such exploration. This paper uncovers the role of competitive pricing effects arising from market congestion in creating incentives for exploratory behavior among myopic customers. The intuition is that since established higher-quality supply units are expected to be more popular and, hence, more congested, they naturally demand higher prices at a competitive equilibrium than new supply units, effectively incentivizing customers to explore. This paper presents a comprehensive analysis of the extent to which this intuition holds and the extent to which exogenous incentives for exploration are necessary for such markets.
To investigate this question, we define a novel competitive equilibrium notion for markets with evolving public information about supply units. The key result establishes that, under a tightly characterized market regularity condition, the ratio of equilibrium matching value to the system optimal value is bounded by the aggregate congestion level, indicating that congested regular markets inherently incentivize exploration. We also show that this bound is optimal in the worst case. Furthermore, we show that in markets with linear information transition structures, the equilibrium achieves the first-best system-optimal value regardless of the congestion level. Finally, we address the problem of designing optimal price interventions to align market equilibrium with the system-optimal solution. Overall, our results inform market designers grappling with the concern of incentivizing exploration in various online marketplaces, shedding light on when interventions may be (un)necessary.
Bio: Vijay Kamble is an Assistant Professor of Information and Decision Sciences at the University of Illinois Chicago, with a courtesy appointment in Computer Science. He obtained his Ph.D. in Electrical Engineering and Computer Sciences at UC Berkeley in 2015, after which he spent two years as a post-doctoral scholar at the Society and Algorithms lab at Stanford University. His research is on the design and optimization of online platforms and marketplaces. In particular, his research (a) develops new decision-making algorithms and incentive mechanisms that effectively leverage data to optimize market operations and (b) designs new algorithmic frameworks to address concerns of fairness, ethics, and bias in AI-driven market decisions. His professional experiences include working with Technicolor Labs and advising firms like Stitchfix, LivSYT, and others on algorithmic automation and AI use.
