Chenyang Li
My papers
I work in applied microeconomics theory, mechanism design, and industrial organization.
Can you make people work without directly supervising them? The answer is yes if you have multiple agents by creating an architecture where they supervise one another. Complication arises from the fact that, the agents may collude and jointly deviate to no effort and no peer supervision. This paper models the collusion formation process and characterizes the conditions under which the collusion may or may not occur. The central insight is: if the principal can limit the communication among the agents, it is much easier to deter collusion.
I study two ways of joint deviation: voting and commitment. When all agents are directly connected in a communication network, deviation by voting can be stopped if and only if the threshold of passing the vote is sufficiently high. Deviation by commitment, however, cannot be stopped. However, if the principal limits the initial communication network to a ``ring'', the joint deviation can be deterred when the passing threshold is at least three people no matter the total number of players. Commitment can be stopped when there are at least six agents in the department. The negotiation power of an arbitrary individual in an arbitrary network can also be calculated by an algorithm. The findings give us insights into firm management, and political control, and also designing mechanisms for controlling corruption.
With Yangguang Huang (HKUST) and Si Zuo (Cornell)
Earning a good reputation is crucial for the survival of new firms on online retailing and service platforms. With a dynamic price signaling model, we show that a high-quality firm can signal its unobserved quality by setting a lower introductory price than its low-quality counterpart. After accumulating sufficient favorable reviews, the high-quality firm will raise its price and enjoy a quality premium. Using data from Zaihang, a consulting service platform, we find empirical evidence that experts with high unobserved ability indeed adopt low introductory prices and exhibit a rising price dynamic over time. We use the performance of the expert on another platform as an instrument for the expert's ability on Zaihang to provide evidence that the relationship is causal. Our empirical findings reject alternative models in which firms do not know their own types, or consumers can observe firm types.
With Michael Waldman (Cornell) and Haimeng Hester Zhang (IESE)
Observation of real-world markets suggests that many products are produced at below efficient built-in durability levels, and/or new products are introduced quickly which inefficiently reduces the useful life of durable products. Most of the prior literature on this subject explains these observations employing monopoly/market power models, but a number of the markets that exhibit these behaviors are competitive. We consider models in which consumers have time-inconsistent/present-biased preferences, as first put forth in the seminal analysis of Strotz (1955), and show that present-biased consumer preferences can lead to equilibrium durability below efficient levels and inefficiently quick new-product introductions, even in competitive markets. We also investigate circumstances in which market power aggravates these distortions. In addition to deriving these theoretical results, we relate our theory to recent regulatory changes in the light bulb industry, as well as to the behavior of the well-known Phoebus light bulb cartel of the 1920s and 1930s.
With Yangguang Huang (HKUST) and Si Zuo (Cornell)
With the rise of e-commerce, more chain stores have opened online sales channels. For one chain, there are usually one online store and many offline stores. Online stores may cannibalize the sales of the existing physical stores because of their advantage in lower shopping costs. On the other hand, the online sales channel is usually a tool for advertisement, which may expand the offline store's market. Also, the existence of the online store could increase consumers' utility in shopping in the offline store since they could check the product information before shopping in the physical store. From our novel daily revenue data of 380 offline stores from 2016 to 2020, we separately identify and estimate the countervailing cannibalization effect and the informative effect of opening up online branches on offline stores. We first use offline or online exclusive exogenous demand shocks (rainy days, Covid-19, and online shopping festivals) to provide solid evidence of these two effects. We then separately estimate these two effects by a nested logit model. We find that the cannibalization effect dominates the informative effect in most cases. The electronics category has the largest cannibalization effect, while the cosmetics and jewelry category has the smallest effect.
In many game-theoretic models, it is common to see multiple equilibria. There are extensive literature on identifying which equilibrium is more likely to occur, arguments like the focal points, evolutionary convergence, learning and Et cetera. But all of them have certain limitations. In this paper, I propose a new setup that enables us to model the equilibrium selection process. Instead of players best responding to each other's actual strategies, I assume that players best respond to their beliefs about the opponents' strategy. On the other hand, the belief is generated from a belief formation function that may take any observables and map that into a specific belief about the opponent's strategy. This way, we may construct models describing how payoff irrelevant signals and off equilibrium play may shape future outcomes and choosing equilibrium. Then I introduce the outer games to model how exogenous shocks to an inner game are generated. I allow players in the outer game to affect the game structure and belief formation signals of the inner game and thus control equilibrium outcome exogenously. I use several examples to illustrate this new setup and how it is different from more conventional sequential games. I finish this paper by discussing further regulatory assumptions on the belief formation functions.