Working Papers:

‍First They Came for the Others: A Theory of Divide-and-Conquer (with Jinyuqi Huang and Wooyoung Lim)

Abstract: Divide-and-conquer tactics often succeed not through mechanical coordination failures, but through epistemic friction regarding an aggressor's underlying intent. When an attacker strikes a first target, bystanders must infer whether the assault represents a localized grievance or a systemic campaign. If the attack is rationally interpreted as particularized, bystanders abstain, prompting the isolated victim to surrender. We demonstrate how higher attack costs and lower correlation between victims' fates facilitate this division. We then study how behavioral responses, rhetoric, treaty commitments, and downstream defense networks modify this inference.

Top Trading Cycles in Large Markets: The Asymptotic Irrelevance of Priorities (with Olivier Tercieux)

Abstract. Top Trading Cycles (TTC) is Pareto efficient and strategy-proof and explicitly uses agents' priorities. Although TTC favors higher-priority agents in each round, we show that this priority advantage vanishes as the market grows large under a canonical random model of preferences and priorities. In the limit, TTC produces assignments with virtually the same incidence of justified envy as Random Serial Dictatorship (RSD) -- a mechanism entirely blind to priorities. This stark asymptotic equivalence implies that TTC effectively fails to satisfy standard fairness criteria in large markets, casting significant doubt on its practical appeal for balancing efficiency and fairness.

Learning Against Nature: Minimax Regret and the Price of Robustness (with Longjian Li and Tianling Luo)

Abstract: We study how a decision-maker (DM) learns from data of unknown quality to form robust, “general-purpose” posterior beliefs. We develop a framework for robust learning and belief formation under a minimax-regret criterion, cast as a zero-sum game: the DM chooses posterior beliefs to minimize ex-ante regret, while an adversarial Nature selects the data-generating process (DGP). We show that, in large samples of n signal draws, Nature optimally induces ambiguity by choosing a process whose precision converges to the uninformative signals at the rate 1/ √ n. As a result, learning against the adversarial DGP is nontrivial as well as incomplete: the DM’s ex-ante regret remains strictly positive even with an infinite amount of data. However, when the true DGP is fixed and informative (even if only slightly), our DM with a robust updating rule eventually learns the state with enough data. Still, learning occurs at a sub-exponential rate—quantifying the asymptotic price of robustness—and it exhibits “under-inference” bias. Our framework provides a decision-theoretic dual to the local alternatives method in asymptotic statistics, deriving the characteristic 1/ √ n-scaling endogenously from the signal ambiguity.

Dynamic Market Design forthcoming in Econometric Society monograph 2025 World Congress, Vol 2, Ch 3.

Abstract: Classic market design theory is rooted in static models where all participants trade simultaneously. In contrast, modern platform-mediated digital markets are fundamentally dynamic, defined by the asynchronous and stochastic arrival of supply and demand. This chapter surveys recent work that brings market design to this dynamic setting. We focus on a methodological framework that transforms complex dynamic problems into tractable static programs by analyzing the long-run stationary distribution of the system. The survey explores how priority rules and information policy can be designed to clear markets and screen agents when monetary transfers are unavailable, and, when they are available, how queues of participants and goods can be managed to balance intertemporal mismatches of demand and supply and to spread competitive pressures across time.

Allocating Students to Schools: Theory, Methods, and Empirical Insights(with Julien Grenet, Yinghua He), Ch. 4, Handbook of the Economics of Matching

Abstract: This chapter surveys the application of matching theory to school choice, motivated by the shift from neighborhood assignment systems to choice-based models. Since educational choice is not mediated by price, the design of allocation mechanisms is critical. The chapter first reviews theoretical contributions, exploring the fundamental trade-offs between efficiency, stability, and strategyproofness, and covers design challenges such as tie-breaking, cardinal welfare, and affirmative action. It then transitions to the empirical landscape, focusing on the central challenge of inferring student preferences from application data, especially under strategic mechanisms. We review various estimation approaches and discuss key insights on parental preferences, market design trade-offs, and the effectiveness of school choice policies.

Pandora’s Box Reopened: Robust Search and Choice Overload (with Sarah Auster)

Abstract: This paper revisits the classic Pandora’s box problem, studying a decisionmaker (DM) who seeks to minimize her maximal ex-post regret. The DM decides how many options to explore and in what order, before choosing one or taking an outside option. We characterize the regret-minimizing search rule and show that the likelihood of opting out often increases as more options become available for exploration. We show that this “choice overload” is driven by the DM’s fear of “selection error”—the regret from searching the wrong options— suggesting that steering choice via recommendations or cost heterogeneity can mitigate regret and encourage search.

Optimal Auction Design for Dynamic Stochastic Environments: Myerson Meets Naor (with Andy Choi)

Abstract: Allocation of goods and services often involves both stochastic supply and stochastic demand. Motivated by applications such as cloud computing, gig platforms, and blockchain auctions, we study the design of optimal selling mechanisms in an environment where buyers with private valuations arrive stochastically and are assigned goods that also arrive stochastically, and either buyers or goods can be held in a queue at costs until allocation. The optimal mechanism dynamically leverages competitive pressure across time by managing the queue of buyers and inventory of goods, using reserve prices that increase with the number of buyers in the queue and decrease with the number of items in inventory, and an auction to allocate the goods.

Data-Driven Monitoring and Deterrence in a Changing Environment, (with Jinwoo Kim and Konrad Mierendorff)

Abstract: We study a dynamic model in which a principal monitors agents based on historical data of infractions. This data informs when and at what intensity to monitor; the monitoring decision, in turn, selects the collected data, shaping the principal’s future learning. We analyze this feedback loop using a bandit model in which the underlying monitoring environment evolves according to a hidden Markov process. Because data collection is endogenous, how the principal uses this information is critical: surprisingly, a myopic approach renders historical data completely valueless. By endogenizing the agent’s incentives, we demonstrate that the principal’s purely informational motive to explore serves as an endogenous commitment device. This inherent drive to gather data compels persistent vigilance, strictly lowering the equilibrium infraction rate and restoring the power of deterrence.

Prestige Seeking in College Application and Major Choice (with Dong Woo Hahm, Jinwoo Kim, Se-jik Kim, and Olivier Tercieux)

Abstract: We develop a signaling model of prestige seeking in competitive college applications. A prestigious program attracts high-ability applicants, making its admissions more selective, which in turn further increases its prestige, and so on. This amplifying effect results in a program with a negligible quality advantage enjoying a significant prestige in equilibrium. Furthermore, applicants “sacrifice” their fits for programs in pursuit of prestige, which results in the misallocation of program fits. Major choice data from Seoul National University provides evidence for our theoretical predictions when majors are assigned through competitive screening—a common feature of college admissions worldwide.

Monotone Comparative Statics without Lattices (with Jinwoo Kim and Fuhito Kojima):

Abstract:  The theory of Monotone Comparative Statics (MCS) has traditionally required a lattice structure, excluding certain multi-dimensional environments like mixed-strategy games where this property fails. We show this structure is not essential. We introduce a weaker notion, pseudo lattice property, and preserve the theory’s core results by generalizing the MCS theorems for individual choice and Tarski’s fixed-point theorem. Our framework expands comparative statics to pseudo quasi-supermodular games. Crucially, it enables the first MCS analysis of mixed strategy Nash equilibria and (trembling-hand) perfect equilibria.

Statistical Discrimination in Ratings-Guided Markets (with Kyungmin Kim and Weijie Zhong):

Abstract: We study statistical discrimination of individuals based on payoff-irrelevant social identities in markets where ratings/recommendations facilitate social learning among users. Despite the potential promise and guarantee for the ratings/recommendation algorithms to be fair and free of human bias and prejudice, we identify the possible vulnerability of ratings-based social learning to discriminatory inferences on social groups. In our model, users' equilibrium attention decisions may lead data to be sampled differentially across different groups so that differential inferences on individuals may emerge based on their group identities. We explore policy implications in terms of regulating trading relationships as well as algorithm design.