Publications
Dynamic Voluntary Contribution to a Public Project under Time-Inconsistency with Murat Yılmaz
Journal of Economic Behavior & Organization, January 2018, 145: 114-140.
Working Papers
Dynamic Many-to-One Matching [Last Updated: January 2025]
Abstract. We study many-to-one matching markets in a dynamic framework with the following features: Matching is irreversible, participants exogenously join the market over time, each agent is restricted by a quota, and agents are perfectly patient. In such markets, a form of strategic behavior emerges: The side with many slots can manipulate the subsequent matching market in their favor via earlier matchings. In such a setting, a natural question arises: Are there situations under which such manipulation incentives are ruled out, and thus we can analyze a dynamic many-to-one matching market as if it were a static market? We provide sufficient conditions under which the answer is yes.
Dynamic Pricing under Loss Aversion: To Match Own-Price, or Not? with Murat Yilmaz
Abstract. We consider a monopolist who is facing loss-averse buyers over two periods, with an uncertain demand in the second period. Monopolist cannot commit today to a future price, and posts the price for tomorrow only after the demand realization. We incorporate best-price provision policy into the monopolist’s problem, in the form of a most-favored-customer (MFC) clause, which is an own-price match guarantee. We solve for the dynamic pricing problem of such a monopolist under loss aversion both with and without an MFC clause. We show that for a wide range of discount factors, the monopolist optimally adopts MFC if and only if the buyers are loss-averse. Also, the result holds for any discount factor when buyers are sufficiently loss-averse. We also make inferences on the residual demand, which reflects the set of buyers who choose to wait for tomorrow to purchase.
Designing the menu of Licenses for Foster Care with Diana E. MacDonald [Last Updated: January 2025]
Abstract. This paper studies the menu of licenses designed by the child welfare agency to screen foster parents. We develop a two-sided matching model with heterogeneous agents, search frictions, private information, and a designer who coordinates match formation through a menu of contracts. We focus on incentive-compatible contracts, examine optimal transfers, and analyze sorting patterns that arise in equilibrium. We establish three main results: (i) foster parents in different licenses will never care for the same group of foster children, (ii) complementarities do not ensure that Positive Assortative Matching (PAM) will arise in equilibrium, thus we provide additional condition that guarantees PAM, and (iii) the equilibrium transfer scheme is not unique and does not affect the unique equilibrium sorting pattern. Moreover, our results suggest that an optimal menu of licenses must not only account for the child’s attribute (as it is in practice), but also for other characteristics of the market such as parents’ types, supply of parents, and supply of children.
Dynamic Matching: Shall we commit to not competing?
Abstract. Appointments occur sequentially in various many-to-one matching markets such as college admissions and entry-level professional labor markets in the US. Although literature rations early matchings with a competition argument—that is, institutions compete to hire better agents early on—early matchings in some markets (e.g. job market for Finance Ph.D. candidates) show inferior quality. This paper introduces a suitable notion of dynamic group stability and shows that institutions can rationally benefit from committing to match worse early on. Such a commitment lowers the competition for better agents in the subsequent market.
Best-Price Provision under Loss Aversion with Murat Yilmaz
Abstract. We consider a seller who is facing loss-averse buyers over two periods, with an uncertain price in the second period. We incorporate the decision to adopt a most-favored-customer (MFC) clause into the seller's problem, and solve for the optimal price and the optimal MFC decision. We show that for sufficiently patient buyers, the seller optimally adopts MFC if and only if the buyers are loss averse. Also, regardless of the patience, the seller optimally adopts MFC if and only if the buyers are sufficiently loss averse. Moreover, MFC negatively (positively) affects the expected volume of trade when buyers are (not) loss averse.
A Dynamic Search and Matching Model of Foster Care: Theory and Evidence with Diana E. MacDonald [Draft coming soon]
Abstract. This paper examines the two-sided dynamic matching problem in the US foster care system, where foster parents and children form reversible matches that may separate, persist, or transition to adoption. We first present new empirical facts on match transitions using administrative data. We then develop a two-sided search and matching model to rationalize these facts and generate predictions. Our findings show that a financial penalty on adoption exacerbates the intrinsic disadvantage faced by children with greater care needs and discourages high-quality matches from transitioning to adoption, highlighting important policy implications for improving permanency outcomes in foster care.
Skill Accumulation in Sequential Matching Markets with Vladsylav Nora [Draft coming soon]
Abstract. This paper investigates dynamic matching markets where agents' attributes evolve based on earlier matchings. We model a two-period setting involving students, colleges, and employers, exploring how Positive Assortative Matching (PAM) emerges and its implications for efficiency and inequality. Our analysis reveals that while PAM maximizes output given a fixed attribute distribution, decentralized markets may generate suboptimal outcomes due to inefficient human capital accumulation. We provide sufficient and necessary conditions for PAM to always be optimal and demonstrate how centralized interventions, such as committing to non-PAM strategies, can reduce inequality and potentially improve welfare. An illustrative application highlights the practical relevance of these results.
Work in Progress
Desert Preferences and Moral Hazard in Teams, with M. E. Ertuğrul & Murat Yilmaz
Screening Loss Aversion through Price-mathcing Policy, with H. Basaran & Murat Yilmaz
Allocation of State Grants for Colleges: Case of Kazakhstan, with A. Taken