Financing Rivals (with Jorge Lemus), R&R at the Journal of Economic Behavior and Organization.
We study how a lender with market power shapes competition between liquidity-constrained rivals by choosing contract terms strategically. In our framework, two firms compete for a prize and a monopolistic lender decides which firms to finance and on what terms. We show that the lender finances both rivals when their internal resources are similar, but finances only the more constrained firm when asymmetries are large. When financing both firms, the lender optimally caps credit because full funding induces an inefficient investment arms race that lowers the surplus available for extraction. We also show that, even when the lender can use fees, caps, and interest rates, an optimal contract sets interest rates to zero and relies on non-distortionary instruments instead. Thus, the impact of common lenders on competition depends critically on the initial distribution of resources.
Dynamic Peace Renewal (with Benjamin Matta), submitted.
We develop a dynamic model of peace maintenance in which an existing state of peace must be repeatedly renewed through costly interventions. Players receive private stochastic opportunities to intervene. Each intervention resets a finite peace window and lowers the loss associated with eventual war, which occurs if the window expires without intervention. The unique symmetric Markov Perfect Equilibrium is characterized by a recursively determined sequence of thresholds. Within each peace window, players strategically delay intervention, free riding on others’ future opportunities and de-escalating only as expiration approaches. Across windows, declining benefits from successive renewals lead to later intervention thresholds, lower renewal probabilities, and longer waiting times conditional on renewal. A social planner, by contrast, internalizes the shared benefits of intervention and uses every opportunity until a terminal level of de-escalation is reached.
Provider and Worker Responses to Supply-Side Audits in Social Insurance (with P. Celhay, N. Figueroa, and B. Oteiza).
Social insurance often relies on third-party gatekeepers to certify eligibility, but planners cannot easily verify whether those certifications are accurate. We study this problem in the market for medical sick leave using the universe of sick leave certificates in Chile from 2018-2022 (about 22 million records). We estimate how a sharp increase in enforcement changes behavior on both the supply side (physicians) and the demand side (workers). We combine matched difference-indifferences and regression discontinuity in time on physicians, and a design that addresses non-random patient exposure. Enforcement produces a large and persistent drop in issuance among the highest-issuing physicians, with smaller spillovers to similar non-sanctioned physicians. Workers who were more exposed to these physicians reduce sick-leave use, but the final effects are smaller than the supply contraction, consistent with imperfect substitution on the worker side across providers. Workplace networks strengthen these effects, suggesting that information and access to substitute certifiers are channeled through coworkers. The findings show how substitution and spillovers shape the ultimate impact of monitoring policies.
We study a market in which the buyer has no information about product quality, while the seller has private probabilistic information about it. Buyers observe price and can procure an inspection, which provides valuable information about the good for sale. With costless inspections, there is no separating equilibrium. We then show that when information acquisition is costly, there is a separating equilibrium that satisfies the intuitive criterion, in which high prices signal high quality and furthermore, the dynamic separating equilibrium showing higher separating prices than the static one. Finally we discuss the implications of time-on-the-market on separating equilibria. Specifically, when there is only one asset on sale over both periods (therefore both price and time-on-the-market can signal quality) there is no separating equilibrium even if single-crossing is satisfied. The key to this result is that the second-period buyer cannot observe why the asset did not sell in the first period. Notably, the failure to sell can be attributed to overpricing or an unfavorable inspection outcome. Therefore the copycat behavior is more attractive to the poor-quality seller because he benefits more from an increase in buyer beliefs than his high-quality counterpart. Allowing only the first-period buyer to acquire information on quality, we show the existence of a separating equilibrium in which high prices and time-on-the-market signal high quality.