Working papers:
Learning, Arms Race, and the Optimal Monitoring of Adaptive Offenders (new draft coming soon)
I study a dynamic arms race with asymmetric information and its impact on enforcement design. An agent chooses offenses and invests in evasion technologies, and a monitor invests in detection technologies. These investments determine the monitor's ability to detect offenses. I analyze how enforcement policies affect investment dynamics, revealing a trade-off: Deterring detectable misbehavior leads to a costlier arms race due to increased investments by the monitor and more evasion by the agent. When this cost is too high, the optimal policy tolerates high levels of detectable offenses in order to avoid the arms race. The model's applications include digital security, smuggling, money laundering, and tax evasion.
Designing contracts for technology procurement (new draft coming soon)
This paper studies the problem of optimal procurement in the presence of uncertainty about relevant production technology. A buyer chooses a symmetric procurement mechanism that depends on this technology and the set of relevant sellers. Each seller observes his production cost and the trade mechanism before choosing production technology. The main result of this paper shows that the optimal mechanism induces mixing in the technology adoption by the least efficient agents' types whereas the most efficient ones adopt only the technology that is most likely to succeed. Applied to a generalized auction setting, mechanisms involving mixing induce more efficient trade (in expectation) and a more aggressive bidding behavior by the least efficient types.
Work in progress:
Resource allocation in the presence of moral hazard and endogenous adverse selection (Joint with Esteban Muñoz)
This paper studies the problem of resource allocation in the presence of moral hazard. An agent exerts effort and privately chooses resource allocation between two types of capital: one that increases the productivity of effort and one that reduces its cost. Our analysis provides conditions such that the agent's problem exhibits complementarity between effort and productivity. In this case, we show that the agent under-allocates resources to increase his productivity. The main result of the paper provides sufficient conditions on the production problem such that the agent strictly benefits from the allocation being private information. The model can be applied to several economic environments, such as technology procurement, product development, and time allocation in labor settings.
Efficiency, contagion and the nature of decentralization in criminal networks
Criminal groups often choose their network structure as a response to constraints from their environment such as the nature of their activity and policing. These constraints lead to a trade-off between exposure, which favors more decentralization, and efficiency which requires more communication and centralization. This paper provides a setting to study the properties of optimal hierarchies when a criminal group faces a risk of disruption. I show that more efficient policing leads to less centralized structures. The main result of this paper shows that the nature of this decentralization depends on the efficiency losses from inefficient communication. The model predicts that when the group's activity requires more coordination, decentralization leads to splits in the network leading to cell-based structure and small operational units. Conversely, activities requiring less coordination decentralize vertically and favor hierarchical structures. These results contribute to explaining the differences in network structures between drug trafficking and mafia organizations.