Dhruva Bhaskar

Research

Working Papers

Regulation Design in Insurance Markets (with Andrew McClellan and Evan D. Sadler). R&R @ AER

Regulators often impose rules that constrain the behavior of market participants. We study the design of regulatory policy in an insurance market as an optimal delegation problem. A regulator restricts the menus of contracts an informed firm is permitted to offer, the firm offers a permitted menu to each consumer, and consumers choose contracts from offered menus. Under an ordering on consumer types and firm information, the regulator can fully leverage the firm’s information by forcing the firm to offer specified additional options on each menu. Several extensions illustrate the applicability of our framework.

Dynamic Screening and the Dual Roles of Monitoring.

This paper studies optimal monitoring of agent behaviour in long-term self-enforcing relationships with adverse selection and moral hazard. At the beginning of the relationship, monitoring is used to screen agents by deliberately incentivising low types to shirk in order to detect and fire them. After high types have been identified, monitoring is used both to incentivise effort from high types and to provide off-path (dis)incentives for low types looking to build a reputation early in the relationship. This gives rise to a key trade-off: providing cost-efficient incentives to high types versus screening low types faster. The optimal contract uses probation after screening: high types are overworked and excessively monitored to facilitate better screening. Monitoring decreases over time as the agent's track record improves.

Resource Allocation with Positive Externalities (with Evan D. Sadler).

In several common allocation problems, transfers are unavailable, but incentives are partially aligned because the allocation to one player entails positive externalities to the other. We study the extent to which a designer can exploit this alignment when allocating a finite resource between n players. We identify a natural mechanism that partitions the type space into a collection of intervals and allocates the resource among players in the highest reported interval. While interim allocations are identical for all reports in the same interval, the exact allocation depends on the lowest reported interval. This novel feature is a crucial source of incentives. We show that our mechanism is optimal in special cases, and it is approximately efficient when n is large.

Dynamic Inspection and the Value of Information.

This paper studies a dynamic inspection game between a principal and agent with hidden type. The principal employs an agent with private information about his type, and strictly prefers his outside option to employing a low type. Every period that the agent is employed, the principal can perfectly monitor the agent's actions at a cost. The agent can either exert effort or shirk. The high type's preferences are aligned with the principal so he strictly prefers to exert effort. The low type finds it costly to exert effort, but may find it worthwhile to do so in order to deceive the principal. I show that the game generically has a unique equilibrium, which is a war of attrition between the principal and the low type. The main result finds that monitoring has value for the principal if and only if he is sufficiently pessimistic enough about the agent's type. Monitoring decreases as the agent's reputation grows. Eventually, either the low type is caught and fired, or the principal trusts the agent enough that he stops monitoring forever.