Maxim Senkov

Abstract. 

In a continuous-time moral hazard problem, an agent chooses to start shirking either if the multistage project is completed or if the project is unlikely to be completed before the final date. A principal wants to convince the agent to incur effort as long as possible and can design the flow of information about the progress of the project to persuade the agent. If the project is sufficiently promising ex ante, then the principal commits to providing only the good news that the project is accomplished. If the project is not promising enough ex ante, the principal persuades the agent to start incurring effort by committing to provide not only good news but also bad news that a project milestone has not been reached by an interim date. We show that it is optimal for the principal to promise immediate provision of the good news and to release the bad news at a deterministic date - an interim deadline. The model sheds light on a supervisor-supervisee relationship in scientific research.

Microtargeting allows politicians to design policies that target narrow groups of citizens by using fine-grained voter data. How does this practice affect the welfare of voters? The paper provides a model where the incumbent and the challenger compete for office by offering targeted benefits to voters. Voters differ in their affinities to the politicians and the value they derive from targeted benefits. We compare social welfare when politicians can make offers to voters with specific characteristics to a scenario where they choose the mass of benefits they are ready to provide, and then voters receive them uniformly at random. In the first scenario, which models microtargeting, politicians offer targeted benefits to swing voters who put a relatively high value on them. Moreover, microtargeting forces the incumbent to make more generous offers by making the challenger more effective in getting supporters. Both effects are present when the incumbent is sufficiently popular compared to the challenger. We present the conditions under which microtargeting benefits voters.

Abstract. 

This paper shows that a biased principal can strictly benefit from delegating to a misaligned agent. We consider a “delegated expertise” problem in which the agent has a learning advantage relative to the principal.  We show that it is optimal for a principal who is ex ante biased towards one action to select an agent who is less biased. Such an agent is more uncertain ex ante about what the best course of action is and would acquire more information. The benefit to the principal from a more informed decision always outweighs the cost of a small misalignment.

Abstract. 

We study a Bayesian persuasion model in which the state space is finite, the sender and the receiver have state-dependent quadratic loss functions, and their disagreement regarding the preferred action is of arbitrary form. This framework enables us to focus on the understudied sender's trade-off between the informativeness of the signal and the concealment of the state-dependent disagreement about the preferred action. In particular, we study which states are pooled together in the supports of posteriors of the optimal signal. We provide an illustrative graph procedure that takes the form of preference misalignment and outputs potential representations of the state-pooling structure. Our model provides insights into situations in which the sender and the receiver care about two different but connected issues, for example, the interaction of a political advisor who cares about the state of the economy with a politician who cares about the political situation.

Abstract. 

In a persuasion model, we show that the receiver's ignorance regarding the correlation between the dimensions of the state of the world can benefit her. The receiver chooses a binary action and the preferences of the sender and receiver regarding that choice depend on different dimensions of the two-dimensional state of the world. The receiver has a “simplistic worldview”: her prior belief is such that she perceives the two dimensions of the state of the world as independent when they are correlated. We characterize the condition on state of the world distribution that pins down if the posterior of the receiver with a simplistic worldview is either less or more optimistic than the true one. Receiver's simplistic worldview either restricts or expands what sender can do and induces a welfare redistribution either from sender to receiver or from receiver to sender, respectively.