We propose a generic computational framework for solving large-scale infinite-horizon, discrete-time dynamic incentive problems with persistent hidden types. First, we combine set-valued dynamic programming techniques with unsupervised machine learning to determine irregularly-shaped feasible...

We examine dynamic contracts when output has negative environmental effects and the manager (agent) can invest to build up ESG capital and mitigate the externality. The incentive component of the optimal contract rewards based on cash flow and ESG capital when the principal is risk neutral; and...

Previous literature documents that mutual funds' flows increase more than linearly with realized performance. I show this convex flow-performance relationship is consistent with a dynamic contracting model in which investors learn about the fund manager's skill. My model predicts that flows...

We examine optimal dynamic contracts when the firm's production generates harmful pollution undermining its productivity. The optimal contract rewards for financial performance and penalizes pollution. The combination of both contract sensitivities incentivizes the agent's effort and...

This paper develops a framework to study general equilibrium implications for an economy in which agents are allowed to have dynamically inconsistent time and risk preferences. This framework accommodates, but is not limited to, the following settings: (1) non-exponential discounting; (2)...

This paper considers a long-term relationship between two agents who undertake costly actions or investments which produce a joint benefit. Agents have an opportunity to expropriate some of the joint benefit for their own use. The question asked is how to structure the investments and division...

We present a continuous-time agency model under mean-volatility joint ambiguity uncertainties, where both the principal and agent exhibit Gilboa-Schmeidler's extreme ambiguity aversion. For this, we extend the martingale method well known in the agency literature, by allowing not only the mean...

The main result of this paper is that, in continuous time games with imperfect monitoring it is better to average information over time rather than respond at every instant. The two main reasons why it is better to introduce delayed response to signals are that it helps to (1) loosen...

Consider a setting in which a principal induces effort from an agent to reduce the arrival rate of a Poisson process of adverse events. The effort is costly to the agent, and unobservable to the principal, unless the principal is monitoring the agent. Monitoring ensures effort but is costly to...

This paper considers the eff ects of a two-period interaction on the decision of a principal to delegate authority to a potentially biased but better informed agent. Compared to the (repeated) one-period case, the agent's first period actions may also signal his type which in turn impacts wages...