Zachary Breig

Ph.D. Candidate – Department of Economics

Phone: (714) 588-1776


Research Summary

Research Statement

Job Market Paper

Delegation with Continuation Values

In many real-world principal-agent problems, contingent transfers are not possible. This paper shows how to optimally delegate actions to an informed agent when the principal cannot require payments. Instead, they are able to commit to actions in the future that affect both parties’ payoffs. I use these results to study a model of delegating decisions sequentially, and a model of delegation in which the principal can use measures that are costly to both parties. When choices are delegated sequentially, the principal takes advantage of the agent’s superior first period information by limiting the agent’s choices in the second period. When the principal can commit to mutually detrimental measures, I show a case in which the principal uses these measures to incentivize the agent not to take biased actions. I then give general conditions for when the principal does not use this option, and instead simply delegates an interval of choices to the agent.

Work in Progress

Prediction and Model Selection in Experiments

This paper compares the predictive performance of several models of risk aversion and time preferences in experimental settings. Models are evaluated on the basis of out of sample prediction rather than in sample fit. Some models predict behavior better than others, and these are often not the models which have the best in sample fit. For preferences over risk, with the exception of very small sample sizes, allowing the estimation procedure to select between constant relative risk aversion and constant absolute risk aversion improves prediction beyond that of a single model. Moreover adding a behavioral parameter such as loss aversion improves prediction further. This contrasts with time preferences, where adding the present bias parameter β worsens prediction for all sample sizes, despite improving fit significantly. The methodology is easy to implement and interpret, and results can be used by both experimentalists and applied modelers.

Repeated Contracting without Commitment

I study a dynamic model of monopoly sales in which consumers are finitely lived. I characterize the equilibrium when the monopolist can only commit to a contract in the current period, and compare it to the equilibrium when the monopolist can fully commit and when she can sign long term commitments with renegotiation. I then extend the model to a repeated setting, in which one long-term monopolist interacts with a series of short-term consumers. I show that the monopolist is able to attain the commitment outcome for the previously mentioned commitment types, and provide a folk theorem which gives the payoffs of a sufficiently patient long-term player. Finally, I show that commitment in the repeated game is not necessarily monotonic in exogenously provided commitment: for some parameter combinations, a monopolist who can only sign one period contracts can implement the full commitment outcome in equilibrium, while a monopolist that can sign long term contracts that are renegotiated cannot.

(Bed)Time inconsistency

with Matthew Gibson and Jeffrey Shrader

We are running a study to adjudicate between two alternative models that explain time inconsistent behavior: naive quasi-hyperbolic (or β-δ) preferences and optimistic expectations about time use. For either of these models, individuals can exhibit dynamically inconsistent choices over, for instance, the timing of real-effort tasks. In the case of hyperbolic preferences, this is because the costs associated with doing the task are discounted differentially if the person is deciding whether to do it in the near or far future. If the individual is optimistic about her schedule, she might believe that she will be able to do a task at a given time when that time is far in the future, but when the event arrives, she finds that she is busier than expected. We have designed and piloted an experiment that identifies which of these two models is causing time inconsistency for a given individual. The key difference we exploit is that when given information about their behavior, the hyperbolic discounter will have additional demand for costly commitment devices while the optimistic person will have lower demand. Pilot results support the optimism model as the dominant model of behavior in this case, and a full-scale study (funded and IRB approved) will be run in the spring of 2017.