Optimal Discrete Decisions when Payoffs are Partially Identified

• Econometrics

Date and Location

Monday, November 6, 2023, 3:40 PM - 5:00 PM
Blue Room, 1113 Social Sciences and Humanities

Abstract

We derive optimal statistical decision rules for discrete choice problems when the decision
maker is unable to discriminate among a set of payo  distributions. In this problem, the
decision maker must confront both model uncertainty (about the identity of the true
payo  distribution) and statistical uncertainty (the set of payo  distributions must be
estimated). We derive e cient-robust decision rules which minimize maximum risk or
regret over the set of payo  distributions and which use the data to learn e ciently
about features of the set of payo  distributions germane to the choice problem.We discuss
implementation of these decision rules via the bootstrap and Bayesian methods, for both
parametric and semiparametric models. Using a limits of experiments framework, we
show that e cient-robust decision rules are optimal and can dominate seemingly natural
alternatives. We present applications to treatment assignment using observational data
and optimal pricing in environments with rich unobserved heterogeneity.

Subscribe to Upcoming Seminars

Click here to receive weekly notice of all upcoming seminars via email.

Individual seminar and workshop announcements are circulated via separate mailing lists. To subscribe, send an email to the relevant list:

• Click here to subscribe to the Departmental Seminars.
• Click here to subscribe to the Agricultural Economics Workshop.
• Click here to subscribe to the Development Workshop.
• Click here to subscribe to the Econometrics Seminars.
• Click here to subscribe to the Environmental and Resource Economics Workshop and the Environmental and Energy Economics Workshop.