Learning and Decision Making

A common theme is choice without priors, mostly based on the the solution concept minimax regret.

Publications:

Choice under Complete Uncertainty when Outcome Spaces are State-Dependent (joint with Clemens Puppe), Theory and Decision (2007, forthcoming, available online)

Working papers:

Decision Making in Uncertain and Changing Environments (June, 2009) joint with Andriy Zapechelnyuk, Department of Economics and Business Working Paper 1160, Universitat Pompeu Fabra

Minimax Regret and Strategic Uncertainty (April, 2008) joint with Ludovic Renou, Department of Economics and Business Working Paper 1087, Universitat Pompeu Fabra

Decision Makers Facing Uncertainty: Theory versus Evidence (June 11, 2007) joint with Paolo Giordani and Sanne Zwart.

Why Minimax Regret - a brief nontechnical overview (October 23, 2006)

Distribution-Free Learning (December 30, 2006) WP ECO 2007/1, European University Institute

Minimax Regret with Reduced Variance - The Correlated Binomial Average Rule (September, 2006)

Minimax Regret in Practice - Four Examples on Treatment Choice (joint work with Patrick Eozenou and Javier Rivas) (June 2006). Here is the matlab program associated with this paper and the instructions on how to use it.

Eleven - Designing Randomized Experiments under Minimax Regret (this version March 12, 2007, earlier version as  WP ECO 2006/2) My paper cites the 15 November 2005 version of a paper by Joerg Stoye Minimax Regret Treatment Choice with Finite Samples which we provide a link to in order to clarify priority.

How to Minimize Maximum Regret in Repeated Decision-Making , Mimeo, 2003.

Updating Strategies Through Observed Play – Optimization Under Bounded Rationality (joint with Ross Cressman), Discussion Paper No. B–432, Sonderforschungsbereich 303, University of Bonn, 1998.

Bounded Perception and Learning How to Decide, Discussion Paper No. 972, The Center for Mathematical Studies in Economics and Management Science, Northwestern University, 1991.