Speaker: Prof Timothy Christensen
Affiliation: New York University / University College London
Location: Room 309, Steele Building (#03), UQ St Lucia campus
Abstract
We consider the problem of learning treatment (or policy) rules that have welfare guarantees in target populations that are similar to, but possibly different from, the experimental population from which the data are sampled. This paper makes two main contributions. First, we provide a formal sense in which policies that maximize social welfare in the experimental population are robust to shifts in the distribution of potential outcomes (but not covariates) between the experimental and target populations. Hence, policy learning methods that have good regret guarantees in the experimental population, such as empirical welfare maximization, are externally valid with respect to shifts in potential outcomes. Second, we develop new methods for policy learning that are robust to shifts in the joint distribution of potential outcomes and covariates between the experimental and target populations. Our methods may be used with experimental or observational data (i.e., where treatment is possibly endogenous). Many of our methods can be implemented with linear programming.
About the presenter's meeting
Prof Timothy Christensen will be visiting the School of Economics from Tuesday 30th August to Friday 2nd September 2022. While here, he will be using Room 520A, Level 5, Colin Clark Building. If you would like to meet with Prof. Christensen please contact Dr Fu Ouyang who will be his host while at The University of Queensland.
About School Seminar Series
The School of Economics General Seminar Series is held on Fridays. These are in-person and presented by a range of guest researchers from around Australia and internationally.