The Econometrics of Inequality Sensitive Treatment Effect Estimation

Decision makers often have a strong interest in understanding the distributional effects of different practices, but existing analyses of these experiments often either fail to adequately consider these issues or lack a clear economic framework for evaluating them. The research program of the EquiStat Lab seeks to bridge this gap by developing a statistical framework for analyzing distributional impacts in interventions, building on modern welfare economic theory. This will provide a more comprehensive and rigorous approach to evaluation that brings both efficiency and distributional concerns to the analysis.

This material is based upon work supported by the National Science Foundation under Grant # 2313969. Principal Investigator: Eduardo Zambrano.

The research program consist of three main research projects.^[1]

Project 1: Protected income and inequality aversion

The first component, written in collaboration with Marc Fleurbaey, involves developing a methodology for determining an evaluator's social preferences based on the tradeoffs between the well-being of different individuals that the evaluator considers acceptable. This component is now complete, and it is available at this link.

Project 2: Social preferences under uncertainty and ambiguity

The second component involves incorporating this social welfare information into a statistical estimation framework, using an egalitarian equivalent representation of the evaluator's preferences, a concept akin to the certainty equivalent representation in expected utility theory.

Project 3: Inequality sensitive optimal treatment assignment

The third component involves applying this framework to the analysis of optimal treatment rules in randomized controlled trials, using Bayesian, maximin, and minimax regret criteria, and testing the developed methodologies on several well-known trials that exhibit considerable treatment effect heterogeneity. Interventions that exhibit treatment effect heterogeneity can be difficult to evaluate and summarize in terms of welfare. The methods developed in this project provide specific, quantitative guidance for how to do this evaluation.

Stay tuned

Through this website we plan to share code that will assist the interested reader with the following tasks: (i) conducting simulations to assess the effectiveness of the statistical and econometric tools we create, (ii) estimating egalitarian equivalent treatment effects and determining optimal treatment rules using both experimental and observational data, and (iii) visualizing the core theoretical results achieved throughout our research.

Disclaimer