Synthetic Controls for Experimental Design
Abstract
Abstract This paper studies how to design experiments when the experimental units are large aggregate entities (e.g., markets), and only one or a small number of units can be treated. We propose a class of non-randomized experimental designs using the synthetic control method, which jointly selects treated units and untreated units to serve as a comparison group. We analyze the properties of estimators under these synthetic control designs and develop new inferential techniques. We show that, in experiments with aggregate units, synthetic control designs can substantially reduce estimation bias relative to randomized experiments.
Fields of Study & Subjects
Advanced Causal Inference Techniques
Economic and Environmental Valuation
Publication details
- Journal / source
- The Review of Economics and Statistics
- Publisher
- MIT Press
- Pages
- 1–39
- Publication date
- 21 July 2026
- Language
- EN