Comparative Politics and the Synthetic Control Method
Causal Question / Estimand
The effect of a discrete, large-scale intervention on a single aggregate unit — here the effect of the 1990 German reunification on West German per-capita GDP — as a template for quantitative comparative case studies.
Identification Strategy
Same synthetic-control logic as AbadieEtAl2010-SyntheticControlMethods: a convex combination of OECD donor countries reproduces West Germany’s pre-1990 GDP path and predictors, and the post-reunification gap is the estimated effect. The paper’s distinctive contribution is conceptual: it shows the conventional regression comparison is also a weighting estimator with weights summing to one, but those weights are unbounded and extrapolate — making SCM’s non-negativity restriction the key safeguard. Adds a cross-validation procedure for choosing predictor weights ().
Key Assumptions
Convex-Hull-Restriction, long close pre-treatment fit, No-Anticipation, SUTVA (no spillovers to donor countries, no confounding contemporaneous shocks), unaffected donor pool.
Threats to Validity
Donor pool contaminated by units exposed to similar shocks; treated unit outside the donor convex hull (interpolation bias); over-fitting. Inference via in-space and in-time placebo tests.
Setting / Data
1990 German reunification; West German vs. OECD-country per-capita GDP, 1960–2003.
Key Claims
- SCM makes the choice of comparison units explicit, systematic, and based on observed pre-treatment similarity rather than the researcher’s judgment.
- Unlike regression, SCM cannot extrapolate outside the donor support; the regression comparison hides extrapolating weights.
- Reunification lowered West German per-capita GDP relative to the synthetic counterfactual.
Connections
- Builds on: AbadieEtAl2010-SyntheticControlMethods
- Synthesized by: Abadie2021-UsingSyntheticControls
- Contrast with regression under Ignorability (allows extrapolation) and DiD Parallel-Trends. See also SCM
Citation
Abadie, A., Diamond, A., & Hainmueller, J. (2015). Comparative Politics and the Synthetic Control Method. American Journal of Political Science, 59(2), 495–510.