32.2 Key Features of SCM

• SCM estimates asymptotically normal parameters in linear panel models when the pre-treatment period is sufficiently long, making it a natural alternative to the Difference-in-Differences model (Arkhangelsky and Hirshberg 2023).
• SCM is superior to Matching Methods, as it matches not only on pre-treatment covariates but also on pre-treatment outcomes. SCM differs from Matching Methods because:
  • Matching methods focus on covariates.
  • SCM constructs a synthetic unit by matching on pre-treatment outcomes.
• SCM can be implemented under a Bayesian framework (Bayesian Synthetic Control) to avoid restrictive priors (S. Kim, Lee, and Gupta 2020).

References

Arkhangelsky, Dmitry, and David Hirshberg. 2023. “Large-Sample Properties of the Synthetic Control Method Under Selection on Unobservables.” arXiv Preprint arXiv:2311.13575 2.

Kim, Sungjin, Clarence Lee, and Sachin Gupta. 2020. “Bayesian Synthetic Control Methods.” Journal of Marketing Research 57 (5): 831–52.