Munich Personal RePEc Archive

Revisiting the Synthetic Control Estimator

Ferman, Bruno and Pinto, Cristine (2016): Revisiting the Synthetic Control Estimator.

WarningThere is a more recent version of this item available.

Download (956kB) | Preview


We analyze the conditions under which the Synthetic Control (SC) estimator is unbiased. We show that the SC estimator is generally biased if treatment assignment is correlated with unobserved confounders, even when the number of pre-treatment periods goes to infinity, and in settings where one should expect an almost perfect pre-treatment fit. While our results suggest that researchers should be more careful in interpreting the identification assumptions required for the SC method, we show that, with a slight modification, the SC method can substantially improve in terms of bias and variance relative to standard methods.

Available Versions of this Item

Logo of the University Library LMU Munich
MPRA is a RePEc service hosted by
the University Library LMU Munich in Germany.