32.3 Advantages of SCM

32.3.1 Compared to DiD

  1. Maximizes similarity between control and treated units (including unobservables).
  2. Useful when no untreated unit closely matches the treated unit.
  3. Objective selection of control units, reducing researcher bias.

32.3.2 Compared to Linear Regression

  1. Avoids extrapolation (no regression weights outside of [0,1]).
  2. Provides transparent weights, explicitly showing control unit contributions.
  3. Does not require post-treatment outcomes of the control group (reducing risk of p-hacking).

32.3.3 Additional Advantages

  1. Selection criteria provide insights into the relative importance of each donor unit.
  2. Prevents overfitting, since post-intervention outcomes are not used when constructing the synthetic control.
  3. Enhances interpretability, since the synthetic unit is constructed using observable pre-treatment characteristics.