Estimate penalized synthetic control models
and perform hold-out validation to determine their
penalty parameter. This method is based on the work by
Abadie & L'Hour (2021) .
Penalized synthetic controls smoothly interpolate between
one-to-one matching and the synthetic control method.