Shrinkage Estimation in the Random Parameters Logit Model
We explore the properties of a Stein-like shrinkage estimator that combines the fully correlated and uncorrelated Random Parameters Logit model(RPLM). Monte Carlo experiments show that shrinkage and pretest estimators can improve upon the fully correlated RPLM estimator.