SMS: Bayesian Estimation of the Multinomial Logit Model: A Comment on Holmes and Held, "Bayesian Auxiliary Variable Models for Binary and Multinomial Regression" Response