Note on the Uniqueness of the Maximum Likelihood Estimator for the Tobit Model