A Maximum Likelihood Estimator based on First Differences for a Panel Data Tobit Model with Individual Specific Effects

This paper proposes an alternative estimation procedure for a panel data Tobit model with individual specific effects based on taking first differences of the equation of interest. This helps to alleviate the sensitivity of the estimates to a specific parameterization of the individual specific effects and some Monte Carlo evidence is provided in support of this. To allow for arbitrary serial correlation estimation takes place in two steps: Maximum Likelihood is applied to each pair of consecutive periods and then a Minimum Distance estimator is employed.