Simulation Analysis Of Dynamic Discrete Choice Models

The main content of this paper is using a variety of Monte Carlo experiments to analyses two estimators which are maximum likelihood estimator and the conditional choice simulation estimator suggested in V. Joseph Hotz, Robert A. Miller, Seth Sanders and Jeffrey Smith(1994)[13]for the parameters of dynamic discrete choice models. We first introduce the dynamic discrete choice models and the conditional value function and we also list the steps simulation data generated. Then we introduce the two estimators which are maximum likelihood estimator and conditional choice simulation estimator for the models. At last, we establish the homologous models and simulate the data through the example of the replacement of GMC bus engines in Rust(1987)[6]and use the maximum likelihood estimator and the conditional choice simulation estimator to analyse the results.