Estimating promotion probabilities of navy officers based on individual's attributes and other global effects

The basic objective of the CCMT Officers project is to develop an econometric model that captures the promotion trajectory for the individual officers, conditional on past behavior and performance as well as on past (and current) economic and policy conditions. The model is also conditional on staying in the Navy the minimal time for promotion. The estimated results are also used as a forecasting tool for the policy maker. The econometric framework used here is a two-step first order Markov model that accommodate for time dependent information, for the cohort information, for censoring problems with the data as well as incorporates macro economic and policy level information. In the first step the conditional probabilities of staying or leaving the Navy are estimated. Conditional on staying in the Navy enough time to be eligible for promotion, in the second step of the estimation, the promotion probability within the next year (conditional on all the independent variables) is estimated. The estimation model is an Information-Theoretic, Generalized Maximum Entropy model that is non-parametric in the distribution.