Air Force Pilot Retention: Evaluating the Results of Alternative Models

Abstract

This study developed two alternative models of the relationship between Air Force pilot retention decisions and demographic, educational, economic, and Air Force institutional factors. A particular focus of the study was an evaluation of the effects of a bonus on pilot retention. A neural network model helped identify several variable nonlinearities and interactions that were incorporated into a logit model estimated using data from 1980 to 1988. The logit results showed over 80 percent (45 out of 56) of the explanatory variables (including relative military to civilian compensation, airline hiring, and Air Force pilot manning) to be significant (p ≤0.10). Results from the logit analysis revealed that Air Force pilot continuation behavior, since the 1989 introduction of the pilot retention bonus, is highly elastic, compared to the prior nine-year period (2.335 vs. 0.529). Elimination or reduction of the bonus would seriously affect future Air Force pilot force management.