Many studies of the military retirement system are based on models whose structures are likely to be changed by the policy interventions that they analyze. Such models could lead to seriously biased predictions of the retention effects of alternative retirement systems. This report examines the adequacy of the existing retention models for retirement policy analysis, quantifies their limitations, suggests improvements, and develops a simulation methodology to test the suggested and future improvements. It also examines the importance of paying analytical attention to the inputs of the retention models.

Author: R. Yılmaz Argüden