Plans to reduce the number of military personnel by 25% have stimulated public interest in the social representativeness for the U.S. military and in the effects of voluntary military service on civilian labor market outcomes. The military is the largest single employer of young people and veterans benefits constitute the largest federal subsidy to education and training. The purpose of this project is to analyze the impact of military service on the earnings of veterans after leaving the military. What is at issue is whether or not military service, on average, has a positive or negative effect on success in the civilian labor market. Errors in the scoring of military entrance examinations in the late 1970's increased the probability that low ability applicants would be accepted. This exogenous shift in acceptance probabilities provides a natural experiment that can be used to estimate the effects of voluntary, military service on civilian labor-market outcomes. The effects of military entrance standards on the social composition of the armed forces will also be analyzed. The empirical strategy is to match armed forces administrative records on military applicants to military records on entrants, and finally to longitudinal social-security earnings records. The matched data set will consist of a long panel containing information on earnings, test scores, veteran status, and demographic characteristics. Semi-parametric estimators will be developed for estimating the effects of military service in nonlinear models with unobserved individual effects.
