Military Career Outcome and Lifespan of 6 Classes of Annapolis and West Point graduates: causation and selection effects.

Among military officers higher rank associates with life expectancy. This may be causation: benefits of higher rank may cause life to last longer - or selection: robust health helps making it to top ranks. We investigate graduates of 1949, 1950, 1951 of the US Naval Academy (n=2206) and US Military Academy (n=1719), with 42%, 49%, 49% equally distributed survivors, focussing on men with 20+ years service, when men could retire with benefits.Variation in major intervening variables in this sample is minimal. Beyond the expected positive association between final rank and life span we find mortality differentials by rank peaking around age 75 then decreasing. This pattern supports selection hypothesis. Modelling unobserved heterogeneity by a frailty variable suggests that levelling off of differential mortality at higher ages is caused by differential loss rate by final rank. Trajectories to different final ranks and different lifespans start drifting apart early.