This paper adds to the literature on the relationship between military service and long-term real earnings. Based on a regression discontinuity design it compares the earnings of age cohorts containing British men who were required to undertake post-war National Service with later cohorts who were exempt. It also compares age cohorts containing men who were conscripted into military service during the first half of WWII and those with later spells of conscription. It argues that, in general, we should not expect large long-term real earnings differences between conscript and non-conscript cohorts since important elements of the former received military training and experience of direct value in the civilian jobs market. In the case of call-up during WWII there is even more reason to expect that there was no major disadvantages to those conscripted. This occurred largely because their pre-military job status was preserved due to the employment of substitute women workers who acted as a temporary employment buffer thereby protecting serving men's positions on the jobs hierarchy.