tobacco was rationed during wwii, so less for the general populace but it was also included as part of military service. sounds like it would be really difficult to measure the effects of limiting it for one subset, subsidizing it for another and then trying to determine the effect on both as a set.