War, What Is It Bad for? Family Fixed Effects Estimates from Prisoners and Post-War Suicides

Evan Roberts, University of Minnesota

This paper adds new evidence on the causal effects of wartime experiences, by examining the life expectancy impact of three different wartime contexts: New Zealand and American prisoners of war from World War I, and New Zealand veterans who died by suicide following service in World War I. Although New Zealand and American prisoners experienced the same category of event, the distribution of spells of imprisonment was quite different. Sampling from different nations provides greater variation in the length of captivity. Suicide decedents provide a distinctly different lens on these issues, as they are sampled from an event occurring after the war. For all three groups, i use linkage to vital registries, military personnel records, and genealogical databases to construct control groups of siblings, both veteran and non-veteran, of the prisoner; and non-related control cases matched on pre-war characteristics. The paper demonstrates how family controls can be used for robust causal inference in historical population studies. Substantively we find that the impact of being a POW on life expectancy was significant. Compared to either the general population, or the sibling and soldier controls men who experienced even short terms of wartime incarceration lived several years less than their peers. Suicide in New Zealand veterans generally occurred in middle age, and the life expectancy impacts of death by suicide were significant.

No extended abstract or paper available

 Presented in Session 14. Mortality: The Aftermath of Disease