The Demographics of Pandora's Box: An Empirical Investigation of the Determinants of Who Dies in War

Gartner and Segura propose to explore the demographics of casualties in war. Their goal is to understand better the system of relationships and choices at the individual, social group and institutional levels that produce the observed distributions of battle casualties across groups in the society. They plan to examine whether and why casualty representation by group, as well as the larger military manpower pool, varies on a variety of dimensions -- region, socio-economic status, religious preference, race, and ethnicity -- within and among conflicts. Their previous research shows that variations in casualties influence a variety of domestic political process, including individual, mass, and elite opinion formation and elections.

Gartner and Segura theorize that: 1) an understanding of who dies in war requires an understanding of who fights. 2) Initially, those who fight and die emerge from a peactime personnel process that is endogenous to the lessons of the previous conflict, the duration of the peace, economic variables and other factors. 3) Casualties provide information that leads to a wartime process of personnel selection. 4) In each situation, there are individual and institutional factors that affect the demographics of those at risk in war. 5) The relative influence of the peactime and wartime processes is a function of the duraciton and intensity of the conflict and the preceding peace.

Gartner and Segura focus empirically on post-W.W.II American military manpower, and casualties from the Korean, Vietnam and Gulf wars. They use three distinct types of data. First, they use data from the National Archives and Records Administration, including the Korean Conflict Casualties file and the South East Asia Casualties file. Second, data on unemployment, population, and other socio-economic variables are drawn from the Millennial Edition of the Historical Statistics of the United States. Finally, new data, including a continuous measure of civilian-military pay differentials and serial data on inductions, aggregated by race, religion, region and ethnicity are collected from archival sources available from the Department of Defense and state draft boards.

Three sets of hypotheses are tested. First, testable arguments are offered about the demographics of manpower across the entire post-W.W.II era, drawing on general concepts such as economic performance and institutional aspects of conscription. Then, they examine the role of casualties, both as a dependent variable and, when lagged, as an independent variable. Finally, they address specific issues about African-Americans and Latinos in Vietnam and Korea.

For each group of interest (minority, religion), Garner and Saguaro employ a cross-sectional time series analysis to determine, first, if there is variation between the group's share of the total population, military manpower pool, and wartime casualties. Second, they use multivariate analysis to determine which factors best explain the deviation between these representations. Finally, they attempt to identify the determinants of variation in these deviations across time and space. For example, they examine African-American representation in society, the military, and among causalities nationally and across states, and in each case across time.

The normative and policy implications of the study are considerable. A principal tenet of democratic thought is that the benefits and costs of membership in the body politic are visited equally on the citizenry. If one or more social groups bear a disproportionate share of these costs, and this inequity is endemic in the institutional arrangements of society, then there are clear implications for the legitimacy of the system and it ability to use force effectively and in a manner consistent with the preferences of its population.