My main purpose in this article is to investigate the military spending correlates, using international cross-sectional data for 1984. This perspective broadly mirrors that of A. Maizels and M. K. Nissanke, who investigated the determinants of military spending for a cross section of developing countries, and R. P. Smith, who examined a cross section of industrialized countries.' In contrast, my effort embraces both developed and developing countries as well as market and centrally planned economies but distinguishes countries on the more relevant basis of whether they are in conflict and whether they belong to alliances such as the North Atlantic Treaty Organization (NATO) and the Warsaw Pact. An advantage of cross sections such as this is that it makes it possible to distinguish broad patterns in defense spending across the world as a whole. A disadvantage is that one cannot infer dynamic behavior from cross sections in which case strategic theories of arms expenditure cannot be investigated satisfactorily.2 The same applies to dynamic versions of the neoclassical theory of military expenditure.3 However, international cross sections and national time series are more likely to be complements than substitutes in research. In the next section I propose a simple model of military expenditure that is investigated in Section IV by cross-sectional data for 137 countries. The model is based on the premise that military expenditure deters war and departs from previous theoretical efforts in that it applies state preference theory to determine the demand for military spending. In Section III I consider other variables that, in terms of the model, may also affect military spending via its shadow price. Since the sample consists of cross-sectional data I pay particular attention to various specification tests that are usually ignored. These suggest that the results are not independent of the method used for normalizing