THE REGIONAL AND GLOBAL LEVEL OF AGGREGATION

In a regional level analysis, O’Loughlin and Anselin (1991) indicate that African states interact primarily with their immediate neigh­bors, and that either conflictual or cooperative behavior “beyond that (contiguous) level is rare and insignificant for the African system as a whole.”

On a more global level, Levy (1984) explores a possible linear association between the number of great powers (system size) and war for the extended temporal span of 1495-1975.

Gleditsch and Hegre (1997) show that the probability of war in a politically mixed dyad must be higher than the probability of war between two non-democracies, and the relationship between democracy and war at the system level must be parabolic. Thus, increasing democratization initially produces more war, and the reduction of war takes effect only at a higher level of democratization. Crescenzi and Enterline (1999) measure the statistical relationships between democracy, democratization, and war, conducting their analysis on both the regional and global levels.

There is a theoretical school emanating from the Correlates of War project that proposes to examine the impact of “structural clarity” of the system's hierarchies and coalitions and its possible impact on the war- proneness of the international system. Clarity would be a function of the similarities and isomorphism of interstate coalitions resting on trade, alliances, and diplomatic bonds, and the extent to which these configurations are highly similar. Similarly, when we look at hierarchical patterns, high clarity would rest on the extent to which pecking orders in terms of military and industrial capabilities, com­mercial dominance, and diplomatic prestige should show the same approximate patterns. Despite the plausibility of the argument, almost no systematic empirical results have emerged.