ALTERNATIVE COLLECTIVE‐GOODS MODELS OF MILITARY ALLIANCES: THEORY AND EMPIRICS

Abstract

How should the defense activities of allies be aggregated to determine the alliance‐wide level of defense? Two alternative models – best shot and weakest link – are contrasted with simple summation of defense spending or manpower for aggregating allies' defense efforts. We extend the joint product model to include these methods of aggregation, and devise an empirical procedure to test between best‐shot and weakest ‐link models. We apply this test to four alliances: Triple Alliance (1880–1914), Triple Entente (1880–1914), Warsaw Pact (1963–1987), and NATO (1961–1987). The testing procedure can be applied to other collective choice situations.