We saw in this chapter that differences in domestic politics can profoundly alter the risks, costs, conduct, and aftermath of war. None of the central hypotheses examined here, nor the strong evidence in their favor, could be right if we focus on states rather than leaders and internal politics. We saw that the expectation of high audience costs influences which disputes are pursued and which never arise. We saw that audience costs also influence the likelihood that a regime, once engaged in a crisis, will escalate or back down. We saw that, contrary to standard views, weakness does not always invite aggression and that, indeed, sometimes the weak are particularly likely to be the initiators of conflict. We saw reasons that democratic governments, with their reliance on a large coalition of backers, are more selective about the wars they are prepared to fight and they try harder to win than their small-coalition, autocratic counterparts. We also found that demo¬crats are more likely to fight over policy issues and over regime change than are autocrats so that even the reasons for fighting differ because of variations in domestic political institutions.

Perhaps most surprisingly we found that domestic politics shape the will to win. Given a stark choice between winning a war by spending a lot on it or saving money but ensuring defeat, democrats are much more inclined to sink extra resources into the pursuit of victory than are autocrats. We looked at evidence that tests the logic behind these claims and we found them to be consistent with the historical record. Each of these observations (as well as the findings regarding the democratic peace reported in chapter 14) would be considered false under structural, unitary actor explanations of international affairs. This is because each of these observations involves looking inside the state, breaking the unitary actor assumption, and allowing an important part of the action in international affairs to arise from factors that differ from state to state and from regime type to regime type.

War is so terrible a feature of international relations that I cannot leave the subject with¬out emphasizing the importance to all of us of improving our understanding of this deadly phenomenon. The past few decades have seen real progress in our theoretical and empirical understanding of the logic behind warfare; most of that progress arises from shifting our focus to the incentives and constraints faced by national leaders and away from treating the state as a unitary actor. Much, of course, remains to be done.