Politicians appear to anticipate that the public will hold them accountable for war deaths. Yet, little is known about why some politicians openly oppose costly conflicts while others do not and the difference this makes to their electoral fortunes. Examining U.S. Senate elections from 1966-1972, we find that state-level casualties, military experience, and a variety of other factors affect candidate positions on the Vietnam War. Challenger and incumbent positions are negatively related, suggesting that strategic considerations play a role in wartime policy formation. We also find that war plays a role in elections. Incumbents from states that experience higher casualties receive a smaller percentage of the vote, an effect ameliorated when the incumbent opposes the war and his or her opponent does not. Wartime casualties, we conclude, influence both the perceived cost of the war and its salience, affecting both candidate positions and elections, suggesting that selectorate/electorate-type arguments about war and domestic politics can apply to the US system.
Nearly 30 years after the conclusion of the Vietnam War, it has become conventional wisdom that the U.S. population remains extremely sensitive to wartime casualties, and that this sensitivity serves as a check on the behavior of decisionmakers. That casualties matter to citizens and leaders alike appears obvious, but what is less clear is the process that connects the two. Specifically, politicians act as if they might be held accountable for war deaths, yet little is known of the process that makes this possible. What evidence is there that the public is willing or capable of holding decisionmakers accountable for the human costs of war?
Virtually all analyses of war and domestic politics focus on presidents, prime ministers, or states as unitary actors. On one hand, this makes sense, as foreign policy operates primarily at the national level. Yet there are a number of reasons to explore the effects of war at sub-national levels of aggregation. First, democratic leaders with sub-national constituencies frequently influence foreign policy, and some systems, such as the U.S., have a disaggregated policy process that includes politicians with sub-national constituencies-e.g., U.S. Senators. Second, arguments about the electoral accountability of leaders to policy outcomes are not theoretically bound to operate exclusively at the national level, and testing these arguments on a population of sub-national actors helps to determine better the penetration of international affairs into domestic politics. Finally, the domestic effects of international politics, and in particular wartime casualties, dramatically vary both geographically and temporally within a single conflict. An advantage of disaggregating the political processes of nation states to explore the effects of war on politics is that it allows us to capture this variation.
We argue that wartime deaths, which vary considerably across space and over time, represent a, if not the, most visible cost of a nation's involvement in war and serve to highlight both the relative importance of the conflict to its citizenry, as well as the successfulness of the effort. As casualties increase, ceteris paribus, voters are less satisfied with the status quo. As approval declines, incumbents seeking reelection are more likely to pay a price if they are perceived to be responsible for, or continue to support, the current war policy. Candidates openly opposed to current policy may reap electoral benefit, particularly as costs mount. During periods of international conflict, proximate casualties and candidates' positions on the conflict augment previously established domestic political variables in predicting electoral outcomes.
Casualties also influence the wartime positions of both incumbents and challengers. Challengers might choose to differentiate themselves by articulating opposition to a status quo policy, particularly one perceived as increasingly costly. …