War comes with terrible costs both in terms of money and lives. Do voters punish incumbents for these costs? Much of the existing literature on the effects of war deaths on public opinion toward incumbents and their war efforts suggests that the answer is yes. We test this proposition on data from a non-US case: Canada's war in Afghanistan. We estimate models of the effect of local war deaths on incumbent support using individual-level panel data from the 2006, 2008 and 2011 Canadian Election Studies and aggregate district-level data from the 2008 and 2011 general elections. In none of our models do we find support for the conclusion that war deaths decrease support for candidates of the governing party. Instead, we find evidence at both the individual and district levels that support for Conservative party candidates is higher in districts that experienced war deaths.
