Published online by Cambridge University Press: 25 January 2021
In this article, we analyze a unique set of state-level monthly survey data covering the eighteen months preceding the 2006 election to estimate (1) the relative effects of national and local conditions on the strength of challenges to incumbent senators and governors and (2) the effects of these challenges on incumbent popularity and, ultimately, vote shares. The analysis confirms several of the basic components of the theory that the strategic behavior of candidates and campaign contributors amplifies the effects of local and national conditions on election results, thereby enhancing electoral accountability. But it also uncovers a striking difference between the two offices. Even taking the strongly pro-Democratic national climate into account, the election context had a strong tendency to reduce the approval ratings of Senators, while it had an equally strong tendency to increase the approval ratings of governors. We speculate as to what might account for this difference.