What sorts of policy decisions do voters pay attention to, and why? And how does rational voter attention affect the behavior of politicians in office? We extend the Canes-Wrone, Herron and Shotts (2001) model of electoral accountability to allow the voter to choose when to pay costly attention to learn the consequences of the incumbent’s policy. We find that the voter will pay the most attention in close races, when the instrumental value of information about incumbent performance is greatest. When the voter’s cost of attention is ‘intermediate,” she will pay different levels of attention to different policies, and generally pay more attention to an unpopular policy than a popular one. This asymmetric attention can improve accountability by giving a moderately strong incumbent an opportunity to be “caught in the act of being good” after choosing an unpopular policy. However, it can also harm accountability by leading a moderately strong incumbent to categorically avoid the unpopular policy to evade voter scrutiny. This negative effect can be sufficiently strong to harm the voter’s overall welfare, providing a novel rationale for why subsidizing information about government performance may unintentionally harm voters. Finally, we find that rational voter attention can never induce “fake leadership,” i.e., a moderately weak incumbent choosing an unpopular policy precisely because it draws more scrutiny, hoping that the voter discovers an accidental policy success.

scoco@ceibs.edu