We present a model of (re)elections in which an incumbency advantage arises because the incumbent can manipulate issue salience by choosing inefficient policies in the policy dimension in which he is the stronger candidate. The voters are uncertain about the state of the world and the incumbent's choice of policy. Under complete information they would reelect the incumbent if and only if the state is sufficiently high. Undesirable policy outcomes may be due to either a bad state or the incumbent's choice of inefficient policies. The incumbent uses inefficient policies in intermediate states, whereby he creates uncertainty about the true state in such a way that voters are better off in expectation reelecting him. Hence the equilibrium exhibits an incumbency advantage that stems from asymmetric information and the use of inefficient policies.