On political ambiguity and anti-median platforms

This paper constructs a new, information-based explanation for political ambiguity and the success of anti-median platforms. It argues that voters' and candidates' correlated preferences about the appropriate policy combined with ambiguous platforms can help candidates with non-median preferences increase their support and even win against a median candidate. I show how ambiguity can arise in a standard citizen-candidate setting where voters have different preferences, in its extension with primaries, and even in a Condorcet jury model where disagreement arises only from differences in voters' information. The paper also offers a formal framework that allows for dog whistle politics. The model illustrates how ambiguity can have important negative welfare implications. Specifically, I show that despite having ex-ante identical preferences with voters, politicians may choose ambiguous platforms even if voters would be keen on banning them.