Policy platforms, campaign spending and voter participation

We model electoral competition between two parties in a winner-take-all election. Parties choose strategically first their platforms and then their campaign spending under aggregate uncertainty about voters' preferences. We use the model to examine why campaign spending in the United States has increased at the same time that politics has become more polarized. We find that a popular explanation -- more accurate targeting of campaign spending -- is not consistent. While accurate targeting may lead to greater spending, it also leads to less polarization. We argue that a better explanation is that voters preferences have become more volatile from the point of view of parties at the moment of choosing policy positions. This both raises campaign spending and increases polarization. It is also consistent with the observation that voters have become less committed to the two parties.