There is substantial evidence on the influence of political outcomes on the business cycle and stock market. We further hypothesize that uncertainty about the outcome of a U.S. presidential election should be reflected in pre-election common stock returns. Prior research pools returns based on the party of the winning candidate, assuming that the outcome of the election is known a priori. We use candidate preference (i.e., polling) data to construct a measure of election uncertainty. We find that if the election does not have a candidate with a dominant lead, stock market volatility (risk) and average returns rise.