Attack Politics

This paper is concerned with the dynamics of competition in American political campaigns, in particular with how the type of message a candidate sends is affected by the stage of the campaign, the status of the candidate, and the competitiveness of the race. The paper examines the allocation of political campaign resources between positive and negative advertising in television spots. Political strategists and analysts consider that negative advertising is often a riskier strategy: while negative ads tend to be very effective, these advertisements may also backfire against the sponsor of the advertisement. This insight yields a testable prediction: front-runner candidates should focus on positive advertisements while the underdog should run a more negative campaign. Using a unique dataset on television advertising during the 2000 elections, three pieces of preliminary evidence support this hypothesis. First, incumbents, who are often frontrunners, in Congressional races tend to run more positive campaigns, relative to challengers. Second, incumbent campaigns, relative to challenger campaigns, tend to be more positive as the degree of incumbent safety increases. Finally, using daily data on advertisements during the two months preceding the election, the degree of negativity in Bush’s campaign, relative to Gore’s campaign, was negatively correlated with Bush’s lead in daily tracking polls. These results are also relevant in the context of the classic tournament model studied in the labor contracts and incentive theory literature. The reason is that, while testing the implications of tournament theory with one (positive) action has proven difficult in the literature, the empirical study of any incentive structures with more than one action, in particular with negative actions (where an agent may work to decrease the performance of his competitor), is non-existent.