Swiftboating: Misleading Advertising in Presidential Elections

The term "swiftboating" arose out of advertisements aired in the 2004 presidential campaign and has now come to refer to any untrue political advertising. I develop an equilibrium model of voting behavior and candidate advertising in a presidential election with misleading and non-misleading advertising. I estimate the model using a unique dataset I created by systematically quantifying misleading statements in political advertisement videos in the 2008 election using FactCheck.org and PolitiFact.com. A candidate chooses misleading and non-misleading advertising in order to maximize his expected electoral votes in exogenously-determined battleground states.

I find that the two candidates spent over $38 million on misleading advertising. At the levels of advertising observed in 2008, I find that the gross effect of misleading advertising is, on average across television markets, more than twice as large as the gross effect of non-misleading advertising in terms of increasing vote share. I calculate that the shadow price of one electoral vote to be $9.7 million for McCain and $15.2 million for Obama. In a counterfactual analysis where candidates cannot air misleading advertising, three states (Missouri, Indiana, and North Carolina) have different electoral outcomes, and the equilibrium level of advertising falls by 21%, with the decrease in non-misleading advertising accounting for 26% of that change. In a second counterfactual, I repeal the laws against misleading advertising in the six battleground states that had them in place in 2008, and I find that misleading advertising in those states increases by 50%, and this change leads to a different electoral winner in the state of North Carolina.