A central challenge in estimating the causal effect of TV advertising on demand is isolating quasi-random variation in advertising. Political advertising, which topped $14 billion in expenditures in 2016, has been proposed as a plausible source of such variation and thus a candidate for an...