How and When to Use the Political Cycle to Identify Advertising Effects
54 Pages Posted: 20 Jun 2019 Last revised: 7 Aug 2019
Date Written: August 5, 2019
A central challenge in estimating the causal effect of TV advertising on demand is finding quasi-random variation in advertising. Political advertising in the US has been proposed as a plausible instrumental variable because political spending on television has skyrocketed in recent elections, topping $4 billion in 2016. We take a multi-category approach to assessing how and where this instrument is valid and useful. We characterize the conditions under which political cycles theoretically identify the causal effect of TV advertising on demand, highlight potential threats to the exclusion restriction and monotonicity condition, and suggest a specification to address the most serious concerns. To characterize "where'' the approach might be most useful, we test the strength of the first stage of the political ad instrument, using both a single instrument and also optimal instruments obtained by machine learning. For the majority of commercial advertising categories, our findings suggest that researchers should consider using weak-instrument robust inference, as first-stage F-statistics are less than 10 for at least 202 of 274 product categories. Political advertising has the largest first-stage F-statistic for categories that typically advertise exclusively locally, such as automobile dealerships and restaurants. Failure to use the suggested specification leads to results that suggest violations of exclusion and monotonicity in a significant number of categories. Finally, we conduct a case study of the auto industry. Despite a very strong first stage, the IV estimate for this category is imprecise and includes zero.
Keywords: Advertising, Advertising Effectiveness, Political Advertising, Causal Effects, Instrumental Variables
JEL Classification: L10, L11, M31, M37, C26, C23, C81
Suggested Citation: Suggested Citation