American Politics Workshop with Anthony Fowler
DateFebruary 2, 2015
Time3:00pm
Location
4357 Bunche Hall
Contact
Contact Information
Belinda SunnuPhone bsunnu@polisci.ucla.edu
Presenter:Anthony Fowler, University of ChicagoTitle:”A Bayesian Explanation for Incumbency Advantage”Abstract: Incumbents perform significantly better in elections just because they are incumbents, yet the most commonly proposed explanations for this phenomenon are unsatisfying and inconsistent with empirical evidence. First, I review previous evidence and introduce new data to demonstrate that previous explanations are unlikely to account for the incumbency advantage, its changes over time, or its consistency across offices. Next, I introduce a new explanation that is parsimonious and consistent with existing empirical evidence. If voters lack perfect information about candidates in an election, incumbency is an informative signal of quality, and voters will update their beliefs accordingly — producing an incumbency advantage where low-quality incumbents who barely win office receive a significant electoral bump in support because of their incumbency status. These claims are formalized through a decision-theoretic model where voters receive noisy signals of candidate quality. Finally, I experimentally test this explanation by providing voters with information that removes the informative value of incumbency—their incumbent’s vote margin in the last election. When voters learn that their incumbent barely won office, they are significantly less likely to support reelection. The results suggest that this simple theory of Bayesian learning explains a meaningful portion of the observed incumbency advantage in American elections. I conclude by discussing the ways in which this explanation can explain the dramatic rise of incumbency advantage in the last half of the 20th century, variation in incumbency advantage across countries, and even a negative incumbency advantage in some contexts.
Event Details:
Parking | Directions
Please register here:
