Candidate Quality, Incumbency, and Election Outcomes in the United States

In today's politically polarized era, how much does candidate quality matter in elections? Spatial models predict that valence factors, such as candidate quality, matter less to voters as differences between the parties increase. In this paper I examine the link between candidate quality and incumbency effects, and how the importance of each has changed over time. I estimate that candidate quality explains about one-third of incumbency effects, and that incumbency status explains only about one-fourth of candidate quality effects. I show that while incumbency effects peaked in the 1980s and slowly declined since, candidate quality effects gradually increased from the 1950s until 2010 despite significant polarization, but then experienced a sharp decline over the past decade. I also show that the decrease in competitive elections over time -- and particularly after 2010 -- has reduced the share of elections where candidate quality effects can plausibly alter an election's outcome.