Editor's Choice: Measuring Candidate Quality using Local Newspaper Endorsements

I construct a measure of candidate quality differentials using local newspaper endorsements. I argue that political endorsements made by newspapers can be used as expert opinions that reflect both a newspaper's partisan bias as well as quality differences between the candidates in an election. Using a dataset of 21,094 local newspaper endorsements, I simultaneously estimate the quality differences between candidates in 6,432 elections, along with a dynamic measure of the partisan bias of 368 local newspapers. After validating the endorsement-based measures of quality and bias, I use the candidate quality differential measure to reexamine the relationship between candidate quality, incumbency, and election results. I conclude with a discussion of how the new endorsement-based measures have a strong potential to help advance our understanding of the importance of candidate quality in electoral politics and governance in the United States.