Elite signaling and efficient markets : the evaluation of endorsements in presidential primaries
Abstract
Endorsements are an integral part of American political campaigns, but despite their ubiquity, there exists only sparse literature evaluating either the impact of endorsements on electoral outcomes or the circumstances in which endorsements are offered. I hypothesized that the primary factor influencing a politician’s decision to endorse was a perceived increase in a candidate’s chance of success; thus politicians are more likely to endorse candidates who have demonstrated a real potential for winning the election. Using the 2012 Republican Presidential nomination and the 2008 Democratic Presidential nomination as case studies, I regressed daily proportions of endorsements given on prediction market share prices from the Iowa Electronic Markets. I used an Almon distributed lag model to account for the delay between a politician observing an increase in a candidate’s chance of success and their subsequent endorsement. Though there was a weakly positive association between prediction market share prices and endorsements shares, I found little evidence to support the claim that politicians systematically offer endorsements in response to increases in a candidate’s probability of success.