Voters seem to react more to relatively smaller scandals by high-quality officials compared to low-quality ones

Strategic Opposition Research. Benjamin Ogden & Alejandro Medina. Texas A&M University Working Paper, June 2020. https://static1.squarespace.com/static/5d51841f673aa70001587348/t/5eed1373f54506483b4d82db/1592595316471/MO_2_19_20.pdf

Abstract: We develop a model of strategic opposition research within a campaign. A candidate faces an opponent of unknown relative quality. After observing an unverifiable private signal (e.g., rumor of a scandal), the candidate chooses whether to undertake opposition research, attempting a costly search for verifiable bad news, and then whether to reveal what the research found to the voters. Increasing the ex-ante quality of an opponent deters opposition research, but also increases voter response to any given revelation in equilibrium because the voter knows the (unobserved) private signal was sufficient to launch research. This "Halo Effect" can explain both why voters seem to react more to relatively smaller scandals by high-quality officials compared to low-quality ones, and why even high-quality challengers may want to raise the cost of searching their backgrounds, despite their expected lack of scandal. This effect may be sufficiently strong that parties prefer lower expected quality candidates on average. These results also rationalize the mixed empirical literature showing that exogenously generated negative information about candidates (i.e., experiments) tend to show smaller effects on voter behavior than endogenously generated negative information over the course of campaigns (i.e., surveys).