Voter choice and issue salience: environmental preferences and the 2016 Presidential election

A large body of evidence suggests that identity-derived political affiliation is increasingly driving environmental preferences. We consider a variation on the reverse question: under what conditions might issue preferences change voters’ party choice? Academic literature predicts that voters are most likely to change their party affiliation when: (1) a party’s platform is distinct and transparent; and (2) the issue is important and personally salient. We argue that the explicitly anti-environmental campaign message of Donald Trump in the 2016 U.S. Presidential election fulfills the first condition; for the second condition, we exploit the plausibly exogenous spatial variation in EPA Superfund sites to generate a source of exogenous variation in the personal saliency of environmental issues. Our empirical analysis, conducted at both the individual- and countylevel, presents evidence on the relationship between Superfund and environmental preferences, establishes a robust causal link from Superfund to voter behaviour, and finally explores the possibility of heterogeneous effects via differing issue salience by age and/or income cohort. We find robust evidence that the presence of a nearby Superfund site did indeed reduce the number of votes for Trump. Specifically, our results imply that almost 490,000 voters that would have otherwise voted for Trump changed their Party vote choice based on their Superfund-induced environmental preferences in the 2016 election. Furthermore, we find evidence of hetergeneous effects of Superfund on voting behavior associated with household income, but not with voter age cohort. In particular, we find that the effect of Superfund on support for Trump grows as household income increases from well below the poverty line to moderately low levels of around $30,000-$40,000, and then tapers off or declines. Significantly, these moderately low-income voters are precisely the income group that are also most likely to vote for Trump, suggesting that educational campaigns aimed at lower-income households to increase personal saliency of environmental issues could potentially have disproportionately large political effects.