Between Withdrawal and Engagement: Disentangling the Effects of COVID-19 on Turnout

This project marks the first test of the household-level turnout effects of living with someone who died from COVID-19. Drawing on policy feedback and threat literatures, I argue that the theoretical effects of close COVID contact are unclear. I link death records with the registered voter file to identity Americans who lived with a decedent, leveraging a triple-differences design to distinguish the effect of household COVID deaths on turnout from non-COVID ones in Minnesota, North Carolina, and Washington State. Household COVID deaths reduced turnout in all three states, though the effect was different (and larger) than non-COVID household deaths only in Washington. I thus conclude that while COVID deaths clearly lowered household turnout, they operated largely through the opportunity costs associated with household deaths more generally rather than a distinctly political response to the government’s handling of the pandemic.