The Impact of Procedural Information Costs on Turnout

• In Event: Voter Turnout

Abstract

Formal institutional rules determine who can vote and when, and extensive research has shown that these rules directly affect voter turnout. Yet, we know relatively little about the beliefs that citizens hold about these rules and how this affects their likelihood to participate in elections. In a context where media increasingly report about disinformation campaigns that specifically try to discourage certain groups of voters from casting a vote by disseminating incorrect or misleading information about electoral rules and procedures, a more systematic investigation of the issue seems warranted. While incorrect procedural beliefs are likely to increase the cost of voting for an individual and thus, potentially reduce voter turnout, they will affect different societal groups differentially. Especially low-information citizens with little or no prior exposure to elections are likely to be most responsive to incorrect procedural beliefs. In order to assess the effect of procedural information costs, I leverage a natural experiment from the U.S. context. Using a large, original dataset of state voter file data from 18 U.S. states with a regression discontinuity design, I compare voter turnout among young people – arguably a group including many low-information citizens – whose 18th birthday falls shortly before or after the registration deadline in their respective state. While those slightly older face no additional informational hurdles, those slightly younger with a birthday after the registration deadline, but before Election Day face higher procedural information costs due to uncertainty about whether one has to be 18 by the registration deadline or by Election Day to vote. As expected, I find a clear drop in turnout right at the registration deadline, showing that turnout for those who turn 18 after the registration deadline is up to 9 percentage points (or 25%) lower than for those who are slightly older. This effect persists over time, with young people who do not vote in the 2008 presidential election 37 percentage points less likely to vote 8 years later in 2016. I also show that alternative explanations such as formal rules that prevent 17-year-olds from registering or concurrent administrative acts that may make it easier for 18-year-olds to register when obtaining their driver’s license are unable to account for this substantial drop in voter turnout.

Author

• Alexander Held, University of British Columbia