The Effect of Income on Turnout: Evidence from Public Salary Records and Administrative Voting Data

How does income affect voting behavior? Existing work is clear: there is a large turnout gap between high- and low-income voters. A widely-held assumption derived from this between-person relationship is that changes in any given person’s income should cause a change in that same person’s political participation. Recent work has interrogated this assumption using within-person economic windfalls, however, and found no subsequent effects on turnout. Using public records requests for data on the salaries of public employees and other administrative data on over 570,000 people, I examine whether enduring changes to annual income between 2014 and 2020 affect political participation. The public employee data span everyone from custodial workers to college professors and cover a total of over 13,000 different occupations. I link the salary records to administrative data on voter registration and turnout to track within-person changes in income and participation over time. Results suggest that for a given person, changes in annual income correspond to extremely small changes in turnout. The effect is at most a fifth of the size of the gap illustrated in aggregate, cross-sectional designs between similar income levels. Data on salary trajectories within occupations suggests turnout patterns may be more a function of selection than individual economic incentives. These findings have implications for our understanding of what drives political participation as well as the types of policy interventions that might affect it.