A Single Table Buried In One Student’s 1976 Dissertation Makes A Powerful Case for Automatic Voter Registration

In the last few months, there have been several developments that have made widespread automatic voter registration (AVR) in the United States a real possibility. Liz Kennedy, my colleague at Demos, argues that the core feature of AVR is to shift the burden of registration from individuals to the government. Because registration is a key impediment to voting, such a shift has the potential to dramatically shift American politics.

Some have questioned whether an automatic system of registration would dramatically change turnout. However, the idea that individuals (not the government) are responsible for registration is an idea that took hold in the 19th century, largely to depress turnout among immigrants and workers.

The Rise of Registration. From Richard John Carlson, “The Effects of Registration System On Presidential Election Turnout In Non-Southern States: 1912–1924.” (1976)

When registration requirements were taking hold, regimes for registration were in flux. In most early systems, counties had registration systems resembling automatic voter registration, in that town and county officials, not voters, were responsible for maintaining the registration lists.

Interestingly, this creates a natural experiment that was the subject of a 1976 Ph.D. political science dissertation by Richard John Carlson. He explores differences in turnout between 1912 and 1924 across counties with no registration, non-personal registration (a proto-AVR system) and personal registration (the dominant system today). The results, show below, offer an important lesson: turnout is indeed dramatically lower when comparing counties with registration to those that do not require registration. But, more interestingly for today, the counties with non-personal registration have higher turnout than those with personal registration.

Proto-AVR counties have higher turnout than personal registration counties. From Richard John Carlson, “The Effects of Registration System On Presidential Election Turnout In Non-Southern States: 1912–1924.” (1976)

This gap between proto-AVR counties and personal registration counties averages 4.8 points in the four elections studied. Carlson also finds large gaps within states between proto-AVR counties and personal registration counties.

Within states, proto-AVR counties have higher turnout. From Richard John Carlson, “The Effects of Registration System On Presidential Election Turnout In Non-Southern States: 1912–1924.” (1976)

Combined with the other evidence that automatic voter registration bolsters turnout and the powerful effect higher turnout would have on policy in this country, this makes a strong case for AVR being a progressive priority.