New report by Eric McGhee and Mindy Romero. From the summary:
Automatic voter registration (AVR) laws take advantage of transactions at government agencies where applicant information can be captured and repurposed to register citizens to vote. Implementation varies, but the core idea is always to make eligible residents actively decline registration if they do not want it. The reform movement is still quite young, so there is little information about how effective AVR has been in the U.S. context. Oregon was the first to adopt a law generally agreed to be AVR, and of the 21 states that have adopted by the time of this writing, only 18 have actually implemented the reform and 12 of those did so by the 2018 General Election (see Figure 1). Thus, systematic evidence of the effects of these laws has been limited.
In this report we begin to address this gap by analyzing the effect of AVR on registration rates in four out of the 12 states that have implemented AVR thus far (Oregon, Colorado, California, and Delaware). We limit ourselves to registration rates because higher registration is arguably the most basic intended consequence of the AVR reform. Our data and methodological approach also permit us to examine key underrepresented groups that many election officials and voter advocacy organizations have hoped AVR would draw onto the registration rolls: Latinos, Asian Americans, and young people.
Our results suggest the consequences of AVR depend on the state and the type of outcome. The reforms appear to be very effective at making the DMV a primary method of registration. In many cases, the reform has reshaped the boom-and-bust pattern of registration in a typical election cycle by ensuring that registrants are steadily added throughout the year and voter records are kept updated. This will likely have the long term effect of a sizable increase in registration.
However, with one exception the AVR programs we look at are quite young, so this process of steady registration increase has not had much time to have an effect. Thus, the evidence for increased registration so far is ambiguous in the context of recent elections that have seen enthusiastic registration and turnout even without reform. More assertive approaches to AVR that have been adopted in Oregon and Alaska might be more successful at increasing registration rates. However, such systems come with important concerns about accidentally registering non-citizens; the challenges in any given state could be considerable.
In addition, we see no clear evidence that AVR has closed gaps in registration between historically overrepresented and historically underrepresented groups. The data for underrepresented groups sometimes show their strong use of AVR over other methods of registration, but without clearly demonstrating an improvement in their registration that was directly caused by reform. Neither do we see clear signs of a worsening position for these groups due to AVR; instead, the evidence is simply too ambiguous at this point to make a firm conclusion.
In short, AVR is promising but the conclusions we can make at this point are not firm. The COVID-19 pandemic has temporarily cut off access to many of the government agencies that serve as touch points for AVR. But when AVR access begins again, it will be important to continue monitoring it as it develops over time.
There’s a related webinar at noon pacific. Registration info here.