The Case for Back-End Opt-Out Automatic Voter Registration

A sign marks the location of a polling place in Janesville, Wisconsin.

Download the PDF here.

Automatic voter registration (AVR) is taking the country by storm. In 2016, Oregon became the first state to implement AVR, and the program’s great success set off a wave of AVR across the country.1 According to the National Conference of State Legislatures, 17 states and the District of Columbia have now adopted some form of AVR, while bills to establish the policy have been introduced in several additional states this year.2 In March, the U.S. House of Representatives even passed a bill—the For the People Act—to establish AVR at the federal level.3 The question is no longer if lawmakers should adopt AVR, but how these programs should be implemented.

AVR policies adopted by states and proposed by Congress vary in scope and design; most notably is the distinction between front-end and back-end opt-out.4 In a front-end opt-out system, the potential voter registrant is given the opportunity to opt out of being automatically registered at the time of the agency transaction. For example, in applying for or renewing their driver’s license, potential voter registrants are informed of voting eligibility requirements and given the opportunity to decline registration—often by checking a box on a paper or electronic form—indicating their desire not to be registered to vote. Unless they decline, their information is electronically transferred to state election offices in order to be added to the voter rolls.5 In a back-end opt-out system, potential voter registrants are given the opportunity to opt out of registration after the agency transaction. Here, an individual’s information is automatically transferred by the AVR source agency to state election offices. State officials then use information already on hand to confirm potential registrants’ eligibility, after which time a mailer is sent to potential registrants informing them that they will be automatically registered to vote unless they return the mailer indicating their desire not to be registered.6

To be sure, implementing AVR in any form is better than maintaining the status quo. The traditional process for registering to vote is outdated, inefficient, and confusing. Regardless of its specific scope and design, AVR streamlines the voter registration process while ensuring that voter registration lists are accurate, up to date, and secure. Still, many experts agree that back-end opt-out AVR is the ideal policy model for reasons detailed below:

  • Back-end opt-out is more effective than front-end opt-out: Research has shown that back-end opt-out AVR models result in more voting-eligible people being added to the voter rolls, compared with front-end systems. For instance, in Oregon, which uses a back-end model, 94 percent of eligible individuals who interact with the state Department of Motor Vehicles (DMV) have been registered to vote through AVR.7 By comparison, in California, which relies on front-end opt-out, only about 60 percent of eligible people who interact with the DMV have been registered to vote through AVR. The success of back-end AVR can be attributed to its default nature, whereby eligible individuals are automatically registered to vote unless they take affirmative steps to decline registration. The extra effort required—namely, having to fill out and return a mailer indicating a desire not to be registered—makes it less likely that a voting-eligible person will opt out.
  • Back-end opt-out is more efficient than front-end opt-out: As previously described, in a back-end opt-out AVR system, a potential registrant’s information is automatically transferred to state election officials upon completion of an agency transaction without any effort on the part of the individual. All the real work—including confirmation of an individual’s eligibility—is done by state officials.8 Alternatively, front-end opt-out systems require potential registrants to take additional time during agency transactions to review eligibility requirements and either attest that they are eligible to vote or check a box if they prefer not to be registered.9 Moreover, although employees at source agencies are trained for carrying out AVR services, they are not necessarily experts on election-related issues and may therefore be ill-equipped to answer questions regarding voter registration, resulting in confusion or delay.
  • Back-end opt-out is more secure than front-end opt-out: AVR programs—back-end and front-end models alike—tend to operate out of government agencies that regularly collect citizenship information, which means that it is already highly unlikely for someone who is ineligible to vote to be automatically registered through AVR. Even so, in a front-end opt-out system, officials rely partly on individuals’ attestation that they are eligible to vote.10 Individuals with limited English proficiency may have difficulty interpreting eligibility requirements, and agency officials may be unable to sufficiently translate, leading to potential unintentional errors.11 Through a back-end AVR model, however, the burden falls largely on state officials—rather than the individual—to confirm eligibility using reliable information that the state already possesses.


AVR is a game-changing policy that has been shown to expand and diversify the electorate while facilitating voter participation.12 Although any AVR is better than no AVR at all, wherever possible, lawmakers should adopt the back-end opt-out model.

Danielle Root is the associate director of Voting Rights for Democracy and Government at the Center for American Progress.


  1. Rob Griffin and others, “Who Votes With Automatic Voter Registration?” (Washington: Center for American Progress, 2017), available at
  2. National Conference of State Legislatures, “Automatic Voter Registration,” available at (last accessed May 2019); Brennan Center for Justice, “Automatic Voter Registration Bills, 2015-Present,” available at (last accessed May 2019).
  3. For the People Act of 2019, H.R. 1, 116th Cong., 1st sess. (March 14, 2019), available at
  4. Brennan Center for Justice, “Policy Differences of Automatic Voter Registration,” available at (last accessed May 2019).
  5. National Conference of State Legislatures, “Automatic Voter Registration.”
  6. Liz Kennedy, Lew Daly, and Brenda Wright, “Automatic Voter Registration: Finding America’s Missing Voters” (New York: Demos, 2015), available at; Sean McElwee, Brian Schaffner, and Jesse Rhodes, “Automatic Voter Registration in Oregon” (New York: Demos, 2017), available at
  7. AVR NOW, “Truths About Automatic Voter Registration,” available at (last accessed May 2019).
  8. Henry T. Berger, “Automatic Voter Registration – Recommended policy for New York State,” October 16, 2018, available at
  9. Sean McElwee, “Re: AVR Impact on State Voter Registration” (New York: AVR NOW, 2019), available at
  10. Berger, “Automatic Voter Registration – Recommended policy for New York State.”
  11. Ibid.
  12. Griffin and others, “Who Votes With Automatic Voter Registration?”; Danielle Root and Liz Kennedy, “Increasing Voter Participation in America” (Washington: Center for American Progress, 2018), available at