Eric McGhee, PPIC research fellow, testified before the Little Hoover Commission in Sacramento today (May 25, 2017). Here are his prepared remarks.
Good morning Chairman Nava, Vice Chairman Varner, and distinguished committee members. My name is Eric McGhee, and I am a research fellow at the Public Policy Institute of California, where I study voter turnout and electoral and political reform more generally.
California’s voter turnout in 2014 hit record lows in both the primary and the general elections. Turnout recovered some in 2016, but languishing participation remains a concern moving forward. I was invited to your panel to help put California’s turnout in broader context and to give some thoughts on potential solutions. In doing so, I will address three questions:
- Who votes, and who doesn’t?
- What does California’s turnout look like over time and compared to the rest of the country?
- What should we expect from recent reforms, and what else should be tried to improve turnout?
Question #1: Who votes and who doesn’t?
California’s registered voters do not look like the population of all adults in the state. They are older, whiter, better educated, and a little wealthier; they are less mobile, more rooted in their communities, and more likely to own their own home; and they are more likely to identify with one of the two major parties and less likely to identify as independents.
Given California’s large immigrant population, citizenship is also an important factor for voting here. A substantial portion of the Latino and Asian-American populations are not citizens, and many are undocumented and so ineligible to ever become citizens. Historically that has meant significantly lower participation rates among these communities.
Yet the role of citizenship in California’s turnout is changing rapidly. Most of the growth in both the Latino and Asian-American communities now comes from children who are citizens because they are born in the United States. That means a growing share of these communities is eligible to vote, which in turn diversifies the state’s voting-eligible population.
These changes are happening faster here than in other states. Figure 1 shows the share of California’s Latino and Asian-American populations that are eligible to vote, compared with the same eligibility rates in other states. In the early 1990s, California Latinos were less likely to be eligible than were Latinos in other states. Since then, California Latino eligibility has steadily increased and now slightly exceeds Latino eligibility rates elsewhere. Roughly the same is true for Asian-Americans.
In addition to these demographic factors, PPIC research by Mark Baldassare has shown that voters differ from nonvoters in their opinions on policy issues (Baldassare 2016). Likely voters are about evenly split in their approval of a larger, more activist government that spends more money. Nonvoters, by contrast, are clearly in favor of government involvement across a number of issues. For example, 7 in 10 nonvoters want more government action on income inequality; just half of likely voters feel the same.
Question #2: What does California’s turnout look like over time and compared to the rest of the country?
California’s turnout has dropped about 20 points in the last 30 years. To understand the causes and implications of this trend, it is important to break the problem into two separate parts: 1) registration among those who are eligible; and 2) turnout among those who are registered.
Figure 2 shows that California has a registration problem compared with other states. The state’s flat registration rate—always holding steady between about 70 and 80 percent of total eligible residents—masks a relative decline that started in the late 1990s. California registered at higher rates than the rest of the country in the 1990s, but by 2016 had fallen about five percentage points behind.
In addition to the relative registration problem, turnout in California’s midterm elections—when the state votes for governor and other statewide offices—has been falling, while turnout in presidential elections has largely remained flat (Figure 3). This has created a widening divide between the two types of elections. In contrast to registration, this problem is not unique to California. Other states have experienced similar midterm turnout declines. But that does not excuse the problem so much as tell us that it is part of a larger national pattern.
A deeper analysis of these trends suggests different explanations for each one. The first trend—the decline in the relative registration rate—can mostly be explained by the growing Latino and Asian-American populations. No other combination of demographic characteristics, nor the state’s declining competition in statewide elections, comes close to explaining as much as this single change. Latinos and Asian Americans register at the same rates in California and other states. But in any given election, they register at lower rates than non-Hispanic whites or African Americans. Thus, as they become a larger share of the eligible voter population, their lower registration patterns pull down the overall registration rate more in California than in states that are not diversifying as quickly.
In contrast to the registration trend, the decline in midterm turnout is largely a function of the changing behavior of young people. Young people continue to vote in presidential elections, but they are increasingly likely to skip midterms. California’s expanding Latino and Asian-American populations play almost no role here: once registered, these groups have been voting at consistent rates over time. And unlike registration, it is the changing turnout rate of young people that has had the largest effect, not any change in their share of the registered population.
Question #3: What should we expect from recent reforms, and what else should be tried to improve turnout?
California has recently passed a wide range of reforms meant to increase voter turnout. Most of these are meant to ease the registration process, while at least one is meant to make it easier to vote.
The registration changes consist of four main reforms. First, California has allowed 16- and 17-year-olds to “pre-register” with a placeholder record that will only become official when they turn 18. Second, California has established an entirely electronic online registration system that makes it easy to find out how to register and to fill out an application, saving even the trouble of finding a stamp. Third, the state has adopted a “conditional” registration system that permits voters to register and vote in a single trip to the county registrar after the normal registration deadline has passed. And finally, the state has committed to an automated registration system that has the potential to register voters by default when they visit the Department of Motor Vehicles.
The evidence for the efficacy of these reforms varies. Pre-registration is meant to facilitate registration, but research suggests it also helps increase turnout among those who pre-register because it engages them at an impressionable time in their lives (and often in a group setting in their high school or community) (Holbein and Hillygus 2015). Online registration offers significant administrative benefits, but it has not necessarily increased registration or turnout by much (McGhee 2014). For its part, conditional registration is likely to pick up some number of people who miss the registration deadline but decide to vote at the last minute as they get swept up in the excitement of the election. But evidence of its effect on turnout is mixed: some studies have shown a substantial effect, while others have shown something smaller (McGhee 2014).
The final registration reform, automated voter registration, carries perhaps the largest potential to increase registration rates. If implemented properly, the law could increase registration rates by as much as 14 percent in the first year (McGhee and Romero 2016). This may go a long way toward undoing the underrepresentation of Latinos and Asian Americans in the registration rolls. It may also help alleviate some of the need for conditional registration, as more people will be registered throughout the year and do not need to sign up at the eleventh hour. But as I will discuss below, these new registrants need to be mobilized to vote or they may stay home anyway.
In addition to these registration reforms, California is moving toward broad vote-by-mail implementation by enacting a reform first adopted by Colorado in 2012. All voters will get vote-by-mail ballots by default. They can either mail in those ballots or drop them off at any of several drop boxes or “vote centers” that replace neighborhood precincts. If voters lose their vote-by-mail ballot they can have a new one printed at a vote center, and the vote centers will be open for early voting several weeks before Election Day. In short, the new law gives all voters the chance to vote by mail if they want to.
Studies of the effect of this system suggest it saves a great deal of money by limiting the staff and equipment required to operate lightly trafficked polling places (Gronke and Miller 2012; Hall et al. 2012; Folz 2014). The reform may also increase turnout somewhat, though the findings there are mixed (Stein and Vonnahme 2008; Hall et al. 2012; Folz 2014). That said, there is some suggestive evidence that vote-by-mail registrants are more likely to be repeat voters, returning to vote in future elections after showing up in the first one. This pattern is especially notable for young people, meaning it might help alleviate the midterm turnout decline discussed earlier.
Taken together, these reforms place California at the national vanguard for voting access. But we should be wary of complacency. There remain significant issues of implementation and follow-through if these reforms are to increase turnout to the maximum extent.
For example, the ultimate impact of California’s automated voter registration system is dependent on the number of DMV customers who agree to be registered. The new system will certainly be an improvement because it will register any eligible customer who does not actively decline. But it will also require customers to affirm their eligibility to vote before they can be defaulted into registration. While a sensible failsafe measure, this extra step also risks reverting the process back to something like the current system, where DMV customers who skip the section on registration remain unregistered. True default registration would require everyone to answer the eligibility question. In the absence of such a requirement, the success of the new system will hinge in part on how aggressively the eligibility question is pressed upon potential registrants (McGhee and Romero 2016).
There are also a number of implementation challenges for the new Colorado model of voting. One complicated issue concerns how many vote centers to make available for a given population. Since the goal is generally to open fewer vote centers than precincts, it is possible that too few will be opened and voters will have trouble finding a convenient one. Research on the effect of distance on voting has found turnout declines up to 5 percent for distances up to 10 miles from the precinct (Dyck and Gimpel 2005). But since vote centers are more flexible than precincts—they will accept all potential voters no matter where in the county they reside and will be open for weeks rather than just one day—voters may more often find themselves in close proximity to a voting location at a moment when they have some free time to cast a vote.
There are important reasons to be careful about a wholesale switch to the Colorado model of voting. Young people currently choose vote-by-mail less often than older voters because of confusion about what address to use and other issues. If they do choose to vote by mail, they are less likely to send in their ballots on time (Romero 2014). And many people of color are wary of the vote center model because they do not trust the reasons behind it (Romero 2016a, 2016b). That raises the potential for problems with some of the very communities the reform is trying to reach.
Caution is therefore in order. Fortunately, the law is structured to provide it. Under California’s version of the Colorado voting model, each county decides whether to switch to the new system, and even the counties allowed to make the change will be phased in over time. This offers numerous opportunities to assess the rollout and make any necessary adjustments. Given both the promise and potential risks of the Colorado model, it is important to manage the phase-in to ensure the reform is going as expected.
Even if the reforms work as intended, they must still be coupled with aggressive mobilization. While it always makes sense to mobilize as many people as possible in every election, Latinos, Asian Americans, and young people clearly need special attention. Simply placing eligible residents on the voter rolls, or making voting easier, will not solve the whole problem. Outreach will be an important ongoing part of any solution. Research suggests such outreach will be more successful if it is embedded in the communities it is trying to mobilize, with communication by members of the community in ways others in the community understand and relate to (García Bedolla and Michelson 2012). This is especially important because Latino and Asian-American immigrants and their children are least likely to register and may be more culturally and linguistically distinct.
In any case, such mobilization work is hard and must be sustained over many election cycles. Yet it is the effort in many ways best suited to the source of the problem. Many of the biggest legal obstacles to voting in California are now gone or are in the process of being removed. This offers a fresh opportunity to get Californians to engage with the process and make their voices heard.
REFERENCES
Baldassare, Mark. 2016. “California’s Exclusive Electorate: Who Votes and Why It Matters.” San Francisco: Public Policy Institute of California.
Dyck, Joshua J., and James G. Gimpel. 2005. “Distance, Turnout, and the Convenience of Voting.” Social Science Quarterly 86 (3):531-48.
Folz, David H. 2014. “Vote Centers as a Strategy to Control Election Administration Costs.” SAGE Open 4 (1):1-10.
García Bedolla, Lisa, and Melissa R. Michelson. 2012. Mobilizing Inclusion: Transforming the Electorate through Get-Out-the-Vote Campaigns. New Haven: Yale University Press.
Gronke, Paul, and Peter Miller. 2012. “Voting by Mail and Turnout in Oregon: Revisiting Southwell and Burchett.” American Politics Research 40 (6):976-97.
Hall, Steven R., Joseph Losco, and Raymond Scheele. 2012. “Convenient Turnout: A Case Study of the Indiana Vote Center Pilot Program.” International Journal of Business and Social Science 3 (8):304-12.
Holbein, John B., and D. Sunshine Hillygus. 2015. “Making Young Voters: The Impact of Preregistration on Youth Turnout.” American Journal of Political Science 60 (2):364-82.
McGhee, Eric. 2014. “Expanding California’s Electorate: Will Recent Reforms Increase Voter Turnout? Technical Appendix.” San Francisco: Public Policy Institute of California.
McGhee, Eric, and Mindy Romero. 2016. “What to Expect from California’s New Motor Voter Law.” San Francisco: Public Policy Institute of California. Romero, Mindy. 2014. “Disparities in California’s Uncounted Vote-by-Mail Ballots: Youth, Language Preference, and Military Status.” UC Davis: The California Civic Engagement Project.
———. 2016a. “The California Voter Experience: Vote-by-Mail vs. the Polls.” Davis: UC Davis California Civic Engagement Project.
———. 2016b. “The California Voter Experience: Why African-American Voters Choose to Vote at the Polls or Vote-by-Mail, and How They Perceive Proposed Changes to California’s Voting System.” Davis: UC Davis California Civic Engagement Project.
Stein, Robert M., and Greg Vonnahme. 2008. “Engaging the Unengaged Voter: Vote Centers and Voter Turnout.” The Journal of Politics 70 (02):487-97.