Civic Engagement

Making Voting Easier Doesn’t Increase Turnout

To the surprise of many, making the act of voting easier hasn’t actually led to higher voter turnout. To increase turnout, we need to get more people interested in politics.

Increasing Voter Turnout: It’s Tougher Than You Think Increasing Voter Turnout: It’s Tougher Than You Think In this 15-part series, election experts from government, academia, and the private and nonprofit sectors will weigh in on important questions, including: What can the social sector do to improve voter turnout in the United States?

The act of voting in elections is the fundamental act of participation in our democratic system. But year after year, millions of Americans fail to show up at the polls. Why don’t people vote?

In their landmark study of voting published in 1980, political scientists Steven Rosenstone and Raymond Wolfinger argued that people didn’t vote because it was difficult. They focused mainly on the process of registering to vote—a task that in the 1970s was in many ways more onerous than the act of voting itself. Some states required voters to register weeks in advance of election day; others required them to travel great distances to a county office. These requirements created tangible barriers to voting.

From this perspective, the cause of low voter turnout was the high costs of voting. Rosenstone and Wolfinger therefore argued that the solution to increasing turnout was to make voting easier by easing registration requirements.

Making voter registration easier came with an added benefit, according to some. Reformers believed that such changes would not only increase turnout, but level the political playing field as well. In 1988, political scientists Richard Cloward and Francis Fox Piven argued that easing registration requirements would lead to greater socioeconomic equality in the composition of the electorate. Thus, registration reform would be the magic bullet that could fix a variety of electoral ills.

In the four decades since Rosenstone and Wolfinger made the call to ease registration requirements, the government answered their charge. The 1993 National Voter Registration Act explicitly directed states to allow voter registration at motor vehicle agencies and through mail-in procedures. With the passage of the Motor-Voter Act, the act of registering to vote became a trivial matter. The high costs identified by Rosenstone and Wolfinger disappeared.

Furthermore, the act of voting became easier as well. In the last 20 years, election officials have enacted a number of different reforms designed to increase turnout by easing restrictions on the casting of ballots including early voting, the relaxing of stringent absentee balloting procedures, and universal voting-by-mail (VBM) in states such as Oregon and, most recently, Colorado. Taken together, the changes over the last 40 years have greatly reduced the direct costs of registration and voting; it has never been easier to cast a ballot than it is today

But the effects of these reforms have surprised scholars and reformers alike. The recent wave of electoral reforms does not seem to have had any significant effect on voter turnout. And there is even evidence that some of the new reforms may have actually decreased turnout.

A brief survey of some of these studies paints a clear picture. Two different research teams concluded that the Motor-Voter Act did not change turnout rates in the United States. Likewise, making the act of voting easier does not seem to have increased turnout rates consistently. My research with political scientists Nancy Burns and Michael Traugott found that VBM increased turnout slightly in Oregon. But political scientists Thad Kousser and Megan Mullin conclude that VBM actually decreased turnout in California. In a similar vein, political scientist Barry Burden and his colleagues have argued that early voting thwarts traditional mobilization activities, thereby counteracting any gains that could rise from lowering the direct costs of voting. The balance of evidence is clear: lowering the direct costs of voting does little if anything to increase turnout.

Moreover, contrary to the predictions of Piven and Cloward, lowering the costs of voting might skew the electorate in important ways. Research by myself and others indicates that the wave of voting reforms have had significant, if unintended consequences, on the types of people who turn out to vote. In particular, reforms designed to make voting easier actually magnify the existing socioeconomic biases in the composition of the electorate.

These two unexpected consequences have a common cause. The problem, I believe, is that when we talk about the “costs” of voting, we have been thinking about the wrong kinds of costs—the direct costs of registering to vote and casting a ballot. Most politicians and scholars have focused reform efforts on these tangible barriers to voting, making it easier for all citizens to vote, regardless of their personal circumstances. But, as I have argued elsewhere, the more significant costs of participation are the cognitive costs of becoming involved with and informed about the political world. Studies of voting from the last 60 years make this point clear. Political interest and engagement, after all, determine to a large extent who votes and who does not.

By focusing on the wrong kind of costs, we have lost sight of the real effects of reforms such as early voting and voting by mail. Rather than stimulating the unengaged, who are relatively deficient in political and economic resources, reforms to make voting easier mostly serve to keep existing voters voting. Thus, any slight increase in turnout works by ensuring that politically engaged voters continue to come to polls election after election. Nonvoters are left behind. Without engaging this segment of the US public, turnout will remain low.

What then, can be done? If we are serious about increasing turnout, we need to change the nature of the conversation about voting reform. Institutional change alone will not bring about a democratic electorate. Put simply, no matter how low the direct costs to casting a ballot are set, the only way to both increase turnout and eliminate socioeconomic biases in the voting population is to increase the engagement of the broader public with the political world. Political information and interest, not the high tangible costs of the act of voting, are the real barriers to a truly democratic voting public.

[The last names of Piven and Cloward were swapped. Correction made on 2-21-16]

Tracker Pixel for Entry


  • You swapped Piven and Cloward’s last names in paragraph four, otherwise an excellent and provocative article.

  • What’s the right age to begin cultivating this important behavior?  Is it 12?  or 10?  Will civic engagement now be an integral part of public K-12 education in a way it hasn’t been to date?

    It’s interesting to note that studies show a person has decided whether or not to breastfeed by middle school, so the importance of this crucial public health practice has to be explained when kids are quite young.  Maybe the same is true for encouraging active citizenship.

  • Adam: I think you misunderstand Piven and Cloward’s efforts (and you leave out evidence that the National Voter Registration Act has improved registration or turnout in some studies). More importantly, the NVRA needs to be implemented well to work. The problem wasn’t the policy intent so much as the poor execution of it by states and local officials, and severe neglect under the Bush Administration’s DOJ. To this day, many states still haven’t implemented NVRA-mandated programs properly. This is what might cause the unintended consequences you find. Indeed, you are right that passing laws alone is not enough, but that is at least partly because some laws are not properly implemented or designed to target the non-participating groups. The latter is a serious problem, which gives rise to the perverse consequences (let alone unintended); however, good policy can help. To wrap up, I fear that much of the work being done by some foundations is not fully aware of your warning. However, in this brief column,I don’t think you diagnosed the source of the policy problem quite right.

  • JC. Thanks for pointing out that Piven and Cloward’s last names were swapped. We fixed it.

  • Phil Keisling's avatar

    BY Phil Keisling

    ON March 14, 2016 04:40 PM

    though I greatly respect Dr. Berinsky’s bonafides as an elections scholar, I have to take strong issue with his characterization that there’s a “clear picture” from the academic research that oft-called “vote by mail” ballot delivery systems like Oregon’s have little—or even a negative—effect on the turnout of voters.

    If anything, the opposite is more likely true: that of all the actual and potential “innovations” out there,  the potential effect of this one (compared to so much else being advocated for, much of which I actually support) is like the difference between lightning, and a lightning bug.

    The Oregon system—in use since 2000, and now also used by Washington (2012) and Colorado (2014)—is best understood as one in which 100% of registered voters are automatically sent their ballots, two weeks prior to Election Day, via the U.S. Postal Service. Voters can then return their ballots either back through the mail, or by taking them to one of literally hundreds of official ballot drop sites, anywhere in the state.

    Those scholars—though not Dr. Berinsky, to his credit—that disparage this as “mandatory” or “compulsory” voting by mail fail to grasp the separate and discreet steps involved in the “voting act,” steps that I became quite familiar with during my 9 years serving as Oregon Secretary of State (1991-99).

    By design, this reform takes those already at what I’ll call “Step 1” of the voting process—being registered voters—and automatically gets every one of them to Step 2 (connecting with their ballots). Of course, it’s then critical that these voters complete   Step 3 (marking the ballot) and (most importantly)  Step 4, of actually casting/returning/voting it.

    Recognize this multi-step reality when it comes to voting,  and several other key things quickly become apparent, too. Oregon’s Universal VBM ballot delivery system is also profoundly different than “Absentee ballots for everyone.” Absentee ballots require an application—and in most states approval viz specific legal criteria—long before Election Day. And in stark contrast to 47 other states,  America’s three “Universal Ballot Delivery” states literally eliminate the distance between Steps 1 and 2 for 100% of their registered voters, no longer forcing them to either physically go their ballots, or to apply for (and qualify for) an “absentee ballot”.

    What these systems can’t do—and have never pretended to do—is to get people to Step 1 (voter registration) in the first place. Nonetheless, most academic studies of voter turnout use as a measuring stick of all eligible citizens, including the unregistered— leading to critiques that often are the equivalent of “Sure, he can bat .450; but he still strikes out a lot!”

    Grasp what these 3 states’ systems are—and aren’t—and the real-world evidence becomes compelling that, at a minimum, this idea is worth far more analysis and discussion for its electoral “innovation quotient” than the painfully short shrift (less than a paragraph) given in SSIR’s otherwise excellent 15-part series.

    In the 2014 midterms, the turnout rate among active, registered voters in the US—as per the Election Assistance Commission’s 2015 report —was a paltry 48%. In the 16 battleground states that year, it averaged about 56%. Oregon’s rate was 71%—and as a decidedly “non-battleground” state. Colorado—which used this system for the first time in 2014—was the only other U.S. state to exceed 70%. (Indeed, no other “battleground” state even came close on this yardstick).

    In the 2010 midterms, Oregon and Washington were the only two states to exceed 70% turnout against ARVs,  vs. a national average of just over 50%  And the performance of Oregon and Washington in party primary elections has been just as notable:  rates consistently close to 40% in recent cycles, vs about 20% of active registered voters in a typical state (and even single digit turnout rates in some). Those who’ve studied American elections carefully—not to mention candidates who run for partisan office— keenly understand that in today’s politics,  for the vast majority of Congressional, Senate, Gubernatorial, and state legislative races in the U.S, winning the dominant party’s primary election is the key. The general election contest itself is largely a formality.

    So what about the scholarly research that, notwithstanding these realities, purportedly proves the Oregon system having little—or even negative—effect on voter turnout? Look a bit closer, and suddenly there’s a lot less than meets the eye.

    For example, the specific study that Dr. Berinsky cites—which he did along with colleagues Drs. Michael Traugott and Nancy Burns—was published 15 years ago, in 2001. Oregon had used its new system in just one regular election cycle (the 2000 presidential contest)—and it’s in presidential contests, due to the many other turnout-inducing dynamics present, that a reform like this will logically show the least impact..

    Sadly, 3 presidential, 4 midterm, and 8 party primary elections later, this arguably remains the most comprehensive Oregon-based studies. The only other specific studies Berinsky refers to, from the mid-2000s, use select precinct data in a few California counties. Left unmentioned – perhaps because it might have muddled an otherwise clear picture? – are more recent studies, such as one by Georgia State’s Dr. Sean Richey that found a whopping 10% turnout difference in Oregon across 30 years of elections.

    While voter turnout has steadily declined (amidst great lamentations) in midterm and primary elections in the last 10+ years—and in 2014 took a huge downturn viz 2010— Oregon’s has increased, even as we’ve lost our battleground status and had fewer truly competitive races. It went up dramatically in Colorado in 2014 compared to 2010 (when that state was also a battleground). And just last November, when California’s San Mateo county went to Universal Ballot Delivery for all municipal elections—there were dramatic voter turnout increases, especially among younger and Latino voters.

    I will readily admit to being a policy advocate here. However —and this is key—I am not trying to “prove” the opposite to be true: i.e, that the evidence is “clear” that this system dramatically improves turnout. 

    My larger point is a subtler (but also more important) one: that the research to date, especially in the last 5-10 years, has been far less comprehensive—not to mention,  far less definitive—than it’s typically portrayed to be in dismissively brief paragraphs like the one noted here.

    At the very least, the performances of states like Oregon, Colorado, and Washington should be provoking far more curiosity among researchers than they have shown to date. In my own professional career, I’ve drawn conclusions based on what I considered to be the best available evidence at the time—and then changed my mind as new data was revealed, or when a new framework revealed itself that allowed for new insights. It is my fervent hope that knowledgeable experts—perhaps even Dr. Berinsky—will have the curiosity and the willingness to re-visit this subject, even at the risk of unsettling what’s been widely viewed as conventional wisdom.

    Finally, those who might want to learn a bit more about Oregon’s system—including some recent research involving voter turnout by age using state-by-state voting records— can do so in the Jan/Feb 2016 issue of the Washington Monthly:

  • Zoe Harris's avatar

    BY Zoe Harris

    ON May 13, 2016 07:07 PM

    Great article and I certainly agree with you. But what about the wealth of research supporting election day voter registration as a means to increase turnout? It will also be interesting to study the relationship between automatic voter registration and turnout.

  • Mark Anderson's avatar

    BY Mark Anderson

    ON May 15, 2016 04:22 PM

    Phil Keisling’s comments, and his article in the Washington Monthly, are well worth reading.  He is a thoughtful individual & served our state well as a public servant - that rare public official who actually believed in servant-hood for the good of the whole. 

    As a fellow Oregonian, I have mixed feelings about automatic voter registration, as well as vote-by-mail.  Not because these don’t increase participation - but precisely because they do - by many who, in the past, couldn’t be bothered enough to register & show up to vote.  According to the last stats I saw on this, a number actually went to the trouble of un-registering, and the majority didn’t have a party preference.  Yes, there have been many barriers, but in an age when there is online registration & computer access in public libraries, these various barriers are much less. 

    At the heart of this is the question whether votes by those not engaged enough to know what they are voting for improves decision-making - proxies being active registration & voting.  On the one hand, participatory democracy is something I believe in.  On the other, the only thing worse than no vote may be an uninformed one.  (Having done research on medical ‘informed consent’ many of those who ‘consent’ don’t really know what they’ve consented to - they just go along with what the doctor or family says ... often to their own detriment.  This on an issue that can literallly involve their own life-or-death.  Ditto for voting.)

    I note that those who have pushed for auto-registration & vote-by-mail recently - Democrats - do so in large part because they believe that will increase Democratic votes. 

    And yet many of those Dems strongly oppose the most direct form of participatory democracy - the initiative/referenda.  Seems to me they want it both ways - give power to lower-motivated, lower-engaged, lower-informed individuals in the name of democracy - by electing Democratic officials, but don’t actually engage in the process by directly enacting legislation - or vetoeing it.  Dems have actively subverted referenda by adding an ‘emergency’ clause to legislation that is not an ‘emergency’ in the ordinary sense of that word, precisely because they know the chances are high that a majority of engaged voters will overturn that legislation (the most recent example of being overturned was the driver’s license to unauthorized (illegal/undocumented) aliens so those aliens could go to work - which under federal law is illegal).

    And with respect to registration, the current Dems have actively blocked the many attempts by the Independent Party (of which I am a member) to achieve ‘major’ party status, and thwarted implementing that accomplishment by refusing to honor that party’s wishes as to how to run its primary.  Again, not really interested in increasing registration per se - but in registration as a Democrat.  (Automatic registration, including those who don’t choose a party, may dilute numbers enough that the Party may be de-certified.)

    I note that the largest group of unrepresented voters in Oregon & the US are those independent third-party members, as well as unaffiliated, who constitute a plurality - disproportionate numbers of Democrats & Republicans who don’t really represent them, who regularly follow Party partisan policies, not what a plurality, much less a majority, of voters want.

Leave a Comment


Please enter the word you see in the image below:


SSIR reserves the right to remove comments it deems offensive or inappropriate.