In 2014, just 41.9 percent of the voting-age citizen population of the United States voted. But the people who voted are not only in the minority, they form an unrepresentative minority. Millions of Americans are too young to vote. Others are disenfranchised felons, unable to vote for health reasons, missed registration deadlines, stuck at work, dissuaded by voter ID laws. In many salient ways, voters are not like nonvoters: voters are richer, whiter, and older than other Americans. And my new report, Why Voting Matters, shows how their votes produce a government that caters to their interests—and how boosting turnout would lead to a more representative democracy.

Political scientists once accepted the idea that voters were a “carbon copy” of the nonvoting population. In 1999, Benjamin Highton and Raymond E. Wolfinger summarized this consensus, writing that, “simply put, voters’ preferences differ minimally from those of all citizens; outcomes would not change if everyone voted.” More recently, though, that view has come under attack. Jan Leighley and Jonathan Nagler, a pair of political scientists, argue that gaps between voters and nonvoters are real and have widened, and that the divergence in their views is particularly acute on issues related to social class and the size of government. However, measures that examine a one dimensional left-right axis obscure these divides.

Census data on the 2014 midterm elections quantifies some of these gaps. While 52 percent of those earning more than $150,000 voted, only 24 percent of those earning less than $10,000 went to the polls. That divide is further magnified by age. Among 18-24 year olds earning less than $30,000 turnout was 17 percent in 2014, but among those earning more than $150,000 and older than 65, the turnout rate was nearly four times higher, at 65 percent. There were also racial gaps in voter turnout. In 2014, 46 percent of white voters turned out to vote, compared to 40 percent of black voters, and just 27 percent of Asians and Latinos.

Percent Voting, by Family Income Bracket (2012 and 2014)

Census Bureau, 2013 and 2015 Demos Calculations (Sean McElwee)

Policy Preferences

It’s not just the demographics of voters and nonvoters that differ; so do their views. Four questions from the American National Elections Studies (ANES) data show a stark divide on issues related to economic inequality. Nonvoters tend to support increasing government services and spending, guaranteeing jobs, and reducing inequality—all policies that voters, on the whole, oppose. Both groups support spending on the poor, but the margin among nonvoters is far larger. Across all four questions, nonvoters are more supportive of interventionist government policies by an average margin of 17 points.   

Net Support of Policy, Voters and Nonvoters

ANES 2012 (SeanMcElwee)

Factoring in wealth and race only widens these gaps. Between white voters and nonwhite, nonvoters, there was a 42-point net gap on these questions, and between rich voters and poor nonvoters, the gap was 50 points on average.

Net Support of Policy, Rich Voters and Poor Nonvoters

ANES 2012 (SeanMcElwee)

Measuring these differences with other data sets produces similar results. I took numbers from Pew and YouGov comparing registered voters with the non-registered population. These polls were not taken close to elections, so registration can serve as a rough proxy for the voting and nonvoting population. The polls show the same dramatic differences. In every instance, net support for greater government intervention in economic affairs was higher for the non-registered populations—sometimes dramatically so. For instance, while net support for free community college was 7 points for the registered population, it was 46 points within the non-registered population.

Net Support of Policy, Registered vs. Not Registered

YouGov, Pew, Demos (Sean McElwee)

Since nonvoters tend to be younger, less white, poorer and more mobile than voters, this isn’t entirely surprising. But one reason these findings are so striking is that voters and nonvoters hold broadly similar views on a range of other controversial issues. Christopher Ellis, an assistant professor of political science at Bucknell, tells me that gaps on issues like abortion, immigration, and gun control are comparatively modest (he is supported by Pew research). But economic issues are different.

That’s a remarkable finding, because there’s a mass of evidence to suggest that the views of voters play a hugely significant role in changing policy. Brian Newman and John Griffin, who have found that voters are “almost always more conservative” than nonvoters, have argued that, “increases in turnout may lead to greater policy liberalism.” While many scholars have focused on the rather pedestrian question of whether turnout would benefit Republicans or Democrats, Griffin and Newman argue that even beyond the party differences, “voter ideology substantially affects the way Senators cast votes.” They note, for example, that, “in states where voters are more conservative than nonvoters, Senators tend to be more conservative.” Other recent studies bear out the conclusion that, lobbyists and wealthy donors notwithstanding, elected politicians really do tend to act in accordance with the views of their electorates.

Differences in Ideology Between Voters and Nonvoters

Positive numbers indicate that nonvoters are more liberal than voters. (Source: Griffin and Newman, 2005)

One researcher, for example, found that higher turnout among the wealthy changes the legislative agenda: Policymakers spend less time on bills relating to housing, welfare and healthcare. They’re also likely to pass anti-predatory lending statutes, expand children’s health insurance, or increase the minimum wage. Conversely, another study finds that higher turnout among the poor leads to higher spending on welfare programs. Counties with higher turnout receive more funding from the federal government, while districts with lower turnout have less influence on the policy positions taken by their representatives.

Income Bias and Anti-Redistribution Bias

Higher income bias indicates more high-income voting; higher anti-redistribution bias indicates that voters are more likely to oppose redistribution than nonvoters. (Source: Henning Finseraas, 2007)

Electoral Turnout and Redistribution, 13 Countries (Available Years 1979 – 2000)

Vincent Mahler, 2008

The Mobilization Gap

There are many reasons why people don’t vote, but research suggests that three factors are particularly crucial: registration, unions, and parties.

Registration is a barrier that exists in the United States but not in any other country in the developed world. Political scientists have shown that the requirement to register dramatically reduces voter turnout. This effect primarily hurts the poor. One study finds that, “states with restrictive voter registration laws are much more likely to be biased toward upper-class turnout.” But if so, America is moving in the wrong direction. In the wake of a recent Supreme Court case restricting the Voting Rights Act, states have begun to pass an increasing number of restrictive voter ID laws, with racially disparate impacts.

There are other barriers to voting as well. Counties with large black populations are less likely to have access to early voting. Felon disenfranchisement disproportionately reduces turnout among potential voters of color and those with low-incomes, and explains some of the decline in voter turnout over the last few decades. (There is a strong correlation between the pervasiveness of racial bias in a state and the ease of access for a state’s voting system.) Conversely, a recent study suggests that easier access to voting does indeed boost turnout.

Unions were once key catalysts for voter turnout, but without them, turnout has slipped. In a recent study, Jasmine Kerrissey and Evan Schofer find that, “union membership is associated with many forms of political activity, including voting, protesting, association membership, and others. Union effects are larger for less educated individuals, a group that otherwise exhibits low levels of participation.” This is important because these activities, like voting, tend to be dominated by the wealthy. However, as unions have declined in power and influence, so has turnout among low- and middle-income people.

It’s not just unions that are no longer moving blue-collar voters to the polls. Parties have systematically failed to mobilize low-income voters. Between the 1960s and the 1980s, political scientists Steven J. Rosenstone and John Mark Hanse argue, half the decrease in turnout can be attributed to less effort by political parties to contact voters. The rate at which the Democratic Party reached out to high-income voters increased nearly four-fold between 1980 and 2004. In 2004, high income Americans were nearly three times more likely to be mobilized by the parties. In tight elections, however, when parties compete and Democrats are forced to mobilize low-income voters, the turnout gap declines.

It may be that those who have the most to lose from lower turnout among low-income voters least understand its significance. Although the parties have dramatically polarized in the last few decades, lower-income people and nonvoters remain largely unaware of this dramatic shift, and may be inclined to believe that there are few differences between the parties. In their book, Leighley and Nagler find that “those in the top income quintile see a larger difference between the candidates on ideology than do those in the bottom quintile.”

If the public wants to increase turnout, this research suggests a clear set of tools for accomplishing that. First, remove the barriers to registration. Vigorous enforcement of the National Voter Registration Act, combined with same-day or automatic voter registration, could substantially increase access to the polls. Second, create more competitive elections, to incentivize parties to mobilize low-income people. Third, since the decline in unionization has also contributed to the decline in low and middle-income mobilization, unions need to be revived or replaced. (Or, to flip that around, increasing turnout may be the best hope unions have for revival.)

Inequality creates a worrying double-bind: Low-income people become more supportive of interventionist policies, even as they drop out of the political system. The result is a troubling divergence between the economic views of the population as a whole, and the policies that voters and the politicians that they elect tend to favor. The simplest way to reverse that is to mobilize the great mass of potential voters who don’t currently head to the polls. Universal registration, Jan Leighley said, would lead to a “more serious conversation about economic inequality, and one that included a wider range of views.”


Higher turnout may lead to policies that somewhat better reflect the views of poor and middle-class Americans, but other factors will continue to favor the affluent. Members of Congress tend to come from white-collar occupations, and may therefore be less likely to support policies that benefit blue-collar workers. They are also wealthier than the average citizen, and perhaps unsurprisingly, more favorably inclined to policies like eliminating the estate tax that yield disproportionate benefits for the wealthy. More affluent Americans remain more likely to contact members of Congress, to work on campaigns, to donate to funds, or to have social networks that include elected officials.

Median Income at Different Levels of Political Participation

Spencer Piston, ANES 2012

If higher turnout cannot change all of these factors, though, then lower turnout certainly exacerbates them. In low turnout elections, politicians are incentivized to cater to the interests of a small portion of the general public—in a system with near-universal voting, they might take into account a wider array of interests. After all, for what shall it profit a politician, if she shall gain a Super PAC, and lose her own seat?