Author Essays, Interviews, and Excerpts, Books for the News, Politics and Current Events

Voting by Mail? Read an Excerpt from “Democracy in America?: What Has Gone Wrong and What We Can Do About It”

With fears of COVID-19 keeping some away from in-person polling areas, is it time for all elections to be held by mail? Would this increase overall voter turnout even during times that aren’t faced with a pandemic? Could it make voter turnout more representative? In Democracy in America?: What Has Gone Wrong and What We Can Do About It Benjamin I. Page and Martin Gilens examine how, even though mail-in voting can increase turnout in currently registered voters, it is not enough on its own to handle low, and often unrepresentative, voter turnout.

book cover


It is easy to come up with reforms that would lower individuals’ costs of voting and thereby increase voter turnout. The best single reform would be universal, government-administered registration, about which we will have more to say in a moment. Short of universal registration, we could at least allow same-day registration at polling places when people show up to vote. Online registration and registration updating were shown in the 2012 California election to increase the number of voters, especially among young people.

After more (preferably all) Americans are registered, we could make it much easier to vote by holding elections on a holiday rather than a working day; by allowing for early voting or universal absentee voting; or by holding all-mail elections. Longer in-person voting hours, more polling places and voting machines, and perhaps Election Day voting centers in high-density population areas would help as well.


All of the aforementioned reforms would tend to increase the number and proportion of Americans who vote. When it comes to our central objective, however—making government policy more responsive to average citizens—the number of Americans participating in politics is actually less important than the representativeness of those who participate. Since universal participation is probably not feasible, we want government officials to listen to a fully representative set of citizens. As a statistician might put it, we want elections to be decided by an “unbiased sample” of Americans, not by a sample that is biased toward the affluent or any other particular group.

This point is easy to understand if we think about an imaginary “reform” effort that focused on registering customers at high-end jewelry stores and luxury automobile dealerships. This might well increase registration and turnout a bit, but it would make the current class bias in voting worse. Most of the new voters would be affluent. Affluent Americans are already overrepresented in the electorate. Why work to increase their overrepresentation?

Unfortunately, something similar is true of several reform proposals for increasing election turnout: they may increase participation, but they do not do much to decrease—and in some cases they would actually increase—biases in participation.

Research indicates, for example, that the “motor voter” law (the National Voter Registration Act of 1993, or NVRA), which made it easy to register at the Department of Motor Vehicles and other public offices, may well have increased overall registration, but it has only modestly reduced the bias of the voting population toward higher education, income, and age groups. By contrast, it would probably reduce participation biases substantially to fully enforce provisions of the NVRA as applied to the Affordable Care Act, by offering voter registration through federally facilitated Health Benefit Exchanges.

Election Day registration, which clearly has a positive impact on overall turnout, appears to reduce biases in turnout between individuals with the lowest and highest incomes, at least in those states that previously had large income-based gaps in turnout. Online registration mobilizes more young people, but it does not have much effect on other electoral biases.

Again, all- mail elections tend to increase overall turnout modestly, but mainly by making it easier for people who resemble current voters to participate, not by mobilizing chronic nonvoters into the electorate. All-mail elections do not make the electorate more representative of the voting-age population.

Even more disappointing so far have been the results of early voting, both in person and absentee. This popular reform, now common in many states, is most likely to help individuals who are already registered and more likely to vote in the first place. Early voting can actually decrease turnout and the likelihood of voting, if not paired with reforms such as Election Day registration. Moreover, political campaigns appear to react to early voting systems in ways that tend to undermine the reforms. Even more money is required to run an effective campaign, and there have been many postelection court cases questioning the validity of ballots cast early, especially those cast via absentee ballot. Unless same-day registration is already in place, early in-person voting can exacerbate existing gaps between the resource-advantaged (wealthy or highly educated people) and the resource-disadvantaged, particularly when campaigns spend more money to strategically target the most-likely early voters.”

One extremely disadvantaged group consists of people who were once convicted of a felony and have served prison time. Many states either flatly bar ex-felons from voting or impose difficult requirements before they can regain the franchise. This means that millions of Americans—about 5.8 million, including roughly one in every thirteen African Americans—are unable to vote because of a felony conviction. This slows the reintegration of ex-prisoners into society and has negative effects on whole communities. Since the United States has many more prisoners in proportion to its population than any other advanced country does, felon disenfranchisement leaves a large set of our citizens unprotected, with no way to hold politicians accountable for their welfare. These laws need to be changed.

Benjamin I. Page is the Gordon Scott Fulcher Professor of Decision Making at Northwestern University and the author or coauthor of several books. Martin Gilens is professor of Public Policy at the UCLA Luskin School of Public Affairs. He is the author of Why Americans Hate Welfare: Race, Media, and the Politics of Antipoverty Policy and Affluence and Influence: Economic Inequality and Political Power in America.

Democracy in America?: What Has Gone Wrong and What We Can Do About It is available now on our website and from your favorite bookseller.