Back to top Jump to featured resources
Article filed under The Public Good

American Voice 2004: Is it true that our voter turnout is much lower than that of other democracies?

| Written by David Morris | No Comments | Updated on Jun 1, 2004 The content that follows was originally published on the Institute for Local Self-Reliance website at

Q. Is it true that our voter turnout is much lower than that of other democracies? Don’t Americans care enough to vote?


Throughout our history, Americans have passionately defended and fought for the right to vote. But we exercise that right far less than citizens in other countries. For example, from 1945 to 1998 voting turnout in the U.S. ranked 139th of 171 countries.[1] Our turnout rate of about 50 percent compares to the 73 percent average for the 36 established democracies.[2]

Is this bad? You bet it is. Is this really bad? The jury’s still out. Many argue that a low voting turnout is a symptom of disaffection and cynicism and that in the long run it undermines the legitimacy of elections and support for democracy. Others argue that our comparatively low voter turnout is a result of barriers to voting, not a lack of civic pride in voting.

Here’s some background. In the 19th century voting turnout in the U.S. achieved levels of 75 percent, comparable to European levels today. But at that time the franchise was limited to white men over 21, mostly with property. After World War I, the number of people eligible to vote expanded, primarily because women were given the right to vote, while the proportion of people who actually voted fell. The women’s vote, in terms of turnout, did not catch up to that of men until 1980.

After the Civil War, and especially after World War I, southern states made it all but impossible for blacks to vote. As recently as 1952, according to the U.S. Census Bureau, only 4 percent of southern blacks of voting age voted.

Voting turnout rose to 62.8 percent in the Kennedy-Nixon election of 1960, a 20th century high. (In comparison, at the height of the Great Depression in 1932 only 53.4 percent of the voting age population cast a ballot in the Roosevelt-Hoover election.) The Voting Rights Act of 1965 dramatically increased the participation of African-Americans, but the voting turnout dropped to 55 percent in 1972 after 18 year olds gained the right to vote, and has stayed slightly lower than that for most elections since then.

Turnout for Congressional elections in non-presidential election years is lower still, hovering between 36-39 percent since 1980.

Voter turnout in the United States (and virtually all other countries) has been drifting lower throughout the last 20 years. This has occurred despite a major increase in registration effected by the 1993 National Voter Registration Act (commonly called the motor voter law). That law required states to register voters when they get drivers’ licenses, Aid to Families with Dependent Children, Medicaid or other state services.[3] According to the Census Bureau, 40 percent of those who registered to vote between 1995 and 2000 did so at a state motor vehicle department.[4]

Although the motor voter law spurred registration, many observers believe that our unique state registration laws, first introduced in the 1920s, present a significant barrier to voting. By one estimate, registration laws in and of themselves may account for half the difference between U.S. turnout rates and those of other advanced democracies.[5]

Another factor accounting for our comparatively low voter turnout is that election day is a work day. This may be an accident of history. Back in 1872, federal law established election day as the first Tuesday after the first Monday in November. At the time most people worked on Saturdays. Thus voting on a Saturday would not have been more convenient. Church leaders opposed Sunday voting because elections often were accompanied by drinking and gambling.

Whatever the original reason for choosing Tuesday, by now it is deeply embedded in our psyche. By 1996, 46 out of 50 states held their primaries on Tuesday.

Not surprisingly, one recent study found that countries that vote on the weekend or on a holiday have a 6 percent higher turnout than would otherwise be expected.[6]

After the 2000 presidential election ex-presidents Gerald Ford and Jimmy Carter sponsored the National Commission on Federal Election Reform, a group of politicians, academics, and attorneys. The Commission’s report, To Assure Pride and Confidence in the Electoral Process, recommended that election day and Veteran’s Day be merged into a single national holiday.[7] Advocates argued it would be appropriate to use the day set aside to honor those who fought to defend our freedoms as a day to exercise that freedom.

Turnout levels are conventionally measured by dividing the number who vote by the total voting age population.[8] Some commentators argue that the recent decline in voter turnout is not because fewer people want to vote but because fewer people are eligible to vote. By eliminating non-citizens and disenfranchised felons,[9] and adding military personnel and civilians living outside the U.S., Michael McDonald and Samuel Popkin have estimated the voting eligible population. They concluded that voting turnout as a percentage of the population eligible to vote has decreased very little from the 1970s to the 1990s. What has changed dramatically is the number of people ineligible to vote (see chart).


The primary factor behind the remarkable increase in disenfranchised Americans is the increase in prisoners and more narrowly, the increase in imprisonment for drug offenses. Almost 5 million Americans, more than 2 percent of the voting age population, are today ineligible to vote because they have been convicted of a felony. Some 1.4 million of these are ex offenders who have completed their criminal sentence.

Six states, Alabama, Florida, Mississippi, New Mexico, Virginia and Wyoming exclude more than 4 percent of their adult population from voting. In Alabama and Florida, almost one third of all black men are permanently disenfranchised.

Some states are reexamining their voting bans on inmates and ex-offenders, but more states are moving in the opposite direction. In 2000, Massachusetts and New Hampshire disenfranchised inmates. In 1999, Oregon disenfranchised federal inmates, in l997 Colorado disenfranchised federal inmates and parolees and in l998 Utah disenfranchised all inmates.[10]

Voting turnout guru Mark Franklin suggests that the most interesting piece of what he calls the “turnout puzzle” is not the decline of voter turnout but the stability of voting turnout. In virtually all countries, voting turnout varies little from election to election. Which means, to Franklin, that the propensity to vote is established early in people’s lives. When new voters arrive in large numbers turnout drops until they gain (or don’t gain) the voting habit. Franklin’s research suggests that throughout the world, voting turnout is slowly declining. He is unsure that a nation can do much to substantially increase voter turnout.

If voter turnout were to rise overall, would it affect election outcomes? Some observers argue it wouldn’t. In 1992 Ruy Teixeira asked the question “What if they gave an election and everybody came?” His answer? “Not much”.[11] A 1995 study concluded, “our data support the conclusion that voters and non-voters do not seem to differ substantially in their attitudes on public policy issues.”[12]

Martin Wattenberg of the Center for the Study of Democracy disagrees, especially with respect to non-presidential elections. He tells of seeing a bumper sticker after the historic GOP takeover of Congress in 1994 that read, “Newt happens when only 37 percent of Americans vote.”[13] Wattenberg then examined what would have happened if all voters had turned out in the same proportions regardless of how much formal education they had had. “ If turnout rates had been equal among all education categories the Republican share of the vote would have fallen from 52.0 to 49.2 percent.” Wattenberg also estimated the impact of higher turnout by gauging the likely behavior of registered nonvoters. “Overall, had all registered citizens gone to the polls, the GOP vote share would have been reduced by 2.8 percent”, he concluded. “Applying this loss uniformly to all districts yields an estimate of only 206 seats won by the Republicans which is 24 less than they actually won and l2 short of a majority.”

Which is why both parties are feverishly engaged in get-out-the-vote efforts targeted at demographic groups that might favor their candidates.

Voting turnout differs dramatically by income, education, age and marital status and in most cases these divergences are widening. Only 59 percent of those aged 18-29 report voting while those 55 and over had a voting turnout above 80 percent. Turnout among those with incomes above $50,000 is 89 percent while it is only 59 percent for those with incomes less than $10,000. Almost 63 percent of married folks vote; only 45 percent of single folks do.[14]

When we add these characteristics together, the voting turnout differences can be quite substantial. Wealthy, college educated, older whites vote at the rate of 91 percent while only 22 percent of young, poor, minority group members who did not finish high school vote.

Get-out-the-vote efforts should be successful because the majority of people who don’t vote say they would like to vote. More than 60 percent of those who don’t vote, according to the Census Bureau, offer a variety of reasons why they did not in the last election.[15] One could imagine a get-out-the-vote campaign that addressed these obstacles.

On the other hand, a third of those who don’t vote consciously decided not to vote, either because they did not prefer any candidate (13%) or because they were simply not interested (16.6%).[16] Get-out-the-vote campaigns probably would be less successful in this group. And after all, the right to vote gives us the right not to vote too. Right?

[1]Voter Turnout: A Global Survey. International Institute for Democracy and Electoral Assistance

[2] International Institute for Democracy and Electoral Assistance, Voter Turnout: A Global Survey. Established democracies are defined in this study as countries that have been democracies for over 20 years and have a population of at least one-quarter million people.

[3] National Voter Registration Act of 1993. States that do not require voter registration or permit same day registration are exempt from the law. This includes Minnesota, North Dakota, Wisconsin, Wyoming, New Hampshire and Idaho. California, Illinois, Pennsylvania, Michigan, Mississippi, New York, South Carolina, Vermont and Virginia initially refused to comply with the law and challenged its constitutionality. A series of cases in the federal courts have confirmed the constitutionality of the law. For more information see

[4] U.S. Census Bureau. Press Release. July 28, 2004.

[5] G. Bingham Powell, “American Voter Turnout in Comparative Perspective”, American Political Science Review 80:17-44. 1986. After 1972, Wisconsin, Wyoming, Idaho, Maine, New Hampshire and Minnesota allowed voters to register on election day. These states, plus North Dakota, which in 1951 abolished formal registration completely, are among the states with the highest voter turnouts. Some countries make voting mandatory. This is true of Greece, Belgium, Australia and Luxembourg. Australia adopted compulsory voting after its turnout fell to what the nation thought was a scandalously low rate of 58 percent in 1922. Since then Australia’s turnout has often been above 80 percent. When the Netherlands abandoned compulsory voting in 1972, turnout dropped by 11 percent and later stabilized at 8 percent below former levels. Also see Frances Fox Piven and Richard Cloward, “Why Americans Don’t Vote”, Pantheon Books 1988. Excerpt available here. And Jason P.W. Halperin, “A Winner at the Polls: A Proposal for Mandatory Voter Registration”, Legislation and Public Policy, Vol. 3(69) 1999-2000.

[6] Mark Franklin, “Electoral Participation”, In Lawrence LeDuc, et. al. Eds. Comparing Democracies. Thousand Oaks. CA: Sage. 1996. Franklin also notes however, that “when countries adopt weekend voting their turnout does not appear to increase nor does turnout appear to drop when they move to weekday voting”. Thus there may be something in the culture of voting in different countries that is separate from when the voting takes place. See Mark N. Franklin, The Voter Turnout Puzzles. Trinity College, Connecticut. 2002.

[8] The science of estimating voter turnout is somewhat muddled by the different ways in which the Federal Election Commission and the Census Bureau count voters. The FEC uses voter rolls from state election offices while the Census Bureau asks people whether they are registered and whether they have voted. Their estimates vary, sometimes widely. For example, the FEC found that the number of registered voters increased after the motor voter act while participation among registered voters declined. In 2000, 76 percent of the voting age population was registered, up from 70 percent in 1992, but only 66 percent of registered voters voted, down from 78 percent in 1992.( Federal Elections Commission, Voter Registration and Turnout Statistics). Curtis Gans, director of the Committee for the Study of the American Electorate (CSAE), argues that the reason for this discrepancy is that the FEC’s number for registered voters is inflated. The motor voter law made it harder to track the number of registered voters by prohibiting states from removing non-voters from voting rolls until two federal elections have passed. (Center for the Study of the American Electorate, Battleground State Mobilization Efforts Propel Voter Turnout Slightly Upward, November 10, 2000, and personal communication, 7/30/04.) This inflates registration numbers because people who have died or moved remain on the rolls. CSAE estimates registration at around 67 percent of eligible voters Center for the Study of the American Electorate, Final Post Election Report, 1998. The CSAE’s estimate is similar to the U.S. Census Bureau’s estimate of 69.5 percent of the voting-age citizen population (similar to the U.S. Census Bureau’s estimate of 69.5 percent of the voting-age citizen population). According to the Heritage Institute, some states have regulations that automatically register homeowners and drivers, making it more complicated to keep non-citizens off the voting rolls. Todd F. Gaziano, Election Reform, Heritage Foundation, March 14, 2001.

[9] Michael P. McDonald and Samuel Popkin, “The Myth of the Vanishing Voter”. Only Maine and Vermont allow prisoners to vote. Most states restore the right to vote when a sentence and probation have been completed, but thirteen states take voting rights away for life. The argument against allowing felons to vote is simple: people who don’t follow the rules shouldn’t get a say in making them. Revocation of voting rights is part of the punishment for certain crimes. It is specifically allowed by the 14th Amendment and does not reduce a state’s congressional representation. Todd F. Gaziano, Election Reform, Heritage Foundation, March 14, 2001. The Supreme Court has found that even if felon disenfranchisement laws have a disproportionate racial impact, Congress does not have the right to change such laws unless they were created with a discriminatory intent (Hunter v. Underwood, 471 U.S. 222 (1985)). The 2001 report of the election reform commission noted, “A strong case can be made in favor of restoration of voting rights when an individual has completed the full sentence…including any period of probation or parole.” When Senator Harry Reid proposed an amendment to the Help America Vote Act that would have enacted that suggestion, it failed. Rebecca Perl, “The Last Disenfranchised Class”, The Nation, November 23, 2003.

[10] Losing the Vote. Sentencing Project. October 1998.

[11] Ruy A. Teixeira, The Disappearing American Voter. Washington, D.C. The Brookings Institution. 1992. Cited in Martin P. Wattenberg, Turnout Decline in the U.S. and Other Advanced Industrial Democracies. Op Cit.

[12] Sidney Verba, Kay Lehman Schlozman, Henry E. Bracy, Voice and Equality: Civic Voluntarism in American Politics. Cambridge, MA: Harvard University Press. 1995.

[13] Martin P. Wattenberg, Turnout Decline in the U.S. and Other Advanced Industrial Democracies. Center for the Study of Democracy. 1998.

[14] Christine Barbour and Gerald C. Wright, Keeping the Republic: Power and Citizenship in American Politics. 2000. “By 1996 a significant marriage gap had opened up with respect to turnout. Because of the substantial decline in marriage rates, this is one of the chief demographic correlates of the decline of turnout.” The Civil Rights Project, Harvard University, Democracy Spoiled, July 12, 2002

[15] A 1996 Census Bureau study found that 21.5% could not take time off from work/school; 15% sick, disabled, family emergency; 4.3% had no way to get to the polls; 11.1% were out of town; 4.4% forgot to vote.).

[16] Then there is the not insubstantial number of people who thought they had voted but whose votes were not counted. The CalTech/MIT Voting Technology Project estimated that 3 million eligible voters who thought they were registered were unable to vote in 2000. About 1 million people did not vote because of problems with the lines, hours or locations of polling places. An unknown number of absentee ballots were not counted. CalTech/MIT Voting Technology Project, Voting: What Is, What Could Be, Appendix 2, Estimating Lost Votes, July 2001. At least 1.9 million votes in the 2000 presidential election were “spoiled ballots” and were not counted according to a Harvard University analysis