Vanishing Voters

Electoral participation is vital to political democracy. Yet in the past quarter century our rate of voting participation has dropped sharply and shows no signs of rebounding. In 1988 just 50.2 percent of voting-age adults voted for President, down from 62.8 percent in 1960. Voting for lesser offices, chronically lower than presidential voting, has fallen dramatically as well. In 1986 only 33.4 percent of the voting-age population participated in House elections. The last time half the eligible population cast ballots in House elections in a presidential year was 1972.

Turnout in 1988 came startlingly close to the depressed levels of 1920 and 1924, the all-time lows for the twentieth century, when a majority of the voting-age population did not vote. If present trends continue, in 1992 a majority of voting-age adults will again sit out the presidential election. Yet unlike the early 1920s, when turnout rates dropped after the Nineteenth Amendment expanded the electorate to include women, no abrupt event in the current electoral era explains the non-participation of half the citizenry.

Scholars do not agree on why voting participation has dropped so sharply since 1960. Several popular theories suggest different remedies -- or in some cases no remedy. There is, however, one well established fact. Though non-voting has spread through all social classes, in our time the poor, the uneducated, and the young are least likely to vote. E.E. Schattschneider, a leading American political scientist in the 1950s and 1960s, presciently suggested in his classic essay The Semisovereign People that those who vote in America may constitute "the largest, most broadly-based, ruling oligarchy in the world." That ironic characterization still aptly describes American politics.

Explaining Voting Decline

Ruy Teixeira's Why Americans Don't Vote, the work of Paul Abramson and John Aldrich, and Raymond Wolfinger and Steven J. Rosenstone's Who Votes? exemplify one leading school of thought. They explain the propensity to vote mainly in terms of voter traits, such as income, occupation, education, and partisanship. A rather different, more conspicuously historical and structural approach, can be found in Walter Dean Burnham's work. Burnham and those influenced by him, such as Paul Kleppner and the historian Michael McGerr, argue that the dynamics of turnout decline since 1960 are linked to the overall path of party and electoral change since the late 1890s. A variation on this view, exemplified by Frances Fox Piven and Richard Cloward's Why Americans Don't Vote, embellishes Burnham's approach to argue that low turnout since 1960 results from explicit political efforts to keep poorer, less educated, and minority voters out of politics.

Others writing in this broad tradition hold that turnout has fallen, particularly among lower-class voters, because the Democrats have moved to the right, abandoning their working-class and poor constituencies as they have become more like Republicans. For these constituencies, party politics now only offers "echoes," not "choices," according to Thomas Ferguson and Joel Rogers, among others. However, as I argue, important differences between the parties persisted, but the policies adopted by Democrats in the sixties weakened their ties to their historic constituencies and thus depressed turnout of their potential base.

Many political scientists of diverse schools now see a microeconomic cost-benefit calculus in the decision to vote. Voting, while easy, does cost time and may require some sacrifice of income or leisure. Sorting out the candidates and issues also takes time and energy; these are "information costs." Yet the individual benefits of voting are nearly zero, since the actual contribution of a single vote to a policy outcome is obviously extremely small. A strictly rational view would predict zero turnout, since if each person precisely calculated his or her costs and benefits, no one would vote. Of course, in crude form this insight borders on the tautological: if someone does not bother to vote, the cost must not be worth the benefit. But when married to an analysis of voter traits, the cost-benefit view can be illuminating, since certain traits help voters to pay the "costs of participation." And when incorporated into historical-structural interpretations, the cost-benefit perspective shows that the evolution of electoral politics has periodically changed the individual logic of voting.

The "Voter Traits" Approach

In Who Votes? Wolfinger and Rosenstone refined a standard proposition, namely, that "haves" are more likely to vote than "have-nots." They found that the most potent predictor of voting was not income or occupational status but education. The more their years of schooling, the more likely Americans are to vote. By contrast, past a certain threshold level, income has no impact on turnout. (While occupational status has a more powerful effect, it is nowhere near as great as the effect of education.) Education appears to be so powerful a predictor because it promotes civic-mindedness and better enables citizens to follow politics and navigate the complexity of voter registration.

Plausible as these propositions were, they also posed a puzzle. Because the population has become better educated, turnout should have risen since 1960. Also, outside the South the average presidential turnout from 1840 to 1896 was about fourteen percentage points higher than average presidential turnout from 1900 to 1984, yet twentieth-century Americans are better educated.

A VERSION OF THE "voter traits" approach, exemplified by the work of Teixeira, resolves the puzzle. This approach links turnout to changes in attitudes, such as depth of partisanship and sense of political efficacy. Teixeira proposes that recent turnout decline reflects a crisis in the "system of the 1950s," when the relatively strong partisan identities created in the 1930s and 1940s still persisted. In those years, voters had a high sense of political efficacy and consequently were willing to pay the costs of participation.

Political parties in the United States, unlike most advanced industrial democracies, do not work hard to mobilize voters. Compared to Europeans, few Americans are formal party members. America is unique in the registration burdens it places on voters. Nonetheless, the relatively strong partisan identities left over from the New Deal and a correspondingly high sense of political efficacy compensated for these obstacles to participation and helped to produce the modern turnout peak that occurred in 1960.

Since 1960, though, voters have lost their previous sense of partisan identity and political efficacy. The American population has become more mobile, more single, and on average younger, all voter traits that tend to lower turnout. But Americans have also become better educated and more prosperous, which should increase turnout. In 1960, for instance, about half the voting age population had less than a high school education; by 1980 only 26 percent fit into that category, and the number with 16 or more years of education had nearly doubled. So changing demographic traits could not fully explain turnout decline; changing attitudes, according to Teixeira, were the key.

Teixeira reports that in 1960 only about 15 percent of the voting age population reported agreeing with two standard statements used in surveys: "People like me don't have any say about what the government does" and "I don't think public officials care much what people like me think." By 1980 about 32 percent of the population agreed with both statements, while the proportion who disagreed dropped from 61 to 34 percent. The percentage reporting strong partisanship dropped from 36 to 26 percent.

Teixeira proposes that these attitudinal changes resulted in large part from the turmoil of American politics since 1960 and, to a degree, from the rising influence of the broadcast media. John F. Kennedy's assassination, the Warren Commission, the Chicago riot of 1968, George Wallace's third-party run in 1968, the assassinations of Robert Kennedy and Martin Luther King, Jr., Vietnam, Watergate, Vice President Agnew's difficulties with the law, President Nixon's resignation, Ford's pardon, Carter's ineptitude -- these events apparently left voters skeptical and uncertain, without strong partisan identities or a sense of political efficacy. Increasingly, media campaign professionals took control of the interpretation of events, and while television became more important, the percentage of Americans who read newspapers declined. Watching TV does not appear to be a perfect substitute for reading. When people read less, they are more likely to find politics confusing. The net result of all these changes was lower turnout.

The strength of Teixeira's analysis is conceptual and methodological. He takes into account the demographic approach of Wolfinger and Rosenstone and the work of others, such as Abramson and Aldrich, who emphasize voter attitudes. In this new synthesis, political disorientation resulting from turmoil and from a decline in campaign newspaper reading overwhelmed the "upgrading" effect of demographic changes, such as greater education. The idea that turnout decline since 1960 reflects the erosion of an earlier "system," the system of the 1950s, is a coherent way to make sense out of the diverse demographic, electoral, and political facts of the last three decades.

The Rise of Strong Parties

But where did Teixeira's "system of the 1950s" come from? Why did political parties evolve into organizations that did not work hard to mobilize voters? What were the origins of personal registration and other electoral practices that increase the "costs of participation"?

Burnham and his school trace the collapse in turnout since 1960 to a long historical shift in electoral politics. By the mid-nineteenth century, the professional party politicians who began to revolutionize American politics in the 1820s were producing an average presidential turnout of 74 percent, up from 25 percent at the beginning of the century, when the electorate was much more limited. That achievement was all the more remarkable in light of rapid population growth. Turnout in non-presidential elections was also very high, apparently averaging about 68 percent. Through torchlight parades, festivities, and marching companies, party professionals created a politics that made partisanship the crucial determinant of an adult's political identity During political campaigns they involved entire communities of men, women, and children in a continuous, public display of partisanship.

The Civil War -- in part a war between the Democratic and Republican parties -- only deepened the hold of partisanship. In post-bellum decades partisan identity was so strong and deep in Northeastern and Midwestern states that political independence in an adult male was widely considered effeminate. A wildly partisan press reinforced such attitudes. As Kleppner argues, ethnicity and small-town and religious values also reinforced partisanship, since parties, at the state level, often consciously sought to appeal to different religious and ethnic groups by staging legislative quarrels over temperance, parochial education, and Sunday closing laws. Not surprisingly, presidential turnout reached record highs during these decades, between 78 and 82 percent, even as the voting age population expanded.

While there was vote fraud, most analysts do not believe it was so widespread from 1840 to 1896 as to account fully for the difference between nineteenth- and twentieth-century presidential turnout. Indeed, since parties were competitive, they had strong incentives to monitor each other and to keep fraud to a minimum.

Reform and Retrenchment

By the 1890s three key groups came to see this highly participatory political system as dangerous. Because American electoral democracy so effectively mobilized ordinary people, it had always potentially threatened concentrations of wealth. That potential threat became more palpable at the end of the nineteenth century as disaffected economic groups, such as the Knights of Labor and farmers' alliances, turned to electoral politics, culminating in the Populism of the 1890s.

To antiparty reformers and to Protestant, middle-class Americans, the ubiquity of patronage and the emphasis on spectacle and display also seemed a threat to rational government. They wanted to reduce the role of parties and rely more on disinterested, nonpartisan administration to cope with the strains of urban life, industrial disorder, and immigration.

Finally, to conservative Southerners, a vigorous, unfettered party politics endangered the stability of the South's social hierarchies. From 1868 to 1892 both white and black presidential turnout in the South was at least as high as it is now and probably higher, despite violence and other efforts to restrict turnout. The Populist strategy of building a class-based, cross-racial coalition of poor farmers threatened conservative Democrats and their economic allies.

Through gradual changes on a number of fronts, the groups that were dissatisfied with high participation prevailed. In the pivotal 1896 election, the Democrats embraced some of the Populist rhetoric but lost the White House for nearly two decades. The ensuing realignment left the Democrats strong inside the South, but Republicans strong in every other region, and as a result created enough regional one-party dominance to reduce popular interest in politics, particularly state and local elections. The reduced stimulus of less party competition weakened the hold of what Kleppner calls "party norms" on the electorate. Turnout dropped.

The elections of 1896 also set the stage for attacks on earlier electoral traditions. The sway of the two parties in their different regions made it easier to change the rules of electoral politics. In the South, after the collapse of Populism, Bourbon Democrats were free to revive white supremacist violence and to push blacks out of politics. But the new rules they imposed, including poll taxes and literacy tests, excluded poor whites as well.

Outside the South, new rules also made participation more costly. Legislatures established personal registration during workdays. At that time workers had neither an eight-hour day nor an hour off for lunch. Between 1900 and 1930 the percentage of counties outside the South with personal registration jumped 72 percent, according to Kleppner. Nor did legislatures require registration opportunities to be fairly distributed by neighborhood. As Piven and Cloward stress, personal registration depressed worker presence in politics, so that rational politicians increasingly directed their appeals to middle-class concerns. In turn, the absence of populist or collectivist appeals continued to discourage worker involvement in politics until the New Deal.

ONE-PARTY POLITICS in the states also heightened the attractiveness of Progressive reforms aimed at weakening parties further. These new provisions for referenda, recalls, party primaries, and nonpartisan elections changed the previous partisan simplicity of politics. They also, if unintentionally, raised the "information costs" of political involvement.

McGerr convincingly argues that politics became more culturally distant from ordinary voters. Party politics once physically involved "the people" in floats, parades, and public gatherings lasting for days of political song and speech. But the new style was more remote. It was an "advertised politics," consciously modeled on mass marketing techniques. In their private worlds, voters would presumably ponder their choices as voter-consumers. The press also changed. Now nonpartisan papers responsibly arrayed facts about politics before a passive electorate.

These changes created a new political context for voting. The handful of Northeastern and Midwestern states and cities where political machines remained had diminishing influence. The machines were isolated remnants of the nineteenth-century system. The addition in 1920 of millions of relatively apolitical female voters sharply depressed presidential turnout to roughly 49 percent in 1920 and 1924. However, the drama of the 1928 Smith-Hoover contest, followed by the New Deal, rekindled political passions, bringing presidential turnout up over 62 percent in 1940. But the New Deal left intact the Southern regime, the registration rules obstructing participation, and a lower level of partisanship in the population as a whole.

Nonetheless, from the 1940s to the early 1960s, the parties again resembled the "team" parties of the nineteenth century. The alliance between Democrats and organized labor's political action committees introduced a new approach at the national level to the mobilization of voters. But since the early 1960s, the Democratic-labor alliance has been undermined by labor's growing weakness, and "advertised politics" has become increasingly dominant. While serious issue differences between the parties persist, campaigns are obsessively organized around the promotion of a candidate's persona, constant fund-raising, the development of campaign momentum by paid consultants, and the use of broadcast media. Paradoxically, the permanent campaign, in Sidney Blumenthal's phrase, has demobilized the electorate.

The work of Burnham, Kleppner, McGerr, Piven and Cloward, and others thus casts a searching light on the larger historical conditions that explain why postwar American voters would respond to confusing political events and to television by losing their sense of partisanship and political efficacy The party identification and sense of efficacy that characterized the "system of the 1950s" now look relatively fragile. After all, the Civil War, the turmoil of the post-bellum decades, and the Compromise of 1877, when Democrats and Republicans brokered a presidential election behind closed doors, did not shock Americans into weaker partisanship and political disorientation.

Ideology and Democratic Decline

Do conflicts over issues and ideology have anything to do with the collapse in turnout since 1960? Burnham and many working in his tradition argue that since the New Deal, and especially since the 1970s, the Democrats have moved to the right, abandoning both working-class and middle-class Americans concerned about corporate power and government provision of economic security In response, these people have stopped voting, or they have never started when they have come of voting age. In "The Turnout Problem" and "The Eclipse of the Democratic Party," Burnham explains the move to the right as the result of the inherent instability of a center-left party in a weakly politicized market society In Right Turn, Ferguson and Rogers emphasize pressure from business on the Democrats to move to the right to lower the cost to business of welfare-state measures in the face of international competition.

Yet, as voters recognize, the philosophies of the two parties are different. The Democratic Party did not turn less liberal; it embraced a different version of liberalism that has demobilized its potential electorate. Ira Katznelson and Margaret Weir argue that the Keynesian approach adopted during the Kennedy-Johnson era, which emphasized market-led growth rather than political alleviation of unemployment through increased public spending, and the Great Society, which conceived poverty as a residual, largely racial problem in an otherwise healthy economy, undercut a potential class-based alliance among black and white voters. Democrats targeted social policies on the supposed few unable to get into the mainstream education, labor, and housing markets because of poverty or racial discrimination.

This renewed emphasis on welfare policy departed from the approach of Roosevelt and Truman. Rather than offering protections to the majority, the Democrats now seemed chiefly concerned about the minority below the poverty line. Much of the increased non-voting may simply result from the Democrats' letting down their historic working- and middle-class constituencies, not by moving right, but by substituting a new approach to welfare for the old one. The Democrats fumbled their chance to rebuild their majority coalition.

THE DEMOCRATS MIGHT NOT have suffered politically if the economy had continued to perform well. But when inflation, the deficit, and competitiveness emerged as issues in the 1970s and 1980s, the Democrats could no longer present their social policies as essential for prosperity Rather, they have presented them as moral imperatives but fiscal luxuries. Such policies were affordable in an expansive era, but not in times of austerity. The Democrats have forgotten how to attract key constituencies by making arguments other than "compassion." Consequently, they have proved unable to resist calls for less government, budget-balancing, and deregulation. Fighting to retain the social policies they created has seemed irresponsible at worst and backward-looking at best, if not impolitic in the face of a tax revolt.

The Democrats' party organization has deepened their difficulty in developing new policies that might rebuild their historic constituencies. The national Democrats are now a fairly cohesive legislative party. They have organized themselves around their congressional power, particularly in the House, establishing caucus government and a firm grip on the committees. But to be a legislative party requires congressional Democrats to get reelected, and to do that they need "permanent," well-funded campaigns and predictable electoral bases. These imperatives make it risky and difficult to adopt new approaches to social and macroeconomic policy. The permanent congressional campaign absorbs time and energy and compromises policy commitments. Building a war chest primarily from well-heeled donors to scare away potential challengers is not an activity that encourages deep reflection on social change.

Implications

Can the turnout decline be reversed? Some observers think it cannot, except under very unlikely conditions. Others are more optimistic and activist. The "voter traits" approach implies that raising turnout may be virtually impossible, especially if the attitudes Teixeira describes have taken on a life of their own. The continuing turnout depression in the last decade, after the political turmoil of the 1960s and 1970s subsided, gives little basis for optimism. The historical-structural approach is also discouraging. The evolution of electoral politics may have permanently reduced the participatory potential of American electoral politics. Party politics, nineteenth-century style, is clearly impossible. In the present context, rational politicians will not agree to give up their candidate-centered approach to elections and submit themselves to the discipline of an inner circle of national party leaders. Social democratic politics, approximated by the Democrats' alliance with the CIO between 1936 and 1948 and the AFL-CIO in the 1950s and 1960s, appears unlikely. Many Democrats now see labor as too weak to be much good to them; political professionals now widely regard Mondale's alliance with labor in 1984 as a major cause of his resounding defeat.

The Piven and Cloward approach, with its emphasis on the demobilizing consequences of personal registration, does offer a clear prescription: scrap all personal registration and other unwarranted barriers to voting. Approaches that emphasize the effect of the Democratic Party's moderation also suggest a simple solution: go left. Yet there are good reasons for doubting that these ideas would have much net impact.

Piven and Cloward have tried to carry out a strategy based on their understanding of the causes of low participation. They have founded the Human/SERVE Campaign (Human Service Employees' Registration and Voter Education Campaign) to push for simpler registration procedures and to offer registration assistance to clients at welfare offices, motor vehicle bureaus, and other public agencies. Proposed federal legislation would encourage "mail-in" registration and require the states to provide "motor-voter" registration (enabling citizens to register to vote when obtaining or renewing a driver's license).* The total, five-year cost of this reform, according to an estimate from the Congressional Budget Office would run between $215 million and $250 million.

Currently, several states and the District of Columbia have strong forms of motor-voter registration, combining the two kinds of registration on the same form or containing a "prompt question" on a motor vehicle form that triggers voter registration assistance by the registry clerk. About twenty other states are actively considering similar programs.

The Human/SERVE campaign for registration reform presumes that once registered, people are very likely to vote. About 80 to 85 percent of registrants vote, Piven and Cloward believe. Some data, however, suggest that voting by registrants is down 15 percent since 1960. But Piven and Cloward make a convincing case that this apparent drop in registrant voting results from failures by state election officials to purge registrants who have moved or died. Census surveys continue to show a strong link between registration and voting. So getting rid of legal obstacles to voting should significantly increase turnout, just as Piven and Cloward claim. The increase might be as high as 11 percent of the eligible electorate, or about 20 million voters.

Easing registration procedures is surely one step toward fixing the turnout problem. But changes in other national policies are necessary as well to alter the sense of inefficacy and alienation that undergirds non-voting. Voting turnout might increase if we pursued a new generation of policies that fostered a sense of civic membership. Universal worker training and retraining opportunities might restore a link between the citizen as worker and the citizen as voter. Improved primary and secondary schools would help restore voter confidence in public institutions generally. I am not offering a policy platform, only emphasizing that policy ideas should be weighed for their contribution to rebuilding the sense of civic efficacy that invites participation in politics. Registration reform alone, if Teixeira is right, will not motivate people to vote.

Non-voters' policy preferences are, moreover, a slippery question. Piven and Cloward want registration reform not for aesthetic reasons; they have a political agenda. They want to bring poor people back into politics to move the Democratic Party to the left. Yet their critics, including Teixeira, have convincingly pointed out that even extraordinarily high turnout rates among poor, Hispanic, and black adults eligible to vote would not have won the 1988 election for Dukakis and Bentsen.

Teixeira argues that the broad, downward trend of non-voting is not limited to the poor; it is far more widespread. To be sure, this point is irrelevant to the Piven and Cloward strategy. Simply winning elections for the Democrats is not their goal. They want to change the agenda of public debate and get the parties to disagree over social and economic policy in new ways. Such change, in turn, would open up American politics to previously foreclosed possibilities. In such a context, turnout rates might surpass those of the 1930s and 1950s.

Implicitly, though, Piven and Cloward are assuming that non-voters are likely to be farther to the left than voters and more likely to vote Democratic -- if Democrats begin responding to economic needs of current non-participants. Surveys do not show strong policy differences, however, between voters and non-voters. Non-voters, on balance, are only mildly pro-Democratic. There may well be serious measurement error in surveys of non-voters. Non-voters who cooperate with surveys are probably not typical of the larger population of non-voters, who are cynical and politically alienated. But this survey evidence cannot yet be dismissed. It has striking implications, as well, for the position that turnout will increase if the Democratic Party moves left or if it develops a different approach to social policy. Non-voters do not secretly have intense partisanship, held in reserve until some change in the political agenda or some invitation to participate unlocks their passions. They do not vote in part because they lack strong political affiliations. Partisanship among the non-voters would have to be constructed, and that will be difficult.

The Future of Non-Voting

If the past is a guide, transforming our participatory structure will require a debate about turnout. Years of intense debate about the political functions of parties preceded the last, great transformation of party politics. Between the 1870s and the 1900s, such magazines as the Nation, the Atlantic Monthly, and Harper's Weekly discussed the dangers of strong party politics. The editors of large newspapers grew hostile to parties as they responded to the emergence of local reform organizations, such as the City Reform Club established in New York City in 1882 by Theodore Roosevelt. A rising class of social scientists and academic leaders, gathered into professional organizations, also attacked parties. In time, too, military officers (who had organized themselves into their own special associations) turned against strong parties, seeing them as obstacles to the professionalization of the armed services. The hostility to democracy that informed much of this late nineteenth century debate now seems offensive. Yet we live, ironically, in a world created in part by such public argument against the old electoral regime.

There will be no contemporary debate if the collapse in turnout since 1960 is taken lightly. We cannot say, as some have, that non-voting reflects complacency and even contentment among the population. As Piven and Cloward drily remark, "no one has satisfactorily explained why 'the politics of happiness' is so consistently concentrated among the least well off ." Others say that non-voting prevents fascism and demagoguery, yet there is no tradition in America of caesarist politicians succeeding on a national scale. Finally, certain writers insist non-voting is healthy because high levels of participation overload democratic government. This argument appeals to those who see the general public as made up of grabby people willing to bankrupt government with insatiable demands. But some of the grabby people are those with the power to abuse deposit insurance or to alter the tax code to their benefit. Broader electoral participation would curb such self-seeking demands on our scarce public resources. In the end, it is awfully hard to contend that non-voting is good for democracy

The "party of voters" would do well to appreciate its self-interest in bridging the divide that separates it from the "party of non-voters." While turnout has dropped, the demands on government have grown in this century If democratic government is to take on the tasks demanded of it, it needs to seek out the voices and the votes of people who now feel they simply do not count.

As in the past, the expansion of participation can help to turn the subjects of administration into citizens capable of self-government. It can create a hardy sense of membership in a political community. Our electoral politics now fails to realize the empowering possibilities of democratic life. We need to recreate the popular ownership of electoral politics that we once had -- indeed, that America pioneered. Otherwise, the likely low turnout in the 1990 and 1992 elections will be accepted as routine, and America will increasingly look like Schattschneider's "broadly-based oligarchy."