5
THE ABSENT VOTER
If the people fail to vote, a government will be developed which is not their government.… The whole system of American Government rests on the ballot box. Unless citizens perform their duties there, such a system of government is doomed to failure.
—PRESIDENT CALVIN COOLIDGE, 1926
In the presidential election of 1888, W. W. Dudley, the treasurer of the Republican National Committee, distributed a circular on committee stationery that instructed party operatives in Indiana to “divide the floaters [bought voters] into blocks of five, and put a trusted man with necessary funds in charge,” being sure to “make him responsible that none get away and all vote our ticket.”1
Although Dudley denied having engaged in the actual buying of votes, this incident fueled the perception of an American politics afflicted by an epidemic of voter fraud. The New York Evening Post typically claimed in 1888, “There are thousands of voters in both parties who wait in every important election to be paid, even to vote the ticket of their choice. A high Democratic authority has placed this number at one-third in each party in many sections of the state.”2 Despite anecdotal rather than systematic and analytical evidence of voter fraud, reformers still perceived that vote-buying and other illicit practices corrupted elections in late nineteenth-century America and compelled respectable citizens to impose countermeasures.
Reformers of the time found it difficult to reconcile their conflicting instincts to simultaneously expand and purify the vote. The democratic impulse led to the innovations of primary elections, the direct election of senators, votes for women, and the initiative, referendum, and recall. Yet voter turnout declined gradually in the early twentieth century, partly through the countervailing effect of reforms to protect the vote from fraud and abuse: government-issued secret ballots, personal registration requirements, literacy tests, poll taxes, at-large elections, and the disenfranchisement of aliens, felons, and Native Americans.
The problem of nonvoting was both institutional and individual as Americans turned away from politics in the early twentieth century. The irony is that Americans reduced their voting at a time of expanding government influence over people’s lives, and growing affluence and educational attainment, characteristics typically associated with heightened participation in politics.
Revisions in the system of voting and elections fit within a broad context of social and political change. Highly ethnocentric notions of nation and citizenship had guided the shaping of the franchise since the early republic. These concerns expanded in the late nineteenth and early twentieth centuries when millions of largely Catholic and Jewish “new immigrants” from eastern and southern Europe poured into America’s cities, became voters, and found themselves recruited by allegedly corrupt urban political machines. The same impulses that drove white, Protestant Americans to back Prohibition, immigration restriction, and campaigns against drugs, smut, and vice motivated efforts to purify politics.
These anti-pluralists sometimes joined forces on political reform with secular activists dismayed by the baleful influence of a new force in American life: the wealthy corporations, whose money greased the gears of party organizations. A New York Times editorial in 1905 denounced “the trusts and great corporations” that “continue without hindrance to furnish large sums of money for the use of party committees in electoral campaigns.… If the trusts can elect Congresses and Presidents with their money they can at pleasure guide the activities that make the laws or palsy the arm that enforces them.”3
Until the 1880s, Americans routinely voted openly by voice or, more frequently over time, through preprinted party ballots, emblazoned with distinctive different colors, that they handed to election judges at the polling place. Elections turned less on persuading doubtful voters and more on drilling and mobilizing partisans who could be readily identified by the color of their ballots. Brawling, rioting, and even murderous violence marred voting at the nineteenth-century polls. “In the middle decades of the nineteenth century, eighty-nine Americans were killed at the polls during Election Day riots,” writes historian Jill Lepore. On one gruesome election day in Louisville, Kentucky, in 1855, at a time of anti-immigrant agitation, rioters massacred twenty-two voters, mostly Irish and German immigrants.4
For reformers, party ballots and the mass mobilization of voters confirmed George Washington’s warnings about divisive party politics that subverted the historic ideal of autonomous voters devoted to the common good. Reformers further objected to party ballots as a ripe source of fraud. Unscrupulous politicians could buy voters and verify that they stayed bought by casting the right color ballot. American opinion-makers had fretted about electoral fraud from the earliest days of the republic. But the campaign against fraud became urgent as the growing, heavily immigrant working class mobilized by political machines seemed to validate Alexander Hamilton’s warning about the rise of purchasable votes in America’s cities. As historian Melvin G. Holli explained, “The inferior moral fiber of foreign-born and lower-class electorates and their representatives, and the dereliction of upper classes was one of the most pervasive themes in the early Conferences for Good City Government” sponsored by the National Municipal League.5
Reformers pushed for states to adopt the secret ballot pioneered in Australia in 1856. An 1889 editorial in the New York Times said that the abolition of open party ballots would “make it impossible” for fraudsters “to tell whether any single vote that they had purchased or contracted for was really given in the way they wish,” and “would completely destroy their base but profitable occupation.”6 The so-called Australian ballot, printed by governments, not parties, would include the names of candidates. It would be distributed only at the polling place and cast in private by each voter. By 1908, thirty-nine of forty-six states and the territory of Arizona had adopted some form of the secret ballot.7
The candidate-centered ballot printed a heading for each office, with candidates listed below, typically identified by party. The party-centered ballot printed a separate heading for each party, with its candidates for each office listed below. It most closely resembled the ballots previously distributed by the parties. It facilitated straight-party voting, making it difficult for voters to split their tickets. However, some states with candidate-centered ballots also included an option to vote for a straight-party ticket. By 1908, most states were using either the party-centered ballot or the candidate-centered ballot with a straight-ticket option. Although states eventually moved away from party-centered ballots, some still maintained a form of straight-ticket voting, and third parties still had to struggle for ballot spots.8
Reformers alone could not have engineered the turn to the Australian ballot without backing from Democrats in the South and primarily Republicans in the North. The nation’s major parties had something to gain from a transition to official, secret ballots that spared them the expense of printing and distributing their own ballots and made elections more orderly and predictable than before. The Australian ballot system, moreover, placed a disparate burden on third parties and independent candidates or rogue party factions who lacked guaranteed places on official ballots. Each state adopted its own ballot access laws, which the two major parties, with their guaranteed places on the ballot, typically wrote to impose burdensome deadlines and signature requirements for third parties and independent candidates. In the presidential election of 1936, for example, five nationally organized parties—Socialist, Communist, Union, Prohibition, and Socialist-Labor—failed to get ballot positions in nine to twenty-nine states.9
The adoption of the Australian ballot eventually dampened the turnout of voters, a generally welcome result for purity crusaders who hoped to staunch what one reformer called “the admitted evils of ‘tramp’ suffrage” and a “licensed mobocracy.” The secret ballot weakened the incentive for partisans to pay or compel voters to back the party ticket, when operatives could not readily verify voter compliance. Yet the secret ballot may have prompted a new kind of corruption, paying likely opposition voters not to vote, which operatives could easily verify, thereby eliminating the problem of bought voters who would not stay bought. James K. McGuire, the chairman of the Executive Committee of the New York State Democratic Party, charged in 1900 that “under the new ballot law you cannot tell how a man votes when he goes into the booth, but if he stays at home you know that you have got the worth of your money.” Paying potential voters to stay home, he said, “is the only sure way to buy votes.”10
Likely much more important than voter payoffs, whose numbers remain speculative, the lack of verification reduced the incentives for parties to mobilize the ranks of purportedly loyal voters. It made voting a dull, industrial-style event, even as other forms of diversion emerged in the early twentieth century through the radio, movies, amusement parks, and automobiles. Studies by political scientists have found that by the early twentieth century, secret ballot laws had an independent effect of depressing voter turnout, notably in states that relied on candidate-centered ballots. The lack of voter participation fell most heavily among lower-income voters, which was precisely as the reformers had intended.11
The adoption of secret ballots led to the innovation of voting machines to make voting more private and secure than before. Cloaked behind a curtain, voters would pull the levers for their favored candidates, with gears in the machine preventing multiple voting. The counter inside the machine tallied the votes. By 1929, about half the states had adopted machine voting.12
To curb further the undue influence of urban political party machines and establish a bipartisan, scientific basis for government, many municipalities adopted a city manager form of government under which an elected council appoints a city manager responsible for the day-to-day administration of the municipality. From its inception in Staunton, Virginia, in 1908, 321 cities had adopted the city manager system by 1924, limiting the influence of voters over the operations of their local governments. Municipalities with the city-manager system tend to have lower turnout than other municipalities with direct voter control over government, due to the remoteness of the city-manager form from voter control.13
Cities and counties shifted from district-based elections for local legislative bodies to at-large elections in which all voters in the jurisdiction selected all members. This was yet another means for restricting the power of political machines and their immigrant and minority loyalists. At-large elections submerged minorities in white-controlled electorates and favored politics driven by money rather than low-budget tactics like door-to-door canvassing. By 1960, 60 percent of cities with a population of more than ten thousand people elected members of their councils citywide, while only 23 percent elected them from districts, some of which had African American and immigrant majorities. The remaining 17 percent had mixed at-large and district systems. The consensus of statistical studies indicates that ward elections through the mid-twentieth century provided minorities better opportunities to elect candidates of their choice than citywide or at-large elections, although the effect could be minimized by the racial gerrymandering of districts.14
As part of these initiatives to insulate local elections from the mobilization of voters by political machines, reformers decoupled the timing of local and national elections. The reformers extended a nineteenth-century practice of holding local elections on off-years and dates prior to the November elections for national and state positions. To diminish the influence of political parties, they also adopted in many cases officially nonpartisan elections for local offices, in which candidates competed without party nominations or labels. As with other reforms, however, party machines that believed their superior organizations would most effectively mobilize voters lent support to these reforms. Today, about 80 percent of U.S. cities hold municipal elections on days other than the national election day.15
Although the evidence does not suggest that nonpartisan elections by themselves reduce voter turnout, scholars have established that off-cycle elections substantially reduce the participation of voters. According to political scientists Zoltan L. Hajnal and Paul G. Lewis, “The existing evidence suggests that turnout in city elections may average half that of national elections, with turnout in some cities regularly falling below one-quarter of the voting-age population.” They conclude that “ultimately, important policy decisions are being made by local officials who are chosen by a relatively small and likely unrepresentative group of citizens.”16
Off-cycle elections expand both the influence of well-organized party machines and of nonparty groups, like labor unions, chambers of commerce, trade associations, and good-government organizations with special interests in local election outcomes. At the 1908 meeting of the National Municipal Association, Robert S. Binkerd of the Buffalo Municipal League foreshadowed the rise of independent groups in local politics. He noted that with the reform of local elections, “there is need, in every community, for a non-partisan organization which shall, primary after primary and election after election, present to the press and voters the facts concerning office-holders and candidates.” Of course, competing interest groups had their own conceptions of the relevant facts and different levels of resources available to influence elections.17
In 1955, Charles R. Adrian, a scholar of municipal politics, concluded that for low-turnout, off-cycle elections, “highly organized groups, whether of the nature of old-fashioned city machines or of special interest groups of any type, will thereby be able to control the government, for the lighter the vote the easier it is for such groups to win. They have a solid nucleus of dependable voters. A small turnout does not result in the same percentage distribution of the vote among the various segments of the population as would be found in a large turnout.” A recent study found that low-turnout, off-cycle elections disproportionately reduced the turnout of minority groups. Thus, an inherently discriminatory process is built into the system for electing most local officials in the United States, largely beneath the notice of the American people.18
To combat manifestations of alleged fraud and corruption that could not be cured by official secret ballots—repeat voting, voter impersonation, and voting the graveyards—reformers imposed a barrier between voters and their vote: personal registration requirements. Unlike the secret ballot, personal registration laws had a long history in the United States. Massachusetts began the practice in 1800, with a few other states soon following. But personal registration requirements revived in the 1860s and became nearly universal in the early twentieth century. The registration process was often intrusive and onerous. In Ohio, for example, in addition to age, citizenship, and place and term of residence, applicants had to attest to their place of birth, their occupation, the name of their employer if they had one, their marital status, and whether they were “the head of their family.”19
In the North, to throttle the urban machines, states targeted personal registration laws to the cities, often exempting rural areas where fraud was supposedly uncommon and voters and election officials often knew one another. A 1909 state court decision upholding a registration law aimed at New York City held that “it is a matter of common knowledge that the great centers of population offer peculiar opportunities for fraud and corruption in the conduct of elections, which can be prevented or minimized only by special regulations adapted to the conditions out of which these evils arise.”20
In most large cities, voters had to renew their registration annually, biennially, or within four years. Few states or municipalities had registration offices open all year, and most imposed stringent registration deadlines prior to an election. States frequently revised their registration laws, sowing confusion and erecting new hurdles for voters to surmount. A 1924 compilation found that substantial percentages of otherwise eligible urban voters remained unregistered, including 47 percent in cities with annual renewal requirements and 31 percent in cities with permanent registration. It is unknown how many of those unregistered city dwellers would have voted in the absence of registration requirements.21
Even as states largely abandoned re-registration laws by the mid-twentieth century, voters still risked exclusion through lengthy residence requirements for registration, early deadlines before elections, and the purging of registration rolls for failure to vote, even for periods as brief as two years. Some states automatically purged nonvoters, while others provided notices to registrants before removal. In every state, purged voters could re-register with appropriate documentation. A 1963 report by the President’s Commission on Registration and Voting Participation identified nonvoting, not fraud, as the major flaw in American elections. It found that restrictive registration laws disenfranchised millions of Americans. It recommended that residence requirements should not exceed six months, that registration should not end more than three or four weeks before Election Day, and that no registrant should be purged for failing to vote for any period of less than four years. The report also recommended that states should provide the option to register by mail. Most states, however, would not implement these recommendations until after enactment of the National Voter Registration Act of 1993. But the issue of restrictive registration requirements would continue to provoke litigation in federal and state courts.22
In the South, registration laws shifted disenfranchisement from the point of voting to the point of registration, erasing, not just restricting, the black franchise. Southern registration laws imposed poll tax and literacy test requirements and ceded discretion to all-white local officials for adding and subtracting registrants. Black leader Thomas E. Miller, a former member of Congress from South Carolina, dubbed the local registrar the “Emperor of Suffrage.” L. C. Coulson, a critic of the registration restrictions in the Alabama constitution of 1901, said, “Under the law as it now stands, white men steal the vote; if we adopt the new constitution, white men will steal the voter.”23
African Americans turned without success to the federal courts for relief from the registration discrimination in the South. In 1895, a panel of the Fourth Circuit Court of Appeals, presided over by Chief Justice Melville Weston Fuller of the U.S. Supreme Court, overturned a lower court decision that had permanently enjoined the enforcement of South Carolina’s 1894 restrictive registration law. Fuller’s majority opinion said that the courts should avoid entanglement in political issues and that any remedy must be found in the same southern political system that denied African American voting rights, not in the judiciary. The U.S. Supreme Court dismissed an appeal by the plaintiff as moot because the election in which he alleged a denial of his franchise had already taken place.24
In 1898, the U.S. Supreme Court turned away a challenge to the administration of Mississippi’s registration laws. A unanimous court rejected the plaintiffs’ claim that registrars used their discretion to block African Americans from registering. The court ruled that “discriminations by the general government or by the states against any citizen because of his race … must be the result of the constitution or laws of the state, not of the administration of them.” It found that Mississippi’s registration laws, which included literacy and poll tax requirements and a grandfather clause, “do not on their face discriminate between the races, and it has not been shown that their actual administration was evil; only that evil was possible under them.” The Court found that the registration provisions and their administration law targeted not race per se but only “the alleged characteristics of the negro race,” and reached both “weak and vicious white men as well as weak and vicious black men.” Five years later, the Supreme Court extinguished any hope of relief from discriminatory registration laws when it rejected black plaintiffs’ challenge to Alabama’s registration laws and their administration.25
Commentary from the first decade of the twentieth century reflected the failure of African Americans to win the restoration of their voting rights in the courts. “Under existing court decisions and present legal practice,” the Atlanta Constitution announced in 1904, “the southern states have lawfully accomplished the disenfranchisement of the Negro.” In a 1910 analysis in the Harvard Law Review entitled “Is the Fifteenth Amendment Void?” Arthur W. Machen, Jr., noted that the Supreme Court “has in no single instance held any state or federal statute or the act of any state or federal officer to be in conflict with the Amendment; and no case in that court can be found which would have been decided differently if the Amendment had never existed.”26
Southern lawmakers exploited the institution of the primary, designed to expand democracy, as another means for extinguishing black voting. In the solidly Democratic South, where victory in the Democratic primary was tantamount to election, states adopted majority-vote rules that required one-on-one runoffs unless a single candidate gained a majority vote in the initial primary. This requirement guarded against the possibility of nominating an African American because white voters split their ballots among multiple white candidates. Most southern states also restricted primaries to whites only. To avoid adverse court decisions against government-sponsored white primaries, the Democratic Party typically adopted the white primary as a party, not a state, institution.
For reformers who sought a purified American politics it was not enough to change the conduct and administration of elections. They would also purge the electorate of people they deemed unfit for the responsibility of voting, no longer women but now illiterates, felons, aliens, Native Americans, and insane or “feeble-minded” individuals.
Driven by the fear of voting by Irish and German immigrants, literacy tests originated in the North, led first by Connecticut in 1855 and then by Massachusetts in 1857. The literacy test, said an editorial in the Connecticut Courant, was aimed at the “miserable tribe of low, hack, partisan, demagogic, hollow, hypocritical democrats” who “would give the state into the arms of foreigners, or drunkards,… or irresponsible grog-shop scoundrels and poor-house rum-blossoms.” Other northern states, provoked by fears of immigrants and other untrustworthy voters, adopted literacy tests in the late nineteenth and early twentieth centuries.27
By 1924, eighteen of the forty-eight states had literacy tests in place, eight in southern states and ten in northern ones. The practical effect of literacy requirements on the opportunity to vote was substantial. According to the Illiteracy Commission of the National Education Association, there were 10 million American illiterates in 1924 and an additional 10 million semiliterate people. The army literacy tests in 1917–1918 found that 25 percent of men tested—nearly 90 percent of whom were white—were illiterate and another 5 percent were semiliterate.28
In the South, to disenfranchise black voters, common practice required of black applicants not just reading but interpreting documents like the state constitution or having to answer obtuse questions that no one could be expected to comprehend. In 2014, an instructor at Harvard administered the Louisiana literacy tests used for black voters through 1964 to a group of his students. The questions were so confusing and the time limit so constricting that not a single Harvard student passed the test.29
In the early republic, many states did not draw distinctions in their constitutions between inhabitants and citizens, and the federal government authorized suffrage for aliens who met property requirements in the Northwest Ordinance of 1789. When older states began replacing suffrage standards based on the status of property with standards based on the status of people, they backtracked on both black and alien suffrage. Still, new western states that needed population infusions authorized alien voting. In the 1840s, the territory of Wisconsin pioneered a compromise form of alien voting that granted voting rights to immigrants who declared their intent to became American citizens. In the acts establishing the Washington, Kansas, Nebraska, Nevada, Dakota, Wyoming, and Oklahoma Territories, Congress enabled these territories to institute noncitizen voting, and most did so. It did not, however, authorize such voting rights for the territories of California, New Mexico, and Utah, acquired after the Mexican War of 1845, perhaps because the treaty settling the war gave Mexicans the choice to remain citizens of Mexico or to become citizens of the United States.30
Opportunities for alien voting peaked again in the 1870s when most of the states incorporated into the union after 1800 and most territories permitted some form of alien suffrage. For at least some years in the nineteenth century, aliens could, in principle, vote in about forty states. The wave of nativist reaction to the new immigration of the late nineteenth and early twentieth centuries began another reversal of alien voting. In about nineteen states at the turn of the twentieth century, alien suffrage survived often in the modified Wisconsin form of requiring a declaration of citizenship intent.31
These provisions for limited alien voting toppled during the next two decades amid anti-immigrant sentiment. Buttressed by popular reaction and the prevailing “racial science” of the time, which found new immigrants mentally and morally inferior to older stock Americans, Congress enacted temporary immigration restriction in 1921 and a permanent law in 1924. The Johnson-Reed Act of 1924 limited European immigration to about 150,000 individuals per year with nationality quotas based on the origins of the American population in 1890 that heavily favored western and northern Europe. The act required entry visas with photographs—another facet of control—and excluded Japanese and other Asians. Four years later, according to political scientist Leon E. Aylsworth, “For the first time in over a hundred years, a national election was held in 1928 in which no alien in any state had the right to cast a vote for a candidate for any office—national, state or local.” Congress repealed the quota law of 1924 in 1965, but since that time no state and only a few scattered localities have authorized alien voting.32
The extinguishing of noncitizen voting was coupled with restrictions on opportunities for certain aliens to obtain American citizenship and thus qualify for the ballot. Federal court decisions strictly construed an 1870 law that opened opportunities for naturalization only to those identified as “a white person” or person of “African descent.” The unspecified meaning of “white” in the law left open the still unsettled question of how to distinguish “whites” from “nonwhites.”
The U.S. Supreme Court interpreted the word “white” according to the supposed commonsense views of native-stock Americans, which excluded people of Asian and most Middle Eastern descent. Immigration officials struggled to decide whether Hispanics who appeared to be of mixed European and indigenous descent qualified as “white.” In 1923, the Supreme Court ruled that even “a high caste Hindu, although of the Caucasian or Aryan race, is not a white person within the meaning of the naturalization laws,” which “were to be interpreted in accordance with the understanding of the common man.” Unlike Asians, the Court recognized that “immigrants from Eastern, Southern and Middle Europe, among them the Slavs and the dark-eyed, swarthy people of Alpine and Mediterranean stock,” merged into the white population “and lose the distinctive hallmarks of their European origin.” Four years later, without a hint of irony, the Court ruled that a child born in Mississippi of Chinese ancestry could be classified as “colored” and consigned to legally segregated black schools.33
Through the Dawes Act of 1887, the Curtis Act of 1898, and the Burke Act of 1906, the federal government began a process of assimilating Native Americans into the mainstream of American life. These laws granted American citizenship to Indians who had accepted allotments of land from the government and agreed to live separately from their traditional tribes. The acts transferred millions of acres of land from Indians to white settlers in the process. From 1887 to 1934, lands owned by Native Americans dwindled from 138 million to 48 million acres. The Indian Citizenship Act of 1924 granted federal citizenship and, in principle, the right to vote to all American Indians. However, by 1938, seven states with substantial Indian populations still effectively prohibited them from voting. This was achieved, for example, by denying voting rights to Indians who maintained their tribal affiliations, lived on reservations, or were considered under the “guardianship” of government. Some states excluded from voting “Indians not taxed,” mimicking a clause of the original Constitution that excluded such individuals from being counted in the U.S. Census.34
In Arizona, where nearly one-sixth of all Native Americans resided, the state supreme court ruled in 1928 that as “wards of the United States,” Arizona’s reservation Indians fell under the state’s prohibition of suffrage to any “persons under guardianship,” placing these citizens in a similar category to those classed by the state “non-compos mentis, and insane.” The court found, “The man who for any reason is exempt from responsibility to the law for his acts, who cannot be trusted to manage his own person or property, certainly as a matter of common sense cannot be trusted to make laws for the government of others, and placing him under the guardianship of another conclusively establishes that incapacity” for suffrage.35
Barriers to Native American voting began to break down through statute and court rulings in the 1940s. In 1948, the Arizona Supreme Court reversed its 1928 decision and ruled that in excluding “persons under guardianship” from voting, the framers of the Arizona constitution did not intend to include Indians indiscriminately in that category. Rather, they “had in mind situations where disabilities are established on an individual rather than a tribal basis.”36 Simultaneously, a three-judge federal court ruled that under the Fourteenth and Fifteenth Amendments, New Mexico, another state with a substantial Indian population, could not indiscriminately disenfranchise “Indians not taxed.”37 However, statutory prohibitions on Indian voting were not repealed until 1950 in Idaho, 1951 in South Dakota, 1957 in Utah, and 1960 in Minnesota.38 Even in the post-repeal era, Indian voting rights remained precarious as states used mechanisms such as at-large election systems and gerrymandered redistricting plans to limit the impact of the Indian vote, prompting litigation under the Voting Rights Act of 1965 and its amendments.39
In 1923, political analyst Frank Kent debunked the myth that “universal suffrage” prevailed in America. Rather, Americans “are hedged about by voting requirements, rigid in some States, lax in others, but all designed to do two things—first, promote the orderly and honest conduct of election, and, second, eliminate certain citizens who do not measure up to State standards.” The laws excluded not only aliens but many classes of citizens: felons, the “feeble-minded” and insane, soldiers who lacked permanent residence, Indians on government reservations, paupers, delinquent taxpayers, and the uneducated. He noted that voting restrictions in most northern states are “a natural protective” against the “tides of immigration,” which impresses on them “the desirability of keeping the vote out of the hands of the wholly illiterate and unworthy.” Restrictions in southern states resulted primarily from “the dread of the negro vote.” An untainted motivation to purify politics, he said, prevailed in only a few states with small black and immigrant populations. In states such as New Hampshire and Connecticut, the people “simply believe in restricting the vote to those citizens who have at least a certain degree of intelligence or some small stake in the community.”40
Kent noted with dismay that fewer than half of American adult citizens voted in the presidential election of 1920, and accurately pointed to the collapse of turnout in America since the late nineteenth century. From 81 percent in the first post-Reconstruction election of 1880, the percentage of people eligible to vote by state laws pertaining to sex, age, and citizenship who turned out in the presidential contest had plummeted to 49 percent in 1920, with most of the decline coming after 1896. Turnout for congressional elections in midterm years declined comparably, falling from 66 percent in 1882 to 36 percent in 1922, with again most of the decline occurring after 1896.41
Kent claimed that the dismal turnout rates of his time could not be explained by suffrage restrictions alone, but rather that “the great bulk of those who do not vote, are not those who cannot vote, but those who will not.” Kent was writing at a low ebb of turnout in the United States and in the early years of women’s suffrage, when women’s turnout trailed behind men’s. Study of the broader chronology shows that at least in the North, Americans did not stop voting, contrary to Kent’s implication and much scholarly opinion. Turnout in the North rebounded robustly in 1928 after women largely caught up with men in their political participation. Northern turnout in the presidential election hit 66 percent in 1928 and averaged 69 percent through 1964, only 15 percentage points below the 84 percent northern average in its 1880 to 1900 peak. Changes in the institutional context of voting are sufficient to explain most of this relatively modest drop in northern presidential turnout, without resort to broader cultural and political explanations.42
The state of North Dakota provides an instructive, controlled study on the influence of structural changes in voting access on voter turnout. Unique among the states, North Dakota never had a personal registration requirement or literacy test, and it adopted the secret ballot prior to its first participation in presidential elections. Thus, any changes in voter turnout in presidential elections from the late nineteenth century to the early to mid-twentieth century occurred independent of changes in electoral rules. Despite the state’s sparse population and its late entry into the union in 1889, absent any institutional changes in registration or voting, presidential turnout in North Dakota averaged 72 percent from 1928 to 1964, essentially unchanged from the 74 percent average from 1892 to 1900.
Scrutiny of turnout in midterm congressional elections adds shading to this picture. In the North, congressional turnout rebounded only modestly, averaging 51 percent from 1930 to 1966, compared to 70 percent from 1882 to 1902. Thus, the shrinkage in northern turnout primarily impacted so-called core voters, those who habitually participated in elections, even without the allure of a presidential contest. This decline in core voting cannot be explained by institutional changes alone; it also reflects the erosion of the mass mobilization by political parties of their loyal followers, the rise of alternative means for influencing governance through organized interest groups, and the proliferation of unelected administrative agencies at every level of government. In North Dakota, despite a lack of institutional changes, core voting declined by about 10 percentage points from 1930 to 1966 as compared to the period from 1890 to 1902. The loss of core voters reduced turnout especially among minorities, young people, and the less affluent, not just in the South but also in the North.43
Through the mid-twentieth century, the South dragged down national turnout rates, lagging far behind the North in both presidential and congressional elections. From 1928 to 1964, southern presidential turnout averaged only 31 percent, well less than half the northern average of 69 percent. For midterm congressional contests, southern turnout during this period averaged merely 15 percent, less than a third of the northern average of 51 percent. This southern turnout decline reflected lack of general election competition in this Democratic Party–dominated region, as well as the uniquely restrictive registration and voting laws that blunted insurgent movements and disenfranchised some low-income whites and most African Americans.
With American democracy by the early twentieth century suffering from southern exclusionary tactics and absent voters, much of the outside world was moving in the other direction toward an expanded democracy. For the first time, the United States was falling behind other democratic nations in opportunities for voting. Significantly, by the early twentieth century, the United States had reverted not to a white man’s republic but, with the enfranchisement of women, to a white citizen’s republic. State laws and constitutions had extinguished alien voting. Native Americans were largely excluded from full political participation, and so too were African Americans in the South, where most black people still resided. In the North, literacy tests and personal registration requirements limited voting opportunities for nonwhites and less-affluent citizens generally.
Across the globe, elites in many nations could not resist pressure from below for the suffrage without unacceptable repression. Yet democracy proved to be a fragile commodity that remained precarious worldwide, even as it survived imperfectly in the United States. In the “Golden Age” of democracy after World War I, the number of functioning democracies soared from a handful to some twenty-nine nations. Then, from 1922 to 1942, the number dwindled to only twelve, almost all of them in western Europe or among English-speaking nations such as Canada, the United States, and Australia.44