Professor William G. Ross, Cumberland School of Law, Samford University

The most complicated bit of governmental machinery which the modern world has to exhibit is that which is employed in the selection of the chief executive officer…for the United States…It is almost marvelous that any people should have preserved political unity for a century under such a loose and decentralized system of election of its chief magistrate.

John W. Burgess, “The Law of the Electoral Count,” 3 Political Science Quarterly 633 (1888).

Every four years for nearly two centuries, the United States has played the political equivalent of Russian Roulette when it has conducted a presidential election. This year, the gun finally exploded. How much injury it has inflicted on the body politic is not yet clear.

History demonstrates that this year’s electoral cliffhanger is not a freak accident but rather is merely the latest in a long series of remarkably close elections which often have ended in doubt and acrimony after considerable conflict. This history, which is briefly reviewed below, indicates that the turmoil over the present election was predictable and that a similar mess will almost surely occur again if the present system is not reformed.

Although the House of Representative’s selection of a president in 1824, the disputed election of 1876, and the close and questionable Kennedy-Nixon contest of 1960 have received much attention in recent weeks as news media have grasped for examples of other elections that have caused trouble, many Americans are unaware that a remarkably large number of other elections have produced close results that nearly created political chaos.

The very first seriously contested presidential election, in 1800, was thrown into the House of Representatives after an Electoral College tie between Thomas Jefferson and Aaron Burr that resulted from a double-balloting anomaly in the Constitution. Jefferson was elected only after much maneuvering. The 1800 election, produced one positive result (aside from Jefferson’s victory) insofar as it led to the 1804 enactment of the Twelfth Amendment, which eliminated double balloting.

The amendment retained, however, the curious procedure for election by the House of Representatives (voting by states) when no candidate receives a majority of the electoral vote. When no candidate received a majority of electors in 1824, the House elected John Quincy Adams even though Andrew Jackson had received a plurality of both the popular and electoral votes.

In four more elections during the next half century (1836, 1856, 1860), the shift of a relatively small number votes in a few states would have thrown the election into the House of Representatives. In six other elections (1828, 1840, 1844, 1848, 1864, and 1868), the shift of a small number of votes in a few states — fewer than 20,000 in the first four — would have resulted in the election of the other candidate. In most of these situations, the winning candidate would not have received a plurality of the popular vote. In three of these elections (1836,1844, and 1860), the final outcome was determined by the electoral votes of only one state, New York.

In 1876, the outcome of the election between Republican Rutherford B. Hayes and Democrat Samuel Tilden was in doubt for months as the result of the submission of double slates of electors in four states. In Florida, where both Republicans and Democrats alleged fraud, the governor certified Hayes electors, but a court ruled in favor of Tilden electors. In South Carolina, where fraud allegations also were rampant, Democrats submitted a slate of electors even though state officials had certified the Republican electors. In Louisiana, which had two governors,there were two competing canvassing boards and two conflicting sets of returns. And in Oregon, one Republican elector was challenged on a technicality. Congress eventually settled the dispute only by creating an electoral commission composed of eight Republicans and seven Democrats who voted on strict party lines, making Hayes the winner by the margin of one vote, even though he had lost the popular vote.

Only four years after the notorious election of 1876, the now largely forgotten election of 1880 nearly produced another crisis. Republican James A. Garfield’s victory over Democrat Winfield Scott Hancock hinged on Garfield’s capture of New York’s 35 electors, which Garfield won by a margin of only 20,000 votes. Rumors of Republican vote fraud in New York were rampant during the days following the election, and the chairman of the Democratic National Committee and the leader of Tammany Hall seriously considered contesting the election.

Hancock did not concede for five days, and even then he publicly suggested that vote fraud had occurred in New York. Although Hancock believed for the remainder of the life that Garfield’s victory was fraudulent, a book about the 1880 election has concluded that Hancock’s “essential conservatism caused him to acquiesce in the results.” In particular, Hancock probably feared that turmoil would be particularly injurious to the nation so soon after the turbulence of the 1876 election. Moreover, the reputation of the Republican Hayes Administration for honesty and efficiency also may have made a Democratic contest of the election impracticable.

Although no one knows whether the outcome of the 1880 was really the product of fraud, a Republican diplomat stationed in South Africa had written to Garfield during the summer of 1880 to propose that the G.O.P. import nine thousand voters into New York on election day from heavily Republican states that adjoin New York, and another six thousand from Canada. No reply from Garfield is known.

The 1880 election is also significant because it produced history’s most narrow popular presidential vote, with Garfield winning by fewer than 8000 votes. The 1880 election therefore nearly became another election in which the winner of the electoral vote did not carry the popular vote.

Gilded Age presidential politics was a dicey affair. Like the election of 1880, the elections of both 1884 and 1888 hinged on the outcome in New York. In 1884, Grover Cleveland defeated James G. Blaine in New York by a margin of barely one thousand votes out of more than a million cast. Cleveland won the nation’s popular vote by only 23,000 votes. In 1888, Cleveland was defeated by Benjamin Harrison, who carried New York by a margin of only one percent. Cleveland, however, won by more than one hundred thousand popular votes nationwide.

In the next three elections, the victor (Cleveland against Harrison in 1892 and William McKinley against William Jennings Bryan in 1896 and 1900) won by a comfortable margin of both popular and electoral votes. In each one of these elections, however, a shift of fewer than 75,000 votes in strategic states would have changed the election’s outcome.

In 1916, another election hinged on a narrow vote in only one state. Woodrow Wilson, a Democrat, defeated Republican Charles Evans Hughes by only three thousand votes in California, where 13 electoral votes tipped the election to Wilson. Since Hughes carried most of the East, many newspapers declared Hughes the victor in early editions on the morning after the election and were as embarrassed as the television networks in the 2000 election when returns from the Midwest and West threw the election into doubt.

For another two days, the outcome was uncertain while votes were counted slowly in various closely contested states, especially in California, which the Democrats had at first conceded to the Republicans and in which the ballot format retarded tabulation efforts. Meanwhile, the national chairmen of both parties retained legal counsel in anticipation of litigation over the election, and the Democratic chairman wired state and county chairmen in various states to warn that Republicans were “desperate” and to urge that they personally guard the safety of ballot boxes. In words that could have been written during the past month, the American Review of Reviews in December 1916 observed that when “a presidential election is to be decided by the result in a single State, where the vote is so close as to be in doubt for a number of days, it is not surprising that there should be anxiety as well as suspense.” Hughes did not concede until more than two weeks after the election. When Wilson finally received Hughes’s telegram, he drily remarked that “it was a little moth-eaten…but quite legible.”

Many factors may have contributed to Hughes’s decision to refrain from contesting the election. First, Republicans may have feared that a challenge to California’s votes would result in a Democratic challenge to the outcome in Minnesota, which Hughes had carried by only 396 votes; Wilson’s victory in Minnesota would have enabled him to win the election without California’s electoral votes. Moreover, Republicans do not appear to have had substantial evidence of irregularities in California. Hughes’s decision to refrain from contesting the election also may have reflected ambivalence over whether he wanted to be president. Hughes was one of the very few persons ever to have been genuinely drafted to run for president. He made no effort to obtain the Republican nomination and resigned very reluctantly from the U.S. Supreme Court to accept it. Hughes correctly believed that the next president would have to lead the nation into war and that the incumbent party would suffer as the result of wartime strains. Hughes also may have been particularly reluctant to plunge the nation into an election crisis during a time when the nation was involved in international crises that clearly presaged war. Finally, Hughes may have felt that the result was fair insofar as Wilson won the popular vote by nearly 600,000 ballots, a margin of three percent.

After 1916, no election was close again until 1948. In 1924, however, the Electoral College created much anxiety during the months preceding the election. Although Republican Calvin Coolidge won by a modest landslide of popular and electoral votes, there were indications until late in the campaign that Robert M. LaFollette, running as a Progressive, might well carry enough Midwestern and Western states to throw the election into the House of Representatives. Speculation about various unsavory deals that might occur in the House abounded throughout the nation for several weeks. LaFollette, however, carried only his home state of Wisconsin.

In 1948, 1960, and 1968, the nation was narrowly spared having a winner who had lost the popular vote and/or the election decided by the House of Representatives. Moreover, the winner of the1968 election nearly lost the popular vote, and ballot ambiguities prevent any certain declaration about who won the popular vote in 1960.

In 1948, a shift of only about 24, 294 in three states (16,807 in Illinois, 8,933 in California, and 3,554 in Ohio) would have resulted in the election of Republican Thomas E. Dewey rather than Harry S Truman, even though Truman won the popular vote by more than two million. With third-party States’ Rights candidate Strom Thurmond winning 39 electoral votes, the election would have been decided by the House of Representatives if Truman had lost two of these three states to Dewey.

In 1960, a shift of only about four thousand votes in Illinois and fewer than 25,000 votes in Texas would have enabled Richard M. Nixon to defeat John F. Kennedy. Since Harry F. Byrd received the electoral votes of fourteen unpledged electors from Mississippi and Alabama, plus the vote of a faithless elector from Oklahoma, the election would have been have thrown into the House if Kennedy had lost either Illinois or Texas together with any one of several closely-divided states that had enough electoral votes to deprive Kennedy of victory but not enough to enable Nixon to win.

Rumors of fraud in numerous states, particularly closely divided Illinois and Texas, encouraged Republicans to consider contesting the election. Since Nixon conceded to Kennedy on the day after the election and later encouraged the belief that he had stood aside only in order to prevent a constitutional crisis, many commentators in the wake of the 2000 pointed to Nixon’s prompt concession as an example of good sportsmanship — often with the implication that Gore should be so magnanimously acquiescent. The myth that Nixon simply rolled over and played dead was recently shattered, however, by David Greenberg, a columnist for the online magazine Slate. Greenberg has demonstrated that Republicans investigated vote fraud charges, demanded recounts in eleven states, and filed lawsuits in Illinois, New Jersey, and Texas.

The popular vote was also very close, with Kennedy receiving an official plurality of only 112, 827 votes. Since there was some ambiguity about whether Alabama votes for Democratic electors were actually intended to be cast for Kennedy, whose name was not on the Alabama ballot, some scholars have suggested that Nixon actually won the popular vote. Since the intent of Alabama Democratic voters is impossible to discern, however, it remains impossible to say who actually won the popular vote.

In 1968, a shift of 53,024 votes from Nixon to Hubert H. Humphrey in New Jersey, Missouri, and New Hampshire would have thrown the election into the House of Representatives, since third-party candidate George C. Wallace won 45 electoral votes. A shift of slightly more than 100,000 votes in California would have had the same result. Moreover, the transfer of only a quarter of a million votes would have given Humphrey a plurality in the popular vote.

In 1976, a shift of only 5,559 votes in Ohio and 3,687 in Hawaii would have resulted in the election of Gerald R. Ford rather than James Earl Carter. And a shift of less than 12,000 votes in Ohio and Delaware would have produced a tie that would have sent the election to the House of Representatives. As in so many past elections, the outcome of the 1976 election hinged on the electoral vote of New York, which went for Carter by the relatively narrow margin of fewer than 300,000 votes. According to Electoral College experts Neal R. Peirce and Lawrence D. Longley, Republicans made allegations of voting irregularities in New York and demanded a recount on election night, but later decided not to proceed with an extended investigation. If an automatic recount had not confirmed Carter’s victory in Ohio, Republicans might have investigated rumors of fraud and irregularities in Mississippi and Missouri. Republicans decided not to proceed with challenges at least in part because Carter’s popular vote exceeded Ford’s by a decisive 2.7 million.

Many of the closely divided elections have produced proposals for electoral reform. In particular, the 1968 and 1976 elections resulted in serious congressional consideration of abolition of the Electoral College. In 1979, a majority of senators — 51– actually voted to abolish the Electoral College, but this fell short of the two-thirds majority needed for a constitutional amendment.

Proposals for abolishing the Electoral College will almost certainly revive next year if, as seems likely, the winner of this year’s election will have lost the popular vote. The expansion of popular democracy through the elimination of racial, gender, and economic disenfranchisement and the development of the “one person, one vote” principle in legislative apportionment suggests that Americans today will tolerate this anomaly far less gladly than did the Americans of 1824, 1876, and 1888. The actual occurrence of this anomaly would also provide the abolition movement with an impetus and an urgency that the movements following the 1968 and 1976 elections lacked.

But while abolition of the Electoral College would prevent the embarrassment of electing another president who had failed to win more popular votes than his or her opponent, it could actually exacerbate turmoil if the national vote were close. Rather than re-counting the votes of one or a few close states, the votes of all fifty states would need to be recounted. Imagine the Florida mess multiplied fifty-fold! Principles of federalism also favor retention of the Electoral College.

Accordingly, any serious consideration of electoral reform, with or without abolition of the Electoral College, needs also to consider various other reforms. In particular, both federal and state officials need to acknowledge that presidential elections affect critical national interests and therefore need to be regulated more carefully by federal authorities. The confusion over the format of ballots and counting procedures demonstrates that some type of more uniform system is needed for designing ballots and determining the procedure for counting them in a predictable and politically neutral manner. More standardized procedures for voter identification likewise are needed.

In addition to providing more unity to state procedure, any consideration of reform also needs to take account of congressional procedures. In particular, the creaky system of election by the House in the event of a deadlock needs careful scrutiny. When the Constitution was adopted, election by the House seemed more democratic because senators were not chosen by the people. The procedure which gives each state one vote also made more sense at a time when states’ rights were more pronounced. Today it might make more sense to allow every member of Congress to vote to break the deadlock.

Reform proposals during the next year are also likely to include plans for providing a more refined procedure for determining the people’s choice through run-off elections or a system used in some European countries that permits voters to list their choices in order of preference.

Ultimately, of course, no system can prevent close elections, provide for every legal contingency when a close election occurs, eliminate all fraud, or foreclose rancor arising from a closely contested election. The chaos arising out of the present election has amply demonstrated, however, that the present system invites uncertainty and turmoil. After the settlement of the present election, the nation will certainly turn its attention to efforts to devise more coherent and politically prudent procedures for presidential elections.

December 4, 2000

William G. Ross teaches Constitutional Law and Constitutional History at Cumberland School of Law, Samford University, Birmingham, Alabama. He is a 1979 graduate of Harvard Law School. He welcomes comments on his columns at JURIST@law.pitt.edu