The great voter-fraud myth

When we think of voter fraud, we generally think of the grift depicted in Preston Sturges's 1940 film The Great McGinty: The drifter McGinty accepts a party operative's offer of two bucks for a fraudulent vote and, enterprisingly, zips through 37 precincts, casting 37 votes (and billing 74 bucks). McGinty gets away with it because the poll guardians can only ask him to sign his name. We use the same system in my precinct in California and in many other states where there's a premium on getting people to vote: Once you've registered, all you have to do is show up, say who you are, and fill out a ballot.

Self-styled vote reformers say such simple safeguards don't suffice, that we need stronger forms of identification and registration to vote, because McGintyesque scenes occur routinely at our polls. Apparently, that's why Bush administration officials wanted to fire certain U.S. attorneys who didn't believe many McGintys existed offscreen. And with good reason: They don't. Histrionics about voter fraud, now as in the past, don't draw attention to a priority problem. Instead, they serve to draw attention away from the real scandals of our electoral system.

Vote reformers argue that Americans need more and better documentation if we want to exercise our right to vote. They play on fears that, as Federal Election Commissioner Hans von Spakovsky put it, "the Democratic Party and its alter ego organizations" like the NAACP and the League of Women Voters, working particularly in states with large immigrant populations, are drawing on their "long and studied history" of voter fraud to steal elections from good, solid, Republican candidates.

But this is bunk, through and through. Suppose there were widespread fraud under our current system. What would you see? You'd expect, for one thing, a much higher turnout in states not requiring ID to vote, as McGintys flooded the polls. But in 2004, according to, two states which require ID, Alaska and South Dakota, ranked among the top ten states for voter turnout, while six of the states with the lowest turnouts do not require identification. There is no obvious correlation.

But if that's not good enough, let's take another view. If there were widespread fraud, you would also expect results to differ dramatically from predictions. If it were true, as one Republican lawyer remarked to Karl Rove in a public forum last year, that the Democratic Party "rests on the base of election fraud," you would expect Democrats to win elections out of proportion to public support for their policies. But if anything, the pattern runs the other way: at no time since the late 1980s has less than half the population supported the proposition that "government should care for those who can't care for themselves," and indeed the proportion of Americans supporting that idea has generally run above two-thirds. Yet Republicans campaigning against this presumption managed to maintain electoral majorities nationwide through much of this era.

So the circumstantial evidence isn't good. Nor has direct evidence of systematic or consequential voter fraud shown up despite considerable effort to find it among the millions of votes cast. As Attorney General Alberto Gonzales has noted, the Department of Justice "made enforcement of election fraud and corruption offenses a top priority" starting in 2002. Yet in three years the Department of Justice found only two cases of double voting and 12 cases of voting by non-citizens--some in states requiring voter ID, some not. Such cases appear to have resulted from error or ignorance, not systematic corruption, and in any event have no clear relation to stricter voter ID laws.

The divergence of rhetoric from reality resembles that of a hundred years ago, when reformers first supported registration laws. Although the reformers talked about "corruption," they didn't really mean vote-buying or repeat voting. They meant the wrong kind of people voting: "Universal suffrage," one reformer noted in 1903, meant "'tramp' suffrage"; it meant "licensed mobocracy."

Perhaps most egregiously, reformers obsessed over picayune and hidden corruption while ignoring the commonplace, open-air corruption that festered in front of their faces. The Democratic candidate for Pittsburgh's mayoralty, George Guthrie, lost in 1896 but claimed bitterly that he had won "a majority of the free and unbought voters." And indeed, in that era, one or two McGintyesque repeaters were caught in Pittsburgh. Yet such minor fraud--though real and probably common enough in those days of higher turnout--paled in comparison to the wholesale disfranchisement of voters carried on at the same time. After years of keeping the black vote down through terrorism, southern Democrats set about doing it by law, rewriting their state Constitutions to include new restrictions such as registration, residency requirements, and poll taxes. While, in keeping with the Fifteenth Amendment, those laws did not explicitly mention black voters, they nevertheless severely reduced black turnout. States of the North and West followed suit, seeking to disfranchise immigrants. Thus did Americans openly corrupt their democracy, and vote reformers spoke little, if at all, against it.

For this reason Americans--and indeed all citizens of newly security-minded nations--should balk at these new calls of "papers, please." Requirements designed to restrict voting, enforced at the discretion of polling-place minders, leave a tremendous amount of power in the hands of those workers. One hundred years ago, such workers commonly found blacks less suitable for voting than whites and immigrants less suitable for voting than native-born. And voter turnout then began to fall to its current, lamentable level. This is the real fraud of American democracy--not too much voting, but too little. To make it a priority that fewer people vote is madness--if it isn't anti-democratic partisan activism meant to thwart the will of the citizenry.

Along with the new immigration bill, which would create a legal caste of guest-workers bearing documentation of their second-class standing, and the REAL ID law, which would require that we provide extraordinary proof of citizenship to drive or buy booze, the new push for voter identification threatens simply to codify in law what we already do: treat certain classes of people poorly. Until now, Americans have generally had the decency to rebel against writing such prejudices into law even as we continue to practice them. The gap between reality and ideal has helped remind us that we want better for our children, who should do as we say and not as we do. Under these laws we begin to surrender even the shield of our hypocrisy.

By Eric Rauchway