Does America Really Have a Problem with Ghost Voters?

Earlier this year, the Public Interest Legal Foundation, a nonprofit law firm founded to "fight lawlessness in American elections," issued a report with what, on the surface, looked like alarming statistics: according to US census data, in 141 counties, across 21 states, the number of people registered to vote far outnumbered residents who were eligible to vote—in some cases, very far. In Franklin County, Illinois, for example, voter rolls outnumbered eligible voters by 90 percent. In other words, if you are registered to vote in Franklin County, there is a good chance that you are dead, or don't live there anymore.

There's no actual law against these imbalances—federal election law prevents counties from removing voters from the rolls without confirming that they aren't actually ineligible. But PILF is nevertheless threatening to sue the counties if they don't clean up their books. Claiming that their report's findings are so egregious that they constitute a threat to the integrity of the country's elections, PILF argues that the counties are violating a section of 1993's National Voter Registration Act, which requires local officials to make reasonable efforts to remove dead or ineligible voters from their rolls.

All this hue and cry is a little shrill, and it's unlikely that PILF's legal haranguing will actually come to anything. But the problem the group points out is real, and widespread. And while these ghost voters may not pose the existential threat that PILF suggests, they are signs of a systemically bad bureaucracy that has the potential to cripple the democratic process nationwide.

PILF is just the latest in a long line of organizations to name, shame, and even sue counties for having more voters on the books than residents over the past five years or so. In 2011, a similar campaign called out 14 counties in Illinois; another group took 24 counties in Iowa and 20 in Colorado to task the next year; and in 2014, Judicial Watch put 11 states and the District of Columbia on notice for having voter rolls overloaded with people who didn't exist.

A more expansive survey of the problem, conducted by the Pew Center on the States in 2012, found that about one in eight voter records in the US are inaccurate, out-of-date, or duplicates. Contrary to what conservative groups like PILF suggest, Pew found that these errors did not appear to result in incidences of voter fraud. But they nevertheless were inefficient, and put a drag on taxpayer time and money.

J. Christian Adams, PILF's president, believes that the problem is even more widespread. Adams, who served in the Justice Department's Voting Section from 2005 to 2010 and who has since been involved in a number of campaigns to rectify voter rolls, believes that while his team only caught 141 hyper-inflated counties in their recent report, there may be more than 200 counties nationwide with similar imbalance figures.

That's not a huge number, given that there are 3,143 counties or county equivalents in America, so 200 only accounts for 6.4 percent. But national averages show that only about 70 percent of eligible voters are actually registered, a figure against which many more counties may be implicated in voter roll inflations. Within individual states, the proportion of counties with more registered than eligible voters can be quite high. PILF's report identifies 24 offending counties in Michigan, for example, or nearly a third of the counties in the state.

PILF and other organizations concerned with this issue claim that the errors and oversights on the voter rolls make it easy for people to commit fraud at the ballot, diluting ballots, disenfranchising citizens, and corrupting the electoral system at the core of our democratic government. There's often a conspiratorial or paranoid tinge to these warnings, suggesting that the current Democratic president and his allies on the left are opposed to cleaning up voter rolls because it allows them to cheat their way into power. Implicitly, and in at least one campaign explicitly, the conclusion is that as long as voter rolls remain cluttered and inaccurate, there's a reason to enact new voting restrictions, like the voter ID laws that have been shown to disenfranchise many eligible voters, many of whom are poor or minorities.

In recent years, most direct claims of voter fraud, by any means, have proved false. And where people do vote as someone they're not, the issue was usually a clerical error, like signing a voter roll on the wrong line, usually as a relative whose name would show up proximate to your own, rather than a vote by an illegible voter or a double vote by anyone. So it seems we're pretty well protected against fraud by existing measures, with or without dead voters on the rolls.

And despite a few reportedly successful cases against counties with inflated voter rolls, it's actually pretty hard to prove that a voter registration system isn't making due efforts to keep up with changing population. Taken together, it's easy to see PILF's report as overstated and attention-mongering, and dismiss the issue of inflated voter rolls as an odd but ultimately innocuous consequence of outdated record-keeping.

In reality, inflated voter rolls are a problem, for reasons far less sexy than voter fraud and fundamental electoral legitimacy.

"If you have someone on your list that should no longer be on your list, and you're still sending them information in the mail, you're still drawing precinct boundaries thinking they live in a particular place," said David Becker, director of Election Initiatives at the Pew Charitable Trusts. "There are costs that go along with that… [Also], when someone has moved, that means there's a voter somewhere who's not getting the information they expected to get. They might not know where the proper place to vote is. They might not know when election day is. They might not know what information is on the ballot. Those problems can lead to provisional ballots, which are very costly to process, [and] can lead to lines at the polls—all things that we want to avoid."

"There is [also] a legitimate integrity issue," he added. "I don't think [we've] seen any evidence of people voting who shouldn't be voting. But certainly the perception of integrity with having ballots go to addresses where they're not supposed to go is one that should be addressed."

Most counties do want to get obsolete names off of their voter rolls, but claim they don't have the resources necessary to launch a full review, and often can't get funding or support from state legislatures. Thanks to the parochial nature of data—and the petty nature of state bureaucracies—county elections officials also have very limited access to information from other jurisdictions, limiting their ability to do simple cross-reference searches that could tell them who shouldn't be on their rolls.

A few new tools have been developed in recent years to help states with this kind of analysis. Chief amongst them, explained Becker, is the Electronic Registration Information Center, a cooperative endeavor launched in 2012 by seven states concerned with voter roll inflation. States that join ERIC provide voter registration and Department of Motor Vehicle data, which is then compared to death records and soon, US postal service addresses. In exchange, ERIC member states—there are now 12—pay a minimal service fee and agree to send out registration information to eligible individuals who are not yet on voter rolls in the interest of getting both accurate and high figures in local electoral head counts.

John Lindback, the executive director of ERIC, argues that the service has proven effective at cutting down on the costs of bad voter rolls. Minnesota and Washington, for instance, have both reported a 40 percent reduction in returned mail, for instance. And only two ERIC states, Colorado and Louisiana, showed up on the PILF's recent list of states with electorally profligate counties.

"I was going to guess that those were the two states," said Lindback. "I don't think either state has done any list maintenance yet with their ERIC [data] for internal reasons. Colorado has some issues with their DMV data and Louisiana is a relatively new member… They're getting very close to doing their first set of list maintenance [though]."

This suggests that, when ERIC and similar tools are used, they can adequately and thoughtfully solve the legitimate problems of voter roll inflation, and eliminate a grounds for paranoia about voter fraud (and support for worrying polling restrictions), all at low, pay-for-themselves costs. So promoting ERIC may be a better tactic, for anyone legitimately concerned about the functionality and legitimacy of American elections, than the PILF's lawsuits, and alarmism.

"We've seen in three years that at least the ERIC states are seeing a real benefit," said Becker. "And as more states continue to join, I think were going to see that this problem is not entirely solved—but is getting solved… And we can move on to other issues."

Follow Mark on Twitter.