The Extremely Weak Evidence of Voter Fraud in Crawford, the Indiana Voter ID Case

As I've been working on revisions to this article, I've gone back to look at the evidence of voter fraud presented by the state of Indiana in Crawford, the voter identification case that the Seventh Circuit, on a nearly party line vote, recently refused to rehear en banc. In the majority panel decision written by Judge Posner, we learn that the state conceded that there was no evidence of any prosecution ever in Indiana for impersonation voter fraud (as opposed to absentee ballot/vote buying fraud, for which there was a good deal of evidence---but the voter id law exempted absentee voting). Here is how Judge Posner responded to the lack of evidence:

But that lacuna may reflect nothing more than the vagaries of journalists' and other investigators' choice of scandals to investigate. Some voter impersonation has been found (though not much, for remember that it is difficult to detect) in the states that have been studied, and those states do not appear to be on average more "dishonest" than Indiana; for besides the notorious examples of Florida and Illinois, they include Michigan, Missouri, and Washington (state). Indirect evidence of such fraud, or at least of an acute danger of such fraud, in Indiana is provided by the discrepancy between the number of people listed on the registered-voter rolls in the state and the substantially smaller number of people actually eligible to vote.

There's a lot to criticize about this paragraph, including the fact that there's been a great deal of attention nationally from the DOJ to the question of voter fraud in recent years, and still almost none of it has been found (again, in contrast to absentee ballot vote buying, which has been found and successfully prosecuted). Much fraud also has nothing to do with voter impersonation fraud, relevant to the id issue. But I thought it most curious that Judge Posner provided no evidence or citation to these so-called "notorious" cases of voter fraud in other states.

So I went back to the district court opinion, at 458 F.Supp.2d 775 (S.D. Ind. 2006), and here is the discussion, in full (footnote omitted), of the issue in that case:

The State has also produced evidence of published books and media reports discussing allegations and instances of in-person voter fraud in several other states. See Larry J. Sabato & Glenn R. Simpson, Dirty Little Secrets 292 (1996) (noting that documentation of in-person voter fraud often occurs only when a legitimate voter at the polls hears a fraudulent voter trying to use her name, as happened to a woman in California in 1994); John Fund, Stealing Elections 64 (2004) (noting in the St. Louis fourteen dead people "voted" in the 2000); State's Ex. 2, p. 23 (describing recent U.S. Department of Justice investigations into election fraud, which, as of August 2005, had resulted in 52 convictions); State's Ex. 3, pp. 4-5, 19 (court findings that in the State of Washington's 2004 gubernatorial elections more than 1,600 fraudulently cast ballots, including 19 ballots cast by dead voters, six double votes, and 77 votes unaccounted for on the registration rolls); State's Ex. 4, pp. 2-4 (joint task force findings describing instances in the 2004 elections in Wisconsin where individuals voted twice by using fake names and addresses*794 and citizens who told investigators that they did not vote, even though the report showed that someone voted in their names); State's Ex. 6, pp. 42-43 and State's Ex. 7, pp. 3-6 (describing an investigation by the Missouri Secretary of State after the 2000 elections of two of counties which revealed over 1,000 fraudulent ballots, including at least 68 multiple votes, 14 dead person votes, and 79 vacant-lot voters, with another 200 sites requesting further review); State's Ex. 10, pp. 1-2 (newspaper reports that dozens, possibly hundreds, of people who lived outside the city limits illegally cast votes at the polls in Miami's mayoral elections in 1997); State's Ex. 11, p. 1-2 (Johns Hopkins University study which found that in Maryland at least 63 votes were cast in the name of deceased individuals between the 1980's and 2004). The State has produced newspaper reports recounting that in recent elections votes were cast in the names of dead people in Georgia, Illinois, and Pennsylvania. See State's Exs. 12-14, 18. The report from the Commission on Federal Election Reform (known as the Baker-Carter Commission) recently concluded that "there is no doubt that [in-person voter fraud] occurs." State's Ex. 1, p. 18.

Upon examination, it is shocking how weak this is. Sabato and Simpson's book, and Fund's book, are full of anecdotes, many of them about allegations of voter fraud that have not been proven, and in fact have been disproven. (The single best debunking of these urban myths about vote fraud appears in Lori Minnite's analysis for Project Vote. Don't be put off by the fact that this is for an advocacy organization. This is careful, well-researched and even-handed scholarship.) Many of these allegations have nothing to do with the kind of in-person voter fraud that an id law would prevent. Many involve no illegal conduct. Take for example, Washington State, where, as we learned from fired U.S. Attorney John McKay, there was no evidence of any crime being committed by the felon voting, much less voter impersonation fraud that would support the need for an id law like Indiana's. (From the Seattle Times article: "Ultimately, Sullivan told McKay 'there was no disagreement on the task force' that a federal case could not be brought on the felon voters. Most of the felons who voted in the 2004 election, according to McKay, received ballots in the mail from the state of Washington. Therefore, he said, it would have been extremely difficult to prove in court that they knew it was unlawful for them to vote, but did it anyway")

Or take the case of the Maryland study. The district court cites to State's Ex. 11 for the following proposition: "Johns Hopkins University study which found that in Maryland at least 63 votes were cast in the name of deceased individuals between the 1980's and 2004." That exhibit, however, is to an article in the Baltimore City Paper which investigated the Hopkins study by some computer science students. That article shows how little voter fraud there actually was over a 25 year period in Maryland:

The Hopkins data suffers similarly from scrutiny. On May 25, Peck, a quiet, neat, dark-haired young man whose natural environment seems to be the glow and hum of computers, gave City Paper a printout of 62 of the 63 names (one was inadvertently left out). "It is what it is," he remarked about the data quality as he handed over the printout in the offices of Hopkins' Information Security Institute, on Wyman Park Drive. "I can't vouch for its accuracy."

Seven people on the Hopkins list were quickly found to be among the living. Although Social Security information recorded them as deceased, records of local court proceedings show otherwise--, in one instance, a wife explained over the phone that "my husband is still very much alive." In 17 other cases, the only confirmation that the person died, and when, came from Social Security--and its death information is unreliable, as several cases already show. Another dozen people have death certificates on file at the Maryland Archives in Annapolis, but, due to restricted hours for public searches, City Paper was unable to view them to confirm dates of death before publication.

Ultimately, City Paper's research concentrated on what was left of the Hopkins list: 26 currently registered voters whose deaths, and dates of death, were confirmed through additional sources: medical-examiner records, obituaries, legal records of wills and estates, or surviving family members. Each of the 26 is recorded as having voted since death at least once, in elections dating back to 1991. Each remains on the active voter registration list for Baltimore City, even though many of them died in the 1980s.

Though their deaths are confirmed, the votes cast by most of these 26 dead voters remain uncertain, because the needed election records are no longer available. Called voter authority cards (VACs), each has a voter's name and other information on it, and each voter signs one at the polling place on election day.

Since a corpse can't sign, a signature on a dead voter's VAC would suggest the corresponding vote was fraudulent, although it's also possible that a dead voter's VAC was mistakenly handed by an election judge to a living voter to sign, and the voter in question didn't make sure the card bore the correct name. This could happen, for instance, if the living voter's polling place is the same as a similarly named dead voter--John Doe Jr. is alive and going to the polls, and John Doe Sr. is deceased but still registered to vote at the same polling place. No signature on a dead voter's VAC, on the other hand, suggests that the vote was mistakenly recorded as cast when, in fact, the dead voter just didn't vote. This could happen if, for example, a dead voter's VAC was stuck behind a living voter's VAC, and after the living voter signed the top one, both were recorded as having voted.

VACs are retained for 22 months after an election, so they're available only for voters in the 2003 primary and the 2004 primary and general. Out of the 26 confirmed dead voters, three participated in these elections. City Paper provided Lamone with information on those three voters. She discovered that two of the VACs in question were unsigned; the other was signed, but name's-the-same factors come into play. The other 22 simply can't be checked. While their apparent postmortem votes are suspicious--as is their continued presence on the active voters list--neither error nor fraud can be ruled out as the cause.

Upon closer examination, I'm shocked that the district court granted summary judgment on this "evidence" of voter fraud, and that it passed muster before Judge Posner, someone usually interested in what the empirical evidence actually shows.
We'll see if this case goes up to the Supreme Court.
UPDATE: Pages 7-16 of this excellent amicus brief by the Brennan Center covers much of this same ground. Bob Bauer reacts to this post too.

Posted by Rick Hasen at May 2, 2007 08:45 AM