Of Voter Identification, Fraudulent Votes, and Closed-Out Voters

I hope you will all agree that increasing the stringency of voter identification requirements at the polls may prevent some unlawful votes, but will also result in some perfectly lawful votes not being cast, as voters, from various causes, do not get the right identification in their hands by election day.

What it seems, naively, should determine the policy for identification stringency is minimizing the sum of unlawful votes plus lawful votes that would otherwise have been cast but were not as result of the stringency. Naive or not, this is the voter identification policy test for which I will argue.

Both spurious votes and lost votes go equally to the principles underlying democracy. Suppose that there were a voting machine bug that occasionally added a spurious vote by duplicating a genuine vote. Another related bug deleted genuine votes, and did so a little more frequently than the vote duplicating bug. Would anyone contend that we should exterminate the first bug before training our resources on the second – on a theory that spurious votes are more serious than vote deletions?

Despite the negative answer I assume to this question,  it turns out that one side of the debate on voter identification seems not to think that treating lost votes and spurious votes on a par is the correct approach to policy analysis. At least it is easy to get the impression that many who argue for greater stringency are either unconcerned about qualified voters dropping away from the polls or, at least, find it acceptable if that is the price to be paid for eliminating a speculative number of cases of unlawful voting. Is there any defensible argument supporting a policy that will predictably keep plural qualified voters from the polls for each instance of unlawful voting it prevents?

The special evil of voter fraud: One possible argument is that voter fraud is a serious offense. Someone who, e.g., votes twice exhibits a utter hostility to law, fairness, and democracy. It is a high culpability crime deserving of punishment. It is also a crime that should be prevented where reasonably possible. I, for one, would subscribe to the proposition that making a crime more difficult to commit as a practical matter is sometimes both effective and just.

Photo identification card requirements will prevent some instances of voter impersonation. (They will not always prevent, e.g., voting with the driver's license of a brother who is unable to get to the polls.) If a birth certificate or passport is required to obtain the photo i.d., the requirement would further deter any non-citizen voting there might be. Depending upon details of the law, it might also prevent people from voting in a state in which they are not domiciled or city in which they do not reside.

These benefits in combating the particularly repugnant evil of unlawful voting, the argument continues, are worth the cost of some lawful votes lost – even in greater number than the unlawful votes prevented.

A hypothetical deflates the appeal of this argument. Late one night polling place officials take aside the last two members of the voting line outside town hall. After showing their credentials, they explain to the voters that they are confident that the man in front of them in line is planning to vote a second time, presumably through impersonation. Not sure that they will be able, under the rules, to turn him away at sign-in, they propose to close off the line, just in front of the intending malefactor. They will give the disingenuous excuse that it will be after their latest permitted voting time. They ask the consent of the two tail-enders.

One of the two asks “Why isn't it enough that he can be prosecuted for voter fraud? Why should I give up my right to vote?” The chief official replies, “You aren't losing any rights; you're just being a good citizen and not exercising your right to stop this fraud. Letting this creep vote again, would be a terrible failure of the process.”

The two voters would and should be unconvinced. Their two votes should be counted, even though that would mean that the bad guy “gets away with” his nefarious plan – so far, at least, as getting his tainted vote in and counted. The culpability of the fraudulent voter will be taken into account in the severity of the criminal punishment, and, it might be added, that his voting may well convert what would have been at most a criminal attempt into a completed crime. Of course, we cannot, even in this case, be assured that the fraudulent voter will be prosecuted. The lawful voters, however, should not be asked to bear the burden of this failure of the criminal justice system.

This hypothetical actually makes the argument from the special evil of unlawful voting look better than it is by focusing only on the most egregious sort of unlawful voting. In fact, stringent identification requirements at the polls may prevent merely negligent or even entirely innocent unlawful votes.

Suppose Anne moves out of New York in August, and stays temporarily in Vermont until closing on a house in New Hampshire two days after the election. Not knowing where she will be eligible to vote, Anne checks with a lawyer, who advises Anne that she should vote in New Hampshire based on her having physically moved out of New York, her fixed intention to make New Hampshire her permanent home, and her contract to buy a home there, binding as of election day. Now on the basis of my, admittedly faded, knowledge of the law of domicile, I think Anne's lawyer either misunderstood Anne's recitation of the facts or misunderstood the law. Because Anne was not actually resident in New Hampshire before the election, she was not domiciled there. Having no intention to live permanently in Vermont, she was not domiciled there either. So her domicile remained in New York until she actually started living in New Hampshire, and she should have voted in the Empire State. Under these facts, and New Hampshire's new identification requirement, it is very unlikely that she could have voted in that state had she tried. In this case, then, the identification requirements would have prevented an ineligible vote that would not have been fraudulent. It would not even have been negligent, as Anne did what a reasonable person would do to determine where to vote.

(Were this the presidential primary election, Anne would have been disenfranchised. Not yet eligible to vote in New Hampshire's first in the nation primary; her eligibility for New York's primary would have long expired before that state's date. Had Anne moved in the opposite direction in the period between the primary of New Hampshire and New York, it would seem as if she could lawfully vote in both.)

On the same facts, without the call to the lawyer, Burt would, I think, be negligent if he tried to vote in New Hampshire. It is only Cindy, who aced her Conflict of Laws exam, who would be engaged in an intentional fraud had she sought to vote in New Hampshire on similar facts.

Neither negligent voting at the wrong polling place nor innocently so voting are very common, but they are surely much more common than fraudulent in-person voting. So if it is too high a cost that two lawful voters not vote to prevent one fraudulent vote, it is a yet more unacceptable cost taking into account that most of the unlawful votes would not be fraudulent, at least if that means with intent to deceive.

Not a deprivation of rights, merely an inconvenience: I should really go on to the remaining positive arguments for preferring the stopping of an unlawful voter to the maintaining of the voting status of an active lawful voter. I will interrupt the program at this point, however, because I can almost hear, in response to my criticism of the “special evil” argument, the aggrieved shouts of defenders of stringent identification requirements: “Yes, it would be too high a cost if we took away the rights of lawful voters, but identification laws do not do that. They only require voters to go through a procedure reasonably designed to assure that they are who they say they are and that they are citizens of the state in which the seek to vote. If they decline to go through that procedure, that is their choice.”

This is a fair response if, but only if, the procedure in question is of trivial cost in time and effort. There may well be some who become election dropouts because they cannot be bothered to take a five minute walk to the motor vehicles office and go through a ten minute wait in line. At the other extreme, there are those who don't own a car and are forty miles from the closest office, or must deal with a vital records office in their birth state that is a nightmare of bureaucratic misfeasance and delay.

The pro-stingency rhetoric suggests that almost everyone is at the mild inconvenience end of the spectrum, and almost no one at the serious burden end. My own suspicion is that the median voter who lacks the newly required identification is, in fact, inconvenienced to an extent that, if monetized. would exceed the poll taxes long since declared unconstitutional. I also suspect that the very substantially burdened voters well outnumber the unlawful votes that will be prevented.

This empirical question need not, however, be resolved. The high stringency states can put to rest the objections by the critics of the “minor inconvenience” counter-argument, by insuring that it really will be only a minor inconvenience. These states can deputize a phalanx of state paid identification facilitators to work with voters who have difficulty meeting the requirements. They could provide transportation to DMV offices and pay all required fees (e.g. for a birth certificate copy). If the “mere inconvenience” claim is correct, even the most thoroughgoing facilitation program will cost very little. I see no problem at all with a state voter identification card requirement if the state can truly proclaim, “We will make it a state responsibility, at state expense, to see that no qualified voter is without an identification card, and we will do so before the next election.”

Unless I am mistaken, however, no state comes close to doing this. My suspicion is that this is because a voters' assistance corps would would conflict with the political ulterior motive behind stringent identification requirements.

Those who do not do whatever it takes to get the required identification do not deserve to vote: Perhaps states don't need to pair their stringent identification requirements with voters' assistance corps because no one is deserving of a vote who is not willing to make the effort to comply with a reasonable identification procedure, however much effort that might happen contingently to require in the particular case. Therefore we don't need to count those otherwise lawful voters who do not vote as the result of our new stringent identification procedure, because, though lawful voters, they are not deserving voters.

I have to say here that I am uncomfortable with the whole idea of undeserving voters. I don't think that the concept can be squared with any sound theory of democracy. The better occasion for that discussion, however, would be the stronger use of the concept where the lack of desert is supposed to strip the citizen of the right to vote, as in the case of felony disenfranchisement.

In the present case, failure in procuring an identification card, desert plays a weaker role, justifying de facto rather than de jure disenfranchisement. Their supposed lack of desert is supposed to make invisible, for policy evaluation purposes, their loss of access to the polls.

There is one thing I hope we can all agree upon with respect to the sort of desert one could arguably lose by putting insufficient effort, or insufficiently well directed effort, into getting a voter identification card: if there is any such concept of desert at all, it is not a strict liability concept. That is, it cannot be that there is no desert without success. But if lack of success does not imply lack of desert, there will very likely be a significant number of cases of deserving citizens who do not get to vote only because they have failed to work their way through the new identification requirements in time. For these voters we have yet to see an argument for not counting them full weight in making the policy decision about a particular stringent identification scheme.

Perhaps if the stringent identification state institutes a robust and well funded voters' assistance corps as sketched above, there will be an argument that anyone who fails to do so much as to contact the corps would be undeserving. I retain theoretical concerns, but go ahead, red states, set up the corps.

Vote cancellation argument: There is another defense of stringent identification that focuses more on the unlawful vote than the culpability of the unlawful voter. One online apologist for stringency contended that the only acceptable level of unlawful voting is zero. He argued that a single unlawful vote cancelled out his vote, and so, in effect, disenfranchised him. This argument requires an initial bold assumption that the ballot of the unlawful voter is going to be, candidate for candidate and proposition for proposition, the opposite of his. I am not sure what “opposite” is when there are three or more candidates on the ballot for a given position. Put that aside because the suppressed premise of the argument was that the unlawful ballot would be in support of Democratic candidates, the aggrieved voter himself voting Republican. (The most recent cases of prosecuted unlawful voters, if news reports are correct, do not to bear out the assumption that most unlawful voting attempts are made by Democrats.)

Let us, however, give the “cancelled vote” argument its best case and suppose that in the election for President our lawful voter marks for the Republican candidate and there is a single unlawful vote in his state, marked for the Democratic candidate. Let us also completely suppress our knowledge of probabilities and imagine our election to result in a flat tie. The lawful voter could then say, “But for the unlawful vote, my vote would have made the difference, giving all the Electoral College delegates from my state to the Republican. This tie, and the legal quagmire it puts us into, are exactly what would have happened had I not voted and had there been no unlawful vote. So the unlawful vote really has cancelled out my vote.”

Note that every Republican voter, on these facts, can make the same claim, although, of course, none can claim unique victimization, and certainly we cannot say that their votes were all cancelled.

Suppose, now, that in this same state, with its tied vote, there is a perfectly qualified voter who, despite the greatest diligence, has not been able to get her identification card. She is disenfranchised directly and in the standard sense of that word. Had she been allowed to vote, there would have been no tie either. Her candidate, whether Republican or Democratic would have won her state's election.

The lawful voter whose vote was “cancelled” by an unlawful opposite vote has a legitimate complaint, although the complaint has more than theoretical punch only in the astronomically rare event of a tied vote. His is, however, not one wit better a complaint than that of the qualified voter unable to vote because of the stringent identification requirement. Indeed her complaint is stronger. She was not allowed to exercise her fundamental democratic right at all, and her vote, unlike his, does not appear in the totals. (His vote was “cancelled” only in quite a special sense.)

Conclusion: The naive view is correct. In the absence of good empirical evidence that the number of unlawful votes prevented will probably exceed the number of otherwise lawful votes prevented by a specific stringent voter identification requirement, that requirement is unjust.

Appendix on one Moment in the Empirical Debate: The Heritage Foundation is a theoretical bastion for stringency of in-person voter identification. In a piece in USA today, its chief voting security theorist, Hans A. von Spakovsky, failed to address whether prevented lawful votes should count the same, or less, than prevented unlawful votes. He did, however, suggest that in-person fraudulent voting is a major problem. How good was von Spakovsky's evidence?

He does not give us a list of voting fraud convictions. (For all the other sorts of fraud of which I was aware as a frauds prosecutor and a white color defense lawyer, a very long list of convictions could be produced very quickly.) He does not even give us a list of indictments. Instead he relies upon the authority of the Supreme Court. Justice Stevens (!) in Crawford v. Marion County Election Board, 553 U.S. 181 (2008). von Spakovsky:

Opponents who say there is no voter fraud are wrong. As the Supreme Court noted in 2008 when it upheld Indiana’s photo ID law, we have a long, documented history of voter fraud in this country — and it could make the difference in a close election.

Perhaps there is some “opponent” somewhere who has said that there is no voter fraud, but the point that opponents keep making is really that there is not enough voter fraud to justify the effective disenfranchisement of mostly poor, minority voters. Does the Supreme Court authority he cites supply the empirical evidence von Spakovsky needs?

Von Spakovsky's “could make a difference in a close election” would seem to be a large step in the right direction, and it is nearly a direct quotation from Stevens's opinion. In the opinion, however, it followed a discussion of fraudulent absentee ballots, not polling place fraud. Assembly line fraud of absentee ballots is a great deal easier than the retail impersonation of another (perhaps dead or moved away) voter. Of that, the opinion says:

The only kind of voter fraud that SEA 483 addresses is in-person voter impersonation at polling places. The record contains no evidence of any such fraud actually occurring in Indiana at any time in its history.”

Slip Opinion at 11. (http://www.supremecourt.gov/opinions/07pdf/07-21.pdf)

The long and well documented history of in-person voter fraud to which von Spakovsky refers is mostly to be found in footnotes 11 and 12 of the opinion. The first and most detailed is sourced to one of boss Tweed's operatives who reported giving whisky to men who voted as many as four times in different stages of shaving off their facial hair. This occurred in 1868.

The other footnote is introduced by “occasional examples have surfaced in recent years.” The text of footnote 12 indicates that most of the evidence of fraud relied upon by the District Court was debunked by the amicus brief of NYU's Brennan Center and others “because much of the fraud was actually absentee ballot fraud or voter registration fraud.” However, the footnote continues, “there remain scattered instances of in-person voter fraud” Apparently the best example of such scattered instances was that of 19 “ghost votes” that turned up in a Washington State gubernatorial election in 2004. A single one of the 19 was “confirmed,” by an investigation. That investigation was not by the attorney general of the state, a district attorney, election officials, or a police agency, but by a reporter for the Seattle Post Intelligencer. Now I think the Seattle PI is a pretty good newspaper, but one would like to see a little more authoritative substance to hardest single piece of evidence that we find tracing back von Spakovsky's suggestion that in-person voting fraud is a serious current national problem.