by hilzoy
Via Matt Yglesias, the NYT:
"A federal panel responsible for conducting election research played down the findings of experts who concluded last year that there was little voter fraud around the nation, according to a review of the original report obtained by The New York Times.Instead, the panel, the Election Assistance Commission, issued a report that said the pervasiveness of fraud was open to debate.
The revised version echoes complaints made by Republican politicians, who have long suggested that voter fraud is widespread and justifies the voter identification laws that have been passed in at least two dozen states.
Democrats say the threat is overstated and have opposed voter identification laws, which they say disenfranchise the poor, members of minority groups and the elderly, who are less likely to have photo IDs and are more likely to be Democrats.
Though the original report said that among experts “there is widespread but not unanimous agreement that there is little polling place fraud,” the final version of the report released to the public concluded in its executive summary that “there is a great deal of debate on the pervasiveness of fraud.” (...)
A number of election law experts, based on their own research, have concluded that the accusations regarding widespread fraud are unjustified. And in this case, one of the two experts hired to do the report was Job Serebrov, a Republican elections lawyers from Arkansas, who defended his research in an e-mail message obtained by The Times that was sent last October to Margaret Sims, a commission staff member.
“Tova and I worked hard to produce a correct, accurate and truthful report,” Mr. Serebrov wrote, referring to Tova Wang, a voting expert with liberal leanings from the Century Foundation and co-author of the report. “I could care less that the results are not what the more conservative members of my party wanted.”
He added: “Neither one of us was willing to conform results for political expediency.”"
Note that that last email is not a comment in response to the Times' questions; it's an email sent to a staff member of the commission. I leave to your imaginations what earlier questions or remarks by the commission or its staffers might have prompted one of the study's authors to write: “I could care less that the results are not what the more conservative members of my party wanted.”
Commentary below the fold.
I take elections seriously, and I care both about making sure that any citizen who wants to vote can, and that people who are not citizens can't. These two concerns are at odds. A lot of measures that would prevent people who are not eligible to vote from voting would also keep people who are eligible to vote from voting. Imagine, for instance, that we required (say) that registered voters produce five forms of identification before they could vote. This would make it a lot more difficult for someone who was not eligible to vote to commit vote fraud, but it would also result in a lot of people showing up with only three or four kinds of identification and being turned away. And if those people had jobs to go to, or kids to get out of child care, and couldn't manage to get back to their homes, put together the requisite papers, and go back to the polling place, they would not vote. (This is particularly likely if their polling place has very long lines.)
So coming up with a set of rules governing voting will always involve some sort of balancing act. You want to make it as hard as possible for people who aren't eligible to vote to cast a ballot, while not making it unduly difficult for people who are eligible, and these two goals are opposed to one another. (The best solution, I think, would be a hard-to-forge national ID card, but I digress.) In figuring out how to strike that balance, knowing exactly how much of a problem voter fraud actually is is crucial. If there isn't much of a problem, then we don't need to adopt "solutions" that deter legitimate voters; if there is, then we do. It's also important to know whether more stringent requirements really do deter eligible voters, since if they don't, then we can adopt them without worrying about their effect on turnout among people who are actually eligible to vote.
It's hard to get good information, though, since this issue is politicized. The people who are most likely to be deterred from voting by onerous requirements are the poor, who are a natural Democratic constituency. This gives Democrats a reason to overstate the problem, and Republicans a reason to understate it. It also gives Republicans a reason to overstate the problem of voter fraud, and Democrats a reason to understate it, since the need for more stringent ID requirements and so forth only exists if there's a real problem of voter fraud.
Personally, I've always sided with the Democrats on this one. You might think that this is because I am a Democrat, but I actually have reasons for thinking as I do. To see why, it's crucial to bear in mind that voter fraud is only one version of election fraud. Election fraud includes not just getting dead people to vote, but things like: stuffing ballot boxes, rigging voting machines, discarding ballots that have been cast, etc. Adopting measures that deter voter fraud in particular do not deter election fraud more generally: requiring that people produce massive amounts of ID before voting in no way decreases the likelihood that their ballots will be "lost" by corrupt election officials.
Now: when I ask myself what I'd do if I wanted to rig an election, voter fraud is one of the last options I'd choose, since it has a number of pretty striking disadvantages. (Note: I think this was much less true earlier, before the advent of things like computerized voter rolls.) Specifically:
* It's difficult. Voter fraud requires that you either successfully register a whole lot of people who are ineligible to vote, or that you find out the names pf people who are registered but who are unlikely to vote and equip a bunch of people with their ID, etc. This is a lot harder than simply making ballot boxes disappear in precincts where your opponent is likely to do well, or corrupting a voting machine, if you do it in the numbers necessary to alter the results of an election.
* It involves a lot more people than other forms of election fraud, if (again) you want to alter the results of an election, as opposed to seeing whether your dog can successfully vote for a fraternity prank. An actual individual has to cast each and every fraudulent vote. Even if you have people going around voting all day long (without election workers catching on?), you'd need a fair number of them to alter the course of most elections. And every person you involve makes your plan more vulnerable to discovery. Again, much simpler just to disappear the odd ballot box.
* It's vulnerable to discovery on other fronts. Voter rolls can be checked. People you think won't vote can show up. Moreover, it's a lot harder to explain voter fraud as an accident. People don't accidentally register to vote when they are ineligible, nor do they accidentally show up to vote in someone else's name. By contrast, people do accidentally misplace ballot boxes, and so forth.
For all these reasons, if I wanted to rig an election, I would vastly prefer to use other methods. Again, I think this might not have been true before voter rolls were computerized: when the list of registered voters in a given ward consisted of someone's handwritten entries in a large book (or whatever), and might never be checked, it would have been a lot easier for a corrupt registrar to simply enter a whole bunch of fake names on the rolls. But nowadays, voter fraud seems like a pretty unlikely way of trying to rig an election, given the existence of a lot of much easier alternatives.
Still, I am a Democrat, and so I might just think this because I am ideologically predisposed to. To decide the question, we need not just my a priori speculations, but actual facts. And this is why what the administration did was so damaging: here, as elsewhere, they are not content to present the facts accurately and defend their responses to them; they try to slant the facts themselves. If there is no serious problem with voter fraud, then that's something genuinely worth knowing, since it means that we do not need to take any further steps to address it. (Note the 'further' in that last statement: I'm not saying that we don't need to do anything to ensure that only eligible voters vote, just that if there's no serious problem, then whatever we're doing now is working pretty well, and we don't need to do anything else.) And that's what the administration is trying to hide from us by altering this report.
In deciding whether to adopt more stringent requirements, as I said above, we also need to know whether they actually do affect voter turnout. The commission has been active on this front as well. From the NYT article:
"And two weeks ago, the panel faced criticism for refusing to release another report it commissioned concerning voter identification laws. That report, which was released after intense pressure from Congress, found that voter identification laws designed to fight fraud can reduce turnout, particularly among members of minorities. In releasing that report, which was conducted by a different set of scholars, the commission declined to endorse its findings, citing methodological concerns."
I take the integrity of elections seriously. If voter fraud were a serious problem, I'd want to know. But I want to form my views on the basis of the facts, not vice versa. And the fact that this administration is doing its best to stand in my way makes me angry, not so much for me, but because the result of imposing more stringent requirements on voting than the facts warrant is that actual eligible voters will not cast their ballots. And the very same concern that makes me care about voter fraud also leads me to find the idea of deterring voters for no good reason unacceptable.
"It could be that I am simply smarter than everyone who has ever worked in government so no one has ever thought of this idea before. However, I doubt that."
I wouldn't doubt for a moment that you might be smarter than anyone assigned to such a project. But I strongly suspect that such a project would be strongly resisted, by liberals. Which is why I think you would have to convince them.
"Seb, I'm trying really hard to believe that you have real concerns about voter fraud, as opposed to a desire to reduce minority voting. However, it would be easier to do that if you showed a little interest in actual voter fraud investigations. Alternatively, you could continue to argue that we cannot ever measure voter fraud without imposing radical identification requirements."
You may have confused me with someone who has actual power. I regret to admit that I don't. As for radical identification requirements for voting, lots of other countries require some sort of actual identification. It isn't unheard of. I'm not saying that is dispositive--just that it isn't radical.
Posted by: Sebastian Holsclaw | April 12, 2007 at 07:22 PM
Who are the authorities who can order that a business not give employees a paid day off, and what authority did those authorities use?
I assume it was based on a law saying you can't offer someone anything of value as enticement to vote. A paid day off is something of value.
You could give all your employees Election Day as a paid day off, naturally. But a law like that would prohibit you from conditioning it on actually voting. It's a fine distinction, I grant you.
Posted by: Steve | April 12, 2007 at 07:40 PM
"I assume it was based on a law saying you can't offer someone anything of value as enticement to vote."
Is it illegal in some places to do that, rather than it being illegal to offer someone anything of value as enticement to vote in a specific way?
I'm not disagreeing: I just hadn't realized or been aware of that, if so. Does anyone have a pointer on how widespread such laws are? (Yes, I can google, but I'm feeling lazy, and not to mention crappy, so I'm asking first.)
Posted by: Gary Farber | April 12, 2007 at 07:48 PM
Sebastian... "I lived out of my car for two months for God's sake."
Was it a Caddy or a Hummer?
Posted by: LWM | April 13, 2007 at 12:20 AM
Vote for your favorite song
My kid's choir
instead of
cleek doing moody stuff
I'm a balding-headed link ho now, give me some love, or else you will never actually hear Thullen.
Posted by: DaveC | April 13, 2007 at 12:34 AM
Gary: Is it illegal in some places to do that [offer someone anything of value as enticement to vote], rather than it being illegal to offer someone anything of value as enticement to vote in a specific way?
Yes; in Virginia and Iowa at least, and I believe in many places.
The rationale is, I think, that there are many precincts in which an overwhelming majority of the voters vote one way or the other, so that offering an enticement of value has the effect of an enticement to vote in a specific way.
Posted by: Nell | April 13, 2007 at 01:15 AM
Common Sense, the procedure you describe is cluster sampling. (As you doubtless know.)
Which means I imagine Republicans would have two problems with it:
One: if the objective is to find out if voter fraud is a widespread problem, it would be only sensible to make clear that the data in the survey can't and won't be used to prosecute or investigate any individiduals. (The UK Census has an unbreakable rule that it doesn't share its raw data with anyone, for just that reason.) I can see Republican politicians calling it an "investigation into voter fraud" rather than "a survey of voting patterns" and then complaining that the people carrying it out were "shielding criminals".
Two: Many conservatives appear to have locked themselves into the belief that cluster sampling is not an effective scientific method of getting data, because if they did believe it, they'd have to believe that around 700 000 people in Iraq have been killed as a result of the US invasion/occupation, and they don't want to. (For all I know, this will be a short-lived rejection, but where it exists, it seems to be absolute.)
Posted by: Jesurgislac | April 13, 2007 at 03:51 AM
"Sebastian... "I lived out of my car for two months for God's sake."
Was it a Caddy or a Hummer?"
It was a VW Rabbit, thanks for wondering.
Posted by: Sebastian Holsclaw | April 13, 2007 at 11:30 AM
Sebastian... "I lived out of my car for two months for God's sake."
Was it a Caddy or a Hummer?
Wow, was that spectacularly crass. Sorry about that, Sebastian.
Posted by: Anarch | April 13, 2007 at 07:52 PM