At the Monkey Cage today, Justin Levitt has a nice summary of the evidence — or better, the nonevidence — on voter impersonation fraud. The headline is that he has found all of 31 credible episodes since 2000. Some are proven, some aren’t, but that’s 31 out of, he says, about 1 billion votes cast in this period.
In other words, here’s yet more evidence that the kind of fraud that could be prevented with voter identification is essentially nonexistent.
There is voter fraud in the U.S., but not of the kind that voter ID is supposed to prevent. Most current ID laws — Wisconsin is a rare exception — won’t stop fraud with absentee ballots because measures requiring ID at the polls push more people into the absentee system, where there are plenty of real dangers. Nor will it prevent vote-buying, coercion, fake registration forms, voting from the wrong address or ballot-box stuffing by officials.
These types of voter or election fraud have been documented. But to believe that polling-place voter impersonation is a real problem, you also have to believe that those responsible are super geniuses (because unlike all other election crooks they never get caught), and that these masterminds have chosen the most difficult, inefficient and clunky ways to steal elections.
Advocates of voter ID claim the costs of the system are minimal. But given how little benefit there is in deterring a nonexistent crime, even very small costs — say, the hassle for even a small number of people to get IDs they don’t have, or the tiny increments of time it takes for polling place workers to enforce the provision — just aren’t worth it.
As I’ve said, some partisan Democrats make implausible allegations that entire communities would be disenfranchised by voter ID, and there doesn’t seem to be any evidence so far to support that. But given how little potential gain there could be from this policy, even mild harassment of a few people seems an unwise cost to pay.
Generally, there probably is a trade-off in some cases between making it as easy as possible for legitimate voters to cast their ballots and making fraud difficult. But in the specific case of voter ID, given it’s dubious value, there isn’t much of a trade-off. Such measures do nothing about real election fraud, so if they keep legitimate voters away from the polls … let’s say if they keep more than 31 legitimate voters away over 15 years or so … then they are a bad idea.
[Cross-posted at Bloomberg View]