Voter ID laws are again in the news today. This issue is becoming tiresome, since rational people know very well that the movement to require positive identification at the polls is nothing more than a Republican strategy to deny the vote to people assumed likely to favor Democrats. This becomes particularly obvious when one realizes that many voter ID laws severely restrict what identification is acceptable, disallowing student IDs, for example.
It’s time to admit to ourselves that our elections are not perfect. Recounts seldom yield vote counts identical to the original tabulations. Voting machines malfunction. Ballots are poorly designed and mislead voters into casting votes other than as they intended. Ballots are marked ambiguously. And, although vote tampering is less prevalent than formerly, it is still possible through any number of mechanisms.
In-person voter fraud, the problem that voter ID laws are ostensibly intended to foil, is either exceedingly rare, or we are unbelievably bad at detecting it. It is difficult to document any instances at all, and it is incomprehensible that it has ever swung an election. If one really wanted to steal an election, having people cast votes on behalf of others is an incredibly inefficient and stupid way to go about it. It would be the equivalent of draining a lake one thimble at a time. The integrity of our elections would be better enhanced by developing better methods of voting (including better voting machines), better ways of reporting vote tallies, and stationing more election judges at the polls.
There is no reason to believe that a law requiring positive ID to vote will prevent fraudulent voting by more than a handful of people. On the other hand, many people who lack an approved ID will simply stay home on election day or will cast a provisional ballot that probably will never be counted. In a given state, there are likely to be tens of thousands of people—perhaps hundreds of thousands of people—disenfranchised thereby. These people are disproportionately poor, nonwhite, or elderly. They represent a natural constituency of the Democratic Party.
It is helpful to think of voter ID laws by using an analogy to medical testing. When a person appears at a polling place, a test is performed to determine whether that person is qualified to vote. Usually, the test is simply a matter of comparing the person’s self-declared name against a voter list. Voter ID laws require particular forms of identification in order to establish a name that can be compared to the voter list. If we think of allowing a person to vote as a positive test, then voter ID laws seek to prevent false positives. Rejecting an otherwise qualified voter for lack of a specific form of identification becomes a false negative.
A voter ID law eliminates—tries to eliminate, at any rate—false positives at the cost of creating false negatives. No honest observer believes that there are now many false positive test—perhaps a handful in any election. How many false negatives—that is, how many fully qualified voters—have to be disenfranchised to present the handful of false positives? As an example, a judge has just ruled that Wisconsin cannot employ its new voter ID law in November’s election. It has been reported that this affects 300,000 Wisconsin voters who would otherwise be disenfranchised. Is it really worth taking the vote away from, say 60,000 citizens to stop one individual pretending to be someone else. If you’re a Republican, I suppose this does indeed seem like a reasonable tradeoff. In fact, voter ID laws are not really intended to present false positives. They are intended to create large numbers of false negatives.