After the 2000 election, some commentators observed that if the Electoral College had previously been eliminated and the president had been chosen by popular vote, the drawn-out political battle in Florida would have been replicated in every state. Here's Sen. Mitch McConnell, R-Ky. (in an essay included in the 2001 anthology, Securing Democracy: Why We Have an Electoral College):
The difference between Al Gore and George W. Bush in the national popular vote was about 500,000 (less than that, even, in the first few days after the election). That is a difference of less than .5% of the votes cast. A few thousand votes here and a few thousand votes there could have changed that election result. The Electoral College served to center the post-election battles in Florida. Without it, I fully expect we would have seen vote recounts and court battles in nearly every state of the Union.
There are several problems with McConnell's logic. The first is the familiar truism that hard cases make bad law. Because most elections are not as close as the 2000 election, it would not make much sense to let this example determine all our thinking about how presidential elections should be conducted. Even if the 2000 election had created an occasion for recounts all over the country, the republic would have survived. Might the recounts have awarded the presidency to the wrong man? Conceivably, through human error or chicanery. But at least the country would have been trying to find the top vote-getter and award him the presidency. In 2000, we didn't even try because that wasn't how our system worked. It's not obvious to me that picking the guy with fewer votes on purpose is preferable to picking the guy with fewer votes by accident.
Nor is it clear that a difference of 500,000 votes nationally (actually, it was more like 544,000) poses a likelier recount scenario than a difference of 537 votes in a single state, as occurred in Florida. Quite the contrary. The greater the numeric gap between one candidate's votes and another's, the less confidence authorities would likely have that the cumbersome business of recounting votes could be performed more accurately than the original count—regardless of how narrow the gap were proportionally—or that the recount could change enough votes to alter the outcome. Bigger numbers also pose a greater obstacle to corruption. It's harder to steal a lot of votes than it is to steal just a few.
Remember, also, that under a popular-vote system, all votes are equal. Under the Electoral College's winner-take-all allocation (in all states but Maine and Nebraska), votes in big states count more than votes in smaller states because they can leverage a lot more electors. A recount in the right state might therefore tempt partisans simply because it could turn a small difference in votes into a large difference in the Electoral College. This, George C. Edwards III observes in his excellent new book Why the Electoral College Is Bad for America, is precisely what happened in Florida:
[T]he election was so close that adding or subtracting a few hundred votes might realistically have changed the election outcome. A recount thus had a plausible possibility of altering the outcome. If the election had been by popular vote, George W. Bush would have had to find about a thousand times as many votes—a daunting task.
A final argument for switching to a popular vote is that it would likely necessitate imposing a uniform set of rules governing recounts that would apply to every state (at least in presidential elections). That standard would presumably reflect a national consensus, sparing us the tendentious debates in 2000 about the virtues of one system's method of dealing with hanging chads as compared to another's.
America's Worst College Archive:
Aug. 11: "Let's choose our president by popular vote."