Partisan considerations aside, it strikes me that two of the remaining contested Senate races seem to have their processes exactly backwards. In Georgia, Saxby Chambliss and Jim Martin are heading for a runoff, despite the fact that voters expressed a fairly clear preference for the incumbent. I suppose it’s possible that Chambliss will fail to add 0.1% of the remaining third-party voters in the runoff — and as a Democrat, I hope that happens — but the more likely outcome is that it will prove to be a waste of time and resources that merely confirms the results of the initial round. And in the event that Obama’s victory shakes up the race and allows Martin to win over voters who had supported Chambliss in the first round, suddenly we’re not talking about a runoff so much as a new election.
In Minnesota, meanwhile, you could make a good case that a runoff would be clarifying. Coleman and Franken have effectively tied with 42 percent each, which means that a runoff would allow the 16% of the electorate who voted for someone else the chance to express their second choice. Instead, as happened in Florida in 2000, the election will be decided by a vote margin that will almost certainly be less than the margin of error of even the most accurate counting method.
In the unlikely event that I were ever given the opportunity to write a state constitution, I would probably go with a modified version of the Georgia model. That is, require a runoff if no candidate reaches a certain level of support, but set that threshold below 50 percent, perhaps at 45. That way, runoffs would be reserved for elections in which a third-party candidate garnered significant support, not one in which a few gadflies chipped away enough of the frontrunner’s vote total to keep him or her below a majority.