Posted: February 25, 2009
Coin-flipping: Is it a good idea?
In 2006, Indiana University law professor Mike Pitts published an entertaining law review article in which he argued that, when an election is so close that it amounts to a "statistical tie," the judiciary should acknowledge this by not even attempting to determine the "true" result. Heads or Tails? A Modest Proposal for Deciding Close Elections, 39 CTLR 739 (Dec., 2006). Instead, Pitts argued, they should just flip a coin to determine the winner and be done with it. Pitts argument was an obvious reaction to the morass of 2000, but similar sentiments have been echoed in the New York Times and regional newspapers since the Minnesota Senate recount began. But is coin-flipping really a good idea, or just a provocative one?
First let me say that I became infatuated with the coin-flipping idea when I first heard about it. Here, I said to myself, here is a way to shut up all the lawyers and judges and have the thing decided by a truly neutral and objective process (when it comes to resolving election difficulties, lawyers and judges are in my view generally the problem, not the solution). The idea also appealed to a certain independent streak in me that does not like being asked to accept a result that has no statistical validity. Finally, I thought a coin-flip would preserve some degree of respect for the democratic process in the mind of the average person, who would be spared by the coin from having to see more photos like the infamous magnifying glass photo of 2000 (a photograph that by itself probably did more to discredit the voting process than the actions of any candidate or court). I did recognize that the coin-flipping would cause a great deal of angst in those who want to be able to idealize the democratic process in their minds, but that was a price I was willing to pay (or have them pay) to keep the average people (including myself) from becoming even more cynical.
However, the honeymoon is over and, thanks to a Moritz colleague, I can no longer subscribe to this idea. The main problem is not the basic idea of coin-flipping but, as my colleague pointed out, the implementation. Specifically, it would be difficult if not impossible to implement the idea in a way that will accomplish the goal of eliminating troublesome litigation. If you pass a law that says election contests within a certain margin will be determined by a coin-flip, then people will file lawsuits to influence the result of any earlier recount proceedings and push the result across that margin (recounts occur before election contests in almost every state). You have not eliminated the troublesome litigation, you've only pushed it back so that it occurs at an earlier point in the process.
Mike Pitts is no dummy and thought of this problem. The solution he proposes is this: Pretend that the "triggering margin" is 0.5% of the total vote, so that any race won by fewer than that margin would be decided by a coin flip. Now, in a statewide race candidate A loses to candidate B by 0.6% of the vote. Ordinarily this would mean candidate A could still get a coin flip by closing the margin in a recount by just a little over 0.1% of the vote. However, under Pitts' regime, this 0.1% would not be enough. Instead, in order to get a coin flip the recount would have to show not only that candidate A came within the coin-flip margin, but that candidate A actually got more votes than candidate B. In other words, candidate A would have to move the count by 0.6% of the vote, a formidable task in a statewide election where the sample size is large enough to absorb most types of errors without changing the result. This would squelch most attempts at disputing the result in an election contest, and also attempts to use litigation to influence any recounts.
Trouble is, candidates can get around this litigation-squelching effect by pushing the litigation still earlier into the process, so that it occurs on the eve of the election when polls show the result will be close, or even during the counting process itself. Because a final count has not yet been determined, Pitts' rule would not apply and there would be nothing to discourage candidates from filing a series of mandamus actions to influence whether the eventual count falls within the coin-flipping margin. And this kind of last-minute, catch-all litigation is a bad thing, worse than the kind that occurs after the election, because at best it delays the count and at worst makes a muddle of it or even influences administrators to adjust counting practices to ensure the "right" candidate wins. Furthermore, the original goal of avoiding protracted and often silly litigation is not accomplished.
Of course, you could take Pitts' idea still further by prohibiting this type of litigation, but that does not make sense on a policy level because it could lead to the absurd result of not being able to stop administrators from counting votes in a way that everybody agrees is wrong.
There are, of course, other alternatives to coin-flipping in close races, such as automatically holding a new election. However, holding a new election is not as easy as it sounds, and can be downright unfair. As Moritz Fellow Steven Huefner discusses in a 2007 Harvard Journal of Legislation article, new elections are very expensive both for the states and the candidates running. Remedying Election Wrongs, 44 HVJL 265, 296 (2007). For candidates, this means that the candidate with the largest "war chest" left over after the original election has a big advantage in getting out the vote and conducting other types of campaign activities that he or she did not necessarily have in the original election. Furthermore, it may be that in the original election the vote of one candidate was buoyed up by a charismatic candidate of the same party running in a more visible election, in the way that many Democratic candidates were buoyed up by Obama this November. It is not necessarily fair to conduct another election in which that charismatic candidate is no longer running. In general, turnout will generally be lower in the new election because it involves fewer candidates, and it will involve a different mix of voters, so it can hardly be labeled a reasonable facsimile of the election that should have been. These are just some of the problems with holding a new election.
In conclusion, there is no good way that I know of to resolve a close election, only a choice of evils. Coin-flipping and automatically holding a new election are some of the most evil choices so, by process of elimination, traditional election contests start to look pretty good. They are not without their faults, but there is simply nothing better for now, until technology and record-keeping procedures improve to the point where making a meaningful determination of the result of a very very close election becomes possible.