Wednesday, June 20, 2012

When a Vote Ends in a Tie

Some economists have trouble answering the question "why do people vote?". The problem comes from the fact that the chance that one person's vote is pivotal to an election is so small that if there is any cost at all to voting, most people shouldn't bother going out to vote, but if voting is costless, everyone should vote. Of course, the easy way out is to assume people just get utility from the act of voting. I particularly like the idea of the "swing voter's curse" (e.g. Feddersen and Pesendorfer). The idea is that if some part of the population is "partisan" and votes no matter what, uninformed independents should cast ballots in order to just cancel out the partisan vote and the rest should abstain in order to give the informed independents a better chance of deciding the election. This type of model predicts strategic abstention, decent sized voter turnout (but still not really large enough unless you assume a whole lot of partisans), and can yield more voting in national contests and more abstention in local down-ballot contests even when voting is costless.

However, sometimes it's nice to remember that just because a single voter is almost never pivotal does not mean there are never elections decided by one vote. This morning the Wall Street Journal published a front page article titled "Elections are a Crap Shoot in Texas, Where a Roll of the Dice Can Win." It is all about recent elections in Texas that ended in ties. Sure, the elections were local, and voter turnout was very low, but it's still neat to see these examples. It would be interesting to see statistics on how many elections end in an official tie and have to go to some tie-breaking rule. If the Journal found several examples in Texas this year, there are bound to be more.

According to the article, it seems that in Texas both candidates need to agree to a game of chance or a run-off/second election needs to be held. My first thought when reading the article was "If you really think you're the better candidate, why risk losing on a flip of the coin? Why not just demand a second election?" But then, of course, you just had an election and tied. In addition, run-off elections are costly (even for a very small town a cost of 10,000 dollars is cited, not to mention the extra campaigning time which is costly to the candidates, as well) and I can see an equilibrium where voters punish candidates who request run-offs by voting for a different/the other candidate if the marginal benefit of picking the "right" candidate does not exceed the cost. Thus, candidates actually could have a weakly better chance of winning by agreeing to put their future job in the hands of chance; thus, we should see both candidates readily agree to roll the dice.

No comments:

Post a Comment