Monday 29 September 2014

Electoral lotteries

The Timaru Herald asked me for comment a few weeks' back on my non-voting stance. They used about one line from the email below; I'll share the rest with you here.
“I have not voted in any election since the 1997 Canadian Federal election. I also do not buy lotto tickets. In any large-scale election, the odds of making or breaking a tie, which are the only ways that you can change the election result, are so vanishingly small that you might as well buy a lotto ticket and promise to give any winnings to your favourite charity. And buying lottery tickets is a poor way of supporting charities. Work by Casey Mulligan and Charles Hunter in the United States shows that only one out of every 89,000 votes cast in Congressional elections changed the outcome. They went back and checked every election going back over a hundred years. You would have had to have voted 89,000 times to expect to change one outcome. You could vote diligently over thousands of lifetimes and still not expect to change an election. New Zealand results would not be much different: your vote would have to be the one making or breaking a tie in a critical electorate like Epsom, or the one pushing your preferred party to or over the St-Lague quotient threshold. Neither are particularly likely. So why bother?
Most explanations of why people vote fail. If it’s civic duty that drives voting, there are many better ways of doing one’s civic duty. Voting is not a particularly effective use of one’s time in achieving any reasonable civic objective because of the infinitesimal chance that your vote changes the outcome. The best explanation I have seen is that many people just seem to enjoy voting as a means of self-expression. But it is a bit of a rigged game. If you don’t vote, you’re told you can’t complain about the outcome because you didn’t take part. But if you do vote, and your side loses, or if your side wins but does things you don’t like, you’re also told you can’t complain about the outcome: “We won, you lost, eat that,” as one Minister was reported to have said. I prefer not to play. 
It seems reasonable to object by asking what would happen if everybody else stopped voting because of such calculations. But if turnout were sufficiently low, I would probably go and vote. The odds of changing an election outcome where there is low turnout are much higher than when turnout is well above 70%; when turnout is low enough, voting makes sense even if you do not enjoy voting for its own sake. While it is fun to imagine what would happen if they held an election and nobody came, it is an exceptionally unlikely outcome. The calculus here discussed does mean that, all else equal, those prone to being very bad at assessing the statistical likelihood of things, pivotal votes included, are disproportionately likely to vote and so parties have less incentive to cater to the policy wishes of the numerate. While this is unfortunate, my one vote would not change it.
 As a concluding comment, I would urge those people who insist on voting to vote well. Philosopher Jason Brennan argues that we have no specific duty to vote, but that if we do vote, we have a duty to vote well. Voting well does not mean supporting any particular party or set of policy positions, but it does mean that a voter must look carefully at each party’s proposed policies and assess whether the policies would actually achieve the goals that the voter wishes to further. This is hard work, but the costs imposed on the public if you vote without such careful assessment seem to outweigh whatever civic benefits obtain from higher turnout.”
I tweeted Brennan's line on voting well on election day. I hope that doing so wasn't illegal under New Zealand's bizarre election-day legislation.


  1. The real problem is that no matter who you vote for a politician wins. Douglas Adams got it right, anyone who wants to be a politician should not be allowed to become one.

  2. Can we assume that you don't believe in ethical consumerism? I note that you failed to mention today's logic in your article about Fair Trade (Oct 2013), questioning the effectiveness of the certification and margin rather than the vanishing likelihood that your dollar tips the balance between an ethical company succeeding or going bust.

    Returning to the topic of elections, isn't it egotistical to demand to be pivotal? If an opponent wins by 110 to 100 votes, then don't 101 of the total 210 votes have effect?

    p.s. My inner maths geek wants to see the odds re-calculated for the St Lague formula. Surely they would drop by at least one order of magnitude?