Thursday, March 29, 2012

Against Compulsory Voting

Justin Wolfers likes the fact that Australia has compulsory voting.
I share Tom Friedman's view that the divisive nature of U.S. democracy is due to non-compulsory voting. But fixing that requires a mandate.
I respectfully disagree.

Less divisive it may be, but it offends me deeply as a statistician.

Why?

Well, who are the marginal people who vote under a mandatory system but not a compulsory system?

It's the people who didn't care enough to turn up of their own accord under a voluntary system.

Now, some of these people might actually have a firm view of the world, but just be feeling lazy or ambivalent. Maybe we really do want their opinions.

But a large number of the people who you're forcing to vote either

a) know virtually nothing about politics

b) genuinely don't give a flying fig

or both. If those people rationally decide to not vote, that's an entirely sensible decision.

How on earth does the decision-making system improve by forcing these people to pick a random answer? You're just intentionally adding noise to the process.

I remember my uncle had a mother who was senile and in a nursing home. He went in on election day to take her in to vote, only to be told that she'd already voted with the rest of the nursing home in the morning. Who did she vote for? Who the hell knows! She didn't know. Possibly someone told her who to vote for. Possibly she voted for the candidate suggested to her. Possibly not, too. But her completely random vote counted, just as much as the guy who read the paper every day. You can rest assured about that.

Lest you think that these people make up an insignificant number of votes, consider the following:

In the 1998 Australian federal election in the seat of Lindsay, there was an independent candidate who stood for office named 'Steve Grim-Reaper'.

Yes, really.

Without delving into the details of his policies, let's assume for the sake of the argument that people voting for a guy called 'Grim-Reaper' are essentially voting for a joke candidate. Let's further assume that the people voting for 'Grim-Reaper' might, if the 'Grim-Reaper' weren't running, vote for anyone at all. They are pure noise in the electoral process.

So how many people voted for the Grim Reaper in 1998?

1,043, or 1.36% of the electorate.

This isn't even counting the additional 4467, or 1.94% of the electorate, who voted informally (i.e. didn't bother to fill out the ballot properly).

Now, let's look at the seats that changed hands at the 1998 election. How many of these were cases where the margin of victory was less than the number of people in Lindsay who appeared to be voting as a joke?

In Bass, Tas, the margin was 0.06%.
In Dickson, Qld, the margin was 0.12%
In Kingston, SA, the margin was 0.46%
In the Northern Territory, the margin was 0.57%
In Stirling, WA, the margin was 1.04%
In Patteron, NSW, the margin was 1.22%

Six seats, where the victory was within the margin of joke voting. What a triumph!

In the most recent federal election, in 2010, the Labor Party ended up forming a coalition government with a majority of only one seat.

Meanwhile, the seat of Corangamite, Vic, was decided by 0.82% of the vote, and the seat of Hasluck, WA, was decided by 1.14% of the vote.

It is entirely possible that not only the outcome of a few seats, but in fact the outcome of the entire 2010 election, was decided by morons voting at random.

Justin Wolfers is a highly-trained economist, and a very competent statistician. It would amaze me if he weren't offended by this kind of forced noise in the voting process.

Even if it increases the civility of debate, it seems like a pretty steep price to pay.

No comments:

Post a Comment