The Narrowing, (not) by Dean Koontz

Graph of two-party preferred preferences since the campaign started

The Narrowing. The idea that during an election campaign voters return to the incumbent government. Supposedly the reality of an actual vote, as opposed to mere opinion polls, triggers voters’ fear of the unknown. As this graph shows, if there is a Narrowing, it’s bloody tiny this time around.

The Narrowing is nothing but mythology.

In the 2001 campaign, Kim Beazley started from behind and gained 5% before voting day — not enough to win, but enough to give him hope for next time. That’s a shift against the incumbent party, of course.

Of course that loose-mouthed thug Mark Latham went and screwed all that up. But this time we can see what the electorate really thinks of Howard now that they’ve got a credible alternative.

As the graph from Possums Pollytics shows, yes, you can sort of see a little sign of The Narrowing. But that gentle glidepath has to cross that line marked “50”.

Yes, the Coalition might be able to claw back enough support to win. As long as the election is on 28 July 2008.

I gather the election is sooner than that.

4 Replies to “The Narrowing, (not) by Dean Koontz”

  1. I noticed that if the starting point of the data was the 19th of October instead of the 15th then possibly the regression would be widening instead of narrowing. I remember from my trading days regression lines were very bad predictors because of starting points and chosen data set. Just an aside 🙂

    That said, there’s 5 points of pain for the Libs consistently sitting there! Ouch!

  2. @Marc Lehmann: Oh absolutely, even the half-subject of statistics I did some million and a half years ago told me that linear regressions are pretty much useless when dealing with data that doesn’t change in linear ways. Still, it’s useful to give an indicator of “Is it going up or down?”

    I think the real message of the graph is the one the analysts have been saying all along: the polls are consistently showing Labor to be roughly 10 percentage points ahead of the Coalition, and they have been all year.

  3. I was surfing Dynamic Business website today (they were pitching add space to us) and i hit a Howard versus Rudd poll. For what you would think is a right of center user base of that website that poll still had Rudd ahead 5 points. That’s about when I thought I should hit the centrebet website only to see a dead cert $1.35 price vs $3.20 for Libs. That’s the trader in me 🙂

  4. @Marc Lehmann: There was a comment in Crikey the other day (look it up yourselves, people!) to the effect that if the polls are this certain about Rudd winning, then there’s still some good prices to be had.

    OTOH, as of right now Rudd has narrowed to $1.28 on Centrebet. And they’ve turned their home page into two big “Bet Now!” buttons.

Comments are closed.