I've asked a circle of friends to make their prediction of the Election Result. It will be interesting to see the range of expectations.

Given that I've asked them to make predictions, I guess I'm honour bound to make a prediction myself.

(Friends: please make your own predictions before you read below to see what I think!)

Despite earlier doubts, I've come to a fairly firm expectation of an ALP win. But, as I've repeatedly said, it'll be hard for the ALP to gain seats.

To win government, some seats will have to swing a long way. Now: that may be on the cards since there are reports that the Liberals are doing private polling of seats like Kooyong and Goldstein. See Possum's Pollytics article on the private polling here.

To get into that territory, the ALP would be gaining around 30 extra seats to get a total of 90 or more, with a swing around 10%.

That may yet be possible, but it's not my current expectation.

I reckon that the young are turning against the Liberals. And the so called Doctors' Wives. But in the mortgage belt seats with margins of 4-8%, the swings won't come so quickly.

This is backed up by a seat by seat analysis of the betting market put together by Simon Jackman.

For those of you unfamiliar with the concept of betting market analysis, here's a quick summary: The major pollsters rarely do a statistically valid sample of any individual seat (1000 voters). Studies have shown that the best way to get a prediction is to look at how people are betting. Those who bet are usually well informed, and they gather all the information they can: Polls, studies of how the seat is changing, anecdotal evidence. For example, Malcolm Turnbull's seat of Wentworth looks marginal on 2.5%, but the punters know that his margin was reduced in the last election by a strong independent liberal candidate. So it's not as marginal as you might suppose.

Jackman's most recent analysis shows a seat by seat expectation that the ALP will gain 14 seats... possibly only taking them to 74 seats out of the Parliament of 150 seats. If that were true, the ALP would only get in if they got the support of the likely two independents.  (Both former Nationals, but good "Country Socialists" who might support the ALP.)

The current state of the market can be found here. (I suspect that this document may be liable to change. As of 14th October, there are 14 coalition seats in the top left quadrant. That's the quadrant which shows coalition seats that are more likely than not to fall to the ALP.

In making my prediction, I'll go with those 14 seats, and add the four seats closest to 50/50 betting as likely to fall: Bowman, La Trobe, Herbert, Deakin and Stirling.

That means I'm predicting that the ALP will win fairly narrowly by ending up with a total of 79 Seats.

Things might tighten or widen during the campaign, so I reserve the right to change my mind. But as of today, the day the election was called, my prediction is that the ALP will win 19 new seats and take government with a majority of four seats.

Comments (0)
Anthony Holmes October 14th, 2007 10:26:09 PM

No Comments Found
 Add a Comment
Web Site:
Comment:  (No HTML - Links will be converted if prefixed http://)
Remember Me?