Antony Green takes my prediction to task. His email follows (I've added his
links); my response will appear in a day or two.
I think you are taking the pendulum far too literally as a guide to the election.
For instance, you list
as the only Gold Coast seat to leave the Labor fold.
Now if you go back through the last decade of elections, the two seats Labor always had a chance of winning were
and Southport. The difficulty Labor always faced was that in Judy Gamin and Mick Veivers, the National Party had two well known and hard working sitting MPs.
Then along came the 2001 election. The swings in
were large, but not as large as in seats like Mudgereeba, which came from the clouds.
I would say Broadwater is the hardest seat for Labor to hold on the Gold Coast. It includes all the exclusive estates around Hope Island. You could also find either
Mudgereeba falling before Burleigh. If you know the history of some of the Queensland seats over the last decade, then some of the margins in previously marginal Labor seats do not seem to make a lot of sense.
Another thing to remember is that at previous big change elections, like 1974 and 1989, newly elected sitting MPs had a good record of holding on against the swing. There were Liberals who won previously safe Labor Brisbane seats in 1974 that were not dislodged until the mid 1980s. Similarly, some of the northern suburban seats Labor narrowly won in 1989 were the seats they narrowly retained in 1995 against awesome swings.
The point I am making is that the swings at the 2001 election were remarkable and not very uniform, and past history in Queensland also suggests that when you get a change of sitting members, swings at subsequent elections can also produce surprises.
Now personally, I think there is no sense talking about two-party preferred at the moment in Queensland. The 1998 result produced a pattern of swing that was determined entirely by which party ran third in each electorates. There was a 4.6% swing to Labor in two-party contests, but a 5% swing against Labor in seats where the final contest was between Labor and One Nation.
At university, you are taught about the “paradox of voting”, or “Arrow’s Impossibility Theorem” as the economists call it. The 1998 Queensland election came closest to being an example of this theoretic possibility as I have ever seen..
There may be something you can class as a statewide swing at the election on 7 February, but I would not expect seats to fall in the order determined by ticking up the pendulum. While as Malcolm Mackerras rightly argues, the extra seats that fall often cancel out the seats that don’t, there is
no intrinsic law that makes this happen. It just tends to happen.
I no doubt will get calls the day after the election asking what my two-party preferred vote for the election is. To anyone who might think of doing this, don’t bother. I don’t even have a two-party preferred vote for the 2001 election. It just doesn’t make a lot of sense to calculate a two-party preferred vote when you have a dozen seats where Coalition candidates finished third and even fourth on the primary vote.
Labor in all probability will lose seats. There are seats like Keppel they may yet gain. But just don’t expect seats to fall in the order of any of the various electoral pendulums floating around at the moment.
Bugger: accidentally uploaded a rough draft of my response. (Second time I've
done that!) Will repost when it's finished.