I've been saying regularly that national opinion polls are reflecting people's belief about who will win the upcoming Australian election and not providing an accurate assessment of voting intentions. That's in spite of the unsound question asking whether voters have already made up their mind for whom they will be voting returning a strongly positive response.
To get an accurate view of what's going on one must look at what's happening on an individual seat level and factor in differences between how states will vote. Western Australia, for example, is looking much stronger for the Coalition than, say, Tasmania and it's possible they could actually pick up a seat there - though I haven't allowed for that in my calculation. Everyone who is predicting a "Ruddslide" are applying the high swing to Labor predicted by opinion polls uniformly across the nation. Unfortunately for them I don't expect the swing to be uniform at all.
I'm going to be the first pundit out of the blocks to predict a Coalition win with the following number of seats for each party:
To those looking to enjoy the sight of John Howard's concession speech I can only suggest that it'll be Don's Party all over again.
While the government's strength in the west is clear the irony is that it will be his home state of Queensland that will lead to Kevin Rudd joining a long list of Labor leaders who have come off second best in a head to head contest with John Howard. While the PM has been looking calm and relaxed over recent weeks, Kevin Rudd has been looking decidedly stressed and it's my view that it's because he knows what internal party polling is telling him regarding Labor's election chances.
Here's my assessment of the state by state result:
The following are the seats that I expect the government to lose:
NSW: Dobell, Eden-Monaro, Lindsay, Parramatta
QLD: Blair, Bonner, Moreton
SA: Kingston, Makin, Wakefield
TAS: Bass, Braddon
The key seats that I expect the government to hold, which are currently under threat are:
NSW: Bennelong, Wentworth
VIC: Deakin, La Trobe
WA: Hasluck, Stirling
Of course, I only need to get two seats wrong for the government to fall, as I expect that the two independents would side with Labor should there be a hung parliament and I'll be having a little bet on that outcome as soon as a market is available.
Bennelong and Wentworth look bullet-proof in spite of the wishful thinking of many that Howard and/or Turnbull could lose their seat. Solomon may well be the seat I've called incorrectly but both Labor and Liberal party polling shows that it will stay with the Coalition so I'll go with that assessment. Herbert requires a 6.3% swing to fall, a big ask in any election, and the Coalition will be pumping resources into the place and promising a whole heap of good stuff to keep the seat on board. Both Labor and Liberal polling suggests that Victoria will remain with the status quo and while Deakin and La Trobe will be closely fought battles I expect them to stay in Liberal hands. In WA, Hasluck and Stirling are pretty close calls but the mining boom in that state will help them remain with their incumbent members, especially after the expected full court press in that state by the Liberals. It's also possible that the Coalition could pick up Swan and Cowan, both of which are held by narrow margins by their Labor members.
The example of WA is why people shouldn't be applying a swing uniformly across the nation.
So there you have it. I'm predicting that the government will be returned by the narrowest possible margin. It will be an interesting campaign indeed.