The importance of context

Survey the mainstream media, and you get a picture of a stalwart government exposing the softness of a feckless opposition. Labor must win 16 seats and yet is at risk of losing more in WA. There are even pro-Liberal blog posters suggesting Rudd will be sent to the scrapheap. Rudd's worst outcome will be like Whitlam in 1969. Unless he makes an absolute hash of things he will win no matter what.

Therein lies the problem. Survey the political landscape and the thing that is exposed is the lie of the land. Labor need 16 seats to win, but hold an extraordinary primary vote of around 48%. The Lib/Nats have only got above 40% on a handful of occasions. That primary vote, even allowing for Green defectors from 2004, suggests they are going to take around 5-6% of the primary vote off the Lib/Nats unless something dramatic happens. The stratospheric figures are based in NSW and Victoria where most of the seats are.

The Lib/Nats have been forced to run an in-out leaking campaign, creating the effect that the overall vote is either soft or concentrated in seats with little influence on the outcome. Anyone with a calculator can tell you that either someone's polls are dodgy or this is patently not true.

On paper, the Lib/Nats should be seeing WA as the great saviour. But even here they are reliant on dubious polling to push the idea the individual electorates are not following the state trend. The fact that WA statewide is swinging 5-7% to the ALP is a nightmare scenario for the government. Given WA is full of Lib seats, which seats are in danger - most probably Canning? If this is how the great pro-AWA boom state treats them, what about the weaker states?

The best poll in any marginal on the list for the government is Braddon - here they are just 2% behind. Given the Mersey hospital, the pulp mill and the tax cuts and the negative advertising on unions, Labor still lead with a swing of 3-4%. Ouch! I suspect they're not thinking about Bass next door, because Mr Ferguson is heading to the bottom of the adjoining strait. In fact, not one external poll has been released for a seat within the 16 seats that is not in WA that says the Libs are in front. We won't mention seats like Leichhardht in North Queensland, where you could swear it was Labor with the 10 point margin.

As for Bennelong and Wentworth, in a conventional election, Howard and Turnbull's opponents would be playing tagging roles keeping their key ministerial opponents at home. Because of the current climate, Bennelong, Wentworth and Sturt all feature winnable battles for Labor against ministers.

Conventional wisdom is predicting a polarising election. One very interesting nugget of information found in Newspoll's issues survey from last week found that 19% of voters preferred an 'other' party to run the environment - i.e. the Greens. An election run primarily on climate change, water and the pulp mill could actually lead to a understatement of the Green vote, rather than an overstatement. It may also lead to Green preferences going back to the Libs rather than straight to Labor.

Remember that Howard held the first week - mainly because Labor has set itself for a long campaign and kept its powder dry. The only potential risk for Labor is it is seen as unrepresentative, as that counters the 'out of touch' theme which Labor lays its store in. Labor has issues from interest rates to workchoices, climate change, health, child care and nuclear power to run on. It also has an awful lot of money left to spend.

The key is whether these polls swing back to Labor when their issues are on the table. If this is the Lib/Nat's high watermark, they are currently behind in all but 1 seat on the fabled 16, and in deep trouble in Dobell, Deakin, La Trobe, McEwen, Corangamite, Leichhardht, Sturt, Blair. That's not counting the great worries in seats like Ryan.

Another couple of bad weeks for Labor may tell us something, at the moment we should suspend judgment while the parties gear up for a long campaign