Plenty of room for Rudd's souffle

Galaxy's latest poll, which either shows we are all on Mars or David Briggs' outfit is polling the Andromeda system, provides just a dribble of sustenance for those Liberals bravely clinging to the-historic-16-seat margin argument as a reason why Labor will not win.

It is true that 16 seats is a very large buffer to make up. But it is by no means unprecedented. Labor oppositions have traditionally come off shockingly low bases, hence they have needed two steps to get close enough to win power. Whitlam won 18 seats in 1969 with a 7% swing. Hayden, while winning only 13 seats in 1980, did so from an appalling position. He received a 4.2% swing, mainly from minor party voters swinging behind him.

Hayden's result would give Rudd a statistical tie - 51.5 to 48.5 is the exact point on the pendulum where there is no clear majority. Whitlam's result gives Labor a 54.4to 45.6 win - in short a landslide. This is also almost exactly the current average of the four major polls.

History shows that when the opposition gets an insufficient swing to govern, the average swing since 1949 is 2.25%. That swing would give Labor 49.6%. Clearly, this is not your average election. Rudd's standing and perceived competence is daylight compared to Latham and he is on track to at least pinch votes from the Liberals. A conservative estimate would be a swing of 4.5% - and even that includes 1.8% for the anti-Latham swing, 2.25% for the average swing and only 0.5% for other factors.

Latham's result in 2004 was appalling. Labor achieved just 37.6% on primaries. Interestingly, because of the post-modern third party politics era, that awful primary left Labor at 47.3% of the two-party preferred vote. The Liberal-National vote cannot go much higher than its 2004 position. Labor can thus move from a terrible position to a very good position with a relatively small change in primaries.

The minimum primary vote swing from the polls is 4.4% (Galaxy) ranging up to 10.4% (AC Nielsen). Using the unscientific approach of avergaing these results gives a 7.4% swing - i.e. 45% on primaries. That is in line with a 1969-style swing.

Both in polling and qualitative terms, this election's trend looks more like 1969 than 1980. Rudd is not taking votes solely from minor parties - in fact he may be losing some of them back to the Greens. We know that 35% of unionists voted for Howard last time - bet they will not do that after being labelled thugs. We also know that there is a big swing among mortgage holders who may well switch straight from one major to another. This points to Labor gaining the majority of its new votes from the Liberal-Nationals and putting it in the box seat to take advantage of Green preferences to win marginal and safe seats alike.

On this basis, I predict that Labor will get around 45% of the primary vote, the Liberal-Nationals around 40.5% and the Greens around 9%. Labor to win 54.7-45.3% on two-party preferred around 90 seats.