One opinion poll on its own is basically meaningless, but since the budget there have now been four – one each from Galaxy, Newspoll, Morgan and ACNielsen. Can we now say something sensible about a trend?

At the micro level, it looks confused. One poll (Newspoll) shows a post-Budget movement to Labor, the other three a movement to the Coalition. But the macro-level message is unmistakable: the last three months of poll movements, before and after the budget, have been statistically insignificant.

Labor’s support is holding steady at somewhere around the 57% mark (two-party preferred). (As usual, Bryan Palmer’s graphs are indispensable.)

There are all sorts of reasons why that lead might not hold up for the next six months – the most obvious being the brute fact that, in our modern experience, federal oppositions just don’t get 10% swings. But the actual evidence for any easing of Labor’s position is very slim.

In light of that, it is absurd for the prime minister to maintain, as he did this morning, that Australians are “contemplating a change of government”. On the available evidence, they’re not contemplating, they’ve decided. They might change their minds, but something will have to make them change.

Glenn Milne, a commentator I don’t often agree with, argues today (presumably without having seen the Nielsen figures) that Howard’s air of complacency is hurting the government, and feeding an “insider” perception “that Rudd’s overwhelming lead in the polls is an aberration.”

Citing “senior Liberals”, Milne says “Howard must embrace the obvious; he must go out and get the message across that Rudd is the undisputed frontrunner for the election and that at this late stage of the cycle a Labor victory is more likely than not. He must reassure the electorate that he knows he’s in trouble…”.

Of course, insiders are not always wrong. The evidence for a movement back is slim, but not completely absent; both Nielsen and Galaxy have the government in a better position than it’s been since the beginning of the year. (Nielsen’s headline result actually shows no swing, but in light of its primary vote figures that is not believable; the two-party-preferred swing is probably 1 or 2% to the Coalition.)

But the government will have to do much better than that to deal itself back into the game.