From today’s Crikey email:
I’m something of a sceptic when it comes to predictions. As I’ve argued at length, we don’t really have access to the sort of data we need to do so accurately, and on a broader level, elections are once off events and public opinion is constantly changing. This is little understood, and we see avoidable errors constantly in commentary on elections. It took some time for some pundits to realise that it wasn’t the exact 16 seats on the blue side of the pendulum Labor needed to win, but any 16 seats. Though such egregious mistakes aren’t universal (and perhaps their authors are too busy sending flirty emails to bother with a spot of psephological reflection), we still hear constantly “analysis” which assumes that the 2004 results are some sort of baseline.
In a sense they are, but in a real sense they’re not. Someone else (and I apologise for the lack of attribution but I’ve read far too much commentary during the course of this campaign to remember where it all came from) used the analogy of a test match. If Australia scores 150 in its first innings, its starting point in the second innings isn’t 150 but 0. Each and every innings is independent of each other one, just as each toss of a coin is.
Of course, the analogy isn’t exact. In one way, it’s inexact because unlike in successive innings, where you’d imagine that the batting line up is going to be the same under almost all circumstances, the players in each election are often different.
Labor, as I’ve argued since April this year, was probably going to win in 2004 had they had a credible leader. Mark Latham wasn’t that leader, but aside from his brief reappearances in the Fin Review, he’s not a factor this year. Despite all the millions of words that have been written in this election year, the story is relatively simple. Howard succumbed to hubris and introduced WorkChoices, and Labor’s vote surged past the Coalition’s when Rudd became leader. And that’s where it’s largely stayed. For all the talk about undecided voters, there’s more evidence around that most campaigns don’t make too much difference to the final result.
Of course, the “media narrative” doesn’t always reflect what the available evidence tells us, as it’s likely to be blown off course by minor fluctuations and shaped by self-interested spin from both parties. So we’ve seen in the dying days of this campaign incredible claims that Labor could lose with 53% of the 2PP vote. Not likely. It’s worth noting that this sort of talk – largely from a desperate Coalition – is probably more helpful to the ALP than the Libs. Similarly, the leaks from Labor strategists down playing their chances are a classic play to regain the underdog crown, negate complacency and gee up the troops. It astonishes me that they’re not reported that way.
The truth of the matter is that John Howard succumbed to hubris when he thought he had Kevin Rudd’s measure on the day Rudd was elected leader. He’d have done much better to have done what he did with Beazley – ignored him and got on with the business of governing. But he couldn’t resist the urge to take Rudd down and almost a year’s worth of shrill and tedious ranting and raving from the government and endless wedges and rabbits has been his own undoing.
I think John Howard himself believes he’s going to lose.
So predictions? I don’t have any for you. My educated guess is that Labor will win by quite a respectable margin. My tip in Richard Farmer’s comp is Labor with 87 seats – meaning a 2PP of just under 54%. My own reading of the campaign is that it’s been in the bag since Wednesday last week when Rudd called a halt to the spendathon. Labor may do better, or a little worse, but I’d be immensely surprised if Rudd isn’t claiming victory on Saturday night. Your guess, of course, is as good as mine, and that’s the beauty of a democracy!