UPDATE Saturday 2.25pm
Nothing has happened to make me revise my views from yesterday. A preview this morning from Peter Brent, one of our best psephologists, tips “a comfortable Coalition victory,” guessing about 85 seats (one of Brent’s great virtues is that he admits when he’s guessing). I’m thinking very much the same ballpark, if a tiny bit lower – maybe 84 on the Coalition side of the ledger, of which a couple will probably be rural independents (Bob Katter at least, and maybe Tony Windsor or Rob Oakeshott).
That means 66 on the Labor side, a gain of eight, but I include in that Greens and Labor-leaning independents. The actual Labor total would be more like 62.
The Senate is anyone’s guess. My estimates came out at Coalition 31 (down two), Labor 27 (up two), Greens nine (down one), Xenophon four (up three), LDP/Family First two (unchanged), Palmer and ex-Palmers one (down two), others two (unchanged). But that’s very rough.
There’s been a bit of commentary around to the effect that the whole point of the double dissolution was to improve the government’s Senate position, so if it returns another dog’s breakfast of a Senate it will be a failure. But I think that’s misguided; the purpose of the double dissolution was simply to hold an early election without screwing up the timetable for the future. The government wasn’t going to get control of the Senate whatever happened (nor was the opposition), and everyone who had their head screwed on right knew that.
It’s not the Senate result that matters, it’s whether the win (assuming it is one) in the lower house will be big enough for Malcolm Turnbull to claim vindication. And on that question, the spin doctors will be out in force from tonight onwards.
As most readers will be only too well aware, Australia goes to the polls tomorrow [to pass judgement] on Malcolm Turnbull’s Liberal-National coalition (hereafter just “Coalition”) government.
A very quick recap: Australia’s lower house, the House of Representatives, is elected from 150 single-member districts on the typical British pattern, with the refinement of preferential (or “instant runoff” voting). There is also a powerful Senate, with 12 members from each of the six states and two for each of two territories.
The Senate is usually elected in rotation, half every three years, but on this occasion all senators are being chosen at once under a special procedure called a “double dissolution”, last invoked in 1987. The (somewhat complex) mechanics of that procedure also account for the unusually long duration of this year’s election campaign.
At the last election, in 2013, the Coalition won 53.5% of the two-party-preferred vote and 90 lower house seats. The Australian Labor Party won 55 seats, and another five went to minor parties and independents (although they had collectively 21.1% of the vote).
Working out Labor’s target this time is a little tricky, so bear with me. Of its current 55 seats, one (Charlton) has been abolished in a redistribution, but the Liberal-held seat of Barton has been restructured to such an extent that it is now clearly a Labor seat. Another two seats with Liberal members, Dobell and Paterson, now also have notional Labor majorities, but only by tiny margins, so I will continue to count them as Liberal seats.
To Labor’s baseline of 55 I would then add the one seat held by the Greens (Melbourne) and two held by independents (Denison and Indi), on the basis that all three are more likely to support a Labor government in the event that they hold the balance of power. That brings Labor to an effective 58 seats, meaning it needs to gain 18 to win a majority.
Looking at that equation the other way, to the Coalition’s current 90 seats we can add the new seat of Burt and the Palmer-held seat of Fairfax (Clive Palmer is retiring, and without him it is a safe Coalition seat), and subtract the restructured Barton, already mentioned. It can also probably count on Bob Katter, who holds the seat of Kennedy for his eponymous Katter’s Australian Party. That gives a total of 92, meaning it could afford to lose 16 seats and still hold government.*
A gain of 18 seats would require a uniform two-party-preferred swing of 3.8% (Antony Green has the pendulum here). But because the Coalition was the party making gains last time, most of its marginal seats have new sitting members, who can expect to benefit from a “sophomore surge”. So the real swing that Labor needs is probably a little over 4%.
I join with most observers in thinking that that target is out of reach. My colleague William Bowe calculates an average of 50.9% for the Coalition in the opinion polls, representing a swing of 2.4% [sorry, 2.6%] – which according to the pendulum would produce a gain of just seven [should be eight] seats.
There is a heavy concentration of seats with margins just above that point, so in reality I suspect a 2.4% swing would do a bit better than that for Labor. On the other hand, the movement in the last week seems to be towards the government, particularly with the uncertainty created by the British referendum result, so I think 2.4% is probably on the high side of expectations. A swing of between 1.5% and 2.0% seems more likely, for a net gain to Labor of between five and ten seats.
That’s still a close result, and I don’t discount the possibility of Labor doing significantly better. But I don’t think the seats are there for it to win a majority, or even particularly close to one. If Labor was going to win government, you would expect there would be eight or ten seats at least where it looked comfortably ahead, and I’m not seeing that.
Trying to pick specific seats is always fraught, but the only ones I’d say Labor should be confident about would be Paterson (-0.3%), Dobell (-0.2%), Capricornia (0.8%), Lyons (1.2%), Solomon (1.4%) and maybe Eden-Monaro (2.9%). I’d also give them a better-than-even chance in Petrie (0.5%) and Hindmarsh (1.4%), and a fighting chance in Page (3.1%) and Macarthur (3.3%).
There’s also some chance of gains in Western Australia, where Cowan (4.5%), Hasluck (6.0%) and Burt (6.1%) are all possibilities, but none of them look easy. As against that, Labor has five seats in Victoria at risk – McEwen (0.2%), Bendigo (1.3%), Chisholm (1.6%), Bruce (1.8%) and Melbourne Ports (3.6%) – so while it’s probably the favorite in each of them, you’d have to expect at least one will fall.
There’s also the chance of seats falling to third parties: Labor has seats under threat from the Greens, while the Coalition has seats at risk to rural independents and to the Nick Xenophon Team in South Australia. In each case they are not expected to alter the balance between Labor and Coalition, but if there is a stronger-than-expected underlying swing then they could produce a parliament where neither side can be confident about its majority.
That would be bad news for Turnbull; if the result gets anywhere near hung-parliament territory, his position within his party will become very difficult indeed, since many (perhaps most) MPs only supported him as a guarantor of electoral success. But the long campaign seems to have played to his strengths, and last week’s news from Britain was a gift from the gods, so I’m guessing he will be out of danger – at least for the time being.
I’ll update tomorrow with any last-minute news, and maybe some thoughts about the Senate.
* Note: a loss of 17 seats would mean a tie, an unlikely but not impossible occurrence (why do people insist on having even-numbered legislatures?). That would probably result in the continuance of the Coalition government, since Cathy McGowan in Indi is the least obviously attached of the independents, but it would have a precarious existence.