Photo: National Archives of Australia, A1500, K21454
WARM THE SET and cool the tinnies. While 2009 was a shocker of a year for Australian election watchers – with just one (albeit an interesting one) in Queensland in March – the next twelve months will yield four fat contests, including the fattest of all, a federal election. Tasmania and South Australia vote in March, the federal poll will probably be held somewhere between August and October, and Victorians cast their ballots in November.
General expectations about all but Tasmania are clear: Labor incumbents will romp home. According to online betting markets, in South Australia, Victoria and federally there is only about one chance in five of a Coalition/Liberal win. The Tasmanian contest is viewed as much closer, with Labor slightly favoured to hold on.
I don’t see it this way. In fact, I’m tipping two changes of government in 2010.
One of those is in Tasmania. At the 2006 election Labor won a majority again, to the surprise of most, with fourteen seats against the Liberals’ seven and Greens’ four. The government has been in trouble for a while now, but it is difficult to imagine a defeat because the Liberals are very unlikely to get more seats than Labor and the Greens combined. But if the Liberals emerge on election night with a clear plurality of seats (that is, more than any other party), the Greens may decide to support them – for a while at least. I see this as the most likely outcome.
Over in South Australia is the youngest state Labor government in the country, at eight years old. Its 2006 win set records and it might be expected to hold on easily – except for two factors.
Since the change of federal government in 2007, state and territory elections have seen big swings against incumbent Labor. This has happened in the Northern Territory, Western Australia, the ACT and Queensland, and in most cases followed a huge win at the previous election. We can recognise the existence of a pattern – factor number one – without fully understanding its causes, so I reckon SA will suffer a huge swing in 2010. From 57 per cent of the two party preferred vote (its 2006 result), though, it should still be safe, except for factor two.
South Australia has a unique electoral redistribution mechanism. For the last two decades the Electoral Office has, after every election, adjusted the boundaries to make that result retrospectively “fair” – so that if the previous election had seen equal support for both parties they would have received the same number of seats. The idea is to make the next contest “unbiased.” The aspiration is admirable, but it is not without problems and generally works in favour of oppositions.
This mechanism was one reason Labor took government in 2002 with around 49 per cent of the vote against the Liberals’ 51. (The other was independent Peter Lewis’s betrayal of his conservative electorate.) In fact, if the 2002 booth-by-booth results were replicated in 2010, the Liberals would probably win by a large margin.
In a close contest vote-wise, the South Australian Liberals could prevail with a minority of the vote. I still favour the Rann government to win in 2010, but I’m not nearly as sure as other observers.
Next up is the federal election, where virtually no one expects the Coalition to prevail. Not only do governments in this country usually get at least two terms, but Kevin Rudd is the most approved-of (“popular” overstates it) prime minister in our history and polls have consistently put his government a long way ahead. But what might be a reasonable expectation of the result?
The Labor opposition’s massive poll margins throughout 2007, sometimes more than 60 to 40, did deflate to 52.7 to 47.3 on election day. Something like this could happen in 2010. Then again, many electors who decided at the last minute to stick with the Howard government probably opted for the status quo over the unpredictability of change. With Rudd and Labor now the “business as usual” option, the trend might be the other way. Then again, probably some people simply wished to get rid of Howard, are already tired of the new government and are willing to go Liberal again.
There is also the Tony Abbott factor. The current opposition leader has many attributes, but he does not radiate assurance and overall continuity, something opposition leaders generally must. The combination of likability and a hint of danger could work both ways. If on polling day the prospect of his becoming prime minister seems real, he could do very badly. But if Rudd seems destined for a thumping win, Abbott could get a decent encouragement vote.
Another variable is that the 2007 outcome, eighty-three seats out of 150 for Labor, was a historically poor return for a 52.7 per cent vote. (In 2004 an almost identical vote reaped Howard eighty-seven seats.) Redistributions in New South Wales and Queensland have increased the government’s notional position to eighty-eight seats. As well, the many Labor seats gained in 2007 will be boosted by the personal votes attracted by the new MPs, a reason governments generally do well in the marginals. So the government could increase its majority despite a small swing to the Coalition.
All in all, a federal Labor victory is highly likely, and an increased majority (that is, greater than eighty-three seats for Labor) is more likely than a decreased one.
Finally there’s Victoria. Since taking over mid-term, John Brumby has been a very popular premier, recent surveys have his government a long way ahead – further ahead in fact than at the last election – and the opposition does not impress.
But it’s important to remember that by November the Victorian government will be eleven years old. The Queensland government was about that vintage when it went to the polls last year. Twelve months earlier, Queensland Labor was 60 to 40 in the polls. At the previous election they had won 56 to 44; on election day Anna Bligh won with 51 to 49. If either of those shifts is replicated in Victoria, Brumby is in trouble. I expect opinion polls to narrow markedly in Victoria in the coming year (and probably, at some stage, for the opposition leader to change).
This all adds up to opposition parties doing better than most commentators expect in all three state elections in 2010. I slightly favour the Liberals to form government in Tasmania, and Labor to lose in either South Australia or Victoria, with a small possibility of them losing in both.
It is foolish to make predictions without equivocations. So let me equivocate: these are my best guesses at what will happen. I obviously all but ignore the day-to-day goings on in each jurisdiction because they are not usually very important, and even when they are, the implications are difficult to measure. But outright blunders can occur, and if, say, the Victorian Liberals install Chopper Read as leader, all bets are off.
At the end of the year I will graciously accept lampoons from folks whose predictive prowess has proved superior to mine. Fence-sitters, however, need not apply. •
Peter Brent is a researcher at ANU and editor of Mumble.
5 Comments
Oh dear, if even Peter doesn’t quite understand the SA redistribution methodology it really does need explaining. Unlike just about every other boundaries commission, the Electoral Districts Boundaries Commission has been given an outcome requirement – it is required to produce a set of seats that will, “as far as is reasonably practicable”, allow a party to win a majority of seats (and therefore government) only if it also wins the support of a majority of voters across the state. This “fairness requirement” was put into place to address the fact that geographical concentration of Liberal voters in country areas meant that previous commissions had unknowingly regularly created a biased set of seats. (Federally, Labor was similarly disadvantaged in city areas in the 1970s.)
The EDBC does look backwards after the election to assess whether the party that won government did indeed have the support of a majority of voters. But then it looks forward to the next election to draw a set of seats that will produce a fair outcome (i.e. will not be biased towards one party) at the next election. It isn’t concerned to make the last lot of boundaries “retrospectively fair” – you can’t change boundaries once they are in effect.
I am not sure why Peter thinks the system “generally favours oppositions”. That is not an opinion I have heard from the parties here. In fact by ignoring any advantages of incumbency there might be an inbuilt advantage to the party with the most seats – the government.
Finally if you transfer the 2002 (and 2006) results to the boundaries that will operate in march at the 2010 election (and thereafter for the next 4 years) you find that at a 50:50 election outcome the Liberal Party was not advantaged at all. And at a 50.9% 2PP result – i.e. the 2002 result – on the new boundaries the Liberals would win just 24 of the 47 seats – just a bare (but fair) margin, not a large one.
I reckon you’re on the money in South Australia. There’s an awful lot of angst about, especially over water. After Massachusetts, nothing is certain in politics. A lot of people see Rann as arrogant and soiled goods. Redmond’s a bit rough and ready but regarded as a cleanskin. I was at lunch yesterday with 21 members of the chattering classes and it’s doubful if there were more than one or two votes for Labor among them.
For all that, the Liberal team is pretty woeful and, looked at objectively, the government really should be returned. It could scrape back in - or suffer utter rout.
Hi Jenni. I know (a) you’re writing a PhD on this and know much more about it than me and (b) are a proud South Australian. However …
I appreciate that projections are made of future demographic changes and so on, but if I understand correctly, calculations are based solely on how people voted at the last election. What else could they be based on? Or are elections before that used as well?
I believe it favours oppositions for reasons outlined here. However, as the SA landslide was at the last election perhaps this effect will be minimal.
Finally, do you mean you have plugged in the 2002 booth results into today’s boundaries? (As opposed to simply assuming uniform swing and plotting 50.9 against today’s pendulum.) If so, then I stand corrected. I haven’t been able to do that, which was why I inserted “probably”.
Cheers, Peter
Hi Peter, as you say the boundaries are tested out using the voting data from the previous election, although in 1993 there was such a landslide that the most recent data couldn’t be expected to show a normal pattern of voter support for the parties, so earlier figures were used to see how the marginals looked. The data are adjusted to take into account demographic changes expected to occur by the time of the next election.
Your point about the Commission reducing a “real” ALP lead of 9-10% back to 7% highlights the fact that one of the things the Commissions need to look at is poor responsiveness that creeps into small scale systems. With a 2PP result of 56.8%, the ALP should be able to suffer a swing of up to 6.7% and still hang onto government, but would it be fair to anyone if there was a much bigger swing and they still held onto government? Poor responsiveness in the state electoral pendula and also in the smaller states’ federal pendula can really disadvantage parties if they need to win much more than 50% of the vote to win half of the seats or the median seat. Poor responsiveness not addressed in the federal redistribution in NSW means the ALP there needs to win more than 51.3% of the 2PP vote just to win half the seats; in SA the Liberals need to win more than 52% to win the median seat federally. My feeling is that we should see anything more than about 51% as bias. Commissions can address poor responsiveness but only if they are allowed to look at the political effect of the lines they draw.
As for the estimates, yes I have re-sorted booth and declaration vote data for state and federal elections into the new boundaries. Not widely distributed as not many people (non-MPs) are all that interested, but available on request!
Cheers,
Jenni
Correspondence being continued by email …
2 Trackbacks
[...] just my opinion. Posted 9 February 2010 by Peter. Comments and trackbacks are open. Follow the comments feed. [...]
[...] in all this is mildly encouraging for the Victorian leg of 2010 anticipations. Posted 14 February 2010 by Peter. Comments and trackbacks are open. Follow the comments feed. [...]