I just noticed the betting market odds have moved slightly.
At the beginning of March 2005 Centrebet had the Coalition on $1.45 for the win, and Labor on $2.55. Giving implied win probabilities for the 2007 election of 64 per cent for the Coalition and 36 per cent for Labor.
This morning the odds are $1.50 for the Coalition and $2.40 for Labor. The implied win probabilities are 62 per cent and 38 per cent respectively.
The Australian (also here) has the latest Newspoll results from last weekend. The full set of statistics should be available on the Newspoll website later today.
Those who follow this site closely would know there has been some volatility in the opinion polls lately. Both Morgan and ACNielsen have produced a recent poll favourable to Labor, although Morgan’s most recent poll saw a return to the post 2004 election status quo. The question we have pondered is whether the two pro-Labor polls from early March 2005 were the beginning of a trend or just random noise.
Today’s Newspoll suggests the latter interpretation is correct. It increasingly looks like those two polls from early March were blips in an otherwise unchanged pro-Coalition electoral mood since the 2004 general election.
The headline statistics for last weekend are:
Note: While this poll suggests a slight movement in opinion towards the Coalition, this is likely to be random noise in the opinion polls. I would want to see a couple of other polls bounce in the same direction before I was convinced there was a genuine movement to the Coalition in public opinion.
Check out my opinion poll charts (you may need to hit the refresh button on your browser to get the latest charts).
The post poll reporting includes some beauties: errors of fact and overly excited analyses.
The key Labor message has been the result is a kick in the pants for the government over interest rates. According to this analysis, the swings to Labor have been greatest in the seats with the largest mortgages. For example, Damien Murphy said this,
Voters living in so-called McMansion estates that are changing the demographics of south-western Sydney have given Labor more ammunition to attack the Federal Government over interest rates after the by-election in Werriwa on Saturday.
In suburbs such as Hinchinbrook, Cecil Hills and Prestons, where many are paying off mortgages of a least $550,000, the successful Labor candidate, Chris Hayes, attracted swings of up to 15 per cent, a substantial improvement on Mark Latham’s vote in October.
Mr Hayes said the by-election, in which Werriwa remained strongly Labor, had put an end to the drift away from traditional Labor-voting patterns. The ALP was taking heart from the McMansion votes, he said.
Glenn Milne and Andrew West made the same 15 per cent claims. At least Milne and West did not mention Prestons, where Labor went backwards on the primary vote. Excluding the 36 formal votes collected by the special hospital team, the largest swing to Labor in the primary vote for a booth was under ten per cent.
A number of journalists portrayed the weekend win as an 11 point two-party preferred swing to Labor. However, the provisional two candidate swing in this by-election cannot be compared with that for the 2004 election. Without an endorsed Liberal candidate in the by-election, this is not an analogous comparison. As Peter Hartcher said, ‘when there is no second party, “two-party” results become a meaningless measure.’
To be fair, the Coalition has been no less self serving in its analysis. Murphy’s piece includes this quote from the Prime Minister.
Mr Howard congratulated Mr Hayes but said the result was a below-average performance for an opposition in a by-election. He said oppositions had an average 6.5 per cent swing towards them in by-elections since 1949.
While Howard is technically right, his analysis confounds those by-elections where the punters get to send a message to the government of the day, and the one horse race by-elections like Werriwa. Also, the by-election result suggests a real vulnerability for Howard. If a 25 basis point increase in interest rates drives a two and a half percent strengthening in Labor’s primary vote (albeit in the context of a one horse race by-election), further interest rate increases could be diabolical for the Coalition. This is particularly so as opinion polls suggested around half the residents of Werriwa saw the recent rate increase as a broken 2004 election promise,
However, it was hubris to portray such vulnerability as the end for John Howard. Milne got beyond himself with this analysis.
… the swing to Labor was about 15 per cent. If even half of that was replicated in the mortgage-belt seats of Sydney, Melbourne and Brisbane that kept Howard in government on October 9, it would be a catastrophe for the Coalition. Howard and other leading government figures will, of course, argue that there was no Liberal candidate in Werriwa, so the big vote for Labor can in no way be seen as a reflection on the Coalition or its policies. Except the Liberals did effectively run a candidate in Werriwa. His name was James Young. He was once a staffer for former Howard government minister Jackie Kelly, who holds the adjoining seat of Lindsay. He publicly offered himself up for the Liberal vote, backed Howard repeatedly, displayed the Prime Minister’s photograph at polling stations and endorsed himself on his how-to-vote cards as the candidate standing for “your Liberal point of view”.
And yet he got just under 8 per cent of the vote. The only conclusion is that the Howard brand has suddenly gone bad among Liberal as well as Labor voters in Werriwa.
Not everyone agreed with Milne. Hartcher saw things more ambiguously.
Labor’s big, long-running problem is that the effects of 14 years of uninterrupted economic growth have recast much of Labor “heartland” as contestable territory.
David Burchell, a political analyst at the University of Western Sydney, said yesterday that “Werriwa was the seat that could have been the next domino to fall to the Coalition, and it remains so.”
Labor’s feeble victory means that the “blue circle” of Coalition seats that has started to encircle Sydney remains a real danger to Labor’s survival. So John Howard cannot take the Werriwa result as any kind of endorsement, but neither can Labor.
Another who got the spin right is the anonymous Labor source quoted by West.
One federal caucus member said the result should not have been such a surprise.
“I say the outcome was solid but hardly spectacular,” he said. “With interest rates up and no Liberal candidate, this result is really about the minimum we could have achieved. Anything less would have been terminal, I think, for Beazley.”
And another who got it right was Antony Green, quoted by Murphy.
The Herald election analyst Antony Green said it was difficult to read anything substantial into Labor’s improved showing because it was too little, and too far away from the next federal election.
“There were six by-elections in 1994 and none of them had any bearing on John Howard’s victory two years later,” he said.
In my view, the Werriwa by-election result was not unexpected. If there is any long term message, it is not a strong one. While it exposed Howard’s vulnerability to rising interest rates, it was no harbinger of a golden age for Labor under Kim Beazley.
One thing is certain, the reporting of the by-election shows that if you stare at the tea leaves long enough you will see what ever you want to see.
Ho hum. The votes from polling booths are almost counted. Although the AEC has not counted postal and pre-poll votes, it looks like Labor will win with an increased primary vote over Mark Latham’s in the 2004 election. No surprises there. My (pre-election) assessment was that it was at least even money that Hayes would do better than Latham. And as it is almost certain Hayes will get more than 50 per cent of the primary vote, there is no need to count preferences.
But the primary vote for some of the other candidates was a little unexpected.
My tip for second place in the primary vote was James Young. He came second, but with around 8 per cent of the vote his performance was much poorer than I anticipated. In the 2004 election the Liberal candidate got 35.11 per cent. As the quasi-Liberal in the by-election, I thought Young would get at least half that. It is entirely possible that the final two-party preferred result (if it gets counted) may not have Young on the other side of the ledger as the AEC (and I) anticipated.
Another surprise (for me at least) was the Christian candidates. At the 2004 election there were no Christian candidates in Werriwa, although Tan ran in 2001 and got 2.46 per cent of the vote. This time they have managed more than eight per cent between them.
More surprising was the Pentecostal aligned Family First candidate, Sykes, appears to have whipped the mainstream protestant aligned Fred Nile Group candidate, Tan. In the 2004 Senate election for NSW, the Fred Nile Group beat Family First 2.38 per cent to 0.56 per cent in the seat of Werriwa. It looks like Sykes will get his money back but Tan will not. [Note: candidates who score more than 4 per cent of the primary vote get their $350 nomination deposit back and (if the 2004 rate has not yet been indexed) $1.94397 for every primary vote they receive.]
Crikey had tipped Deborah Locke to be the candidate to run through the middle of the pack. But with less than four per cent of the primary vote she is not going to get her money back. (I was actually a little surprised when Crikey anointed Locke, I considered her an unlikely outcome).
Australians Against Further Immigration (AAFI) also benefited from the leaked Liberal vote (not to mention the donkey vote). Like the Christians, AAFI was absent in 2004. Janey Woodger contested Werriwa in 2001 and got 1.76 per cent of the vote. This time she should score more than four per cent, enough to get her money back.
Another who might get his money back (if postal and pre-poll votes go his way) with an initial count primary vote just under four per cent is Joe Bryant.
Even the One Nation candidate, Charles Doggett, has seen his primary vote increase by around one percentage point from a base of 2.36 per cent in 2004.
Last time donkey vote beneficiary, Sam Bargshoon, has suffered a reversal of fortune. His primary vote is down from almost five percent in 2004 to around one per cent in 2005. May be there is a donkey vote lesson here for AAFI. Bargshoon’s hospitalisation in the last week may also have been a factor.
But not everything was a surprise.
More predictable was the performance of the 19 year old Greens candidate, Ben Raue, who got under 6 per cent of the vote. In the 2004 election he got 3.13 per cent. At one point some were spruiking a 15 per cent primary vote for Raue. But Werriwa is traditional Labor. It is not the (small-l) liberal, middle-class Labor of Cunningham. And it is definitely not the latte-left Labor of inner-Sydney or inner-Melbourne.
Local radio comedian Mal Lees got a respectable two per cent of the primary vote. Not a bad performance for a joke candidature.
Ex-Labor, and now prospective Liberal, candidate 23 year-old Ned Mannoun should be happy with a vote of almost two per cent.
The Progressive Labour Party’s Pat McGookin scored just a touch under one per cent.
Mike Head from the Socialist Equality Party got much the same result as last year — less than one per cent.
Another who got much the same as last year was Robert Vogler (albeit last time he ran in Wentworth and last time he had the donkey vote position in that seat).
I think Mr Aussie-Stone has just chalked up his 17th Federal election loss and he did not even crack one per cent of the primary vote. Sometimes experience is not enough! No surprises there either.
Update: Others with post poll analysis include the Poll bludger, AusPolitics.Net and Andrew Bartlett.
Scott Bennett, Gerard Newman and Andrew Kopras, from the Parliamentary Library, have completed their comprehensive analysis of the 2004 general election (but be warned: this link is around 4.6Mb). This paper is seriously worth reading.
The key trends they identified are:
the figures over the past two decades suggest that the problems of the Labor Party have been as much long-term as a number of particular electoral failures a measure of Labor’s decline is that it is now the major beneficiary of the compulsory allocation of all preferences on House of Representatives ballot papers the Liberal Party now has a strong grip on the seats surrounding Sydney, and if it maintains this, it will continue to be difficult for the Labor Party to regain office Labor has seemingly lost touch with many of its traditional supporters, a touch that it needs to regain the shifting electoral strength of the Coalition partners suggests future non-Labor victories are quite likely to see the Liberals gaining a House majority in their own right, and Labor not only needs to improve its House performance, for its performance in Senate elections since the last increase in the size of the Senate has been weaker than its performance in House of Representatives elections.
One of the cautions to come out of the paper is the importance of looking at the primary vote in the opinion polls in the light of the Peter Walsh aphorism, ‘a party that cannot get a vote above 40 per cent is unlikely to win a Commonwealth election.’
Perhaps election analysts have become used to focusing on the two-party preferred vote, and now tend to put insufficient weight on the first preference figure. The two-party preferred figure, after all, with its over-simplified reduction of the contest to just government versus opposition, is less complicated for the analyst to deal with. On polling day 2004, the Weekend Australian gave the final Newspoll figures with a story headed: ‘Latham within striking distance’. This was despite the fact that Newspoll gave the Coalition’s first preference lead as six points (39–45 per cent). Labor’s Newspoll figure in the final six polls had fluctuated between 39 and 41 per cent, while the Coalition readings were between 43 and 46 per cent. The Weekend Australian indicated why the election was being called in this fashion. Having noted that the parties were apparently equal on the two-party preferred vote, the journalist explained that although the Coalition lead on first preferences would ‘normally’ see it returned to office, it was the fact that people’s likely preference allocation would strongly favour Labor that ‘could neutralise the primary vote’. Clearly the two-party preferred vote was being focussed on rather than the first preference vote.
The Newspoll figures during the three months prior to the election showed little alteration in the first preference margin between the parties, nor was there any evidence of any voter volatility. The figures suggested, then, that as the Coalition’s first preference vote was healthy, the most likely result was a Government victory.
The researchers devote some space to analysing Labor’s failure over the longer term. They suggest that the 2004 result is as much evidence of a Labor decline since 1990, as evidence of a stand-alone campaign failure. Among the challenges for Labor is reconciling its two constituencies — one (small-l) liberal, cosmopolitan and middle class, the other traditional working-class. An interesting observation in this vein was that the dissonance between Labor voters and Labor candidates is wider than for Coalition voters and Coalition candidates. The paper quoted Katharine Betts of the Swinburne University of Technology.
… the gap between voters and candidates is very much wider for Labor candidates and their voters. Overall Coalition candidates are quite close to their voters whereas Labor candidates are quite distant from theirs. On average the gap between Labor candidates and Labor voters is almost three points wider, in terms of percentage points, than it is between Coalition candidates and Coalition voters.
Judith Brett is quoted as saying the Labor party is seen by the public as:
a self-serving, faction-driven political machine, filled with professional politicians who place the survival of themselves and their factional colleagues above the interests of the people who vote for them.
The movement of the Sydney urban fringe to the Coalition is another striking element of the analysis. In the following diagram the Liberal seats are shaded for each election.
All-in-all, an interesting paper.
