Oz Politics Blog

Morgan’s poll of 1904 voters was taken on the weekends of 7-8 and 14-15 of January. The headline prediction was a national two-party preferred vote of 54 per cent for Labor and 46 per cent for the Coalition. This is Labor landslide territory, should the vote be replicated at the next election.

The primary vote predictions, compared with the last Morgan poll in 2005, are as follows:

• Coalition — 39 per cent (down 0.5 per cent)
• Labor — 43 per cent (up 3.5 per cent)
• Australian Democrats — 2 per cent (up 0.5 per cent)
• Greens — 8 per cent (down 2.5 per cent)
• Family First — 3 per cent (unchanged)
• One Nation — 1 per cent (down 0.5 per cent)
• Other — 4 per cent (down 0.5 per cent)

Although the most recent Newspoll went north for the Government, Morgan went south. The difference in the predictions from the two polls is so great that at least one of the polls is wrong. It looks like it will be a while before we get a handle on the performance of the parties after Christmas.

Just out of interest, I have added linear regression trendlines to this week’s graph.

• The upper trendline is Newspoll, which had the Coalition on just under 54 per cent in December 2004, dropping to 48 per cent in January 2006 (R²=0.46).
• The lower trendline is Morgan. It had the Coalition on 52 per cent in December 2004, dropping to just over 46 per cent in January 2006 (R²=0.42).
• Updated: The steeper trendline starting in the middle of the other trendlines is ACNielsen. It began in January 2005 with the Coalition on just over 52 per cent, and concludes in December 2005 with the Coalition on just over 45 per cent (R²=0.43).

We have discussed the systemic pro-Labor bias in the Morgan series, when compared with the Newspoll series. However, the first two regression lines suggest strong agreement between Newspoll and Morgan: the Coalition’s predicted two-party preferred vote has declined by between 5.5 and almost 6 percentage points over the last 14 months.

Even if you take the more optimistic number from the government’s perspective, a 5.5 per cent decline from the 52.7 per cent TPP vote in the November 2004 election, it leaves a national vote of 47.2 per cent. This would be a landslide win for Labor.

The usual graphs are here. You may need to hit the refresh or reload button on your browser to see the latest graphs.

Correction: When I saw the Crikey link to this post I realised it had an arithmetic error (now fixed). How embarrassment! Thanks to Crikey’s Charles Richardson for not drawing attention to the error.

According to the West Australian Alan Carpenter will be elected unopposed as the Leader of the State Labor Party, and hence Premier of Western Australia.

According to today’s ‘Tiser,

DEFENCE Minister Robert Hill is expected to announce his resignation from politics within days, possibly tomorrow.

The Advertiser understands the South Australian Liberal senator has decided to close the door on a distinguished parliamentary career and move to New York as Australia’s Ambassador to the United Nations.

Update 20 January 2006: More reshuffle rumours, also here

Now that was something I have not seen in a while. The next New South Wales election is due on 24 March 2007 and Centrebet is posting odds. It is paying $1.65 for a Labor win and $2.10 for a Coalition win. The implied win probabilities are 56 per cent for Labor and 44 per cent for the Coalition.

The last time I saw odds on the next NSW state election was in August 2005. Interestingly, I have not seen odds on the three state elections before NSW: South Australia, Tasmania and Victoria; all in 2006.

The first Newspoll of the year has Howard in front.

The headline prediction from last weekend’s poll was a national two-party preferred vote of 52 per cent for the Coalition and 48 per cent for Labor. This is the first time Newspoll has had the Coalition ahead since the first week in October 2005.The Coalition’s primary vote was 45 per cent and Labor’s was 37 per cent. Such a national vote would see the government returned comfortably.

It looks like Christmas with the folks and two weeks at the beach has erased national unease with Howard’s industrial relations reforms. I thought that Senator Andrew Murray’s analysis would prove correct: the industrial relations reforms would not impact on a significant number of people before the 2007 election. As a result, I thought the uncertainty that drove the polls in the last three months of 2005 would dissipate. If Morgan and ACNielsen replicate this poll result, it will have happened much faster than I had anticipated.

(Note: technically this result is not statistically different from the previous result. We could be witnessing nothing more than random noise. That is why it is important to wait and see whether the result is replicated before coming to a firm view about the voting intention of the nation after the Christmas break).

The usual graphs are here. You may need to hit the refresh or reload button on your browser to see the latest graphs. (I will update the remaining graphs when Newspoll publishes the full dataset).