Tuesday, November 18, 2014

Poll Roundup: G20 No Help To Government

2PP Aggregate: 53.1 to ALP (+0.8 since last week, +2.1 in five weeks, highest in four months)
ALP would easily win election "held right now"

A fairly quick update as I have a lot of other work to do, which may sometime over the weekend or so include putting this piece of pseudo-scientific garbage in its place.  This week the federal government hosted the G20 meeting in Brisbane.  This would have been expected to at least allow the government to dominate the media cycle and demonstrate its international credibility.  However, things haven't been that simple.  Climate change was seen to dominate the agenda (probably to a greater extent than it actually did), and the isolation of Vladimir Putin by various Anglosphere leaders was still an anticlimax compared to the threatened "shirtfronting".  Even when Australia signed a generally well-received (if probably not well understood) free trade agreement with China, the Prime Minister still ran into Parrot problems.

But doubtless the worst moment of the conference for the Coalition was the PM's widely slammed address in which he was seen as not only rabbiting on to world leaders about irrelevant domestic issues, but also doing so in a way that whinged about the known unpopularity of government measures such as GP co-payments.  It was a free hit for the Opposition and it was duly and effectively taken.

None of this is to say that the issues surrounding the G20 are necessarily the cause of what we have seen in polling this week, or even that we have conclusive evidence that things are worse for the government than they were two weeks ago. Also, it might be that aspects of Australia's stance that were criticised in the last few weeks look better down the track.  At least, however, if the Government was hoping for even a modest quick bounce in its favour from hosting this major event, it hasn't arrived.

This week's polls

This week saw no change from Essential Research (52:48 for the third week in a row.) Morgan's last-election result was 54:46 to Labor, an increase of just half a point, and equivalent to a 52.5 from anyone else (the respondent-allocated figure was 55.5).  The big result was Newspoll, which put out a 55:45.  Although this was only a one-point shift in the released 2PP from the previous fortnight, in this case the primaries (39-36 to Labor) implied that it was really 55, not a 54 that looked like a 53.  Indeed, neither Newspoll nor Morgan this week have showed the primary-vs-2PP discrepancy I talked about in the previous episode.

The Newspoll was also a surprise as it was a shift up for Labor from an already above-trend figure.  It's possible that this is just a randomly friendly run for the ALP from Newspoll or even that Newspoll is on a bit of an off-track excursion, (See my recent concerns about Newspoll's high Green/Others vote figures.)  For the time being, my aggregate isn't fully buying the strength for Labor of the recent Newspolls, but has still gone to 53.1 for Labor, the highest reading for four months.

The smoothed tracking graph now shows 2.1 points of blowout in Labor's lead over the last five weeks.  This is looking more and more like a government that is intrinsically unpopular rather than only unpopular in response to specific events.

The Newspoll was the third 55:45 to Labor for this term, the previous two coming in May and June during the budget blowout.  The ALP's primary lead of three points was greater than in any of the blowout polling, and was the party's best since a single poll in July 2010, when Julia Gillard had just been installed as Prime Minister.  Prior to that, Labor had held similar leads under Kevin Rudd until March 2010.

While the voting intention figures are similar to the blowout phase, Prime Minister Tony Abbott's net satisfaction rating is "only" -19 (36-55) and not so bad as the  ratings he recorded in May through to early July (which bottomed out at -31).  Bill Shorten polled his best netsat for nearly three months (-2 (39-41)) and Shorten's six-point better-PM lead (43-37) was his highest since late June.

Last week's Essential figures also saw Abbott's personal ratings worsening slightly (down three points to net -11) while Shorten was steady at -1. Essential had Abbott "leading" as Better PM, but only by two points, which is not enough for an incumbent.

G20 and China FTA Polling

The most interesting attitudinal polling this week was Essential's series on the G20 and the China free trade deal.  The FTA with China was asserted by shockjock Alan Jones to be failing "the pub test", but it passes the Essential panel test with 44% approving and 18% against.  Support was strongest from Coalition supporters (58-9) and reasonably good from ALP supporters (44-21) with Greens (31-27) and Others including PUP (34-35) equivocal.

The deal was seen as skewed in China's favour overall (35-12) with the Australian government (52%) nominated as the party most likely to benefit in Australia (hmmm) and Australian workers least (25%).  Predictably voters want to have their cake and eat it too, supporting greater access to Chinese markets (61-12) but opposing reduced restrictions on Chinese investment (52-23).  It seems that voters ideally want an FTA with China in which we trade freely, but they don't, and that despite their lukewarm support for the FTA signed, most voters are missing the point of what an FTA actually is.  Voters were generally sceptical (62-16) that the G20 was anything but an "expensive talkfest".

Essential also had some useful stuff on refugee intake, showing that major party supporters are most likely to overestimate Australia's current intake while Greens supporters are more likely to underestimate it.

Any new polling over the coming weekend will be added in updates to this article.  I expect this series to be back in a fortnight and I already have a title: Poll Roundup: Deeply Unpopular By December.

Friday ReachTEL Update
A new federal ReachTEL has been released with a 53:47 to Labor 2PP; I model it as 52.7% to Labor off the decimal primaries; either way, it makes no difference to my current aggregate, and also no difference when I apply the weekly reset.  The leadership ratings see Abbott record his worst "Very Poor" rating (36.1) since June, although the 2.9 point increase there comes partly off "Poor", which is down by 1.8.  All up I convert Abbott's ratings to about -15 on the Newspoll scale.

Bill Shorten's combined "poor" rating (37.7) is at its lowest level since February (when he was still quite new to the job), though the gain was entirely made by Satisfactory, with Shorten's "meh!" factor as high as it has been since March.  All up I convert Shorten's ratings to about -3 on the Newspoll scale.

ReachTEL's G20 ratings reveal that the number of voters rating Tony Abbott's performance at the G20 poor or very poor (41.7%) is massively higher than the proportion who believe the summit did not produce positive outcomes for Australia (22.2%).  Also of interest is a question on Jacqui Lambie leaving the Palmer United Party - PUP supporters think she shouldn't, but Labor, Liberal and Green (especially Liberal) supporters tend to believe that she should. 


  1. Deeply Unpopular By December

    Nice title

    1. Ta, it comes from a quote by a nameless Coalition source in AFR in mid-April just before the Budget.

  2. I wonder what a question about the ISDS provision in the China-Australia FTA would yield (now that it's started to hit the news - which it hadn't really when this polling was done) and whether asking this question (with explanation of what ISDS entails) before asking whether the respondent approves of the FTA as a whole would skew the result of the latter question.

  3. Yes the interesting thing here is that 62% felt qualified to give an opinion on the FTA when most likely 0.01% (if that) knew what was actually in it.

    If respondents are asked a question about a specific feature before being asked a question about the agreement as a whole that would nearly always skew the result (positively or negatively depending on the feature included). In the case of something like ISDS it would also come down to how the explanation was framed - it would be possible to explain it in a way that caused either a positive response or a negative response, though I suspect the latter is easier. The real test is whether asking the same overall question about the FTA a little while after news about a feature of an agreement breaks produces a cooler response than before that news was known. Perhaps there will be FTA polling from other pollsters.

    (For those wondering, ISDS = investor-state dispute settlement, which gives corporations rights to sue countries that have signed FTA-type agreements but that then restrict those corporations' abilities to trade in certain ways.)