Poll Roundup: The Turnbull Slide Continues

2PP Aggregate: 51.8 to ALP (-0.1 since last week)
Labor would probably win election "held now" with a small majority

Almost four months since its re-election, the Turnbull coalition government continues to poll ordinarily.  It has still not led in a single poll by anyone, and only now are there some signs that Labor's lead might be stabilising.  With a new bad news cycle hitting the government, oh, about twice a day on average lately, there might even be relief it is not worse ... yet.  An implicit stoush between PM Turnbull and Tony Abbott over the history of rapid-fire shotgun negotiations might have looked like fodder for further movement, but Newspoll suggests it was seen as just beltway stuff.

Most of our polling still comes just from Essential and Newspoll though this week a third player, Morgan, dipped a very tentative toe in the water with one of its irregular phone polls.  The only voting intention result released was 55:45 to ALP, easily Labor's biggest lead from anyone so far.  Given the small sample size and the lack of adequate documentation of 2PP method or primary votes, this only counted for 10% of a normal poll in my aggregate, but even that made some difference.



The last two Newspolls have been both 52:48 to Labor.  The last few Essentials were 52, 53 and 52, but the last 52 looks like it was very close to being a 53.  I aggregated the Newspolls at 51.9 and 52, and the Essentials currently in the frame (which is this weeks and the week before lasts) slotted in at 52.2 and 52.4.  (Last week's was 53.4, and it will count towards next week's reading - see the methods page for how this works.)  I should note that The Australian helpfully published finer breakdowns of the Newspoll votes for One Nation and NXT, so I have back-calculated those Newspolls for which that data is available, though in most cases that made no different.

While I have the Coalition up from 48.1 to 48.2 on the single-week reading (assuming that there aren't any more polls this week), the smoothed aggregate graph is a one-way street.  It's extremely dull viewing except for ardent ALP supporters, and it looks like this:

Still, early days ...

Turnbull Is Less Popular Than Trump?

One Ben Cantwell on Twitter has just picked up the amusing statistic that Malcolm Turnbull's Newspoll netsat of -28 (29-57) is worse than Donald Trump's Real Clear Politics net favourability aggregate (currently at -25.2).  This isn't really comparing oranges with oranges, firstly since favourability is a lower bar than satisfaction, and secondly since Newspoll is among the harsher measures of Australian leader popularity. (In BludgerTrack, Turnbull's aggregated netsat across all polls is only -16.4, though it is bound to take a hit when this latest Newspoll is added.) You can find polls where Trump is doing much worse than -28, but still, the comparison is a surprising one.   Very few recent polls have not had Trump's favourability higher than Turnbull's current Newspoll satisfaction rating of a feeble 29%.

That 29% is more notable because it is lower than the 30% satisfaction rating Tony Abbott had when he was dumped (although Abbott's net rating, -33, was worse than Turnbull's present -28).  Indeed Abbott only polled three sub-29% satisfaction ratings in his whole career as PM.  These came in a row in early 2015 following the Prince Philip knighthood, the Queensland election defeat and the spill motion against Abbott's leadership.  For comparison, 27% were satisfied with Bob Hawke just before he was finally removed by his own party, 36% with Kevin Rudd, and 28% with Julia Gillard.

Bizarrely, we are starting to see some hemi-demi-semi-serious discussion about bringing back Tony Abbott!  Let's not forget that it was only by listening to the electorate for once and throwing Abbott under the bus that the Coalition got itself re-elected at all.  An Abbott revival would be history repeating as farce and the ultimate debasement of the idea that party leadership at election time means anything at all.  Abbott is not Kevin Rudd the second; voters never understood why Rudd was removed in the first place and so the hankering for his return was natural as a matter of fairness if nothing else.

The Silver Lining

There is something interesting developing in the relationship between Turnbull's netsat and the Coalition's 2PP.  Falls in Turnbull's netsat hurt the Coalition much less than the same falls for any previous Prime Minister of the Newspoll era.  On average, Turnbull has to lose 15.7 netsat points for the Coalition to drop one 2PP point.  This compares with 7.0 netsat points for Abbott, 6.2 for Gillard, 8.6 for Rudd (first time), 6.4 for Howard, 5.8 for Keating and 6.0 for Hawke.  The average 2PP reading for a party whose leader is rating this badly is a mere 46.3%.

There are a few possibilities, and we'll probably need to wait til ReachTEL comes back to find out which one is true.  One is that a lot of people are just mildly dissatisfied with Turnbull, with very few strongly disliking him.  A second is that a lot of people really dislike Turnbull, but do so from the Right and therefore aren't open to voting Labor.  A third is that a lot of people are really fed up with the PM, and would in theory be open to voting Labor, except they are even more fed up with Labor.  I think it's more the first two than the third, as the statistic I call the "disconnect index" (the minimum proportion of voters who must be preferring a party while being dissatisfied with its leader) is at modest levels by the standards of the last few years, though it is starting to creep up.

Meanwhile, Bill Shorten is on a net -15 (36-51), his ratings having barely moved since June, and Turnbull is rated "better PM" 42-32, his closest lead over Shorten so far.  Last fortnight's better PM reading (45-30 to Turnbull) was also interesting, because only once had a PM with such a lousy netsat held a better PM lead so large.  Paul Keating (then -27) had a 24-point better PM lead over Alexander Downer in December 1994.

Essential actually had Turnbull rebounding to a mere -3 net approval this month (38-41, showing how different Essential and Newspoll scales are, or how different the Essential voter panel is) with Shorten the same (37-40), and Turnbull ahead 41-28 as better PM.

Other Polling

As usual, there are dozens of issue questions over at Essential, and those interested can read them all over there.  Here are some that I found interesting:

* Labor, Coalition and Others voters rate "ISIS and Islamic radicalisation" the biggest threat to global peace and security in a field also including global warming, Russian aggression, inequality, overpopulation and Trump.  For Greens voters, ISIS and the Islamists come in third behind Trump and other forms of global hot air.

* While many voters have no view, those who do have a view support reestablishing the ABCC, 36-16.

* The Turnbull Government has a net -27 confidence rating for its ability to get stuff done "given the Senate".  The question design does seem prone to invite agreement compared to, say, one that also mentioned that governments can get some things done by gazetting them without Senate approval.

* Australia is quaking in its boots about the small risk of Donald Trump becoming President.  This is not unusual; previous polls (including pre-Trump) have found that Australian Liberal voters are about as pro-Democrat as Australian ALP voters.

* 61% of the panel thinks multiculturalism has been a net positive for Australia; 23% disagree. Predictably Others are most likely to whine.  24% think a migrant family should be rejectable purely based on their religion, and 56% disagree.  (The question "which religion" might have been an interesting followup to see if the concern was purely about Islam).

* Essential again found that voters overestimate how many Muslims there are in Australia, though awareness of the real number has improved.  Information about the actual number makes little difference to concern levels, which drop from 53-42 to 47-47 post priming.

There has been a general run of rather xenophobic results from the Essential panel, and the Morgan phone poll provides quite a different view.  However, the difference is probably caused directly by the polling method.  People are much less likely to express xenophobic sentiments in a live phone poll because they do not want the interviewer to think they're a racist.

Those Fickle Millennials

It was only about six weeks ago that I realised "millennials" are more or less the same people that not so long ago were called "Generation Y".  This probably shows about how much attention I pay to generational stereotyping and how little regard I hold it in.  Anyway, the relevance of said cohort to this website can be found in an AFR article which referred to the difficulty of connecting with millenials politically.   The claim is that those born between the early 80s and the very late 90s have no brand loyalty and churn through products, and that the same thing applies to how they interact with politics.

I have a different view, one that is closer to that of Labor's George Wright than of anyone else quoted in the article.  You don't get the votes of younger voters if you don't appeal to them, and all generations are like that. The millennial cohort - together with various others - discards political products so quickly not because they are compulsive churners, but because the political products are crap.