Tuesday, November 17, 2015

Poll Roundup: Terror Bounce Kicking Labor When They're Down?

2PP Aggregate: 53.6 to Coalition (+0.8 in a week, +7.2 since Abbott was PM)
Coalition would win election "held now" with increased majority

Another three federal polls are out this week and the Coalition's aggregated polling lead just keeps on growing.  The relentless rise of the Turnbull regime has to peak sometime, but for another week the answer to the question "when?" is "not just yet."  There's now no sign on the smoothed tracking graph that it is even slowing down:

This week so far we've had a fourth consecutive 55:45 from the Turnbull-friendly Morgan series (56-44 respondent allocated), 52:48 from Essential and 56:44 from Fairfax-Ipsos (57-43 respondent allocated).  The latter was remarkable as the highest result for the Turnbull government in any poll so far, although Ipsos has so far displayed just a slight lean to the Coalition (about 0.6 of a point on average).

Taking into account the primaries and the apparent current house effect of Morgan (which is showing some signs of waning in that Morgan stays at 55 while the others mostly increase) I aggregated the Morgan at 52.8, the Essential at 51.6 and the Ipsos at 55.7, which produces a new reading of 53.6% to Coalition.  The increase of 0.8 of a point (0.9 since the weekly reset) reflects that Ipsos had a three-point 2PP jump in four weeks, while the other two polls stayed the same.

In recent weeks my readings for the Coalition have been several tenths lower than other aggregators, so we'll probably see 54+ and maybe even 55+ from other sites this week.  (Edit: 54.4 from Mark the Ballot and Bludger Track, 54.9 from Phantom Trend). For instance I assume that the house effects of polls other than Morgan have stayed the same under Turnbull as they were under Abbott.  I also don't downweight Essential (which has been the most Labor-friendly pollster of the Turnbull age so far) as much as some aggregators do.  But even on my slightly cautious reading of the Turnbull surge, the Coalition ticks another two boxes this week.  Firstly, for the first time in its term after the small matter of two years and two months, it's ahead of its 2013 election result!  Secondly, the Coalition now has the same lead under Turnbull as Labor had when Abbott was removed.  (In reality though, since the Coalition only needs a shade under 49% 2PP for a 50:50 shot at winning, the lead now is much larger.)

The Ipsos result attracted both concern and scepticism from ALP supporters online.  The same had also been true of the 53:47 in Ipsos last month, which was rubbished in some quarters but sooner or later repeated by Essential, ReachTEL, Morgan phone and Newspoll.   This one looks more like a bouncy outlier, except that there is something obvious that could be driving it: the Paris terror attacks (which were breaking news with wall-to-wall coverage on the third day of the Ipsos sample).  That said, half of Morgan's sample was post-Paris, and Morgan showed no sign of it.  Ipsos is the bounciest of the currently active pollsters, but not by a massive amount: its monthly polling cycle often makes it look bouncier than it is.

We'll need polling from other sources to resolve whether the Paris attacks have impacted on Australian polling, but it wouldn't be at all odd if they did.  Although the direct impact on Australians is small (with one Australian injured and apparently none killed) there's a sense that this attack is a big deal for all Western nations since it shows that IS, like al Qaeda before it, has an ability to create large attacks inside the West.  For some comparisons with bounces from past terrorism incidents see this previous article; the Abbott government ended up gaining 2.3 points over six weeks following the MH-17 shooting-down, although it's unlikely it was all down to that incident.  In this case, because the government's ratings were rising anyway, untangling causes and effects will be even harder.

With nothing that Labor tries to do against Turnbull having even the slightest impact so far, issues that put national security back in the spotlight are the last thing Labor needs.  This is not just because it is a Coalition strength area but also because it gives Turnbull further opportunities to refresh perceptions of political language by avoiding Tony Abbott's ever-predictable macho rhetoric and incessant "death cult" references.

Leaderships and party perceptions

Cue car crash jokes - there was more woe for the ALP in the Ipsos leadership matchups.  Malcolm Turnbull has a massive net approval of +53 (69-16) while Bill Shorten languishes on -28 (29-57).  The 81 point gap isn't an all-time record as incorrectly reported by the Age (Hawke-Peacock in Oct 1984 had a 102 point gap) but it's still the largest in the Nielsen/Ipsos series for a very long time.  Turnbull's net approval exceeds Rudd's best in the series (+52, noting that Nielsen didn't poll in Rudd's first five months) and if there has been a PM with a lower disapproval rating, it would probably have been Hawke in his first term.  (Hawke's highest approval was 75% in November 1984).  Shorten's rating is again his worst in this series to date.

Predictably, Turnbull flogged Shorten 69:18 as preferred Prime Minister (the biggest lead since Rudd over Brendan Nelson in May 2008, 70:17 - Newspoll had had a 73:7 lead earlier that year, but Nielsen were not polling at the time.)

I typically dismiss preferred-leader scores for their house bias to the incumbent, their status as a composite and lagging indicator, and their poor relation to election outcomes.  However, I've just had cause to look at the predictiveness of bad preferred leader scores for Opposition Leaders, in the context of a bad rating for Tasmanian Opposition Leader Bryan Green.  I find that at state and federal levels combined, Opposition Leaders who poll below 20 on a preferred leader score after more than 18 months in the position are almost always toast.  Of eighteen such cases in the past 30 years, two-thirds were removed by their parties before the next election.  Of the six who made it to the election, only one succeeded, and even then only because a seat result from the original election was annulled.
The present situation is unique, but Shorten is in some very poorly performed company.

Essential surveyed party attributes and found little change in perceptions of the ALP since last year.  For the Coalition however they found voters viewed the party as more moderate, having more vision and a better team of leaders but also (unsurprisingly) more divided than in the previous survey.  In comparing the two parties, Essential found that the ALP generally compared pretty well on the questions asked.

Other Polling

Essential also surveyed issue importance and found economic management down (but still top of the list) and national security unsurprisingly up.  Also up were a fair taxation system and housing affordability while "leadership" was down now that voters have a leader they are happy with.

Offshore detention of asylum seekers had strong support (54-31, with only Greens voters against) even although the question design made no mention of processing.  Essential found 25% of respondents said the current government was too tough on the issue, 29% too soft and 31% taking the right approach.  Coalition (32-11) and Others (42-22) voters tended to prefer too soft, Greens of course preferred too harsh (73-13) while Labor voters (35-28) seemed to be engaging in random tribal disagreement (I suspect the left wing being more sincere than the right).

Ipsos asked about the GST and these results have been reported plenty elsewhere.  The results are similar to those for NSW Premier Mike Baird's electricity leasing policies, in which the raw question gets a strong disapprove but when the full setting of the proposal is explained, responses become more balanced.  In this case, voters (51-42) support a GST increase if it is accompanied by tax cuts and compensation for households with incomes below $100,000 per year.

It's four weeks since the last ReachTEL so there might be one of those soon; either way I'll try to squeeze in another episode around work commitments after Newspoll next week.


  1. Dear Kevin,
    Has there ever been an opposition leader on a Newspoll of 14% who did not resign ?


    1. Yes. Brendan Nelson recorded many Better PM scores of 14% or below (down as low as 7%) over a period of several months before being removed. The other two Opposition Leaders to get down to 14 exactly were Simon Crean and Malcolm Turnbull. Crean resigned and Turnbull was removed before either could face another Newspoll.