Wednesday, December 9, 2015

Poll Roundup: Turnbull Bounce Finally Peaks

2PP Aggregate: 53.4 to Coalition (-0.5 in a week)
Coalition would win election "held now" with similar majority to 2013

Almost three months since Malcolm Turnbull took over from Tony Abbott as Prime Minister, the massive surge in Coalition polling has finally hit its first speedbump.  The odd behaviour of a couple of pollsters compared to the rest has made it hard to say exactly when this happened, but my figures now have the Coalition peaking at a revised 54.0% two-party preferred at the end of the week before last.  On that basis I estimate that the Coalition gained 7.6 points over ten weeks.  I have this as the second largest polling surge in such a period in Australian polling history (the largest being at least ten points in a couple of months for the doomed Whitlam government during the lead-up to the 1975 dismissal).

Following the initial jump when Tony Abbott was removed, some of the pollsters have followed very different scripts.  Newspoll, Ipsos and ReachTEL have all returned results consistent with an increasing lead over time but Morgan showed a much greater surge initially then no real change thereafter.  Essential didn't move much after the initial surge and has mostly just wobbled around the same value for several weeks.  Overall, Morgan has averaged 2.2 points above my aggregate while Essential has averaged 1.1 points below, but in Morgan's case the difference is trending down while in Essential's it has increased greatly in the last four weeks.  This has in fact been Essential's biggest off-track excursion of this term.  A few weeks back I was adjusting Morgan by two points and Essential not at all, but for now I am adjusting both by 1.5 points.

Last week's Morgan was 55:45 by last election preferences (for the fifth fortnight in a row) and this week's Newspoll was 53:47; Essential had 51:49 both times.  Considering the primaries and the recent behaviour of Morgan and Essential I counted the Morgan as 53.3, the Newspoll as 52.6, last week's Essential at 52.5 and this week's Essential at 52.3.   None of those are consistent with recent form and the result is a small dip in the Coalition's lead:

Next week may well be the final active polling week before the summer break and so we won't get a reliable idea of where the story goes from here for quite a while.  (Some other aggregates: Phantom Trend has 52.6, Mark the Ballot has 54.5, Bludger Track 53.8.  Mine will go down to 53.2 on weekly reset this Saturday if there are no other polls before then.)

It's no surprise the tide has finally stopped rushing the Coalition's way.  The new regime has been challenged first by continuing questions about Special Minister of State Mal Brough's role in the Ashby/Slipper saga.  The most pertinent question is why Turnbull appointed Brough knowing him to have this liability, and one possible answer is that Turnbull reckoned he needed every last vote he could get.  There are others, and they are much less flattering.

There's also been some silly stuff with dumped minister Ian Macfarlane's attempt to defect to the National side of the LNP, apparently in order to oblige the PM to give the Nationals an extra minister (namely him).  Although this cunning plan may well be stymied ("you can check out any time you like, but you can never leave"), it serves as a preview of the kind of post-spill dummy-spits that are likely to distract the Coalition now and then all the way to the ballot-box. It may raise more serious questions about the relationship between the Coalition partners.

While the Coalition's polling seems to have peaked for now it is still too early to say much about what it predicts about the next election.  By about March the honeymoon aspect should have well and truly gone away, and if the government is still in front then, then history gives Labor very little chance of winning.


The gloss finally came off PM Turnbull's Newspoll ratings this week too, with a sixteen-point netsat plunge from +38 (60-22) to +22 (52-30).  +22 is still, of course, better than any other PM managed for the last six years.  The sixteen-point hit isn't the highest for an incumbent PM this year (Tony Abbott took a 19-point hit back in February) but it is still the equal twelfth highest of all time.

Meanwhile Bill Shorten has bigger problems, recording a career-low netsat of -38 (23-61).  This is even worse than Abbott's low as Opposition Leader (-36, 27-63) and is the worst netsat since Simon Crean polled -39 in May 2003, and the worst satisfaction rating since Crean polled 22 in November of that year (just before being ousted.)  Further strengthening the Crean parallels noted last time, Shorten's better PM rating matched Crean's all-time low of 14% (Turnbull leads 60-14).

Essential, which only polls leaderships monthly, had a less dramatic picture with Turnbull down three points from +36 to +33, and Shorten stable on -20.  Turnbull has a +6 rating among Labor voters while Shorten rates deeply negatively among all non-Labor parties.  Turnbull leads Shorten 54-15 as better PM, compared to 55-14 last month, and enjoys confidence in representing Australia overseas by an impressive margin of 64-23.

Essential had preferred leader polls for each party with Malcolm Turnbull making a modest gain since September while Bill Shorten slips to a miserable 13% as preferred Labor leader, just behind Anthony Albanese and Tanya Plibersek (though Plibersek's standing comes heavily from minor party voters).  Which brings me to ...


One of the current sidebar Not-A-Polls asks readers who they think should lead Labor to the next election.  This is a live question in my view, whether Labor will do anything about it or not.  Early in the poll Jason Clare (who is generally not included in pollster readouts) was the leader, with Shorten taking the lead briefly on Sunday, and since then Albanese has been in front with Clare second.  The dominant theme, however, is that about four-fifths of voting readers do not think Labor should keep Shorten til the next election.  That is even though I have phrased the question so as to give as much scope for a response based on strategy instead of personal preference as possible.  I haven't put an end date on this not-a-poll yet but it should run well into next year.

The other Not-A-Poll, on pollster reliability, ends in a few weeks.

Other polling

Essential repeated their September poll on the fate of Tony Abbott, and the December results seem to suggest increasing belief that the former PM should get out of parliament, either now or at the next election.  Support for him leaving Parliament is at 49:32 compared with 41:41 in September, while Coalition voters have slipped in their desire to retain him from 60:27 for staying to 48:37.  However, much of this shift may be caused by changes in the question design that have rendered the comparison moot.  The option of saying that he should "resign" (actually, that should be "retire") at the next election wasn't in the September readout.  Some voters may well have not wanted Abbott to resign immediately and cause a by-election.

North Sydney

I have a separate article on the North Sydney by-election won by the Liberals' Trent Zimmerman, with a two-candidate-preferred vote headed for about 60:40 against the independent Stephen Ruff, and a primary vote that has just passed 48%.  Commentary on the result has tended to confirm the widespread view that by-elections are a sort of Rorschach test, and also the less widespread view that many Twitter lefties are gullible people who are happy to do the work of Tony Abbott in making the result look like some kind of massive flop.  The loss of a personal vote for a long-established sitting member, the usual by-election tendency to kick the incumbents, the increased number of parties and the presence of a vaguely competitive indie - all these things explain a lot but probably not quite all of the 13% primary vote swing.  Throw in some blowback over local council mergers and it's not clear if there is anything left to account for.  Those driven to park their primary vote with micro-parties for whatever reason have preferenced Zimmerman enough to avoid a close 2CP contest.  It is not as if hordes of voters started putting the Liberals last.

On the other hand, there's no evidence that the popularity of PM Turnbull contributed to an above-average result.  The 2CP "swing" is headed for about 6%, which is typical for government vacancies where the government is polling at a similar level to its previous election result.  The comparison is dubious because it is Liberal vs Ind rather than vs ALP, and indies can be more dangerous in North Sydney, but Ruff wasn't especially dangerous as such indies go.

So, technically a harmless and predictable result, but one that can be spun as worse than it actually is.

If my suspicion that next week is the last polling week of the year appears correct, then I will have an annual polling review up next week.

No comments:

Post a Comment