Tuesday, March 8, 2016

Poll Roundup: How Far Will Turnbull Slide?

2PP Aggregate: 51.0 to Coalition (-0.3 in a week, -2 in three weeks)
Coalition would win election "held now" with much reduced majority

The downwards trend I've been tracking in the Coalition's aggregated federal polling has continued. It's not so much the case that the polls keep getting worse, but more that it was longer since they got a really good one.  In the last two weeks Essential (which has leant to Labor on the whole since Turnbull became PM) has put out a couple of 50:50s, and Newspoll (which in its new incarnation has also leant to Labor, though not by much) has followed up its shock 50:50 from a fortnight ago with more of the same.  Morgan had a very slight improvement for the government, up to 53:47 by last-election preferences, but this may be more evidence that its house effect is hanging around than anything in the government's favour.  Anyway with all these included my aggregate now falls to 51.0% to Coalition - more or less exactly where it was in the immediate aftermath of last September's removal of Tony Abbott.  Here's the smoothed tracking graph:

There's increasing talk of a July 2 double dissolution, which would mean the election is now less than four months away.  Normally, if you offered a new first-term government a 51:49 poll lead four months out from an election it would take that in a shot, since governments of any age in such a position have a very high historic chance of re-election.  Again though, it's the pace at which the Coalition has shed support that would be causing a lot of concern. Some of what has happened in polling this year is the natural coming down from any leadership transition bounce, but that far from guarantees that that process is even finished.

It wasn't only the 2PPs that were of interest this week.  Had the Greens' vote crashed and burned in the wake of their support for Senate reform there might have been more pressure to reconsider the concept, but the party's vote didn't change in Newspoll and reverted to its normal recent level of 10% in Essential.  The Greens lost two points in Morgan, but that was from a figure of 15% (which nobody sanely believed in the first place) and the likely cause of it was the addition of the Nick Xenophon Team to the menu in states outside South Australia.

NXT polled 5% nationwide in their first such polling measure, implying 3.5-4% outside SA.  If such a level held up at a double dissolution this would put them in play for the odd Senate seat outside their SA stronghold.


The key leadership figure this week is yet another netsat fall for Prime Minister Turnbull, who is now on a Newspoll net rating of just +3 (44-41).  Turnbull is still far more popular than Bill Shorten, but he's dropped 35 points in the four polls taken since last November.  This loss of 35 points in three and a half months is exceeded only by Paul Keating in 1993 (43 points in just over three months), John Howard in 1996 (36 points in six weeks and 43 points in twelve weeks) and Howard again in 2001 (38 points in six weeks).  The 1996 Howard example comes with a big asterisk too, because Howard was falling from the career-high +53 netsat he had jumped 24 points to reach in the immediate aftermath of the Port Arthur massacre.  It is not at all normal then for a PM to lose this much popularity this fast, but then again it is not that normal for them to have it in the first place.

PM netsats are closely connected to 2PP polling, so while Turnbull's current rating still isn't a problem, the Coalition would not want to be fighting a campaign with him below, say, -10.

Other Newspoll results were less interesting, with Turnbull retaining his 55-21 "better Prime Minister" lead over Bill Shorten and Shorten himself moving up slightly to a -25 netsat (30-55).  Essential also had a big fall for Turnbull (down 14 points in a month to +10 (45:35) while Shorten didn't move (still -20) and Turnbull's lead as "better PM" shrunk eight points but was still a healthy 29 points (48:19).

Other polling

Essential last week had a poll on Senate reform, which after a very long preamble came up with a 53:16 thumbs up from respondents.  There was a report that a union-commissioned Essential had "found" that Greens voters opposed the reform, but Essential advises me this was actually a ReachTEL.  The question wording from the former poll has not been published, which is generally a bad sign for the credibility of a commissioned poll question.

Essential respondents showed that they are less inclined to regard economic management as one of the most important issues.  (It's been overtaken at the top by health, though I am sceptical that health issues motivate votes to quite the extent that voters cite them as doing so.)  Essential also found that Coalition and Labor voters resemble each other, and Greens and Others voters resemble each other, on house prices.  Reactions to negative gearing have a usual partisan split, except that Others voters aren't sure they like it but don't want to change it either.  Essential also polled on gender equality with a predictably non-equal response.

That's just a quick instalment for this week as I am again very busy with work.

No comments:

Post a Comment