Tuesday, October 27, 2015

Poll Roundup: Newspoll Smells The Coffee

2PP Aggregate: 52.4 to Coalition (-0.1 since last week, +6 since Abbott was PM)
Coalition would easily win election "held now"

There are only two polls to deal with this week so far (for last week's including Friday's ReachTEL see Investments Attack On PM Fails) but one of them is a rare and special case. This week's Newspoll has set not one but two new all-time records!  One of these records is bad for the Shorten-led Labor Opposition (but not as bad as the many comments on Shorten's dire ratings are making out), while the other can be spun almost any way you like.

Over the past few weeks there has been debate about whether the Coalition has a meaningful 2PP lead or whether its lead is barely over 50:50.  With results from Newspoll, Essential, and the very early Galaxy and ReachTEL results falling in the latter category, the sceptics still held out after Ipsos joined Morgan in the former.  Then ReachTEL late last week joined Ipsos on 53:47.  This week Newspoll, which in its new incarnation has been somewhat Labor-friendly, came out at 52:48 to Coalition, meaning that for now at least this debate is over and the Coalition has a solid lead.



I aggregated both the Newspoll and the Essential at 51.8 after considering the primaries.  The week to week readings, as backdated after a couple of methods adjustments noted in recent articles, are bouncy:


However, the smoothed graph looks like this:


What initially looked like a 5-point shift to the Coalition at the time of the leadership change has been revised down to 4.4 points as it has become clearer that odd behaviour by the Morgan poll was exaggerating it.  However, since then, another 1.6 points have so far been added.  There is no way yet of knowing at what point this will flatten out, but at the moment the Coalition would win an election easily with hardly any seat losses.

It must be stressed again that this is bounce polling.  In terms of predicting the next election it isn't all that meaningful.  However, it is better to be polling such numbers than to not be polling them.  Had the Coalition only poked around 50 for a few weeks then started to fall back (as Labor did when Rudd was returned) then there would now be a damaging narrative that the change at the top was a shuffling of deckchairs and that the government was on its way to defeat.

Essential had a poll question regarding whether Turnbull and Morrison would be better economic managers than Abbott and Hockey.  50% said better, 10% worse, and 25% much the same.  Here Coalition voters have overwhelmingly accepted the transition as good for the economy (67:9) further suggesting that Abbott isn't being missed.  Only the Others category (34:16) - which includes those who have defected from the Coalition in annoyance at the removal of Abbott - seems even close to having reservations about all this.

Essential also had a question finding 45% to 26% believed Australians would be better off if unions in Australia were stronger.  (Coalition supporters dissented, 28:41).  This is somewhat surprising given that last week Essential found trade unions to be among the least trusted institutions, and also given that support for the Trade Unions Royal Commission is up slightly.  It seems that voters view unions both as necessary and as prone to corruption.

Essential asked a question about support for a national vote on same-sex marriage that was primed with cost information.  The impact was rather dramatic, with support for a plebiscite (or similar) slipping from 67:21 to 43:41 when the cost was mentioned.  So much for the view that the cost of a plebiscite is not seen as a serious issue.  This is also a good example of how the use of leading information (a common tactic in commissioned polls) impacts on poll results.

Newspoll Record 1: Netsat Gap

This week Malcolm Turnbull has a netsat of +35 (58% satisfied, 23% dissatisfied).  This is the highest netsat for an incumbent PM since Kevin Rudd polled the +35 (63:28) in October 2009; two weeks before Rudd had polled +43.  It's also the first PM satisfaction rating over 50 in five and a half years.  To find a dissatisfaction rating so low we have to go back to Rudd again in April 2009.

Bill Shorten, meanwhile, polled a netsat of -32 (26:58).  This equals his previous worst, but his satisfaction rating of 26 is not only a new personal low, but also the lowest for any opposition leader since Turnbull in the wake of the Ozcar forged email blunder (June 2009).  Even Tony Abbott never got quite that low as Opposition Leader.  As @sorceror43 commented on Twitter, Abbott always had a loyal right-wing base.  Shorten, however, is not liked by the Left, the conservative Right or swinging voters, so his natural "base" is small: rusted-on ALP supporters who aren't especially left-wing.  There is potential for his satisfaction rating to fall further.

The difference between the two, of 67 points, set a new all-time Newspoll record, beating the previous record of 66 points (Rudd +57 vs Nelson -9, Feb-Mar 2008, in a probably rogue post-election poll with a Labor 2PP of 63%). This record does come with an asterisk because of Newspoll's recent change of methods, which seems to have slightly reduced the natural "undecided" level in Newspoll leader ratings.

What does this record mean?  Well really, not a lot.  Turnbull is in "honeymoon phase" and it remains to be seen whether he will be a popular leader in the long term.  Unpopular Opposition Leaders had a very long history of always either being removed or losing, until Abbott.  However, because Opposition Leader ratings have such a weak relationship with the 2PP standing of their party, this history tells us little about how electable Shorten is.

The poll also saw another thumping for Shorten as better Prime Minister, this time 63:17.

Newspoll Record 2: 2PP/PM Netsat Disparity

This week the Coalition has a 2PP of 52% but Turnbull's netsat is +35.  This is a serious discrepancy.  Normally if a government has a 2PP of 52%, the PM's netsat will be about +14.  Or, if the PM has a netsat of +35, the 2PP will be about 55%.

Indeed (as seen on the graph below) this is the highest ever netsat recorded for a 2PP of 52% or lower (the previous highest was +30).  Or, another way of putting it, this is the lowest ever 2PP recorded for a netsat of +35 or higher (the previous lowest was 54).  It's quite an outlier:


This week's result is the red dot on the right-hand edge of the blue ones.  When Abbott was PM, we now and then saw results near the opposite side of this chart (terrible PM netsats but only slightly behind on 2PP).

What does it mean?

There are at least three different plausible takes on this result.  The first is that it is about Abbott.  Turnbull as a popular PM would normally be producing landslide party ratings, but voters who like Turnbull (at least for now) still distrust the government because it was so awful under Abbott.  If this is the case, then should Turnbull continue polling high personal ratings, these will translate into larger leads down the track (although these leads will probably contract as the election approaches).

The second is that this is about Turnbull.  The pessimists on the Coalition side (and they are not just the Andrew Bolt types on the hard right) are worried that Turnbull polls high ratings because he is approved of by lefties who would not actually vote Liberal.  If this is the case then massive approval ratings for Turnbull will probably never convert into large 2PP leads.  Worse, if the gloss comes off Turnbull personally, the government might be less resilient than under a leader whose appeal was less cross-party.

The third is that this is about the circumstance. Since prime ministerial netsats appear to drive, and move slightly ahead of, 2PP voting intention, it may be that the Turnbull bounce has further to go.

These three takes each create their own prediction for what will happen to the Coalition's vote share if Turnbull's netsat now stays about where it is.  In order, they predict gradual increase over some time, no change, and a quick increase.

North Sydney

A by-election for North Sydney has been set for December 5.  At this stage I don't see any reason to doubt that the Liberals will retain the seat, whether the ALP contests or not.  I don't see enough for me to get stuck into to justify a separate thread for the seat at this stage, so I refer readers for now to Poll Bludger and Tally Room.

No comments:

Post a Comment