Tuesday, July 17, 2018

Poll Roundup: National Narrowing As Super Saturday Approaches

2PP Aggregate (2016 Preferences): 51.7% to ALP (-0.5 since last week, -1 in six weeks)
With One Nation preference adjustment 51.1% to ALP
Closest position since October 2016
Labor would probably still win election "held now" but it would be close

Even psephologists have trouble with counting to big numbers sometimes. Like two.  Normally I release a new post in this series every second Newspoll week, but two weeks ago on Newspoll Monday I had an inconvenient distraction.  By the time I'd got through that and a couple of days of work my mind was so much elsewhere that I had forgotten it was time for another Poll Roundup.

Anyway, another Newspoll week has come and gone and all the current streaks noted on my Newspoll records page have continued.  These streaks are: the Coalition for most 2PP losses in a row (now at 36), Malcolm Turnbull for most Better PM wins in a row (now 56), Bill Shorten for most negative netsats in a row for an Opposition Leader (now 69) and Malcolm Turnbull for the third-most negative netsats in a row for a PM (now 47, and he has overtaken Julia Gillard for second place for longest stay in negativeland by time.)

What we are seeing lately though is a definite narrowing in voting intention polling compared to most of this government's term.  The government's 2PP score by 2016 preferences was in the 46s and sometimes 47s for all of last year and through to April this year.  Since then it has consistently been at least in the 47s and is now having a nibble at the 48s.  Assuming that 2016 preferences may in fact favour the Opposition, currently two-party voting intention is very close indeed.  Here's the smoothed tracking graph by 2016 preferences:

In the time since the previous episode Newspoll has released a 48 for the government followed by two 49s, Essential has gone 46-48-48-49 and there was a ReachTEL at 48 and an Ipsos at 47.  My aggregated estimates using 2016 preferences for all these were 47, 48.1 and 47.9 from Newspoll, 46.1, 48, 48 and 48.9 from Essential, 47.2 from ReachTEL and 46.5 from Ipsos.  After putting all those in I have the government at 48.3 by 2016-election preferences, or 48.9 if applying a modest adjustment for One Nation.  

The differences between Newspoll 2PPs and those obtained from their primaries using 2016-election preferences continue as a result of Newspoll's methods change; I now have the average difference at 0.88 points.  ReachTEL and Ipsos use respondent preferences, but lately Ipsos has shown a much weaker advantage to the Coalition on these compared with other polling in this term.   (That is the main reason why BludgerTrack has lately looked more like my 2016-preference aggregate and less like the One Nation adjusted version.)

Super Saturday

What does it all mean for the by-elections?  In my early piece on Super Saturday (which was rather more optimistic for Labor than the run of seat polling since has suggested, for what little either might be worth) I noted that there is a relationship between by-election swings and how a government is polling nationally at the time.  What is notable about this relationship is that though the percentage of variation explained is only modest, the slope of the graph is reasonably steep.  For every point of gain in national polling, a government adds 0.67 of a point to its by-election results in Opposition-held seats.  The change in national polling since I wrote that piece knocks half a point off Labor's expected performance in Longman and Braddon.

If they stick to their schedules this is the last we've heard from Newspoll or Essential prior to the five by-elections on July 28. We're overdue for a national ReachTEL but we might not get too much more evidence about any further narrowing before the votes are counted.

Longman has looked close through the campaign, because of the likelihood of One Nation driving preferences to the Coalition, with the available seat polls supporting this.  The revelations about Trevor Ruthenberg's misnamed military medal won't necessarily sink his ship, but at best they put him in the same boat of sloppiness as Susan Lamb.

As for Braddon, the vibe seems to be that Brett Whiteley is winning, but hard evidence for that has been hard to track down (here's my Braddon preview which includes a great many words about prospects that basically boils down to "dunno").  I think Braddon is very unpredictable; either side might plausibly win 55-45.

There's a widespread view that a win in one or both by-elections could be followed by a snap federal election in September.  I think the Coalition would want to be sure of the reasons why it won and that they were repeatable nationwide before taking that risk.  Voters often don't respond well to the cynical calling of early elections for perceived advantage.  However it's not impossible for a government that has polled badly all term to pop up a winner in such an election.  The Whitlam government did this in 1974, but had much more reason to force an election.

Equally, just because voting intention has closed up for now doesn't mean we should expect the Coalition to take the lead.  In late 2012 Labor under Gillard staged a big rally but it only lasted a few months before they went back to losing heavily again.

At this stage I expect to have live blogging of the by-elections on this site. 


This section is brought to you by YouGov Galaxy's David Briggs, who as I write this article thrilled the crowd (well the lefties anyway) with a false claim on Sky that no Opposition Leader had ever recovered from a Better Prime Minister deficit as large as Bill Shorten's current 19 points.  In fact Tony Abbott trailed Kevin Rudd by 22 points, 31-53, in the Newspoll of 5-7 July 2013, with Rudd enjoying a massive but very short-lived honeymoon bounce after getting the Precious back.  Two months later, Rudd was smashed at the election.

Yes of course that's different, but there is other evidence that trying to use Shorten's current ratings to predict defeat is silly.  Andrew Peacock went to the 1990 election with a final netsat of -37 and a Better PM deficit to Bob Hawke of 31 points.  Yet the Coalition got a 50.1% 2PP vote, which would probably be enough to put Shorten in the Lodge if Labor gets it this time round.  Abbott, as noted, won massively, but in November 2012 he had a netsat of -36 and trailed Julia Gillard as better PM by 14 points.  Pretty much nobody was saying Tony Abbott could not win!  I couldn't even get him declared "embattled"!

The other problem is that we don't know very much about the personal ratings histories of winning Opposition Leaders.  There have only been three of them in Newspoll history.  Older polling history gives us another three, but Fraser had been made caretaker PM by the time he won, while Hawke's Opposition Leadership was very brief.  The sample size is just too small to say that Opposition Leaders can't win from certain personal ratings.  Before Abbott, no federal Opposition Leader had come back from a netsat worse than -12.

None of this is to deny that Shorten's position is fragile.  The Coalition declined to throw Abbott under the bus when he was on the nose in late 2012, because Abbott was ahead where it counted (the 2PP).  While Labor stays ahead by that measure I do not expect that Shorten would be rolled.  However, should the Coalition move into the 2PP lead, the wheels could fall off quickly.  That's especially if the Coalition wins one or both of the Braddon and Longman by-elections.  The recent noises by Anthony Albanese might be seen as an insurance strategy in case Shorten does have to be cut loose.  (How they would do it given the "Rudd rules" is another issue, but I suspect some way would be found.)

Perhaps a loss in Braddon will be survivable on the grounds that Tasmania is a bit different and nobody will really understand why the loss happened (I'm finding it a very hard seat to make any predictive sense of) but perhaps even there a post-defeat narrative might emerge.  A possible damaging narrative in that case would be that Braddon voters wanted to get on with building the economy, and if tax cuts helped with that well and good, and that Shorten's tax message wasn't playing well in areas hungry for jobs.

Anyway, recent Newspolls have seen Malcolm Turnbull break into single-digit negative netsats.  This week he's at -8 (41-49) and last fortnight -6 (42-48).  The -6 was his best result since March 2016.  Shorten, meanwhile, remains in the -20s, currently at -24 (32-56).  Especially given Shorten's recent taxation policy bungle, his ratings are remarkably static - his satisfaction score has now been 32, 33 or 34 for the last 19 Newspolls in a row.

Tax Polling

Lastly, I am not going to reproduce any polling figures about the government's proposed business tax cuts because the polling on this issue has all been useless!  One would think that if people commissioning polls wanted to know what voters thought about reducing business taxes for companies with a turnover over a certain amount, they would, er, just ask that.  Really, how hard could that be - assuming that you really want to know the answer.  Instead, all polls I have seen on this issue (almost all commissioned) have loaded their preambles and questions with all manner of distractions (usually of a sort that will obviously skew the result in some direction or other) and as a result we can have no sound idea what the public think about the matter.

Group Wirrah Award For Fishy Polling
Awarded To Everyone Who Polled The Issue Of Company Tax Cuts in Mid-2018


  1. Kevin,

    given the margin of error perhaps narrowing is not the right word maybe??

    1. The narrowing since April is more or less certainly real. When there are multiple polls confirming something the effective margin of error goes down. In this case it is based on 15 polls of which only 3 have been as bad for the government as what it was polling before.

      Smaller movements like the half-point shift since last week don't necessarily mean anything. They could be random noise or they could be real. A common mistake is to dismiss all movement within the margin of error as insignificant, but in fact any movement implies a different probability distribution. A further point is that Newspoll has been displaying an under-dispersed pattern of variation (it tends to change less from poll to poll than a poll with its sample size should), meaning that applying standard MOEs to it is unsound.

  2. Myself and a few of my mates have taken revenge on the phone pollsters harassing us up to six times a night by giving the most stupid answers we can think of. Tonite I was a 25 to 35 yo female who liked Brett Whitely but voted Labour and was going to vote for a fourth party on the basis of leadership. I am 50 something, loath Whitely, the libs and labor with equal measure.

    Granted most of these polls are party polling, but the number of people I hear of doing this is increasing, rapidly. There is going to be some head scratching done on Sunday 29th. Something for you to consider in your prognostications. Sorry for being a pain, but the pollsters have outdone their welcome.