Monday, June 18, 2018

Federal Newspoll Records Page

Introduction (June 2018)

This weekend, the Coalition government under PM Malcolm Turnbull trailed on two-party preferred vote for a 34th consecutive Newspoll, an all-time outright record.

As there has been so much interest in federal Newspoll records this year I thought I would start a Newspoll records page, a resource page which I will update and expand as time permits.  Some of the material has previously been published on the Newspoll Wikipedia page, which I fixed up after finding it contained a large number of errors caused by confusion between satisfaction scores and Better Prime Minister.

Suggestions for new categories are welcome, as are corrections.

This may seem like a very nerdy cricket-statsy kind of page full of numbers without much connection to what they mean.  If there is a conclusion I wish to be drawn from all this it is that drawing predictive conclusions from record feats in Newspoll is difficult.  The records are awash with cases of impressive results polled by leaders who turned out to be flops, and poor results polled by leaders who were unexpected winners.  As I commence this page (June 2018) it's especially important to note that changes in Newspoll methods have made long runs of similar results more likely, and dramatic poll-to-poll swings less so.  (This made PM Turnbull's choice of consecutive 2PP losses as a metric to justify removing PM Abbott especially unwise.)  As a result there are three currently ongoing record streaks, probably the most important of which favours Labor while two others favour the Coalition.  


The first Newspoll was released in November 1985.  In the early days Newspoll's regular federal series polled just primary votes for the parties and the satisfaction ratings of the leaders individually.  Better Prime Minister (often called "preferred prime minister" was polled in the leadup to the 1987 and 1990 elections, and then continuously from July 1991 (except possibly for three polls in July-August 1994, for which data are not on the old Newspoll site.)

Two-party preferred (2PP) estimates were published during campaigns from 1993 onwards, and then continuously from Nov-Dec 2002 onwards.  For other Newspolls it is possible to construct 2PP estimates using the primaries and the preference flows from previous elections, but because of rounding issues these are not necessarily exactly the same as the numbers Newspoll would have got. At least two sets of estimates for the early days exist, one by me (which hasn't been published in full yet), and one slightly different set by Peter Brent (which occurs on his Old Mumble site, but with a formatting error that causes some figures not to line up correctly.) 

The frequency of Newspoll releases has varied over time, and also varies through each year because of the Christmas-New Year break.  See How Often Have Federal Newspolls Been Released? for more information on this.  It means that records for winning and losing streaks often don't compare like with like.

Newspoll was historically a live-interview telephone pollster calling only landlines.  From July 2015 The Australian, which owns the Newspoll brand, contracted the administration of the brand to Galaxy Research.  At the same time there were major changes in the conduct of the poll, including an increase in sample size and a switch to a hybrid interview method using robopoll (automated) landline polling and also using online polling to reach voters who do not have landlines.  The poll also became much less "bouncy" (even after considering sample size), and no explanation for this (also a property of Galaxy's previous federal polling) has been publicly available.

In late 2017 Galaxy Research was acquired by YouGov (becoming YouGov Galaxy) but continued to administer Newspoll by the same methods.  A significant preferencing method change commenced just before the acquisition was announced, but I do not believe it was connected with the acquisition. 

For its 2PP calculations, Newspoll has generally employed the usually very accurate method of last-election preferences.  During 2004 it employed respondent preferences, resulting in a serious failure at the 2004 election at which it had the 2PP at 50-50 in its final poll.  YouGov Galaxy now employs an undisclosed mixture of previous federal and state election preference flows in estimating the 2PP, which produced a sharp change from November 2017 onwards when One Nation flows were adjusted to resemble those seen in recent state elections.


Within this page, I make use of both published 2PP figures and derived last-election 2PP estimates.  I also overwrite the 2004 respondent-preference figures with derived preferences.  Over time I intend to upload a full set of my derived 2PP estimates.

Where a record relies in part or in full on derived 2PP estimates, that result appears in italics.

Streaks that have not yet ended are shown with an asterisk.

There is no set number of results to be included for each category.  The general standard is three plus as many more as I consider worth including.

I do not have exact polling dates for all polls in the last few years readily to hand.  There may be some slight date errors (a day or two) with these but they will be checked and corrected over time.  For early polls taken over multiple weekends in a month I have in some cases for now just listed the month.

Records up to date as of the Newspoll released 17 June 2018.

On to the records ...

2PP Winning And Losing Streaks

Most 2PP Wins In A Row (Government)

53 Labor (Rudd) Jan 2008-April 2010
37 Coalition (Howard) March 1996-Aug 1997
19 Coalition (Howard) Dec 2002-Sep 2003

Most 2PP Losses In A Row (Government)

34* Coalition (Turnbull vs Shorten) Sep 2016 -
33 Labor (Gillard vs Abbott) April 2011-September 2012
30 Coalition (Abbott vs Shorten) Sep 2016-June 2018
27 Coalition (Howard vs Beazley(1)/Rudd (26)) Nov 2006-Nov 2007
25 Labor (Hawke(20)/Keating(5) vs Hewson) May 1990-Mar 1992

Labor (Hawke/Keating) 1990-1992 had the longest trailing sequence in terms of time.

Most 2PP Losses In A Row (Opposition Leader)

37 Beazley (Labor) vs Howard March 1996-Aug 1997
29 Turnbull (Coalition) vs Rudd, Oct 2008-Nov 2009
19 Crean (Labor) vs Howard Dec 2002-Sep 2003
16 Nelson (Coalition) vs Rudd, Jan-Sep 2008

(Note: Beazley's losing streak is 28 polls if using the Peter Brent 2PP estimates, which give the 9-11 May 1997 poll as 50-50, whereas my estimate is 51-49 to Coalition.  There is no right or wrong answer here as there is a subjective issue with dealing with the preferences of One Nation, which did not contest the 1996 election.)

Highs and Lows: Voting Intention

Highest 2PPs (Government) 

63 Labor (Rudd), 29 Feb-2 Mar 2008
61 Labor (Rudd), 18-20 Apr 2008
60 Coalition (Howard), 12-14 April 1996
59 has occurred seven times, all within the first year or so of the Howard and Rudd governments.

Lowest 2PPs (Government)

40 Coalition (Howard), 25-7 May 2007
40 Labor (Keating), 20-22 Aug 1993
41 Coalition (Howard) 13-5 Apr 2007, 11-13 May 2007 and 31 Aug-2 Sep 2007
41 Labor (Gillard), 2-4 Sep 2011 and 27-29 Apr 2012

Highest Primaries (Coalition)

55 (Howard - Government) 12-14 Apr 1996, 10-12 May 1996, 24-6 May 1996
54 (Hewson - Opposition) 20-22 Aug 1993
54 (Howard - Government) 29-31 March 1996

Highest Primaries (Labor)

54 (Hawke - Government) 12-14 June 1986
53 (Hawke - Government) 5-7 June 1986
52 (Rudd - Opposition) 16-8 March 2007 and 25-7 March 2007

Lowest Primaries (Coalition)

31 (Nelson - Opposition) 29 Feb-2 Mar 2008
33 (Nelson - Opposition) 13-15 June 2008
34 (Howard - Government) 26-8 June 1998
34 (Nelson - Opposition) 18-20 Apr 2008
34 (Turnbull - Opposition) 16-18 Oct 2009
34 (Turnbull - Government) Feb 23-26 2017 and Nov 9-12 2017

Lowest Primaries (Labor)

26 (Gillard - Government) 16-18 Sep 2011
27 (Gillard - Government) 8-10 July 2011, 19-21 Aug 2011, 2-4 Sep 2011, 27-29 Apr 2012
The worst Labor primary in opposition is 33, on 23-25 Nov 2011 (Crean), 11-13 Apr 2003 (Crean) and 3-5 Dec 2004 (Latham)

Highest Primaries For Each Minor Party

Nationals: 9% (31 Mar - 2 Apr 2000) Note: Nationals primary released separately from Nov 1999 on.
Greens: 16% (29-30 May 2010)
Democrats: 17% (16-18 March 1990)
One Nation: 13% (26-28 June 1998)

The highest combined vote for parties excluding Labor and the Coalition is 29%, which occurred four times in 2017: Feb 1-4, Feb 22-25, Apr 19-22, July 5-8.

Leadership Satisfactions

Highest peak satisfaction rating (PM - only highest rating for each leader is shown)

71 Kevin Rudd (Labor), 18-20 April 2008
67 John Howard (Coalition), 10-12 May 1996
62 Bob Hawke (Labor), 24-26 Jan 1986

Highest peak satisfaction rating (Opposition Leader - only highest rating for each leader is shown)

68 Kevin Rudd (Labor), 16-18 Feb 2007 and 11-13 May 2007
66 Mark Latham (Labor), 19-21 March 2004
55 John Hewson (Coalition), 17-19 Jan 1992

Lowest career-low satisfaction rating (PM - only lowest rating for each leader is shown)

17 Paul Keating (Labor), 20-22 Aug 1993
23 Julia Gillard (Labor), 2-4 Sep 2011
24 Tony Abbott (Coalition), 6-8 Feb 2015

Lowest career-low satisfaction rating (Opposition Leader - only lowest rating for each leader is shown)

20 Alexander Downer (Coalition), 2-4 Dec 1994 and 16-18 Dec 1994
22 Andrew Peacock (Coalition), 19-21 May 1989
22 John Hewson (Coalition), 11-13 March 1994
22 Simon Crean (Labor), 28-30 Nov 2003

Leadership Dissatisfactions

Highest peak dissatisfaction rating (PM - only highest rating for each leader is shown)

75 Paul Keating (Labor), 3-5 Sep 1993
68 Tony Abbott (Coalition), 6-8 Feb 2015 and 20-22 Feb 2015
68 Julia Gillard (Labor), 2-4 Sep 2011

Highest peak dissatisfaction rating (Opposition Leader - only highest rating for each leader is shown)

69 Alexander Downer (Coalition), 2-4 Dec 1994
67 Andrew Peacock (Coalition), 16-18 March 1990
64 John Hewson (Coalition), 11-13 March 1994

Lowest career-low dissatisfaction rating (PM - only lowest rating for each leader is shown)

(This record and the next are invariably recorded early in a leader's career.)

11 Kevin Rudd (Labor), 18-20 Jan 2008
12 John Howard (Coalition), 15-17 March 1996
22 Malcolm Turnbull (Coalition), 20-22 Nov 2015

Lowest career-low dissatisfaction rating (Opposition Leader - only lowest rating for each leader is shown)

10 Kevin Rudd (Labor), 8-10 Dec 2006
12 Alexander Downer (Coalition), 27-29 May 1994
15 John Hewson (Coalition), 20-22 April 1990
15 Kim Beazley (Labor), 12-14 April 1996

Net Satisfaction

Highest peak netsat  (PM - only highest rating for each leader is shown)

+57 Kevin Rudd (Labor), 29 Feb-2 Mar 2008
+53 John Howard (Coalition), 10-12 May 1996
+38 Malcolm Turnbull (Coalition), 20-22 Nov 2015

Highest peak netsat (Opposition Leader - only highest rating for each leader is shown)

+55 Kevin Rudd (Labor), 16-18 Feb 2007
+51 Mark Latham (Labor), 19-21 Mar 2004
+34 Alexander Downer (Coalition), 8-10 July 1994

Lowest career-low netsat  (PM - only lowest rating for each leader is shown)

-57 Paul Keating (Labor), 20-22 Aug 1993 and 3-5 Sep 1993
-45 Julia Gillard (Labor), 2-4 Sep 2011
-44 Tony Abbott (Coalition), 5-7 Feb 2015

Lowest career-low netsat (Opposition Leader - only lowest rating for each leader is shown)

-49 Alexander Downer (Coalition) 2-4 Dec 1994
-44 Andrew Peacock (Coalition), 16-18 Mar 1990
-42 John Hewson (Coalition), 11-13 Mar 1994

Most consecutive positive netsats (PM)

102 John Howard (Coalition), Aug 2001-Sep 2005
53 Kevin Rudd (Labor), Jan 2008-Apr 2010
28 John Howard (Coalition), March 1996-May 1997
16 Bob Hawke (Labor), April 1987-December 1987

Most consecutive positive netsats (LO)

42 Kim Beazley (Labor), July 1998-Feb 2000
26 Kevin Rudd (Labor), Dec 2006-Nov 2007
25 Kim Beazley (Labor), March 1996-April 1997
25 Mark Latham (Labor), Dec 2003-Nov 2004

Most consecutive negative netsats (PM)

109 Paul Keating (Labor), Jan 1992-Feb 1996 (entire career)
52 Julia Gillard (Labor), March 2011-June 2013
45* Malcolm Turnbull (Coalition), March 2016 - 
36 Tony Abbott (Coalition), Dec 2013-Sep 2015

Most consecutive negative netsats (LO)

67* Bill Shorten (Labor) Feb 2015 - 
65 Tony Abbott (Coalition) Nov 2010-Sep 2013
56 John Hewson (Coalition) April 1992-May 1994
49 John Howard (Coalition) Nov 1985-May 1989

Better Prime Minister

Better Prime Minister correlates with voting intention but also skews towards incumbent PMs.  Historically, the incument PM leads as Better PM about 76% of the time

Largest leads (PM - only largest lead involving each opponent pair is shown)

+66 Rudd (Labor) vs Nelson, 29 Feb-2 Mar 2008
+51 Howard (Coalition) vs Crean, 28-30 Nov 2003
+51 Rudd (Labor) vs Turnbull, 27-29 Nov 2004
+49 Turnbull (Coalition) vs Shorten, 20-22 Nov 2015

Largest defecits (PM  - only largest deficit involving each opponent pair is shown)

-20 Keating (Labor) vs Downer, 8-10 July 1994
-18 Abbott (Coalition) vs Shorten, 5-8 Feb 2015
-16 Keating (Labor) vs Hewson, 20-22 Aug 1993
-13 Howard (Coalition) vs Rudd, 16-18 Mar 2007 and 19-21 Oct 2007

Most consecutive wins vs same opponent (PM)

54* Turnbull (Coalition) vs Shorten, Sep 2015 -
47 Howard (Coalition) vs Crean Dec 2001-Nov 2003
46 Howard (Coalition) vs Beazley, Jan 2005-Nov 2006

Considering all opponents, John Howard recorded 139 consecutive wins from June 2001 to Feb 2007.

Most consecutive wins vs same opponent (LO)

23 Rudd (Labor) vs Howard, Feb-Nov 2007
10 Abbott (Coalition) vs Gillard, June-Nov 2011
9 Hewson (Coalition) vs Keating, May-Aug 1992
8 Shorten (Labor) vs Abbott, Nov 2014-Apr 2015

Poll To Poll Change

Largest 2PP swing to government (Excludes cases where an election was held between polls)

+8 to Labor (Keating), 6-8 Nov 1992
+7 to Coalition (Howard), 28-30 May 2004 (+6 by respondent preferences) - rogue poll
+6 to Coalition (Howard), 21-23 Sep 2001
+6 to Labor (Rudd), 29 Feb - 2 Mar 2008
+6 to Labor (Gillard replaced by Rudd), 27-29 June 2013

Largest 2PP swing to opposition

-8 against Labor (Keating), 20-22 Aug 1993
-8 against Labor (Keating), 23-25 Sep 1994
-7 against Labor (Hawke), 19-21 Sep 1986
-7 against Labor (Hawke), 19-21 May 1989
-7 against Labor (Keating), 10-12 Feb 1995
-7 against Labor (Rudd), 30 Oct-1 Nov 2009

There were three six-point swings against the Howard government.

Largest netsat gain by PM 

+24 Howard (Coalition), 10-12 May 1996
+20 Howard (Coalition), 31 Aug-2 Sep 2001
+19 Hawke (Labor), Feb 1991
+18 Howard (Coalition), 21-3 Sep 2001
+18 Howard (Coalition), 18-20 Octoer 2002

Largest netsat loss by PM

-24 Howard (Coalition), 22-4 Feb 2002
-23 Gillard (Labor), 4-6 March 2011
-20 Hawke (Labor), Feb 1988
-20 Keating (Labor), 20-22 August 1993
-20 Rudd (Labor), 30 April - 2 May 2010

Largest netsat gain by LO

+23 Hewson (Coalition), 22-4 Nov 1991
+23 Beazley (Labor), 19-21 Oct 2001
+18 Abbott (Coalition), 10-2 Sep 2010, immediately following election
+16 Hewson (Coalition), July 1991

Largest netsat loss by LO

-40 Turnbull (Coalition), 26-8 June 2009
-36 Downer (Coalition), 5-7 Aug 1994
-31 Latham (Labor), 2-4 April 2004
-25 Howard (Coalition), 29 Nov-1 Dec 1985

There are no others of -20 or worse but several in the upper teens.

Largest Better PM margin increase by PM vs same opponent

See this poll roundup for detailed discussion of this record.

+17 Howard (Coalition) vs Beazley, 10-12 Mar 2006
+15 Keating (Labor) vs Hewson, 19-21 Mar 1993, following election victory
+15 Howard (Coalition) vs Beazley, 16-18 Oct 1998, following election victory
+15 Howard (Coalition) vs Latham, 19-21 Nov 2014, following election victory
+15 Rudd (Labor) vs Turnbull, 26-8 June 2009

Largest Better PM margin decrease by PM vs same opponent

-23 Keating (Labor) vs Hewson, 20-22 Aug 1993
-18 Howard (Coalition) vs Crean, 22-24 Feb 2002
-14 Howard (Coalition) vs Beazley, 17-19 Oct 1997
-14 Howard (Coalition) vs Beazley, 19-21 Oct 2001
-14 Rudd (Labor) vs Abbott, 23-5 July 2010

(primary vote changes to be added)

Thursday, May 31, 2018

Not-A-Poll: Worst Opposition Leader Of The Past 45 Years: Round 1

For comments on the Braddon poll see Braddon guide.

Admin note: apologies for delays in comment clearing as I am not currently receiving notification emails for comments - this is a global Blogger issue which will hopefully be fixed soon. Also I have had a report that at least one reader can see the old Not-A-Polls but not the new ones.  If anyone else is getting exactly this problem please report it at , preferably with browser detail + whether you are using a mobile phone.


Following on from a long-running Not-A-Poll series in which Gough Whitlam was this site's pick for the best Prime Minister of the Last 45 Years, Worst Prime Minister only needed a single round for the knockout.

The left-wing skew of this site's reader-base was again apparent, but the result leaves little room for doubt that Tony Abbott would have won the thing anyway.  He finished almost fifty points ahead of John Howard, a result that barely moved through the time the Not-A-Poll was running. Even with the left-wing skew, Howard's second place might be surprising to those who are accustomed to seeing him rate highly but there is an argument for it out there.  Kevin Rudd eventually managed to beat Julia Gillard among the ALP contenders, while the very miserly totals for Paul Keating and Bob Hawke are interesting.  Keating was reviled as PM during his tenure (not only by the right but by a fair slab of the left as well) but nobody much hates him anymore.

Worst Opposition Leader

I'm now starting a Worst Opposition Leader series that will run for an unknown number of months.  The rules are:

* I have split the leaders into two groups - one including those who at some stage were Prime Minister and one including those who have not yet been PM.  The reason for this is to find out something interesting even if views of leaders' qualities as PM colour opinions of their role as Opposition Leader.

* Once the winner of each group has been determined, the two group winners will contest the Grand Final.

* Any leader who scores over 50% in a round will immediately win that bracket.

* At the end of each round in each bracket, the following are all eliminated
- the candidate in last place
- any candidate who would be mathematically unable to win from that position in a preferential election
- any candidate with less than 8% of the vote

* Because John Howard had two widely separated Opposition Leaderships, one of which ended in a landslide victory and the other one of which wasn't so successful, Howard appears in Group 2 twice.  The threshhold applies to both his options combined, but each has an individual threshhold of 4% as well.

* Ties will be broken against the leader who ceased to be Opposition Leader longest ago.  A tie in the final will be broken against the leader who was never PM.

* I reserve the right to make any changes necessary to the results, and to declare a result at any time, to deal with any cases of multiple voting that are detected.

Gee there are some duds in group 1.  It brings back memories, it does.  After this is finished I think I'll run a Best State Premiers series. 

Voting on round 1 for both groups is open in the sidebar until 6 pm 30 June.  For mobile phone users you may need to scroll down on the main page and click on View web version to vote.

Wednesday, May 30, 2018

Poll Roundup: Unwanted Records On Both Sides

2PP Aggregate: 52.3 to Labor (last-election preferences) 
51.8 with One Nation adjustment 
Coalition has improved 1.2 points in last eight weeks and now in best position since late 2016
However Labor would still almost certainly win election "held now"

It's been a little while since the last federal poll roundup (highlighting the major issue of Newspoll preferencing changes); I meant to do one in Budget week but was simply too busy with other things.  In the last five weeks the Turnbull government has kept the modest improvements in national polling that it made during April, but there has not been clear evidence of anything more.  Despite a lot of media excitement about the possibility of a quick election off the back of some possible success in the Super Saturday by-elections on July 28, we are so far not seeing anything in aggregated polling to get so excited about.

The recent national polls have been:

* Two Newspolls with headline figures of 51-49 then 52-48 to Labor.  As the previous article notes there has now been official confirmation that Newspoll is using a preference distribution for One Nation that is derived from recent state elections.  My aggregate's headline figure uses last-election preferences, and on that basis I aggregated these at 52.1 and 52.9 to Labor respectively.  The 51-49, as with the previous 51-49, was off primaries that would have normally come out to 52.4 by last-election preferences, but the fact that the published 2PP was 51 again suggested something a bit closer.  (An alternative view is that Newspoll might have changed methods twice, but we can't conclude that reliably off just two polls if so.  Also, this week's difference between the two methods was "only" 0.9).

Sunday, May 27, 2018

2018 Braddon By-Election

BRADDON (Tas, ALP 2.2%) 
By-election July 28
Justine Keay (ALP) vs Brett Whiteley (Lib), minor party candidates to be declared
Cause of by-election: Resignation caused by Section 44 ineligibility
Outlook: Historic patterns suggest Labor should retain, although early polling is poor.

With the far-off date for the Super Saturday by-elections now announced, I've decided that's reason enough to put up a guide post for the Braddon contest.  This article will be updated up til polling day with I may do similar posts for some of the other seats, but if so that will probably be only after credible polling is conducted.  (And no, a ReachTEL seat poll of Longman asking people how they would vote in a federal election rather than a by-election isn't what I had in mind here.)

At the moment many people are talking about the by-election date and whether this is some kind of political stitch-up to harm Labor or whether it is all Labor's fault because its ineligible MPs did not resign or at least arrange to be referred sooner.  By the by-election day the parties won't still be talking about that, at least not if they have the slightest sense.

We still don't know for certain that the by-elections are on, since in theory the Government could now go to an early general election in August or early September without needing to run a conspicuously long campaign for it.  In this case the by-elections would be cancelled.  I don't think this is likely at all, but at least cancelling the by-elections would now give the Coalition some kind of plausible argument for going early should political circumstances favour it in the next two months.

The electorate

Note: The by-election will be fought on the old boundaries. Labor won in 2016 with a 52.2:47.8 margin.  Many sources are still giving the post-redistribution margin of 1.6%.

Braddon is a regional/rural seat covering north-western and western Tasmania as well as King Island.  It includes the small regional cities of Devonport and Burnie and the large town of Ulverstone, the rural north-west (Smithton, Wynyard) and the west coast mining and tourism towns (Queenstown, Zeehan, Strahan).  The latter were added in the 2007 redistribution, to Labor's benefit.

Braddon was Liberal-held from 1975 to 1998 (for longer histories see Poll Bludger and Tally Room).  Since then it has been a swinging marginal seat, not far behind its volatile neighbour Bass.  Braddon has changed parties at five of the last seven federal elections, but Labor has had the better of it in that time winning five times to the Liberals' two.  Both the Liberal wins in that time, in 2004 and 2013, were linked to backlashes against Labor over forestry issues (Mark Latham's too-green forests policy and Styx trip with Bob Brown in 2004 and the state Labor-Green government's "peace deal" in 2013). 

Especially in the 1980s and 1990s, Braddon was very socially conservative but it has changed greatly in the last 20 years.  However, it remains economically vulnerable with very low school completion rates and median incomes, as noted in the Poll Bludger preview. 

Burnie and the west coast tend to vote strongly Labor, while the Liberal vote is strong in the far north-west and in the rural settlements surrounding the coastal cities and towns.


Justine Keay was elected for the first time in 2016.  She had previously stood as a minor Labor candidate in the 2014 state election, and had been a Devonport councillor and an electorate officer.

Keay is much more educated than the average Braddon resident, with a degree collection in geography and history, environmental management (x2) and psychology. In other respects she ticks boxes for empathy with Braddon's battlers (seventh-generation Tasmanian, had to move interstate for work, former checkout operator and Newstart recipient, etc).  Keay was targeted over her comments about asylum seeker policy in one interview in the 2016 campaign, an attack which in the end had no effect whatsoever (perhaps because of Braddon's very low level of ethnic diversity.)   As an MP Keay has portrayed herself as a folksy fighter for locals and the disadvantaged.

Keay had to resign her council seat to contest the 2016 election, a big risk as the seat appeared winnable but difficult.  It turned out that Labor's legal advice was wrong both on the question of whether councillors could be federal MPs, and on when a candidate had to finish renouncing their citizenship.  Keay was the first of the five known slow-renouncers to come under scrutiny (we were talking about this here in August, and well done to commenter Dave O who correctly predicted the outcome).   She said that she held on to her inherited UK dual citizenship so long because it was a remaining connection with her late father. She has maintained that she was a victim of a change in the interpretation of Section 44 after following the best possible legal advice, though I would call it more a clarification by the court and say that Labor's advice was blind to an always obvious risk.  Keay's interview with commercial radio presenter Brian Carlton - a longtime Keay sceptic on the matter - has been reported elsewhere as "torrid" but to my ears at least there was plenty of stirring back and forth in that one.

Main Challenger

The Liberal challenger is former one-term MP Brett Whiteley (who doesn't seem as off and running with the online presence bit as Keay just yet).  Whiteley was also a state MP for Braddon from 2002 to 2010, unseated (and outspent) after two terms by a within-party challenge from businessman Adam Brooks.  As a state MP Whiteley was regarded as a hard-right firebrand. This aspect was less conspicuous during his federal tenure, though he did at one stage call for drug-testing of welfare recipients.  Since his defeat, Whiteley has worked as an advisor to Angus Taylor.

In the leadup to the 2016 election Whiteley had rather average personal ratings in polling, and he was not able to turn the personal vote advantage of replacing a long-term Labor incumbent to his benefit.  Reports I receive of Whiteley vary - he clearly has a deep passion for political life, but people not fully on his side of the fence sometimes tell me that they found him abrasive or dislikeable in person.  In the 2016 campaign Whiteley was the target of a bizarre sign vandalism episode with an unknown "artist" superimposing bondage masks on his pictures and changing his name to "GIMP".  It will be interesting to see if this recurs.

Other Challengers

The Greens candidate is Jarrod Edwards (announcement), an Indigenous land management supervisor with the Tasmanian Aboriginal Centre.  Edwards, a new candidate, is a fairly prominent local spokesman on Aboriginal heritage issues including damage to West Coast middens.  The Greens polled extremely poorly at the state election in this seat getting only 3.6% but have decided to have a go anyway (maintaining their record of contesting every federal by-election since Warringah 1994).  The party polled 6.7% in the seat at the last federal election.

The Shooters, Fishers and Farmers candidate is Brett Neal, a third-generation farmer from Yolla.  Neal polled 459 votes at the state election for this seat, at which the party polled 2.54% in Braddon.  Probably the Shooters vote was on the low side at the state election because of competition from the Lambie Network.

Other Possible Contenders

By-elections frequently attract large fields of eccentric no-hopers, in part because candidates do not need to live in the seat to contest it.  Hopefully this will be less of an issue here because of the number of by-elections around the country at the same time.

The Jacqui Lambie Network and Pauline Hanson's One Nation have announced they are not contesting.

The Recreational Fishers Party polled strongly in 2016, but this turns out to have been just a none-of-the-above type vote as the party sank without a trace in the Senate count.  The party has since been deregistered (and ditto for the failed Renewable Energy preference-harvester attempt).

An interesting prospect could be fisherman Craig Garland.  Garland, a self-styled anti-politician, ran a homespun campaign around seal relocation and salmon farming issues in the state campaign, polling nearly 2000 votes and outlasting the Greens in the cutup.  The vote for Garland was very concentrated in the coastal booths in the far north-west.


Keay is likely to campaign heavily on health, education, employment and regional services and to continue with the federal election theme of a Turnbull Liberal government being the wrong fit for low-income Tasmania.  Populist attacks on tax cuts for the rich are staple fare for Keay backers already.

Whiteley has so far flagged that he will try to run on his ability to deliver the bacon for his electorate.  He has aimed to highlight concrete achievements during his previous term as a government MP, while various Liberal proxies (eg Senator Eric Abetz) have waded in with attacks on Keay as having achieved nothing concrete in Opposition and being prone to waffle about what she has done.  The Keay campaign has been quick to adjust to this attack and ward it off with specific examples, but the Liberals have continued pressing the point that Keay cannot deliver Labor campaign promises because she is not in government.

The Greens greatly struggled for any kind of policy oxygen at all in Braddon at the state election and the plight of endangered subspecies of cute fluffy birds on King Island might not provide them with any this time either, but it will be interesting to see if Peter Whish-Wilson has any success in drawing policy commitments on the issue from the majors.  It will also be interesting to see if the choice of a new candidate refreshes their appeal at all given that the party has been prone to preselect serial candidates in the past.  However Braddon remains very difficult territory for the party.


A Sky News ReachTEL taken around 2 June had the Liberals ahead 54-46 based on a respondent-allocated preference distribution.  Primary votes with "undecided" redistributed came out to Liberal 48.2 Labor 34.5 Green 6.6 Ind 7.2 Other 3.5.  This would also come out to 54-46 by last-election preferences for this specific seat. The question wording did refer specifically to a by-election but I do not know if it named candidates. The primary votes are very strong for the Liberals so it will be interesting to see what other polling emerges (if any). ReachTEL federal polling in Tasmania has a history of overestimating the Liberal vote and underestimating the Labor vote, but the final ReachTEL in this particular electorate was quite accurate at the 2016 election if previous-election preferences were used. (This is unlike the adjacent electorate of Bass where all polls, whether by ReachTEL or Newspoll/Galaxy, were terrible.) Also the final ReachTEL was quite accurate at the 2018 state election.

The Liberal Party was spruiking a commissioned MediaReach federal poll taken 11-12 April which supposedly showed them with a 53-20 primary vote lead in Braddon, or about a 14% two-party swing.  However the sample size of this poll was only 756 statewide, meaning c. 150 per electorate, which is meaningless given the issues in seat polls we have seen in recent years.  (I also have other concerns about that poll including an implausibly low Wilkie vote in Denison and an implausibly high vote for the Greens in Braddon, as well as the usual concern that the full wording of all questions asked has not been released).

There have been various reports that internal polling as of early June is vaguely similar to the ReachTEL although none of these have given figures or mentioned a named source

The ReachTEL was criticised in some quarters because of showing high support for company tax cuts, which was considered atypical of a low-income electorate.  However the poll does not seem to have asked specifically about tax cuts for rich companies.  Also, Braddon is different to low-income city electorates.  It has a long history of being very pro-development and desperate for jobs.  The poll also showed strong opposition to allowing refugees on Nauru and Manus Island to settle in Australia.  I have not been able to find comparable polling from this electorate in the past.

Seat polling in Australia has been unreliable at both the last two federal elections and should be treated with great caution. It shouldn't be dismissed as evidence entirely, but individual seat polls shouldn't be weighted very highly either.  A review of the 2016 federal election seat polling by Simon Jackman and Luke Mansillo found that seat polls had the average errors of samples one-sixth their actual sample size, which means, for instance, that the Braddon 54-46 ReachTEL should be treated as having a margin of error of about 6% on the 2PP, and up to 8% on the major party primaries.  Even when multiple polls of the same seat had similar results, that result was sometimes wildly wrong.  At the 2013 election, seat polling skewed massively to the Coalition. Changes in national polling may well be more significant.


Election betting is not reliably predictive, and this was recently shown in Tasmania with the Hodgman state government being returned with a majority, easily, despite this having been at odds at one stage as long as $15.

When this article was written all bookmakers had Labor around the same mark (eg Sportsbet and (bookie) 1.40, Ladbrokes 1.44) but varied greatly in their generosity level on the Liberals' chances ((bookie) 2.80 Sportsbet 2.40 Ladbrokes 2.00).  These could be converted to between a 58% and 67% implied expected chance of a Labor win.

Following the 54-46 ReachTEL, Sportsbet was soon at 1.32/2.65 and (bookie) 1.55/2.35.  As of 5 June, Sportsbet had closed to 1.50/2.15 (an implied 59% Labor chance).

As of 17 June, Sportsbet 1.66/1.90 (bookie) 1.55/2.35 Ladbrokes 1.60/2.00, so the range of implied chances is 53% to 60% - the markets are very unsure about this!

Note: References to the name of one bookie have been removed above and replaced with "(bookie)" because their SEO person hassled me asking me to link to their site.

Prospects: Will The State Election Be Repeated?

A much longer and wonkier "prospects" section than normal here, for reasons that will soon be apparent!

As noted in my previous overview of the by-elections, governments virtually never win by-elections in Opposition held seats.  The average pattern in such seats is for a small swing to the Opposition, but there is a lot of variation, meaning that in the case of a close seat a Government win is possible, despite the history of this generally not happening.  However, add in the fact that Keay is recontesting (not the case in nearly all previous such by-elections) and that the federal government is still struggling in polling, and one might on average expect a swing of, say, 3-4 points to Labor.  That is, however, "all else being equal", so what are the reasons why it might not be?

The main argument being made for a Liberal victory is the theory of a rub-off effect from the recent state election triumph of the Hodgman Liberal government.  In Braddon, the Liberals recorded a 56.1% primary to just 27.3% for Labor, which if repeated would translate to about a 62-38 2PP pasting.  However, majority government was at stake in the state election, which it won't be at the by-election.  The by-election is also unlikely to see a repeat of the cashed-up campaigns from gambling interests against Labor and the Greens.  And also, Tasmanians have a long history of often sharp divergences between state and federal voting intention.  For instance, the 2013 federal election saw Labor narrowly win the 2PP in the state (despite lucking out on the seat distribution) but just six months later Labor were thrashed in the state election.  Then Labor picked up a 6% swing at the 2016 federal election, then they were thumped again in the state election earlier this year.

I've been very interested in trying to test this idea of a rub-off effect between state and federal elections (in either direction) in general.  Unfortunately it's difficult to test in the case of state elections rubbing off on federal by-elections, because there are so few federal by-elections contested by both parties these days.  So the best I can do is look at the link between state and federal general elections and just note that a by-election should offer improved prospects for an Opposition.

For Tasmania for the last fifty years, I've found that there does appear to be a weak relationship between performance at general elections of either kind and the election of the other kind that follows it (whether closely or otherwise).  After translating the Tasmanian state elections to estimated 2PPs if voters all gave preferences (which they don't, but we need some basis to compare the two election types fairly) it looks like this:

This relationship firstly explains only 10% of the variation in results of following elections, and secondly isn't actually statistically significant by itself.  Thirdly, the slope of the line is rather shallow.  From what I've seen looking at other states, it seems that it is significant on a national basis, but I still have to reconfigure some of the other-state data to 2PP estimates that take account of preference flows from the Greens (my original attempt used differences in major party primary votes).

In any case,  we can't say that a good performance in a state election causes a party to do well in the next federal election, or vice versa.  Historically there have been runs in Tasmania of the Labor Party doing badly at both levels or well at both levels at the same time.  An example of the former was the 1980s, where the state Liberal Party had electoral success based on undivided support for the Franklin Dam project, which divided the ALP.  The Liberals also polled strongly in the state at federal elections at this time as federal Labor was seen as interfering in Tasmania's state rights.  In the 2000s there was a different pattern in which Labor did generally well at both state and federal levels (despite the loss of two federal seats in 2004).

In recent times we have seen mixed-up results,  so whether the pattern applies at the moment is debatable.  Even if it did, the predicted Labor 2PP score in Braddon in the next federal election would be about 47, but this is a by-election. After throwing in the expected advantage to the Opposition (a few points given that the incumbent is recontesting) this pattern still doesn't predict a Liberal victory.

The rub-off effect theory also importantly maintains that the state result is more likely to rub off on the by-election because the by-election is so soon after the state election.  The delaying of the by-election has taken some of the oomph out of this argument, but how much oomph was present to begin with?  If this claim was predictive at all we would expect that the "swings" from a state election to a federal election (or perhaps vice versa) would be smaller if the gap between the two was shorter. And this is what we have seen in Tasmania in the past:

Even off such a small number of data points this is a very significant relationship: it looks like state election voting tends to contaminate federal election voting in Tasmania if a federal election follows less than a year after a state one.  (The average swing from state election to state election in Tasmania is about six points, the highest of any Australian state.)

However, when I looked at other states I found that, oddly, NSW and WA both showed a similarly strong historic relationship while Victoria, SA and Queensland all showed none at all.  A good example of why Victoria doesn't was the strong performance of the Keating government shortly after Victorians elected the Kennett government, and clearly in reaction to the agenda of the latter.  Furthermore no state showed this relationship in the reverse direction (state election results tending to be more like federal results in that state if the state election followed soon after the federal election.)  So, hmmm ... state election rub-off might help Whiteley, but it might not, and it's impossible to really say how much it will help him if it does.  It's a reason not to be too confident re Labor's chances, but it's too messy to read too much into.

Prospects: Other Issues

One other argument being made by upbeat Liberals regarding this seat is the rate of employment and economic growth in Braddon.  However this is exactly the same reason Whiteley was supposed to hold his seat in 2016, but he still lost.  A more likely reason for a rebound would be that the 2016 Liberal results in Tasmania generally were poor.  This might be attributed to the innovation-based campaign the Government ran being received poorly in the regions, and also to anxiety over health and other services.

As for the issue of Justine Keay herself having caused this by-election by having to resign, there is no evidence from the three previous such by-elections that this actually does any harm (if anything, possibly the contrary, though the evidence there is messy too).  I don't think the parties will still be talking about this much on election day - the campaign is more likely to be about who will do the best for the electorate, and what the electorate thinks of the current federal government.

One further factor that may assist Whiteley is government infrastructure spending in the area arising from the 2018 Budget.  Whether this has done enough to convince voters the Turnbull government is taking the region seriously remains to be seen.

Braddon is a very difficult seat for a contender outside the major parties to be competitive in.  The low primary vote for the Greens is one reason for this, and another is that the major party primaries tend to be close together.

Friday, May 18, 2018

Cable car catch-up

Not everyone notices when I post links to updates on the sidebar, so for those who read from the top, just a very quick note that I have updated my old article on polling on the proposed kunanyi/Mt Wellington Cable Car following the release of two new polls by groups opposed to the project.

I thought I'd highlight this with a note at the top because I'm actually mildly annoyed about it.  Up til now it has only been the prospective developers of the cable car through time who have engaged in the usual silly commissioned-poll games involving misleading polls with biased preambles.

Now it's both sides.

If you respect the mountain, you should also respect the facts.

Thursday, May 17, 2018

EMRS: No Real Change Since Election

EMRS: Liberal 47 (-3.2 from election), Labor 30 (-2.8), Green 14 (+3.7) Ind/Other 8 (+1.2)
Interpretation: Liberal 49.5 (-0.7) Labor 31.5 (-1.3) Green 11 (+0.7) Ind/Other 8
Outcome if election "held now" based on this poll: Liberal majority government (c. 13-9-3)
However it is unlikely in practice Greens would be in a position to regain Bass so quickly.
Poll provides no evidence that any party's support has changed.

The December 2017 EMRS poll, taken three months out from the 2018 state election, proved to be completely unpredictive of the outcome.  It had a 17% swing against the Government (the actual swing in the end was 1%), a 3.2% swing to the Greens (the actual swing was 3.5% against) and an 8% vote for the Jacqui Lambie Network (who in the end got 3.2).  If a poll taken three months out is predictively worse than useless, what can we say of one taken two months into an expected four-year term?

Wednesday, May 9, 2018

The Expected "Super Saturday" By-Elections

Today's four resignations from the House of Representatives following the Section 44 disqualification of Labor Senator Katy Gallagher is expected to trigger a day of at least five by-elections, or at least a cluster of by-elections close to each other.  The following seats are affected:

Braddon, TAS (ALP, 2.2%)
Fremantle, WA (ALP, 7.5%)
Longman, QLD (ALP, 0.8%)
Mayo, SA (Centre Alliance vs Lib, 5.0%; Lib vs ALP 5.4%)
Perth, WA (ALP, 3.3%) 

See The Tally Room for detailed histories of these seats.  Also see the Poll Bludger thread for Perth.  All seats will be contested on the old boundaries, irrespective of redistributions.

It's possible that given the strictness of the High Court's ruling, other MPs may come under pressure to resign or be referred to the High Court (note: as of Friday the media are suddenly all over the Anne Aly story, which has been known via Jeremy Gans' Twitter comments for months), though the Coalition may not be in any great hurry to hunt down any more and invite more scrutiny of its own remaining unclear cases.  The by-elections are not just a nuisance for Labor, but also for the Coalition, which must either throw resources into contesting them seriously or else chicken out and leave voters wondering what all the fuss was about.

Australia has never had a day with five federal by-elections before, so it would be quite a novelty.  Three were held on the same day in 1981 and 1984.  In 1994 four were held across three weekends following a cluster of resignations, but the resignations came on different days.  At state level, NSW has often held multiple by-elections on the same day.

Saturday, May 5, 2018

Legislative Council 2018: Hobart and Prosser Live And Postcount

Hobart: CALLED (7:15) Rob Valentine (IND) retain
Prosser: Howlett (Lib) defeated Lambert (ALP) by 887 votes after exclusion of Mav (IND).

Welcome to my live coverage thread for the 2018 Legislative Council elections for Hobart and Prosser.  After the craziest week in the Lower House in decades, we now come to the voting for two Legislative Council seats - Hobart, where Rob Valentine faces his first defence and Prosser, a new seat created by a redistribution.  The left currently has the numbers in the Legislative Council, care of four Labor MLCs and four left-wing independents, and that's not changing unless the Liberals can pull off a big upset in Hobart.  Indeed, should Prosser go badly for them, the balance will become even worse for the Hodgman government.  By the way, should a party-endorsed candidate win either seat then the party representation will reach a new all-time high (see the chart at Poll Bludger to see how the parties ebbed and flowed in the last several years.)

Comments will follow below the dotted line, scrolling from the earliest upwards. All the seats will be covered together.  I'm leaving this bit of text at the top to try to prevent colours from the heading running into the main text.

Wednesday, May 2, 2018

Mayhem On Day 1 As Hickey Nicks The Chair!

(NOTE: My very old 25 vs 35 seats article has been updated.)

The opening of the Tasmanian parliament on May the 1st was meant to be a routine affair.  After the election of the new Speaker we were expecting to start off with the ritual parliamentary theatre of a Greens no-confidence motion over the Liberals' failure to disclose any pokies-related donations prior to their re-election in March.  It seems to be the Greens' lot in life lately to have their thunder stolen but in this case they won't mind.  Former Hobart Lord Mayor Sue Hickey has decided that starting her parliamentary career on the backbench was not acceptable, and she's nabbed the Speakership instead.  

That part is by no means unprecedented.  In 1992, Ray Groom's Liberals won 19 of the then 35 seats and they nominated the flamboyant Michael Hodgman (Will's father) as Speaker.  However, Liberal Graeme Page and a colleague voted for Page and Page was elected Speaker with Labor and Green support, 18 votes to 17.  Previous Labor Speaker Michael Polley is generally credited with hatching the plot.  In this case there had been some speculation that the former Liberal Speaker Mark Shelton could do the same thing (if he wanted) but the Greens poured cold water on it.  While I did tweet that this year it would only take one renegade Liberal to repeat the dose, that tweet was better classified as a bit of stirring at Hidding's expense than a serious prediction.