Friday, 19 January 2018

Farrar's Honeymoon Scam

Kiwiblog's Smooth Operator David Farrar

Diamond life, lover boy
He move in space with minimum waste and maximum joy
City lights and business nights
When you require streetcar desire for higher heights
No need to ask, he's a smooth operator
Smooth operator, smooth operator
Coast to coast, LA to Chicago, western male
Across the north and south to Key Largo love for sale
Melts all your memories and change into gold
His eyes are like angels but his heart is cold
- Sade


Over recent weeks, National Party agent provocateur David Farrar has managed to profoundly shape mainstream media analysis of the Post-Election Mood.

In two highly influential Kiwiblog posts, Farrar set out to aggressively heighten expectations of the new Ardern Labour Government's impending Poll performance (What sort of poll boost should the new Government get? November 6, 2017 - published some 2 weeks before the very first poll was released) and then subsequently went out of his way to ignore the first two Post-Election polls,  instead waiting 5 weeks for the third poll to emerge, before declaring that Labour had conspicuously failed to live up to expectations (No real bounce for Labour in first Colmar Brunton poll December 10, 2017).

Farrar's interplay of  arguments (both explicit and implicit) can be summarised as:

(1) Incoming governments traditionally enjoy a huge Honeymoon surge of post-Election support

(specifically Farrar claims that in the immediate wake of the 1999 and 2008 General Elections,
both the Clark Labour and Key National Governments experienced a remarkable 11 percentage point spike in ratings.  In his Nov 6 post he explicitly states that these spikes occurred in the very first post-Election poll: "In 1999 the Labour Government won 39% of the vote. The next poll had them at 50%. So they got an 11% bump from winning ... In 2008 National won 45% of the vote. The next poll had them at 56%. SO also an 11% bump from winning." - (thus making it all the more curious that he would subsequently go out of his way to  avoid the very first post-2017 Election poll).  In his Dec 10 follow-up post, Farrar employs somewhat looser terminology, suggesting these spectacular surges of support were registered in "the polls" (plural) immediately after the 1999 / 2008 General Elections. "In 1999 the incoming Labour Government got an 11% bump in the polls, as did the incoming National Government in 2008."  Regardless of the particular post, however,  journalists were clearly in no doubt that he was situating the emergence of this honeymoon bounce specifically in the immediate few weeks following the respective Elections)

(2) This massive Post-Election Poll Bounce comes largely or entirely at the expense of the Opposition Bloc and in particular the Major Opposition Party

(explicitly stated in regard to 2008 (Dec 10 Post) where Farrar contrasts the putative 11 point bump to the incoming Key Govt with Labour's 6 point fall.  But also more broadly implicit in Farrar's notion that these alleged post-Election spikes occur because "people like to give a new government a fair go" (Nov 6 Post). The idea of a new Administration enjoying a honeymoon, a spontaneous outpouring of goodwill, with voters feeling generous and giving them the benefit of the doubt implies a segment of Opposition supporters swinging to the Govt.  Certainly, journalists John Roughan and Heather du Plessis Allan inferred this message from Farrar's posts and both essentially went on to embroider what the Kiwiblog host had originally woven)

(3) Such a Poll Bounce failed to materialise in the immediate aftermath of the formation of  the 2017 Labour-NZ First-Green Government

(in his Nov 6 post, Farrar suggests that if the Ardern Govt "follows (historic) form, then the next public polls should show Labour at 48%."   Having seeded this expectation of a dramatic 11 point boost in the media, he then waits 5 weeks for the Colmar Brunton to emerge before spinning Labour's 2 point spike as "a disappointment indeed for Labour." (No real bounce for Labour in first Colmar Brunton poll, Dec 10)

(4) This failure is unprecedented in Modern Political History

(5) The reasons for this alleged failure are two-fold: (a) In 2017, "there was no clear vote for change as happened in 1999 and 2008" and (b) Labour "have had a pretty shambolic start to Government" (Dec 10 post)

(6) None of this augurs well for the survival /  longevity / future electoral prospects of the Ardern Govt

Media UpTake

As so often over recent years, Farrar's carefully-contrived narrative quickly gained wide currency among MSM Notables. Despite the central involvement of both Farrar and segments of the Fourth Estate in the murky 2014 Dirty Politics scandal, journalists still seem more than happy to take his claims at face value and to widely disseminate them throughout the media.

First off the mark (anticipating the early December Colmar Brunton) was Spinoff's vaguely liberal-left Senior Editor, Toby Manhire (who showed due deference to the Kiwiblog host by describing him as National's "polling wizard"):

The first big poll for ages is due. What would be a good result for Labour ? Typically a new government can expect a bump in support after a victory. As David Farrar noted recently at Kiwiblog, 1999 and 2008 both saw the incoming government register a lift of 11%. "If this government follows form, then the next public polls should show Labour at 48%", he said.

Manhire was obviously a little sceptical:
Labour at 48% would of course be a miracle. That would mean they'd literally doubled their poll numbers in just over three months ... It's a bit different, too, given Labour did not win a plurality of the election tally.

And yet, in the end, Manhire still allowed his expectations (and therefore those of his readers) to be broadly shaped by Farrar's narrative, going as far as suggesting a mild fall for the Nats in the upcoming poll would constitute an excellent result for them:
Just over 44% of voters ticked National. If they can hit 42% or better on Sunday, they'll be thrilled ... Certainly (Labour will) expect to go higher. Anything under 40% will be a disappointment. Anything above National will be a relief. Anything over 45% will warrant champagne.

Now, far be it from me to suggest Manhire allowed himself to be played like a finely-tuned stradivarius here, but in his subsequent Dec 10 post, the Kiwiblog host transformed this whole exchange into a kind of feedback loop, using Manhire's inflated (and Farrar-influenced) predictions to declare that, yes, National certainly were thrilled with the Colmar Brunton result while, yes, Labour certainly were "disappointed indeed".  

The Shock-Horror headline-juggernaut No Traditional Poll Bounce for New Ardern Govt ! really took off, however, in the immediate wake of the release of the Colmar Brunton on Dec 9 and Farrar's Kiwiblog post the following day: the major precepts of which became tightly woven into the subsequent plethora of Op-Eds, Editorials and media reports. Something for which Farrar and the National Party can be supremely satisfied.

Not bothering to critically scrutinize Farrar's lines in any way, Senior Herald writer John Roughan decided instead to joyfully run with them like a small child through the fields with a kite (Polled voters back National's economic legacy):
A new government normally gets a good lift in the first post-election poll. It stands to reason. A certain number of those who didn't vote for it find the sky hasn't fallen, they're finding it interesting to have a change and while they don't regret their vote, should a pollster phone to ask what party would get their vote if an election was held today, they go with the change.
For all those reasons Labour would have been expecting the polls to wipe out National's election victory by Christmas. I thought they would, I suspect National thought they would too. But it hasn't happened ...
We're in uncharted political water. Never before in my lifetime have we had a government led by a party with fewer seats than its main rival ... No party in living memory has gone out of government in as strong a position as National is now.

By a striking co-incidence, Heather du Plessis Allan's core contention (Labour has shortest honeymoon ever) also closely echoed the Farrar Narrative:
The first poll since taking office will likely have been gutting for Labour  ... You'd have expected a bigger bump for Labour.You'd have expected voters who didn't like the idea of change to realise change isn't that bad ... But any approval of all that work — and it is a lot of work — hasn't translated into "Yes I'll vote for you next time".
Which is a surprise, because the last two new Governments - John Key's and Helen Clark's - had much bigger jumps in support after their first election wins.
It  feels like Labour's honeymoon is over. In fact, it probably didn't have one. It probably used up its honeymoon during the election campaign. That sucks for Labour. John Key's went on for years ... To turn these poll numbers around, Labour can't afford much of a holiday.
(Farrar set up Heather du-Plessis Allan in a very similar way to the Toby Manhire feedback loop, cheerfully repeating and reinforcing her dutiful regurgitation of his lines in a somewhat later Dec 17  Kiwiblog post)

No real surprise that over at the impeccably neo-liberal National Business Review, Chris Keall similarly relied on the core Farrar message (No Honeymoon Bounce for Coalition):
The Coalition appears to have missed out on the traditional honeymoon bounce
 As did Audrey Young at the Herald:
The disappointment for the Government is that it has not received a bounce in support which Labour did after the 1999 election and which National did after the 2008 election.

By this stage, the MSM echo-chamber had transformed Farrar's little cluster of assertions into a received wisdom largely beyond question.

Henry Cooke at Stuff (No post-election bounce for Labour in first major poll):
Post-election poll reveals Labour has missed out on traditional support bounce ... In 1999 and 2008 the new Labour and National governments both enjoyed substantial polling bounces.
TVNZ Political Editor, Corin Dann (New post-election poll shows Labour has missed out on traditional support bounce):
There has been no big poll bounce for the Labour led Coalition government ... In 1999 and 2008 the incoming governments both experienced strong poll bounces after taking office.
Left-leaning commentator Chris Trotter at both Stuff (Lack of post-election bounce an ominous sign for new Government) and on his own blog Bowalley Road (Bringing down Labour's Warrior Queen):
If Jacinda was anticipating a “post-election bounce” in the polls, then she and her colleagues will find it hard to avoid feeling ever-so-slightly jumpy ... National’s voters clearly remain unconvinced by the new government’s arguments for change. Certainly, this poll has registered nothing like the decisive 10 percentage-point shift in voter allegiance that followed the election of Helen Clark in 1999, and John Key in 2008. Branded by its enemies as a “coalition of the losers”, the Labour-NZ First-Green Government is beset by legitimacy issues entirely absent from previous MMP configurations.

Others in the media endorsing the Farrar line in one way or another included Senior Spinoff Writer Simon Wilson, former National Party Cabinet Minister Wayne Mapp (also at Spinoff), the Scoop team of writers, and the Editor of regional Paper Hawkes Bay Today who, despite adopting a more nuanced approach than many in the MSM, still accepted the core Farrar meme:
So, the coalition as a whole did not receive the normal bounce a new government gets

Meanwhile, on the blogosphere ... Tim Watkin at Pundit, Bomber Bradbury at the Daily Blog and The Veteran at No Minister each, to one extent or another, repeated the received wisdom originally woven by Farrar and subsequently embroided by the MSM (albeit with a slight twist in Bomber's case). And no prizes for guessing that right-wing Whaleoil took to the Farrar thesis like a large aquatic sea mammal to water (although, as a player in the Judith Collins faction, he couldn't quite resist taking a few potshots at English, Joyce and Bennett along the way):
Both Clark and Key experienced an 11 point bounce. This is a bad result for Labour. It also shows that National's support is rock solid.

To be fair, the occasional dissenting voice did make an appearance:

The Standard's indomitable Mickysavage was having none of it:
The Herald appears to be attempting to create a meme, that the honeymoon is over and this Government is in trouble.  Nothing could be further from the truth.  The relief that I have sensed in the community since the change of Government has occurred is palpable.  And the right track wrong track result will inevitably lead to further support to Labour unless it changes dramatically. (See Mickysavage's detailed critique of the Roughan and HDPA analyses and also his Latest Colmar Brunton poll result post)
And both panel members on Nine to Noon's Monday Politics programme, Stephen Mills (from the Left) and Matthew Hooton (from the Right), explicitly challenged the accuracy of the Colmar Brunton results while veteran commentator Colin James raised some doubts in a more implicit manner (I'll briefly touch on the arguments of all three in the Conclusion).

Even Cooke and Roughan were prepared to add a brief caveat that internal polling from both Labour and National was believed to have the two Major Parties much closer ... but, crucially, none of this was allowed to intrude into the main thrust of their analyses or the attendant headlines.

Obviously, Farrar had closely co-ordinated this whole strategic campaign with Bill English's Office. Consequently, a smattering of media reports on the Colmar Brunton simply quoted English enthusiastically pushing the Farrar line:

From the Herald:
English said a new Government would normally get a bigger bounce."I think the Government would have expected that with a lot of favourable publicity from the formation of the Government and the newness of it, that they'd get a bigger bounce in support from the public but I think they have had such a messy start out of not being able to do things they'd do."

From TV3 Newshub:
Bill English told Q & A he expected a "bounce" for Labour but their "messy start" to taking power had hurt them in the public's eyes and this was reflected in the poll result.

Astonishingly, the media embrace of Farrar's honeymoon meme was so pervasive that even PM Ardern herself was forced to implicitly accept it. While emphasising in her Monday morning interviews that her Government "certainly continued to hold a mandate", she implicitly conceded that the Coalition had failed to receive "an automatic bounce simply by being there" and instead would need to earn people's support over time.

Finally, to end this overview of the burgeoning MSM consensus around the Farrar narrative (before I move on to the core Poll analysis designed to test it), notice that both Roughan and du-Plessis Allan have erroneously convinced themselves that the early Dec 2017 Colmar Brunton was the very first poll conducted after the Ardern Coalition took Office.

They weren't alone: Kathryn Ryan (RNZ), Tracy Watkins (Stuff), Audrey Young at the Herald, an Otago Daily Times editorial writer, and the author of an anonymous report in the Herald were all under the same misapprehension.  Why ?

It seems likely they were caught out by Farrar surreptitiously moving the goalposts between his Nov 6 and Dec 10 interventions. In his Nov 6 post, Farrar had placed his entire analytical emphasis on situating the emergence of the respective honeymoons in the very first polls of the Clark and Key eras ... and thus, in turn, anticipating his own upcoming comparative analysis between these initial post-99/08 polls and the very first post-election poll of 2017.  Unfortunately for Farrar, the first three post-2017 election polls - the Roy Morgan, the UMR, (and even his own Curia) - were far too positive for the Ardern Govt. Hence, he was forced to wait it out a few weeks for the more neutral results of the December Colmar Brunton, in the process leaving a number of journos more than a little confused - indeed almost floundering - in his wake.

Farrar's Modus Operandi

David Farrar's authority with the mainstream media chiefly derives, of course, from his impeccable connections with the upper echelons of the National Party. More specifically, though, in regard to the matter at hand, he's also widely considered something of a "polling wizard" (to repeat Toby Manhire's fulsome description). John Key's equally effusive endorsement of the National Party pollster on Election Night 2014 can surely only have cemented that reputation. And further, given his frequent comparative opinion poll posts at Kiwiblog, the media also, no doubt, assume Farrar possesses a near-monopoly on historic polling data that isn't always that easy to track down.

It's not entirely surprising, then, that various journos - strapped for time and bereft of resources - gratefully seize and regurgitate the Spin-Meister's handy little (made-for-the-media) analyses ... with little if any critical thought. A process probably further reinforced by a certain degree of Press Gallery Group-Think, where sweeping assertions and uninformed speculation quickly harden into conventional wisdom.

All of which, unfortunately, is an acute problem because Farrar tends to play fast and loose with reality: he cherry-picks, he obfuscates, he "creates happy mischief". Essentially, he indulges in a creative sleight-of-hand that always - but always - favours and flatters a National Party with which he is, of course, very closely aligned.

After a decade of reading these Kiwiblog poll analyses, if I was to summarise his trademark modus operandi, I would suggest Farrar tends to weave together: (a) one or two little nuggets of unimpeachable fact (to add a sense of plausibility), (b) a degree of outright fabrication, (c) omissions and silence on crucial contextual issues that might otherwise undermine his narrative, (d) what might be called creative ambiguity.

In essence, Farrar employs ambiguity and omission to mislead and manipulate the media - dancing merrily through that precarious space between fact and fiction. He delights in leading time-starved journos up the garden path, leaving very strong (and very false) impressions, while never quite explicitly spelling them out (thus allowing himself just enough wiggle room for plausible deniability). In short: he sets out 2+2 in such a way that visiting journos will almost inevitably conclude the answer is 5.

Call me old fashioned, then, but rather than taking Farrar's sweeping claims at face value, I've decided to put them to the test in the following analysis. Proceeding chronologically through the last seven changes of government, I explore the relevant poll data to determine (1) whether or not incoming governments before 2017 really did enjoy dramatic surges of support in the first post-Election polls (as Farrar and the Media assert) and (2) to what extent the Ardern Govt's poll performance paled by comparison (in the way that, once again, the ubiquitous Kiwiblog host and his MSM acolytes would have us believe).

Last Four Incoming Govts of  20C:  A Huge 11 point Surge ?

While some segments of the media have strictly confined themselves to repeating Farrar's very specific claims about the immediate post-Election Mood in 1999 and 2008, others have subtly expanded the notion of a huge honeymoon surge to the extent that it's almost become an axiom - a kind of instantaneous iron law of modern New Zealand electoral politics.

Several journos, for instance, assert that it's "traditional" for new governments to enjoy a spectacular upswing in immediate post-election support (Chris Keall, NBR / Corin Dann, TVNZ / Henry Cooke, Stuff), others say it's what "normally" happens (John Roughan, Herald / Hawkes Bay Today Editorial / Anonymous Herald report), or "typically" occurs (Toby Manhire, the Spinoff), or is simply "expected" (Heather du Plessis-Allan, Herald), and "uncontroversial" (Tim Watkin, Pundit). Farrar himself encouraged this sort of slippage in a tweet the night Colmar Brunton released their Early Dec 2017 poll: "Great result for National as new Government normally gets 11% poll boost". 

All of which implies an entrenched honeymoon tradition stretching back deep into the late 20th Century.

Here I test this (very sudden) widespread assumption in the media by drawing on a pool of historic opinion poll data to delineate public opinion immediately following the final four Turnover elections of the 20th Century.

Did any of these four Incoming Governments enjoy the sort of spectacular 11 point surge implied by Farrar and key parts of the Media ?

Apparently not.

(1) Kirk Labour Govt

1972 Election         First Poll (Herald-National Research Bureau)

Lab     48.4                51.0           + 2.6
Nat      41.5               39.0            - 2.5
SC        6.7                 5.0             - 1.7
Other   3.4                 5.0            + 1.6


(2) Muldoon National Govt

1975 Election        First Poll (Herald-National Research Bureau)
Nat      47.6               44.0            - 3.6
Lab     39.6               41.0            + 1.4
SC        7.4                 8.0             + 0.6
Values  5.2                 7.0             + 1.8

(3) Lange Labour Govt

1984 Election        First Poll (Heylen-TVNZ Eyewitness News)
Lab      43.0               46.9             + 3.9
Nat       35.9               35.3             - 0.6
NZP     12.3               10.9             - 1.4
SC        7.6                  6.9              - 0.7
Other   1.6                   0               - 1.6 

(4) Bolger National Govt

1990 Election        First Poll (Heylen-TVNZ One Network News)
Nat       47.8                50.0               + 2.2
Lab      35.1                32.0                - 3.1
All        14.3                17.0               + 2.7
Other    2.9                   0                   - 2.9

Clearly, when exposed to the relevant poll data, this newly-minted media meme (subtly encouraged, of course, by Farrar) disintegrates before our very eyes. There are simply no empirical grounds to sustain it.

Far from receiving any sort of euphoric honeymoon bounce, Muldoon's incoming 1975 National Govt  actually lost support in the first post-Election Poll - much of the swing seemingly heading towards Labour and Values on the Left. Meanwhile, the new 1972 Kirk-Labour, 1984 Lange-Labour and 1990 Bolger-National Governments only enjoyed relatively mild bounces of 2.6,  3.9 and 2.2 percentage points respectively. Nowhere near the magnitude we've been led to expect.

What's more, neither the Lange Govt's bounce nor (probably) the Bolger Govt's bounce came principally at the expense of the main Opposition party in the way implied by prevailing MSM mythology.

1999: Incoming Clark Labour Govt:  A Huge 11 point Surge ?

Did the incoming Clark-Labour Government enjoy a decisive 11 point shift in voter allegiance in the first post-Election poll, as Farrar explicitly asserts in his Nov 6 post and as sections of the media have dutifully repeated ?


The first Poll conducted after the 1999 change of government was the late January 2000 National Business Review-Compaq poll (variously described in the media at the time as the first "major" poll and the first "big" poll out since the Election. Compaq had produced 6 nationwide Polls every year for the NBR since 1993). The second and third post-Election polls (TV3-CM Research and One News-Colmar Brunton) were carried out about a month later towards the end of February.

Clark Labour Govt

              1999 Election                    First Poll (National Business Review-Compaq)
Labour                38.7                     45.0                 + 6.3
Alliance               7.7                        6.7                   - 1.0   
L+A Coalition     46.4                     51.7                 + 5.3  

Green                    5.2                      7.1                   + 1.9
Left Bloc              51.6                     58.8                 + 7.2

NZF                      4.3                       3.0                  - 1.3
United                  0.5                         0                   - 0.5

National              30.5                      29.0                 - 1.5
ACT                     7.0                       5.0                   - 2.0
Right Bloc           37.5                     34.0                 - 3.5

Minor Parties      6.0                       4.2                   - 1.8

Yeah, Clark's Labour Party (and the Lab-Alliance Coalition as a whole) certainly enjoyed a more impressive bounce than those experienced by the four previous Incoming Governments, but the fact remains this honeymoon surge was only half the size suggested by Farrar.

2008: Incoming Key National Govt:  A Huge 11 point Surge ?

This is where Farrar's carefully-woven tale really starts to unravel.

Let's remind ourselves of his core claim (Nov 6 post):
In 2008 National won 45% of the vote. The next poll had them at 56%. SO an 11% bump from winning. So both recent changes of Government has (sic) seen an 11% bump for the winning party. This is because people like to give a new government a fair go

And to re-summarise a few MSM Notables:

John Roughan (Herald):
A new government normally gets a good lift in the first post-election poll

 Heather du Plessis-Allan (Herald):
... the last two new governments - John Key's and Helen Clark's - had much bigger jumps in support after their first Election win
Henry Cooke (Stuff):
In 1999 and 2008 the new Labour and National governments both enjoyed substantial polling bounces
Toby Manhire (The Spinoff):
Typically a new government can expect a bump in support after victory ... 1999 and 2008 both saw the incoming government register a lift of 11%
Chris Trotter (Bowalley Road):
 ... the decisive 10 percentage point shift in voter allegiance that followed the election of Helen Clark in 1999 and John Key in 2008
Corin Dann (TVNZ):
In 1999 and 2008 the incoming governments both experienced strong poll bounces after taking office

Right, so did the incoming Key National Government really enjoy - as Farrar and the MSM explicitly assert - a massive 11 point spike in the first post-Election poll ?

Not on your Nelly.

The first Poll (Late Nov 2008 Roy Morgan) was conducted 2-3 weeks after the Election - precisely the point when you'd expect an effusive outpouring of this generosity of spirit (as evoked by Farrar and the MSM) toward the fresh, new, untainted Government.

And yet the Nats in this RM actually found themselves down 0.9 points, the broader Right Bloc were also down (- 0.5) while the Govt as a whole (Right Bloc + Maori) rose by a mere 0.7 points. The Left, meanwhile, were up 0.9 points on their Election result with the broader Opposition bloc rising by 0.3. 

This first post-Election poll, then, very much recorded a No Change / Steady-as-She-Goes result - with the broader Blocs and most of the individual parties essentially flat-lining. This mysterious Spectacular 11 point Poll Bounce of Popular Mythology ? ... Nowhere to be seen.

Key National Govt

              2008 Election                    First Poll (Roy Morgan Late Nov 2008)
National           44.9                       44.0                  - 0.9
Right Bloc        49.5                       49.0                 - 0.5
Govt Bloc         51.8                       52.5                 + 0.7

Labour             34.0                      32.5                  - 1.5
Green                 6.7                        9.5                  + 2.8
Left Bloc          41.6                      42.5                  + 0.9
Oppo Bloc        45.7                      46.0                  + 0.3

Well, OK, but let's cut young Farrar some slack here and assume that he's simply made an "inadvertant error" and really meant the second post-Election Poll. Surely that one must have recorded this huge double-figure surge for the National Government ?


The second poll was also a Roy Morgan - conducted in early-mid December and released about 6 weeks after the 2008 Election. It recorded a very mild bounce to the Key National Govt of around 2 percentage points, with the Left remaining steady and the Opposition down a mere 1 point on their Election result. Pretty similar to the first post-Election polls of 1972 and 1990.

Oh dear. Farrar's now beginning to take on the appearance of a dodgy Used Car Salesman with an equally dodgy toupee, a 5 o'clock shadow and a cheap polyester suit. Much of the media, meanwhile, are looking ... well ... if not a little foolish exactly then at the very least, just a teensy weensy bit too casual, credulous and naive. And yes (in some cases) partisan.

But, hang on, let's maintain our generosity of spirit here and assume that this has all been some sort of "ghastly misunderstanding" ... after all it's plausible, is it not, that Kiwiblog's enfant terrible just became a little confused in all the excitement and he was actually referring to the 3rd, 4th or 5th Polls to be carried out in the wake of the 2008 General Election. I mean, for the love of God, surely to goodness at least one of these must have recorded a decisive 11 point shift in voter allegiance to National ? Surely ???

To take these three polls in order, the answers would have to be:

Nope, Nein and Nyet.

The 3rd, 4th and 5th polls (released 11-15 weeks after Election Day) each recorded relatively mild bounces roughly on a par with the incoming 1984 Lange Government's first Poll result. Key's National rose by between 3.1 and 3.6 percentage points in these 3 Polls, with the Left remaining pretty steady (down 0.1 - 1.1 on its Election result) and the broader Opposition Bloc falling by between 1.7 and 3.2 points.

Once again, no sign of a massive 11 point spike ... and let's remember this is now more than 3 months after the 2008 General Election !

2017: Incoming Ardern Labour Govt:  No Surge ?

Curious, indeed, that having specifically pinpointed the very first post-Election polls of 99 and 08 as the definitive evidence for these putative surges in new government support, the prominent National Party operative should suddenly find himself "otherwise engaged" when the first two Polls - the (Early Nov) Roy Morgan and (Mid Nov) UMR - emerged following Winston's October 19 announcement. His silence remains deafening to this very day.

Farrar may not have gone quite as far as claiming the (Early Dec 2017) Colmar Brunton was the very first post-announcement poll to emerge (after all, he needs to maintain plausible deniability), but as we've seen that's certainly what more than one journalist inferred from his posts.

So, call me a reckless daredevil if you wish, but I think it's probably high time we took a cool, calm, sober look at that first Roy Morgan, that first sampling of public opinion after Winnie's announcement, the Poll that (apparently for both Farrar and the MSM) Dare Not Speak its Name.

Ardern Labour Govt 

                                       2017 Election      First Poll  (Roy Morgan  Early Nov 2017)

Labour                                   36.9                    39.5                  + 2.6
Green                                      6.3                     10.0                  + 3.7
L+G                                       43.2                    49.5                  + 6.3
L+G+NZF                             50.4                    54.5                  + 4.1  

National                                44.4                     40.5                  - 3.9
Nat+ACT                              44.9                     41.0                  - 3.9

Cast your mind back to the six previous post-Turnover polls (1972-2008) and these figures are already starting to look impressive for the new Ardern Government. The Left rising by a pretty hefty 6.3 points, the Government Bloc as a whole up by 4 and both the Major Opposition Party and broader Opposition Bloc correspondingly down by 4.

To see just how impressive, let's systematically compare these Roy Morgan results with the first post-Election Polls following the previous changes of government:

Swing in First Post-Election poll: 1972-2017 (Comparisons)

                                      1972        1975        1984        1990        1999        2008        2017

Major Govt Party          + 2.6       - 3.6         + 3.9       + 2.2        + 6.3        - 0.9       + 2.6

Govt Bloc                     + 2.6       - 3.6         + 3.9        + 2.2       + 5.3        + 0.7       + 4.1

Major Oppo Party         - 2.5       + 1.4         - 0.6         - 3.1       - 1.5         - 1.5        - 3.9

Oppo Bloc                    - 2.5       + 1.4         - 1.3          - 0.4       -5.3         + 0.3       - 3.9

First, in terms of the Major Party of the incoming government, Ardern's Labour (+ 2.6) enjoyed a mildly above-average surge. Both Clark's Labour (+ 6.3) and Lange's Labour (+ 3.9) had enjoyed bigger post-election boosts, but all three incoming National Parties (Muldoon - 3.6 / Key - 0.9 / Bolger + 2.2) recorded poorer performances, while Kirk's Labour experienced precisely the same spike as Ardern's.

Under MMP, of course, polling numbers for the Government Bloc as a whole assume more intrinsic importance. And it's here that the Ardern Coalition begins to come into its own - the new Govt's + 4.1 boost was second only to the Clark Coalition's + 5.3, mildly up on the Lange Govt's + 3.9 and well ahead of the initial poll performances of the other four incoming governments.

And yet probably even more impressive for the new Ardern Govt was its unrivalled ability to carve into the support-base of its Major Party competitor. National's 3.9 point fall represented a far worse performance than that suffered by almost every other Major Opposition Party in an initial post-Turnover-Election Poll. Only Moore's Labour Party in 1990 (down 3.1) fared anywhere near as poorly. And crucially, even then much of the swing away from Moore's Labour appears to have headed toward the Alliance on its Left rather than across to the new Bolger National Govt. In stark contrast, not only did the 2017 English National Party suffer a heavier loss than all of its Major Opposition Party predecessors but these former National voters were far more inclined to head directly across the political divide to the Govt Bloc. A very clear 4 point swing from Right to Left took place in the immediate wake of the 2017 Election.

The story's much the same when comparing initial poll performances of the Opposition Blocs as a whole - although, once again, the post-1999 poll begins to standout as not only unusually good for the Clark Govt (+ 5.3) but also unusually poor for the broad Opposition Bloc (- 5.3) (though not for the Shipley National Party itself (- 1.5)). A straight 1999-2017 Opposition Bloc comparison, however, is just a little bit deceptive because the former involved a much greater variety of Parties (including NZF) than the latter.  If we narrow the 1999-2017 comparison to just the National-ACT Right Bloc core of the Opposition, then the Right's fall in 2017 (down 3.9) remains a little worse than their loss in 1999 (down 3.5).

Swing in First Post-Election poll: 1972-2017 (In Order)

Major Govt Party     Govt Bloc     Major Oppo Party     Oppo Bloc      Right Bloc        Left Bloc

   1999   + 6.3          1999   + 5.3       2017   - 3.9          1999    -5.3        2017  - 3.9       1999  + 7.2
   1984   + 3.9          2017   + 4.1       1990   - 3.1          2017   - 3.9        1975  - 3.6       2017  + 6.3
   2017   + 2.6          1984   + 3.9       1972   - 2.5          1972   - 2.5        1999   - 3.5      1984  + 3.9
   1972   + 2.6          1972   + 2.6       1999   - 1.5          1984   - 1.3        1972   - 2.5      1972  + 2.6
   1990   + 2.2          1990    + 2.2      2008   - 1.5          1990   - 0.4        1984   - 0.6      1975  + 1.4
   2008   - 0.9           2008    + 0.7      1984   - 0.6          2008   + 0.3       2008   - 0.5      2008  + 0.9
   1975   - 3.6           1975     - 3.6      1975   + 1.4         1975   + 1.4       1990    + 2.2    1990   - 0.4

Overall, then, the new Ardern Government's performance in its first post-Election poll was actually highly impressive. Of the seven incoming governments, Ardern's enjoyed the 2nd highest Govt Bloc and Left Bloc bounce and the 3rd highest Major Party of Govt bounce. Meanwhile National and the Opposition / Right Bloc suffered something of a shocker - Bill English's National Party recording the steepest plunge for a Major Opposition Party since polling began. Precisely the same can be said of the broader Nat-ACT Right Bloc.

In other words, if the MSM had bothered to closely scrutinise Farrar's central claims (rather than simply taking him at his word), then the headlines generated should have been the complete inverse, the absolute antithesis, the polar opposite of No real bounce for Labour (Kiwiblog), No post-election bounce for Labour (Stuff), No honeymoon bounce for Coalition (NBR) or New Post-Election Poll shows Labour has missed out on Traditional Poll Bounce (TVNZ).

As a brief aside, let me highlight just how comparatively popular the Ardern Coalition was - not in terms of the bounce itself - but of the overall percentage support in this first post-election poll. After all, the clear subtext underlying the claims made by Farrar and the media is that, in this immediate post-election period, the Ardern Govt has been unusually unpopular, fragile, precarious, indeed almost illegitimate.

Overall Govt Bloc vs Oppo Bloc % in First Post-Election poll: 1972-2017
     Govt Bloc          Oppo Bloc              Lead

2017       54.5                41.0                   + 13.5
2008       52.5                46.0                    + 6.5
1999       51.7                37.0                    + 14.7
1972       51.0                39.0                    + 12.0
1990       50.0                49.0                     + 1.0
1984       46.9                42.2                     + 4.7
1975       44.0                41.0                     + 3.0

As you can see, taking  a 54.5% slice of party support in the Early Nov Roy Morgan, the Ardern Govt was more popular than any previous new government since regular polling began in the late 1960s. Furthermore, their lead over the National-ACT Opposition Bloc (+ 13.5) was second only to the incoming Clark Govt's + 14.7, was slightly ahead of Kirk Labour's + 12.0 and was way ahead of the other four incoming governments in their first post-election polls. In other words, about as legit a new Govt as you're ever likely to see.
 Right, so back to the bounce ...

A week or so after this first Roy Morgan emerged, the second post-Election poll (a UMR) essentially confirmed its findings

Ardern Labour Govt 

                         2017 Election              First Poll                        Second Poll 
                                                            (Roy Morgan)                       (UMR)

Labour                36.9                            39.5     + 2.6                   39.2     + 2.3
Green                    6.3                            10.0    + 3.7                    8.9      + 2.6
L+G                     43.2                           49.5    + 6.3                    48.1     + 4.9
L+G+NZF           50.4                           54.5     + 4.1                    54.6     + 4.2

National               44.4                           40.5     - 3.9                     40.7     - 3.7
Nat+ACT             44.9                           41.0     - 3.9                       No Data

Although I don't have the UMR stats specifically for the ACT Party (and therefore can't provide a Nat-ACT Right Bloc result for UMR) I think you can be pretty sure the Right Bloc fell by roughly the same percentage points as National (somewhere in the minus 3.5 to 4.0 neck of the woods). Thus virtually echoing the Roy Morgan in terms of the all-important Govt vs Oppo result.

Make no mistake, these were two very good polls for the Ardern Govt over successive weeks.

(And then, of course, there's the little matter of Farrar's own Curia Internal polling for National. According to Tracy Watkins, Curia's findings were little different from Roy Morgan and UMR in placing the Nats just marginally ahead of Labour)

But no, despite having placed his entire analytical emphasis on the very first polls of the Clark and Key eras (Nov 6 post), Farrar decides the Roy Morgan, the UMR (and it seems his own Curia) are far too positive for Ardern, the Government and the Left and thus decides to wait it out, hoping like hell for some rather less impressive results that he might be able to weaponise to National's advantage.

Finally, 11 weeks after the 2017 Election and 7 weeks after Winston's Oct 19 announcement, Farrar seizes on the newly-released Colmar Brunton which presents a rather more neutral, somewhat less positive picture.   Even then, though, it has to be said that despite Farrar diverting the media's attention towards the perceived failure of the Ardern Coalition to achieve his (as we've seen, entirely mythical) 11 point spike, the Colmar Brunton was in fact quite positive for Labour and the Left - the latter now matching National and Opposition Bloc support. Overall, though, we're essentially talking a No Change Poll result.

                                     2017 Election      Third Poll  (One News-Colmar Brunton Early Dec 2017)

Labour                                   36.9                    39.0                  + 2.1
Green                                      6.3                     7.0                    + 0.7
L+G                                       43.2                    46.0                  + 2.8
L+G+NZF                             50.4                    51.0                  + 0. 6 

National                                44.4                     46.0                  + 1.6
Nat+ACT                              44.9                     46.1                  + 1.2


Prominent National Party operative David Farrar has very successfully managed to sell the MSM a bogus honeymoon meme. This, in turn, has generated a whole series of negative headlines for the Ardern Coalition ... reinforcing, in the process, some of National's key attack lines around the alleged fragility and illegitimacy of the new Government.

It'd probably be going a little too far, I think, to suggest that a Machiavellian Farrar brought to bear all the innumerable dark arts of messaging, comms, social psychology and public relations when devising his various rhetorical strategies. That would be crediting his two Kiwiblog posts with a degree of sophistication that they don't, quite frankly, possess. But in his own relatively crude way, he was able to successfully weave a dodgy little tale of woe for the Govt using his trademark blend of fact and fiction, as always playing on the ambiguity that lies between.

The nub of Farrar's Honeymoon Scam is this: Both explicitly (Nov 6) and implicitly (Dec 10), Farrar left visiting journalists with the distinct impression that the two previous incoming governments - 1999 Clark Labour and 2008 Key National - had enjoyed massive double figure spikes of support in the very first post-Election Poll. At a bare minimum, journalists went away from Kiwiblog with the impression that these honeymoon surges emerged in the immediate wake of these elections - that is, the first few weeks.

Yet, as we've seen, Farrar's claims were essentially fraudulent.

For 1999, we actually have no idea of the immediate post-election mood because that first (NBR-Compaq) poll wasn't carried out until fully 2 months after the Election. What we do know is that in this earliest sampling of the post-99 mood, the bounce for Labour (and for the Labour-Alliance Coalition) was roughly only half the magnitude claimed by Farrar.

For 2008, on the other hand, we have a very precise picture of public opinion immediately following the Election. And it bears no resemblance whatsoever to Farrar's sweeping claims. Far from a spectacular surge in support for the new Key National Government, the first post-Election sampling (just a fortnight or so out) suggested voter sentiment hadn't shifted at all ... if anything, National and the Right Bloc were very slightly down. And subsequently, the 2nd, 3rd, 4th and 5th post-Election polls (released 6-15 weeks after Election Day) recorded only a very modest bounce for the Key Govt. Three months had elapsed since the 08 Election and the very best poll National could point to only had them spiking by less than a third of the 11 points claimed by Farrar.

Having set up and seeded these fraudulent claims about the post-99 and 08 mood (and despite having placed his entire emphasis on anticipating a comparative analysis of the very first post-election polls), Farrar then goes out of his way to avoid the first post-2017 Election polls (while, of course, keeping stum about his own Internals for National) because they're just a little too positive for Labour and - even more so - for the Ardern Government as a whole. Instead, he twiddles his thumbs until the more neutral Colmar Brunton result emerges ... and Voila ! ... the Ardern Coalition Fails to get Traditional Massive Poll Bounce genre is born. And for a few weeks, it's a headline juggernaut that takes on a life of its own.

Does that mean Farrar entirely concocted the idea of these 11 point surges out of whole cloth ?  Not at all - he's both a little too clever and a little too cautious for that. Plausible deniability is everything.

Instead, Farrar's particular sleight-of-hand was to mislead journalists by conveying the clear impression that he was citing the very first post-election poll - regardless of polling company (ie "the next poll had them at 50%" (1999)/ "the next poll had them at 56%" (2008)) whereas, behind the scenes, what he ultimately decided to do was employ figures specifically from the first Colmar Brunton polls ... and that's an entirely different kettle of fish.

These Colmar Bruntons did indeed record significant double figure surges for the Clark and Key Governments but - and here's the crucial point - they were conducted not weeks but in fact a matter of months after the 1999 and 2008 Elections. (The post-2008 Colmar Brunton that Farrar cherry-picks, for example, was the seventh !!! post-Election poll to emerge and was released more than three and a half months after the election). They cannot even remotely be compared to polling carried out in the immediate aftermath of Winston's late October 2017 announcement of a change of government.

One might add, incidentally, that - quite apart from these spikes occurring far later than Farrar led journalists to believe - the 2000 and 2009 Colmar Bruntons that he relies on almost certainly possessed an historic bias toward National and the Right. Back during the Clark Labour Government, for instance, Rob Salmond and Keith Rankin presented persuasive evidence that the CB fairly consistently overstated support for National and the political Right by somewhere around 5%.

And then head back to the late February to late July 2009 Colmar Bruntons that Farrar cherry-picks and you'll find these also pretty much bordered on National-friendly outlier territory. Throughout that period, for instance, the Nats were averaging 4 points higher in the CBs than in the Roy Morgans. And this took place in a more general context where National's polling over the entire first term of the Key Govt is widely deemed to have been grossly overstated (indeed, there's evidence that each of the polling companies altered their sampling / weighting to fix this problem immediately following the 2011 Election result).

Finally, let me anticipate a possible riposte from some of our more partisan Tory chums.

It might go something like this:
OK, we may well accept that the Ardern Govt enjoyed an impressive surge immediately after voters became aware of a change of Govt on Oct 19. But surely the significance of the Colmar Brunton is that this was a remarkably short-lived honeymoon. To the extent that National is up in the CB and the Left and broader Govt Blocs down (relative to the RM and UMR - though still up on the Election result), it's clearly the End of the Golden Weather for the Ardern Coalition and that, indeed, does not augur well for their long-term prospects.

 Well you know, that's an entirely plausible scenario. But there are other equally plausible scenarios. 

The Colmar Brunton, for instance, may well prove to be something of a National-friendly outlier.  Certainly neither Stephen Mills nor Matthew Hooton have put much store in its findings:

Stephen Mills (from the Left), Nine to Noon Politics, RNZ (11 Dec 2017):
I would take this poll (Colmar Brunton) more seriously if it was from Reid Research - who had an excellent campaign. But given the completely implausible backflip that Colmar Brunton did (during the election campaign) ... I think this poll's wrong on the National vote - it's too high - and I think it's wrong on the margins between the two parties which is now very close. It's inconsistent with the Roy Morgan poll, it's inconsistent with the UMR poll, it's inconsistent with Tracy Watkins' column which said that National's internal polling had National just marginally ahead. It's inconsistent with the surging Right-track / Wrong-track numbers ... And also on Jacinda Ardern's leap (in the Preferred PM ratings) ... I think it's also kind of out of line with events ... And I think if Tracy Watkins is right about National's polling and Bill English has seen that polling (then) it's pretty disingenuous of him to attribute "National's good numbers" to the messy start from the Government ... (this Colmar Brunton) does set a narrative that National is doing really well when I think in fact there has been a shift between the two major parties (in Labour's favour) and they're now almost deadlocked. 
 Matthew Hooton (from the Right), Nine to Noon Politics, RNZ (18 Dec 2017):
I think the Government is going to have momentum and ... I don't think for a second National's at 46% (as Colmar Brunton would suggest). So I think the idea it (National) can just hope the Government fails and that Winston Peters will be involved in some scandal and all will be well is wrong.

Much like Stephen Mills, veteran political columnist Colin James places a particular emphasis on both:

Ardern's leap in the Preferred PM ratings:
Prime Minister Jacinda Ardern had a clear lead over National leader Bill English in all three polls as preferred prime minister or in positive-v-negative ratings.

and on the surging Right track / Wrong track numbers in all of the post-2017 Election polls:
Another measure suggests that voters are comfortable with the (Ardern Coalition) arrangement. All three polls ask if voters think the country is heading in the right direction or on the wrong track. Morgan recorded 66.5% positive to 20% negative, UMR 65%-17% and 1 News Colmar Brunton 51%-26%. These are substantially higher than before the election.
Roy Morgan recorded an 8 point surge in Right direction responses (up from 58% to 66%) in the wake of the Ardern Govt's formation (with the Roy Morgan Government Confidence Rating jumping a "substantial" 15.5 points).

UMR's figures, meanwhile, suggest:
... historically high levels of voters believe the country is now on the ‘right track’.  UMR has run this right track-wrong track poll since the mid-1980s and the current 17 percent ‘wrong track’ reading is the lowest ever recorded. Whatever its critics may say, the Ardern Government appears to be supported by a large well of public good will.

More broadly, Colin James implicitly questions (as, indeed, did Mickysavage) any monocausal reliance on just a single data point.  By averaging the three post-Election polls (RM/UMR/CB), he finds that the Ardern Government's majority over National has grown from 6 to 11 percentage points since the Election ... and concludes that:
Voters have broadly endorsed the new government

All of which raises the more momentous question: How did previous incoming governments fare in public opinion polling over their entire first terms. Did their immediate post-election ratings represent a high-water mark and was it all downhill from then on ? (as some in the media and blogosphere have already claimed) or did their ratings essentially flatline over the long-term ? continue rising ? fluctuate wildly ? or did these new governments lose and then subsequently recover support ?

That's something I'll be exploring in a future post. But why do I have a nagging feeling that Farrar's already planning to spin a few dodgy little Blue-friendly sorties in this regard ? And of course where Farrar goes ...  the highly-distinguished doyens of Her Majesty's Mainstream Media are sure to follow.

(NOTE: I'll add a few Footnotes when I have time)