Poll Roundup: Albanese Honeymoon Phase Polling

Enough polls have now appeared for the new Albanese Labor government that I think it is worthwhile summarising where things are at.  The limited polling data available is all over the place but on average points to a substantial honeymoon boost in favour of the new Albanese Labor government.  

While we haven't seen this kind of honeymoon polling for a while, it has been historically normal for governments elected from opposition.  A phase with a new government averaging above 54-46 occurred for six of the previous seven new governments, the exception being the Abbott government which never got above about 53% on aggregation in its early months and hence never matched the 53.5% 2PP it had been elected with.  (This indicates the extent to which the election of the Abbott government was a repudiation of Labor's Rudd/Gillard/Rudd shambles rather than a positive endorsement of the alternative.)  

Lopsided honeymoon polls get a lot of attention but whether we are still seeing them in twelve months, six months or even three months remains to be seen.  The longest runs of mostly 54+ 2PP polling were for the Rudd government elected in 2007 (just over two years), the Howard government elected in 1996 (14 months) and the Hawke government elected in 1983 (most of the next two years though with a brief spell below after about five months).  The Menzies government's post-1949 phase lasted about eight months while the Whitlam 1972 and Fraser 1975 victories were good for about four months of basking on laurels.  (The difference is that Fraser's government then mostly kept the polling lead until a year and a half in, while Whitlam's soon lost the lead altogether.)  

In contrast, governments that are re-elected don't get long honeymoons of any size and usually don't get noticeable honeymoons at all.  A few months of decisive government leads occurred after Holt's win in 1966 and Fraser's in 1977 (but those were lopsided wins to begin with), after Keating's surprise win in 1993 and after Howard's come from behind victories in 2001 and 2004.  

At the moment though, the new government appears (though there is no specific polling on the causes of its lead) to be benefiting from hitting the ground pretty smartly and with some intent to Do Things, and also from the fact that being a new government is fun.  It's fun because nobody blames you for anything, but you also get to use the machinery of government to discover amazing facts about your predecessor.  And here we haven't just had the usual story about how the previous mob were secretly trashing the nation's finances, but rather something more humiliating in the form of the bizarre Scott Morrison "secret ministries" saga which is consuming what little oxygen the new opposition might have otherwise somehow found.

Is honeymoon polling predictive?

In a word, no.  For instance here is how the first Newspoll 2PP (estimated where necessary) of each government's term since 1987 compared with the result at the end of the term.  New governments are shown in green:

Even though the three new governments polled more strongly than the rest, absolutely nothing to see here.  However these polls are not strictly like for like as since 2013 Newspoll has tended to take a couple of months off before returning to the field after each federal election.  Here's the same graph using the state of government Newspolls at around 2.5 months out from the election:

Again nothing to see here (this is not even remotely significant for such a tiny sample.)

However, if this honeymoon can last a few more months it could be significant in the Victorian election.  Federal drag theory suggests that the eight-year old Andrews government could expect a serious whacking and could possibly even lose without a Coalition federal government to rail against, but federal drag effects are muted if a same-party federal government is popular, and might even disappear completely if the federal government is polling in the high 50s 2PP at the time.  If the feds can keep themselves out of trouble til November, the historic chance of the opposition winning drops sharply. 

Good Polls Are Easy Under Albanese

We've so far seen the following voting intention polls:

* A Newspoll released at the start of August had Labor ahead 56-44 off primaries of  ALP 37 Coalition 33 Greens 12 One Nation 6 UAP 2 IND/others 10.  This was the highest combined vote for non-major parties in any Newspoll ever, but is scarcely surprising following the record non-major vote at the election.  I get 55.7% as the expected average 2PP off these primaries so 56-44 is the expected rounding.  

* A freshly released Resolve poll with Labor on 42 Coalition 28 Greens 12 One Nation 5 UAP 2 IND 8 others 3.  It is not clear whether the high Independent vote is a honeymoon for the teals, a result of Coalition voter disenchantment following the loss, or whether Resolve has reverted to having Independent on the readout everywhere.  If the latter, it is information the SMH/Age or the pollster should be publishing.  No 2PP was published but I convert the primaries to a massive 61.3-38.7 lead to Labor.  (In the old, bouncier Newspoll, Rudd's honeymoon phase included a 63 and a 61, but I don't think I have ever seen a federal poll with a Coalition primary as low as 28.)  Resolve has some history of volatility so far so this could be an outlier, but outliers are detected through the rear-view mirror.

* Morgan, as I am fond of saying on Twitter, will never not be a mess.  I could even make a hashtag for it, #MWNNBAM.  I have so far located seven apparently weekly federal Morgan readings covering early to mid June and mid July to mid August, of which one consists of a full poll report release, one of a full release of primary votes in a video, two can be inferred by changes from the previous two, two are 2PPs only released in a video, and one is a graph dot.  In sequence, the released Labor 2PPs are remarkably steady: 54-54-53-53.5-54-52.5-53.  

The two detailed early and mid June results are covered by a note in the mid June report that suggests the 2PPs were based on 2019 preferences, but I couldn't replicate this in either case (I got 52.2 for both samples, which Morgan had as a plausible (given rounding) 53 and a definitely not plausible 54).  The preferencing basis of the later numbers is unclear.  By 2022 preferences I got 53.2 for the inferred early June sample (Morgan 54), 53.1 for the mid-June sample (53), 54.6 for the inferred late July sample (54) and 53.4 for the early August sample (52.5).  

Anyway these are the primary votes I've found:

June c. 6-12 Labor 34 L-NP 37 Green 12.5 Ind 7.5 ON 3.5 UAP 1 other 4.5 

June 13-19 Labor 36 L-NP 37 Green 11 Ind 4.5 ON 4 UAP 0.5  other 7

July 24-31 Labor 36 L-NP 36 Green 12.5 IND 7 ON 3.5 other including UAP 5

Aug 1-7 Labor 37 L-NP 38.5 Green 11.5 IND 6 ON 3.5 other including UAP 3.5

Given that, in a very brief interval, Morgan has had the Coalition primary 10.5 points higher than Resolve, it seems likely that one of Morgan and Resolve if not both has been a long way out!  Anyway, none of Morgan's readings so far have been considered worthy of induction into their historic tables or the site's header bar.  

* Essential has not yet released any voting intention numbers, making it impossible to benchmark its leadership ratings.

* A Dynata tracking poll on 14 June had primary vote numbers (after redistributing undecided) of Labor 36.1%, Coalition 32.8, Greens 12.9, One Nation 4.5, UAP 4.3, IND/other 10.5.  I estimate the 2PP at 55.6% to Labor.  However Dynata is yet to be accurate at any election I have watched it at, and I have concerns about its methods given that it has claimed to be nationally representative and properly weighted while publishing a 2022 exit poll that had Labor on 41%, 8.4 points higher than they actually got in the election being exit polled.  

Leaderships

This is also a familiar honeymoon polling trope (and again the exception is Tony Abbott): new federal governments often see huge approval ratings for the new Prime Minister, underwhelming positive approvals for little-known opposition leaders, and outlandish Better/Preferred Prime Minister leads.  Better Prime Minister polls skew massively to incumbents anyway, but especially so when they are up against relatively unknown opponents.  The most dramatic case of this effect occurred when Brendan Nelson (a somewhat surprising pick with little authority and up against a stratospherically popular Kevin Rudd) polled Better Prime Minister figures starting with 11%, 9% and 7% (the latter against Rudd's 73%).  This was considered to say something about Nelson himself, but it mainly just reflected that he led a Coalition that was getting smashed on voting intention by a political rockstar opponent.  

Anthony Albanese's first Newspoll satisfaction rating as Prime Minister was a net +35 (61-26).  The 61 was hailed as the highest satisfaction in a first post-election Newspoll, which it was, but this is partly a result of Newspoll methods changes that have reduced the don't know rate on approval ratings, and partly also a result of the poll taking 2 months and 10 days from election day to final sample.  Rudd had started with a net +48 (59-11 with 30 rather than 13 undecided), and 2 months and 23 days after winning he registered net +55 (68-13).  John Howard in 1996 recorded net +53 (67-14) after 2 months and 10 days, though that was his fifth Newspoll and the rating included a disaster bounce for the Port Arthur massacre.  Overall it's clearly not correct to say Albanese is the most popular new PM ever (especially not if one goes back to pre-Newspoll days and Hawke) but he is far from doing badly either.

Peter Dutton started out with a net -4 (37-41).  That may sound OK but it is in fact the worst debut and the first net negative debut for a first-time Opposition Leader in Newspoll history.  The net rating debuts for previous Opposition Leaders were Hewson +18, Downer +19, Beazley +21, Crean +5, Latham +15, Rudd +31, Nelson +17, Turnbull +25, Abbott +5, Shorten +8, Albanese +3.  However, many of these had their first rating taken almost as soon as they entered the job, not two and a half months afterwards.  Only Downer (after an excellent start) hit net negatives about as fast as Dutton has, though Nelson took not much longer.  The bad news for Dutton is that Opposition Leaders frequently record career-high net satisfaction ratings, or nearly so, in their first few months in office, before politically opposed voters have started to dislike them.  On the other hand, it may be that Dutton comes to office sufficiently disliked by voters for the left parties to begin with that his net rating is more resilient to the usual rapid increase in disapprovals.

Albanese held a 59-25 lead as Better Prime Minister but this isn't exceptional given a 56-44 2PP and the example of Nelson shows that Dutton could be doing a lot worse there.  (I suspect that Dutton has more respect from voters who are distinctly right-wing than a more wishy-washy leader would have.)  A similar theme carried on to the Resolve PPM figure which showed Albanese leading Dutton 55-17 - this is from a poll which had among the largest government polling leads ever seen.  So long as he's not responsible for the voting intention being like that then Dutton is doing OK to be on 17% given that and not, say, 10.

Resolve had similar leader approval ratings to Newspoll: Albanese net +39 (61-22) and Dutton net -7 (30-37).

The first few Essential readings for Albanese had his honeymoon apparently deflating over time: net +41, net +32, net +27.  This hasn't been matched by the only other series with more than one data point, the insufficiently transparent Morning Consult leader tracker, which has had Albanese more or less stable at his current net +29 for months excepting a brief kick up to +36 in mid-July.  Essential has not yet polled an approval rating for Peter Dutton or a better Prime Minister reading.

Essential has, however, done something that created some confusion: today it released a leadership favourability item that is polled differently to its standard approval questions.  Respondents are asked to grade "how you feel about each of the following people" on a scale of 0 to 10.  Essential then grades responses from 0-3 as negative, 7-10 as positive and 4-6 as neutral.  Absent of any evidence of calibration (eg also asking the same respondents the normal leader approval question to see how many of those giving a leader a given score would also approve of their performance) I would think that 4/10 is actually a negative response while 6 might or might not be a positive.  

Some people on social media tried to compare Albanese's 43% "positive" favourability with his 61% in Newspoll and hence conclude his popularity was crashing, but even without the methods difference this would still not work - the Essential rating has 34% of respondents who were neutral, don't know, unsure who Anthony Albanese is or lost in the rounding, while Newspoll had only 13% with no view either way.

Anyway the favourability ratings as dubiously defined were (in the form net rating followed by positive then negative):

Albanese +20 (43-23) 

Dutton -8 (26-34)

Littleproud -6 (21-27)

Bandt -14 (23-37)

Lambie 0 (27-27)

Hanson -26 (22-48)

Adam Bandt's low rating here is mildly surprising given his high profile and his party's strong election performance (similarly weak net ratings for Greens leaders in the past have come for lower-profile leaders early in their careers).  Bandt does have a polarising approach and personality and it's hard to say how voters "feel about him" would relate to what they thought of his performance.  

Unsurprisingly with the honeymoon voting intention polling has come similarly rosy assessments (mostly) of the new government's performance, with Resolve's respondents rating it well ahead of the Coalition as best equipped to handle the economy.  A few less flattering assessments can be found from Essential's panellists now and then but when 58% of them jump at red meat like "Australia’s economic system is broken and the government needs to make fundamental changes to sort it out" it's hard to know what to make of that.  

At this stage there are not enough non-Morgans for a voting intention aggregate to be especially useful but I do intend to run one in this term and probably start it by the end of this year.  

I will update this article with any more federal polls that appear in the next week or so that I feel inclined to comment on, and then there will be a new roundup sometime down the track when I have something different to say, or more of the same.