As Gillard recovered, so can Abbott?

Advance summary:

1. The view has been advanced that since Julia Gillard has become much less unpopular, then so can Tony Abbott.

2. This view does not automatically follow, because Gillard and Abbott occupy different positions and became unpopular for mostly different reasons.

3. The record since the late 1960s shows that the average recovery in net satisfaction for both Prime Ministers and Opposition Leaders between their worst rating and the next election is about 20 points.

4. However, very unpopular Opposition Leaders are usually removed, and those who make it to elections have so far lost them even if they had ceased to be unpopular by election time.

5. There is one case of an Opposition Leader recovering from a rating worse than Abbott's to a positive net rating within two months, but without winning the election.  (Andrew Peacock, 1984)

6. However, the general historical trend is against the likelihood of Tony Abbott recovering to a positive or even near-zero rating before the next election.



--------------------------------------------------------------------------------------------------------------
Having apologised a few days ago for writing so many articles exploring Tony Abbott's unpopularity, I hereby disapologise (but I do have three pieces on Abbott-free subjects in the works, and there's probably a Tasmanian EMRS poll coming up, so the Abbott focus will decline eventually).  On TT I used to average about one or two articles a month, but for whatever reason I'm finding that here I am able to write much more often.  Possibly this is because I no longer spend as much time arguing with people on Tasmanian Times!  Anyway, if the average so far of an article every few days is maintained (and we'll see about that, since I'll only write when I pretend I have time, and have something I consider worth saying to say) it stands to reason that the same subject will sometimes inspire a series of articles.
-----------------------------------------------------------------------------------------------------------------

The genesis of today's post is a simple comment by Bernard Keane, one of the Crikey house pundits, yesterday:

"Abbott can turn his bad numbers around. Gillard turned around even worse numbers. But he has less time to do so."

So, the question is, do Opposition Leaders turn around bad numbers as easily as Prime Ministers?  I expected that the answer would be no, because PMs and Opposition Leaders become unpopular for different reasons.  PMs often become unpopular because implementation of government policies involving short-term pain for long-term gain annoys voters, as do mistakes that have actual consequences.  Opposition Leaders may now and then cop flak for stupid policies or bad responses to government policy, but by and large what they do affects voters less, and I'd expect that when they become unpopular it's more often because of their personality or their ineffectiveness.

And indeed, the history of the Newspoll era (see bottom graph at Mark the Graph for an example) shows that since 1985, once Opposition Leaders have become very unpopular, they tend to flop about on the floor for a while, and bounce back a small to moderate degree, but sooner or later they get removed and replaced.  In the Newspoll era, every Opposition Leader who has polled a netsat of -16 or worse, has then stayed negative until their tenure ended.

In a long piece on TT I mentioned also that the example of Fraser (recovering from -24 to win) is a rather misleading case, because he was only unpopular for a few weeks during a crisis.  But looking back at the full history of Morgan Gallup data from before Newspoll, I've found a rather startling example of Opposition Leader recovery.  Although the leader in question did not win, and was in fact removed from office inside a year like all the others who have been so unpopular, his fate provides an instructive example that sometimes Opposition Leaders can bounce back big and quickly from extremely awful ratings.

But usually, they don't.

Unpopular Prime Ministerships

The following table breaks down cases since 1968 in which a Prime Minister has become unpopular (even only fairly mildly or briefly) into the parliaments in which that has occurred.  The cutoff is a netsat of -10.  Thus, there are multiple listings for PMs who were unpopular in different parliaments.  Also, the year in brackets after the PM's name is the year that parliament ran for (they were not always in office for all of it).  Although Fraser was formally PM during the 1975 campaign I have treated him as if he was Opposition Leader and as if Whitlam was still PM, for the purpose of this exercise.  Hence the asterisk.

The "Trough" column gives the worst netsat that PM recorded in that Parliament (I use Morgan Gallup until Newspoll existed, and Newspoll thereafter.)  "Peak" is the highest netsat that PM recorded in that Parliament that came after the trough.  "Election" is their netsat in the last poll before the election.  "P-T" is the difference between peak and trough.  "E-T" is the difference between election and trough.  "Result" I trust is self-explanatory.  

 

The two PMs who were removed while unpopular (Gorton was removed while not unpopular) were both removed at their lowest point, so there is no peak after the trough.  It can be seen that Gillard's recovery is already around the average recovery for PMs who made it to the subsequent election; it may or may not yet increase above that level.  It can also be seen that where an unpopular PM recovers, they will usually fall backwards several points from the peak of that recovery by the time of the next election.  

The most spectacular recovery is of course that of John Howard, who polled -36 (Abbott's current rating!) in March 2001, but gained 20 netsat points in a single poll from his handling of the MV Tampa issue, and another 18 in a single poll after S11.  Not far behind him comes Keating during his losing 1993-6 tenure, but not everybody gets to face off against 1994-model Hewson or Alexander Downer, and there's no doubt Keating's recovery from the horrors of -57 was assisted by those two dud opponents.

Unpopular Opposition Leaderships

Here's the same again for Opposition Leaders:

The top result is the standout and the one I referred to earlier.  When Bob Hawke called the seemingly endless 1984 campaign, the once reasonably popular Liberal leader's ratings had been sliding and in late September he too was in what we now know as Abbottland (with apologies to Edwin Abbott Abbott.)  Morgan Gallup's week by week polling then shows that in the first two weeks of October, Peacock recorded two consecutive -47s (20-67 then 19-66).  However, as the campaign continued, Peacock rebounded rapidly.  He went into the sole televised debate at -13, defeated Hawke in that debate, and came out the other side at +14, a 61-point rebound in just seven weeks. (It's possible that the frequency of polling by Gallup Morgan, and their sample size for those weekly samples, exaggerated the size of the bounce, but it still would have been a whopper under any pollster.)  Peacock then still had a positive netsat when he quit facing party disunity, and was replaced by Howard the next year. 

It's easy to see that this remarkable recovery could have convinced the Coalition that Peacock was a very "usable" replacement for the struggling Howard going into the 1990 election.  The change to Peacock did improve the Coalition's vote share slightly but for whatever reasons, Peacock Mark II was a total popularity dud. In The Abbott Factor I argued that Peacock's unpopularity may have cost the Coalition that election.

None of the other bounce-backs are in anything like the Peacock class, and we also have near the top of the table cases like Fraser (assist from the Governor-General) and also a couple of tenures in which the incumbent was never that unpopular to begin with (but I put the bar at -10 to avoid excluding Howard's winning term and risking a claim of cherrypicking based on my previous research.)

Some other things to note here:

1. Even with me putting the bar high enough to allow Howard's 1996 win to limbo underneath it, the success rate for these unpopular Opposition Leaderships that even reached an election is a putrid 2 from 10, compared to 6 from 10 for Prime Ministers.  As I have mentioned many times before, the evidence is that going to an election with an unpopular Opposition Leader is either a tacit admission of failure or else a reason for failure to occur.

2. Unpopular Opposition Leaders are far more disposable than unpopular PMs, with a much higher proportion of bootings.  Excluding Latham, who resigned ostensibly for health reasons but would probably have been booted soon enough had he not done so, the average "trough" rating for Opposition Leaders who have been booted is the magic number again, -36.  However, Opposition Leaders are seldom booted at their nadir, and those booted are on average allowed to recover by 13.5 points before they are disposed of.  Probably, this represents the delayed appearance of action after panic.

3. The two Peacock losses (both respectable efforts, but losses nonetheless) were the only examples of an Opposition Leader more unpopular than Abbott is now being allowed to contest the next election by their party.  And in both cases the Coalition had no choice about keeping Peacock since his worst netsats were both during the campaign.

4. Unlike Prime Ministers, Opposition Leaders tend to record their bounce-back "peak" rating during the subsequent election campaign. 

5. Unpopular Opposition Leaders who are allowed to make it to the next election recover about as strongly as unpopular Prime Ministers, on average regaining about 19 points, although with massive variation.

There is a way of reading the above that seems not too bad for Abbott.  If we assume that both Gillard and Abbott are typical leaders, and assume that we are at Gillard's post-trough peak, and that Abbott has gone as far down as he is going to go, then we might not be too surprised to see them both at about -20 come the next election.  I think the Coalition would take that if they could get it, firstly because it would neutralise their current leadership disadvantage issue, and secondly because an unpopular PM is a bigger drag on their party's standing than an unpopular Opposition Leader.

But we do not yet know whether Gillard's rise will continue, we do not yet know if this really is the bottom for the budgie smugglers, and there would also be a major danger in deciding to retain Abbott on the basis of the track record of unpopular leaders bouncing back by an election.

That danger is that it may well be that the unpopular leaders who were sacked, would not have recovered as well as the ones who were not, had they been retained.  So if a party decided to keep an embattled leader (hmmm, I wrote the word that time without even thinking about it!) on the basis of the history of leaders who were kept - when otherwise they would have removed them - they could be making a very big mistake.  Using historical psephology like this to make decisions is fraught with these sorts of problems. 

There are also all kinds of factors that can be brought to bear on data like the above.  For instance: whether leaders were unpopular for a long time or a short time, whether they were unpopular near the election or far from it, whether they had ever been popular in the first place and so on.  But introducing too many of these into a predictive model will just overfit it and render it useless.

Intuitively, Abbott's problem is that he has never been popular to begin with, and his unpopularity is increasingly about his personality and flaws, rather than policy.  He had an underwhelming four month honeymoon period, a post-election bounce that lasted a solitary Newspoll, and from there on it's been as downhill all the way as a naturally bumpy event-influenced dataset is likely to get.  All very well to point out that Keating's ratings were worse, but Keating was running an economically troubled government that was getting very long in the tooth, while Abbott has the supposedly simple task of taking down a government that he and many right-wingers insist to be worse than Whitlam's.


Anyway, I make the following prediction: that if Abbott is retained and the government goes full term without collapsing, he will not get a Newspoll netsat exceeding -10 before the next election.  This is not a firm call that it can't happen, but I would rate his chances of making it back to above -10 in this term, conservatively, at less than 20%.  

There is some view that the Coalition should get rid of Abbott now and start the new year afresh as per the usual killing season model.  I don't know, but I suspect this would be risky.  Opposition Leaders do not need long in the job to build up their credentials if they are already well-known quantities, and if Abbott really is terminal, then keeping him in the job a few more months will provide time to be sure of this, in the form of continuing bad ratings and Labor taking the 2PP lead.   It will also provide more time to think about who on earth should be the replacement.  Changing too early provides too much time for the replacement leader to stuff things up, and increases the chance of replacing one failure with another. 

------------------------------------------------------------------------------------------------------------------
 Followup to previous

Thanks to very many readers for the great response to my previous article Embattled Abbott Thirty-Six Below The Wave, a rather provocative fusion of psephology and low-level activism (complete with a bit of fun Twitter stirring thanks to some helpful folk from Poll Bludger and elsewhere) in which I decided that the evidence warranted taking a pot-shot at media inconsistency in use of the magic E-word.  Thanks even to the guy on Twitter who called it a "Load of rubbish" (and still retweeted it!) and the person of unknown but ostensibly male gender on the Bolt blog who posted a post I found hilarious about it there (but still linked to my site!) 

A day after the article (and I claim no credit for this) Abbott was described implicitly as "embattled" (or at least soon to become so) in a very much mainstream piece by Paul Kelly in The Australian:

"THE re-election of Barack Obama is littered with lessons and traps for Australian politics but offers an insight into the successful tactics embraced by Julia Gillard and the emerging problems for Tony Abbott. 
 
There are three standout messages: the power of the negative campaign, the changing nature of the culture war and the narrative that defines the embattled leader." (my underline)

(NB If that's all you can see before the paywall, try typing the article title into Google News then clicking on the first hit that comes up.  Usually works for me.)

The article states that the Liberals have blundered by failing to contest the marketing of their leader, thereby allowing Labor to do it for them:

"Frozen by their lead, the Liberals sat pat and allowed Labor to recast their leader's image" (in bold, four words I wish I'd written!)

The article copped some flack for closing with a somewhat Catholic-sounding line about Gillard's 
need for forgiveness from the electorate (which some Poll Bludger readers parsed as Gillard's need for forgiveness from News Ltd) but despite the way it's worded there's a point there.  Voters do need to decide that they are not really that angry anymore about the perception of a no-carbon-tax broken promise and the mess Labor made of explaining the removal of Rudd.

(I add that there are some things in Kelly's commentary on the US election that I consider to be debatable.  For instance he argues that Mitt Romney "never recovered" once he was defined by the Obama campaign.  I agree that advertisements like "Firms" played a big role in ensuring Romney went into the campaign well behind, but his first debate performance did reverse a lot of negative perceptions, mainly by giving some impression of humanity and self-deprecating humour.  Only it wasn't enough, and I think that's not just because the Republican campaign was technically poor, but also because the case for removing the incumbent was weak.

I am not convinced that had the vote been "a referendum on Obama's weak record", Obama would have lost.  There is strong evidence that the voters simply did not believe Obama was to blame for the state of the economy; eg CNN exit polling shows that when voters were asked who was more to blame for it, 53% said George W Bush to just 38% saying Obama.  This suggests that outside Republican loyalists, the view that it was really Obama's fault found little traction.  Indeed, quite a few Romney voters weren't prepared to blame Obama.)

--------------------------------------------------------------------------------------------------------------------
Some standards commitments on this site

On TT, two sides of my politics-related persona occupied largely different spheres.  The mild-mannered numbers-minded analyst who I call "Jekyll-Bonham" wrote the articles as house psephologist, and the forthright contrarian ("Hyde-Bonham") kicked heads in the comments.

On this site, this separation doesn't exist and both these kinds of output get equal billing.  (Hyde-Bonham pieces will quite often be tagged "no pseph content" on Twitter so that people who are afraid of him can avoid this nasty character.)  And in pieces like Embattled Abbott, you may from time to time get something that was probably written by both.  Lest anyone suspect that my politics is trumping my psephology too much, I do make the following commitments:

1. This site will not cherry-pick which results to publish. Whenever I finish investigating a substantial pseph matter, whatever it is, I will publish the results as soon as I can, whether it's politically convenient to me or not.  (The exception is where I do work that is covered by a confidentiality requirement.)

2. Whatever pseph findings I publish will always be my actual opinion.  Though I will quite often present advocative arguments that rely on my pseph research, at no stage will I say anything about a pseph matter that I don't believe to be true in order to do so, nor omit anything that I consider true and relevant.

And of course, if you ever find a factual error in one of my articles, drop me an email, and if it really is an error I will fix it.

---------------------------------------------------------------------
Update (20 Nov): changed "no carbon-tax broken promise" to "perception of a no carbon-tax broken promise" as despite the current carbon pricing model being more or less universally colloquially referred to as a "tax", its actual status as technically such is at the very least disputed. See also Gillard's pre-election commitment to seek a market-based carbon mechanism here.