Friday, May 10, 2013

The Abbott Factor Revisited: Abbott's Ratings and Labor's Decline

 Advance summary:

1. This article covers changes in the relationship between Tony Abbott's net satisfaction rating and Labor's two-party preferred status.

2. Until September 2012 both Abbott's standing and Labor's tended to decline at the same time, although this did not always happen.

3. In late 2012 there was a period in which Abbott's standing declined while Labor's improved, which has now reversed.

4. There is currently very little evidence that Tony Abbott's unpopularity is a significant barrier to the Coalition winning the next federal election.




----------------------------------------------------------------------------------------------------------
With the Tasmanian Legislative Council elections now over, it is time to get back to federal issues, last covered on this site over a month ago.

There hasn't been that much to see in that time.  Labor's two-party preferred position crashed in early February around the time of the Week From Hell (early election date announcement, Peris "captains pick", ministers not contesting next election, vanishing mining tax, Swan figures gaffe etc) and crashed some more following the abortive spill in late March.  There have been weak signs of a possible recovery over the last month or so, but Labor hasn't broken into what I think of as the "competitive zone" (closer than 46:54) and even those have now reversed.  The Coalition's aggregated position across all polls is at least in the 55s, there is no sign of polling life for Labor whatsoever, and the clock is ticking down with now just over four months to go.  Strategic signs are that Labor has the white flag up and is just trying to minimise seat losses (not very effectively thus far), and for most analysts the main question now is just how big a defeat it will be.  Even on that question, those who were certain this would not be a rerun of, say, 1975 in margin terms are now less certain, or should be.

Indeed, the Gillard Government's polling position is rapidly approaching that of the Howard Government in 2007, when viewed on the basis of points behind for time to go til the election.  On my rolling Newspoll-only averages, the Gillard Government was at 44% with four months and one week to go; Howard was at 43.6% one week closer to the election.  Howard's government did recover to suffer only a moderate defeat in the end, and while Labor may do likewise, I would not take it for granted that they will. 

Betting odds are blowing out, as well they might; while we have seen the government in these sorts of polling positions before, the closer the election is to being held, the more predictive a large gap in polling becomes.  Attribute and issue based polling questions show that voters often approve of what the government is doing in principle, but when asked to comment on the government's implementation, they switch to responding through a partisan prism.   If you're a progressive reader and really want to get thoroughly depressed about the extent to which the Coalition is running psychological rings around Labor as demonstrated by attribute polling, the view from Gordon Graham's site is even grimmer than this one.  Mr Graham recently offered the following extra-cheerful assessment of Labor's position:


Belief that the government is still competitive - essentially an article of faith rather than evidence - comes now not in the form of belief that its polling position will recover, but instead in belief that the polls are all Just Plain Wrong.  (Mobile phones vs landlines and uncommitted/refused rates are the two commonest lines of objection - I may look at the second one in detail sometime soon as it is something that deserves serious and objective analysis).  Nothing the government does rebounds to its favour in the eyes of the public, and while I am not going to absolutely call an election four months out, it shows almost every possible psephological sign of being a Dead Parrot. 

But it was not always thus.  Last spring there was a resurgence in Labor's position that saw them rapidly improve from the mid-40s to around 48% two-party preferred.  Many reasons for this improvement were suggested, but the one most widely anticipated in advance was that the Coalition's margin had been inflated by fears about the impact of the carbon tax, that then failed to be justified.  At the same time Tony Abbott's personal ratings, which at that time had gone down in step with the Government's position (the reverse of the normal pattern for Opposition Leaders) remained in freefall. 

In those far-off days I explored the matter of Abbott's unpopularity and its links with the Coalition's polling position in a number of articles:

The Abbott Factor: Opposition Leader Ratings and Party Standing
Newspoll Provides More Evidence that the Abbott Elastic has Snapped
Embattled Abbott Thirty-Six Below The Wave
As Gillard Recovered So Can Abbott?

The last two of these, while still hopefully of some use, are feeling a little bit sorry for themselves in places these days.  They plead that they had no idea that Julia Gillard would call an election seven months from the date, or that Simon Crean of all people would force a leadership challenge on behalf of Kevin Rudd, which would then fail to actually happen. However, at least one aspect of Labor's current problems - that its recovery-period polling was inflated by unrealistic expectations the party would deliver a budget surplus - was forseeable.  The tentative prediction concerning Abbott's future Newspoll ratings in "As Gillard Recovered"; unwisely framed in terms of single poll results, had a near-death experience in the near-rogue Newspoll of late March, when it went within two points of being falsified.  Currently it sits six points from the edge.  (That prediction was that Abbott would stay in Newspoll double-figure negative netsats right through til the election.  I don't much like its chances of surviving now, especially given how bouncy Newspoll has been.)

(Indeed as my later article on polling and the "Spill Cliff" noted, Abbott's recovery from bad polling is the equal second-greatest by an Opposition Leader, now trailing only Andrew Peacock's in 1984.  Both Abbott's and Peacock's recoveries have one thing in common: the election date being called a long time in advance.)

An observation of these articles was that the link between Abbott being unpopular and the government being unpopular had stopped being causal.  At that time the government was becoming more popular while Abbott was becoming less, and it seemed quite possible that that trend would simply continue until a point at which Abbott had to be removed.  But just as Abbott could become less popular while the government became more so, so the breaking of the elastic between these two things meant the reverse could happen too: the government could become more unpopular while Abbott became less.

And it is that we have been seeing since late January.  Ugly graph time!


This ugly graph (click for a larger version) shows a comparison between Abbott's net satisfaction (points in blue, figures on the left) and Labor's 2PP (points in red, figures on the right) since the 2010 election. Numbers across the top (below the years) are simply a time sequence of Newspolls.  Because the Labor 2PP axis is stretched, it looks much bouncier compared to the Abbott netsat than it actually is.  In both cases, I have used rolling averages to smooth out a lot of the bouncing that would otherwise obscure the pattern.

Labor came out of the election with roughly 50-50 2PPs and Abbott with roughly even average ratings (partly on the back of a post-election bounce that lasted precisely one Newspoll).  Both dropped sharply from around poll 9 in the sequence following Labor's announcement of a carbon price.  At around poll 13 (big number 1), there is a split: Labor's ratings fall through the floor and then rebound while Abbott's netsat falls slowly. While the two lines are both going down on average from the election to big number 2, the Abbott line is declining steadily while the ALP line is rising and falling.

Big number 2 is the "elastic snap" that I talked about last spring, although it can be seen that the relationship was never all that causally tight anyway.  At this point the Labor 2PP heads up while the Abbott netsat keeps going down.  This is actually not the first time this had happened - the first recovery in Labor's fortunes, starting in October 2011 and ending in March 2012 after Julia Gillard defeated Kevin Rudd's challenge, displayed a similar pattern.

Big number 3 is the point early in 2013 when the relationship suddenly reverses: Abbott starts going up, Labor starts going down (the Week From Hell is the first big drop and the Spill Cliff is the second).  Actually, on Newspoll at least, Labor peaked before Abbott began improving.

We can put a straight line through either the Labor or the Abbott graphs and get a pattern going down over the period, but since September, we are seeing the more normal relationship for Opposition Leaders, as discussed in the initial Abbott Factor.  If the government goes up, the LO goes down, and vice versa, although historically it isn't that strong a connection. 

I believe there are a few main causes of the pattern we have seen this year, in which Abbott has been able to improve his ratings after a past history mostly of coasting downhill:

1.  Abbott's critique is increasingly seen as partly vindicated.  While his doomy predictions about the impact of the carbon "tax" have not eventuated, his party's attacks on Labor's ability to deliver a balanced budget are seen as justified because Labor will not deliver one.  His party's attacks on Labor instability are seen as justified because the government has been internally disunified, culminating in the failed "spill".  If voters have disliked Abbott for being too negative - which Galaxy's female-voters poll found to be a bigger turnoff than his social-issue positions -  but start to think that some part of his negativity was right, they are more likely to give him a free pass (at least to convert a negative rating to a neutral or a neutral to a positive).

2. Abbott's performance in interviews has improved.  Avoiding them for a while may have helped, but he seems to have shed the hapless, grinning naughty-schoolboy persona of last year and adopted a more serious and reserved demeanour, together with a lower gaffe rate.  There was widespread frustration from left-wingers on Twitter when Leigh Sales' most recent interview with Abbott failed to demolish him in a similar fashion to their previous encounter, but it's not really the interviewer's fault if the subject fails to say anything stupid.

3. Abbott is under no pressure, save for that of making sure he seals the deal, and preferably does so with a respectable seat margin.  From February on, with Labor's ratings in freefall, it has been clear that Abbott's leadership is not under threat, since the party would be bonkers to roll him while they are in an objectively winning polling position.

It is hard to now see Abbott's unpopularity as still being any real barrier to the Coalition winning the election.  As pointed out in the original Abbott Factor, the damage caused by even a very unpopular Opposition Leader to their party's standing is typically limited to a point or two, meaning that it is only an issue when the 2PP picture is close.  Most elections are close, but this does not show signs at this stage of becoming so.  Not only is the 2PP picture not currently close (even with Abbott's now moderate unpopularity factored into it), but Abbott is not currently all that unpopular anyway. 

2 comments:

  1. U never factored in Abbott's court case or the 'Ashbygate'affear Doc.

    ReplyDelete
  2. Indeed not because I greatly doubt either of them will do him damage. Slipper is in disrepute and no-one likely to vote Coalition cares if/that there was a conspiracy to bring him down. The One Nation stuff is ancient history involving a discredited party and Ettridge is representing himself. Hard to see that going anywhere in a hurry but even if it does no-one will care.

    ReplyDelete