Sunday, December 29, 2013

Unpopular State Premiers Have Dire Historic Fates

An updated version of this article is available: see Unpopular State Premiers Still Have Dire Historic Fates (2020)
---------------------------------------------------------------------------------------------------------



Today's post is brought to you by the letters C and B and the magic number -20.  Newspoll finally released the first state polling for Western Australia of any reputable kind since Colin Barnett emphatically won a second term in the state election way back in March.  Given the magnitude of Barnett's victory earlier this year, with an 8.8% swing and a net gain of seven seats (plus two for the Nationals), it might be expected the Barnett government would still be on cloud nine nine months later.  But no, the poll (taken between October and December) shows the government with a feeble 51-49 lead.  It also shows Colin Barnett with very bad personal ratings (34% approve, 54% disapprove) and that Labor's Mark McGowan is now rated better premier by six points, 43-37.

There has been speculation recently that Barnett is on the skids, but some of it has seemed, in theory, questionable.  For instance, should anyone really care  that Standard and Poors has downgraded WA's credit rating when that agency has a recent history of predictive failure?  And if the answer is really Troy Buswell, one would hope that the question was pretty stupid.  Now, however, we have something concrete: Colin Barnett is not well liked.  It came as some surprise to me to find that his -20 is in fact the worst netsat by a WA Premier in the admittedly spotty and discontinuous 27-year history of Newspoll ratings.



When discussing federal politics, I've often pointed out that Prime Ministers seem to be able to get away with more or less any level of unpopularity - not just as a freak event but seemingly routinely.  Hawke won after a term in which he'd been down to -25, Keating from -44 and Howard in successive terms from -31 and -36.  Since 1986, the break-even term-low for an incumbent PM (the rating that gives them a 50% chance of retaining office after the next election) has been a remarkably bad -27

At state level, it's remarkably different.  The following is a table showing every State Premier since 1986 who has polled a term-low Newspoll netsat of worse than -10, together with their fate at the end of that term.  (These are by no means the worst ratings ever recorded - for instance short-term Tasmanian Premier Harry Holgate made it down to -59 in Morgan polling before Newspoll started.)


It's not a pretty picture for the Barnett fans.  24 of these 29 unpopular premierships have ended in electoral defeat or in some form of pre-election (or in one case immediately post-election) removal somewhere along the jumped-to-pushed spectrum. (A note re Groom: as Premier of a majority government he had promised to govern in majority or not at all, and resigned so his party could continue in office in minority.) One of the four wins was so underwhelming that the winner (Bartlett) was gone within a year.  All four wins were by first-term Premiers and three of them were from ratings less bad than Barnett's is now. I haven't checked all past Premier ratings to see what the break-even point is (ie at what term-low netsat does a Premier have a 50% chance of winning the next election) but I suspect it is somewhere round zero, and it's certainly well up from where Barnett is now.

The part of the chart above Barnett is especially striking: only one of sixteen previous Premiers to have copped -20 or worse at any time has finished that term triumphant. 

If we look at the history of Better Premier ratings at state level there's not a lot of good news for Barnett there either.  In general Better Premier shows the same pattern at state level to federal level: there's a huge advantage to incumbents, so if the Opposition Leader is ahead then either the Premier or their government is on the nose, usually both.  However, at state level it's more common than at federal level for the leader of a defeated government to hang around for a while, and if their replacement is off to a rocky start some odd figures can emerge.  Carmen Lawrence was considered better as Premier than Richard Court by a massive 61:27 soon after Court took office, and Joan Kirner led Jeff Kennett by a point in his first poll as Premier.  When well established Premiers have trailed in a two-way Better Premier question, they have never won the election afterwards.  Iemma, Rees, Keneally, Bligh, Bannon, Arnold, Olsen, Rann all trailed and all lost or were removed.

What I have not looked at closely yet, and may examine later, is the interplay between the personal ratings of Premiers and their party's 2PP polling - a complicated issue in many states because of OPV, three-cornered contests and the patchiness of published 2PP results.  It's no surprise that many of the above Premiers were not only unpopular but in charge of governments that were trailing massively and clearly doomed.  The important thing is that they were not all like that.  For example, at the time that Iemma and Brumby became about as unpopular as Barnett is now, the 2PP polling in their states was 50:50.  If, as at federal level, bad leader netsats are an advance indicator of worsening 2PP polling, then a bad rating for the Premier in the context of an average 2PP isn't something to be ignored.

Why the difference between state and federal patterns?  I believe one reason is that state governments are nowhere near as resilient to bad polling as federal governments.  Voters grumble about federal governments a lot, but see dismissing even an unpopular federal government from office as a big thing.  Changing state governments, should some reason arise, is a much less dramatic step, and so the incredible recoveries from shocking polling that we see governments pull off federally are distinctly rare events at the next level down.

Another possible reason for this is that the federal picture significantly impacts upon state elections.  It is a massive disadvantage for a state government to be of the same party as the party in power federally.  We don't know yet whether this will still be a problem for Barnett in 2017, but it has plenty of potential to drive some troubling ratings for his leadership and his party in the meantime. 

I won't be writing off the Premier of WA just yet, though this time I won't go so far as to try to passionately deter others from doing the same.  Just because there is a history of leaders not successfully recovering by the time of the next election from ratings like this does not mean it can't be done (eg Tony Abbott becoming PM despite the history of lastingly unpopular Opposition Leaders always losing).  Plus it is early in the term with plenty of time to turn things around given the chance.  Bloated by federal factors as it was, Barnett's win earlier this year was by an enormous margin and Labor will be needing to win seats on margins of around 9% (and on smaller margins but with their opponents being assisted by sophomore effect) to win government.  That task does become easier if a few independents help out by winning safe Coalition seats. 

The obligatory disclaimers noted though, the historical record says two very different things about bad personal ratings for PMs and Premiers.  Bad personal polling for Prime Ministers is just a challenge to be overcome, and enough of them have done so down the years.  Bad personal polling for State Premiers, on the other hand, is usually an advance sign of political doom.  After this result, should Barnett fail to make it to the next election, or lose it, no-one should be at all surprised.

No comments:

Post a Comment

The comment system is unreliable. If you cannot submit comments you can email me a comment (via email link in profile) - email must be entitled: Comment for publication, followed by the name of the article you wish to comment on. Comments are accepted in full or not at all. If you submit a comment which is not accepted within a few days you can also email me and I will check if it has been received.