Wednesday, December 6, 2017

EMRS Says The Wheels Are Falling Off

EMRS (Tas State) December: Lib 34 ALP 34 Green 17 JLN 8 IND/Other 7
Appears to be lowest Liberal primary for 11 years
Interpretation (based on historic skew) Lib 35.5 ALP 37.5 Green 14 JLN 8 Others 5

Modelled seat results based on this poll if election "held now": hung parliament with 10-10-4-1 (Liberal, Labor, Green, JLN) with next most likely outcome 9-11-4-1
Rolling aggregate of all state polls 12-10-3-0 
Rebecca White increases Preferred Premier lead over Will Hodgman to 13 points

If the December EMRS poll is to be believed (see also the helpful trend tracker), the Hodgman Government is currently headed for a Campbell Newman-like reversal of fortune at the 2018 Tasmanian state election.  Having won a massive victory from Opposition at the 2014 state election, the current poll suggests Hodgman's government, much like Newman's, could be going straight back where it came from and that election night could be carnage with incumbents losing all over the place - to Labor, the Greens, the Lambie Network and their own party.  On a like for like basis (which is rather difficult to follow through old EMRS poll reports) this seems to be the Liberals' lowest primary in an EMRS poll since August 2006.

It's only one poll, so we shouldn't get carried away (and it would not be surprising to see reports of different figures in Liberal internals, as has been the case in the past).  It's also by a pollster that only really gets tested once every four years (apart from an average performance on the same-sex marriage survey) so perhaps it's all a load of nonsense.  But if the government is looking at anything near this in its internals then it would be very concerned.  To make matters worse on a headline which shows Labor on level terms for the first time since mid-2009, EMRS has a history of underestimating the Labor vote at elections.  There's not a lot of time left before the state election and probably won't be any more parliamentary sittings, so it's all down to the skill of the party machine in convincing voters not to vote Labor.  Oh wait, Pembroke, moving right along then ...

The poll is the second recent EMRS poll to include the Jacqui Lambie Network, who are now clearly running in at least Braddon, albeit with a rather low profile set of candidates.  For this reason I have had to play about with assumptions regarding the Others vote to separate out JLN (just as I did with One Nation before it turned out they were iffy about running - it seems they might be again now, but who knows.)  At this stage, I am assuming the distribution of the Lambie vote will follow the Senate election vote, although it's possible JLN won't even run in Franklin or Denison.

After applying historic adjustments for EMRS' tendency to overestimate Others and the Greens and underestimate Labor, my model of the current poll looks something like this.  The JLN factor makes it very rubbery, as does the lack of recent electorate breakdowns:

In this model, the Liberals lose a seat to Labor in Bass; the concentration of their vote with Michael Ferguson and Peter Gutwein prevents them from exploiting Andrea Dawkins being a long way short of quota.  In Braddon, the Liberals lose two seats (to Labor and the Lambie Network); although the JLN ticket will leak a lot, that sort of lead would be too much.

In Denison, the Liberals are well short of a second quota.  One risk is that the high share of that electorate that says "independent" to the pollster are mostly lefties who actually vote for Labor and the Greens in the absence of a Wilkie-endorsed candidate.  A second (lower) risk is that after Scott Bacon bolts in, Labor's #2 and #3 end up with a fairly evenly split vote and are able to stay ahead of the second Liberal with the assistance of Green preferences.  (Having Sue Hickey on board may hope reduce the risk of these calamities, though it could in theory contribute to the second.)

In Franklin, we have the long-mooted loss of one seat to Labor, and in Lyons the Green vote in this model is too high to prevent them from gaining a seat.  Of course, any of these things might not pan out like that off a certain level of vote, but what goes down in one electorate would probably come up somewhere else.  I therefore have 10-10-4-1 as the most likely distribution if votes were cast according to this poll, with 9-11-4-1 the next most likely.  On these numbers, neither major party gets anywhere near a majority.

Although this poll suggests the Lambie Network has enough for a seat, I'd treat that with caution still.  We may soon see Lambie herself replaced as a Senator, and unless she manages to be her own replacement (which is possible) this could be a destabilising event for her brand.

Aggregate Of All Polling

In this business we do not rely on just one poll, especially not when it is an outlying result.  I've been maintaining an aggregate of all Tasmanian poll results available to me.  This is what it looks like when I mix in the current poll, which I've given a weighting of 40% since it is so long since the last one.

The cross-poll aggregate shows what an unpredictable system Hare-Clark can be, because of it being about candidates and not just parties.  In this aggregate the Liberals have only a two-point primary lead, which would seem completely hopeless.  However, after giving away two seats in each electorate to each major party and Denison and Franklin for the Greens, in this model there are three left in play:

* Bass, where it would be very close between the Liberals and Greens for the final seat but I believe preferences would be slightly better for the Liberals.
* Braddon, where it would be the Liberals vs JLN for the final seat but the JLN vote would probably leak enough to allow the Liberals to survive.
* Lyons, where an even enough distribution between two or three candidates within any party could result in either major party beating the Greens, although this isn't very likely.

On balance I'm giving Bass and Braddon to the Liberals and Lyons to the Greens in this aggregate projection, but it does show that there is a freak-save scenario for the Liberals that might work off really small primary vote gaps, if they can keep the swing down in the right seats.  It probably won't happen, though.


Rebecca White is way ahead of Will Hodgman as Preferred Premier, an indicator that normally favours incumbents, 48-35.  It's not rubbing off on her party's primary vote much, but all that really says is that voters for minor parties prefer her to Hodgman.  We don't have any direct measure of White's personal popularity yet but it is probably very high.  Nothing the government has come up with so far to attempt to damage her has worked.  See previous article for history from other states.

What Is The Hodgman Government Doing Wrong?

Lots of little things and the odd medium-big thing?  I haven't found explaining this government's steadily worsening polling easy over the last year or so.  The state's economy seems to be a lot better than it was and there is a lot of development going on, and the tourism sector seems to be doing well too.  On the other hand, Tasmania is historically a Labor state a lot of the time, the federal government's poor standing is a drag on the state government, and the health sector is a continual source of voter concern.  The government also has some pretty average performers on board, and a strange fondness for pushing culture-war politics on issues where the Legislative Council isn't interested.  If I was writing a big-picture story of how the last Liberal majority government lost its majority, it would be a simple one: it picked big fights with the unions and voted itself a 40% pay rise.  With the current government, if it goes the same way it seems like we will need a very long list of mostly little reasons.  Lately for instance we've seen the government's age-attack on an opponent in a by-election, its troubled attempt to take over TasWater, some small-scale public service scandals and so on.  I can't see any one thing over the past four years that goes close to being easy to pick out and say that it was by itself an election loser - if the election is indeed lost.

In the event of a hung parliament, my suspicion is that Labor and the Greens will find some way for Labor to govern.  While White recently ruled out governing with Green support at a Property Council event, there is a history of governments finding some way around these pre-election comments.  The indications are that if the Liberals fail to win a majority and attempt to keep governing, they will be removed on the floor of the House quickly.  Things might be different should the balance of power holder be someone other than the Greens.

If You Don't Like This Poll ...

...Just wait a few days, there might be another one.  ReachTEL have been active in Tasmania over the past fortnight although I have no idea who has commissioned their poll.  Their poll is blighted by the inclusion of the Nationals (who are not known to be running, though a since-disbanded rogue version scored nothing to speak of last time), a forced-answer two-party preferred question (aaargh!) and a seemingly inescapable, not to mention interminable, charity-question nightmare at the end.  Still, it may be of some use as a second opinion if its results see the light of day.

PS Actual Netsats!

In a radio interview with Leon Compton (ABC), EMRS released actual "net favourability" ratings for the three party leaders.  (This is basically the same as "net satisfaction" ratings seen in Newspoll save that voters seem to have been asked if they have a "favourable" view, a wording more common in the US or UK.)

Rebecca White scores +40, Will Hodgman +13 and Cassy O'Connor -5.  Hodgman's rating is very similar to his Newspoll scores before the 2010 and 2014 election, suggesting that the troubles of his party have not rubbed off on him personally.  O'Connor's rating is fairly typical of the few Green leader ratings available through time (in contrast to Nick McKim who had very high ratings in 2010 but very low after a term as a Minister).  White's rating is, of course, through the roof.

Historically it is fairly unusual for Premiers with positive net ratings to lose elections, but it's possible that "favourability" is a softer measure here than "satisfaction".  If it isn't, then Hodgman's moderately good rating might indicate some potential for the Liberals to improve.  Opposition Leader ratings, in contrast, don't have a huge amount to do with voting intention historically, and there is a view sometimes argued by Peter Brent that high leader ratings (federally, think Rudd or Latham) can be connected with soft voting intentions for their party.  It's very odd for an incumbent who is being thumped as preferred leader to still have their own personal rating in positive territory.

The other material added in this interview is confirmation that the Lambie Network is polling especially well in Braddon (around "12-15%"), and confirmation that as usual the government and Will Hodgman are being received better in the north of the state than in the south, while for Labor and White it's the other way around.

Thanks to the reader who supplied the audio from the interview.

No comments:

Post a Comment