Friday, July 22, 2016

EMRS: Hodgman Government Would Lose Majority

EMRS July 2016: Liberal 37 Labor 32 Green 17 "Independent" 12 Others 2
Intepretation: Liberal 38.5 Labor 35.5 Green 14 Others 12
On this poll Hodgman government would clearly lose its majority
Seat projection 11-10-4 (Liberal-Labor-Green) or 12-10-3.
No seats are projected to fourth parties/independents as no prominent fourth parties are yet known to be running

Also now released EMRS May 2016: Liberal 41 Labor 29 Green 21 "Independent" 8 Others 1
Interpretation: Liberal 43 Labor 31 Green 18 Others 8
On this poll Hodgman government would have lost its majority in May as well
Seat projection for May 12-10-3

Aggregate of all recent polling 12-10-3 (no majority)

After an unusually long period of silence, Tasmanian phone pollster EMRS has hit us with not one but two poll readings for Tasmania, taken in May and July.  Also see the trend tracker.  Both these results show the Labor Opposition on the rise and the Hodgman Government's stocks falling, in the July case to the Liberals' lowest level since the aftermath of the 2010 election.  This isn't the first time an EMRS poll has pointed to a minority government outcome, but it is the first time it's happened twice in a row.

The July poll shows a massive rise in the level of support for "Independent" to an EMRS-record level, which is highly problematic given that at this stage there is no prominent fourth party or independent known to be running at the 2018 state election.  It is difficult to avoid suspecting that the July poll - and especially its unusually low (for EMRS) undecided rate - is heavily contaminated by the recent federal election, with voters for Andrew Wilkie or Jacqui Lambie imagining that they would find someone similar on the ballot to vote for in two years' time.  Whether they will, or whether this vote will at best splatter between a number of minor parties without actually electing anyone, still remains to be seen, but the potential for a fourth-party win is out there if anyone can do it.

Nonetheless the trend is not at all the government's friend, with it shedding eleven points on the headline rate since November, and Labor up seven points in the same time.  Labor will be full of beans after winning the Upper House seat of Elwick in May (albeit not against Liberal opposition), followed by unexpectedly strong results in the federal election (especially Ross Hart's 56:44 demolition of Andrew Nikolic in Bass, a result outside the margin of error of every poll conducted in the seat).  Two and a bit years on from the end of the much-disliked ALP-Greens coalition government, Tasmanian voters seem more interested in again kicking the incumbents than in holding grudges.

EMRS as a pollster has a long and stable history of getting the Liberal Party vote pretty close to right while underestimating Labor and overestimating the Greens.  However it is difficult to know quite what to make of its recent surges to "Independent".  It is also difficult to compare present with past EMRS results as I am now firmly convinced (having myself been again polled for both the May and July samples) that EMRS has become a "panel pollster".  It cannot possibly be still random-sampling the entire Tasmanian population; it is doing at least a significant part of its polling by calling the same people over and over again.  This isn't ideal, but then, unfortunately, no Tasmanian polling ever is.

As for that 14% Independent/Others vote, at the recent federal election something interesting happened in the House of Representatives seats with the micro-party vote.  ReachTEL polls had shown high support for the unheralded (and at that stage obscure) Recreational Fishers Party, and also for the Christian Democrats and Renewable Energy Party.  These findings seemed unlikely but in fact proved fairly accurate - these parties averaged 5.6%, 1.9% and 2.3% respectively in the seats they contested, compared to the poll readings of 5.0, 3.3 and 2.6.  Yet all these parties have sunk without trace in the Senate count. So there is a general mood on the ground to give the "big three" (Labor, Liberal, Green) the finger by voting for anyone else, however obscure.  The high "others" vote in this poll might therefore be close to real.  Whether this would coalesce into enough of a vote for anyone specific to actually win a House of Assembly seat is much more difficult to say.

An obvious candidate would be the Jacqui Lambie Network, but the JLN's high below-the-line vote in the Senate election (with quite a lot of that being a personal vote for Lambie herself that then leaks out of the ticket) shows that the Lambie brand would not automatically transfer to a Lambie-endorsed candidate.  Even in the Lambie Network's best division (Braddon, where they polled 14.3%) there is no guarantee a JLN state candidate would be able to rustle up the 8% or so that might be good for a state seat.

Possible results under this poll

The following is a model of what might happen under my interpretation of this poll, but it's a very rubbery model for a number of reasons.  Firstly there have been no electorate-level state polls released for a very long time, so the electorate breakdown of support may be out of whack (especially in Bass).  Over time I will look at reweighting the electorates using what we can glean from the federal result.  Secondly, the "independent" vote is extremely difficult to handle for the reasons discussed above.  We can't treat the Others vote for a given electorate as behaving like a single candidate, but the potential is there for some fourth force to tap most of it in given electorates and potentially win seats almost anywhere, especially if they can take votes from the Greens.

This projection seems rather friendly to the Greens, who stand to pick up a seat in Lyons with a fair chance of retaining their Bass seat off a vote that (after my usual house-effect correction) is not that flash.  But in the absence of serious fourth-party challengers, they would probably do so.

Assuming this poll is somewhere near right, what is the government doing wrong?  The (now largely resolved, but avoidable and expensive) Basslink energy crisis and the health sector are two possible culprits, the latter of which reared its head in the federal election campaign.  Federal drag is also a likely suspect, with a widespread feeling that the federal government is losing interest in the state, and Liberal voters may be (quite rightly) blaming the state party in part for the party's terrible federal outcome.

But there's more out there than that - anecdotally I get feedback that the state government is seen as not doing enough to maintain its relationship with the communities that elected it, and that some people perceive that the government is not really doing much at all.

[Correction: This article previously contained an example of inaction in the area of threatened species classification, but I have since found that changes were in fact finally made in this area shortly before this article was published.]


Unfortunately EMRS only poll "preferred premier" scores; they have at times polled approval ratings (which are vastly more useful) but have not actually released them.  The preferred premier scores show Will Hodgman leading Bryan Green 48-25, down from 52-21 in February and 51-22 in May.  Preferred premier scores skew greatly to incumbents but even taking that into account, it does seem that Hodgman is holding up better than his party given its current dive.  Whether this is because voters like Hodgman, don't think Green is up to it, or are really in many cases Liberal supporters parking their vote in the Independent column, we can't say.  In the absence of separate approval data for each leader this figure doesn't mean a lot.


As well as these two EMRS polls there was also a Morgan SMS poll in late May that had the government on 41%, Labor on 34.5%, the Greens on 17% and Others 7.5%.  As this series has displayed a massive skew to the Greens and against the Liberals this was a rather better result for the government than the surrounding EMRS readings, assuming that it wasn't just evidence of the house effect of Morgan's polling changing for some reason.  That said the Morgans have small sample sizes, SMS is in my view a trashy polling method, and Morgan generally is a brand that has done much to earn a lot of caution.

The following is my combined reading of these recent polls, based on the previous aggregate and weighting the new EMRS at 45%, the May one at 20% and the Morgan at 4%.  The aggregate is coming off much higher values for the government so still treating these two recent bad polls from the same pollster with some degree of caution.  Nonetheless, it still maintains that the government would lose its majority.

(For those wondering re Bass, the Greens might hold it, but micro-party preferences were adverse to the Greens in 2014).

Where do we go from here?

The last Tasmanian Liberal majority government led by Ray Groom won a thumping majority at the 1992 state election but lost its majority at the 1996 poll with a double-digit swing against it.  This government now faces the problem that not only is the federal government of the same party as itself, but also the election results appear to have marginalised the state's influence in Canberra.  Assuming that the Turnbull government does not collapse before the next state election, historic patterns suggest that it will be very tough for a Liberal government in this state to retain its majority.


  1. Based on the recent Federal election, do you think Jackie Lambie take a seat off the Libs in Braddon?

    1. Lambie herself would bolt in, but I am sceptical of whether that translates to any candidate endorsed by Lambie.

  2. You are right re- comments on green tape. The libs were all bark in opposition but they are terrified of the Greens/ENGOs and are really letting down those who voted for them to bring back some balance. Biggest mandate I have seen squandered. You cant just cruise along hoping for a second term....

  3. You just have to shake your head in disbelief at how much the preferred premier score favors Will Hodgman after such a lack lustre demonstration of leadership qualities. Not with standing what you said about it not being a great indicator of a leaders performance or skill, wasn't Lara Giddings polling poorly on that measure against Will Hodgman at this stage in the election cycle? Do state preferred premier polls produce bigger margins favoring the Premier compared to the same type of poll looking at preferred primeminister?

    1. Giddings was trailing Hodgman 22-45 at roughly the same point in the previous cycle though the methods are not the same - at the time they included the Greens leader, McKim, who was on 15; had McKim not been included I expect it would have been something like, say, 30-48. The thing here is that preferred leader ratings are also strongly influenced by party standing, so when a Government is being thrashed in voting intention that wipes out the incumbency advantage. (In federal polls, if the government trails 47-53 then the PPM will normally be about level, for a larger gap the Opposition Leader should be the preferred leader).

      My impression is that normally the advantage for state premiers isn't any greater than for PMs. However EMRS has only been doing two-leader comparisons for a short time so we don't know much about how they compare with other pollsters.

  4. There is now no getting away from the question the media needs to grill Will with, will you continue to govern at the next election under a minority government relying on the Greens for support? No mixed messages or dodgy answers allowed you won't get away with it, the media will need to do their job properly on this one.


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. Comments will be published under the name the email is sent from unless an alias is clearly requested and stated. 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.