Tuesday, December 3, 2013

EMRS: Another Shocker for Labor

EMRS: Lib 49 ALP 22 Grn 19 PUP 5 Ind 4 Others 1
Interpretation (provisional): Lib 49.5 ALP 26.5 Grn 16 PUP 5 Ind/Others 3
Outcome of this poll if election held now: Liberal Majority Government (13-14 seats, possibly 14-7-4) 
State polling aggregate: Liberal 13 Labor 8 Green 4

A new EMRS poll of Tasmanian state voting intentions has been released and it's a pretty similar story to what we've been seeing in state polling for the whole of the last three years.  But at least it's the first one to explicitly canvass support levels for the Palmer United Party (PUP).

Every single poll since the start of 2011 has pointed to a Liberal majority government if an election was held now, and this one is no different. There has been variation between polls in the strength of the Liberals' lead and the division of the punishment between the minority government parties but not one of these polls has yet pointed to another hung parliament. Even when these polls are tweaked to adjust them for the well-established anti-Labor pro-Green skew of EMRS and the apparent anti-Labor pro-Liberal skew of ReachTEL (at least in its local-level federal polling) it makes no difference; all roads point to 13 seats, at least.  You can see the tracking for the headline rates on the EMRS website here but bear in mind the red and green lines are likely to be much further apart in reality than shown.

The EMRS headline rate shows Labor on just 22%. This is Labor's equal lowest headline figure in this term of government; tied with the 22% in August 2011. But when that poll was taken the Liberals were polling a massive 55%; they are now down to a mere 49, with most of the difference going to the newly established PUP. The drop of six points to the headline Labor rate after two consecutive 28s is probably party sample noise; on the other hand, PUP have to get their votes from somewhere, and the rise in the Green vote is consistent with a similar rise in the ReachTEL.

Also partially released today was a National Party commissioned ReachTEL of the three northern seats with a total sample size of 647. The released results were Liberal 52 Labor 26 Green 13 Others (including PUP) 8. The National Party reckoned it might get a big share of the Others action but there's not much cause for optimism there, and nor should they get too carried away with a poll finding that 40% of voters had not ruled out voting National. The experience of various micro-parties at the Senate election showed that findings of that sort are no guarantee of even 1% of the vote.  I've not yet included this poll in my aggregate (full data being a minimum requirement to do so) and doubt it would make much difference to anything anyway.  Probably there are enough state polls in the mix not to need to use data from commissioned polls.

As usual I tweak the EMRS figures to try to get something that I believe is more realistic. Based on past election results the Greens can only count on their locked-in vote share and not on voters leaning to them, and certainly not on “undecided” voters. So I have allocated them 16%. PUP, on the other hand, showed in the federal poll that they can run under the radar of polling before building with a massive last-week spending blitz. On that basis I'm considering their headline vote of 5% to be fair for the time being. The recent federal Nielsen poll, which showed “Independent” on 5% although that category polled less than 2% at the election, suggests that an “Independent” voting intention is an ambit claim; we still don't know yet whether there will be any significant Independent candidates. I've therefore clipped that category and lumped it with generic Others. All that remains is to reallocate what is left over between the major parties. I have been in the habit of splitting it 50:50 but I think that with PUP explicitly named in the poll, those opting for “Independent” are possibly, on balance, lefties, so I've thrown Labor an extra half a point.  

Even so, the poll points towards a very low support level for the incumbent government.  The disappointing aspect of this sort of result is that Australia now has a new federal Coalition government which voters are so far only lukewarm in their approval of.  In the past there has been a history of state results usually favouring the party not in power federally.  But if there is going to be any backlash against the new federal Coalition in this state - the one thing most likely to prevent a Liberal majority government - this poll does not show any sign of it.  The general picture of a government that is too much weighted down by 15+ years incumbency, a poor state economy and an alliance with the Greens for anything else to make much of a difference continues.

A Liberal vote in the very high 40s, if it occurred at the election, would result in majority government. We have consistently seen an uneven distribution of Liberal support with much stronger levels in the electorates of Bass, Braddon and Lyons than in the other three. If the Liberal vote statewide is over, say, 47, then it is very likely they will poll 50 in these three electorates. While it's possible to poll 50% in an electorate and only get two seats because of leakage, the presence of a significant fourth party (PUP) makes this much less likely. The classic leakage scenario (as seen in Bass 2006 when Labor polled 49.6% and only got two elected) involves one of the other parties just over two quotas and the Greens not far below one. If a fourth party is polling even three or four percent (with preferences that partly exhaust and mostly scatter) it becomes very much harder for both the second and third parties to get up.

A PUP vote of 5% is not enough to win any seats anywhere. PUP would probably need about 12% in a seat to have a real chance and they are not there yet. But they are at least on the radar and seem to have some level of state-based support that they may build on. At this stage, those who voted for them at the federal election are not universally rejecting doing so again at state level. That said, not a lot can be read into a result of 5% with an effective sample size of 907.  Taking into account the uncertainties of how "undecided" votes might fall, the real support level for PUP could be 3 or even 7.  A reservation I have about PUP in Hare-Clark is that they may struggle to build vote share for "gun" candidates in each seat.  They may suffer badly from leakage from their ticket if their votes are too evenly scattered.  Strategically, they need to pick a suitable lead candidate in each seat and promote that person very heavily.

The poll does not provide a lot of new information about a possible seat distribution. It does overall point to Labor being slightly weaker and the Greens slightly stronger, and that points to slightly better chances for the Greens in Lyons and Bass, and a slightly more difficult task for Labor in Franklin.  My reading of the most likely seat outcome if this poll was repeated at a state election is 14 Liberal, 7 Labor, 4 Green.  However, the sample size is small compared to the recent ReachTELs, making that a very unreliable reading (especially without electorate data) and the poll is not different enough from other polling to move any seats in my aggregate of state polling.  And here that aggregate is (with the current poll weighted at 30%):

PUP now appear in the aggregate, though they're not yet polling enough to win any seats.  They're not too far away in Braddon though.  I've formed the PUP figure out of the projected Others (including PUP) total by scaling the state PUP total to the 2013 PUP Senate vote.  (This is likely to work better than using the Reps vote because of the impact of Andrew Wilkie on the PUP Reps vote in Denison).  Probably the Other figure left over for Braddon is too low, but it is unlikely to make a difference. 

Lastly, this poll does show a 4-point rise in Lara Giddings' Preferred Premier vote despite the drop in her party's support.  This cancels out the oddity in the previous poll in which she lost seven points while her party remained static.  I'm still not at all sure what all that was about (assuming it was not just a statistically rogue result), but whatever it was, it seems gone.

Note that updates to this article may be slow as I am currently in the process of transitioning to the NBN, among other chaos.
Added 7 Dec:  Poll Bludger Tassie tracking:  The Poll Bludger Tasmanian poll tracking has been updated here and a glitch caused by Wilkie-factor contamination of federal comparisons has been cleaned up.  According to the Poll Bludger tracking the major parties have been more or less stable in their support levels for over a year and the main variation is in the Green vote, which seems to have rebounded slightly after a big dip in the middle of this year.  Overall the PB tracking currently has a very similar position to my aggregate above.  What is interesting is that despite comments about the Liberals losing support, the PB tracker reckons they were never above 50 on any stable basis to begin with (polls showing them at 55 being significantly house-effected.)  

Zucco Quits PUP: Franklin federal candidate Marti Zucco, who was expected to be the Palmer United Party's Denison candidate, has quit the party in a very high-profile fashion.  Zucco started with recorded voicemail messages from Senator-elect Jacqui Lambie threatening to cut him off from party support after conflicts over internal issues.  Then he followed it up by "leaking" (translation: giving to the media) an email from Queensland state PUP MP Alex Douglas that referred to Lambie as symptomatic of "Boganland" and suggested that the PUP higher-ups were doing their best to train her (Pygmalion-style?)  On the face of it these are very damaging revelations about how some in the party might view its own voter base but it's hard to know how (or even if) this news will be received by likely PUP voters in the party's best state seat, Braddon.  (Incidentally PUP voters in Tasmania leaned slightly to Labor over the Coalition on a 2PP basis at the federal election.)

The usual MO of Marti Zucco is to continually build profile by becoming involved in incidents, thus guaranteeing his Hobart City Council seat (chances of him rising higher are perpetually slim).  In this case he has mined an especially rich vein.  

Holy Nauseating Netsats, Batman!  The release of net satisfaction ratings for Tasmanian politicians is a deplorably rare event, which is the main reason I'm reporting such crumbs as these.  The Liberals  commissioned EMRS to poll satisfaction ratings for various politicians, at least some of them statewide, and then selectively leaked gave them to the Advocate.  Figures released are McKim (Green, statewide) -27 (favourable 25 unfavourable 52), Bryan Green (ALP, Braddon) -33 (17-50), Brenton Best (dubiously ALP, Braddon) -25 (14-39), Paul O'Halloran (Green, Braddon) -19 (16, 35).  Two unnamed Labor MHAs placed between Best and O'Halloran, either within their electorates or statewide (it is not clear which).  I expect the Braddon samples were very small, especially if this survey was part of the recent EMRS polling cycle.  While selectively released internal poll results should always be treated with very great caution, there is nothing unbelievable about these figures. However, they can't be translated to chances of winning or losing seats, since under the Hare-Clark system, even if you are unpopular, if all who approve of your performance vote for you, you still get in. 

Some comparisons for past Green leaders at state level are available here.  Bob Brown polled similar Newspoll ratings to these in the early 1990s. 


  1. I suspect things have moved far enough to one side in the course of 15 years that a correction period is needed here as much if not more than it was federally, which is reflected in the data. Also, most of the negatives happening federally have been fairly transient in nature.

    Odd that the National's are running a poll. Tasmania isn't really their normal market; not really remote enough. Hope springs eternal?

    It has to be deeply concerning to Labor that they do not have a single seat with two full quotas. It will be interesting to see how much attention Labor gives during the campaign to attempting to claw back that second seat from the Greens rather than from the Liberals ala Brenton Best.

  2. I would not be so pessimistic if I were Labor. The momentum is certainly not with the Liberals who have dropped from a high of 55% to Reach Tel's 52% and now to EMRS's 49%. In 2010 the poll immediately before the election had Labor on 22%, but they ended up with 37% on the night. I sense that there is certainly a disaffection with Labor but little enthusiasm for Will Hodgman and his team. The Liberals should be coasting home by now after all the Labor shenanigans of the last seven or eight years, but the Liberals are struggling to maintain 50%. I suspect that come election night much of the undecided poll with stay with Labor, taking them towards 30%. This will be helped in Franklin by Julian Amos, an 'old Labor' candidate who will hold traditional Labor voters who haven't liked the Green coalition.

    The other oddity is Hodgman's pledge not to work with anyone else should he fail to win a majority. This means all Labor the Greens (and PUP) have to do is prevent the Liberals winning their majority and they can all look forward to another 4 years of coalition government.

    The last issue worth mentioning is that the Liberals are not very good at election campaigns, and Hodgman particularly does not come across very well in front of the cameras. This, added to Tony Abbott's apparent enthusiasm for dumping on Tasmania, might cause the Libs further difficulties in convincing Tasmania voters to give them a go after 14 years in opposition.

  3. It's worth a close look at the figures from 2010. Feb 2010 had baseline figures on the old EMRS baseline of 30 Lib, 23 Labor, 22 Green, 2 Ind, 23 Undecided. The figures on the same scale now are 41 Lib, 18 Labor, 16 Green, 4 PUP, 3 Ind, 17 Undecided. On the scale most comparable to the current headline rate (49-22-19 etc), the Feb 2010 poll was 39-31-27-3.

    There was then a final (March 2010) pre-election poll with 29 Lib, 21 Labor, 22 Green and a ridiculous 26 undecided on the old headline rate. On the scale most comparable to the current headline rate the March 2010 poll was about 38-33-28.

    The final result was 39-36.9-21.6, so on the rate comparable to the current headline they had the Liberal vote more or less perfect in their last two polls, ALP 4-6 points too low, Green 5-6 points too high.

    In 2010 the "undecided" voters did seem to go more to Labor than Liberal, but less strongly than in 2006. I'm expecting them to split more or less down the middle this time, but yes, if they repeat the split from 2010 then Labor could get around 29.

    Julian Amos is running in Denison.

    As for the no-minority pledge I suspect some way will be found around it if needed so long as there is a pathway to government that avoids the Greens.