Tuesday, May 26, 2015

EMRS: Liberals Regain Ground And Would Be Returned

EMRS: Liberal 46 Labor 29 Green 19 Ind 6
Interpretation (provisional): Liberal 47 Labor 32 Green 17 Other 4
Seat distribution based on poll: Liberal 13 Labor 9 Green 3
Aggregate of all recent state polling: Liberal 13 Labor 10 Green 2

A new EMRS poll of Tasmanian state voting intention was released earlier today.  The headline figures show the Liberals up four points since February, Labor down five, the Greens up four and the remainder down 3.  These changes are statistically significant and suggest strongly that the Liberals have recovered ground since last year's slump (see previous instalment)

They also provide further evidence that a series of Morgan samples (see Wonk Central: Morgan's Tasmanian State and Federal Sampling), that have shown the major parties roughly level, are very strongly "house-effected" in Labor's favour and are not reliable.

EMRS tends to favour the Greens at Labor's expense and as a result the 19% vote shown for the Greens is not reliable; even the 17% that I have adjusted it to as an interpretation score probably reflects a good sample compared to how the party is actually travelling.  The poll was taken from 19 to 22 May (Tuesday to Friday last week) and thus most of the sample was taken after the announcement of the resignation from parliament of Greens Leader Kim Booth.  It's possible there would be a goodwill bounce for the Greens on account of that result.

I am not going to post a detailed breakdown of the electorate results if this poll was taken literally, for the simple reason that we should know by now that EMRS polls with the Greens on scores of 19 are not to be taken too literally, especially not when they have been polling at a level well below that.  For the record if the poll is interpreted literally, then based on a uniform swing from the last state election the Greens would win a seat in each electorate.  The Liberals would as usual lose their fourth seat in Braddon and their third seat in Franklin, and the key seat would as usual be Lyons.  Based on a uniform swing the Greens would have just below a quota in Lyons, with the Liberals (2.80 quotas) just ahead of Labor (1.76) for the final seat.  With a fairly even split of votes between the three Liberals they would be expected to get three seats on those figures, though it is easily within the poll's margin of error (especially considering the lack of recent electorate data) that they might not.  So, probably 13-7-5, but we shouldn't take that projection seriously.

Projection based on adjusted figures

To project this poll based on adjusted figures I have used my usual method of assuming the Greens get no more than their vote prior to distributing Undecided, and splitting Undecided between the major parties.  I have also docked "Independent" one point as this classification continually overestimates the actual Others vote, with contaminants like Legislative Council elections and Senator Jacqui Lambie often confusing voters.

Even this assessment, giving the Greens 17% is probably generous and perhaps it should be a point (or so) lower, but let's see how it goes:

The distribution of the Others vote here is a bit dodgy because Others at the last election were mainly PUP, but PUP are apparently no longer running in state elections (and their vote has collapsed anyway) so the distribution might be different in practice.

There are two electorates worth discussing in detail here;

1. In Braddon the projection shows Labor just ahead of the Greens.  However there is a high projected Others vote in Braddon (see the comment about PUP above) and in practice either Labor would poll higher or preferences from Others would favour the major parties over the Greens, so the second ALP seat there would be clear-cut.

2. In Lyons it looks like the Greens are very competitive, but they would have no incumbent to the Liberals' three and would hence suffer on leakage and probably from the even split in support for the sitting Liberal MHAs.  Also, Others preferences in this electorate were very unfriendly to them at the last election.

While it is not certain that the Liberals would win an election "held now" on these figures outright, it is highly likely.  And in any case, it's an artificial exercise.  If the Liberals went into an election campaign with a 15-point lead, it would be readily apparent to voters that only the Liberals could win and that anything else would mean a hung parliament.  Most likely this would swing votes back the Government's way, as it did for Labor in 2006.


I have been maintaining an aggregate of Tasmanian polling since the last election.  With the now-total demise of PUP as a force I have carried out some complex changes to the Others vote to reduce the influence of its distribution at the last election.  Apart from that the current sample carries a weighting of 40% as, although it is a modest sample, it is the first new state data apart from the dubious Morgans for three months.

It would be nice to get some new electorate-level data, especially given that the final seat in Bass persistently comes out as the crucial factor in whether the government retains office.  Note that the relatively poor standing of the Greens in Lyons compared to Bass is based partly on the seat distribution from a single ReachTEL last year, which is the only electorate-level data that we have.

Anyway this is the current aggregate:

The Greens continue to have a problem in Bass because Labor's recovery causes them to lose their seat even if their vote is unchanged or increases slightly.  This will be exacerbated by going to the next election with a relatively little-known incumbent following Kim Booth's countback.  All the electorates are straightforward in this edition of the aggregate except for Lyons in which all parties seem about even in the race for the final two seats.  However based on not having an incumbent and the unfriendly distribution of preferences in this seat, the Greens would be the one to miss out (it's also possible the aggregate has their support too high there anyway.)

The aggregate is therefore unchanged at 13-10-2.

Preferred Premier

This not-terribly-useful indicator again shows Will Hodgman with a big lead over Bryan Green, 52-24 compared to 48-26 in February.  Preferred leader scores are heavily biased to incumbents, especially when their opponents are relatively new, and there is not a great amount to see here.  Possibly this poll will generate some speculation about the Labor leadership but there is really not a lot to see here.

Causal Factors?

It's always hard to say why polling is moving the way it is in Tasmania, when useful poll readings are so far apart and have fairly small sample sizes.  I suspect the main cause is that the new government is starting to put distance between itself and some of the controversies of its first year. Voters may also believe the state's economy is now improving.  Really there has not been a lot going on in state politics in the last few months that should obviously impact voting intention.

So far I have not seen any really dodgy spinning of the results of this poll, but in all it would be a reassuring result for the Liberals (while still indicating they have work to do to protect their third Franklin seat), and a disappointing result for Labor.  For the Greens there are positive signs but the raw figure is another mirage and the departure of Booth may have affected the result.   It's still a very long way to go til the next election.

Morgan Update (27 May): 

A new Morgan sample is out showing the Liberals leading 41 to 35.5 with the Greens on 19.5.  As this article has noted, the Morgan samples appear very heavily skewed to Labor and the Greens, and with that taken into account the Morgan result is probably not vastly different from EMRS's.  Shortly I will update my assessment of Morgan based on these new results (it won't change much), and then my aggregate as well.


  1. Was the Hodgman government particularly unpopular, or was it simply the federal fiasco dragging them down?

  2. Likely elements of both. The federal government was already recovering by the time of the previous EMRS, and the Hodgman government also polled badly in November.

  3. This is a great poll because it leaves egg on the faces of those political journalists who are making baseless commentary about the likely demise of the Greens. As long as Labor continues to give the impression that it prefers to sit on the fence on contentious issues, the Greens may continue to poll reasonably. At the same time, it is unbelievable to me that the Liberals are up.


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.