Monday, March 16, 2020

EMRS: Liberals Still Ahead Under Gutwein

EMRS March 2020: Liberal 43 Labor 34 Greens 12 Others 11
Interpretation scores not used because of change in EMRS methods
Result "if election held last week" on these raw numbers 13-10-2 (no change), next most likely 13-9-3
Better Premier White 41 leads Gutwein 39 - similar situation to White vs Hodgman

A new EMRS poll of Tasmanian voting intention has been released via the unusual avenue of a FontPR podcast.  (Recommended listening for a lot of new insights into how EMRS operates). This is the first EMRS poll to be publicly released since the July 2019 poll, but in the meantime there were two other polls that were not previously released, in November (very early November - pollster was reported in the field on 31 Oct) and unusually December.  These have now also been released (full report of last three polls here.)  It is unfortunate the earlier polls were not released at the time as they would have usefully informed discussion about the retirement of Premier Hodgman.

The podcast reveals that EMRS have made significant methods changes in recent months, including ensuring at least 35% mobile coverage in their phone polls and making changes to weighting (though this does include the rather risky inclusion of past vote - a partial cause of the 2019 polling failure, but perhaps more justified in Tasmania where it is harder to be confident a sample is even close to representative.)  Based on the late-2019 polls together with this one it looks like this has fixed the pollster's long-standing problem of severely overestimating the Green vote, and at times the Others vote as well.

It's been my habit to adjust EMRS polls for modelling purposes by taking a few points off the Greens, but because of the methods change it appears this is not still the right thing to do and I am just treating the results as they are.

Compared to the July poll this poll shows the Government up five points to 43, Labor up four to 34, the Greens down four to 12 and other/independent down five to 11.  These changes have been largely caused by the methods change.  The results of the November and December polls were similarly weak for Greens and Others, and in fact showed the Government with larger leads than now. The Government led 47-29 in November and 44-31 in December.  The 47-29 lead in November is very different to what we are used to seeing from EMRS, and had the government up nine points on the July poll and Labor down one.

Mostly this would have been caused by the methods change (which may have also corrected EMRS's habit of under-polling sitting governments) but even so it is a striking change from the previous poll.  Perhaps Labor suffered temporary damage over the Madeleine Ogilvie saga, as a result of which one of its seats became an independent seat that provides support to the Government, rendering the Opposition impotent.  The poll does not show any lift in the Government's fortunes as a result of replacing Hodgman with Gutwein (in fact the lead is cut from 13 points to 9), but neither does it show a statistically significant backlash.

Although the Liberals' 43% is well down on their state election result, they would still be likely to retain majority government.  Here is a possible breakdown between the five electorates:

I have not allocated any seats to Others because no others capable of commanding most of the vote in any seat are known to be running.

On these numbers there would be expected to be a close race between Labor and the Greens in Bass.  However because Labor now have two incumbents and the Greens have none, and also because of adverse "others" preferences for the Greens, Labor would probably win it.

In Clark, the Liberals would come close to losing their second seat to Labor, and this is capable of becoming a loss if the Sue Hickey situation causes carnage - but even so, Labor would be disadvantaged by taking only one incumbent to the election.  The scenario is not a very promising one for Hickey winning as an independent either, as she would need to poll 10% to have much chance.   The remaining seats on these numbers are not close and the most likely result would be the status quo despite the sizeable swing against the Government.

There is not much to see with the Better Premier score either, where Rebecca White leads Gutwein 41-39, compared to her leading Will Hodgman 43-42 in December.  Gutwein is not yet established enough for the comparison to be meaningful, but if he was his numbers might be boosted by honeymoon effect anyway.  Overall this result is more of the same, and it's long been clear that White is an outlier in preferred-leader polling.  Normally if an Opposition Leader leads on such polling the government is dead in the water, but in White's case she seems to be very popular personally without that popularity translating to votes for her party.  There is some thought that Labor need to do much more to convert White's popularity to votes by making more presidential use of her appeal.  There may be another in dark socially-distanced corners that says that if she was going to take the party above 35% it would surely have happened organically by now.

With the coronavirus crisis gathering pace in Tasmania as it is everywhere else, this poll is already a snapshot of an older and simpler time (though even last week 9% volunteered coronavirus unprompted as an issue of concern, compared to only 4% for climate change).  The second half of the parliamentary term is likely to be far more consequential for the 2022 election than the first.  If the Government can manage to ensure Tasmania gets off relatively lightly in terms of both health and economic impacts then it has great chances of a third term.  However, as politicians the world over are finding, this virus is no ordinary opponent.

No comments:

Post a Comment