EMRS: The Calm Before The Scandals (Which Are Probably More Calm)

EMRS: Liberal 42 Labor 30 (+1) Greens 13 (-1) IND 13 (-2) (likely to be inflated) Others 3 (+2)

Seat estimate if election "held now" Lib 16-17 ALP 11-2 Green 4-5 IND 2-4

An EMRS quarterly poll of Tasmanian state voting intention taken Feb 14-19 has just emerged, together with an update on the snazzy albeit slightly odd bells and whistles thingy (more on that later).  

This poll was taken before this week's resumption of Parliament in which the government has been attacked over two political class scandals.  The first involved conflict of interest accusations against the Speaker Mark Shelton concerning very large grants for the upgrade of the Bracknell Hall.  Speaker Shelton has used his casting vote to block a referral of the matter to the Privileges Committee, to block a no-confidence motion against him from being brought on, and to block two rather unusual motions that his vote on those motions be disallowed on the grounds of "pecuniary interest".  (I have been unable to find any precedent for pecuniary interest provisions being used in this way.)  In the second, after one of the parliament's longer exercises in dental extraction there was finally an admission that Racing Minister Madeleine Ogilvie had known the former TasRacing boss Paul Eriksson had been dismissed days before she first misleadingly claimed he had moved to Sydney for family reasons.   At times Labor and the Greens were almost tripping over each other trying to work out which scandal to target first.

When polls pre-date parliamentary or similar scandals there is often an assumption that this renders the poll obsolete, but voters don't react to most scandals even a tenth as much as avid politics-watchers might think or wish they would.  Now and then one strikes a collective nerve and has a major impact, but the vast majority are undetectable through the polling rear-vision mirror.  It may be that in another three months the voting intention impact of this week's events will have been about the same as the impact of anything else in the last year, or in other words nothing.  

For the run of EMRS polling since the reopening of borders over the 2021-2 holiday season has the state in a holding pattern with no significant movements: the Liberal Party with a substantial but not majority-winning lead and Labor unable to break out of the low 30s.  If things were to stay that way then the Liberals would go into an election as the only party capable of winning majority government.  Either there would be a 2006/2018 style bandwagon to avoid a minority parliament or there would be some kind of no-majority situation, perhaps one in which for a change the Greens did not have the sole balance of power. But all that is a very long way away.

The poll somewhat surprisingly sees Jeremy Rockliff's lead over Rebecca White as preferred premier shrink a little to 44-36, which is not a very big lead given that this indicator routinely skews to incumbents.  It would be much more interesting to see individual approval ratings for the two leaders.  Rockliff has historically been popular with a very moderate, nice-guy cross-party appeal, but is this still the case especially following his pursuit of an anecdotally disliked AFL stadium proposal?  Rebecca White has at times been very popular (which often used to contribute to unusually competitive Preferred Premier ratings for an Opposition Leader); is her personal popularity recovering?  

Voting Intentions Graphic and a model

The EMRS voting intentions tracker page is fun to play around on.  As a bit of back story here, in the distant past EMRS used to routinely publish the electorate breakdowns of voter support from their polls.  These were too volatile to be that much use in isolation (c. 200 per electorate per poll) but had uses when a series of them were aggregated.  Eventually EMRS stopped releasing them, I think because the Examiner wasn't paying for them anymore.

The new graphic allow for breakdowns of results for the last three years by age, gender and electorate.  However the breakdowns are rather odd in that they show the contribution each component makes to the overall total.  So, for instance, the Liberal vote in the 18-34 bracket being 10% means that 10% of all voters are Liberal voters aged between 18 and 34. This age breakdown is interesting because in the current sample the Liberals are doing better in the 18-34 age group, and worse in the 55+ age group, than in any poll in the last three years, but we will need to keep an eye on more polls to see if this is a thing or just sample noise.

The breakdowns by division allow me to recover voting intentions per seat, but the problem here is that on top of the inaccuracy of small samples, the use of a percentage of all voters means we only get a rough idea of the outcome for each seat.  So, for example, if 11% of all voters are Liberal voters in Bass, then this means the Liberal vote in the Bass sample should be somewhere between 52.5% and 57.5%, provided the number of voters per division is more or less even.  As a result the electorate by electorate breakdowns are even more rubbery than they used to be.

Because I could, I decided to attempt a model of state voting intention based off the current EMRS primaries and the divisional breakdowns for the last four polls.  This involved a fair amount of fiddling around adjusting numbers to make everything sum to 100 and a more precise algorithm would get slightly different results, but anyway here's what I got:

At the last election, fourth-party and Independent support ran highest in Clark and Braddon, but EMRS are finding rather high levels of support in divisions where such forces didn't poll anything to speak of in 2021.  This is the usual problem with including "independent" on readouts when there may or may not be any independent who the voter has heard of running.  But there is a special question at the moment regarding the status of Franklin MHA David O'Byrne.

O'Byrne is a member of the broader Labor Party but was kicked out of the parliamentary caucus after being #metood as soon as he assumed the leadership in mid-2021.  He has generally been listed on the Labor website as a Labor MP (with at least one brief and mysterious interruption) but has been treated by the Parliament as an independent for purposes such as Question Time allocation.  The issue of returning him to caucus is divisive within the party; there is pressure for it to occur sooner rather than later but also some MPs appear to prefer that it does not occur at all.  

It is not clear whether Franklin voters think O'Byrne is an independent or Labor when they are answering EMRS polls.  If O'Byrne is not readmitted he could well run as an independent though he would probably need to hold well over half of his 2021 support to win.  O'Byrne's workrate as a semi-independent has impressed some observers and I am treating him as a potential independent winner until his fate is clarified.

In the above model I have given a seat in Braddon to the independent Craig Garland, who polled 6% in the 2021 election and 7.8% in the federal election (the latter despite adverse publicity over vaccine views, which I suspect nobody in Braddon much cared about).  The last seat in Clark is a mess and would depend on what independents contested.  

Resizing Regret Test

Overall this is a very similar seat estimate to my previous article except that it is more favourable for independents.  The interesting thing though here is the "regret test" factor following the impending restoration of the 35-seat system.  When I apply the above model to the recently repealed 25-seat system, the Liberals retain a majority, holding their two seats in Clark with 1.9 quotas and beating the Greens from behind in Lyons via minor party preferences and candidate-splitting effects.  In my model of the addition of ten seats here, the Government is probably only picking up three of them.  The 25-seat rerun of the next election will be fascinating.

Between now and the next poll we will see LegCo elections for Launceston, Murchison and Rumney.  Of these Rumney is a known two-party contest which may provide some further insight into how the government is going.