Image: J
EMRS: Lib 52 Labor 27 Green 18 Ind 3
Interpretation: Lib 53 Labor 29 Green 15 Ind 3
Outcome “if election was held now”: Liberal Majority
The February 2012 EMRS poll has been released. In recent months I’ve been reluctant to discuss this company’s output at length, firstly because some serious concerns about their polling methods have still not been answered, and secondly because the polls have been saying little that was not already obvious. This poll, however, deserves discussion, not least because it had the potential to weaken the Premier’s hold on her party leadership whether it was accurate or not, but it has not in fact done that.
Lara Giddings can perhaps breathe a sigh of relief that from Labor’s perspective this poll was only routinely awful. In May 2008 Paul Lennon famously polled an EMRS Preferred Premier rating of just 17%, and quit within weeks of that poll. The main reason Lennon’s rating was so bad was that the proportion of voters undecided about the premiership increased, as Labor voters preferring a change to their great white hope David Bartlett (remember him?) refused to back Lennon anymore. It is unknown to me whether the bad poll forced Lennon’s hand or simply reflected his underlying reasons for departing anyway, but that little bit of polling history points to something that could have happened this time, and did not.
Recently, Giddings was the target of leadership speculation based around the idea of a shift to David O’Byrne, although O’Byrne himself admitted only to long-term ambitions, and it is unclear now whether the push was ever really backed by anyone much beyond Kevin Harkins, his mates and a few fairly bored journalists. If there was strong demand from Labor supporters for a new Premier any time soon, this would have been reflected in an erosion of Giddings’ support into the undecided column. With Preferred Premier readings of just 19 points in August and November (essentially reflecting her party’s very poor standing), there was the potential for Giddings to poll something really awful, like, say, 12, which would have made her position untenable. Instead, she is actually up five points to 24. About three points of this change is down to (possibly random) differences in party standings between the two polls. It is not important whether Giddings’ standing relative to McKim and Hodgman has gone up significantly or not. What is important (in a rare case of PP actually telling us something useful) is that it has not crashed, and that the remaining Labor supporters do not seem to want Giddings replaced.
EMRS have changed the order in which they display their results, relegating their raw figures with those massive undecided rates to the second table and making the “Percentage of Respondents Supporting or Leaning Towards A Party after Excluding Undecided Voters” their new headline figure. The public hence are no longer required to make sense of ludicrous suggestions that less than one voter in five still supports the Labor Party (except if their preferred news source doesn’t know a gift horse when it sees one), but the cost is a headline rate that will overestimate the Greens’ support. The reason for this is that election results have consistently showed the Greens do not poll above the EMRS “base” vote and that redistributing undecideds to them is too generous. In my “interpretation” figures at the top of this article I have cut the Green vote back to its “support plus leaning” level (even that is probably too generous) and also docked the Libs a token point because of the possibility that the survey method issues I mentioned six months ago are inflating their standing a little.
The most useful figure for assessing base changes in party support, the “% Support + % Leaning” figure (now table 3) shows Labor up three points and the Greens and Liberals each down two. These changes are not statistically significant, but what is notable is that on any measure the Greens are about where they were in August 2011. Nick McKim has also noticed this, but what he doesn’t notice in his attempt to put a positive spin on the figures is that where they were then was the bottom of a rather large slippery dip. The Greens lost several points of support between February and August 2011 and while it hasn’t got any worse for them since (yet), they have also got none of it back.
The EMRS heading “Tasmanians show little faith in leaders” can very safely be ignored since the poll proves little about whether individual leaders are liked or disliked, and if EMRS want to assess that issue they need to start asking approval ratings questions, which would be far more useful than their obsession with Preferred Premier. A high-teens undecided rate in EMRS’s preferred-premier polls is business as normal for them except around election time, and the reason it shrinks around election time is that people are more engaged with politics – not that they are more supportive of politicians. To illustrate there is nothing to see in this heading, here are the “None of the above” ratings since the last election: 12 (May 2010), 17, 17, 19, 19, 16, 19 and now 18 (Feb 2012). Apart from the figure being lower just after the election there is no trend on that figure at all and I am confident that if major pollsters were polling the state they would get figures that were lower.
An article in The Examiner translates the results as potentially 14-8-3 in the current parliament. The article has attempted to extrapolate from a single round of EMRS’s very small seat samples, which is not a good idea at the best of times, though if we do make the multiply foolhardy assumption that the EMRS electorate samples are accurate then their read of those samples in terms of likely seat outcomes isn’t too far out.
As is my usual custom where two straight polls record similar results, I have simulated results by rolling the two samples together, rounding in favour of the more recent one, and redistributing undecideds away from the Greens to the major parties (but as a new tweak, in Denison I now redistribute half the “Green” undecideds to “Independent”, since if Andrew Wilkie were to be out of Federal Parliament and a candidate at the next State election I suspect he would get those and more.) The result of these samples gives the Liberals between 13 and 16 seats in the current House, Labor between 5 and 8, the Greens between 3 and 5 and some possibility of one Independent. Bass splits 3-1-1 (Liberal-Labor-Green), Braddon 4-1-0 or 3-1-1, Denison any of 3-1-1, 2-2-1 or 2-1-1-1, Franklin 3-1-1 or 2-2-1 and Lyons 3-1-1 or maybe 3-2-0. In this simulation the Greens are actually just shy of their quota in Franklin, but would hold the seat since the major parties would be affected by leakage from their leaders while the Greens would not.
That of course assumes the House has 25 seats at the next election. With the possibility of 35 back on the table, that needs to be examined too, and the outcome of the simulation under the 35 seat system is spectacular: 21 Liberals, 8-9 Labor, 4-6 Greens and one Independent. If a real landslide is on at the next election then a switch back to 35 seats might even blow out the seat margin, mainly because the 35 seat system would turn the easy threes the Libs are polling in the northern seats into fives. Labor’s support has collapsed so badly in the north, much as it did in the eighties and early nineties, that even adding two seats to each electorate there might not save them from seat losses.
Nonetheless I will not be surprised if Labor and the Greens do make a serious attempt to take the House back to 35 seats. An obvious advantage for Labor (especially if the claimed leadership tensions are actually real) is job security for its new MHAs, who don’t really want to have their careers snuffed out after a single term – not that a long time in Opposition is much to look forward to either. Another advantage is that should the gap between the parties somehow tighten significantly before the election, there would be more chance of the Liberals falling just short of a majority in the 35-seat system. Furthermore, if the Liberals fell just short in a 25-seat house, it would be a farce for Labor and the Greens to try to go on governing with just 13 seats between them. 18 seats in a 35-seat system is a different matter; indeed, it is more than they have between them now.
Of course the value of a simulation is only to suggest what sort of result would occur if an election was held right now, and if that actually happened then the fact of it happening so soon would in turn affect the outcome. What we are looking at here only has predictive value should Labor and the Greens make up no ground. But there is nothing remarkable about the percentage breakdowns now being claimed – indeed, they are almost exactly the same as the result of the 1992 state election, following the last time Labor and the Greens tried to share power, and that was a time when Labor was not weighed down by well over a decade in office. And even this far out (or is it?) from the next election, an easy Liberal win with over half of the primary vote is in my view a rather likely outcome.
This poll taken between the 6th and 10th of February may one day seem a relic from a different age. It does not tell us how voters will react to a potentially major event since – the statement by 12 of the 15 Legislative Councillors that they intend blocking the creation of new reserves under the forestry “peace process” unless activist market attacks on forestry company Ta Ann cease. Although this LegCo action is not known to be of the Giddings Government’s making as such, it has been a flashpoint for an remarkable escalation in critical and even angry rhetoric between some Labor figures and the Greens.
Forestry disagreements have historically caused minority governments in Tasmania with the Greens as the crossbench to become unstable, to the point that I’ve in the past argued that instability is the normal condition of Tasmanian minority government, in a way in which it is not in European democracies. A market downturn that appeared to have parts of the industry looking for co-operative partial exit strategies was a great blessing for this Labor-Green coalition in the early days, because it allowed them to work together on other issues productively and stably with forestry no longer holding centre stage.
But now with the peace-deal process under strain (to say the least), and some industry players seeking exit strategies from the proposed exit strategies, it is not clear whether Labor and the Greens can avoid having their relationship dominated by bad feeling and accusations of poor form over forestry issues, and there is pessimism about whether the Government will see out its full four-year term. Perhaps the events of the last few weeks will turn out to be just minor flashpoints, but I doubt it. This is starting to look a lot like a rerun of a movie we watched 21 years ago, with Lara Giddings cast in the improbable role of Michael Field. You don’t need opinion polls to know how movies like that often end.
