Dr KEVIN BONHAM

The succour for the anti-pulp-mill forces in this apparent disaster is that anti-mill voting is extremely strong in areas where the Green vote is already high. If this translates into middle-class suburbs near the city in 2010, which I think it will, then the gains to the Greens in those areas should be enough to ensure the survival of their critically endangered MHA, Kim Booth.


THREE unusually exciting Legislative Council elections have been run and won. The pre-poll musings Peter Tucker and I co-authored can be seen here and herehere. As expected, Michael Aird won Derwent with an absolute majority, but a little less emphatically than had been thought. As expected, the pulp mill factor fizzled and a low-key Kathryn Hay campaign was not enough to stop Ivan Dean getting home on preferences, less comfortably than his 2003 landslide, but in the end not all that close. And, as not expected, the small-town mayor of Latrobe thrashed the city mayor of Devonport to such an extent that it was actually a long-term single-issue political battler who ran an impressive if distant second. So what does it all mean?

Before I get out the microscope (and a warning, this article is long, is rated M and contains frequent statistical scenes) I’d like to commend the Tasmanian Electoral Office on their first-class performance in the prompt determination and online delivery of the results. As William Bowe notes, they could teach many bigger such offices around the country a few things about how to do it.

Derwent

We predicted 55-65% for Michael Aird (Labor), 20-35% for Jenny Branch (tokenly ex-Liberal) and 10-17% for Susan Gunter (Green). The actual figures, subject to final postals, which should stretch the incumbent’s margin, were 51.6, 33.4 and a shade under 15% respectively. On paper then, a somewhat poor result for the incumbent, a very good one for the challenger and business as expected for the Greens. However, Aird had a hideous media cycle in the last few days of the campaign, so adjusting for the few points that must have knocked off his margin, Aird’s result is at the low end of normal expectations, while Branch’s is still pretty good. Had Aird been forced to preferences it would have been a serious embarrassment in a seat which has gone to Labor outright every time for 30 years, and the most important thing for Labor is that that record has remained intact.

The Sunday Tasmanian wrote the less than crushing outcome up as a “scare” for Aird (which is nonsense) and made much of the 26% primary swing against him (irrelevant since he only competed against a Green in 2003). More usefully, it seems Michelle Paine put to Aird that the last time he faced two other candidates he polled 58%, and Aird’s response was that this was different because this was effectively a Labor, Liberal, Green election.

So if Aird’s right, why the discrepancy between this result and the last three-party Lower House contest in 2006 (at which, excluding other forces, Labor polled 66.1%, Liberal 23.4 and Greens just 10.5)? I would expect the Greens to perform better in a three-candidate election than a three-party election, because Greens voters generally look for the party name while more major party voters respond more to personality and profile, but the extra ten points to the de facto Lib who had hardly set the electoral world on fire at municipal level demands some explanation. Branch appearing as an independent on the ballot paper may be some part of it, but Derwent voters have shown little sign of caring about independence much before.

To see where any backlash against Aird may have come from, I’ve constructed a booth-by-booth comparison of the 2006 Lower House results for Derwent with the figures from the 2009 election.

As noted in our preview, Labor outpolled the Greens and Liberals combined at every booth bar three in 2006. Aird only broke 50 in about a third of booths, and was outpolled by Branch in Hamilton and Tea Tree, with Ellendale a tie. The Greens topped no booths but outpolled Branch in three booths – Lachlan, Molesworth and Maydena (the first two were the same two where the Greens outpolled the Liberals in 2006).

Aird’s primary was well down on Labor’s 2006 Lower House primary in every booth. The biggest differences were in Hamilton (36 points), Maydena (28), Abbotsfield, Ellendale and Roseneath (21), Austins Ferry and Pontville (19), Claremont (18) and Gretna (16). The smallest were in Molesworth, Glenora and Lachlan (6), Ouse (8), Gagebrook and Granton (9), and Brighton and Dromedary (10). Molesworth and Lachlan were the party’s two worst booths in 2006 so there was less support there to lose. The big differences occurred in two groups – firstly the Glenorchy City Council booths and secondly a range of small rural booths, however the pattern in small rural booths was inconsistent.

The booths with the biggest differences in favour of Branch were Hamilton (27), Roseneath (21), Abbotsfield (19), Claremont and Pontville (16) and Austins Ferry (15). Branch thus polled extremely strongly in the Glenorchy area, her local council patch where she has strong education connections. She was also strong in some rural areas.

The booths with the biggest differences in favour of Gunter were Maydena (16 points), Bronte, Dromedary and Ellendale (12), Hamilton and Tea Tree (10), Bridgewater (7) and Gretna (6). The least favourable were Roseneath and Glenora (minus fractions), Magra, Lachlan (<1, the latter already a very good booth for them) and New Norfolk North, Westerway, Claremont and Abbotsfield (1-2.5). On this basis (and also comparing Gunter’s differences with Branch’s) the following are apparent: • Glenorchy suburb booths were strong for Branch while differences in non-Glenorchy suburb booths were less pronounced and more favourable to the Greens. • Distant bush booths showed more movement than near-urban bush booths, and the direction of that movement was mostly Green. • Farming booths showed no consistent pattern. I have already explained the first; current forestry issues (eg the Florentine protests) may have something to do with the second. All up, as Peter comments in the Sunday Tasmanian and on Poll Bludger, there is no one strong message for Aird in the results and “no great baseball bat mood”. That said, much of Derwent falls in Lyons, which is the most critical electorate for Labor to hold if it wants to keep its Lower House majority, so even little baseball bats can hurt it. Branch performed respectably enough that she may attract the attention of the struggling Liberal preselectors for Denison (who have still to fill two places) if they think they can make useful inroads in Glenorchy. Michael Hodgman already pulls a large proportion of the Liberal vote in Glenorchy, but that is perhaps by default, because the party preselects so many urban lawyer/business types with little appeal in such suburbs. Windermere

Subject to further postals, Ivan Dean polled 39.2%, Kathryn Hay (“Independent Labor”) 26.8, Peter Whish-Wilson (Green) 16.2, Peter Kaye 9.8 and Ted Sands 8.0. Dean went on to win on preferences with a 55-45 margin. The result is unsurprising, except that Sands polled about as badly as could possibly been expected, while quite a few Dean voters parked their #1 with Peter Kaye. This, plus a surprisingly weak anti-pulp-mill factor in preferences, meant that Dean was able to win comfortably, despite polling a primary that for now is in the sub-40 danger zone. Indeed, Dean’s winning primary in this election is virtually identical with his losing 2007 mayoral primary.

It’s nothing like Dean’s margin when elected in 2003 (without a preference distribution from that one we won’t know, but perhaps a 6% 2CP swing to “Independent Labor”?) However, given the stronger field against him, the Gunns mill issue and the turbulence of his first term, Dean has good reason to be satisfied. The size of his victory can be attributed to making an effort to connect through the media with the mainstream of his electorate across a range of issues during the campaign. Kathryn Hay’s effort level was less clear. Whether she would have beaten him had her campaign been less of a mystery we’ll never know, but it would have been very close.

Ivan Dean topped every booth in Windermere but two. Kathryn Hay predictably topped her old home base in Labor heartland, Ravenswood, but only just – it seems Dean’s anti-hooning message struck a chord. And Peter Whish-Wilson (Green) is most likely the MLC of choice for the 449 electors of his anti-mill home base Hillwood, with 32.3% of the vote.

Comparing 2003 to 2009, one would expect swings against both Dean and “Independent Labor” because the field contained five candidates not four, with the Greens making a serious effort and boosted by a local issue. As it happened Dean’s primary was down 11 points, while the ILP’s was more or less unchanged. The booth by booth picture, however, displays spectacular variations.

Dean’s primary was actually up 0.7 points in the main booth of the officially pro-mill town (PDF link) of George Town and down only 7 points in the other. A solid booth for him was Waverley (down less than 5 points, albeit off a fairly low base), and the damage in Elphin, Inveresk, Newnham and Rocherlea barely made it into double figures. But the recycled baseball bats were out for Dean in Dilston (-22) and it also wasn’t pretty in the Liberal stronghold Norwood (-17), with Invermay, Mowbray, Ravenswood and St Leonards all around the -14 mark.

Hay’s biggest gain compared to Sylvia Smith in 2003 was naturally in Ravenswood (12 points) and there were net small swings to her in most of the blue-collar suburbs, except some in which Dean had minimised his losses. But in Hillwood (-20), Dilston (-10) and Elphin (-7), the voters leaving Dean weren’t exactly flocking to Hay’s cause either, and it was only in the most working-class booths that the combined Hay/Dean primary losses fell within single figures. In Elphin and Invermay, “neither of the above” was up by eighteen points, and in Dilston and Hillwood (the two booths where Dean and Hay combined polled less than half the primaries) by over thirty!

Given that the combined vote for Kaye and Sands was just 0.4 points lower than that for independent Tony Le Fevre in 2003, it is notable that their vote compared to Le Fevre’s was higher in some booths than in others – Dilston (8 points higher), Hillwood (5 points), Elphin (4 points) and significantly lower only in George Town (4 points) and Ravenswood (3). While these differences are partly because Le Fevre himself polled more votes in some areas than others, it is also notable that the standard deviation of the combined Sands/Kaye vote was higher (4.4 points compared to 3) and that the Sands and Kaye votes tended to move together rather than compete with each other. Furthermore, the combined Sands/Kaye vote moved with the Green vote to a statistically significant degree. Thus, parts of the electorate were voting strongly for either Dean or Hay while parts were much more inclined to vote for anybody but.

The Green vote was, of course, the main cause of the losses in the Hay/Dean primary. Getting excited about a 12% “swing” (a near quadrupling of vote) to the “Green” candidate since 2003, however, would be silly. The unendorsed Green, John R Wilson’s, 2003 result was miniscule, perhaps because he was only running to give diehards someone to vote for (I’ll resist the temptation to declare instead that not perpetually writing inane letters to newspapers might help.) But there is more to it than that, because the Greens’ result in Windermere is a large improvement on their recent state and federal election results in the area.

As noted in our previews the Green vote in Windermere tends to lag a couple of points behind the party’s average in the electorate of Bass overall, mainly because the party polls poorly in Ravenswood, Waverley, Rocherlea etc. The Green votes in the area in the State election in 2006 and Federal election in 2007 were both deflated slightly by competition from minor anti-pulp-mill independents (plus the mill issue was already affecting Green vote patterns in both polls), so a fairer guide to what the Greens should get in this electorate ignoring the pulp mill issue is given by the 2002 state results. While 2002 was a good year for the party, that is cancelled out by the advantages for a Greens candidate opposing a small number of other candidates rather than the Greens opposing multi-candidate parties (as mentioned under Derwent).

Compared to the 2002 state election, Whish-Wilson polled best in Hillwood (+12 points), Invermay (+10), Weymouth (+9), Norwood (+5) and Dilston (+4). He did worst in Elphin (-3), Newnham, St Leonards and Rocherlea (-2), and Ravenswood, Rocherlea and Waverley (+1). For the electorate as a whole he did better by 1.8 points. This difference also moves with the difference between the Sands/Kaye vote and the Le Fevre (2003) vote.

It is clear, therefore, that in some booths there were voters voting Green who normally do not do so, and that in these same booths voters were especially likely to vote for Sands and Kaye. But also, there were booths where the reverse operated – not only did Sands and Kaye poll more poorly than Le Fevre (which in isolation might be put down to the latter’s appeal in certain areas) but also Whish-Wilson didn’t reach the 2002 Green vote. In my view this factor is the Bell Bay pulp mill. Adding the Sands and Kaye and Whish-Wilson votes together and subtracting Le Fevre’s 2003 vote and the Greens’ 2002 result gives a statistic which looks like a pretty decent surrogate measure for the impact of the pulp mill on the vote for minor candidates. And it’s worth doing the full roll call on this one, sorted from best to worst:

Weymouth (18)*, Hillwood (17.5), Dilston (11.7), Invermay (7.5), Norwood (6.6), Mowbray (2.2), Elphin (0.9), George Town South (0.6), Inveresk (0.4), Waverley (-1), St Leonards (-1.4), Newnham (-1.7), Ravenswood (-2.7), George Town (-3.2), Rocherlea (-3.7).

(*= LeFevre electorate average used for Weymouth; it was not a booth in 2003)

The figure for Hillwood is inflated by Whish-Wilson’s personal vote in the area and the figure for Ravenswood perhaps deflated by Hay’s. All up though, it appears that the pulp mill only drove a large increase in the anti-Dean/ILP vote in about five booths, most of those small, while in some booths, most of them large, it actually worked slightly in the pro-mill candidates’ favour.

The story of Windermere booth by booth thus has a lot to do with the pulp mill, but when all the booths are added up and allowed to cancel each other out, the mill almost completely disappears. There was a swing from Ivan Dean to Kathryn Hay among habitual major party voters of around six points, which is mainly attributable to Kathryn Hay being a more popular candidate than Silvia Smith. There was a swing from both candidates combined to the Greens of around ten points, that had nothing to do with the pulp mill and everything to do with the latter making a decent effort. There may have been a small net primary swing away from both pro-mill candidates to the remainder of the field attributable to the pulp mill, but if so, my best estimate of that swing is …

drumroll…

…a single point.

And even that could be overstating it.

Another (much simpler) way of looking at it is this. Any candidate with a semi-decent local profile in a seat like Windermere must get some votes. But Ted Sands polled rock bottom despite trying pretty hard, Kaye didn’t get much and most of his preferences went to Dean and Hay, and Peter Whish-Wilson’s vote was just a couple of points up on what the Greens get in the area in a state election in a good year.

Indeed, I believe the strength of Peter Whish-Wilson’s vote was primarily not because of the pulp mill issue itself, but because he was an excellent local candidate for other reasons.

Worse still for the anti-mill lobby, the preferences of anti-mill candidates were not even enough to dispose of the ardent pro-miller Dean in favour of Kathryn Hay. Indeed, when Peter Whish-Wilson was excluded, a surprisingly high 39% of his preferences went to Dean, 56% to Hay and 5% (191 ardent anti-millers who voted for Whish-Wilson, Sands and Kaye in some order and then stopped!) exhausted. In part, an indictment of lack of effort on Hay’s part, but it continues the practice of Tasmanian pro-development candidates ignoring green issues because even when those issues swell the Green primary vote, they have little impact on the way the preferences fall; they are therefore not worth getting wedged over.

In the past, I have frequently expressed scepticism about the impact of the pulp mill issue on voter behaviour. In the leadup to the 2007 federal election I correctly predicted that the mill would just boost the Bass Green vote by several points with little or no impact on the 2PP result. For Windermere I again played down the likely impact of the issue (the huge number of signatures supposedly on the Voters Block in the area notwithstanding) but the net impact was still even less than I expected. Probably, opposition to the mill has become jaded in the confidence that the thing is a farce that will never be constructed anyway, while the usual mantra “jobs, jobs, jobs” continues to make the mill an effective beacon for maintaining support in blue-collar suburbs expecting to be clobbered by rising unemployment as the recession bites.

Another factor may be the mysterious shortage of co-ordinated and high-profile campaigning by the anti-pulp-mill forces such as Tasmanians against the Pulp Mill (TAP), whose website did not cover the election. In a comment at Pollbludger, Tony Saddington writes “Voters were unable to be reminded on the voters block. Strict electoral policy prevented leaflet campaigning on Voters Block and anti mill issues”. I am well aware that the distribution of how-to-vote cards on polling day in LC elections is banned but it would be interesting to know what further restrictions TAP considered themselves to be bound by, and on whose advice, and what other means of campaigning were explored.

The succour for the anti-pulp-mill forces in this apparent disaster is that anti-mill voting is extremely strong in areas where the Green vote is already high. If this translates into middle-class suburbs near the city in 2010, which I think it will, then the gains to the Greens in those areas should be enough to ensure the survival of their critically endangered MHA, Kim Booth.

Mersey

I don’t have all that much to say about this one. Mike Gaffney polled 42.9%, Steve Martin 27.7, Lynn Laycock only 16.1 and Carolynn Jamieson 13.3. Gaffney went on to win on preferences over Martin by a margin very close to 60-40. Gaffney’s support was naturally high (55%) in his home town of Latrobe, but he was able to win every single booth except one (Valley Road, won by Martin). Without the benefit of being on the ground in the area it is hard to say why Laycock performed so disastrously when she is the mayor of much more of the electorate than Gaffney, but presumably even though Laycock gets re-elected unopposed, that doesn’t necessarily mean that she is popular, and clearly Gaffney must have run a very big campaign to do so well. Martin’s performance is amazing given his past form (and it’s not as if the Mersey Hospital has not been a high-key issue before) but it is also indicates that Laycock and Jamieson completely failed to win over those in the electorate who were sceptical for whatever reason (such as his past Labor connections) about Gaffney. Mike Gaffney is relatively young and on the basis of this impressive performance could have this seat for a very long time, if he wants to and he plays his cards correctly. Certainly beats the uncertainties of being up for Lower House election alongside Best and Green.

One thing I do want to note about the Mersey result is that it is a remarkably emphatic rejection by the electorate of the Liberal attacks on Gaffney over his independence or otherwise. Gaffney had last stood for the party in 2002, he had not stood in 2006 and he was able to claim a degree of distance and independence, although Labor voters would still have seen him as their guy. When Laycock appeared as a candidate whose ties to the Liberals could be argued just as easily, it was hard to see anything but shallow opportunism in their initial baiting over Gaffney. I had wondered if there would be a measure of double standard in the electorate over this (those Legislative Council voters who prize “independence” often turning a blind eye to varying degrees of closet Liberalness) but there was none; unlike Branch in Derwent and Hay in Windermere, Gaffney had made a defensible claim to truly independent status and by going after him over the issue the Liberals shot their preferred candidate and themselves in the foot.

Concerning state politics implications, I’ve really nothing to add to Peter’s comment at #24 of the Poll Bludger thread – Labor should be “reasonably satisfied”. The candidates linked to the party in various fashions returned a comfortable if less overwhelming than expected majority (Aird), a competitive performance despite a seriously slack campaign (Hay) and a “stunning victory, plastering the quasi-Lib hope [..] in the process” (Gaffney).

That’s all from me at the sports desk, but a word about what’s coming up next year on this channel. In 2010 Tania Rattray-Wagner attempts to extend the family’s three-term hold on the seat of Apsley, while ex-PLP independent Terry Martin is up for his first defence of Elwick. It’s back to your usual programming – the main suspense should they both run, will be whether anyone bothers to stand against them.