There were no major surprises on the final day of counting as the election finished with ten seats to each major party and five to the Greens.

In Braddon, Brett Whiteley needed to gain preferences over ticketmate Adam Brooks to have any chance of election, but completely failed to do so. Whiteley started the cutup 1425 votes behind Adam Brooks and finished 1650 votes behind him. I did warn that if Whiteley could gain a strong flow of Liberal preferences based on being a sitting member then he had a chance to beat Paul O’Halloran via a sneaky quirk of the Hare-Clark system, but the general trend of the whole election has been that being a sitting member counts for no more on preferences than it does on primaries. In the end the actual margin between Whiteley and O’Halloran was indeed closer than the party totals suggest, as I said it might be – but only by a trivial 47 votes as O’Halloran bolted in by 1172. It was obviously over for Whiteley once the Hiscutt votes split evenly between Brooks and Whiteley. The loss of their chief headkicker will be a serious one for the Liberals, but they have plenty of new talent to make up for it.

The commonest mistake made by commentators regarding Braddon was to argue that because Labor preferences had historically favoured the Libs in Braddon, that therefore Whiteley would catch up and make a contest of it. In comments on other sites, I pointed out repeatedly during the count that this logic was erroneous for two reasons. Firstly, past cases in which this had occurred often involved multiple candidates competing against a single Green, so there was no guarantee a single Lib would outpoll a Green that strongly if it came down to a one-on-one contest. Secondly, even when the Labor surplus did get distributed, if there were still two Liberals left in the race then the Liberal vote would be split and Whiteley might even go backwards.

And thus it was every time a Labor candidate’s votes were distributed. Leakage from Eastley (excluding to exhaust) favoured Liberal over Green 66/34 but O’Halloran outperformed all the Liberals and gained 44 votes over Whiteley. Leakage from Richmond (excluding exhaust) favoured the Libs over the Greens 70/30 but O’Halloran outperformed two Libs and gained 57 votes over Whiteley. Leakage from Broad (excluding exhaust) favoured the Libs over the Greens 58/42 but O’Halloran again beat both Libs and added another 151 to his lead. Finally, Labor votes leaving the ticket when there were no Labor candidates left to throw them to favoured the Liberals 54/46 (nowhere near enough even if Whiteley had been the only Lib to receive them anyway) but O’Halloran thrashed both Libs individually on these and gained a massive 401 votes over Whiteley – this at the point where he was supposed to be (according to the “Labor prefs favour Libs in Braddon” narrative) going quite rapidly backwards.

Thus, in the end, most of O’Halloran’s margin over Whiteley was actually gained on distributions at which the Greens as a party lost votes compared to the Liberals! It is notable that in one of the most erroneous pieces in this year’s crop of post-election analysis, Sue Neales (HERE) claimed that “Independent observers who are expert at analysing Hare-Clark results” and “Experts from most sides of the political fence” believed that Whiteley had the upper hand “because Labor preferences in Braddon traditionally flow to the Liberals before the Greens”. The above demonstrates why projections based on that assumption were so fundamentally wrong, and why these “independent observers” and “experts” Neales refers to really didn’t have a clue – they forgot the fundamental rule that you cannot analyse exclusively by party trends – you must also consider candidates. Hopefully after this very striking outcome (“the intricate beauty of Hare-Clark” as Cassy O’Connor called it once) we will not see this mistake in analysis of Braddon results in the future.

Of course, the views of these supposed experts cited by Neales were actually at odds with the views of practically every expert observer who had published an assessment and put their name to it. Some of us (like me) were warning Whiteley shouldn’t be written off too superficially, but the consensus was that he was up against it. Neales should have stated who her sources were, so the readers could judge for themselves how expert those people were, and also so the sources’ reputations were on the line if their claims turn out to be wrong. There is no confidentiality-based reason why an independent expert’s name should not be cited in such an article, and indeed, if such an expert is unwilling to put their name to their opinion, it isn’t newsworthy enough to be mentioned. That article by Neales aside (and a few premature trumpings-up of the chances of Andrew Wilkie), I was generally impressed by the quality and clue factor of the coverage of the post-election counting. It was much better than it has been in some previous elections and I think the increased volume of quality analysis on the web has contributed to that.

In Lyons, Rebecca White won quite easily in the end, continuing to gain over David Llewellyn every time Liberal votes were distributed, with an eventual margin of 597. It is notable that Labor scrutineering referred to by David Llewellyn and showing him doing very well on Heather Butler’s votes proved to be inaccurate – he did gain on those, but not by much. Whether the sample was too small, or unrepresentative, or whether those sampling made some kind of method error (like failing to check whether the Llewellyn votes went to Polley first) I do not know, but it was surprising to see that information so far off the mark. I think Llewellyn suffered partly from a general “kick the lot out” vibe (which White’s “renewal” campaign tapped very effectively) but also struggled on preferences because of the uncertainty in the electorate about his future. As I hinted at before the election, once rumours get out that a candidate may be retiring or may not be fully committed, it is very difficult to put a cap on those rumours, whatever the truth of the matter, and an incumbent in such a position will so often narrowly lose that I probably should have seen it coming and firmly projected a White victory on the night of the count.

Many expert pre-election predictions of the final total were published, and this time (unlike last time when nobody seemed to get it exactly right) there were quite a few bullseyes (albeit precariously so – there’s quite often one really close one). Charles Richardson and Richard Herr were among those who, like me, correctly went with 10-10-5, and many others were just one seat out (even the 10-11-4ers didn’t end up too far off the money). The funny thing is that when I predicted 10-10-5 a meagre couple of days out (having been very hesitant to make a final prediction before then because of fieldwork commitments and sheer lack of adequate poll data), I did so rather reluctantly and without any intuitive sense that I was getting it right. There were three powerful intuitive arguments at cross purposes: firstly the idea that Labor might be bad but were not quite bad enough to cop the swings needed to drop two seats in any one electorate. Secondly the idea that the Liberals would get the highest vote, perhaps substantially so, and it was hard to credit that Labor would get more seats than them despite everything. Thirdly the idea that the Green vote was so strong (even the Newspoll said so!) that they would win six seats. I thought they had an excellent chance of six, but 10-10-6 doesn’t add up, and overall I thought the case for the Greens getting that many seats was the weakest one. After all, it was only really based on opinion polls, that were all very difficult to fully credit.

I have since seen comments from a few Greens supporters suggesting Nick and co really were going to poll mid-20s until Labor started breaking out the attack ads and the robocalls against them. But this is just post-hoc revisionism, because at the time, the general view being pushed (and by online Greens barrackers in particular) was that the robocalls were a colossal bungle that was only going to hurt Labor more. It seems it didn’t, though it may have dampened any late swing back to them. Furthermore, Labor’s attack on the Greens was quite a feeble one compared to the vicious Exclusive Brethren/TCA/Liberal Party efforts from 2006, and the impact of those on the Green vote at that election, if anything, was rather modest. If any Greens supporters are disappointed by the result, thinking that it should have been much better, then they shouldn’t be. It was a record tally that in my view now was not far short of the best they could have got at this election. The tough part for them starts now: if they support a government that fails they will be blamed; if they support one that succeeds they could be viewed as superfluous.

Aside from the usual problems with projecting the Green vote, it turned out that the EMRS polls again overestimated the gap between the major parties, which again closed up as the undecideds made up their minds (as it does in most elections). One of the biggest sources of inaccurate hot air in the election campaign was the interpretation of polls showing the Libs seven to eleven points in front as being reliable indicators that the Libs would win more seats than Labor. This was parallel to the error made in 2006 by those who assumed there would definitely be a hung parliament just because Labor was polling in the high 30s and low 40s instead of 50s (or because they wanted it to happen). I wrote on 25 Feb when this year’s mainstream media plague of EMRS-equals-scripture fatalism was most out of control:

“But it is far too early to write off the possibility of Labor retaining government as the largest party in minority or at least containing the damage to the messiest possible scenario, 10-10-5.”

The latter is exactly what has happened and the former went within a whisker of occurring (depending upon the will of the Greens of course). By the time you read this, whether Labor tries to survive with 10-10-5 or keeps Bartlett’s commitment to throw in the towel may well be history, as may whether their intentions are decisive, or whether the uncertainty about who would govern next continues for some time. There has been a lot of comment in the area of the constitutional responsibilities of the Governor, and while some of it has clearly been overexcitable, I do not claim to tell you what the answer will be.

A common question that I get is what breakdown a given election would have produced under the old 35-member system. A refresher of previous alternative-universe results (assuming exactly the same vote breakdowns) – the 1998 election (actual result 14-10-1) would have produced 16 or 17 Labor ,14 Liberal and 4 or 5 Green. The 2002 election (14-7-4) would have been 20-9-6. 2006 (also 14-7-4) would have been 19-11-5. And finally, 2010 (10-10-5) would have been 14-14-7. Denison would have gone 3-2-2 to Labor, Franklin 2-3-2, and all the others 3-3-1. In this case the difference in size of the house has no proportional impact whatsoever since both results are 40%-40%-20% (which is actually extremely close to the votes cast). An unusual difference is that while Andrew Wilkie almost won in the 5-seat electorate of Denison, in the 7-seat version he would have had a larger proportion of a primary vote quota but actually would not have got close, since the Greens would have got two over the line and there would not have been half a quota of Green preferences for him to attempt to catch the Liberals on. (Nor would there have been any Liberals to catch – they would have had two quotas locked up too.)

One more comment on Wilkie: following the strong suspicion that Green voters who let their vote exhaust at 5 caused Wilkie to miss out, there have been some suggestions that Wilkie and the Greens should have explored some kind of cross-preferencing arrangement. But quite aside from the likelihood that this would not have helped Wilkie anyway (since it would have forced him to be seen as pro-Green-biased and hence damaged his chances of taking votes from the Liberals) there is a question of what the Greens as a party really wanted concerning Wilkie’s fate. Should the Green MHAs have really cared whether the major parties split 10-9 or 10-10? After all, either way, if they choose to, they can ignore all the shenanigans about Labor and Liberal pluralities and make their own decision about who they are willing to support. And, whatever they say in public about being keen and willing to work with their ex-ticketmate, would the Greens actually really have enjoyed having him in there and potentially competing for their votes, perhaps even forming a new multi-candidate force over time, or would that have just been a recipe for conflicts reminiscent of Tasmania’s Greens-Dems brawling of the 1990s?

The new parliament will be a fascinating one to see in operation after this dramatic clean-out and with the presence of so many new faces, many of them talented and capable and with a lot to offer public life, but thrown straight in the deep end of what may turn out to be a chaotic term in office. The first stages are tough for the new Labor MHAs in particular. But when I said that the messiest possible scenario was 10-10-5 I may yet have been wrong.

The messiest possible scenario could have been the one we have just very narrowly missed out on.

Kevin Bonham looks forward to working with the new Minister in charge of the Tasmanian threatened species list, whoever and from whatever party they may be, but only if they are dynamic, ruthless, scientifically informed and armed with a big red pen.