Tasmania 2018: How Woodruff Won Franklin

It's been a hideous few weeks for the Greens - they lost votes and a seat in Tasmania, were thumped in the Batman by-election and were remarkably anonymous in South Australia.  The party is now facing serious internal recriminations over these poor results.

However there was one rather nifty save amid all this, and before I move on to the other house of the Tasmanian parliament (there are two Upper House seat contests coming up in May) I want to post the instructive Hare-Clark details of how Rosalie Woodruff (Green) managed to retain her seat in a very close contest with Nic Street (Liberal).  This article is naturally rather mathsy and has been rated Wonk Factor 4/5.



Had the Liberals succeeded in Franklin the Greens would have been reduced to one seat.  (Note that this would not have actually stopped Cassy O'Connor from having motions debated - she's advised me that the House of Assembly doesn't require seconders for motions - but I expect it would have been a miserable time for the party anyway.)  As a result of not quite getting there, the Government has only a one-seat majority, so had better hope there are no backbench rebellions in this term!

Given what the government was polling last year, it's pretty amazing that they got as close to holding this seat as they did.  The seat was on a 2014 margin of 2.6% over Labor, but the Labor vote rose to over two quotas, meaning that the only way the Liberals could win the seat was by taking it from the Greens.  In 2014 the Liberals had been very nearly two quotas up on the Greens, so they needed the Green primary to fall and to do so by enough to compensate for the inevitable leakage off Premier Hodgman's massive personal vote.  It turned out that the Green vote in Franklin actually did fall more than the Liberals, but they were able to hang on anyway.

On election night, it appeared to me that while it was going to be extremely close, Street was slightly the more likely of the two to win.  The Liberals had a lead of around 500 votes on party quotas (after deducting their two quotas needed to elect Will Hodgman and Jacquie Petrusma) and while they were obviously going to lose a lot of votes from Will Hodgman's surplus, they were also going to gain nearly as much on preferences from the Shooters, Fishers and Farmers.  Beyond that point, they were a little more exposed to leakage in terms of the number of votes likely to be thrown from their minor candidates, but also Liberal votes don't tend to leak as much as Green votes, so it seemed that the Liberals should still win by 100-200 votes.  But it didn't turn out like that.

As shown by the scrutiny sheet, this is how the battle panned out through the various exclusions and surpluses:

(The count moves from left to right and the numbers on the bottom are just different events - surpluses, exclusions, and part-exclusions.)

The Liberal tally started ahead but initially dropped because of leakage from Will Hodgman's surplus.  Leakage from minor Greens candidates put the race back to even before the Shooters preferences propelled the Liberals to a 672 vote lead.  However, this was mostly lost immediately on leakage from the other remaining Liberal, Claire Chandler.  After the exclusion of Labor's Kevin Midson, Street then led Woodruff by three votes, but the final throw (Alison Standen's surplus off preferences from Midson) gave Woodruff victory by 226.

Another way of looking at it is in terms of the battle between candidates Street and Woodruff:

Street was effectively the third Liberal on a ticket lead by the Premier so his starting primary vote was a lot lower than Woodruff's.  Here the main gains are for surpluses and exclusions from each candidate's own ticket (so the Hodgman surplus and the Chandler exclusion are Street's biggest gains while the exclusion of second Green Richard Atkinson's is Woodruff's).  Although the Liberal Party had a sizeable lead after the Shooters preferences, that vote was split between two Liberal candidates.

A breakup of the different preference types is very instructive concerning what kinds of votes contributed to the result.  Here I refer to a preference vote as "leaking" if it comes from one candidate in a party, could go to another candidate from the same party, but instead goes to another party or exhaust.  Where there is no remaining candidate from the same party, I simply refer to it as a "preference".  So, for instance, the votes from Kevin Midson that did not flow to Alison Standen "leaked" out of the Labor ticket, while those that flowed from Midson to Standen may not have included all five Labor candidates (and might even have originated somewhere else), but at least stayed within the ticket at that point.

The rates at which votes leaked from both the Green and Liberal tickets were not unusual.  What was unusual was the very small share of the leakage that went from each of these party slates to the other.  10% of all Greens leakage in Franklin went to the Liberals; this compares with 19% in 2014 and 42% in 2010.  14% of all Liberal leakage went to the Greens; this compares with 21% in 2014 and 20% in 2010.

With so few leaks from the Liberals going to the Greens and vice versa, the leaks had to go somewhere.  And with only one fourth-party candidate in the mix (and not for all that long) where they went was overwhelmingly to Labor.  This had the effect of continuing to push up Labor's excess over their second quota, such that Labor started just 738 votes over their second quota, but by the exclusion of Midson they were 2193 up (768 was from Hext (SFF) and 687 was net gain on leakage from other parties.)  As a result, votes coming out of the Labor ticket became much more important than they seemed at the start of the count.

Woodruff gained 395 over Street on the Labor votes (including both leaks and the preferences at the end) and won by 226, so had Labor gone substantially backwards on net leakage, she probably wouldn't have made it.  But she also benefited from a change in the behaviour of those Labor votes.

As I noted on the postcount thread, in most cases in the past where Labor preferences have been thrown at the end of a count they have advantaged the Liberals over the Greens. The same is true for Labor leakage in Franklin - over the last four elections almost three quarters of Labor exclusions have seen more leakage to the Liberals.  Yet at this election leaks from all three excluded Labor candidates (Barnsley, Chong and Midson) favoured the Greens over the Liberals.

This probably has a bit to do with what kinds of candidates they were - Barnsley and Chong were both female candidates who are extra-likely to appeal to Green-ish voters (Barnsley from her public service and anti-smoking background, Chong from local government), while Midson, while male and of the Right, is also young.  Indeed, Chong was the one Labor candidate at the 2014 election from whom leaks favoured the Greens.  But I think there's more to it than that.  Just as Labor's pokies policy and generally left-wing campaign took primary votes directly from the Greens, so I believe there would have been plenty of voters who mixed and matched from the Labor and Greens tickets, rather than voting for all of one then all of the other.

The other thing notable here is that Street really struggled through the count to consolidate the Liberal votes under his own name.  A lot of votes (including Liberal votes) were pooling with Chandler rather than Street, and this meant they got another opportunity to leak when Chandler went out.  Partly this reflects the struggle that MPs elected mid-term on recounts often face in getting their profile high enough by the next election. In this case Woodruff was in the same position, but she had had a bit more prior profile-building, eg as the party's federal candidate in 2013.  There was a view that Street's social issue views (as demonstrated for instance in his excellent sheep speech) might get him over the line, but those views most often would have appealed to those with strongly left-wing views who were too committed to the destruction of the Hodgman government to preference any Liberal.