Tuesday, October 30, 2018

2018 Hobart City Council Count (With Some Coverage Of Other Councils)


The number above appears at the top of my coverage to highlight the final informal vote rate for the Hobart City Council councillor count, as a result of absurdly strict formality requirements. Launceston (7.94%) and Clarence (7.24%) are not far behind.  

This level of informal vote as a result of absurd legislation is a farce, an insult to democracy, and a threat to the legitimacy of seats being decided by a handful of votes.  The informal rate was 100 times the final seat margin in Hobart.

I call on the state government and other parties in the Lower House to immediately and publicly commit to fixing this problem.  The current government did not create this problem, but the problem should have been fixed after the last election four years ago.

Coverage follows below.

Note added Saturday night: I will be mostly offline for the coming week (Nov 4-10) so comment clearance will be slow.

--------------------------------------------------------  
Introduction (from Tuesday)

Welcome to my live coverage thread for the Hobart City Council count, which will also have some comments on other councils when I find time to look at them.  My Hobart candidate guide and preview was here and has probably been viewed by about 20% of Hobart voters.  Updates will be added below the dotted lines; check back regularly through the week for comments.  These introductory comments will stay at the top, there are also some more detailed introductory comments at the bottom.



If you have a question about another Council after there are figures available (not before) please ask it in comments. Please note that comment clearance will be slow over the next few days.

During scrutineering I will often have my phone off or on silent.  Media and candidates wishing to talk to me should send a text or email if they have a simple question, or leave a message;  I will get back to them when I can.  TV interviews today will be difficult to arrange and will depend on breaks (if any).   I'll be based at Derwent Park and Elwick and if I come to the C3 tally room then that will probably be late in the night.

Scrutineering and posting comments on councils involves significant downtime from other work and has in the past cost me $150-$200 in travel and phone costs, though I expect those costs could be lower this time.  Donations to cover for some of this are very welcome and can be left via Paypal in the sidebar or email me (address via profile on sidebar) for bank account details.  Please only donate if you are sure you can afford to do so.
-----------------------------------------------------------------------------

Podcast: While waiting for the Hobart button to be checked I have appeared on Ben Raue's Tally Room podcast to discuss this election.  (Link to podcast thread at the Tally Room is here.)

A note on informal votes:

The following are some further details concerning the informal vote rate statewide following the completion of all councillor counts.

Firstly, though the relationship is messy, in general where the number of candidates in a council increased in 2018 compared to 2014, the informal voting rate went up.  If the number decreased it usually (but less consistently) went down:


Secondly considering 2014 and 2018 results combined, there is a strong relationship between the number of vacancies (which is also the number of boxes a voter must number without error) and the informal rate:


Thirdly, ignoring whether the number of candidates has changed or what the number of vacancies is, there is a strong relationship between the number of total candidates and the informal rate:


It's not clear from this alone whether requiring more boxes increases the informal rate for the same number of candidates, or whether it is purely driven by candidate numbers.  However a multiple regression shows that it's both:

Informals = .6033*(Number of vacancies) + .1553* (Number of candidates) - 4.498
(r-squared = .8664)

One might suggest that something about those councils that had high vacancy numbers was driving their high informal rates, except that wasn't the case before all-in-all-out.  So both reducing the number of boxes voters have to number without error, and reducing candidate numbers, should be effective in reducing informal voting rates, but the former is the bigger part of the problem.
----------------------------------------------------------------------------
Hobart Council comments

HOBART SUMMARY

Lord Mayor ELECTED: Reynolds new Lord Mayor
Deputy Lord Mayor ELECTED: Burnet (Green) returns to position
Councillors ELECTED: Reynolds, Burnet (Green), Harvey (Green), Thomas,  Briscoe, Behrakis (new), Denison, Zucco, Dutta (new), Ewin (Green - new), Sherlock, Sexton.
Incumbent Lord Mayor Ron Christie loses seat.

-------------------------------------------------------------------------------------------------------

The tale of the tape: 

This is how the final Hobart count unfolded.  Firstly, as we saw last night in all of Launceston, Huon Valley and Kingborough (but not in Clarence today) Green and left-wing candidates generally did about 15-20% better on the second half of counting than the first.  Conservative candidates did worse, with Merridew in particular almost 30% weaker in the second half.

This meant that Thomas no longer hit quota, and Reynolds' 1.99 quotas put Burnet and Harvey over quota with now significant surpluses, further boosting Ewin.  All this also massively lifted Dutta.  Through the exclusion process Behrakis rose to sixth, and Dutta ended up crossing ninth together with Denison and Zucco. These surpluses elected Ewin.

The six-candidate crunch point for four places that was in the 20% and 50% count now looked like this:

Christie 1326.16
Sherlock 1319.38
Sexton 1105.60
Merridew 1053.60 - excluded

Now Christie is way ahead of Sexton here, but Merridew's preferences broke so strongly to Sexton that his exclusion left this:

Sherlock 1600.93
Sexton 1541.72
Christie 1521.58

As a result Christie, who was 8th on primaries, lost, while Sherlock won from 17th!  Preferences flows from her ticket were not that strong but were enough to make the difference, so this looks like a rare case of a non-Green ticket actually working.

In terms of the overall balance of the Council, it looks like a minor gain to the "green" side (left) since there will be five councillors who I would expect to be quite left-wing, five who will line up on the "blue" side a lot, Sexton in the middle and who knows what Sherlock will do after being elected on a "blue" ticket.  But in terms of the cable car, its position on the council has improved, because instead of Christie (anti-cable car) and a vacancy, we have Behrakis (strongly pro cable-car) and Sherlock (in principle supportive).

Saturday 7:35 ALL OVER - CHRISTIE LOSES SEAT.  Sexton beats Christie for final seat by 20 votes after Merridew drops out, detailed analysis to follow.

Saturday 5:00: If Christie does win, it's possibly significant for the cable car that he has said he may not serve a full term.  On current numbers if Christie gets up instead of Ewin, then on a recount he could well be replaced by Sexton (whose position is neutral) though another possibility is Corr (who is opposed.)  Let's see if he does win and if so who misses out.

Saturday 11:00: We await the final button press sometime today or tomorrow but it should be noted that in all of Kingborough, Launceston and Huon Valley, Green and Green-ish candidates lifted substantially on the final count.  This could be a good sign for Ewin.

Mayoral Preferences: The cutup for Lord Mayor is up.  Of those preferences that went to either Reynolds or Thomas at the various exclusions, Reynolds got the majority from all four anti-cable-car candidates (74.8% in total) and Thomas got the majority from all five pro-cable-car candidates (61.3% in total).  The flow from Denison (55.6% to Thomas) was weaker than from the male candidates, mainly because of gender voting.  A similar pattern with the Deputy Mayor count and preference flows involving Burnet vs Behrakis.  

Hobart 50% Count: The 50% count is up.  Reynolds has lifted to 1.81 quotas and Thomas to exactly a quota, so they are now elected 1 and 2.  Reynolds' surplus now puts both Burnet 3 and Harvey 4 over, and then exclusions from the bottom up put in Briscoe, Zucco, Behrakis and Denison.

This is what the crunch point for the six fighting for four places looks like now:

Dutta 713.89
Ewin 682.45
Christie 662.47
Sherlock 662.06
Merridew 644.82
Sexton 625.36 - eliminated

This time Sexton rather than Sherlock is excluded.  Sexton's preferences put Merridew straight over the top, then Sherlock.  (Sherlock comes up from 18th on primaries, assisted by a weak but nonetheless useful flow from other Liveable Hobart Group ticket candidates.) After their surpluses we are left with the following:

Dutta 801.50
Ewin 772.08
Christie 764.93

There is still time for the 19.46 vote gap from Merridew to Sexton to turn around and put Merridew out instead, which I would then expect to save Sexton and possibly Christie.  The even greater uncertainty is Ewin and Christie for the final seat which is much too close to call.  The reason Ewin is not getting in so easily on this count is that it is Sexton who drops out rather than Sherlock, and Sexton is a much worse preference source for her.

It is still possible for Sherlock to lose in two ways - either by dropping to last at the first exclusion or by finishing 13th if Merridew drops out and his preferences don't put her over Ewin (though I suspect they might).  But obviously her position is much stronger than in the 20% count.  It is also still possible for Dutta to lose at the last exclusion but he is safe from being excluded at the first crunch point.  It also doesn't look like Ewin will drop out there, but her position at the end is basically a tossup if Sexton goes, and possibly even worse if Merridew goes.

We will be awaiting the final button as I am still in no position to call the winners of this six-way contest for four places!  But Dutta and Sherlock are both in pretty good positions, Merridew just needs to hold steady at one exclusion, Ewin is very dicey, and Christie and Sexton are still in it but need to lift.

Thursday 11:42: The 50% count is up and now Merridew and Sherlock win while sitting councillors Sexton and Christie are out.  I am examining this to see how close it is and whether it can be relied on and will post findings shortly.

Landslide Result

Commenter Matt Lyons asks if it would be fair to call this mayoral result a landslide and my answer is absolutely!  At the last election the left did not even make the final two and in this case Reynolds has won by this margin.  Although large margins for established mayors Kennedy and Valentine were common, this is the biggest margin by a new Lord Mayor since at least 1986 (when Kennedy won with what I think was a similar margin over Brian Broadby).  Since then new Mayors have often had close wins:

Freeman 1996 57.4-42.6 vs Valentine (estimate, some votes left unthrown at end of count)
Valentine 1999 51.2-48.8 vs Freeman
Thomas 2011 51.5-48.5 vs Burnet
Hickey 2014 52.5-47.5 vs Thomas

This however is a massive win - inflated by the excessive number of "blue" candidates running for Lord Mayor with the resulting increased exhaust, but nonetheless a remarkably large margin, especially for a candidate only narrowly elected for the first time in 2014.

Oddly, this is also the third time in a row a councillor has won the mayoralty at the end of their first term.

Wednesday 8:33: Final results are up.  On a 2CP basis it's Reynolds 62.35% - Thomas 37.65%.  However that's of votes still in the count, with 17.6% of votes exhausting.  So including exhaust it's Reynolds 51.4, Thomas 31.0, 17.6 exhaust.  That still exceeds the margin in my weighted sample (which I thought was optimistic!)

Wednesday 8:06: Thomas got virtually nothing from Denison.  Now comes the avalanche of Harvey preferences favouring Reynolds which will blow out the margin considerably.  

Wednesday 7:00 Burnet has beaten her two remaining Deputy Lord Mayor opponents combined, saving the TEC hours of counting.  

Wednesday 5:15 After Briscoe, Reynolds' lead has extended so any remaining morsel of doubt (however implausible) has now been totally extinguished.

Wednesday 3:39 Zucco gone and the gap closed slightly.  Now Reynolds 30.93 Thomas 21.06.  Briscoe next to go.  Thomas would have to make very substantial gains here (completely contradicting my limited sampling) but it is much more likely Reynolds will gain or at best for Thomas that nothing happens.  Even if Thomas makes huge gains my sampling of yesterday had Denison's votes not favouring Thomas at all, let alone by as much as Harvey's favour Reynolds.  

Wednesday 2:53 Christie excluded, now Reynolds 29.08 Thomas 17.54.  Zucco next to go.

Wednesday 2:22 Candidate Salpha Dut is on just 2 votes (0.05%) in the 20% count, so will probably finish with about 10.  This is easily the worst Councillor vote I have ever seen for any council.  Dut had no candidate statement, no photo and did not appear to campaign.

Wednesday 1:54 Dutta is excluded and Reynolds did not make as big a gain directly as I thought she might, though a lot will flow indirectly through Harvey.  Currently Reynolds 27.11 Thomas 16.12. 4% exhausted already.

Wednesday 1:26 Three candidates have been excluded from the Mayor count (although a blooper on the TEC site says Alexander has been "elected".  Preferences of Sexton, Alexander and Mallett have assisted Thomas slightly but the last two were the most supportive of him yesterday.  Reynolds' lead will stretch again when Dutta, Harvey and in my sample Briscoe are excluded. Other candidates are making no inroads on Thomas for second so it now looks highly likely to finish as Reynolds vs Thomas with Reynolds winning easily.

Well Done EMRS!

In a world where we are used to seat polling being a dumpster fire, EMRS's brave adventure in polling the Hobart City Council mayoral race for a client has been very successful:


With a sample size of just 319 intending voters who knew who they wanted, for an election which about 40% didn't vote at and 40% of those intending to vote didn't know who for, EMRS has produced a remarkably accurate reading with the top three in the right order, an average error of just 2.1 points and only one error (Denison) exceeding 3.  Remarkably good given the extreme difficulties in polling of this sort of race.

Wednesday morning: Having a bit of a rest after yesterday's 15 hour working day! May do some more scrutineering later today, not sure yet - in the case of Lord Mayor we can just watch the numbers come in as candidates are excluded until Reynolds crosses the line sometime late today or tomorrow.  Although I haven't used the magic word just yet it is pretty much a foregone conclusion.

50% counts for large councils including Hobart are not likely until Thursday and the Hobart count might not finish until Sunday (gulp!) I will be travelling to remote fieldwork on Sunday and it's possible I won't be online by the time the final button is pressed.

A note about 20%/50% counts Some journalists are reporting that councillors are getting elected or excluded based on the 20% counts.  This is false; nobody is formally elected or defeated as a councillor until the final button is pressed with all votes included (which will take a few to several days varying by council.)  However those shown as elected or defeated early in counts will very likely have that fate.  It is those elected or defeated at the bottom, or marked as "remaining in the count" whose fate may change.  "Remaining in the count" just means that on current numbers the candidate would be the last one left unelected (eg 13th for a council with 12 seats).  Because the election would be over at the point with this "remaining" candidate excluded, there is no need to actually exclude them.


Cable car hangs by a thread: Prospects for the cable car to have a majority of at least potentially supportive councillors are not looking good.  In the 20% count only five who are even open to supporting use of public land are elected.  Seven is still possible but only if all of Sexton, Merridew and Sherlock win.  If the cable car cannot get a majority of councillors open to the project then routes still available to it include legal action, hoping somebody changes their mind, or perhaps councillors opposed to the project being absent.  Legal action would be against councillors whose votes would be claimed to have been precommitted against the project, but just as many could be claimed to have been precommitted for.

The 20% count for Hobart: 

Reynolds is the only candidate with quota on 1.7 quotas.  Thomas is second on 0.95 but does so badly on Reynolds' preferences that she puts Burnet (0.82 Q) over the line but not him.  Then it's just exclusions from the bottom up as Thomas, Harvey, Briscoe, Zucco, Behrakis and Denison cross the line.

We then come to the crunch point of the whole count, which is currently the exclusion of Zelinda Sherlock in 14th place with 8 elected:

291.28 Christie
270.68 Sexton
266.91 Merridew
265.88 Ewin
260.97 Dutta
256.93 Sherlock (excluded)

In the 20% count, Sherlock's exclusion elects Ewin and Sexton, and pushes Dutta up to the point where he overtakes Christie with the following final numbers:

317.07 Dutta (wins 11th)
314.56 Christie (wins 12th)
306.27 Merridew (not elected)

However, any of Sexton, Merridew, Ewin or Dutta could be excluded instead of Sherlock.  I would expect Ewin's preferences to elect Sherlock such that one or the other will make it, but we have to wait and see which (perhaps even both, if Dutta goes out instead.)  And even if Sherlock is excluded, Merridew is close enough that he could beat Dutta or Christie.

Christie is obviously not safe at the end so all we can say is that these six are fighting for four positions.  I feel inclined to call the first eight as all of them are elected at least 100 votes clear of elimination.  From a reasonably strong primary vote Christie has become the new Marti Zucco - his performance on preferences is abysmal and this places him at serious risk, though at the moment he just survives.

12:50 It was worth the wait for drama value because in the 20% count HOLLY EWIN IS ELECTED. (She quit the Greens in the final week of campaigning.) Also up in the 20% count are Sexton and Dutta, with Christie only 12th and Merridew missing out 13th.  Examining this count closely to see how prone to change it is, comments soon.

12:40 Still waiting!

10:25 Still waiting on the 20% count.  Hopefully will be up sometime tonight but could be late - heard there's a lot of rechecking of suspected informals still being done.  Note that Lord Mayor counting starts again from 9 am tomorrow, probably with the preference throw.

Tues 8:54:  I have done a weighted sample of 311 Lord Mayor preferences.  In this sample (weighted by the vote strength of other councillors) Reynolds extended her lead from 10.4 points on primaries to 18 points after preferences.  I feel that could be on the optimistic side given unexpectedly strong preferences off Denison, Zucco and Christie but the important point is that my sample does not show Thomas catching at all, let alone by nearly enough.  So there seems to be little doubt about it.  The question I haven't tried to cover off on is whether anyone else who might pass Thomas might be in a position to challenge Reynolds but that seems highly unlikely too.  I expect a long preference throw tomorrow but Reynolds is in a very strong position.

Tues 8:25: Have heard Hobart is getting fairly close to 20% councillor count.  About 79/91 of the way to it to be precise.

Tues 8:11: Hobart mayor count still in recheck, I will do a model based on sampling when it's finished but I expect it to show that Thomas is not catching Reynolds assuming those are the final two.  Harvey won't make the final two, so it would take a lot for any of the others to catch Thomas for second, and even if they did their gap to Reynolds would be larger.

Something worth noting: Denison (staunch Lib) and Reynolds (Green-turned-Ind on a basically Green campaign) are the only women in the field.  1 Denison 2 Reynolds votes are a thing, I've seen a lot of those.

Tues 7:20: Hobart mayor count now doing formality check.

Tues 6:55: Provisional Lord Mayor figures (note the TEC website is down so might be a while before you see them there!) Reynolds 23.4% Thomas 13.0 Harvey 12.2 Briscoe 9.0 Denison 8.6 Zucco 7.9 Christie 6.9 Dutta 6.1 Sexton 5.2 Mallett 3.8 Alexander 3.7. A very big lead for Reynolds and no evidence Thomas will reduce the gap let alone close it.  It looks a lot like an ex-Green Lord Mayor and a Green Deputy will be waiting to see if they have a friendly council or not!

Tues 6:30: Provisional Deputy Mayor figures (something like this will go up soon with the rough primary count finished already!) Burnet 34.8%, Behrakis 13.4, Sherlock 11.1, Merridew 7.1, Stansfield 6.3, Allardice 6.0, Roffe 5.8, Bloomfield 5.0, Taylor (Andy) 4.7, Coats 3.8, Waldhoff 2.0.  Even if no votes exhausted any other candidate would need 71% of preferences to catch Burnet, and votes are exhausting so that's not going to happen.  The margin may close up to whoever finishes second.

Limited Lord Mayor preference sampling so far has been consistent with Reynolds winning with no evidence Thomas will even close the gap, but much more work to do.

Tuesday 6:00:  Now at the Lord Mayor / Deputy Mayor count.  For Deputy Burnet has a massive pile and could be heading for a primary around the mid to high 30s, Behrakis and Sherlock are next but way back.

For Lord Mayor Reynolds has about 23% based on the size of piles with Thomas well behind in second (c. 14%) and Harvey (12%) just behind Thomas.  I am going to assume a Reynolds - Thomas finish and sample preferences on that basis.

Tuesday 4:20: I have now doubled the sample and the second half had a slightly leftier feel - I now have Reynolds topping the poll with over 1.4 quotas, and nobody else on track for quota although Thomas and Zucco might yet lift and make it on the first ballot.  Denison did much better in the second half of my sample and Alexander very much worse.  I now have the top nine as Reynolds, Thomas, Zucco, Behrakis (!), Harvey, Burnet,  Briscoe, Christie, Denison.  Some distance behind fighting for three places on current sampling are Dutta, Alexander, Mallett, Sexton, Sherlock. Merridew, Ewin (Ewin a little behind the others in sample, the rest all about the same).  There won't be anymore Councillor counting til 6-ish and I don't expect to know much more than the above til the 20% sample, so I will probably relocate to the LM/DLM count.

Tuesday 2:20 pm: I have been sampling Councillor primaries in scrutineering and now have a sample of 546.  These could be from anywhere as noted below, so patterns are tentative at this stage.  It is also a very difficult count to sample because nobody much seems to be getting quota, meaning that it is hard to tell who will have surpluses (if anyone much) and a lot of results are not statistically meaningful.  In this sample, Reynolds has about 1.3 quotas, Thomas 1.0, Zucco is just short, and the following have most of a quota: Briscoe, Burnet, Behrakis, Christie, Harvey, Alexander in about that order.  Around 0.4-0.5 quotas in my sample are Denison, Mallett, Dutta, Ewin (!), Sexton and Merridew, with Sherlock and Corr also on the radar.  However with the sample size being so small there may be others who become competitive with more sampling.  A lot of Reynolds' preferences are going to Burnet and Harvey in the event that they are shy of quota.  I intend to sample at least another run of 500-odd before going to the DLM/LM counts, as even for the leading contenders the sample result is only a few dozen votes and may not mean a lot.  The portent for those counts at the moment is that they will be very long and messy. I am not even going to call anyone in as Councillor on this sample, but I will make some calls after my second round if the numbers are similar.

Tuesday 10 am: Record Hobart turnout - as of yesterday Hobart had already broken its turnout record with 57.5% turnout (previous record was 55.5%).  This figure will probably rise a few percent with final numbers. Unfortunately based on past experience several percent of Hobart Councillor ballots are likely to be ruled informal because of overly strict formality rules.

-----------------------------------------------------------------------------

Other Council comments

A note about 20% counts Some journalists are reporting that councillors are getting elected or excluded based on the 20% counts.  This is false; nobody is formally elected or defeated as a councillor until the final button is pressed with all votes included (which will take a few to several days varying by council.)  However those shown as elected or defeated early in counts will very likely have that fate.  It is those elected or defeated at the bottom, or marked as "remaining in the count" whose fate may change.  "Remaining in the count" just means that on current numbers the candidate would be the last one left unelected (eg 13th for a council with 12 seats).  Because the election would be over at the point with this "remaining" candidate excluded, there is no need to actually exclude them.

Saturday 2:00: Button pressed in Clarence and after last night's drama in other councils there ended up being nothing to see in this one, nothing much changed and Brendan Blomeley returns to Council after a 10 year absence with Kay McFarlane defeated by 49 votes.

Friday night: Another thing to keep an eye on in Clarence is that Castillo is currently only 23 behind McFarlane and 26 behind Thurley at the point where Castillo goes out; all these are female.  So if that changes (based on what we've seen in the other councils, Thurley might go out here) then maybe Castillo is still in the mix.  

Friday night: Small councils are more prone to variation between counts than big ones, and in the Huon Valley count Burgess (a winner in the 50% count) got only 55 primaries in the first half compared to 90 in the second half.  This and other changes meant Burgess was excluded 0.22 votes behind candidate Atkins and 0.93 behind candidate Clark.  Although Burgess did very well on preferences of Clark, O'May's position in the final count was much better relative to Burgess so I believe that even had Burgess got through this exclusion point, O'May would still have won the final seat.  

Friday night: What we have seen in the two below could easily flow on to Clarence and save Kay McFarlane as well.  We'll see!

Launceston final count: weird stuff.  In the 50% count, Walker (Green) was excluded 30 votes behind Williams (Green) at quite an early stage.  But in the 100% count, both lifted, but Walker lifted more, by the equivalent of 80 votes.  This meant that instead of either of them dropping at that point, both stayed in while all of Salt, Fitch, Limb, Dakin, Wood and Potter were excluded.  But it also meant preferences going to the Greens split two ways, so Williams was now excluded ahead of Stojansek.  This meant not only Stojansek (a left independent) but also Walker (Green) got up and McKendrick lost.  A bizarre turnaround (and again, stratification!) but a bonus for the left.  

Kingborough final count:  What has happened here is a bit of the stratification I warned about at the bottom - it looks absolutely like the first 50% had more of the early votes and the second 50% more of the late votes.  Left-wing voters vote later and every Green or greenish candidate lifted in the second half.  Mary Midgeley (Green) polled 307 votes in the 50% count but added another 433 in the second half.  Her position then improved by at least another 100 because Richard Atkinson went further over quota.  This put Midgeley over quota before Gideon Cordover (Green) was excluded, unleashing Cordover's preferences to flow to Fox (a Green-ish candidate albeit an independent) with the result that Fox went from losing to being elected 8th.

Friday night: Final results for Kingborough and Launceston are in and both saw quite a bit of movement from the 50% count.  In Kingborough, Flora Fox is now saved and Samantha Scarlett misses out.  In Launceston, Emma Williams now misses out at the cutoff point with Stojansek (who retains), but her preferences put Tim Harris in so her seat is claimed by another Green.  Also, Robin McKendrick is now defeated at the end of the count. Investigating ...

Friday 3:40:  West Tamar and Devonport counts are now completed.  Five to go.  

Friday: As I have continually ranted about, the informal rates for Hobart, Clarence and Launceston are up in provisional figures, even compared to the already much-too-high rates for the large councils in 2014.  An increase was also seen in the separate Glenorchy election earlier this year.  Across other councils, graphs by Kester Takayama show the rate has been more or less steady from 2014 on the whole.  (In 2014 informal vote rates almost doubled statewide because of the increased number of vacancies combined with inflexible formality rules.)

Thursday midnight: All councils bar seven (Hobart, Clarence, Launceston, Huon Valley, Kingborough, West Tamar and Devonport) have now finished.  Close one in George Town with a 3.34-vote result for the final seat, and the Deputy Mayor Tim Harris only able to take his place after surviving three different exclusion points by 4.3, 12.46 and 15.56 votes.   The time frame for the remaining councils is by Sunday.

Thursday 1:50 The TEC display for Mayor of Southern Midlands was slow updating yesterday and some media reported that Tony Bisdee had been re-elected.  He hasn't; Alex Green overhauled him on preferences.  So that's five mayors defeated.

Thursday 1:40: The Clarence 50% count is up and the final position has flipped with Brendan Blomeley now beating Kay McFarlane by 30 votes for the final position, both get over Viney by 55 at the point where he is excluded.  This is still recoverable for McFarlane but obviously one would rather be in Blomeley's position.

Thursday 1:20: David Laugher has won the Southern Midlands Deputy Mayor election against Edwin Batt by 102 votes but has failed to be elected as a councillor by 22.23 votes, while Batt was elected as a councillor.  Laugher cannot be Deputy Mayor since he has not been elected as a councillor.  The council will elect a Deputy from among its members.  This has happened a few times before at Deputy level but never at Mayoral level.

Thursday 10:50:  Huon Valley 50% count is up.  The main doubt there concerns Burgess, who currently avoids exclusion by being 19 votes over Clark then goes on to beat O'May by 33.  The other eight seats look clear.   

Thursday 10:40: Kingborough 50% count is up and the patterns in the 20% count have strengthened so I expect no further change in that one.

Wednesday midnight: Launceston's 50% count is up!  Very little has changed from the 25% count - Stojansek is now 16 behind Brady and 17 behind Williams.  If she goes out this currently triggers a cascade of surpluses electing five councillors with Spencer then beating Brady by about 40 votes (a margin that seems unlikely to shift but still might - in the 20% count Brady was winning).  Two of those candidates, Cox and Soward, will win whoever goes out, but I am not sure whether Williams going out instead of Stojansek would still necessarily leave Brady as the loser.

Overall this is where Councils are at:

20% count: Clarence, Hobart, Huon Valley, Kingborough

50% count: Brighton, Central Highlands, Derwent Valley, Devonport, Launceston, Meander, Northern Midlands, West Coast, West Tamar

Partway through manual count: Burnie, Central Coast, Waratah-Wynyard

The others have all finished.

Mayor Stats:

There is one mayoralty still unresolved according to the TEC website (update: Southern Midlands - Alex Green has won not Bisdee as per media reports yesterday) but some summary comments are worthwhile while waiting for this one.  

Five incumbent Mayors have been defeated - Duncan McFie on King Island, Michael Kent in Glamorgan-Spring Bay, Steve Wass in Kingborough, Ron Christie in Hobart, Tony Bisdee in Southern Midlands.  Only Bisdee made the final two, with McFie finishing 3rd out of 4 on primaries (count went no further), Kent finishing 4th out of 4, Wass 3rd out of 4 and Christie 7th out of 11.  Excepting McFie and Bisdee, the others were the three incumbent mayors who did not top their councils on primaries last time (Christie was elevated mid-term from Deputy.)

There were a large number of vacant mayoralties at this election (ten).  Five were won by men and five by women; previously the split had been 6-4.  That plus the wins by women in Glamorgan-Spring Bay and Hobart means that the number of female mayors has increased by three to eleven.

Two Councillors won vacant mayoralties from off council - Annie Revie (Flinders) and TFGA president Wayne Johnston (Meander Valley).  

Most Mayors elected in a contested ballot are currently topping the poll in their councillor counts, except in the following: Derwent Valley, Glamorgan-Spring Bay, Huon Valley, Northern Midlands, Southern Midlands (in all of these the Mayor is running second, except Southern Midlands where Green was third).  

Wednesday 9:47: Chong has beaten Walker by just 80 votes, 8602 (50.23%) - 8522 (49.77).  

Wednesday 8:08

Chong 38.53
Walker 32.20
Ewington 29.28

Walker survives.  I don't know a lot about Ewington's politics as to whether Walker has any hope here, but it's mysterious that he's made the final two for Deputy while only being provisionally 11th as a Councillor.

Wednesday 6:20 The crazy Clarence deputy race continues:

Chong 26.56
Walker 25.91
Ewington 24.78
Von Bertouch 22.76 (eliminated)

I would expect Von Bertouch's preferences to favour Chong.  If the final two is Chong vs Walker then I like Chong's chances, but if Ewington gets ahead of Walker (which is quite possible) then Walker's preferences might flow strongly to Ewington - even so I think Chong is best placed, but we'll see.
Wednesday 5:45 Launceston closed up a lot on the final exclusion but van Zetten still holds on to defeat Finlay 51.8-48.2. 

Wednesday 4:48 Clarence deputy fight still going - Chong a yawning chasm of 2.3% ahead now.  As for Launceston go back to sleep on that one as Harris's preferences flowed weakly so van Zetten is too far ahead.

Wednesday 3:30 Derwent Valley Mayor is final with a margin of 25 votes.

Wednesday 3:26 Doug Chipman returned as Mayor, defeats Tony Mulder 56-44.

Wednesday 3:20 May still be some life in Launceston Mayor because one of the two remaining candidates to be cut is a Greens staffer (Tim Harris) whose preferences might well flow strongly to Finlay.  It would have to be very strong to bridge the gap to van Zetten though.

Wednesday 3:15 The Clarence Deputy race of musical chairs continues with Castillo next to go out in sixth place and the top three covered by just over 1%.

Wednesday 2:55 Was asked for a comment on Brighton where Tony Foster has again been re-elected massively.  The Brighton count has reached the 50% stage and the margins at the end are not close (nearly 50 votes between 9th and 10th).  So I think there's a good chance that result will hold.

Wednesday 2:34 Janie Finlay has been making appreciable gains in the Launceston mayor race and would now need a 65-35 split off the remaining three candidates if no votes exhausted.  Some will exhaust making this more difficult.  I don't know anything about the politics of the remaining candidates.

Wednesday 1:36 Candidates continue to be cut out from the bottom in the very close Clarence deputy race.  The Mayor race there is being decided on Beth Warren's preferences, which would have to flow extremely strongly to Tony Mulder to threaten Doug Chipman (and I assume they will not.)

Wednesday 1:15 Flinders Island is one of the first to have a finished count.

12:40 Break O'Day and Glamorgan-Spring Bay will not get 20% counts tonight.  The only other council now still waiting is Hobart.

12:32 Incumbents in trouble in Launceston - in the 20% count there Ted Sands, Karina Stojansek and Simon Wood are all not winning.  However at the point where Stojansek goes out she is only 10 votes behind Emma Williams, so perhaps if Williams goes out there instead Stojansek might still win. There are other candidates who could also go out at this point so we need to see more counting at least re Williams vs Stojansek and possibly others elected near the tail end on the current count.

12:24 Dick Adams defeated - the former Lyons MHR has been beaten as Northern Midlands mayor by Mary Knowles on a strong preference flow from Jan Davis.

12:16  The Clarence 20% count is up.  Former state MP Daniel Hulme, who failed to submit a candidate statement, looks gone.  The race for the final seat is currently very close between McFarlane (currently winning) and Blomely (currently missing out).  However at an early junction Viney is excluded a fraction of a vote behind Blomeley and about 11 votes behind McFarlane, so McFarlane has two critical points to try to stay ahead at in the future counts.  I don't know what happens if Viney gets through that exclusion point, so at least the last seat is in serious doubt.  The rest look reasonably comfortable.  

12:02 Finally mayoral figures for Launceston where Albert van Zetten leads Janie Finlay by around 39% to 30%, a gap that looks too large to close especially with most of the other candidates being male.  

In Huon Valley, it looks like Bec Enders will be the new Mayor.  She has 48.08% with 15.13% from Mike Wilson to distribute, some of which will presumably exhaust.

I just checked and the following councils still didn't have 20% councillor counts up: Break O'Day, Launceston, Clarence, Hobart, Devonport, Glamorgan-Spring Bay, Latrobe.  

11:44 A very close mayoral contest in Derwent Valley with Ben Shaw currently beating Paul Belcher by 25 votes - I expect this is subject to further rechecking.

11:37 Sorry to get hopes up for Clarence folk - still waiting for the 20% count to come through to the TEC website.

11:28 Total lack of numbers from Launceston is causing some frustration on social media.

11:19 Janet Lambert, who very narrowly missed a seat for Labor in the state election then was also quite close in the LegCo, has missed out for the third time this year by 110 votes for Mayor of Northern Midlands. (c. 51-49 2CP).

10:54 Hearing Clarence 20% count could be out within half an hour.  Kingborough is already up with veteran Flora Fox currently out by nearly 40 votes. Winter looks like he will have his running mate Jo Westwood as his deputy.

10:09 Controversial sitting Mayor of Glamorgan-Spring Bay Michael Kent is out as he has finished 4th out of 4 on primaries.  He was the only Mayor elected from off council (having never been on the council) last time.

9:46 Dean Winter has presumably won Kingborough mayor with a 46% primary.

8:37 Doug Chipman has a good lead over Tony Mulder for Clarence mayor so will presumably retain there.

Tuesday 8:25: TEC site back up now.  In Circular Head, Daryl Quilliam who ran against Ruth Forrest for LegCo has been given a decent shake as Mayor but will probably get up after preferences.

Tuesday 7:15  The TEC website is down but some figures are being posted on the TEC Facebook page.  Very close fight for Clarence deputy!

Tuesday 2:30: All councillor counts are being data-entered except the following: Burnie, Central Coast, Circular Head, Waratah-Wynyard.  The TEC aims to have 20% counts for all councils tonight, with a degree of prioritisation on finishing counts for the smaller councils quickly after that.

Tuesday 10 am: Every council breaks 50%.  As of yesterday every council had passed 50% turnout.  This previously happened only in 1996 and 2009.  As of yesterday's figures about 4% needed to be added in final counting to break the 1996 state turnout record of 59.5%, so the turnout looks like being close to a record level. 

----------------------------------------------------------------------------

Detailed intro comments: about the count

Automated data entry for Hobart councillors is scheduled to start from 11 am Tuesday (this might be optimistic) but don't expect official results until the first posting on Tuesday night.  I may have some early rough indications from scrutineering before then, depending on the practicality of sampling, and my attentions will be divided between this count and the mayoral and Deputy ballots.

The counts for Lord Mayor and Deputy Lord Mayor (and for mayors and deputy mayors of other councils) will be conducted by hand with counting expected to start from about 2:30 pm Tuesday.  Because of the large number of candidates for both positions these will be slow counts with long preference distributions that I expect to take at least most of tomorrow.  I expect by the end of tonight we will know the frontrunners for the leadership positions but we may have to wait into tomorrow to get much idea of preference flows.  The LM/DLM counts can be difficult for scrutineers to sample because of the way votes are thrown.

For automated entry councils at a couple of stages of the count the Electoral Commission will post current primaries, together with a preference distribution based on the votes available so far.  In 2014 this happened with around 20% counted on Tuesday night, with about 40-50% counted around Wednesday night, and then with a final release at the end of counting late in the week or perhaps Saturday.  If I become aware of any different schedules for this year  I'll edit this post. Between the second release and the final release (when they push The Button and the winners are determined, after a certain artificially programmed dramatic pause for effect - I'm not kidding!) there will apparently be no more releases of progressive primary figures.

The 20% and 45%-ish counts will be indicative only.  For starters, if a candidate is projected to win by fewer than 100 votes based on 20% of the count, it's quite possible that by normal random sample error, that candidate will actually not win.  It might be possible based on these results to call, say, nine of the twelve seats, but not the last few.  They will be waiting a few days.

Random sample error isn't the only reason why results will change.  There's another one, and here it is:

Beware, Big Word: Stratification!

It's very important to be aware that interim primary vote figures for the various Councils using automated data entry won't necessarily be unbiased samples of the whole count.  Votes do not reach the Electoral Commission in random order, but rather arrive on particular days, and within each day's sample they may be further stratified by post office of origin (for instance).  There are some political differences between the pool of voters who tend to vote early in the count and those who vote late in the count. Also, campaign events that happen during the postal voting period can have more impact during some parts of the count than others.  Those who've followed Wentworth (where these changes have been worth 20% on a two-candidate basis!) will be aware of this issue.

The TEC may well make vigorous efforts to reduce this stratification, but nothing will get rid of it completely.

Hence when the TEC release primary figures for various percentages be aware that these are not true representative samples and that past experience shows that substantial changes may occur.  (For instance in 2014 one candidate for one council who was not getting elected on the 20% sample was elected in third place, though that's an extreme example.)  Any result that is even remotely close will therefore not be called on 20% of primaries. 

Please note that I only "call" a seat when I consider it virtually certain that the candidate in question has won.  Calls for Hobart seats will appear in the current summary.  I ask that people not describe me as having called any seat where I have not explicitly used the word "called" for it.

18 comments:

  1. Do you have any idea what the success rate is for incumbents being returned to council? Seems like it would be very high...

    ReplyDelete
    Replies
    1. Haven't checked across all councils. In Hobart historically defeats are pretty uncommon but there were a couple last time.

      Delete
  2. Hi Kevin,

    How is Brendon Blomeley going in Clarence?

    ReplyDelete
    Replies
    1. He has only 8.5% for Deputy - Councillor not up yet.

      Delete
    2. (I say only but that is only 4 points off the lead - very packed Deputy race.)

      Delete
  3. Can you please provide any commentary on the Brighton Council tally?

    ReplyDelete
    Replies
    1. With 50% counted there the margins between the nine who are shown as "elected" and the rest are quite large for the sample size (the smallest is just under 50 votes). So while it's possible someone else will get up I'd say there's less doubt there than for other councils.

      Delete
  4. Thanks Dr Bonham! Can you elaborate at all on the mix of reasons for votes being ruled informal - numbering errors vs. failure to sign vs. deliberate informal etc. (I imagine the proportion of the latter is smaller than in compulsory ballots?) How do you believe the legislation needs to change? Thanks, Sarah

    ReplyDelete
  5. Failure to sign does not count as informal, it counts as a preliminary rejection (ditto for returning the envelope without the flap). In 2014 the biggest group of informals was repeated numbers prior to the number of boxes required for that council, and the next biggest was omitted numbers. These are both unintentional errors. Deliberate informals (blank and "scribble") were about one sixth of all informals, and yes it is smaller than in compulsory ballots.

    The solution is to change the legislation to provide for at least one of the following:

    * Reducing the number of boxes openly required for a valid vote (for instance to 5 consistent with state elections).

    * Create savings provisions so that, for instance, while voters are told to number a certain number of boxes, in practice anything with a unique 1 is accepted as a primary vote and stays in the count until the point of the first error.

    I would probably combine the two. However voters should still be encouraged to number as many boxes as possible.

    ReplyDelete
    Replies
    1. Thanks, interesting. I've since read your post "Council Voting - Please be Careful!", too.

      Delete
  6. On the provisional results, I have run some numbers which are interesting.

    As you said, we now have 13 new mayors and we also have 21 (!) new deputy mayors across the state. 11 mayors and 13 deputies are women, up from 8 and 10 previously.

    On the provisional councillor results, wait for it, a whopping 40.3% (that is, 106) of elected councillors are new. Certainly this puts paid to the notion that councils and councillors are stagnant. Only West Tamar had no new councillors, and Kentish had only one. Notably, Flinders Island council had only 2 councillors reelected, with 5 new councillors (including the only off-council Mayor, except Huon). Four other councils had 5 new councillors. There were 5 Deputy Mayors elected from off-council too, plus Huon where both Mayor, Deputy and all but one councillor did not serve on the sacked council.

    More good news is that the number of women councillors has increased from 83 (31.56%) to 98 (37.26%). No council is without women (previously, Southern Midlands was) but West Coast has only one, Robyn Gerrity (the other incumbent woman councillor retired and no other women ran for election). Flinders Council, which previously had 1 woman, now has 4 and is the only council (other than Glenorchy) with a majority of women. 3 further councils have a 50-50 split.

    Obviously these are not full figures but it's interesting to see that these elections have been pretty monumentous. I believe in the last elections we actually went very slightly down in the number of women elected (not sure about new councillors), so this is good and surprising news.

    ReplyDelete
  7. Would it be fair to use the word 'landslide' to describe Anna Reynolds win (esp. from such a large field )?
    And even though the cable car issue may have inflated the vote of Greens and green candidates in Hobart, what do you make of the vote of Green endorsed candidates in other councils?

    ReplyDelete
    Replies
    1. Yes it would; it's a massive margin for a new Mayor, the left didn't even make the final two in 2014.

      Looks to me like the endorsed Greens are getting two on Kingborough, 1-2 on Clarence, 1 on Huon Valley, 1 currently ahead but shaky on Launceston, 0 on Devonport and West Coast. This as well as 2-3 on Hobart except that the 3rd has quit the party. I might have missed some. Competition from ex-Greens is a problem on some of these Councils so while these results mostly seem a bit average it's also not as bad as it could have been. The Hobart vote is really good given the competition from Reynolds and shows the strength of feeling against the cable car.

      Delete
  8. What happens if there is a 6-6 result when it comes to a final vote on the cable car? I am assuming a council composed of 7 anti cable car councilors are needed to block the thing for good, yet at the moment, everything would have to go right to get the 7 up?

    ReplyDelete
    Replies
    1. Tied motions are lost so 6-6 means application isn't approved. Also any vote is only likely to be binding on the current council, if it can't get anywhere on this one it can always have another go in 2022.

      Note that there are likely to be multiple votes affecting the project - development application, landowner consent etc. Also it might in theory be rejected then resubmitted with changes for a new vote. I don't think 7 councillors can simply take it off the table for the whole term (at least given the current discussion about the Facilitation Act extending to all council land).

      Delete
  9. Now that the final councillors have been decided, I can update the numbers I provided before.
    - We have 14 new mayors, and 11 total female mayors.
    - We have 21 new deputy mayors, and 13 total women.

    There are a total of 263 councillors, of these:
    - 108 are newly elected (41.06%) and 155 re-elected.
    - 104 are women (39.54%) and 104 are men.
    - This is an increase of 21 women from last time.

    ReplyDelete