Sunday, March 20, 2022

South Australia 2022: Classic Postcounts and Legislative Council


On this page:

Expected Labor gain vs Liberal: Gibson (included in tally above)

Expected Liberal retains: Heysen, Dunstan, Unley, Morialta

On other page: Waite, Finniss, Hammond, Flinders.

This page follows classic (Liberal vs Labor 2PP) seat post-counts and also follows the count for half of the Legislative Council.  The Council count does not currently look all that close, but that might change.

Three seats (Hartley, Morialta and Unley) dropped back into the in doubt category on the ABC site overnight.  These are sites where the Liberals were projecting to win before the projection was switched off and still hold healthy leads; I expect these leads will increase and the Liberals won't be troubled holding these seats.  

(Update Monday: the leads in Unley and Morialta came down with further counting to 50.32-49.68 and 50.50-49.50.  It is most likely the Liberals will pull ahead or at worst break even on prepolls and postals so I am still treating those as expected Liberal holds for now, but may add sections below should this not happen as the count goes up or if there are changes to figures.)

(Update Tuesday: The Liberals have surged to the mid-51s in Unley and Morialta putting those beyond doubt.) 

Dunstan (Lib 7.4%)

This is the seat of outgoing Premier Steven Marshall and at present the voters there are precariously close to saving themselves from a likely by-election.  At various stages on the night the ABC computer called the seat for both sides.  Marshall leads Labor's Cressida O'Hanlon (from some accounts an impressive candidate) by just 96 votes (50.33% 2PP) with 54.5% counted.  Postals and prepolls are expected to favour him and boost that to 51.0% or so but when it's this close I'll keep an eye on that.

Sunday 4:05 The addition of ticket votes has put O'Hanlon ahead by 143 (50.49%).  These are votes that would be informal but for SA's savings provisions, which allocate informal votes that are consistent with a lodged preference allocation apart from the error to be transferred as if the voter had followed that lodged preference allocation.  This makes the race much closer - Marshall is still projected to take the lead based on prepolls and postals (especially if one uses Boundary Commission estimates not ABC's) but there is always significant doubt when the other side is this far in front.  

Monday night: With the count up to 60% most of O'Hanlon's lead is gone already, now down to 29.  (50.09%).  I expect Marshall to recover the lead soon.  

Tuesday afternoon: ALP lead down to nine votes.

Wednesday: As expected Marshall now ahead, by 184 votes (50.49%).  69.8% now counted (note that this count finishes around 90 not 100).

Thursday: Marshall has gone over 51 and has clearly retained.

Heysen (Lib 7.6%)

Also in the vulnerable 6-to-10% range is former Speaker Josh Teague's seat of Heysen.  Teague leads Labor's Rowan Voogt with 50.86% 2PP, which the ABC was projecting to come up only into the low 51s.  (61.9% counted).  Most likely a Liberal retain but I'll keep an eye on it for a bit.

Sunday night: Teague has jumped out to 51.6 so I'm assuming this is over and will only follow it again if it closes up.  

Gibson (Lib 10.0%)

Looks like a boilover - this one was close-ish at times but then suddenly, apparently, jumped into the Labor column late at night.  Labor's Sarah Andrews has 51.59% against Liberal incumbent Corey Wingard with 53% counted.  This was projected by the ABC to come down to 50.8% or so but is unlikely to come out to anything that can be pulled back, so probably seat 26 to Labor.

Sunday 4 pm: The ABC has called Gibson, as there is a counting error and the correction will place Labor's lead out of reach.  There are also issues with the count in Unley which I am still assuming the Liberals will hold pending further details.

Legislative Council

The Legislative Council count has been building through the night and has reached 52.8% counted.  The following are the leading parties, in quota totals (one quota = 1/12th of the total, or 8.33%):
Labor 4.453, Liberal 3.960, Green 1.176, PHON 0.503, Lib Dems 0.418, Family First 0.390, Legalise Cannabis 0.266 (various others 0.834 combined)

Labor will win four, the Liberals four, the Greens 1.  At present in the race for the final two seats Labor and PHON are in front.  Labor will also benefit from Greens preferences.  The Liberal Democrats are in notional 12th place but they will probably be overtaken on preferences by Family First.  

I have not yet looked at past evidence of leakage that might harm Labor, but even if they were to start slightly behind the Liberal Democrats, the Lib Dems tend to crawl on preferences in these things and will probably be beatable.  The new Family First however is the smokey here if Labor falls back significantly.  It's not impossible that they'll do OK on preferences among the various woolly centre parties that make up the flotsam and jetsam at the tail end of the party table.  All the same at the moment it would be brave to bet against 5-4 plus Greens and One Nation.  

Updates to follow.  

Tuesday: At 57% counted, very minor changes.  The Liberals are up to 4.034, the Greens are down to 1.14, and the competitive micro parties are all down very slightly.  Again 5-4-1 and One Nation looks likely.  Checking the 2018 spreadsheet I find that Labor lost a net 0.013 quotas on leakage during the throws of various surpluses, which is currently not enough to put them notionally behind LDP or FF.

Thursday: All continues to point to 5-4-1 and One Nation.

Saturday: Labor are falling back to a degree with 87% counted (now Labor 4.432 Liberals 4.141 Green 1.096 PHON 0.498 LDP 0.394 FF 0.365.)  Also the Green surplus is shrinking, though there are also Legalize Cannabis and AJP preferences that I would expect to help Labor.   SA has fully optional preferencing so preference flows are not likely to be very significant here.  

Sunday: See Antony Green's analysis of why the leaders are likely to retain their leads.  Labor outperformed Australian Conservatives (which was a rebadge of the old Family First) on preferences last time with far weaker preference sources and FF are not preferenced by anybody on HTV cards (which will have little follow rate aside from the Liberals' anyway.)  However the revival of the gimmicky FF name may improve their flows to some degree.  

2 April: Labor's lead has shrunk a little, now Labor 4.416 Liberals 4.142 Green 1.087 PHON 0.507 LDP 0.394 FF 0.366 LC 0.250 AJP 0.179.  This is apparently the full count with only any minor changes on data entry to come.   This lead is small enough to not be too confident about Labor's win given how little is known about the properties of this system (now in its second election) and especially since there has not been a Family First presence under that name before, and the new FFP is a Labor breakaway.  If FFP overtake LDP it could be interesting.  However, One Nation will not hit quota so ultimately left preferences should pool with Labor while right preferences remain split two ways, making it much harder for either of the micro-right parties to win.

27 April: The button was pressed with One Nation and Labor winning easily for the expected 5-4-1-1 result, increasing their leads over the LDP with FF making some inroads vs One Nation but not nearly enough to catch up.  See Antony Green's summary.  


  1. Given the bloodbath it looked like on election night, the Libs would probably take 16 seats as a 'not too bad' outcome in the end.

  2. 2/04 comment - so you are saying it looks like neither LDP or FF will get enough preferences to overtake PHON to win?

    1. My current view is that if one does win it is more likely at the expense of Labor, but I don't think either will win. PHON may crawl on preferences but appear to be far enough ahead.