Sunday, March 24, 2019

2019 New South Wales Wrap And Lower House Postcount

The Berejilklian Government has been returned with a very small majority

Expected final result Coalition 48 Labor 36 Green 3 Ind 3 SF+F 3 

Seats changing hands
Barwon (Nat to SF+F)
Murray (Nat to SF+F)
Coogee (Liberal to Labor)
Lismore (Nat to ALP)  link to Lismore thread


This thread will give some summary remarks about the NSW Lower House count (which may be updated if necessary) and will also follow the postcount in a few undecided seats where only two candidates are in contention.  Lismore has thrown up another three-cornered contest that is of special interest and it will get its own thread.


The Coalition government led by Gladys Berejiklian has won a third term in office and at the end of the first night of counting appears to have won a very small majority, although this remains to be confirmed.  This is the first case of a state Coalition government winning a third or later term since the Queensland National government of Sir Joh Bjelke-Petersen won an eleventh term in 1986.  If the result is a majority win, it will be the first time since Tasmania in 1979 that a party has won a majority under its third Premier in the same spell in office.

The election seems to have seen a very low turnover of seats, with the government apparently dropping two rural seats to the Shooters, Fishers and Farmers and one city seat to Labor, with a fair chance of losing a regional seat (Lismore) to Labor or, less likely, the Greens.  Currently, the two-party preferred swing on the ABC's projected totals is running at only 1.5% to Labor in the 77 seats for which there are classic (Coalition vs Labor) 2PPs.  That would leave the Coalition with an impressive 2PP vote in the high 52s (say 52.8) but I expect that to come down to maybe 52-48 once the non-classic seats are included.

The anger in the bush that saw seats fall to the Shooters is so bad that the Nationals have lost Murray with an 8% swing from the mid-term by-election that they won.  Normally by-elections have amplified swings compared to the main event so this is quite extraordinary.

There is currently a high degree of swing variation between seats (a standard deviation of 3.9%) and the Coalition currently has projected swings to it in 26 seats.  On average this would have delivered Labor three or four Coalition seats and the Coalition one Labor seat.  However the swings may have been larger in safe seats as Labor hasn't lost any seats that I can see, while the Coalition has so far only dropped one or two to Labor.  As well as there being very few seats changing hands, there are also very few in serious doubt at this stage, though if something odd happens with the prepolls that may change.

The result is better for the Coalition than any of the campaign polling and supports the view that gaffes by the Labor leader Michael Daley in the final week either blew the election or else made it a lot less close than it should have been.  One could draw a line through the Galaxy (ALP 38% primary) to the Newspoll (35%) to the projected final primary (33.3% according to the ABC) and suspect that Labor's vote was still falling in the final days.  As for that Newspoll, as the only pollster to actually bother with a final-week poll, YouGov-Galaxy carried a heavy burden and will be breathing a big sigh of relief since it looks like they did pretty well here.  More on that when all the numbers are in.

At the previous election Labor was up against Mike Baird, one of the most popular leaders the Coalition has had at state level.  Berejiklian's government is four years older and faces federal-government drag factors that are perhaps even more severe than those provided by Tony Abbott in 2015.  It is also not enough to say that the impression of xenophobia on Michael Daley's part was itself the cause, because Labor was also attacked for much more systematic  "yellow peril" campaigning in Luke Foley's 2015 defeat, so that too was baked into the baseline.  Labor's performance here was just simply bad, and the search for Reasons may commence.

As for the government, its re-election despite being supposedly unable to sell itself to save itself does show that Coalition parties can govern from somewhere near the centre without dooming themselves.  It also raises the question of whether the ability to sell a message is as big a deal as many commentators think, or whether a good economic performance sells itself.

Despite being involved in infighting for much of the term, the Greens have come up with a decent result.  They are tracking for a swing of less than 1% against them, have retained their three seats with good swings, and are vaguely in the mix for a fourth.

Postcount: various seats

Some notes will be added here on seats I'm keeping an eye on but that I don't think need a foldout section is needed.

Wollondilly (Liberal vs Ind) sees independent Judy Hannan trying to catch Liberal Nathan Smith.  However Hannan is about 16% behind and currently needs to gain at a rate of 0.417 votes/preference to win.  This might be achievable if the preferences were all Greens and Labor but they also include 11% One Nation, which are likely to exhaust and/or scatter.  Also as prepolls are added the mountain is likely to steepen as independents tend to do badly on prepolls.  So my view is this is unrealistic.

Update Monday 11:00  The required gain rate has increased further to 0.428.  It didn't shoot up as much as I expected but there are not a lot of primaries left to count now and the One Nation votes look like being the killer for any upset chances so I am no longer monitoring this seat.

Penrith is currently projected to 51.2% to Stuart Ayres (Lib) off a current 52.1% 2PP.  I don't think that's going anywhere but will keep an eye on it.

Postcount: Dubbo

In the seat of Dubbo (Nat, 20.4%), Troy Grant has left the building and his successor as Nationals candidate Dugald Saunders has faced a massive challenge from independent Matthew Dickerson, the former mayor of Dubbo.  As I start this section, Saunders has 35.7%, Dickerson 28.5%, the Shooters 14.3%, Country Labor 15.6%, and the Greens, Conservatives and Flux have barely bothered the scorers.  This pans out to a 50.56% two-candidate vote for Saunders at the moment.  However, only 38.5% of the vote has been counted.

Dubbo had a huge prepoll (29.6% of total) in 2015 and looks like it will do so again; none of the prepolls in Dubbo have been counted yet, though some postals have.  Grant did very well on prepolls in 2015, but more to the point, indies tend to do notoriously badly on them.  Let's see some prepolls but I reckon this will be retained.  Updates to follow.

Sunday 5:30: Dickinson has won iVotes cutting Saunders' margin to 125 votes (50.35-49.65).  However the huge Dubbo prepoll has as expected come in strongly for Saunders.  I project that when it is added to 2CP, Saunders' margin will blow out to about 864 (about 51.5-48.5), a lead which I'd expect to be insurmountable.

Monday 4:35 Antony Green has called the seat but there are no new numbers to see, though I too firmly expect Saunders to win.

Monday 9:05 The Dubbo booth has hit the 2CP pushing Saunders' margin to 960 and there is more to come from the Mudgee booth so I have also called this seat and will no longer be tracking it.

Postcount: East Hills

Well, the Galaxy seat poll was right!  East Hills (Lib, 0.4%) should have been one of the first against the wall in 2015 but was held with the assistance of a vicious smear campaign against the Labor candidate Cameron Murphy.  With any decent statewide swing it should have been one of the first to drop this year as well, but there wasn't a decent statewide swing, and the Liberal candidate Wendy Lindsay leads Murphy.  The current margin is 52.4% to 47.6% (1121 votes) but the count is more advanced on the primaries than the 2PP because of prepolls and one booth that are still to be counted to 2PP.  Once these are included I have Murphy about 750 votes behind, and the gap may close further with further prepolls, but he is still a rather long way behind.  The ABC projects the count to finish at about 50.8% to Lindsay.

Sunday: There are also currently two uncounted booths to primary, Bankstown West Public and Padstow Park Public.  In 2015 Labor gained 216 votes in these booths, and with relatively little shift in the primaries one would expect similar this time.  Of the prepolls, we are yet to get anything from East Hills EM.  In 2015, the East Hills prepoll did nothing.  However in 2015 the Bankstown prepoll favoured Labor by 9.7% (primaries) while in 2019 it did so by 19.5%.  So the hope for Murphy is that this sort of superior performance is also seen in the East Hills prepoll, which if so doesn't put him into the lead, but puts him close enough that he might just get there off absents.   On the other hand, there could be nearly 2000 postals to go, and while early postals tend to be more conservative than later ones, the later ones will still probably help Lindsay.

Labor scrutineers have been making confident noises about the seat based on the incomplete prepolls. The question is whether the sample of prepolls this was based off was representative or not, and I suspect the latter.  We will know much more when the prepoll count is finished at least to primary level.

Sunday 5:20: iVotes were added and favoured Labor by 106 votes, a few dozen better than last time.  The ABC's projection has come down to 50.7.

Monday 9:00: The missing East Hills prepoll wasn't a big one after all; the big one this time was Revesby.  Anyway these prepolls are now all in, except for "declared facility".  My crunching of the numbers based on the preference flow so far gets Lindsay winning by about 1120 before postals, probably more after, which leaves me in no doubt she has won the seat.

Threads for Lismore and the Legislative Council will be added on Sunday.

Donations welcome!

If you find my coverage useful please consider donating to support the large amount of time I spend working on this site.  Donations can be made by the Paypal button in the sidebar or email me via the address in my profile for my account details.  Please only donate if you are sure you can afford to do so.


  1. East Hills Labor still in the hunt

    1. No. I thought they had some chance but this was mainly a result of the East Hills EM prepoll being much smaller than last time. I understand that Murphy has conceded.