Friday, August 21, 2020

NT Election Final Roundup

Counting night approaches for the 2020 Northern Territory election, which I will be covering with live comments here in a separate post and follow-up comments on that thread through the post-count as required.   Normally here for an election I have paid a lot of attention to I would be posting a final polling roundup, but for the NT there's usually precious little to report in that department.  However there are a number of other things that can be commented on in setting the picture.  You can also check out the Tally Room podcast I appeared in today with Ben Raue and Duncan McDonnell.


Almost everything I had to say about polling and this election was said in my earlier piece Divergent Polling In The Northern Territory.  I contrasted a very detailed Territory Alliance internal poll by Mediareach with a lobby group commissioned poll by uComms and found that whereas the former predicted a chaotic parliament with no party likely to be near a majority, the latter was most consistent with a Country Liberal resurgence, Labor retaining a majority, and little joy for the Territory Alliance.  I did not asset that either poll was accurate (though uComms have a mostly good track record) - the truth could be much closer to one than the other, or somewhere in between, or more extreme in some direction than both polls.  Since the article was written, expectations and betting seem to have converged towards the uComms result, but that doesn't necessarily mean that's what's going to happen.

It would, however, be very consistent with my finding in the previous piece that "federal drag" theory (the idea that being of a different party to the one in Canberra helps) is a thing in the NT just as it is in the states.  As bad as the NT economy has been during the present term, a majority government thrown out after one term while the same party is in opposition federally would be a very unusual result from that perspective.  

Beyond a few mutters about internal polling that add little to what was already being said about it, there has been no further sample-based polling.  An unscientific NT News opt-in is currently running at 33% Gunner 56% Finocchiaro and 11% Mills as preferred chief minister, which would strongly reflect the skew of its conservative reader-base and means absolutely nothing as far as the Labor vs CLP contest is concerned.  However the gap between Finocchiaro and Mills might be taken as consistent with the view that TAP has gone off the boil during the campaign. 


I have repeatedly made dismissive noises about the value of betting in predicting federal elections - it seems to be strongly influenced by polls even when the polls are wrong, and fails to improve predictions that could be made using polling aggregates and a simple pendulum model.  However, the NT's lack of polling data makes it a place where betting might be one of the better predictors out there.  Or it might be useless - we will see.  

Labor has continued to firm through the election, now to around $1.35, while the CLP is at $4 and the Territory Alliance has drifted out to $8-$8.50.  Markets on total vote range project primary votes around 37% for Labor, 28% CLP, 18% Territory Alliance, 4% Greens, maybe consistent with about a 52.5% 2PP for Labor assuming the TA preferences split 2 to 1 in the CLP's favour.  This may or may not be in any way accurate, I am just reporting the current market projection.

In seat betting Labor has at times been favourite in as many as 18 seats at once. Currently a market has Labor favourites in all seats except Araluen, Blain and Namatjira (TAP), Brennan, Braitling, Port Darwin, Nelson, Katherine and Spillett (CLP) and Goyder (Purick).  However, the market has significant doubts about nine of Labor's projected 15 wins, while only giving them serious chances in four of their projected non-wins, so on balance it is probably only projecting around 13.5 Labor seats (still a majority, just).  The CLP in contrast would be likely to pick up around 7.5 seats if these odds were correct, because there are many seats it is not favourite in but is second favourite.

[Update 22/8 4:15 pm Labor is now favourite in Port Darwin, but no longer favourite in Daly.]

Seat groupings for election night

There are very few boring seats at this election.  Rather than look at certain seats as key seats I find it useful to break the seats into clusters, and will be using this to inform my election night coverage:

1.  Labor seats on small or negative margins (5): Namatjira (notional CLP), Katherine, Braitling, Brennan, Port Darwin.

2. Labor seats on higher margins that are interesting (7): Drysdale (modest margin), Arafura (modest margin), Arnhem (independent opponent), Johnston (close by-election, which would have been closer but for preference fights between the CLP and TAP - that said, any seat a government can win at a by-election it should hold in a general election), Fong Lim (occupied by defector to TAP), Fannie Bay (Chief Minister at times unpopular and has a key CLP opponent), Barkly (ALP retirement and high-profile CLP candidate, despite huge margin)

3. Labor seats on higher margins that appear safer on paper (6): Sanderson (though some speculation about even this), Casuarina, Karama, Gwoja, Wanguri, Nightcliff

4. Non-Labor seats where Labor appears competitive (3): Blain, Mulka. Daly

5. Non-Labor seats that are interesting but where Labor appears uncompetitive (3): Araluen, Nelson, Goyder (mostly expected to be an easy Purick retain)

6. Safe CLP seat (1): Spillett

All up, in about 60% of the seats Labor has some sort of chance on paper, but so do others.


As of today, at least 53.5% of enrolled voters had already voted, not counting postals that are still to arrive.  Turnout has averaged about 75.5% in recent elections, so unless turnout increases at this election we could see well below 20% of enrolled electors coming out on polling day (compared to 28.1% in 2016).  Something to keep an eye on as numbers come in.  See Antony Green's detailed coverage.

Live comments on the election will start here at 6 pm ACST (6:30 AEST). 

1 comment:

  1. I tried to put money on the overall result about 10 days ago and was told the maximum permitted was $25. In recent years election betting has become a waste of time. Very small amounts permitted, yet these tiny amounts move the odds significantly.