Blogging Senate forecasts and results in the WA Senate re-election until officially declared.

Twitter: @AU_Truth_Seeker


Thursday 5 September 2013

Tasmanian Senate - TruthSeeker forecasts

Here’s the fourth state in my senate series – Tasmania.

UPDATE: I have run an alternative scenario based on a LIB primary vote that is 1% lower. See below for the outcomes of this alternative scenario.

Most likely elected Senate (TAS):



1 GRN, 2 ALP, 3 LNP

I am constantly updating my calculation of primary votes to allow for the latest polls, and slightly tweaked my method for calculating minor party support to allow for more national consistency. Variation is also slightly higher to represent more uncertainty than was previously allowed for.

In particular, the Tasmanian situation is very hard to call, given the small size of the state make it near impossible for any meaningful polls to be conducted. Of the most extensively conducted polls, Reachtel has been inconsistent and downplayed by some. Furthermore, applying House swings to the Senate in TAS is also flawed as in one of the five HoR seats, an independent will poll incredibly well, reducing the accuracy of any Senate estimates. So, avid readers will note I have applied a large +/-10% variance to ALP and LNP. This means the model will not produce a 100% likelihood result unless it is extremely likely that this is the case. Anyway… here are the numbers.


Primary votes:
LIB:       39.3% (+/-3.9%)
ALP:      31.2% (+/-3.1%)
GRN:      14.4% (+/-2.8%)
Minor parties:       15% comprising selected parties:
PUP:      3.3% (+/-1.6%)
SXP:      2.2% (+/-1.1%)
FF:      1.1% (+/-0.5%)
Shooters and Fishers:      1.6% (+/-0.8%)
AFLP:      0.6% (+/-0.3%)
Country Alliance:      0.8% (+/-0.4%)

Likelihood of election:
LIB:      2 elected
ALP:      2 elected
GRN:      1 elected
(Model gives a 99% likelihood to the first 5 being elected)
LIB (3rd candidate):      41%
FF:      28%
SXP:      23%
Shooters and Fishers:      7%
<1% likelihood of election: PUP, AFLP, Country Alliance.

Ideological splits:
4 Left, 2 Right: 23% likely
3 Left, 3 Right: 77% likely

Summary:
- Two ALP, 2 LIB and 1 GRN elected off primaries
- The sixth spot becomes a lottery – I feel like I’m saying this far too frequently.
- Most likely is a 3rd Liberal, with strong outside chances to FF and SXP.
- The Country Alliance vote is high due to the Group A donkey vote.
- There has been a lot of talk about FF in Tasmania, but what’s been overlooked is the SXP. They draw votes from ALP and GRN surpluses and the Left parties, but they also steal votes from the right via Country Alliance, Shooters and Fishers(!) and the Smokers.

As above, and as stated previously my personal expectation is a straight 3-2-1 split. But in applying uncertainty owing to the lack of credible polls, this opens the door for the minor parties. Statistically, I just cannot be certain of the votes of each party and have to apply a large variance as my model foresees unlikely occurrences.

Ironically, if the Liberal Party doesn’t poll enough votes, the 6th spot may be won by either the Sex Party or Family First. If it is SXP, then justice will not have been served as a 4-2 L-R split will prevail, due to votes of conservative groups being “stolen” to elect a party abhorred by conservatives.
In Tasmania, ALP winning 1/5 HoR seats yet the Left fluking 4/6 Senate seats would represent a failure of democracy.

UPDATE: Various other scenarios have been run in Tasmania. Given LIB, ALP and GRN are all on the cusp of obtaining outright quota, the model is sensitive to any change in primary vote. In particular, one scenario assumed a 1% drop in the LIB primary vote and this will result in FF being more likely than LIB to be elected to the 6th spot. Under this alternative scenario, FF gets elected 50% of the time, LIB3 elected 22%, Shooters 11%, Sex Party 10%, Others 7%

Over the coming days, I will publish equivalent summaries for other states. I won’t talk about the methodology – this is largely covered in prior posts. On election day, I will republish my full final results, updated for latest polling data.


Summaries so far:
TAS: 1 GRN, 2 ALP, 3 LNP
WA: 1 GRN, 2 ALP, 3 LNP
SA: 1 GRN, 2 ALP, 1 XEN, 2 LNP
VIC: 1 GRN, 2 ALP, 2 LNP, 1 FF
Running total: 4 GRN, 8 ALP, 1 XEN, 10 LNP, 1 FF

9 comments:

  1. Hi Truthseeker, did you split the Fishing vote in half and distribute it to the new Fishing party? I get an interesting result.

    ReplyDelete
    Replies
    1. I have given 1.64 to Shooters and fishers, and 0.55% for AFLP. As a curiosity, I have re-run my Monte Carlo simulations with 1.1% for each of the two parties.

      I get similar results for ALP, LIB, GRN, SXP, FF... But suddenly PUP's in the game with a 1% chance of election, and AFLP is up to 3%. What outcome do you get?

      Delete
  2. This election is turning up some interesting possibilities. In previous decades, the Democrats would have got those seats in the middle.

    The strangest possibilities are the ones most likely to be thwarted by below-the-line votes. And of course those takes the longest to count, so we may be waiting all next week to see which of FF, Sex Party, No Carbon Tax, Xenophon, PUP and KAP gets seats in the Senate.

    ReplyDelete
    Replies
    1. If you guys are still visiting the site, I'll consider running it post election with updates, based on my Monte Carlo simulations.

      And Tas has a stronger history of BTL voting, owing to its State voting system... If anything, this would tend to favour a 2-3-1 split over a harvesting miracle by FF or SXP.

      Delete
    2. Tas has a strong history of BTL Senate votes but they've never had anywhere near 54 candidates. It's one thing to mark 20 preferences; it's an entirely different thing to mark at least 49.

      We should see a far lower BTL percentage (I'd suggest probably in the 5-10% range) and a far higher informality rate than has been the case in the past. My guess is that this would hurt the Greens, who've benefited from Labor voter ticket splitting in the past (of course, they had Bob Brown in the past).

      Kevin Bonham's blog post has apparently been making the rounds and it's possible that could influence people to reconsider voting below the line. It's also likely to benefit the Sex Party, as they're the only ATL ticket that does not have a shot of electing Family First. My calculations suggest that the Sex Party will win a seat with a primary vote of 2.3% if the Liberals are under 40% (likely) and Labor+Greens are at 48% or higher (less likely but absolutely possible).

      Delete
  3. Truthseeker, thanks for the great site - I can't wait for tomorrow's update. Maybe you could include the continuing senators as well?
    TAS: 1 GRN, 2 ALP, 3 LNP
    WA: 1 GRN, 2 ALP, 3 LNP
    SA: 1 GRN, 2 ALP, 1 XEN, 2 LNP
    VIC: 1 GRN, 2 ALP, 2 LNP, 1 FF
    Continuing: 6 GRN, 13 ALP, 16 LNP, 1 DLP
    Running total: 10 GRN, 21 ALP, 1 XEN, 26 LNP, 1 DLP, 1 FF

    ReplyDelete
  4. Replies
    1. 1.1%, or a range from 0.55% to 1.65%. I think LDP would be newer for Tasmanians, who have sensibly given low votes to the other name-confusing party - DLP. The LDP doesn't trouble the scorer in Tasmania in any of my modelling.

      Delete