Whole Nation
With Inserts
Saturday, 2 July 2016
Friday, 1 July 2016
Senate Predictions
Previously we've had a lot of problems in predicting the Senate results. We found out last election that lower house polling is a very poor substitute for the upper house. The number of niche upper house parties means a lot more of the primary vote goes to "others". I was considering giving an arbitrary doubling to the "other" vote and progressively tweaking it (possibly as a function of number of parties running) to get a better approximation, but instead I've decided to try something else out.
The leader satisfaction polling shouldn't have any relevance to the Senate, yet I'm tempted to use it. It's initially counter-intuitive to use the satisfaction of a lower house member not directly elected to his (or her) position to model upper house voting, but it is not improbable that people who oppose a party in the lower house will do so in the upper house. The advantage of using the satisfaction data is that it is not confined to the two major parties and the few other "likely lads" the pollers include (GRN, NXT, PUP). A high dissatisfaction with both parties should presumably correlate to a higher "others" vote and vice versa.
Unfortunately I cannot find any satisfaction polling at a state or territory level. So I'm going to adjust the LIB and ALP primary senate votes from 2013 for each state by the satisfaction swings, and then eyeball the likely "others" vote.
Firstly, the 2pp + others vote from each state:
Source: https://en.wikipedia.org/wiki/Senate_results_for_the_Australian_federal_election,_2013
Now the satisfaction swing. In 2013, directly before the election, Rudd had a 33% approval rating and 58% disapproval, while Abbott had 44% approval and 50% disapproval. Currently, Turnbull has a 40% approval and 47% disapproval while Shorten has 36% approval and 51% disapproval.
In approval that's a 4 percentage point drop (9% drop) down for the Coalition and a 3 percentage point gain (up 9%) for the ALP. On disapproval its a 3pp drop for the Coalition (6%) and 7pp (12%) drop for Labor.
I'll be using approval, since that's what will lure people to put the major parties ahead in their primary vote. So we'll reduce the Coalition vote by 9% of their past share, and boost the ALP by the same, and calculate the other vote as the remainder:
Because this is a double disillusion election, the quota is a reduced 7.69% for the states and the usual 33% in the territories. So on that prediction the senate would get the following quota allocations:
And the senate would look like this on the primary vote:
Now I doubt the "others" in the ACT are organised enough for any one of them to out poll the coalition, so the last seat in the ACT will be blue, too. Last election no party got more than one candidate beyond their primary quotas in, and sometimes the major parties didn't even get that. In a double dissolution maybe you'd get two, but then the "others" might get in a quick seat too. So lets limit it as a max of one extra seat. There's no easy rule to this - in WA the ALP needed just 2% of the states preferences to gain an second seat and didn't, while in NSW the coalition needed 8.66% and got it. As a general rule though, is seems if a candidate needs more than 3pp to meet a quota they'll fail in a normal election. That's about a quarter of a quota in a normal election, or half in a DD. I'm not sure which to use. So lets hand out seats to anyone who got X.66% of a quota:
Now first pick for an "other" in every state except perhaps SA and QLD has to be a Greens. In SA it's be NXT who out-polled the greens (and ALP) in 2013, and in QLD it's either PUP (who also out-polled the greens), or KAP replacing them. Or former PUP Glenn Lazarus. Or Pauline Hanson. Queensland is just not used to this whole upper house shtick is it? I'm backing PUP doing badly given Palmers' press and losing Lazarus to the Glenn Lazarus Team. I suspect one nation will do poorly on preferences, with most supporters putting them in as a primary and most others ignoring them completely. On the tossup, I'll give a seat to GLT.
Even then, with the reduced quota, the Greens can probably primary vote their way into one seat in all of these, and get damn close to 2 in Vic and Tas, so we'll grant them too. Xenophon almost primaried a double quota in 2013 and has been advertising hard in SA. Plus the general dissatisfaction in the major parties should guarantee two seats. And Family First can normally count on getting Bob Day in, so I'll grant that too.
I'm prepared to leave one seat unknown in each state because even without voting tickets the last one will be a random matter of dropping out, overflow quotas and preferences shooting around like ping-pong balls. So that still leaves 2 seats in NSW, 1 in SA and 2 in WA.
WA is conservative, and the NATs run separate to the LIBS. After the two majors and the greens the NATS came fourth in 2013, but the greens still out-polled them, so I'm thinking another green and a NAT? But then maybe a centrist like NXT or PUP could do well there too. What the heck. NAT and PUP it is.
In NSW David Leyonhjelm should have some traction, so the LDP can get a seat. They did out-poll the greens in 2013 but I can't see them getting 2 seats. Both should just get their quota and have so few remainders they fizzle out. PUP was the next biggest in 2013, but other good bets would be Derryn Hinch for publicity, Shooters and Fishers who seem to pull of surprises now and then, Sex party if the overflow is vry left and Rise Up Australia if its very right. Eeny, meeny, miny... PUP.
The last spot in each state, as I said, I'd leave as tossups for the madness that is the minor preference flows. But for a nominal indication I'll give it to a major party, since we now have exhausting tickets and public media is probably going to mean one of them gets a mark somewhere even in the shortest of lists. I'll allocate it based on which of the parties is closest to an extra quota, with some exceptions.
In SA there are two seats to fill and the major parties have their smallest representations of any state. It is true Xenophon has been undercutting them for ages, but with his two seats allocated I'll give them both a seat.
In Queensland the LNQ got it's 5th seat by rounding up its quota and the ALP only just scraped their 4th on primaries. I could see a minor party getting in here instead, with both majors exhausted. It's got to be PUP, KAP or One Nation, and since I was tossing up between KAP and GLT before, KAP can have the position nominally.
Tasmania's double Greens are ambitions, and Jacqui Lambie is not to be underestimated, even if she can be the Donald Trump of Tasmania at times. The ALP has probably maxed out at 5 and the Libs getting a 5th seems unlikely to me, so JLN can have that last seat.
That comes to:
Coalition (LIB, NAT & NLQ): 30
ALP: 29
Cross-bench: 17
For comparison, after the 2013 election it was:
Coalition: 33
ALP: 25
Cross-bench: 18
So, that's my highly dubious guess, with no real faith in the reliability of my tossup fill-ins.
The leader satisfaction polling shouldn't have any relevance to the Senate, yet I'm tempted to use it. It's initially counter-intuitive to use the satisfaction of a lower house member not directly elected to his (or her) position to model upper house voting, but it is not improbable that people who oppose a party in the lower house will do so in the upper house. The advantage of using the satisfaction data is that it is not confined to the two major parties and the few other "likely lads" the pollers include (GRN, NXT, PUP). A high dissatisfaction with both parties should presumably correlate to a higher "others" vote and vice versa.
Unfortunately I cannot find any satisfaction polling at a state or territory level. So I'm going to adjust the LIB and ALP primary senate votes from 2013 for each state by the satisfaction swings, and then eyeball the likely "others" vote.
Firstly, the 2pp + others vote from each state:
| ACT | 2013 Result | NSW | 2013 Result | |
| Coalition | 34.44 | Coalition | 34.20 | |
| ALP | 33.08 | ALP | 31.56 | |
| Others | 32.48 | Others | 34.24 | |
| NT | 2013 Result | Qld | 2013 Result | |
| Coalition | 41.34 | Coalition | 41.39 | |
| ALP | 32.75 | ALP | 28.52 | |
| Others | 25.91 | Others | 30.09 | |
| SA | 2013 Result | Tas | 2013 Result | |
| Liberal | 27.45 | Coalition | 37.51 | |
| ALP | 24.88 | ALP | 32.83 | |
| Others | 47.67 | Others | 29.66 | |
| Vic | 2013 Result | WA | 2013 Result | |
| Coalition | 40.13 | Liberal | 39.20 | |
| ALP | 32.45 | ALP | 26.59 | |
| Others | 27.42 | Others | 46.82 |
Now the satisfaction swing. In 2013, directly before the election, Rudd had a 33% approval rating and 58% disapproval, while Abbott had 44% approval and 50% disapproval. Currently, Turnbull has a 40% approval and 47% disapproval while Shorten has 36% approval and 51% disapproval.
In approval that's a 4 percentage point drop (9% drop) down for the Coalition and a 3 percentage point gain (up 9%) for the ALP. On disapproval its a 3pp drop for the Coalition (6%) and 7pp (12%) drop for Labor.
I'll be using approval, since that's what will lure people to put the major parties ahead in their primary vote. So we'll reduce the Coalition vote by 9% of their past share, and boost the ALP by the same, and calculate the other vote as the remainder:
| ACT | 2013 Result | + Approval | NSW | 2013 Result | + Approval | |
| Coalition | 34.44 | 31.34 | Coalition | 34.20 | 31.12 | |
| ALP | 33.08 | 36.06 | ALP | 31.56 | 34.40 | |
| Others | 32.48 | 32.60 | Others | 34.24 | 34.48 | |
| NT | 2013 Result | + Approval | Qld | 2013 Result | + Approval | |
| Coalition | 41.34 | 37.62 | Coalition | 41.39 | 37.66 | |
| ALP | 32.75 | 35.70 | ALP | 28.52 | 31.09 | |
| Others | 25.91 | 26.68 | Others | 30.09 | 31.25 | |
| SA | 2013 Result | + Approval | Tas | 2013 Result | + Approval | |
| Liberal | 27.45 | 24.98 | Coalition | 37.51 | 34.13 | |
| ALP | 24.88 | 27.12 | ALP | 32.83 | 35.78 | |
| Others | 47.67 | 47.90 | Others | 29.66 | 30.08 | |
| Vic | 2013 Result | + Approval | WA | 2013 Result | + Approval | |
| Coalition | 40.13 | 36.52 | Liberal | 39.20 | 35.67 | |
| ALP | 32.45 | 35.37 | ALP | 26.59 | 28.98 | |
| Others | 27.42 | 28.11 | Others | 46.82 | 35.34 |
Because this is a double disillusion election, the quota is a reduced 7.69% for the states and the usual 33% in the territories. So on that prediction the senate would get the following quota allocations:
| ACT | 2013 Result | Quotas | NSW | 2013 Result | Quotas | |
| Coalition | 34.44 | 0.94 | Coalition | 34.20 | 4.05 | |
| ALP | 33.08 | 1.08 | ALP | 31.56 | 4.47 | |
| Others | 32.48 | 0.98 | Others | 34.24 | 4.48 | |
| NT | 2013 Result | Quotas | Qld | 2013 Result | Quotas | |
| Coalition | 41.34 | 1.13 | Coalition | 41.39 | 4.90 | |
| ALP | 32.75 | 1.07 | ALP | 28.52 | 4.04 | |
| Others | 25.91 | 0.80 | Others | 30.09 | 4.06 | |
| SA | 2013 Result | Quotas | Tas | 2013 Result | Quotas | |
| Liberal | 27.45 | 3.25 | Coalition | 37.51 | 4.44 | |
| ALP | 24.88 | 3.53 | ALP | 32.83 | 4.65 | |
| Others | 47.67 | 6.23 | Others | 29.66 | 3.91 | |
| Vic | 2013 Result | Quotas | WA | 2013 Result | Quotas | |
| Coalition | 40.13 | 4.75 | Liberal | 39.20 | 4.64 | |
| ALP | 32.45 | 4.60 | ALP | 26.59 | 3.77 | |
| Others | 27.42 | 3.66 | Others | 46.82 | 4.60 |
And the senate would look like this on the primary vote:
| ACT | ALP | |||||||||||
| NSW | ALP | LIB | ||||||||||
| NT | ALP | LIB | ||||||||||
| Qld | ALP | LNQ | ||||||||||
| SA | ALP | LIB | ||||||||||
| Tas | ALP | LIB | ||||||||||
| Vic | ALP | LIB | ||||||||||
| WA | ALP | LIB | ||||||||||
Now I doubt the "others" in the ACT are organised enough for any one of them to out poll the coalition, so the last seat in the ACT will be blue, too. Last election no party got more than one candidate beyond their primary quotas in, and sometimes the major parties didn't even get that. In a double dissolution maybe you'd get two, but then the "others" might get in a quick seat too. So lets limit it as a max of one extra seat. There's no easy rule to this - in WA the ALP needed just 2% of the states preferences to gain an second seat and didn't, while in NSW the coalition needed 8.66% and got it. As a general rule though, is seems if a candidate needs more than 3pp to meet a quota they'll fail in a normal election. That's about a quarter of a quota in a normal election, or half in a DD. I'm not sure which to use. So lets hand out seats to anyone who got X.66% of a quota:
| ACT | ALP | LIB | ||||||||||
| NSW | ALP | LIB | ||||||||||
| NT | ALP | LIB | ||||||||||
| Qld | ALP | LNQ | ||||||||||
| SA | ALP | LIB | ||||||||||
| Tas | ALP | LIB | ||||||||||
| Vic | ALP | LIB | ||||||||||
| WA | ALP | LIB | ||||||||||
Now first pick for an "other" in every state except perhaps SA and QLD has to be a Greens. In SA it's be NXT who out-polled the greens (and ALP) in 2013, and in QLD it's either PUP (who also out-polled the greens), or KAP replacing them. Or former PUP Glenn Lazarus. Or Pauline Hanson. Queensland is just not used to this whole upper house shtick is it? I'm backing PUP doing badly given Palmers' press and losing Lazarus to the Glenn Lazarus Team. I suspect one nation will do poorly on preferences, with most supporters putting them in as a primary and most others ignoring them completely. On the tossup, I'll give a seat to GLT.
Even then, with the reduced quota, the Greens can probably primary vote their way into one seat in all of these, and get damn close to 2 in Vic and Tas, so we'll grant them too. Xenophon almost primaried a double quota in 2013 and has been advertising hard in SA. Plus the general dissatisfaction in the major parties should guarantee two seats. And Family First can normally count on getting Bob Day in, so I'll grant that too.
I'm prepared to leave one seat unknown in each state because even without voting tickets the last one will be a random matter of dropping out, overflow quotas and preferences shooting around like ping-pong balls. So that still leaves 2 seats in NSW, 1 in SA and 2 in WA.
WA is conservative, and the NATs run separate to the LIBS. After the two majors and the greens the NATS came fourth in 2013, but the greens still out-polled them, so I'm thinking another green and a NAT? But then maybe a centrist like NXT or PUP could do well there too. What the heck. NAT and PUP it is.
In NSW David Leyonhjelm should have some traction, so the LDP can get a seat. They did out-poll the greens in 2013 but I can't see them getting 2 seats. Both should just get their quota and have so few remainders they fizzle out. PUP was the next biggest in 2013, but other good bets would be Derryn Hinch for publicity, Shooters and Fishers who seem to pull of surprises now and then, Sex party if the overflow is vry left and Rise Up Australia if its very right. Eeny, meeny, miny... PUP.
| ACT | ALP | LIB | ||||||||||
| NSW | ALP | GRN | PUP | LDP | LIB | |||||||
| NT | ALP | LIB | ||||||||||
| Qld | ALP | GRN | GLT | LNQ | ||||||||
| SA | ALP | GRN | NXT | NXT | FFP | LIB | ||||||
| Tas | ALP | GRN | GRN | LIB | ||||||||
| Vic | ALP | GRN | GRN | LIB | ||||||||
| WA | ALP | GRN | PUP | NAT | LIB | |||||||
The last spot in each state, as I said, I'd leave as tossups for the madness that is the minor preference flows. But for a nominal indication I'll give it to a major party, since we now have exhausting tickets and public media is probably going to mean one of them gets a mark somewhere even in the shortest of lists. I'll allocate it based on which of the parties is closest to an extra quota, with some exceptions.
In SA there are two seats to fill and the major parties have their smallest representations of any state. It is true Xenophon has been undercutting them for ages, but with his two seats allocated I'll give them both a seat.
In Queensland the LNQ got it's 5th seat by rounding up its quota and the ALP only just scraped their 4th on primaries. I could see a minor party getting in here instead, with both majors exhausted. It's got to be PUP, KAP or One Nation, and since I was tossing up between KAP and GLT before, KAP can have the position nominally.
Tasmania's double Greens are ambitions, and Jacqui Lambie is not to be underestimated, even if she can be the Donald Trump of Tasmania at times. The ALP has probably maxed out at 5 and the Libs getting a 5th seems unlikely to me, so JLN can have that last seat.
| ACT | ALP | LIB | ||||||||||
| NSW | ALP | GRN | PUP | LDP | LIB | |||||||
| NT | ALP | LIB | ||||||||||
| Qld | ALP | GRN | GLT | KAP | LNQ | |||||||
| SA | ALP | GRN | NXT | NXT | FFP | LIB | ||||||
| Tas | ALP | GRN | GRN | JLN | LIB | |||||||
| Vic | ALP | GRN | GRN | LIB | ||||||||
| WA | ALP | GRN | PUP | NAT | LIB | |||||||
That comes to:
Coalition (LIB, NAT & NLQ): 30
ALP: 29
Cross-bench: 17
For comparison, after the 2013 election it was:
Coalition: 33
ALP: 25
Cross-bench: 18
So, that's my highly dubious guess, with no real faith in the reliability of my tossup fill-ins.
Subscribe to:
Posts (Atom)