There’s been a lot of attempts by various people to predict the impact of Senate reform in terms of which parties will win and which parties will lose.
Some of this has been smart, and some of it has been quite dumb. Some has focused on the long-term tendencies, while others have been obsessively focused on the outcome under one narrow scenario where the Coalition wins a double dissolution comfortably, or freak out about outcomes that are just as likely under the current system.
I’ve generally tried to avoid getting too much into these questions. In general, I expect that party behaviour and voter behaviour will change in ways which are difficult to predict, and overconfidence in any particular outcome suggests that a person has an agenda to push (usually an anti-reform agenda).
(If you really want to have a better sense of likely outcomes, I can point you to Kevin Bonham, Antony Green and Adrian Beaumount‘s analysis. In short, both major parties and the Greens would have gained seats, in proportion to their vote, but the chances of a Coalition majority are not much greater than the current system.)
I wanted to specifically address a question raised in a bunch of places: does the left “splitting” its vote between Labor and the Greens disadvantage them relative to the Coalition, who run as a single ticket.
(Here’s an example of a claim about preference exhaustion hurting Labor and the Greens, from the newly-renamed Science Party, which makes some ridiculous assumptions that there will be no more minor parties in the post-GVT future).
It is true that, if there are two candidates racing to win a final seat, the candidate in first place is given an advantage if the rate of exhausting is higher. This is true in all preference elections.
It’s often easy to predict how this effect would work in an election for single-member races. In recent NSW and Queensland state elections, the Coalition has usually had an advantage due to the left vote being split between Labor and the Greens, although strong votes for Katter’s Australian Party in 2012 and One Nation in 1998 and 2001 had reversed this effect.
But things get much more complicated in a multi-member race, such as the NSW Legislative Council, the ACT Legislative Assembly or the Tasmanian House of Assembly, all houses with multi-member electorates with versions of optional preferential voting.
Of course, it’s easy to imagine scenarios where the Coalition could get elected, although in practice there have been no cases where Labor and the Greens both missed out due to a split in the vote when a higher preference flow would have given one of them the seat.
On the other hand an efficient splitting of the vote can mean that both candidates remain in the count longer than if the vote was combined on a single ticket, and exhausted votes can lower the quota and elect both of them. If, for example, the third Labor candidate and the Greens candidate are each on 0.7 quotas, that would likely be enough for both to win if exhausted votes lower the effective winning number, but if they had been on a single ticket both would have missed out.
Kevin Bonham has previously analysed this phenomenon. He’s found no cases where Labor and the Greens would have missed out thanks to a split vote – Labor and the Greens are usually competing with each other for the third left seat, not also competing with the Coalition.
The first example of this doesn’t come from OPV – this phenomenon can be found in the 2004 Queensland Senate election. The Liberal and National parties ran on separate tickets, and polled 2.68 and 0.46 quotas.
If these two parties had run as a single group, they would have easily cleared three quotas, but had only 0.14 left over for their fourth candidate, who would have been quickly knocked out, not hanging around long enough to benefit from preferences. Instead, both candidates stayed in the race long enough to benefit from other conservative preferences, and both were elected, giving the Coalition a majority. Under optional preferential voting, both would have likely been elected regardless of preference flows.
A similar phenomenon can be seen in ACT and Tasmanian elections, where each candidate runs their own campaign, and a party’s vote can be distributed across multiple candidates. It’s not uncommon to see a case like that in Ginninderra in 2012, when the Greens vote added up to 0.6 quotas and Labor added up to 2.41. If those votes were in the form of a ticket, the Greens would have won. But instead the Labor votes were evenly distributed between three candidates, and thus all three won. It’s imaginable that they may have been harmed by preference leakage, but the benefits of having multiple candidates stay in the race outweighed this.
Preference exhaustion could be a problem, but it’s particularly likely to be a problem when it comes to relying on minor party preferences, and this is more of a problem for the Coalition than Labor.
In 2013, the vote can be divided up along the following lines:
- 37.71% – Coalition
- 30.11% – Labor
- 8.65% – Greens
- 6.72% – Minor right
- 4.9% – Minor left
- 4.2% – LDP
- 7.5% – Other (including PUP and Xenophon)
The minor right, plus the LDP, would both favour the Coalition with preferences, and they add up to over 10% of the vote compared to the remaining minor left with less than 5%. While the Greens have a strong capacity to hand out how-to-vote cards (and Greens voters are usually the most likely to mark preferences), those right-wing groups will have far more trouble.
In summary, it is false to say that the side of politics which is more unified should benefit compared to the side of politics which is more divided, particularly if that divided side has multiple viable options. And when you are worrying about the effects of preference exhaustion, don’t just think about the Greens but think about all the other small parties.