Saturday, May 9, 2015

Do Proposed Senate Reforms Advantage The Coalition?

(See also Would Senate Reforms Increase The Chance Of A Blocked Senate?)

Advance Summary

1. Concerns have recently been reported that the Joint Standing Committee on Electoral Matters' proposed optional-preferencing Senate system may advantage parties other than Labor, especially the Coalition parties.

2. The reality is that Labor performs poorly under the current system and sometimes loses seats it deserves to win under it.

3. There is no historical, and no convincing theoretical, evidence that the Coalition loses more seats to micro-parties under the current system than Labor.

4. If anything there is some argument that the proposed changes improve the chances of Labor and the Greens acquiring at least a blocking majority in the Senate.

5. That argument, however, assumes that parties would attract the same vote shares under the new system, when the choice of that new system would actually discourage the scattering of much of the right-wing vote among a huge number of micro-parties.

6. All up there is no evidence that the proposed reforms disadvantage anyone, other than removing chances to be elected from micro-parties that don't deserve those chances anyway.


Since the release of the final Joint Standing Committee on Electoral Matters (JSCEM) report, there's been an upsurge of discussion of proposed Senate reforms.  By way of background for those who came in late, the 2013 Australian Senate election produced several absurd and undemocratic results.  These are dealt with in many previous pieces about Senate reform, which can be accessed via the Senate reform tab.  

The proposed JSCEM model would abolish above-the-line group preference ticketing, so that preferences would only flow between party groups if the voter caused them to do so.  It would also make it vastly easier for voters to distribute preferences for specific candidates below the line.  However, as a result of voters being accustomed to just vote 1 for parties above the line, in its early days at least it would be likely to have high exhaust rates in preference transfers, as in the NSW Upper House system.

The initial JSCEM recommendation had support from Coalition, Labor and Green representatives on JSCEM and no apparent opposition from anyone within these parties.  However, cracks in potential Labor and Green support for reform are starting to be suggested.  Heath Aston reports that Penny Wong, Stephen Conroy and Sam Dastyari are "understood" to oppose the reforms, and that union figures too believe the proposal favours everyone but the ALP.

Aston also says that some within the Greens ("an emerging split" - hmmm) are getting wary of the changes, feeling that Greens might lose in some of their weaker states, or just be skinned by some of their supporters, who might be pathologically averse to dealing with the Coalition on anything, especially if their vote is crucial.  A compromise of a 4% primary vote threshhold and retention of Group Ticket voting is hinted at.  However Senator Lee Rhiannon has suggested that the unnamed Greens source quoted is "wildly disconnected from Greens members and MPs" and implied they may be up to no good.  Rhiannon, who is on JSCEM, has stated clearly that the Greens do not support the proposed "compromise".

These reports of internal mutterings are unconfirmed by the forces involved, but they are also unsurprising if true, especially following lukewarm polling on the issue, and they have not been denied by Labor.  It may well be that some on the left have concerns about the precise model involved, but if so it is incumbent on them to propose a better model than JSCEM's.

It's not good enough for any party  to vote against JSCEM's model without putting up one of their own that is reasonably fair.  It's also not good enough for the parties to get into a standoff in which each pushes a model that is not supported by the others, and therefore nothing actually gets done.

The minimum standard for acceptable reform prior to the next election is that it be no longer possible to win from a very small percentage of the primary vote unless that is the will of voters who distribute their own preferences in that way.

If any of the big three parties ends up not supporting reform to this extent, or obstructing it by picking a model not clearly superior to JSCEM's and refusing to accept other viable options, then that party will not merely not deserve to govern.  It will not even deserve to exist.

Which parties have been hurt by the current system?

The belief that the changes would advantage the Coalition seems to stem from the 2013 results.  In a previous article Optional Senate Preferencing: Not An ALP/Liberal/Green Stitch-Up I looked at the exact impact of the changes on the last Senate poll.  Various defects of the current system were responsible for:

* A seat that should have gone to the Coalition in Victoria was won by Ricky Muir (AMEP).
* A seat that probably should have gone to the Coalition in Tasmania was won by Jacqui Lambie (then-PUP) having almost been won by the Sex Party.
* A seat that should have gone to the Xenophon Group and a seat that probably should have gone to Labor in SA were won by the Greens and Family First.
* A seat that should have gone to the Greens in NSW was won by the Liberal Democrats, as a result of party-name confusion to which the large number of parties (partly a product of preference-gaming) contributed.
* A seat that should have gone to Labor in WA went initially to the Sports Party, and then following a by-election caused by the loss of ballot papers and the closeness of the result, to PUP.

So the net result based on the 2013 vote totals is that two seats that probably should have gone to the Coalition, two to Labor, one to the Greens and one to NXG went instead to the Greens, AMEP, PUP (twice), FF and LDP.

We should also consider earlier elections (at least since the collapse of the Democrats vote):

* A seat in Victoria 2010 that should have gone to Labor went to the DLP.
* In SA 2010 Labor were 1% ahead of the Coalition but lost on above-the-line Family First preferences (This might have happened under optional preferencing too; it isn't clear.)
* In SA 2007 the Greens won a seat that possibly should have gone to either Labor or the Coalition (again unclear)
* In Vic 2004 Family First won a seat that should have gone to the Greens
* In Qld 2004 a seat that should have gone to the Greens was won by the Nationals

So all up there is no convincing evidence that this system hurts the Coalition significantly more than it hurts Labor.  If anything, it hurts Labor more.

Is this to be expected?

On the whole, yes.  There is a substantial share of the vote that goes to micro-parties and is likely to continue to do so even without group ticket preferencing.  The micro-parties mostly trade preferences with each other, though some do not, and if there is a perfect preference-harvester (like Muir's in Victoria) going, then it receives votes from left and right micros alike.

The party that suffers the most from this is, then, the party that would have been in line for the next seat based on primaries.  But there is no systematic tendency for that to be either major party.  It's just whoever happens to have the next largest remainder that would have been likely to fill the last seat but for being overtaken by a group-ticket harvester.

It might seem that the Coalition should be likely to have the next largest remainder, if the obvious seats split 2 Coalition 2 Labor 1 Green.  But quite often this doesn't happen.  In SA the NXG vote was so huge that the obvious seats only split 1-1-1, and the Coalition's remainder was the one that survived while Labor lost.  In the original WA poll Labor were beaten by such a margin that the Liberals and Nationals combined had three quotas, making Labor's portion of a second quota the vulnerable remainder.

Furthermore, if Labor does well in an election, then it will probably beat the Coalition on primaries in at least Victoria and Tasmania.  So what tendency exists for the Coalition to seem more vulnerable is largely down to the size of their win in 2013.

And even if it turns out that the Coalition in the long term was pipped by micros 10-15% more often than Labor, so what?  With each side being pipped by micros on a routine basis, the long-term effect of letting the current system continue is that the micros may end up getting the balance of power permanently.  As I noted last week, it may be fun watching the Abbott government trying to herd cats right now, but let's see how Labor might go if, after the next election it has to get its bills approved by Madigan, Lambie, Wang, Day, Leyonhjelm and who-knows-who else as well as the Greens.

Locking in Green seats?

The concern attributed to unions that the reforms would entrench the Greens is hard to understand.  Labor is a traditionally poor performer at preference harvesting because it sits in the middle of the spectrum and hence tends not to receive preferences from ideologically-driven parties once those preferences leave the micro-party maze.  Labor is also often in a position in which it will be fighting for the last seat and hence not excluded, and as a result it is often not attractive for other parties to trade with.  In the last four elections, there has not been one single case of Labor winning an arguably undeserved seat.

Perhaps the Coalition Has More To Fear?

I am not aware of anyone supposed to be dissenting within the Coalition, but the argument was put by Terry McCrann, who argued last year that Senate reform would "lock in an all-but-certain permanent Labor-Greens-Palmer Senate majority".  McCrann didn't feel obliged to provide evidence for his claims.  As it happens, they're exaggerated, but not entirely baseless - if a very short-sighted view of electoral behaviour is taken.

Following the 2010 and 2013 Senate elections, the Coalition has 33 Senators, and Labor and the Greens combined 35.  I've argued above that preference-harvesting (or its flow-on effects such as the LDP confusion) fleeced the Coalition of two seats in 2013 and fleeced Labor and the Greens combined of at least three in 2010 and 2013 combined (not counting any taken from each other).  So without it, based on the votes cast, Labor and the Greens would probably have a "blocking majority" (an alliance with 38/76 Senate seats can prevent all Government legislation being passed, until the government is willing to force a double-dissolution).  It's even possible they would have had a combined outright majority, though with the Coalition governing in the lower house, the difference is minor.

It may seem strange that Labor and the Greens would be half the Senate under OPV following elections at which the 2PP was more or less tied once and a clear Coalition win the other time.  But firstly that's the nature of proportional representation: the Greens often aren't eliminated from counts so none of their preferences would leak en route to Labor.  Secondly, voters vote differently in the House and in the Senate.  On average at the last two elections, the Coalition's primary vote was about six points lower in the Senate than the House, Labor's about three points lower, and the Greens' about the same.

This, though, is a new thing - at the 2004 and 2007 elections, the Senate handicaps for the major parties compared to the House votes were about the same as each other.

It would seem that if anything optional preferencing could give a small advantage to the left based on current voting patterns, because the non-Labor left vote resides mostly with the Greens, who would win seats, while the non-Coalition right vote is scattered between a gaggle of very small religious, hunting/fishing, nationalist/xenophobic and right-libertarian parties whose primary votes are too low to win even under compulsory preferencing and who would just tend to get cut out of the race without much benefit to the Liberals and Nationals.  However ...

System Choice Determines Party Behaviour

This headline is what almost every analysis of the consequences of bringing in optional preferencing is missing.  It doesn't make a lot of sense to say that the right generally would be disadvantaged by the use of optional-preferencing because of the severe fragmentation of the right-micro vote when that fragmentation is itself a product of the current system and would not exist to anything like the same extent under the new one.

Australia simply doesn't need to have three different outdoor-recreation parties and a motorists' party, nor does it need to have both Christian Democrats and Family First, nor does it need to have the Liberal Democrats and any number of little front parties aligned to them, nor does it need to have so many One Nation clone parties as well as what remains of the original.  It also, for that matter, does not need to have two drug reform parties, a voluntary euthanasia party, a Secular Party and a Sex Party, since left-libertarians not attracted to the Greens will generally support all these things.  Such specialised parties exist in ridiculous numbers because the group ticket voting system not only permits them to do so with no effective deterrent bar losing their deposits, but encourages them to do so in order to harvest preferences.  Under optional preferencing if two very similar parties run against each other they destroy both their chances.  It's inevitable that under the JSCEM proposal, most of these micro-parties would merge, run joint tickets, or disappear, driving some votes back to the Coalition.

The Coalition would also have some new strategic options under the system in some states.  For instance, suppose in a state the Coalition on a bad day polled 2.5 quotas, Labor 2.6, the Greens 0.9 and the remaining 1.0 went to various micros.  Under the current system this could well result in a 2-2-1-1 split.  Under optional preferencing, it might look like 2-3-1 (the result that creates a danger of the left parties controlling the Senate), but if the Coalition vote was actually separate Liberal and National tickets polling, say, 1.7 and 0.8 quotas, then it would be 3-2-1.

How About The Greens?

The suggestion is that some unnamed Greens supporters may be nervous about their chances in the states where their votes are historically poor, notably Queensland, SA and WA.  However, since the Democrats ceased being competitive, the only state where the Greens have won seats based on preference-harvesting that they would not have won anyway was South Australia.  The most recent case of this happening was 2013, and that was a rather unusual one because the Greens were able to win on a preference deal with PUP, an ostensibly centrist populist party that has since lost virtually all its support.

Overall the JSCEM proposal wouldn't have much impact on the Greens.  They'd be safer against major-party gang-ups in cases where they polled nearly a quota, but less likely to win by preference harvesting when they polled much less - although their chances in those cases are not likely to be good anyway.  They'd also be safer against repeats of the 2004 Family First win (caused by Labor's willingness to deal with right-wing micros as well as left-wing ones).  Another advantage for the Greens is that because their voters are good at giving preferences when preferencing is optional, they would have more impact on other contests.  We saw this in the NSW upper house election recently.  Perhaps the biggest advantage for the party would be no longer having to trash its brand by dealing with parties that its supporters oppose.

And Xenophon?

There is no doubt that Nick Xenophon was the big loser out of the current system at the last Senate election when his Group polled 1.74 quotas but won only one seat, while the Greens and Family First with less than half Xenophon's primary vote between them won two.  On the 2013 vote shares he would be advantaged by winning a second seat under optional preferencing, and with most of the other micros eliminated his two-seat bloc would become very powerful.  But this is not about "entrenching" him as he is clearly quite entrenched enough already - it is only about whether certain vote shares should get him one or two seats.  Whatever one thinks of the merit of Nick Xenophon dragging a running mate in with him (and after the Ann Bressington nightmare in the SA parliament it's no wonder people are cautious) it's indisputable that the Group deserved two seats and only missed a second because nobody much thought they were worth dealing with.

The argument about Xenophon is only being had because his vote share fell in a very narrow range.  Somewhat lower and he wouldn't have won a second seat under any reasonable system, somewhat higher and he would have won two seats under them all.  He's not likely to poll exactly the same vote continually.

That a Senator or party from a small state can end up having a lot of power is a natural hazard of a state-based system.  We saw the same under the existing system with Harradine.  But it is not something worth risking a permanently chaotic Senate to avoid, especially not when power is then distributed among a bunch of random Senators who may not even have substantial support in their own States.  Ultimately if one small-state Senator does get too powerful then other states can start electing their own state-based Independents - if they really care that much.


The reported Labor concerns about the proposed JSCEM system are misplaced.  If anything, the proposed system frees Labor from an arrangement in which it has performed poorly and lost some seats it deserved to win.  The purpose of the proposed system is to disadvantage micro-parties that do not deserve to be elected anyway, and the system does exactly that.  Evidence for advantage or disadvantage to either "side" is inconclusive at best, and ignores the fact that the new system advantages both "sides" against the micros.

Update 28 May: 

An article by Nikki Savva in the Australian (Conroy and Wong Lead The Charge As Labor Goes To War With Itself - Google title if paywalled) quotes Senator Dastyari as saying:

"Frankly I can't find a single Labor Senator that supports any of this [..] It would be complete madness for Labor to support any proposal that would risk forever preventing a progressive Senate.  I can't see Labor doing this".


"If the Greens want to sign a suicide note and do a deal with the Liberals, good luck to them" [..] "Let's see them explain to their supporters why they are prepared to give up control of the Senate for their own jobs".

Ah yes, that would be the control of the Senate that the Greens do not actually have because they lost it to a bunch of random micros at the last election, meaning that they are only now relevant when they support a Coalition bill, Labor doesn't and not enough crossbenchers do either (this very very rarely happens!)

If there is anything in Labor's comments other than sheer paranoia or opportunism for the sake of it, then those Labor Senators opposing this reform owe it to the public to explain their reasoning with a detailed public paper explaining what they are worried about.  All the rest of us put our ideas out to JSCEM and made them publicly available to be picked over by the committee and by the authors of other submissions.

If Labor opponents of reform cannot at least do the same thing - explain their reasoning properly and in detail to the public, then it does seem they have completely lost the plot - and are basing their "strategy" on the assumption they will lose the next election.

Or it may be there's something deeper at play.  While the argument has not been broached, some Labor figures may be running scared about the ease of cross-party below-the-line voting under the new system.  There may be a worry that the party's ability to deliver seats to, for an obvious instance, the Shoppies, might be at risk if progressive ALP voters start rebelling against the Bullocks of this world.

Who knows what these Senators are actually thinking?  Let's see them go beyond soundbites and tell us!


  1. For what it's worth, Lee Rhiannon has denied the SMH report here:

    1. It is worth quite a bit and I have edited the article accordingly in a few places. Thankyou for the link.

  2. You seem to have chopped off the end of your eighth paragraph. Could you add that text back in? I'd like to read all your righteous fury against any big party leaving the current system in place.

    1. That's really odd - I can only guess that I had the draft open in two windows and somehow saved a pre-final version over what I thought was the final version. (I do this sort of thing sometimes, losing a 5000 word essay and 19 pages of my PhD to two instances of this stupidity in my long-ago student days).

      I don't still have or remember the exact wording but a near-final draft of this article contained a threat that if any of the big three got seriously in the way of Senate reform I would carry a graphic on my sidebar that encouraged readers to destroy that party, and leave it there until such time (however many decades later) as that party was deregistered by the AEC for having less than 500 members. I decided to leave that bit out until such time as it became clear enough that one of the parties was actually serious about opposing reform, before getting quite that nasty!

  3. Micro parties should have 750 signed up member before allowed to field candidates