Senate Reform Has Been Passed

It's all over bar the High Court challenge(s).  On Friday, the Commonwealth Electoral Amendment Bill 2016 finally passed both houses of parliament and was quickly whisked off to receive royal assent.

I've never been a fan of the now presumed-dead group ticket system, but my own campaigning to change it really dates from 30 August 2013.  On that day I discovered that the Family First candidate who had a dream run on the Tasmanian preference allocations, including from Labor, the Greens and various left micros, was an anti-gay extremist.  There was a high risk that socially progressive voters innocently voting above the line for their preferred parties would elect someone whose views would horrify them.  So I wrote this article.  In the end the disaster was avoided by just 821 votes.

I found it was uphill work convincing people to vote from 1 to 54 below the line (let alone the number of squares required in the larger states!) and that party operatives were concerned that I might cause their voters to vote informally.  Around the same time the Truth Seeker site was predicting Senate chaos with numerous micro-party wins based on tight preference flows between micro-parties.

What was most concerning, when the votes were finally counted, was not just that Truth Seeker had been right in a big picture sense, but also that the micro-parties that actually won or nearly won were frequently those that were not even on the radar of those trying to model the election in advance.  Trying to alert people to the consequences of their preference flows and do background checking on which micro-party Senators were going to win was a nightmare when not even expert modellers and not even preference whisperers could determine which micro-parties were going to get up.  Other results of the 2013 election showed that the system was not only absurd on a massive scale, but also a sovereign risk.  The voiding of the WA election would not have happened under the same circumstances but with voter-directed preferences.



That's a brief history of how I became involved in spending the last few years arguing that the Senate system was broken and had to be changed, urgently, to any bearable alternative.  Others - especially Antony Green - have been doing this thing for a lot, lot longer than me.  My thanks to all, for whatever reasons, who have fought the good fight on this issue.

What Has Been Passed

The Act as passed is virtually identical to the final recommendation of JSCEM (see previous article.)  The key features are:

* Group ticket voting is abolished and preferences will only flow to the extent that the voter directs them.
* The ballot paper will direct voters to number at least six parties above the line or at least twelve candidates below.  The voter can preference as many parties or candidates as they wish.
* Votes above the line that do not follow the instructions will be saved and remain formal provided that the number 1 is the lowest number and occurs once and once only.
* Votes below the line that do not follow the instructions will be saved and remain formal provided that the numbers 1 to 6 each occur once and once only.
* The current situation for votes numbered both above and below the line continues: if the below the line vote is formal then it takes priority, but if only the above the line vote is formal then the above the line vote counts.
* Party logos will be allowed on ballot papers.
* There is no explicit provision against encouraging voters to cast a single 1 vote above the line.

Concerning the last I had some discussions with Senator Xenophon about a possible amendment to ban anyone from "advocating" that voters vote formally but contrary to the instructions, but he decided not to proceed with it.  Bob Day and David Leyonhjelm had a go instead, without success.  The question of whether it was legal to put out a "just vote 1" style card was raised many times during the debate in the Senate and I have very little doubt that it will be legal to issue such a card.  Just because such a card will encourage voters to do otherwise than instructed does not mean it is actually misleading, since their vote will still be formal.  However, as Senator Xenophon implies, for most parties there is no obvious strategic advantage in doing so.  Among other things, a party that causes its own preferences to exhaust loses the ability to swap how-to-vote card preference recommendations with another.

Issues

The new system is an enormous improvement on a terrible system that needed to be killed off by any method available (indeed, if a system that exalted backroom deals was killed by them as has been alleged, then I can think of no finer ending).  But we should not imagine everything will run smoothly.

Organising the count under the new system at relatively short notice will be a challenge for an AEC still reeling from the 2013 WA disaster and bringing in new ballot security measures at the same time.  Expect the Senate count especially to be unusually slow and expect some mistakes to happen. What is important to note though is that the risk of an election under this system being wrecked by mistakes (a la WA 2013) is far lower.

The change in the savings provisions for below the line votes may lead to a small number of BTL voters having their votes invalidated by making mistakes in the first six squares.  It's surprising but people do make errors even in numbering from 1 to 6 once each, possibly because they try to vote from both ends and wind up with two 5s or similar.  The savings provisions can be reconsidered for future elections if this is a problem, but if voting BTL be extra careful to include the numbers 1 to 6 once and once each only.

Just-vote-1 cards might be abused to disrupt preference flows (eg by running a just-vote-1 campaign for a party with a seemingly similar appeal to another party) but it isn't a major issue and something where election-watchers will just have to keep an eye on possible abuse.  If anything, the short lead-time to the current election means it's not likely to be a problem right away, because of insufficient time to register front parties designed to game the system.  However I suspect some fine-tuning to legislation of this aspect will be needed down the track.

Some voters are responding to the optional nature of preferencing with relief because they will no longer have to send preferences to any parties they dislike or are unenthusiastic about.  Such voters need to realise that preferencing a party you mildly dislike over one you despise never helps the former to beat any party you have preferenced higher.  A voter might still be strategically refusing to preference in the hope of encouraging the major parties to fight harder for their preference, but they should be under no illusions that in doing so they may help parties they really dislike to win.  Now that preferencing all parties above the line is reasonably easy, voters should encourage each other to go beyond the required minimum number.

Predicting the election based on primary votes on the night may be very challenging.  It will be difficult to work out things like the exhaust rate without access to comprehensive, high quality, organised scrutineering.  It should be obvious quickly which candidates are in serious contention, but it may take longer to be confident about close results.

While Labor has made some complaints about the short lead-in time and quick nature of consultation about exact details of the final model, if they had been in the same position I think they would have done much the same thing.  The further out from the election reform was introduced, the longer the government would have had to deal with a completely uncooperative crossbench.  That said, the slow pace of action on the issue through the government's term did not help.

What Will Happen If There Is A Double Dissolution?

There has been a lot of speculation about likely results in a double dissolution.  The argument run by those fearful of the change (and also many of those not fearful of it) has been that a DD will more or less eliminate micro-parties.  I think this conclusion is unsound.  It will just eliminate those that don't deserve election based on the votes actually cast.

The standard argument goes that only Nick Xenophon is safe.  Day, Muir and Madigan did not poll enough votes last time; Lambie, Lazarus and Wang were elected under PUP but PUP's support has imploded, and Leyonhjelm won as a result of a combination of ballot paper luck and oversized ballots causing confusion.  On this basis seven of the eight are considered likely to lose.

However, this argument is cherry-picking.  It considers a reason why the votes for the PUP Senators could collapse, without considering any possible reasons why votes for Day, Muir and Madigan might rise.  Muir especially has been a surprise packet and may prove to be a cult figure whose vote increases to competitive levels - as, a long time ago, did Nick Xenophon's in the SA upper house.  There's no polling at all, so we don't know.  Furthermore, while PUP's vote has indeed collapsed, we don't know if that transfers to Lazarus and Lambie, or if the electorate perceives them as having been wise to get out - again, no polling.

It is also possible, though not necessarily likely, that new crossbenchers aside from extra NXT Senators could enter.  These might be new independents attracted by the lower quota, or maybe KAP could win a seat in Queensland.  There has been some speculation about Clive Palmer trying to switch to the Queensland Senate, though I suspect he would actually fail.

In my own simulations I found that had the 2013 election been conducted under the new system, 12 non-Green crossbenchers could have won (a more publicised Parliamentary Library piece gets 10 or 11).  With the implosion of PUP - which would have won five seats at a 2013 DD - and the introduction of logos, I don't think we'll be seeing anything like that in 2016.  Nonetheless I think there's about a three in four chance that at least one non-NXT crossbencher will win somewhere.  Of the seven endangered crossbenchers I think Lambie, Day and Muir have realistic chances (Muir might really bolt in or bomb, who knows) and I have no idea about Lazarus.  (Any comments re his hopes from Queensland readers welcome.)  I'm expecting the others to struggle.

It's highly unlikely the Coalition will win a majority or even a blocking majority unless their Lower House result is at least as good as last time, or perhaps even then.  A seven-seat block in any state doesn't look realistic and five seems to be the maximum in SA.  Six seats is also difficult in Tasmania (this would probably require beating Jacqui Lambie, but Tasmania has quite a history of re-electing independents).  It's realistic that if the Coalition wins the Lower House easily, it could get 35-36 Senate seats, but more than that is challenging.

The big winner should be Xenophon.  On 2013 support levels he would be likely to win three seats in SA alone, and there is some possibility of the odd win in other states.  If the Coalition wins the Lower House well then it is distinctly possible Xenophon's team could share the balance of power with the Greens, so that bills pass with the support of one or the other.  Or it is possible that the Coalition could pass legislation with support from Xenophon and one or two others.

On the other hand, supposing Labor wins the election, there would be a good chance of obtaining at least a six-seat Labor/Green block in every state except South Australia.  There would be a distinct possibility of seven ALP/Green seats in this case in one or both of Victoria and Tasmania, and it would be highly likely Labor could govern with the Greens' support (assuming they ever talk to each other again) or at worst perhaps needing the Greens plus NXT.

It will be very interesting to see how much use is made of below-the-line voting.  In theory, the new BTL rules could allow for voters to overturn contentious party preselection decisions.  This is especially so in Tasmania with its history of Hare-Clark preferencing.  If displeasure about the dumping of Senator Lisa Singh to an "unwinnable" position is widespread enough, it's possible an organised and large campaign to convince Tasmanian voters to use BTL voting to save Singh could catch on and deliver enough votes to her to get her above higher ALP candidates on the list in a double dissolution.  It wouldn't be easy, and would depend heavily on the luck of what fractions of quotas different parties ended up with.

Labor's Intellectual Bankruptcy On Senate Reform

I didn't have time during the debate to address in detail all of the spurious arguments raised against reform, few if any of them new, that were raised by opponents.  A number were embodied by Labor's Anthony Albanese during his appearance on Lateline on Friday night.

Albanese somehow managed to at the same time argue that the new system would "raise the drawbridge" to keep out micro-parties while proposing an alternative (threshholds) under which micro-parties not receiving a certain vote would be bulk-excluded and their preferences transferred.  In fact, the alternative he proposed is even harsher on micro-parties than what was actually passed, since they would be excluded from the count no matter what the preference flows to them were like.  It does depend on the level of the threshhold, but in my modelling I've found that micros with 3-4% primary vote support can sometimes legitimately win in double dissolutions under the new system, depending on the position of other parties and the strength of preference flows.  The most commonly suggested threshhold, 4% (based on the figure for getting deposits back, which Michael Maley tells me was a sop to the Australian Democrats), would disqualify possible winners.  A lower threshhold would make it too easy for a slightly winnowed cartel of micro-parties to continue with the same old game.

In any case, if threshholds are the best Labor can come up with, they're not trying.  Keeping the old system but applying a primary vote threshhold is an idea that received short shrift during the initial JSCEM process, because it's undemocratic, arbitrary, probably unconstitutional, and doesn't solve quite a lot of the problems that happened in 2013.  See Michael Maley's JSCEM submission for abundant evidence on this point, and my initial comments, and Ben Raue.

Furthermore, that the new system doesn't really "raise the drawbridge" to exclude everyone but the Coalition, Labor, Greens and NXT can be seen from the history of the system that existed from 1949 to 1983.  That system required full preference flows directed by the voter, and was eventually abandoned because of scandalous informal-voting rates that disadvantaged Labor.  However, it was similar to the new system in that preference flows were natural rather than being allocated by group voting tickets.  Parties that polled below half a quota never actually won, but this did not prevent new parties emerging.  Three parties (DLP, Liberal Movement, Democrats) emerged and won Senate seats under this system, as did four Independents.  My simulations have also shown (at least) that the Greens, PUP, One Nation and NXT would all have emerged under the new system.

Modelling by Nick Casmirri, who has been tweeting results rapidly as I write, has suggested that had a 4% threshhold existed at the 2010 and 2013 elections, the net results would have been 35 Coalition, 27 Labor, 8 Green, 3 PUP, 2 Xenophon and Leyonhjelm.  My modelling of the new system projects 34 Coalition, 28 Labor, 10 Green, 1 PUP, 2 Xenophon and Leyonhjelm for 2010-3 based on raw vote totals, though if the use of logos is factored in Leyonhjelm's seat could go to the Coalition.  So the net effect of the threshhold model could be to weaken the left by three seats compared to what was actually passed, and make no real difference to the range of parties elected.  It would give PUP a more proportional result than the new system, though that has proved irrelevant given the disintegration of the party anyway.  But hey, at least it would hurt the Greens, right?

Albanese also tried to run the argument that micro-parties would be disadvantaged because votes would exhaust, thus "disenfranchising" their voters.  This is a terrible argument that has been disposed of here before.  The real reason micro-parties polling very low vote shares will lose is that voter-determined preference flows - whether voters are allowed to exhaust their votes or not - are not strong enough for parties starting with tiny vote shares to overtake those with several times as many votes.  Only the fake near-100% preference flows created by group ticket voting make it possible for parties to win off tiny vote shares.  Again, under the 1949-1983 system minor parties never won off less than half a quota, though that system had compulsory preferencing.

Another red herring in the interview was the matter of Michaelia Cash's low personal primary vote, a line of attack that had earlier found its way into Stephen Conroy's speech explicitly via a comment in a spuriously-titled Van Badham article - that comment being derived from lame Twitter memes started by crossbenchers.  Much has been made of the fact that some Coalition Senators were elected with below-the-line votes much lower than the BTL votes for certain crossbenchers.  However, a candidate's BTL vote indicates nothing reliable about their personal support levels, for the following reasons:

1. The natural levels of BTL voting vary greatly between parties.  They are typically higher for micro and minor party voters than for the major parties.  This is partly because micro parties attract more voters who are distrustful of parties generally, but also because micro-party voters are often concerned about their own party's preference dealings.

2. Many voters who vote BTL are doing so not because they want to vote for a particular candidate in a party but because they want to order their preferences between parties (as they will be able to do above the line under the new system).  Such voters will frequently vote down the ticket of their most preferred party and therefore all comparisons between lead candidates for one party and non-lead candidates for another are just stupid.

3. Major party voters, and in particular Coalition voters, are often well aware that their vote can help their party win a seat beyond the first.  The concept of a given ticket position being "safe", "winnable" or "unwinnable" is fairly well known.  Given that a voter who might prefer Cash over another Liberal has to vote below the line to demonstrate that preference (whereas by just voting 1 they can ensure their vote passes to her after the first Liberal is elected) we cannot conclude anything about what Coalition voters think of Cash based on the small number who choose to number every box.

The whole enterprise of trying to derive conclusions about personal support from numbers of BTL votes is in most cases pseudoscience.  It is only when a supporting-position Senator tries to run for a Lower House or state seat that we get a real idea of their vote-pulling power.  See also Antony Green.

Among the many poor Labor arguments in the marathon Senate debate I will single out one by Senator Wong.  In also arguing against exhaust, Wong said " [..]I do not think it is a democratic thing to have a voting system that risks between 14 and 20 per cent of votes exiting the system and not staying in play."  The 14-20% claim derives from a comment by Nick Economou .  I had a go at replicating Economou's claim, looking firstly at the percentage of vote values that would be even capable of exhaust based on the 2013 primary votes under the new system. (Vote values that can't exhaust are primary votes that are retained by either elected candidates or the last one to miss out.) I found that percentage to be, oddly enough, 14-20% varying by state (20% for NSW, Vic, WA; 18% SA; 16% Tas; 14% Qld - for completeness, exhaust rates would be tiny in both territories).  However, for 14-20% of vote values to exhaust, it would have to be the case that every vote that was transferred exhausted - which would be what would happen (for instance) if every single voter just voted 1.  Whatever the percentage of voters for micro-parties just voting 1, it will hardly be 100%, and indeed not even all of them voted just 1 under the existing system! So the idea that the system really risks such high exhaust rates at a normal election isn't right.  For a double dissolution, even more of the vote is tied up in quotas for winning candidates and the exhaust rate will be even lower.

But whether or not Wong is right, what is certain is that her party is hypocritical.  It was the ALP which not so long ago brought optional preferential voting into New South Wales and Queensland, in both cases to attempt to hobble the Coalition. With the rise of the Greens as a preference source, Labor ended up hobbling only itself.  In an amusing sign of how quickly one can take part in political debate these days, I sent out the following tweet, which I believe to be my most retweeted thus far:

Within about an hour, Mathias Cormann had read it into Hansard!

With such a poor display by his team including his former leadership rival on this issue, Bill Shorten is on laughably weak ground complaining about the quality of political debate.  At some point, Labor has collectively decided to oppose Senate reform for completely political purposes - most likely with the aim of picking up micro-party preferences (ignoring that micro-parties are not very effective at directing them) or capitalising on very likely implementation problems.  (Edit: Or damage the Greens by portraying them as just another party or naive - see comments.) However in so doing it has had to argue a case that its own keepers of electoral wisdom, like Gary Gray, Alan Griffin and John Faulkner, have been telling it loud and clear is utter nonsense.  It has persisted anyway and done untold damage to public understanding in the process.  We'll be cleaning this up for a long time, but it is better than still having to clean up after a terrible electoral system.

Next: The High Court Challenge

I am not going to cover the High Court challenge in much detail for now.  At this stage, two possible issues look to be the constitutionality of exhaust rates and the constitutionality of above-the-line voting at all.  Given that the Constitution was written when a first-past-the-post system existed it is hard to see how full preferencing could have been intended by its authors.  It is also hard to see why the conclusion of Gibbs CJ in 1984 ("While the Constitution requires electors at a Senate election to vote for individual candidates, it does not forbid the use of a system which enables electors to vote for individual candidates by reference to a group or ticket.") should be overturned.

One oddity that may receive attention is that an above-the-line vote can now never flow to an ungrouped candidate as a preference.  However, disadvantage to ungrouped candidates having been deemed a non-issue before, the point should be made that in fact this system is much better for ungrouped candidates than what has gone before.  They cannot receive ATL preferences, but it is so much easier to vote for them below the line, and ungrouped candidates tended to do lousily on ATL preferences anyway because they had nothing to trade.

Some Hare-Clark fans hope that above-the-line boxes might be struck down entirely, resulting in a return to the 1949-1983 system but without compulsory preferencing, but that doesn't seem terribly likely.

I may cover the High Court challenge in more detail later.

PS More Threshhold Fun: Nick Casmirri has produced some more startling threshhold results.  Because threshholds knock out micro-parties who polled below the threshhold, they in some cases would have caused other minor parties to win off low vote shares (when in reality, those parties were stopped by some micro-party snowballing ahead of them and then losing!)  Thus, using a 2% threshhold, Casmirri finds wins at the 2013 election for PUP off 3.4% in NSW and the Liberal Democrats off the same in Western Australia - neither of these parties actually got near! As bizarre as the old group ticket system was, the addition of threshholds would have made it in some ways an even more surreal twilight world.

See Nick's detailed analysis here.