Yes, rules matter, and they need to be fixed


On Saturday, Fairfax newspapers published an op-ed from Richard Denniss, chief economist at The Australia Institute, arguing against the proposed reforms to abolish group voting tickets (GVT) and introduce optional preferential voting (OPV) for Senate elections.

Unfortunately the article gets quite a lot wrong about the current arrangements, why there is a need for change, and who is supporting the change. I have written about group voting ticket reform a number of times recently (see this and this).

I wanted to specifically address a number of the claims Denniss makes in his op-ed. I’m going to respond to a number of the things he says, below the fold. There’s a lot to say. If you want a high-level summary of why reform is necessary, check out my two links above, or this piece from Antony Green.

Denniss correctly states that the process of deciding on our voting system is very serious, and it’s for this reason that reform is desperately needed. The current Senate voting rules are broken, and if we care about public faith in democracy and fair election results, change is needed.

“If we are to do something as significant as change the way our Senators are elected, it seems obvious that a Senate inquiry in which all of the current Senators, and the public at large, can have their say.”

Throughout his op-ed, Richard Denniss implies that this proposed change has been rushed without the opportunity for debate and discussion.

In fact, the federal Joint Standing Committee on Electoral Matters has specifically held inquiries into this proposed reform, which involved submissions and public hearings, before the current proposal was recommended unanimously by the Joint Standing Committee on Electoral Matters in May 2014. People interested in this issue should take a look.

“…the Abbott government and Greens Senator Lee Rhiannon are reportedly discussing a deal to make fundamental changes to the way senators are elected.”

This is misleading in terms of who supports change, and where the proposal has come from.

The proposal to abolish group voting tickets has been around for many years. The NSW Parliament abolished group voting tickets after the 1999 state election, and Bob Brown was advocating for GVT reform long before Lee Rhiannon entered the Senate.

The JSCEM proposal was unanimously supported, including from Labor, Liberal and Greens members. This report followed on from submissions and public hearings. The proposal was supported by a number of electoral experts and the administrative wings of the Liberal, Labor and Greens political parties.

Denniss also repeatedly implies that the proposal is supported only by Rhiannon, not by other parts of the Greens, but the proposal has been repeatedly endorsed by the party’s governing council, most recently in June, and formed a key part of the party’s recommendations to JSCEM hearings over a number of electoral cycles. Despite opposition from some Labor politicians such as Sam Dastyari, the proposal also has substantial support amongst Labor figures.

Beyond the parliament, a wide range of experts have endorsed the plan, including ABC elections analyst Antony Green, who rarely intervenes in policy debates.

“The US is different. In the 2000 presidential election, George W. Bush and his “hanging chads” defeated Al Gore in the crucial race in Florida. While the dodgy, and privately owned, voting machine played a role, so too did the first-past-the-post system…there are changes on the table that would move us closer in that direction.”

There is a fair comparison between the 2000 presidential election and our Senate voting system, and it can be seen in the 2013 election in Western Australia. When voting systems fail, public faith in democracy is damaged and unfair election results ensue, but Denniss has taken the wrong lesson from Florida in 2000.

There is no comparison between ‘first past the post’ (where voters can only cast a vote for one candidate) and ‘optional preferential voting’ (where voters can cast a preferential vote for as many candidates as they wish). Green voters in Florida in 2000 had to face a choice between casting a vote for their preferred candidate (Nader) or for the candidate who had a chance of winning (Gore). Under OPV, voters have the choice to do both. Sure, sometimes voters will choose not to exercise a preference and ‘exhaust’ their vote, but the important difference is that the voter has the choice.

Electoral expert Peter Brent, in response to Denniss’ invocation of Bush v Gore, pointed out that Gore would’ve been elected President under an OPV system. Only a tiny proportion of Nader voters would’ve needed to exercise a preference for Gore to push him ahead of Bush.

In a lot of ways, the WA Senate election failure was Australia’s own Bush v Gore – a voting system failed to produce a result, at a high cost for the credibility of our electoral institutions.

This election failure was in large part due to the group voting ticket system which Denniss defends. The GVT system tends to produce a lot of possible points in a count where a very small number of votes separate candidates, and the exact order of elimination can produce very different results – it was this process which led to the ridiculous 14-vote gap between two candidates, neither of whom could be elected, which would have decided two Senate seats. Without this tiny gap, the subsequent recount, which pushed the AEC to breaking point, would have never been required.

These kinds of abnormalities don’t happen when voters individually choose their vote, because voters don’t act in a monolithic way, unlike parties exercising complete control over preferences.

This problem hasn’t yet been fixed, and if we go on using the GVT system, this problem will happen again, likely in bigger states, and will serve a harsh blow to the faith of Australian voters in our democracy.

“The apparent impetus for this most unlikely of political allegiances is the result of the 2013 Senate election…Such results have led to widely held view that there is ‘something wrong’ with our voting system and that we “need to do something”.”

Yes, the immediate cause of the current reform process is the 2013 Senate election, but not for the reasons Denniss suggests, and the proposed reforms are not new ideas.

The proposals have been pushed forward by the Greens and non-partisan experts for many years, and already legislated in NSW in 1999. The JSCEM proposal isn’t motivated by a desire to “do something”, but to take a specific action which has been called for for many years.

It is to the shame of the Coalition and Labor Party that, until 2014, these parties continued to support the dodgy GVT system. We should be encouraged that they have finally come around to support the proposal, and not use their support to cast aspersions on an idea that originated outside the Parliament.

“There is no perfect way to elect representatives in a democracy.”

That is true, but that doesn’t mean that we can’t improve our current system. The proposed reforms have already been used in New South Wales for four successive state elections along with many local council elections, and they work well. This isn’t some pie-in-the-sky utopian vision.

“…it’s not surprising that the vast majority of Australian voters trust parties to allocate preferences and opt to vote above the line.”

Denniss acknowledges that it is has been increasingly more difficult to cast an informed, formal vote under our current Senate system. When the only alternative is to give preferences to a large number of candidates (often over one hundred), many of whom are completely unknown, with a high risk of making your vote informal, it’s not surprising that voters choose to vote ‘1’ above the line.

The fact that voters choose to vote above-the-line is not evidence of ‘trust’, it’s evidence of a lack of options. There is a lot of evidence that voters do not trust their politicians, and engagement with political parties is in the gutter. GVTs exist for convenience, not for trust.

“The “problem” that the Liberals and Rhiannon seem determined to solve is that of senators with very low primary votes being elected on the basis of preference flows from other parties.”

Again, Denniss misinterprets what problem the reform is aiming to solve. The problem is not senators being elected with “very low primary votes” – under the NSW system, an Animal Justice Party candidate was elected off 1.8% of the primary vote. Importantly, he was elected on preferences from voters who actually marked those preferences themselves.

The problem is candidates being elected with very low primary votes on the basis of preferences from voters who don’t understand where their vote is going, and from parties with completely different political principles. The GVT system creates a “preference lottery” which encourages a bunch of parties with completely different political principles to throw all their votes into a single pot in the hope that they’ll emerge on the top of the pile.

It also isn’t fair to a lot of other small parties who poll a lot more votes than candidates like Ricky Muir, but don’t get anyone elected thanks to arbitrary preference flows. Sure, it will become harder for very small parties to win without group voting tickets, but they can win if they win over more voters and actively communicate with their voters about how they would like them to exercise preferences.

The system also allows major parties to direct their preferences in ways that don’t reflect the principles of the voters. The ALP elected Steven Fielding to the Senate in 2004, and the Democratic Labor Party to the Victorian Legislative Council in 2006 – based on a preference flow that would have never happened if Labor voters had filled out their preferences themselves. The first example of preference corralling took place in 1984, when the major parties swapped preferences to lock out the Nuclear Disarmament Party’s Peter Garrett.

“…it is worth highlighting that there has been little concern about the very small number of votes required to elect major party senators. A whole 749 people in all of NSW put a [1] next to the Coalition’s John Williams. Despite coming in 20th in primary votes, he took one of the six seats on offer.”

As stated above, the current voting system strongly discourages voters from casting votes for individual candidates, in favour of voting for a party. Unsurprisingly, the vast majority of votes are cast for lead candidates, and subsequent candidates are elected on their preferences. The important point is that voters who cast a ‘1’ for the Liberal/National ticket wouldn’t be surprised that their votes flowed to John Williams, who was #2 on that ticket.

If you want to argue for votes to be cast for individual candidates, not parties, then go for it. I am a fan of the Robson Rotation system in ACT and Tasmania, which involves even less trust in political parties, by getting rid of above-the-line voting entirely and randomising the order of candidates within a party column. But that’s a different argument.

“There are other options. David Leyonhjelm has supported a call by the respected psephologist​, Malcolm Mackerras for optional preferential voting for those voting below the line.”

This is actually part of the proposed reforms, but it isn’t enough on its own.

“And both yours truly, and Professor George Williams have separately argued for a simple threshold test to be added to the current system in which the existing above and below the line system is maintained but only a candidate who receives more than 2 or 4 per cent of the vote can ever be elected.”

This would actually be the worst of both worlds, and justify the criticisms that have been made of Senate reform – it would significantly tilt the playing field in favour of big parties while still allowing them to control their preferences and use them to manipulate results.

We’ve never had an arbitrary threshold in Australian federal elections, for a good reason. While we have quotas for Senate elections, it’s always possible to reach this point with the help of preferences. But under a 4% threshold, a candidate polling 3.99% would be knocked out, while a candidate with 4.01% could be elected, regardless of preferences.

Liked it? Take a second to support the Tally Room on Patreon!
Become a patron at Patreon!


  1. OPV in Victoria in the upper house didnt work because it probably wasnt promoted as well as other states where it has been abused by the majors IE Queensland.
    The reaction of the minors last election showed how much destain they had and was listed in the LNP report.
    OPV does revert very quickly to FPP especially where funding and media that has a biased towards the majors clicks into play.
    I could name most political parties policies etc and their beliefs and objectives but the majority of voters say 30% ALP and 30% coalition will never bother to even check. Years ago I was on the ABC when people were phoning in and saying they had no one to vote for…yet when Steve Austin (ABC) read through some of the websites of minors the callers had no knowledge of who or what they stood for. In ticket votes similar parties stick together and if they cant group enough they past to their major of choice IE …Greens SPA SA CPA etc would go to the ALP, Whereby DLP KAP FF CDP and Aust Christians and occasionally LDP depending on the candidate will do the same. So inevitably the votes land where they should. In 2013 though the introduction of legal agreements caused confusion and led to some disparities. Perhaps that area should have been looked at and addressed and maybe we wouldn’t be in this situation. A vote of 2% across the state is quite a large vote and if say 3 or 4 get that size vote and are ignored by either a bad system or ignorance of the electorate then that is not good for any democracy. You may say why don’t a number of parties merge. But although group one may be socialists and group two may be conservative they do hold very different economic politices on how they will get there. But overall their major objectives and principles arent compromised which is what happened in 2013.

  2. Thanks for your answers Ben and Antony.

    With all of the discussion above, and many(most) of the issues on the table, would be great to get Richard Denniss (@RDNS_TAI over here to see his response (to this and the valuable additional discussion from Kevin Bonham on his blog).


Comments are closed.