The anti-democratic alternatives to Senate reform


Until Labor’s recent cynical turn against reforming the Senate, there was general agreement amongst the main political parties and electoral experts that something needed to be done to deal with the Senate. The number of candidates running for the Senate has skyrocketed, which has made it significantly harder for voters to vote (by making ballots bigger and making it harder to find who you want to vote for) and made it significantly harder for the ballots to be counted (as seen in the 2013 WA Senate count). It has also allowed minor parties on tiny votes to pile up votes and become viable candidates for election.

These aren’t the only reasons why Senate voting reform should happen, but they are the reasons why it’s so urgent and part of the reason why this issue has finally gained support from a major party after decades of major party indifference.

Apart from the current proposal, there have been a number of other “solutions” to these problems which, unlike the current proposal, do nothing to reduce the power and control of the major parties while making life harder for all minor parties, and would genuinely be bad for Australian democracy.

They should give fans of minor parties pause: there are alternative reforms which would make life much harder for small parties, and could be back on the table if the current proposal fails.

The first way that the major parties attempted to clamp down on the spiralling size of the Senate ballot was to raise barriers to entry. Before the 2013 election, the major parties voted through significant increases in nomination fees for both houses. For the Senate, nomination fees increased from $1000 to $2000 per candidate. Bear in mind that groups need to run two candidates in the Senate to get a box above the line, so in practice it cost $4000 per state for a party to run.

It was a total failure. The number of candidates in the Senate increased from 349 to 529, an increase of 51.5%. There was also a 39% increase in the House of Representatives.

Despite this approach failing, the original plan after the 2013 election included further increasing barriers to entry. These included increasing the number of members needed for a party from 500 to 1500, as well as requiring more rigorous processes to verify that those members are actually members. All of this was included in the JSCEM report which also contained the proposal to abolish group voting tickets and introduce optional preferential voting to the Senate. The report also contained a proposal to allow parties to register only in certain states on a per capita basis, which would have lowered the membership numbers for any party wishing to start in any state other than NSW.

These proposed party registration rules reflected how our current system gives a lot of privileges to political parties which aren’t available to independents, and attempt to limit who gets access to those privileges. Parties get to have their party name on the ballot (particularly useful in the Senate where independent groups get nothing except the group letter above the line) and they get to nominate as many candidates as they want from a central office without needing nominators.

(In contrast, independents are required to be nominated by 100 electors in the state where they are running, which sets a barrier but also requires some local connection. Parties, in contrast, are free to run candidates across the country even if they have no presence in a particular state, which allowed the Liberal Democratic Party to run as their lead candidate in Tasmania a person who was also the Mayor of Campbelltown, NSW.

Antony Green has suggested requiring parties to have nominators as an alternative solution to this problem and I tend to agree.)

In the end, the increased party registration rules were dropped from the JSCEM package due to opposition from the Greens, who objected to something that would make it harder for new parties to form. Part of the problem with setting higher administrative barriers for parties is that they are much easier to clear if you have a lot of money (and you are completely exempt if your party has a federal MP), but can knock out genuine parties without a great deal of money. Like with nomination fees, this approach does nothing to stop wealthy dilettantes but is effective at stopping community-based campaigns with small budgets.

The other approach suggested is to impose a threshold on the number of primary votes a party needs to get to be eligible to win seats. This suggestion has been around for quite a long time – Liberal senator Helen Coonan proposed this solution as early as 1999, and she was joined by Democrats leader Meg Lees. More recently, Sydney Morning Herald economist Ross Gittins and former Labor deputy leader Anthony Albanese suggested something similar as an alternative to abolition of group voting tickets.

(Long before the current spring of minor parties, both major parties were looking for ways to lock out even large minor parties. Paul Keating floated the idea of electorates in the Senate which would lock out all but the major parties, and Coonan’s proposal suggested a threshold as high as 80% of a quota.)

At first this approach can be seductive: it would prevent candidates like Ricky Muir with no public profile winning seats, without seeing votes exhaust or increasing the informal vote.

But at its heart it is anti-democratic.

The current system’s problems are various, but one advantage of the current system is that it sometimes benefits small parties, and sometimes big parties. Imposing thresholds would exclusively hurt minor parties while keeping all of the power in the hands of major parties.

While the media (and many politicians) have focused on people like Ricky Muir, the original proponents of GVT reform were not aiming to eliminate candidates on a small vote, but rather produce a more honest, transparent and democratic system.

None of that is fixed by thresholds. Voters still won’t understand their preferences, and small gaps in the vote can produce big differences. It would still be difficult for voters to opt out thanks to a severe assymetry between voting above and below the line.

There is a big difference between quotas and thresholds. A quota is a practical reality – the smallest number of votes which six people can get and not leave enough left over to elect a seventh. A threshold is an arbitrary distortion. And thresholds are set on primary votes, not on the final vote after preferences. If a party polls one vote below the threshold, all the preferences in the world won’t help.

The context of thresholds used in party list systems in Europe and New Zealand is very different. In those jurisdictions, a large number of MPs are elected in a single region (and sometimes the entire Parliament is elected to represent the entire country), so if there was no threshold candidates would be elected on truly miniscule votes. A quota naturally imposes a practical threshold, but it doesn’t set a hard-and-fast barrier, because preferences will flow at different rates at different elections and to different candidates.

It’s also possible that such a move is unconstitutional. Section 7 of the constitution says that the Senate shall be “directly chosen by the people of the State”. While it seems to me that this clause could simply be confirming that senators should be elected, not appointed by state Parliaments (as was still possible in the United States when the constitution was written in 1898), it has been generally interpreted as prohibiting any kind of party list system, and is part of the reason why we still have below-the-line voting.

You couldn’t impose thresholds on individual candidates without knocking out second and third candidates for Liberal and Labor, so it would be necessary that such a threshold would apply to all candidates in a party group on the basis of the party group’s total vote, not to individuals.

Talk about thresholds should worry anyone interested in a healthy system of proportional representation and minor party representation in the Senate.

The proposed model isn’t ideal for minor party representation – like many others, I’d love to see a lower quota – but it still allows for diversity of representation, and for preferences to flow if voters choose to mark them.

If this reform isn’t passed, we should expect to see future attempts at reform, and they may be much less friendly to minor parties.

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


  1. My proposal for thresholds was that they would be applied *after* the surpluses of elected candidates had been distributed, so that they wouldn’t affect the 2nd and 3rd candidates of the major parties. Once all surpluses had been distributed, *then* all remaining candidates who did not have 25% of a quota would be eliminated and their preferences distributed. But this now ancient history – I now support the current reform proposals and I think Labor’s opposition to them, though understandable in terms of party self-interest, is misguided.

  2. Ben,

    Do you think Australia should just have a system similar to New Zealand’s?

    Switching it into the Senate would have issues, because of the low number of seats (though a full 12 seat election or with by slightly increasing seat numbers, would mean if you got something like five or six percent of the vote then you got a seat), and I wonder if in NSW with the larger number of seats (especially if they switched to full upper house elections) if it would fit about right.

  3. MMP can and does produce quite disproportional results because it uses a threshold. Until recently Turkey excluded more than 40% of the electorate from having their votes counted excuse of the threshold. At the last election in Germany the Christian Democrats would have won more seats if some of their supporters had voted for the free Democrats instead of the Christian Democrats.

    I doubt that the high court would ever agree to the serious discrimination of not counting the votes of some electors, for whatever reason.

    Lastly, MMP has a serious constitutional problem in Australia because 1 the constitution commands that senators and MPs be directly chose and 2 the constitution mandates that states have fixed numbers of senators and MHRs. In both Germany and New Zealand the number of representatives fluctuates to deal with overhang seats and balance seats.

  4. I think MMP is a reasonably good voting system. It’s not perfect. I don’t like how some MPs have local representation responsibilities and others don’t, and how local electorate votes don’t make a difference to the overall result, with a few weird examples. I also don’t like how the lack of preferences means votes are thrown out entirely in NZ and Germany.

    I think Hare-Clark (either the ACT-TAS system or the new Senate system) makes more sense for the House of Reps than MMP, partly because voters are more experienced with it, partly for the constitutional issues Alan listed, and partly because of things I don’t like about MMP (listed above).

    Hare-Clark is entirely compatible with our current constitution so could be implemented without a constitutional amendment. They would make districts larger (you could offset that a bit by increasing the Parliament’s size) and some districts would have different numbers of MPs to fit with the number of seats per state, but that’s not a deal-breaker.

  5. kme

    For reasons that are obvious, I prefer to keep legal fictions as far away from the electoral system as possible.

  6. Good article as always Ben. One minor error though. The US Constitution was written in 1787 and came into effect from 1789, not 1898 as you write.

    The 17th amendment, which requires the direct election of senators was ratified in 1913.

  7. MMP is actually quite rare. The user nations are Bolivia, Djibouti, Germany, Hungary, Lesotho, Mexico, New Zealand and Romania. Only Mexico and Germany have larger populations than Australia. None have a larger area than Australia. Only Germany and New Zealand appear on the Economist Intelligence Unit list of full democracies.

  8. Apart from strategy manipulation, it’s equally worth noting that the threshold can have large impacts. At the last German election in 2013 the voters who did not have their ballots counted because of the threshold amounted to 15.67% of the total electorate.

  9. Yes and they were concentrated on one side of the political spectrum, so a significant right-wing majority of the vote produced a majority for the SPD, Greens and Die Linke – it’s only because the Left party is locked out of government that kept the CDU in office.

  10. I have a feeling a lot of commentators arguing against the reforms would be happy with a box ‘a pox on all your houses’ that randomly goes to the leading non establishment candidate.

  11. As far as the federal lower house goes, I am fine with it the way it is. If it aint broke don’t fix it

  12. The last German election seriously discredited MMP in my eyes. The FDP and AfD each polled just under the 5% threshold, so their votes, nearly 10%, elected no MPs, and this allowed the three left parties to win a majority of seats with a minority of votes. Only the SPD’s refusal to be in government with the Left kept Merkel in office. I also dislike the fact that it makes no difference who wins the constituencies, because it’s only the list vote that determines the party numbers. An alternative would be “de-linked MMP”, so that half the seats go to constituency MPs and the other half to list MPs proportional to the list votes, but without regard to the constituency results. In Germany, that would have given the CDU/CSU 384 seats to the SPD’s 150, Left 35 and Greens 31, a much better reflection of the mandate that Merkel got.

  13. The German election didn’t so much as “discredit” MMP as it discredited arbitrary cutoffs. That’s not to say cutoffs aren’t without some merit, but they are a compromise on proportionality. As that election showed, they’re capable of throwing up some weird edge cases.

    I certainly agree that the meaninglessness of constituency contents is the most unsatisfying aspect of MMP. I don’t know about your solution though. For one thing, those constituency contests are decided by first-past-the-post.

  14. Delinked MMP is better known as a parallel electoral system. The user nations are Andorra, Democratic Republic of Congo, Georgia, Guinea, Japan, Jordan, Republic of Korea, Libya, Lithuania, Monaco, Mongolia, Nepal, Pakistan, Philippines, Senegal, Seychelles, Sudan, Taiwan, Tajikistan, Ukraine, Venezuela.

    Parallel systems can lead to widely disproportional results, depending on the ratio of PR seats to district seats and the threshold that is applied to small parties. None of those countries are on the Economist list of full democracies.

  15. The “meaninglessness of constituency results” (for the purposes of determining the overall majority) is a feature – that’s exactly what many people appear to want (eg. when they complain about a party winning Government with less than 50% of the statewide 2PP).

    (They’re clearly not “meaningless” since they do determine that constituency’s representative in the parliament).

    One way to fix the threshold issue might be to only exclude the lowest-voted groups until the total excluded vote is no more than, say, 5%. In the 2013 German case, this would have meant that FDP (4.8%), AfD (4.7%) and PIRATEN (2.2%) were included.

  16. I don’t know what criteria The Economist uses but Andorra, Japan, Republic of Korea, Lithuania, Mongolia and Taiwan (at least) certainly are full democracies.

    I agree that first-part-the-post in single-member contests is inferior to preferential voting. If Germany had preferential voting, the SPD would have won more constituencies by getting Green and Left preferences, producing a more proportional result.

    But I’m not in favour of making a fetish of proportionality. Legislatures in parliamentary systems serve two-functions: as a forum for the range of opinions in the community, and as a platform for the formation of stable majority governments. Strict PR does the first but is bad at doing the second (eg Belgium, Italy, Israel, now also Spain). Pure Westminster (eg the UK) does the second (usually) but at the expense of the first. I support a system that gives a voice to minor parties but is weighted in favour of the leading party. Italy and Greece do this by assigning extra seats to the leading party, but delinked MMP with a 5% threshold does the same thing, as I showed above.

  17. I don’t think sarcasm is appropriate, especially as single party majority governments do lock out oppressed groups. It now becomes appropriate to quote the Fourth Doctor:

    “The very powerful and the very stupid have one thing in common. They don’t alter their views to fit the facts. They alter the facts to fit their views. Which can be uncomfortable if you happen to be one of the facts that needs altering”.

  18. Adam, you list a small subset of the countries which use PR – German PR produces reasonably stable governments without needing one-party majorities.

    Belgium’s problems aren’t due to PR, they’re due to much deeper issues around the federation, and Israel has a very extreme form of PR (and other political issues which mean not all MPs are in a position to be considered to be in government – if you don’t let Arab MPs into government the government needs a super-majority of Jewish MPs).

  19. OK, I won’t pursue this, but I must pick you up on one point, where you say of Israel: “if you don’t let Arab MPs into government.” Israel does not exclude Arab MKs from government because they are Arabs. There have in fact been Arab MKs and ministers in past Labour governments. They are self-excluded at present, because the majority of Arab voters choose to elect communists (Hadash), Arab nationalists (Balad) or Islamists (Ta’al), all of which reject participation in Israeli governments. If Arab voters elected Labour MKs as they did in the past, those MKs would be in a Labour government (and also a Labour government would be much more likely).

  20. If you were going to apply thresholds, you’d obviously apply them after transfers from candidates over quota not before, so you don’t eliminate 2nd/3rd candidates prematurely. Although this is rather academic now.

  21. “But at its heart it is anti-democratic.”

    Ben, you make a good argument in your “choose this or get worse” piece, however, like many others, you seem to be overlooking the inherent attack on democracy in even the current proposed legislation.
    To put it ultra-simply, under the current system parties representing 23% of Australians get together and make a plan to vote for each other so as to elect at least one amongst them, even if that one could be an unpleasant eventuality to some in the group. Each micro party’s motive: the possibility of winning a seat, even at the cost of possibly supporting an opponent, transcends perpetually never winning.
    Micro party supporters who don’t like that can still vote below the line, especially now that optional preferential voting will be allowed there.
    So the new legislation will now be denying approximately a quarter of voters the right to vote for one out of a group of parties, merely because that party is not specifically defined.

  22. I completely disagree that this legislation represents an “attack on democracy”.

    Firstly, if you’re going to count up how many voters would have been unrepresented at the last election, you have to take out those who voted for Nick Xenophon, Palmer United and the LDP, all of whom would have won seats under the last election under the new system. Most of the LDP’s voters were probably Liberals who got confused, but either way you shouldn’t count those voters in the 23%. Those voters add up to about 10% of the electorate, so it’s more like 13%.

    Secondly, adding up 23% of Australians and treating them as a block of voters deserving representation as a block is nonsense. These parties have wildly divergent political views.

    If these parties want to run on joint tickets and support each other’s candidate explicitly then they can do that, and if their voters are willing to vote for those joint tickets then they will have a good chance of winning. I’m sure we’ll see some of that, but we won’t see a single mega-party because their voters wouldn’t go for that because of the wide variance in political views. You say that party is “not specifically defined” – I say that party doesn’t exist.

    You say that the micro party’s motive is to have a chance of winning even at the risk of electing an opponent – but is that the motive of their voters? I highly doubt it. This system is about prioritising what voters want over what parties want.

    The idea that anyone is being denied the right to vote is nonsense. They still have the right to vote for whatever party they want. And if they want to preference all the minor parties, they can do that too.

  23. Ben, a very good point about the real percentage of future unrepresented voters. Something I didn’t realise. However, as I see a lot of the PUP vote being a protest vote which would still protest elsewhere, I see the new percentage as closer to 16.6%.
    Be that as it may, if the fringe vote is above 14.3% then ….

    “You say that the micro party’s motive is to have a chance of winning even at the risk of electing an opponent – but is that the motive of their voters? I highly doubt it.”
    Why on earth not? If you look at a circle of micro parties, from Animal Justice to the HEMP party, Motoring Enthusiasts, Fishing and Lifestyle, Shooters and Fishers, One Nation, Australian Christians, Secular Party, Uniting Australia Party etc, yes we occasionally recognise an abhorrent one, but the average fringe party is odd, amusing but relatively innocuous to the average voter. Why would we sacrifice getting our very own, never before elected party, elected, at the risk of supporting an odd ball party or the rare off chance of an obnoxious party?

    “The idea that anyone is being denied the right to vote is nonsense. They still have the right to vote for whatever party they want. And if they want to preference all the minor parties, they can do that too.”
    Don’t you think this is rather disingenuous? To vote ATL for the major parties a first and second preference vote is probably enough to guarantee success, but to vote for a micro-party to possibly succeed you now have to (correctly) tick off above the line approximately 30 boxes. Something not necessary especially considering the (minority) fussy voters will now have optional below the line voting.

    “This system is about prioritising what voters want over what parties want.”
    This is the biggest con of all. There is this manufactured conspiracy that supposedly exists amongst micro-party organisers done in smoke-filled rooms behind closed doors to dupe their respective supporters. Dupe them in what way? Please tell me how the goals of voters and supporters can be anything but the same?

  24. “Please tell me how the goals of voters and supporters can be anything but the same?”

    Because when voters give their own preferences they look nothing like the deals arranged by the parties. There were 20 different parties that got Ricky Muir to the point where he could get elected on Sex Party and Palmer United preferences. He received 100% of the ticket votes of those 20 parties, but just 6% below the line.

    Same in Western Australia, 100% of 18 party ticket votes to Wayne Dropulich, just 13% below.

    If you look at any election that does not involve ticket voting, voters clearly give preferences to parties they know over those they don’t. Those preferences are totally at odds with the lodged ticket preferences.

  25. I cant see anything in the constitution preventing full rather than half senate elections (im sure a helpful person will point it out if im wrong)

    I think that (along with sensible donation and spending caps) that would fix the senate election a lot further.

  26. Ahh, had thought previously that it was in there.

    Well I don’t see a successful referendum coming any time soon so put that on the wish list.

  27. Single member electorates would not give Belgium or Israel single party governments because they have geographically divided voting groups. In Belgium the Flemish and the Walloons would continue to elect their own parties in their own regions. In Israel, Arab parties would be elected in Arab areas (if they were not the victims of gerrymandering), heavily pro-settler parties would be elected to represent the settlements, Orthodox parties would be elected to represent Orthodox areas, Labour would be elected in Labour areas, etc. Where would the stability come from?

  28. Transcripts of the hearing of Day’s High Court challenge to the constitutionality of the Senate reforms are available on the High court website for Monday and Tuesday of this week.The Justices devoted great effort to understanding what the basis of the challenge was – the general tenor of the questioning did not suggest that they were totally impressed. I am not a lawyer but my guess would be that the challenge is likely to get short shrift

  29. It started off with the Chief Justice asking where the usual two page summary of argument was and being told there was none, and shortly after barrister Peter King being reprimanded for going off the track by making normative comments. It then went downhill from there.
    Plaintiffs Bob Day and Peter Madden really could have spent more time into choosing a skilled or conscientious barrister to represent them.

  30. ..and yes, that is the same Peter King who was defeated in a preselection challenge by none other than the current Prime Minister.

  31. Perhaps the part of the judgement with the most enduring importance: The argument that above the line voting amounts to something other than a direct choice of individuals is untenable.

Comments are closed.