Now it’s urgent: why we need to simplify voting for the Senate

Changes to how senators are elected would improve transparency, make voting easier, and stop parties with scarcely any support slipping into the Senate, says Brian Costar

09 September 2013

Print this article Print this article

Tags: , , ,

Bookmark and Share

A record 529 Senate candidates are contesting this election.
Dirk HR Spennemann

First published 4 Sept; updated 9 Sept

WHILE the final composition of the Senate won’t be known for at least a fortnight, it seems likely that some micro-party candidates who attracted very little primary support will be sitting as senators after 1 July next year. In Western Australia, for example, the Australian Sports Party could gain a place with just 0.22 per cent of first preferences – as could the Australian Motoring Enthusiast Party in Victoria, with 0.52 per cent. The Senate is too important to become a playground for recreational groups masquerading as political parties. System reform, including change to electoral procedures, is urgently needed.

Part of the problem is that the Senate voting system is complex and opaque. Senators are elected by way of a multi-member, quota-preferential, single-transferable-vote, proportional system. This mouthful, which is almost as complex as it sounds, is generally abbreviated to STV, despite the vision those three letters conjure up of a nasty social disease.

Under STV, each state is allocated twelve senators, six of whom usually face election at three-yearly intervals. The territories each have two senators, both of whom face the people at each ordinary election. In the House, a successful candidate needs an absolute majority of 50 per cent + one vote; in the Senate, each candidates must achieve a “quota” of votes – 14.4 per cent at the usual half-Senate elections, like the one on Saturday, or 7.7 per cent at the full-Senate elections that follow a double dissolution of both chambers (which last occurred in 1987).

The system allows the preferences of parties and independents eliminated from the count to be distributed in the same manner as in the counting of votes for the House. The major difference, and a key feature of STV, is that the “preferences” of the big parties are also distributed, even after they have won some Senate positions.

This works in two ways. The lead Senate candidates for Labor and the Coalition attract votes far in excess of what they need to win a place. Those surplus votes are transferred down the ticket and help elect other candidates from the same party. Since major parties rarely win exact quotas, their “surpluses” can also transfer on and help elect other parties’ candidates.

Take the example of New South Wales in 2010. After all the first votes were tallied, the Coalition had 2.3 quotas, which it built to three by harvesting preferences from a clutch of small right-wing parties such as the DLP and the Shooters and Fishers. Labor started with 2.2 quotas and was not so fortunate in reaping small-party support; rather than contributing to a third senator, its 0.2 surplus helped to elect a Greens candidate.

Most small and micro parties get elected by receiving the sugar hit of a partial quota from a big party. (The current exception is Nick Xenophon in South Australia, who again won his own quota.) It is this churning of preferences that can permit candidates who poll just a few votes to remain in the count for long enough to get elected.

A claimed virtue of proportional representation voting systems is that they allocate seats in direct proportion to votes won, so if a party wins 15 per cent of the vote it should receive 15 per cent of seats. The Senate isn’t quite like that because the existence of eight electorates (the states and territories) dilutes national proportionality. The system is best termed quasi-proportional.

Why did the parliament adopt this Senate voting system in 1948? Mainly because, under the old system, a party that achieved even the narrowest majority in a state won all three Senate places. Nevertheless, the opposition leader at the time, Robert Menzies, called the change to the present system a “racket” designed to maintain a Labor majority until 1953. In fact, the Labor majority was swept away at the 1951 double dissolution election.

A controversial feature of the current system is the above- and below-the-line voting option. Above the line, voters cast a valid vote simply by entering the number 1 for the party of their choice; below the line, they must indicate a full set of preferences, which can mean numbering dozens of boxes. Above-the-line is a gift to party managers making preference deals of sometimes bizarre ideological inconsistency. It also means that the donkey vote can be crucial, as the Liberal Democratic Party discovered when it polled 9 per cent in NSW largely because it drew first position on the ballot paper. The below-the-line option raises the risk of pushing up the number of accidentally informal votes.

The above-the-line method was introduced by the Hawke government in 1983 to reduce the traditionally high Senate informal vote. But Labor was also keen to prevent a repeat of the 1974 election, when anti-Labor interests packed the NSW Senate ballot paper with dummy candidates and caused a spike in informal votes. The ruse cost Labor a quota, with devastating consequences for the party in 1975.

ON SATURDAY, voters were confronted with a record fifty-four political parties, 529 Senate candidates and 1188 House hopefuls across the states and territories. An unprecedented number of recently created micro-parties also participated. Some may view this as a sign of a healthy democracy, but they would be wrong. At least in New South Wales, the huge number of tiny right-wing parties running for the Senate was the result of the efforts of a political deal-maker, Glenn Druery, who specialises in constructing mutual preference arrangements in order to increase the chance of a victory by someone who enjoys minimal direct support.

The increased number of parties and candidates contesting House seats inflated the informal vote to a record 5.9 per cent. But it got into double figures in some Divisions, especially where there are large concentrations of people whose first language is not English. Traditionally these are Labor seats.

Reform is needed. First, the test for party registration should be much stricter. At the moment, all a “party” needs is $500, a brief constitution, and a list of 500 supporters entitled to be on the electoral roll. This should be increased to 2000 voters and a $20,000 nomination fee (or $10 per member – just over half the cost of a packet of cigarettes).

Second, the Senate ballot paper should be redesigned to abolish above- and below-the-line voting. Instead, voting should be optional preferential. At a normal Senate election, a voter would need to number at least six squares, and at a double dissolution election at least twelve. This would keep informal voting low and put a stop to secret wheeling and dealing over preferences and unexpected successes for candidates with extremely modest direct support.

Third, unless the lead candidate of a senate group or an ungrouped independent secures in excess of 4 per cent of the first vote, they should be eliminated from the count and their preferences distributed. This might seem harsh, but 4 per cent is the threshold candidates need to receive public funding and have their nomination deposit refunded, and seems a reasonable test of whether they have any real support in the electorate. •

Brian Costar is Professor of Politics at the Swinburne Institute for Social Research.

Related articles

Follow Inside Story

Subscribe to Inside Story’s free weekly newsletter

Just enter your email address below, and make sure you click the confirmation link when you receive an automatic email from us:


  1. Philip Eldridge added this comment on 4 September 2013 | Permalink

    Another Senate voting reform possibility would be to allow (either optional or compulsory) numbering of party/groups above the line. This would also require a rule to ignore parties’ and groups’ own preference directions. But it would also entail acceptance of their own preference order WITHIN each grouping.
    If the present system is to be retained, at very least voters must have easy access to each party/group’s intended preference distribution.

  2. David Collett added this comment on 10 September 2013 | Permalink

    I agree with your reform suggestions except for the $20,000 registration fee.

    There are already enough financial entry barriers for young/small political parties (although I was pleased to see the media giving them some time to put their case this election). I don’t see how adding a further financial barrier for the sake of it benefits the candidates in the new party, its members, the government or the public.

    Maybe auditing the party bank account to show the party has raised $20,000 from individual members would be better than handing over a fee for the sake of a fee, but even if that has been met how is it really relevant?

    If an independent forms a party in his or her electorate and each founding member is able to convince one more person in his or her electorate to vote for the candidate then the 4% min vote has already been met.

    I hope for the day where I can vote for a party who refuses to accept money other than membership fees and is still able to make headway.

    Good article though, I was hoping that cutoff idea would come up (your third point).

  3. Barry Healy added this comment on 10 September 2013 | Permalink

    Brian Costar’s recommendation of raising the deposit and the number of members for a party to be registered is inherently undemocratic and would just lead to a more widely felt alienation from the political system than there is now.

    Allowing people to number preferences above the line would be preferable.

    Also, there is no reason why the AEC can’t publish a list of every party’s preferences and post it on the wall of the polling station for all to read.

  4. Chris Curtis added this comment on 10 September 2013 | Permalink

    To eliminate from the count and distribute the preferences of “the lead candidate of a senate group or an ungrouped independent” unless he or she “secures in excess of 4 per cent of the first vote” would be a disgrace. This proposal, which is being pushed ad nauseam at the moment because the “wrong” people got elected to the Senate, is profoundly undemocratic. That it is so is on plain display in the wording for the number two and number three candidates of the major parties and of the Greens who do not secure the 4 per cent are not to be eliminated.

    Stephen Conroy (with 780 votes), Julian McGauran (with 1190) and Judith Troeth (with 829) were all elected in 2004, yet no one objected to those three getting into the Senate on preferences. Fiona McKenzie of the National Party was elected in 2010 from an initial 0.03 per cent of the vote in 2010. No one objected to that. The fact that their preferences came from within their own group is irrelevant. The single transferable vote is designed to elect individuals, as required by Section 7 of the Constitution, which states senators must be “directly” elected. Under STV, all votes are equal. The vote of someone who supports a minor candidate is not of less value than the vote of someone who supports a major candidate. That voter is entitled to have his or her vote remain in the count until the end. To exclude it or discount it because it went to a minor party candidate is the antithesis of democracy.

    There would be nothing undemocratic even for a person to win a seat from an initial zero per cent. No one wins a Senate seat until they reach a quota – 14.3 per cent for everyone. The vote they start on is immaterial, as is whether or not they get their preferences from their own party or another one. The micro-parties got around 20 per cent of the vote in every state, so it is perfectly democratic that they end up with a senator from every state.

    There is no problem whatsoever with a candidate’s election to the Senate on a tiny initial vote as no one gets there until they reach the quota, 14.3 per cent. The issue is not the size of the initial vote. The issue is the difficulty of voting below the line.

    The solution is to make preferences below the line optional after a certain number. Some suggest this should be the number of candidates to be elected, but that number is too low. The earlier votes become “exhausted” in a the single transferable vote system of proportional representation, the less proportional the system becomes and later seats are filled by people who do not even reach a quota. In the Victorian Legislative Council, where that rule applies, parties typically run five candidates (the number of seats to be filled) even though they have no hope of winning five. This locks up below-the-line votes inside the party and means the above-the-line group voting ticket has more influence on the result. I suggest that preferences be optional after the first 20.

    I have no problem with above-the-line votes remaining, as this has reduced the informal vote, but I would restrict the number of preferences a group ticket can have to 20. That would be an incentive not to create phoney parties that last one election and are heard no more because parties would not be able to control the preferences al the way to 97.

    I do not support preferences above the line, which is the NSW Legislative Council system and which is favoured by the Greens, as it simply entrenches parties rather than individuals and stops you voting ‘1’ for the number 2 candidate of a party, when you ought to have the right to do so. The Greens like it for one reason: they think they will get Labor preferences and well remember that the Victorian branch of the ALP used preferences to put Family First into the Senate in 2004 and the DLP into the Victorian Legislative Council in 2006.

    The kneejerk reaction to the Senate having people elected to it whom other people do not like is to reduce the voters’ democratic rights, via thresholds or above-the-line preferences. There are several steps that can be taken outside of the voting system itself to reduce the number of candidates without reducing voters’ democratic rights.

    Groups run more candidates than can ever be elected. I suggest a sliding scale of nomination fees: $2,000 for one candidate in a half-Senate election (i.e., an independent not in a group and without the right to issue a group voting ticket), $10,000 for a group of two candidates, $200,000 for the third candidate in a group (meaning only the Coalition and the ALP would run one as they the only two parties ever to get a third candidate elected), $300,000 for the fourth candidate in the group and $500,000 for the fifth and sixth candidates each. This move would stop parties padding their ticket with people they know will never be elected. This move alone would have cut 13 candidates from the Victorian ballot without removing a single party.

    Parties should have to be registered for at least two years to run as a group. This move would cut out phoney parties that appear for one election and then disappear. I hesitate to say how many groups this would have cut out this year. I stress this is not to stop individuals from standing. It is to stop pretend parties.

    These steps combined with optional preferences below the line would make it easier for voters to have a meaningful say.

    However, the debate is not really about giving voters a say. It is about a political and media elite wanting to ensure that outsiders do not get elected.

  5. chris gow added this comment on 10 September 2013 | Permalink

    I don’t think you can entirely ascribe the success of the Liberal Democrat party in NSW to the donkey vote. The name of the party is quite misleading and I suspect that the vast majority of its votes came from intending Liberal voters who mistook Liberal Democrat for Liberal National, being first on the ballot paper made this error much more likely.
    The LNP vote in for the HOR in NSW was 47%, in the Senate it was 35%, a 12% gap. In Victoria the HOR vote was 43%, for the Senate 40%, a 3% gap.
    The Liberal Democrat vote in NSW was 8.9%, in Victoria 0.12%.
    One may think NSW voters were inspired by the brilliant policies and campaign of the LDs in that state but I can assure readers no one had ever heard of them.
    So at least 4 of the 40 senators-elect have been elected by quirks of name, ballot placement and the preference system; a disgrace.
    The AEC needs to reconsider the name of the Liberal Democrats and I agree a threshold percentage of primary votes is needed before a party can be elected to the senate.

  6. valerie yule added this comment on 10 September 2013 | Permalink

    Senate preferences

    I followed up some of the parties I favoured,to see the preferences which followed if I voted for one of them above the line.
    I was horrified!

    Costar’s solution is far the best.

    (What happens to the $20,000 nomination fee)

  7. Andrew Brion added this comment on 10 September 2013 | Permalink

    Good article – thanks.

    In an ideal world, it’s probably true to say that anyone who voted above the line really only has themselves to blame if their vote went somewhere other than intended.

    But one must concede that the Senate voting mechanism is probably poorly understood amongst the electorate and, in addition to some education, at least some small change is probably required. However any change should be minimal.

    Most likely – as you suggest – it would be sufficient to simply abolish above the line voting, get rid of party tickets, and simply require that voters number boxes up to at least six (or twelve in the event of a full house election/double dissolution).

    In reality, removing above the line voting then almost certainly obviates the need for a minimum quota (be it 4% or less) as the system becomes far less open to gaming by groups of the most minor parties.

    In fact 4% is likely far too large a hurdle when six candidates are being elected: and it really is an unnecessary diminishing of democracy once voters clearly understand how their votes would be transferred from one candidate to another by having numbered the boxes themselves.

    I’d also be somewhat wary of a large financial impediment to grassroots democracy, and $20,000 is quite a sum – unless it is simply intended as deposit which would be returned and which demonstrated some degree of viability.

  8. Hassan Moussa added this comment on 11 September 2013 | Permalink

    While I agree 100% that the Senate voting system needs to be reformed for better transparency and more credible representation, I feel the House of reps voting system needs to be overhauled. I still feel though that the Australian voting system remains one of the bets in the world. But it is shameful that the size of the informal vote at the Saturday’s election was close to 15% in some Western Sydney seats. With so much focus on Western Sydney, how can between 10% to 15% been so disenfranchised.

    informal was the third political force in Australia. If you consider that the margin that changed government was around 3% this means five times more people did not have their votes counted, whether intentionally or otherwise is not known and will never be known. If you compare though against seats with higher level of education the informal vote reduces to less than 6%. Time to make the process simpler and more expressive of people’s wishes.

    Lets examine some seats (figures taken from AEC website on 10 September 2013), where the biggest winner in the 2013 Federal election was the Informal vote:

    In Fowler (near Fairfield) Chris Hayes seat – Informal was 14.57% or 11,480
    In Blaxland ( Jason Clare’s) – informal vote was 14.28% or 10,887
    In Watson – Tony Burke’s seat – informal vote was 14.26% or 10,928
    In Chifley (Ed Husic’s seat) – informal vote was 13.36% or 11,070
    In Werriwa – Laurie Ferguson seat – informal vote was 13.28 % or 10.011
    In Barton (St George area) result still undeclared in the closest local election ever – informal vote was 12.41 % or 9,477
    In McMahon – Chris Bowen’s seat – informal vote was 11.63 % or 9,229
    In Parramatta (Julie Owens seat) – informal vote was 11.18 % or 8,128
    In Reid – where john Murphy lost – informal vote was 10.07 % or 7,960
    In Banks (where Daryl Melham lost) – informal was 9.81% or 7,654

    But for example in areas of higher incomes and education
    In North Sydney – Joe Hockey’s seat – informal vote was 5.59% or 4,225.
    in Cook, Scott Morrison’s seat near Sutherland – informal vote was 5.97% or 5,033
    In Sydney – Tanya Plebersek’s seat – informal vote was 6.82% or 4,927
    In Hughes (Craig Kelly’s seat) – informal vote was 7.67% or 6,191
    In Grayndler, (Anthony Albanese’s seat) – informal vote was 7.28% or 5,798

    Time for action to reduce informal voting.

  9. Brian Costar added this comment on 11 September 2013 | Permalink

    First, thanks for all comments.

    Some quick responses

    1. I would be happy to have optional preferential voting (OPV) above the line, as a second best option.

    2. The Electoral Act (s 216) already requires the display of group voting tickets at polling places, but few people consult them. If lots did the polling place would be clogged up and create delays.

    3. 2000 members is 0.00012 of the total roll. $20,000 sounds a lot but surely those 2000 are committed enough to the party to stump up $10 each. Some form of refund based on electoral performance would be acceptable to me.

    4. I agree that any threshold should not be higher that 4%–which is still less than one third of a half senate quota, but 2% would be acceptable to me.It may be possible to have different thresholds for full and half Senate elections.

    4. Informal voting in the House is indeed a problem. It has two causes, multiple names on the ballot in Divisions, and the presence of a high proportion of voters from non-English speaking backgrounds. OPV would solve it, but this is unlikely to be supported by the ALP, the Greens or the Nats.


Send us a comment

We welcome contributions about the issues covered in articles in Inside Story. We ask contributors to provide their full name for publication, but if for any reason you need to use a pseudonym please submit your comment to us via email. Because all comments are moderated, they will not appear immediately. Your email address is never published or shared. Required fields are marked *.