Reforming the Australian Senate

by Sarah John // Published June 6, 2014

Having earned a law degree and currently seeking a Ph.D. in Political Science at Finders University, Sarah John is an adjunct professor of Political Science at Cal State Fullerton. She will join FairVote as a research fellow in August 2014. In this guest blog, she explains the current reform debate about Australia’s multi-seat form of ranked choice voting  in federal Senate elections and how there is serious talk of modifying it to a form more like what FairVote has proposed in the United States. 

Australian September 2013 Election, Seats won in the Senate by Party

Australian Senate


Since 1948, Australia has used a version of the Single Transferable Vote (STV, known in the United States as ranked choice voting), a fair representation electoral system, to elect its Senate. Typically, Senate elections are uncontroversial and a little ho-hum. Attention instead centers on the House of Representatives, where the government is formed in Australia’s parliamentary system.

 The September 2013 election proved an exception. Nine parties—including the Sports Party and Motoring Enthusiast Party—and one independent candidate won seats in the Senate. Causing no small degree of controversy, several elected candidates relied almost entirely on preferences flows from other parties to win election. Ninety-six percent of the vote tally for the Motoring Enthusiast Party’s elected candidate came from preferences from other parties. Only 0.5% of voters in the state of Victoria backed the Motoring Enthusiast Party candidate as a first choice, but that candidate still won one of Victoria’s six Senate seats being decided at this election.

Adding to the controversy about the election, the Australian Electoral Commission—the independent non-partisan agency in charge of running elections nation-wide—then lost 1375 Senate ballot papers in the state of Western Australia. These 1375 ballots (out of 1.3 million ballots cast) were enough to alter the election result and the High Court ordered a new election for the whole state, held this April.

In the aftermath of this more “interesting” election than usual, political parties and interested citizens have aired their views to the Joint Standing Committee on Electoral Matters (JSCEM), the federal legislative committee in charge of developing amendment proposals for Australian election law. Proposals for reform of the Senate voting system were many, including the introduction of optional preferential voting, the elimination of “Group Voting Tickets”, increased ballot access and party registration requirements or thresholds that candidates must reach before their preferences will be distributed. More restrictive requirements for the inclusion of parties on the ballot, the abolition of Group Voting Tickets and the introduction of optional preferential voting all look likely to be made law after JSCEM unanimously recommended them in its Interim Report in May. While the motives of the politicians making the recommendations are not entirely pure, overall their recommendations will bring the Australian Senate voting system much closer to the sort of fair voting system advanced by FairVote.

But, first, a bit of background into the unique version of STV used in the Australian Senate and the problems that have evolved from the uniquely Australian practice of forcing voters to rank all candidates in order to cast a valid vote.


The Australian Senate: A Primer

 Modelled on the US Senate, the Australian Senate is a vital part of the country’s legislative process. It possesses powers equal to the Australian House of Representatives, except that budget bills cannot originate in the Senate (as is also the case in the US Senate). The Senate is composed of 12 Senators per state (and 2 per territory), with half elected for their 6 year terms at each 3-year election. Initially elected by simple plurality voting in single member districts, between 1919 and 1948 Senators were elected using preferential block voting in multi-member state-wide districts. But this system delivered large—exaggerated—majorities for one or other of the two major parties. For example, in 1946, the Australian Labor Party (ALP) received 52% of the Senate vote nationwide but won 84% of the seats; in 1943 it received 55% of the vote and won 100% of the Senate seats. Perhaps as a desperate effort to mitigate losses, ALP Prime Minister Ben Chifley, whose government was rapidly losing popularity, almost doubled the size of the Senate and introduced STV (see this piece from John Uhr on the motivations for introducing STV in 1948). 

In 1948, Australia’s two major parties anticipated they would still win the overwhelming majority of seats in the Senate under STV—but more proportionally to each’s vote share. However, over the 65 years STV has been in operation, it has provided a path for minor or break-away parties whose share of the vote is too low to be elected to the single-member districts in the House of Representatives to have a voice in Australian politics. 

With this path for minority representation has come consequences. Over time, the (first preference) vote share of minor party and independent candidates has increased: starting at 5% of candidates when the reform began to 15-20% in the 1970s and 1980s and reaching a record high of 32% in 2013. 

The minor party and independent vote in the House of Representatives has also increased, but remains lower—in part a consequence of widespread perceptions of the Senate as a “House of Review” against the House of Representatives, which is absolutely dominated by the major party in government. Indeed, the Senate’s website describes the Senate as “a house of review and a powerful check on the government of the day.” In this understanding, it is desirable that the Senate be made up of different parties than the House and many voters split their ticket. In the 2013, 21.1% of the first preference vote in the House of Representatives went to minor parties and independent candidates. Only 5 independent or minor candidates were elected to the 150 member House (3.33% of the total seats).

With a fair representation system in the Senate, however, greater support for minor parties also means greater representation. At present, 11 of 76 senators are from independent or minor parties—including 9 from the Greens, which received around 10% of the nationwide (first preference) vote in 2007 and 2010. In July 2014, eighteen Senators from minor parties will take their seats based on the results of the 2013 elections. (A quirk of the Australia system is that whereas the new house members took their seats in November 2013, following the September election, senator-elects won’t take their seats until the July following an election).

Candidates per Seat

With this real opportunity for election, the number of senate candidates nominating has exploded over recent years (from 5 candidates per seat in 1984 to almost 10 in 2007). Ballot access requirements have always been low in Australia (especially by comparison to those in most states in the US) and to nominate for the Senate merely requires a petition signed by 100 registered voters and a US$1800 deposit (this deposit was controversial when introduced last year as it doubled the previous requirement). In 2013, a record 529 candidates ran for the 38 Senate seats across the country (14 candidates for each seat), with 110 candidates on the ballot paper in New South Wales.

While this made for much hilarity on Twitter and Instagram (see this collection of people posing with their “mammoth” ballot papers), under the system of compulsory preferencing, the explosion in candidates ensured that voting was laborious, there were high levels of spoiled ballots, and that a complicated vote counting profess left the result unknown for  several weeks. It also opened up a secondary market in preference dealing.


Compulsory Preferencing and Attempts to Make Preferencing Optional

Normally, STV offers voters the option of ranking several (or all) candidates. Some voters will rank many candidates, and others will rank some. In Australia, voters are required to rank ALL of the candidates—writing the numbers 1 – 110 in the case of the 2013 NSW ballot paper in order to cast a valid vote. This peculiar Australian adaptation of STV is known as “compulsory preferencing.”

Compulsory preferencing raises several concerns and problems. One concern relates to the quality of the votes cast. It is unlikely that many (perhaps any?) voters are (or should be) sufficiently informed to have well-ordered preferences for the candidates from the Bullet Train for Australia Party, the Stable Population Party, the Secular Party, Australia First, Shooters and Fishers, Senator On-Line, the Liberal Democrats, Australian Democrats, Democratic Labor Party AND the Christian Democratic Party (all “real” parties on the NSW 2013 ballot). Nonetheless, Australian voters must express ordered preferences on them for their Senate vote to be counted.

A secondary problem, which has since been solved, with compulsory preferencing was the high numbers of spoiled ballots caused by the full ranking requirement. When numbering a long list of candidates, voters were prone to make numbering errors when attempting to fill in their ballot (skipping or repeating some numbers), resulting in many unintentionally spoiled ballots. Consistently throughout the 1970s and early 1980s, almost 10% of Senate ballot papers were spoiled and did not count in the election, primarily due to numbering errors in long lists of candidates.

The logical remedy for both problems was to remove the requirement that voters express a preference for all candidates (known as “optional preferential voting” in Australia). Until now, this simple change has been extremely unpopular with the larger political parties in Australia, primarily for reasons related to the House of Representatives (It’s complicated, but Australia’s favourite psephologist, Antony Green, explains some of the reasons here).

When in 1983 a major series of reforms were made to the Electoral Act, suggestions of optional preferencing were sidelined in favor of a ticket system where the party you expressed a first preference for would decide the remaining ranking for you. This ticket voting system has itself caused more problems, which were on full display in the 2013 Election.

The ticket system was facilitated by the creation of “above the line” voting. In the system, each party lodged a ticket with the Australian Electoral Commission (known as a “Group Voting Ticket”) stating its rank order of preferences among all candidates on the ballot paper. On Election Day, a voter may merely vote “1” in the square of his most preferred party in the boxes “above the line” (there is literally a thick line dividing two sections of the ballot paper). If his preferences need to be distributed, his vote then flows in the order indicated on the party ticket for which he voted.

Voters are not obliged to endorse a party ticket and can still, if they wish, vote “below the line”, but if they do they must number every individual candidate from every party in their order of preference.

Typically, around 95% percent of Australians vote “above the line”. With up to 110 candidates on the ballot, the reasons a voter chooses to vote “above the line” are not difficult to discern: the extra time required to number every candidate, the difficulty in deciding whether to rank a candidate from the Sports Party, Senator On-Line or the Sex Party at 51, 52 or 53, and the possibility that any missed candidate or mix-up in the order, a scribble or illegible number here or there, may invalidate her vote.

Although ticket voting was hugely successful at reducing spoiled ballots—down to about 3 – 4 % by the late 1980s—the system has, perhaps predictably, created its own problems. The problems with the ticket voting system center on the secondary market it creates in “preference deals”.  With almost the entire electorate voting “above the line”, parties have almost complete control of the voters’ voting choices. Prior to an election most parties barter with one another for favorable preferences on each other’s tickets. Major parties have long cynically made preference deals with minor parties to try to minimize the chances that the other major party will gain a majority in the Senate.

The practice of preference dealing is not new but it is increasingly professional and controversial insofar as it involves very small or “micro” parties. Self-proclaimed “preference whisperer” Glenn Druery makes a living advising small parties on preference strategies and thirty small parties have joined together in the Minor Party Alliance to harvest preferences from each other.

Preference deals increasingly seem to reflect cynical calculations of election chances, with little connection to policy or ideological preferences in the rank ordering of parties. A particularly obvious example comes from the 2013 Australian election. The Sex Party—whose policies include support for voluntary euthanasia, legalized marijuana, reduced censorship of pornographic material, removal of restrictions on abortion and recognition of same-sex marriage—directed preferences to the One Nation Party over the ALP, Greens or Liberal Party – even though the One Nation Party is a right-wing nationalist party founded by Pauline Hanson, which opposes multiculturalism, non-discriminatory immigration policies and free trade. It seems highly unlikely that voters for the Sex Party support One Nation’s policies, or vice versa, ahead of the center-left or center-right policies of the two major parties.

The logic of these preference deals is that if enough small parties, especially ones with appealing names like the Australian Fishing and Lifestyle Party or the Lower Excise Fuel and Beer Party (a real party, running candidates at federal elections in 2001 and 2004), agree to direct preferences to each other ahead of the larger parties then it is possible that one of those parties may be elected (although which is elected is largely impossible to predict), even though each party only attracts a small fraction of the total first preference vote.

This is precisely what happened in 2013 with the election of unemployed 4WD enthusiast Ricky Muir—who pledged to buy a suit after his election—in Victoria from the Australian Motoring Enthusiasts Party. As we have seen, while the Australian Motoring Enthusiasts Party received only 0.5% of the first preference vote in Victoria (the overwhelming majority of which came from above the line), some 22 other parties directed preferences to them ahead of the major parties. Most of the 22 parties, for example the Voluntary Euthanasia Party, had platforms entirely unrelated to the plight of the motoring fanatic. Nevertheless, Mr. Muir will take his seat in the Australian Senate on July 1 and, presumably in a suit, be a key player in the Senate for the next six years.


The Reforms Proposed Today

Despite such outcomes, the general Australian population was not unhappy with the 2013 Senate results and the election of so called “micro-parties” like the Motoring Enthusiast Party or the Sports Party (see page 5 of this report by Essential Research). However, the major parties and those concerned citizens who made submissions to JSCEM were entirely unimpressed.

They proposed a broad swathe of reform measures, many of which were aimed at snuffing the chances of the election of micro-parties. For example, the Liberal Party, Australia’s major conservative party, suggested that “in order to discourage preference deals with distorted and concealed motives” there be “a requirement that before preferences from any party are distributed, that party must have received a primary vote of at least 10% of the value a quota” (Liberal Party submission, p.5). Described as “entirely impossible to justify”, this proposal would ensure that parties likely to received less than 1.4% of the first preference vote (10% of the current quota) would no longer be able to confidently trade preferences.

Fortunately that proposal was not taken seriously. Instead three reforms to the Senate voting system, recurringly suggested in slightly varying forms to JSCEM, were recommended unanimously (JSCEM is made up of members from the Liberal, Labor and Green parties).

Firstly, JSCEM endorsed the abolition of group ticket voting. Abolishing ticket voting would remove the ability of parties to control all of “above the line” voters’ preferences. Each party’s ticket would now only determine the preference order for its own candidates, not any other party’s.  It would therefore remove the incentive for micro-parties to run for office unless they can and will seriously attempt to attract votes in their own right. More importantly, the proposal seriously limits the ability of parties (whether micro-, minor- or major-) to engage in strategic and cynical preference deals with other parties.

Secondly, JSCEM recommended that Australia introduce optional preferential voting for the Senate. JSCEM suggested maintaining both “above the line” and “below the line” voting, with slightly different voting practices in each. An “above the line” vote would remain a party vote, with voters able to rank order as many or as few parties as they wish (so long as they rank at least one party). Voters may simply vote ‘1’ for their preferred party, and their vote would now be interpreted as ranking all candidates for that party, in the order determined by the party’s ticket. However, once all candidates are elected or excluded, the voter’s ballot would exhaust. This, in effect, creates a system of optional preferential party voting for “above the line” voters.

JSCEM recommended “partial” optional voting for “below the line” voters. Voters could, in this proposal, number as many or as few candidates as they wished “below the line”, subject to the qualification—without any logical reason—that each voter express a number of preferences equal to the number of Senators to be elected (6). Since numbering 1 to 6 seems much less arduous than 1 to 100, this measure should reduce instances of unintentionally spoiled ballots and increase the percentage of voters voting “below the line”. In any case, it will reduce the burden of voting for those intrepid 5% of voters who currently use it. Furthermore, the proposals remove the objectionable requirement that voters express preferences for candidates they know nothing—and care to know nothing—about.

Doubts have been raised about the effects of these plans, although they do not seem compelling. Some suggest that voters might be confused between the full, compulsory preferential requirements in the House and the new optional preferential environment of the Senate. Others worry that most voters will not express preferences beyond “1” (this is the case in Queensland and NSW, where optional preferential voting is already used for state elections*) and that this will reduce the proportionality of the Senate results.

The third recommendation of JSCEM was that the requirements to register as a political party be tripled, requiring parties to have 1500 paid-up members who are not members of any other political party. The aim of the proposal is to exclude micro-parties from the privileges of appearing “above the line”. Whether this proposal would have the impact of reducing competition is unclear. If micro-parties are motivated enough to organize 500 members, perhaps many will also be able to organize 1500 members.

Moving forward, a bill incorporating these changes will likely be generated by the Liberal Party in government in time to quickly pass both houses before the next election (due before 14 January 2017). With the support of both major parties (and the Greens), the bill’s success, in Australia’s strictly disciplined party system, is guaranteed. 



In light of the Senate 2013 Election, the time seems to have finally come for optional preferential voting. In no small part, the changed winds are due to changing assessments by the major parties of their interests. Whereas once, the major parties were “terrified” by optional preferential voting, now it looks to be the lesser to two evils. Party tickets and compulsory preferencing were the two means by which minor parties, like the Sex Party, the Motoring Enthusiasts’ Party and the Sports Party, gained their power. Of course, removing the power of the minor parties to win office necessarily bolsters the position of the major parties.

Despite the less than charitable motivations of the major parties, the abolition of party ticket voting and introduction of optional preferential voting has real advantages. STV, with optional preferencing, increases the options open to a voter. Voters can control their preferences: they can keep them assigned to the candidates of a single party by voting a “1” above the line. Or if there are other candidates from other parties they can vote below the line and rank them too.

Even though motivated by the self-interest of the major parties, many of the proposals before JSCEM have the potential to be a plus for democracy and to bring Australia closer to real “choice” voting.


* The reasons for the reticence of NSW and Queensland state election voters to preference more than one candidate are several and largely specific to Australia. In NSW and Queensland lower house elections, which operate under a single member form of ranked choice voting, there is no need for most voters to express preferences because, if they voted for an ALP or Liberal candidate, their first preference will—in most districts—not be excluded and their preferences will not need to be distributed. Another, important, reason voters do not list preferences is because their political leaders urge them not to do so. Cynical short-sighted campaign strategies by the major political parties encouraging voters to “Just Vote One” in the face of numerous minor candidates winning office on the back of preference flows worked. Part 3.1 of this report by Jenni Newton-Farrelly (of the South Australian Parliamentary Library) discusses these issues with much insight. Additionally, in both lower and upper house elections, compulsory voting ensures that some voters, who are at the polls against their will, wish to put as little effort into voting as possible and so vote only “1”. Interestingly, however, the Queensland government  is proposing that all local elections use optional preferential voting, which would be a change for a number of cities that currently use plurality voting.