Executive summary

• In 2016 the Senate voting system was changed to remove the use of group voting tickets; and to require voters to allocate six or more preferences above the line or twelve or more below the line on the ballot paper. The 2016 federal election—a double dissolution election—was the first to be conducted under the new system.
• The change resulted from recommendations from an inquiry by the Joint Standing Committee on Electoral Matters into the 2013 federal election, largely in response to the number of candidates being elected to the Senate from small and unknown parties on very low first preference votes. However, the changes were only legislated late in the parliamentary term, not long before the double dissolution election was held.
• A High Court challenge was launched almost immediately in response to the changes to the Senate voting system; however, the Court rapidly and comprehensively dismissed the case.
• When it was introduced, the new Senate voting system was criticised for a number of perceived problems, including that most voters would continue to vote 1 above the line; that the informality rate would be high; that many more votes would exhaust and not be counted; and that small parties would have no chance of election.
• None of these anticipated problems presented in the course of the 2016 election. Voters quickly adapted to the new system; informal voting rose only a small amount; and that the Australian Electoral Commission was able to implement the new system and count the votes with no major issues eventuating.
• The changes did not arrest the trend of fewer voters giving their first preferences to the major political parties in the Senate election. However, many of those who gave their first preference to minor parties also preferenced one of the major parties with their second and third preferences. While the 2016 Senate election resulted in a record crossbench, much of this is attributable to the reduced quota due to the double dissolution. Yet even under a normal half-Senate election, a number of minor party Senators would have been elected.
• Analysis shows that Australian voters’ use of above-the-line Senate preferences was complex; was not always along party lines; and, in many cases, was not consistent with the first preference party’s how-to-vote recommendations.
  

All hyperlinks in this paper were correct as at January 2018.

Introduction

In 2016, following a long and sometimes acrimonious debate in the Senate, the largest reform to Australia’s federal voting system in more than three decades became law. Despite the haste with which it was introduced and passed through the Parliament, the reform was a response to address flaws in the Senate voting system that became apparent over the previous decade and had been the subject of considerable deliberation.

This research paper outlines the recent history of Senate electoral reform in federal elections, including examining the reasons for the most recent changes to the Senate voting system. It describes how the new system works, and analyses the characteristics of its operation in the 2016 federal election before exploring how well it worked in practice. It concludes that Australian voters adapted well to the additional complexity of the new ballot paper; that the election results were generally more proportional in respect to the vote; and that—so far, at least—the reform appears to have been successful in terms of aims and implementation.

A brief history of Senate electoral system reform

The 2013 Senate election was notable both for the overturned result in Western Australia (WA) due to lost ballot papers, and for the number of previously unknown candidates from small parties who were elected on very small primary votes. Both of these issues received considerable attention from the Joint Standing Committee on Electoral Matters (JSCEM) in its inquiry into the 2013 federal election. In 2016, it was the latter issue that resulted in the most significant change to Australian federal elections in more than 30 years.

The single transferable vote system

Since 1949, the Senate has used an electoral system known as proportional representation through single transferable vote (PR-STV, or simply STV). STV is a family of electoral systems that are both proportional and preferential (although some scholars have argued that STV is at best semi-proportional).[1] In addition to the Senate electoral system, the Hare-Clark electoral systems used in Tasmania and the Australian Capital Territory (ACT) are also STV electoral systems, as are most state upper house electoral systems. Internationally, elections in Ireland and Malta also use STV, but it is uncommon outside Australia.[2]

STV is proportional in that it attempts to ensure that members elected into the multi-member districts are roughly in proportion to the share of the vote they received. That is, if 30 per cent of people vote for Party A and 40 per cent vote for Party B, Party A should win roughly 30 per cent of the seats and Party B should win roughly 40 per cent. In practice there is a limit to the proportionality—for example, ten parties each winning ten per cent of the vote could not be accommodated perfectly proportionally in a district with only six seats.[3]

STV is also preferential in that voters can (or must, in some circumstances) allocate multiple preferences to candidates. It is similar in this respect to the preferential voting used in House of Representatives elections. This feature distinguishes it from most proportional representation (PR) systems in use throughout the world, which are not preferential. In most of these other systems (often referred to collectively as PR list systems), voters only express a single preference for a group of candidates. Candidates are elected from the group in proportion to the vote that the group receives, with a number of different formulas used in translating the vote share to seats.

Introduction of group voting tickets in 1983

The voting system in Australian federal elections in the House of Representatives and the Senate is defined in the Commonwealth Electoral Act 1918 (CEA). Similarly to preferential voting used in House of Representatives elections, the Senate voting system in place between 1949 and 1983 required voters to allocate preferences to every candidate on the ballot paper in order of their preference for the vote to be counted (or ‘formal’).[4]

While this approach worked well when there were few candidates on the ballot paper, by 1983 the informality rate for Senate ballot papers had reached 9.9 per cent (compared to 2.1 per cent informality in the House of Representatives in 1983), mostly due to numbering errors by voters.[5] This meant that almost one in ten ballot papers could not be counted.

The wide-ranging 1983 inquiry of the Joint Select Committee on Electoral Reform (the forerunner to the JSCEM) considered the issue of informal voting on the Senate ballot paper and commissioned an analysis of informal votes from the 1977 and 1983 Senate elections. The analysis indicated that over 75 per cent of informal votes were invalid due to unintentional error, even though the voter’s intent was clear.[6]

The Committee considered several submissions before settling on a combination of recommendations from the then Australian Electoral Office (the precursor to the Australian Electoral Commission) for a ‘list’ system and from the Liberal Party to continue using the existing full preferencing (as opposed to the Australian Labor Party’s recommendation to move to optional preferential voting). The Committee stated:

The ‘list’ system recommended for adoption at Senate elections is designed to simplify the voting process for those electors who are content to record on their ballot paper the preference ordering recommended by a particular party or candidate.

Voters would have the option of recording preferences for all candidates in the normal way, or of placing a tick or cross in a special square on the ballot paper to indicate adoption of a particular party or group list. In the latter case, the vote would be treated as if the elector had numbered every square in accordance with that list.[7]

This ‘tick or cross in a special square’ became the above-the-line voting system.

Due to the negotiations required to pass the changes to the electoral system through the Senate, the legislation that enabled above-the-line voting[8] allowed parties to submit up to three group voting tickets (GVTs).[9] If the party or group submitted more than one GVT, the above-the-line votes for the party or group were divided evenly between the GVTs.

Michael Maley, a former AEC official who was involved in the 1983 changes, describes how the use of GVTs evolved in unforeseen ways, culminating in the 2013 results:

Because it was apparent from previous election statistics that the vast bulk of voters had been following how-to-vote cards anyway, the change was not seen as being a particularly momentous one. But one important point was overlooked, and this was the issue which came to a head in 2013. Up until 1984, the only parties which were able to issue how-to-vote cards were those which had the membership base, field structure and resources to enable them to distribute cards physically at polling places. With ticket voting, on the other hand, every group of candidates on the ballot paper could provide voters with, in effect, a ‘virtual’ how-to-vote card. This ultimately led directly to the phenomenon of ‘preference harvesting’, which enables a host of ‘micro-parties’ to exchange preferences with each other for their mutual benefit.[10]

The vote count in Senate elections and the 2016 reforms

In STV elections a candidate must receive a ‘quota’ of votes, either as first preference votes or surplus votes transferred as preferences from another candidate. The Senate voting system uses a Droop quota (named after English lawyer and mathematician Henry Droop),[11] which is defined as:

There are twelve senators in total in the Senate for each state. As such, in a normal half-Senate election for a state, there are six vacancies and the candidate must receive one seventh of the total vote to be elected. In a double dissolution election, as in the 2016 federal election, there are 12 vacancies and the quota is 1/13 of the vote.[12]

Typically in Senate elections the larger parties receive one or more quotas of first preference votes and any surplus votes flow down the ticket to candidates in the same party or group.[13] The major party candidates may receive very few votes below the line; when they are elected, it is generally from above-the-line votes for their party that have transferred as surplus votes from another candidate from the same party higher up the ticket.

In each election, after all the candidates who have full quotas are declared elected, there are a number of candidates who build up a quota mainly from transferred votes from other groups excluded from the count. Intuitively, those candidates with higher first preference votes might be more likely to be elected once preferences are distributed. Through the use of preference trading, however, it is possible for preferences to be pooled between candidates with very small votes in order to achieve a quota. In some cases, compulsory full preferencing means that when a large party has elected as many candidates as it can, and any of its remaining candidates are all eliminated from the count, those preferences can boost a small party enough to stay in the count and achieve a quota.

In the STV system, the candidate with the fewest votes in each count is excluded. The challenge for any small party is to receive enough transferred preferences each count to stay above the bottom position and remain in the count. If a small party’s top candidate can accumulate some transferred votes each time another candidate is excluded—and stay in the count long enough to be the only candidate remaining when the last candidate for a major party is excluded—they have a reasonable chance of being elected into one of the last vacancies.

Disproportionate Senate election results

The 2013 federal election was not the first occasion that the coordinated direction of preferences led to an unforeseen result where the results seemed incongruous with the first preference votes. The first evidence that coordinated GVT preferencing could lead to an unusual result was in 2004, when Victorian Steve Fielding of Family First was elected to the Senate on a first preference vote of 0.13 quotas, or 1.85 per cent of the total vote.[14] Fielding’s win over a Greens candidate was described by then Greens Senator Bob Brown as a ‘perverse result’ and ‘outrageous’.[15] The unsuccessful Greens candidate called for Senate electoral system reform to allow voters to include multiple preferences above the line.[16]

At the time, election analyst Antony Green, writing in the Sydney Morning Herald, argued that the system allowed party bosses to engineer electoral outcomes:

the results of the election have revealed that the Senate’s voting system, rather than allowing for the expression of the will of the electorate, has fallen under the control of party ‘bosses’ engaging in complex preference deals designed to engineer electoral outcomes.

The problem is ‘above the line’ or group ticket voting, introduced in 1984 to overcome the huge informal vote that had dogged Senate elections. It also offered political parties a wonderful opportunity to control party preferences, as was shown when Labor and the Coalition saw common purpose in ensuring Peter Garrett did not win election for the Nuclear Disarmament Party.[17]

In December 2004, Senator Bob Brown introduced the Voter’s Choice (Preference Allocation) Bill 2004, which would have required voters to number all squares above the line, consistent with the full compulsory preferencing on the House of Representatives ballot paper. Introducing the Bill (which was not subsequently passed), Senator Brown cited the ‘perverse’ outcome in the 2004 election.[18]

Further examples of unforeseen results occurred in the 2010 Senate election, when Victorian John Madigan from the Democratic Labor Party was elected to the Senate on 0.16 quotas, or 2.33 per cent of the total vote.[19]

The 2013 federal election saw two Senate candidates elected on very small primary votes. Ricky Muir from the Australian Motoring Enthusiast Party was elected in Victoria on 0.51 per cent of the primary vote (0.035 quotas), and Wayne Dropulich from the Australian Sports Party was elected in WA on 0.23 per cent of the primary vote (0.016 quotas).[20] The election results were unprecedented in a number of ways, including:

In five out of the six States, a candidate was elected from a party which had never previously been represented in the federal Parliament.

For the first time ever, the seats in one State, South Australia, were divided between five different parties.[21]

The inquiry into the 2013 Senate election

Following each federal election, the JSCEM conducts an inquiry into the conduct of that election. The JSCEM received a reference from the Special Minister of State to undertake an inquiry into the 2013 election, taking into account the following specific issues:

• Senate voting reform
• voter ID (proof of identity) at polling place
• the electoral roll (including the impact of direct enrolments) and public access to the roll
• circumstances surrounding the lost WA Senate votes and
• the feasibility of, and options for, electronic voting.[22]

The JSCEM released its interim report on Senate voting practices in May 2014. In the foreword to the report, the JSCEM Chair, Tony Smith, stated:

Combined with pliable and porous party registration rules, the system of voting for a single party above the line and delegating the distribution of preferences to that party, delivered, in some cases, outcomes that distorted the will of the voter.

The system of voting above the line has encouraged the creation of micro parties in order to funnel preferences to each other, from voters who have no practical way of knowing where their vote will ultimately land once they had forfeited it to the parties’ group voting tickets.[23]

The Chair emphasised that the above-the-line voting system confused voters, and that many voters would not have known where their preferences were going when they voted above the line. While the GVT provided by the parties are publically available, it is unlikely many voters study them in depth due to their complexity.

The Committee made six recommendations in the report: two involved replacing group voting tickets with optional preferential voting above and below the line; one that the AEC should be adequately resourced to undertake a voter education campaign in relation to the change; and three that related to party and candidate registration.[24] The Chair noted that:

This report has been produced at this time to not only provide the Parliament with the time to legislate change, but to enable thorough and adequate information, education and explanation of the improvements to the voting public well in advance of the next election.[25]

The bipartisan Committee unanimously supported the recommendations. The Deputy Chair of the Committee, ALP member Alan Griffin, stated that he believed that the GVT system was not operating as had been intended.[26] The then Shadow Special Minister of State Gary Gray stated that the ‘changes will make the Senate voting system more transparent and will mean a full translation of voter intention to the electoral outcome’.[27]

Legislating the changes

Little more was heard about the matter after the report was released, other than a statement by then newly-appointed Special Minister of State Mal Brough in September 2015 that the Government intended to pursue Senate voting reform.[28] This was interpreted as an attack on the Senate crossbenchers (whose votes in the Senate were necessary to pass the Government’s legislative agenda), and the topic rapidly vanished from public discussion.[29]

The issue again surfaced in February 2016, with media reports that the Government was engaged in discussions with the ALP and the Greens to legislate the Senate reforms before 17 March 2016 in order to pave the way for an early election.[30] Other reports suggested that the Greens would introduce their own legislation if the Government did not, but that the ALP was ‘divided’ as to whether to support optional preferential voting.[31]

The legislation to reform the Senate voting system, the Commonwealth Electoral Amendment Bill 2016,[32] was introduced into the House of Representatives on 22 February 2016 with essentially no official prior notice by the Government. It was amended by the Government almost immediately to allow Senate first preference above-the-line votes to be counted in the polling place. (In its initial form, no Senate results would have been reported until the votes had reached the central scrutiny point, several days after the election.)[33]

On 22 February 2016—the day the Bill was introduced to the lower house—the Government referred the Bill to the JSCEM for consideration, to report no later than 2 March 2016.[34]

The JSCEM took submissions and held a public hearing on 1 March 2016. In its initial form, the Bill implemented the recommendations of the JSCEM interim report to introduce optional preferential voting above the line. The Bill went beyond the Committee’s recommendations in instructing voters to allocate at least six preferences above the line; although if there was at least one first preference above the line, the vote would be formal. Contrary to the Committee’s recommendations in its interim report, voters were still required to allocate preferences to all candidates below the line; however, the allowance for errors in below-the-line preferencing was increased slightly.[35]

The last-minute Committee inquiry resulted in an advisory report, which recommended that the Government also allow optional (rather than compulsory) preferential voting below the line; with a minimum of 12 below-the-line preferences and a savings provision that the ballot paper would be formal if it contained at least six preferences.[36]

The JSCEM advisory report also contained a dissenting report from ALP senators and members, arguing that the Bill was ‘driven not by the democratic interests of the Australian people, but, rather, the political self-interest of the Liberal Party and the Greens’.[37] The dissenting report quoted evidence from Antony Green suggesting that under the Bill the Coalition would win 38 seats at a double dissolution election, which would give them a blocking majority in the Senate.[38]

On 2 March 2016 the Bill was introduced into the Senate. The second reading debate involved speeches from 39 senators and stretched over four days. The JSCEM’s recommended amendments were agreed to, and the Bill was finally passed following the longest continuous sitting without breaks of any kind—lasting 28 hours and 26 minutes—on 18 March.[39] The Bill passed the House of Representatives later that day and received Royal Assent on 21 March 2016.

The High Court challenge

No sooner had the Senate voting changes passed the Parliament than one of the crossbench senators, Bob Day of Family First, lodged a challenge in the High Court against the changes, supported by Liberal Democrats senator David Leyonhjelm. Day claimed that the new voting system ‘will make it easier to get rid of pesky independents like me’.[40] Leyonhjelm set up a crowdfunding campaign to raise $20,000 to pay for the challenge.[41] While Labor opposed the changes, it declined to join the High Court challenge.[42]

In a hearing on 24 March 2016, the Chief Justice of the High Court, sitting alone, ordered that the matter be listed on 15 April 2016 for directions as to referral to the Full Court, and on 15 April the case was referred to the Full Court for a hearing on 2 May 2016.[43] The Court held hearings on 2 and 3 May 2016 and delivered its judgement on 13 May 2016.[44]

As identified by the Court, Day had essentially five arguments that the new voting system was unconstitutional:

• The above- and below-the-line voting process constituted different ‘methods of choosing senators’, forbidden under section 9 of the Australian Constitution.
• Above-the-line voting did not represent senators being ‘directly chosen by the people’, required under section 7 of the Constitution.
• The Droop quota used to calculate which candidates had been elected meant that the votes of one seventh of the electorate did not count towards the election of a successful candidate.
• Votes that exhaust effectively disenfranchise voters.[45]
• In the words of the High Court Justices: ‘a kind of “catch all” proposition repeating the complaint in the previous argument and the complaint of effective disenfranchisement’.[46]

Writing before the judgement was handed down, constitutional law expert Professor Anne Twomey wrote of the merits of the case:

The grounds for this challenge are not strong. For the most part, the arguments being made would apply with equal or greater force to the pre-2016 law, the validity of which has consistently been upheld, albeit by single judge decisions. The 2016 Act is substantively different because it permits optional preferential voting, allowing votes to exhaust. While there is a possibility that this might be regarded as contrary to the irreversible evolution of representative government, the 2016 Act does not mandate that votes become exhausted—they merely give voters a choice as to how their vote can best reflect their preferences. It would therefore be difficult to argue that a Senate elected pursuant to the 2016 Act was not ‘directly chosen by the people’.[47]

The High Court unanimously upheld the changes to the Senate voting system, throwing out Day’s case. The brief judgement rejected all five arguments made in Day’s application, and did not throw a significant amount of new light on the constitutional restrictions on the electoral system. However, the ruling did clear the way for the Government to hold the election under the new system. According to the judgement summary:

The High Court unanimously dismissed both applications. The High Court held that the term ‘method’ in s 9 of the Constitution is to be construed broadly, allowing for more than one way of indicating choice within a single uniform electoral system. The High Court further held that a vote above the line was a direct vote for individual candidates consistent with s 7 of the Constitution. Finally, there was no disenfranchisement in the legal effect of the voting process and there was no infringement of the implied freedom of political communication or the system of representative government.[48]

The Court further ruled that the plaintiff (Senator Day) pay costs. 

How the new Senate voting system works

Political scientists define electoral systems in terms of three variables: the electoral formula (how the votes are counted), the ballot structure and the district magnitude (the number of people to be elected).[49] Compared to the Senate electoral system as it operated from 1984 to 2014 (the GTV system),[50] the new Senate system changed only the ballot structure—leaving the electoral formula and the district magnitude untouched. That is, the new system maintains STV with a Droop quota and transfer values calculated using the inclusive Gregory method,[51] and the number of senators elected at each election remains unchanged.

The ballot paper looks superficially similar to that of ballot papers used between 1984 and 2014, with a series of columns representing groups across the ballot paper and the candidates in each group listed in their column. The ballot paper is divided by a horizontal line, and each group that qualifies receives a box above the line (see Figure 1). Voters can either vote above the line for parties or groups, or vote below the line for candidates. Voters who vote above the line are instructed to complete at least six preferences, and voters who vote below the line are instructed to complete at least 12 preferences (below-the-line preferences are counted as they are entered by the voter). Savings provisions allow votes that have at least one preference above or six preferences below the line to be counted.

Figure 1: Senate ballot paper structure

Note: Groups may be candidates all from one party, or from two parties that have elected to run a ‘joint’ ticket and have their candidates listed in the same column, or as non-affiliated candidates who have chosen to run together in the one column in order to have a square (but not a group name) above the line.

Votes above the line are essentially a shortcut for a below-the-line vote, and are converted into a below-the-line preference order before the vote is counted. In the above example, an above-the-line vote for Group D is counted as a below-the-line vote for all of the candidates in Group D in the order in which they are listed on the ballot paper, top to bottom.

If Group D gets a 1 above the line, and it has four candidates in the group, the first candidate in the group will get a 1 vote, the second a 2 vote, through to the fourth candidate who gets a 4. If the voter then gives their second above-the-line preference to Group F (which has four candidates), the first candidate in Group F will get the next preference in sequence (5), and the next two candidates in Group B will get 6 and 7 respectively—and so on, until all above-the-line preferences are allocated to equivalent below-the-line candidates (see Figure 2).

Figure 2: How above the line preferences are counted on the Senate ballot paper

If the voter skips or repeats any numbers (either above or below the line), the repeated numbers and any higher numbers are ignored. For example if the voter has listed the preferences 1, 2, 3, 4, 4, 5, 6 below the line only, preferences 1 through 3 will be counted. This vote would therefore be informal and not counted, as it has fewer than six below the line preferences.

If the voter allocates preferences both above and below the line, only the below-the-line vote will be counted. However, if the vote below the line is informal, the above-the-line vote will be used. In addition, any ticks and crosses on the ballot paper are counted as a 1.

A high-level explanation of how the vote is counted is that a quota is determined (as discussed above), and the candidates with more than one quota of votes are declared elected. The surplus votes of the elected candidates (their total vote minus a quota) are then transferred to the voter’s next preference. Following the transfer, if any candidates are above quota, they are declared elected and their surplus is distributed to the next preference. If none are over quota, the candidate with the lowest number of first preference votes is excluded and their votes are transferred to the voter’s next preference. The process continues with candidates over one quota being elected and their votes being transferred—or candidates with the lowest vote being excluded and their votes transferred—until either all of the vacancies are filled or there are no more votes to transfer. If the latter happens and there are unfilled vacancies, the candidates with the highest remaining votes are declared elected into any remaining vacancies.

How well did it work at the 2016 election?

Following the federal election on 2 July 2016, the AEC declared the first Senate election result (for the Northern Territory) on 25 July 2016, and the last (for NSW) on 4 August 2016—meaning that despite the new system the AEC completed the count before the deadline of the final day for the return of the writs on 8 August 2016.[52] There were no reports of problems with the Senate count or results, although one former NSW Electoral Commission official suggested that the AEC was being less than transparent about its auditing and certification of the vote scanning process.[53]

Once the AEC began to publish the Senate vote data,[54] a number of analysts started examining the criticisms that had been levelled against the new voting system before the election, including that it would eliminate minor parties; that most voters would just vote 1; and that there would be high vote exhaustion rates and higher informality. These criticisms are considered in detail below.

Above-the-line voting

As in all elections since the introduction of group voting tickets in 1984, the majority of Australian voters chose to vote above the line for the Senate in 2016. While the number of voters numbering preferences above the line was noticeably lower than in recent elections (see Figure 3), more than nine in ten voters numbered preferences above the line.

Figure 3: Above-the-line voting at Senate elections, 1990–2016

Source: Compiled by the Parliamentary Library from AEC data.

Rates of above-the-line voting were slightly lower in 2016 than in 2013 for most states and territories (Table 1). The below-the-line voting rates in both Tasmania and the ACT were substantially higher than those for the other states and territories; however, in 2013 both Tasmania and the ACT also had the highest below-the-line voting in the country. Both of these jurisdictions use Hare-Clark for state or territory elections—a voting experience that is similar to voting below-the-line in the Senate, which likely explains their traditionally high below-the-line voting rate.[55] 

Table 1: Use of above- and below-the-line voting by state

Source: Compiled by the Parliamentary Library from AEC data.

In contrast to the ACT—where below-the-line voting remained high in 2016 but dropped slightly since 2013—the rate of below-the-line voting in Tasmania more than doubled in 2016 compared to 2013, with more than one in four Tasmanian voters numbering preferences below the line. This increase is likely due to the Senate campaign of Labor Senate candidate Lisa Singh, which encouraged voters to vote below the line for Singh.[56]

Use of preferences

At the JSCEM public hearing on the Bill for the new Senate voting system, Labor Senator Stephen Conroy stated ‘I have read figures of 80 to 90 per cent casting a 1 and have no preferences after that’.[57] Under the new system, a ballot paper with just a first preference above the line would still be formal. The data show that this outcome did not eventuate at the 2016 election, however, with only 3.2 per cent of above-the-line voters writing only 1 above the line (Table 2). The highest rate was in NSW, where 5 per cent of those who voted above the line only voted 1. Given that the NSW Legislative Council ballot paper allows a single preference above the line, some voter confusion between the requirements of the respective ballot papers is to be expected.

Table 2: Use of preferences above the line as a proportion of votes above the line

Source: Compiled by the Parliamentary Library from AEC data.

Note: There are small differences between the figures in this table and similar analysis published by the AEC, likely due to the AEC’s use of its own internal vote data for its analysis compared to the published vote data used for this table. The figures above are consistent with other third-party analysis of the AEC’s data.[58]

The vast majority of voters—more than four out of five above-the-line voters—followed the ballot paper instructions and allocated six preferences. The one exception was the Northern Territory where there were only seven columns above the line; here 55.5 per cent of above-the-line voters used six preferences, and an additional 38.7 per cent allocated all seven preferences.

Vote exhaustion

Some parties expressed concerns that the reforms would lead to high levels of vote exhaustion (where votes indicated no preference for one of the candidates remaining in the count and therefore would not count against a final candidate) due to an expected high numbers of 1-only above-the-line votes. These concerns did not materialise at the 2016 election (see Table 2).

For example, in the Liberal Democrats submission to the JSCEM inquiry into the Bill, the party stated:

Inevitably, a significant number of votes cast—especially those cast for parties other than the three major parties—will simply exhaust instead of being allocated according to preferences. Where those votes would previously have helped to elect minor party candidates, they will simply not count toward any vote tally once the six candidates that the voter preferred are eliminated from the contest, ie the votes will be discarded. As an elector, the idea that your vote will be discounted if you do not vote for the government’s preferred candidates should be horrifying.[59]

At least nine of the speeches against the Bill in the Senate—from Labor, Family First and Senator Lambie—also expressed concern about vote exhaustion, including a concern that three million people’s votes would exhaust as a result of the changes (or, as Family First Senator Bob Day termed it, ‘the potential disenfranchising of three million votes’).[60] For example, Labor Senator O’Neill stated:

What they did in that last election was—and there are three million Australians who made this decision with their vote—vote for candidates other than from these larger parties which dominate the political landscape of the country. It looks like the legislation will get through because of this dirty deal. Three million people voted at the last election for candidates other than those from these three parties. Their votes are going to ‘exhaust’. They are ‘exhausted’ simply means they are not counted. They are going to end up not being involved in the election of anyone. Twenty-five per cent of the voting public will end up with no representation of their view of who should be here, and that hardly seems fair.[61]

While all votes—including exhausted votes—are counted to determine the election results, exhausted votes do not count against a successful candidate. This is effectively the same as in preferential voting in the House of Representatives elections, where a vote that preferences the unsuccessful of the final two candidates does not count against the successful candidate’s total vote.

For example, as stated in an AEC report on vote exhaustion following the 2016 federal election:

[It is easy to think of] an exhausted ballot paper as lost or wasted, or to think of exhausted votes as being less effective than fully preferenced ballot papers. There is an argument that the more preferences there are on a ballot paper the more ‘effective’ or ‘powerful’ it is.

However this is a subjective argument. An exhausted vote is by definition formal, and has expressed the electors stated preferences. An elector may reach the conclusion that if the candidates they have numbered are not elected, they do not wish their vote to assist in the election of any other candidates. This is, potentially, as valuable to an elector as their stated preferences.[62]

The High Court action against the new system by then Senator Bob Day, discussed above, also cited the projected exhaustion rate as a problem:

Upon that new approach, the value of the votes of those who follow the instruction is diminished, by reference to the number of parties that are not preferenced [eg 29 in South Australia in the last general election]. That is because in the scrutiny their ballots will be treated as exhausted or not ‘available’ for the transfer of their vote from candidate six in the list [who is eliminated] to the next candidate in the list [ie past six]. The result of that is that such electors record no vote at all, as their vote is denied all real value, and not treated as equivalent to those who did vote either for all candidates by a full preference vote [very few ] or those who record first preference votes in sufficient numbers to elect their first, second or third preferences within the one party list.

The Commonwealth cannot justify that regime: because if voting remains compulsory as it does there is no sound justification for diminishing the value of a vote from some of those votes as against others. Upon the record of the last election the number of persons the value of whose votes is diminished in this way is 25% of the State ‘voting as one electorate’. As pleaded in paragraph 9 of the Grounds this is, nationally, 3,314,174 votes across Senate elections in 2013 as against a total formal vote of 13,380,545 votes.[63]

Election analyst Antony Green has argued that the raw number of exhausted votes is not a good indicator of the issue. In some counts, the AEC excluded the final non-elected candidate in order to determine an order of election for the successful candidates, thus inflating the exhaustion rate. Upon re-examining the ballot papers to assess the exhausted preferences of parties that did not elect senators, Green found an exhaustion rate of around 800,000 votes (around 5.6 per cent).[64] Psephologist Kevin Bonham has agreed with this approach, citing an exhaustion rate of 5.1 per cent ‘at the point at which the winners in the various states were determined’.[65]

Rather than being wasted or discarded, 83.9 per cent of above-the-line votes for minor parties were directed towards a successful party or to the final non-elected party.[66] In addition, up to 70 per cent of voters for the ten largest minor parties by votes marked the Coalition, Labor or the Greens as a second or third preference (see Appendix A). One possible explanation for this may be that voters are aware, at least at some level, of the potential for vote exhaustion, and are voting such that if their vote cannot be counted for a minor party candidate, their preference will count towards a party with a better chance of election and will not be wasted.

A slightly different approach is taken by the AEC in its analysis of Senate vote exhaustion following the 2016 federal election, where a distinction is drawn between the number of ballot papers and the number of votes. The number of votes can be different to the number of ballot papers due to surplus votes being transferred at reduced values. For example, if the quota for election is 80 votes and a candidate receives 100 votes, 20 votes will need to be transferred to the next preference. Rather than choose which of the 80 ballots to keep and which 20 to transfer, all 100 ballots are transferred at a transfer value of 0.2 of a vote each for 20 total votes. If 10 of those ballot papers cannot be transferred to another candidate, they will exhaust—but as they do so at a transfer value of 0.2, only two votes will exhaust. Importantly, because these ballots are transferred after electing a candidate, they still count towards the election of that candidate, even though the votes subsequently exhaust.

The AEC analysis concluded that, using its approach, the national exhaustion rate was 7.5 per cent. Of the exhausted ballots, 86.5 per cent helped elect at least one candidate in the 2016 Senate election, and 56.7 per cent helped elect four or more candidates. The AEC report also noted that vote exhaustion was strongly associated with the number of candidates on the ballot paper, with a higher number of candidates leading to higher levels of exhaustion.[67]

As noted above by the AEC, it is also worth considering that, in some cases, voters may prefer that their vote exhaust rather than see the vote elect candidates with whom they strongly disagree. In this regard at least, vote exhaustion in optional preferential voting is not a ‘bug’ but a ‘feature’.

Informal votes

Prior to the election, it was reported that internal Labor analysis predicted an additional 800,000 informal votes as a result of the new Senate voting system, given the AEC would not have sufficient time to educate voters about the change. The analysis reportedly predicted that the informality rate in the Senate would increase to 10 per cent.[68] The savings provisions built into the change—where votes with at least one preference above the line, or at least six below the line, would be formal—should have prevented this outcome. As the original report on which the article is based is not available, it is not known how the Labor analysis reached this conclusion.

After the voting had begun there were reports of polling officials giving voters incorrect advice—such as having to number ‘up to’ six boxes above the line, or not being able to number more than six boxes above the line. The AEC stated that it was the voter’s responsibility to read and follow the ballot paper instructions, but admitted that—given the scale of the election—it was possible that some issuing officers might have given the wrong advice.[69] Due to the savings provisions, however, such instructions should not have led voters to cast an informal ballot as long as they recorded at least one preference above the line.

AEC figures on Senate ballot paper informality indicate that the number of informal Senate ballot papers did increase slightly on the 2013 election, but by less than 1 percentage point nationally. The final Senate informality rate was just under 4 per cent, or around half-a-million informal ballots (see Table 3). NSW had the highest informality rate at 4.53 per cent.

Table 3: Senate ballot paper informality in 2016 compared to 2013, percentage of total vote

Source: AEC[70]

The AEC also undertook an analysis of types of informal Senate ballot papers, reproduced in Table 4 below. As the AEC does not routinely analyse the types of informal Senate ballot papers, it is not possible to compare the results to previous elections. However, the digitisation of the Senate ballot papers for the 2016 Senate election enabled the AEC to produce the analysis for 2016.

The most common form of informal ballot paper had no identifiable preferences—a category which would include blank ballot papers. These constituted 64 per cent of all informal ballot papers. Unfortunately it is not possible to tell what the voters intended by leaving their ballot papers blank, but they may have included not wanting to vote at all, but wanting to avoid a fine, or not wanting to give any of the candidates on the ballot a preference.

The next most common errors were repeated numbers either above or below the line. In some cases this might be due to ticks and crosses on the ballot paper, which are counted as a 1. (A ballot paper with some candidates ticked and some crossed would be treated as being informal due to having multiple 1s.) Repeated and skipped numbers were common on informal below-the-line ballots—informal only if they had not otherwise had six consecutive preferences starting at 1 below the line. (For example, a below-the-line preference sequence of 1, 2, 3, 4, 5, 6, 7, 7, 9 would have been formal, whereas a sequence of 1, 2, 3, 5, 6, 7 would not.)

Prior to the 2016 amendments to the CEA, skipped and repeated numbers below the line would not have made the ballot paper informal. It is unlikely that many voters knew the details of the complicated pre-2016 ballot paper formality rules, and so it is likely plausible that these were simply errors. Expanding the savings provisions to allow for skipped and repeated numbers, as per the pre-2016 CEA, would have potentially allowed an additional 58,267 (or 10 per cent of the informal ballot papers) to be counted.

Table 4: Types of Senate ballot paper informality in 2016

Source: AEC[71]

These numbers indicate that there was reasonably little misunderstanding of the new Senate ballot paper among voters. The vast majority of the informal ballot papers under the new system would also have been informal under the old system (no preferences, no above-the-line 1, repeated above-the-line 1, and insufficient preferences below-the-line). It is therefore unlikely that the errors were purely a result of the changes.

Senate composition and smaller parties

Much of the criticism of the Senate voting reform Bill related to the expected composition of the new Senate. The main arguments were that it would eliminate the non-major party crossbench from the Senate, and that it would entrench Coalition control of the Senate.

In their dissenting report to the JSCEM’s report into the Bill, ALP members of the committee stated:

This Bill will have the effect of maximising the number of senators elected representing major parties ... This will deprive independents and so-called ‘micro parties’ of votes and prevent new entrants from achieving election to the Senate, thereby entrenching the dominance of existing parties.[72]

Several analysts concluded that the Senate reforms would result in the Coalition having a majority in the Senate after the election. One analysis predicted that the Coalition would win 40 seats and Labor would win only 25 in a double dissolution election under the new voting system.[73] As it eventuated, the Coalition won 30 seats and Labor 26 in the 2016 election.

One of the most politically significant criticisms raised against the new system when it was proposed was that it would eliminate the election of senators from outside the major political parties to the Senate. In testimony at the JSCEM hearing into the Bill, political consultant Glenn Druery stated that it was unlikely that any minor parties would survive or ‘be born’ under the new system.[74]

Clearly, the new Senate voting system did not eliminate minor parties from the Senate. Six non-major parties won, between them, 11 seats in the 76 seat Senate, having received 13.6 per cent of the primary vote (see Table 5).

Table 5: Percentage of first preference Senate vote by party/group

Source: Compiled by the Parliamentary Library from AEC data.

The level of success of small parties in 2016 can be attributed in large part to the double dissolution election, which reduced the quota for election from one seventh of the vote to one thirteenth.

There have been no public counts of the 2016 election using a normal half-Senate election quota, so it is not clear how well the small parties would have fared in a half-Senate election. However, the CEA requires the AEC to undertake a recount under section 282 to provide the Senate with a list of potential short- and long-term senators following a double dissolution election.[75] This recount only includes those candidates who were elected at the full Senate count, but it applies the half-Senate quota. This recount provides some indication of how many senators elected at the double dissolution election would also have enjoyed the same result at a half-Senate election (see Table 6).

Table 6: Senators elected by party under the statutory recount

Source: AEC[76]

The section 282 recount suggests that, under a normal half-Senate election, Derryn Hinch’s Justice Party, Family First and the Liberal Democrats would not have gained Senate seats; and that the Greens, One Nation and the Nick Xenophon Team would have gained fewer seats. However, the section 282 results for One Nation, the Nick Xenophon Team and the Jacqui Lambie Network—whereby they still have senators under the new system—suggest that the reform is not inherently hostile to small parties; and that if small parties have significant public support, they can gain a presence in the Senate.

After the election, the Senate reforms were again blamed for the composition of the Senate—this time for enabling the election of four One Nation senators. Opposition Leader Bill Shorten told the media:

Pauline Hanson’s One Nation party appears to have four Senators. They have four Senators because the Greens political party and Malcolm Turnbull changed the rules in the Senate and then Malcolm Turnbull called a double dissolution. The presence in such numbers of One Nation in the Senate is a direct result of Mr Turnbull and Mr [sic] Di Natale’s actions in terms of their so-called electoral reform.[77]

Kevin Bonham considered the effects that the new system might have had on the success of One Nation, concluding that it was more the double dissolution factor that accounted for One Nation’s success:

By any fair measure, One Nation’s primary vote share and strong preferencing performance compared to other micro parties entitles it to the seats it won. Anyone who disagrees with that has a problem not with the new Senate system, but with preferential and proportional democracy itself. All the same, the fact that we have four new One Nation Senators and not one (given the votes actually cast) is down not to Senate reform but to the government's strategic decision to force a double-dissolution ...

Without knowing what deals might have been done we cannot know for sure how many seats One Nation would have won under the old system—it might even be in theory that at a half-Senate election Hanson herself could have been beaten by some similar right-wing micro-party on a preference snowball. It’s very likely though that PHON [Pauline Hanson’s One Nation] would have won at least one [seat] at a half-Senate election with a good chance of more, and that they would have won multiple seats at a double dissolution. Depending on the deals done, their seat haul at a double dissolution under the old system could have even been as high as seven.[78]

In general, the 2016 Senate election (albeit with a reduced quota) demonstrated that small parties with reasonable first preference votes and strong preference flows can gain seats in the Senate—even in the absence of group voting tickets, and particularly in smaller electorates such as Tasmania and South Australia. Under the reduced quota of a double dissolution election, even a brand new small party such as Derryn Hinch’s Justice Party can be successful in a large state like Victoria in the new system.

However, despite the success of such parties in 2016, it is likely that the new system will, over time, lead to fewer parties contesting Senate elections, although this may take a number of elections to become evident.

Smaller parties that do not have an obvious constituency and do not have the financial or grass-roots support to campaign throughout a state will find it difficult to replicate the results of, for example, the Australian Motoring Enthusiast Party in 2013.

The Senate voting system that operated prior to the 2016 election created incentives for registering multiple parties with catchy names and using those to channel preferences through GTVs to another party, expanding the ballot paper.[79] In the absence of preference deals, single-issue parties (or those with an appealing or provocative party name) will rarely gain sufficient votes or preferences to be elected—particularly in a normal half-Senate election—and will not be able to reliably direct their preferences. Over a number of election cycles, as evidence increases that this type of participation is ineffective, the number of these parties should decrease.

Parties that have a specific constituency—such as those who appeal to certain religious groups, for example Australian Christians or Christian Democrats—may choose to merge rather than risk splitting their vote between multiple parties. Under the previous Senate voting system, these parties could easily trade preferences, pooling the vote of their constituency. Under the new system, these parties will depend on their voters allocating preferences to all of them or have their vote split between the parties. By merging or running joint tickets in the Senate, parties that share a constituency can better consolidate their voter base and increase their chances of election.

This is not to suggest that small or single-issue parties will disappear, but rather that the removal of some of the incentives for the creation of these parties should lead to fewer of them appearing on the ballot paper in the future—particularly at a normal half-Senate election.

In addition to its recommendations for changes to the Senate voting system, the JSCEM interim report on the 2013 federal election also made recommendations in relation to party registration; namely, a minimum threshold requirement of 1,500 unique members to register a political party.[80] These recommendations have not yet been acted on by the Government.

In her second reading speech on the Bill implementing the changes, Senator Lee Rhiannon stated that the Greens objected to raising the registration requirements as it would be a barrier to small parties:

The Greens have stood firm in our longstanding support for the vital role of small parties in our democracy. We did this when Labor, the Liberals and the Nationals voted together to double the nomination fees to run in an election. It was clearly a move to try to knock out smaller parties, and some people have come to realise that that is not the fair way to enhance our democratic process. There was also the controversial attempt to raise the number of members needed to register a party, which the Greens did not agree with, because we recognised that most parties start small with little money and there should not be barriers such as large memberships or high nomination fees put in their way.[81]

By 2016, a total of 66 different parties contested the federal election—almost double the 35 who contested the 2004 federal election (see Figure 4). As noted above, however, it is likely that the new system will, over time, lead to fewer parties contesting Senate elections.

Figure 4: Unique parties contesting recent federal elections

Source: Compiled by the Parliamentary Library from AEC data.

In addition to registered political parties, Senate elections allow non-affiliated groups of candidates to run as a group on the ballot paper, receiving an above the line square. In NSW, which has traditionally attracted the greatest number of candidates and groups to its Senate ballot paper, the number of parties and groups in the past two federal elections numbered in the 40s (although it decreased slightly between 2013 and 2016), and the number of candidates has continued to climb, reaching 151 in 2016 (see Figure 5).

The changes to below-the-line voting allow numbering as few as 12 preferences. This change has certainly reduced the burden on voters, who otherwise would have had to number all 151 preferences below the line on the NSW ballot paper. However, it is arguable that even a relatively well-informed NSW voter would not have sufficient information on all 41 groups to make a rational decision as to whether each was deserving of one of their preferences—particularly given many of these groups would have done little or no campaigning.

Figure 5: Candidates and groups on the NSW Senate ballot paper in recent elections

Source: Compiled by the Parliamentary Library from AEC data.

Proportionality

One of the criticisms of the previous Senate voting system was that it produced disproportionate results, some candidates having been elected with very small first preference votes over candidates with larger votes. As outlined in this paper, there is a limit to how proportional the outcome of an STV electoral system can be; however, it is apparent that the 2016 Senate results were roughly proportionate to the share of the vote that accumulated to respective parties and groups. Parties with larger or medium-sized votes tended to gain more seats relative to their vote share under the new system (see Figure 6). 

Figure 6: Proportion of votes and seats won in the Senate by party/group

Source: Compiled by the Parliamentary Library from AEC data.

Kevin Bonham has also analysed the proportionality of the new Senate by looking at the average number of votes and seats won by each party in each state. He concluded that it was generally proportional, although it might have favoured the left side of politics slightly:

Out of the 72 seats, every party that has averaged over 1/73rd of the vote per state has won at least one seat, and every party that has averaged less has won none. NXT [Nick Xenophon Team] has won three seats instead of a proportional four (mainly because outside SA its 1.8% vote was too evenly spread) and One Nation has won four instead of three (on superior preferencing performance by collecting votes from a lot of right-micros that won nothing).[82] 

Comparing the number of seats and votes won by the major parties to everyone else, it is apparent that the combined seats the major parties won in the Senate closely aligns with the proportion of the vote they received. However, the combined share of seats won by the major parties is consistently greater than the share of votes that they win, suggesting that the Senate voting system has afforded a small degree of advantage to the major parties (see Figure 7). The new Senate voting system has not yet changed that situation; although, one election—and an unusual double dissolution election at that—is not a sufficient basis to discern a trend.

Figure 7: Combined share of seats and votes of the ALP and Coalition, 1987–2016

Source: Compiled by the Parliamentary Library from AEC data.

When considering Senate proportionality, an important factor is the decreasing share of votes that the major parties have received in recent elections. The 2016 Senate election did not slow this trend, with 26.39 per cent of votes not going to the Coalition, the ALP or the Greens (see Figure 8). As outlined in this paper, however, many voters who gave their first preference to a minor party gave their second or third preference to one of the established political parties (see also Appendix A).

Figure 8: Party share of first preference Senate vote, 1987–2016

Source: Compiled by the Parliamentary Library from AEC data.

In general, this decrease in the major party vote in the Senate is not reflected nearly as dramatically in terms of share of seats won, with non-major parties only winning 14.5 per cent of the seats in 2016 (see Figure 9). This is a result of the preferential nature of the system because, while in total around one quarter of the Senate vote is going to non-major parties, that vote is split between many parties (53 of them in 2016), and tends to flow back to the bigger parties on preferences (rather than being an inherent bias towards the major parties).

The results over recent elections show that there is considerable potential for one (or a small number) of non-major parties to collect these votes—much as One Nation did in 2016—and to turn them into Senate seats.

Figure 9: Share of Senate seats won, 1987–2016

Source: Compiled by the Parliamentary Library from AEC data.

Public response to the changes

A survey conducted by Essential Research in mid-July 2016 asked respondents whether they thought that the changes to the system made Senate voting easier or more difficult. A total of 19 per cent of respondents reported that the changes made it easier to vote, and 37 per cent said it was more difficult. Older respondents and those who had not yet completed year 12 were most likely to think it made voting more difficult, and there was not a large difference in perceptions between supporters of different parties.[83]

Asked if the reforms led to a more democratic outcome in the Senate, 20 per cent said it did make it more democratic, 15 per cent said less democratic, 39 per cent said it made no difference, and 26 per cent did not know.[84]

Essential Research had conducted an earlier poll on Senate reform in late February 2016. Essential found that just over half of the respondents (53 per cent) approved of the changes, with a majority of both Coalition and Labor voters approving, while 16 per cent of all respondents disapproved of the changes.[85]

In the final analysis, the relatively low informality rate and the high rate of compliance with the minimum number of preferences above the line suggest that the public did not reject the changes to the voting system.

Conclusion

By any reasonable measure, the recent reforms to the Senate voting system have successfully achieved the parliament’s objectives. Voters appear to have attended to the AEC’s advertising campaign and polling place and ballot paper instructions to allocate six preferences above the line, straightforwardly abandoning the habit of 30 years of Senate voting. While informal voting did increase compared to the 2013 Senate election, the increase was mild in the context of such a major change introduced at such short notice. On the whole, it did not appear the rate of informal voting was a result of the ballot paper changes.

As a double dissolution election, with a halved quota for election, the 2016 federal election is not an ideal test case of the potential of the new voting system to change the composition of the Senate. However, the results of the section 282 recount suggest that small parties will continue to be able to be elected.

Due to the success of some small parties who only just gained seats under the reduced quota (such as the Liberal Democrats and Derryn Hinch’s Justice Party), the likely effect of the new system on the fortunes of very small parties in Senate elections might not yet be apparent. But, in order to be successful, those interested in forming such parties with the hope of being elected to the Senate may need more support in the electorate than was mustered by very small parties in 2013 and 2016. Over several electoral cycles with the new system in place, it is likely that small parties that do gain Senate representation will require substantial grassroots support and/or consolidation.

The new system has not arrested the decline in the major party vote in the Senate observed over the past several decades. Yet voters are certainly not completely abandoning the major parties, and many minor party voters are keeping the major parties high in their preferences. This is positive for both major and minor parties: if a minor party can marshal sufficient support, it stands a chance of gaining seats; but if it does not receive enough votes to get over the line, the votes may well flow to a major party.

From the perspective of the voter, it means that a preference can be lodged for a smaller party for whatever reason, but voters can also be reasonably confident that their vote will count for a party with some chance of electoral success. What is reasonably clear is that voters have embraced the opportunity to allocate their own preferences in the Senate—with preference deals made by parties having much less influence.

Appendix A: Analysis of voter Senate preference choice

The 2016 Senate election represents the first Senate election since 1983 where the majority of voters in the Senate allocated their own preferences between the parties and candidates on the ballot paper.

One of the justifications for the introduction of group voting tickets in 1984 was that most voters wish to follow their preferred party’s how-to-vote (HTV) card, and allowing them to vote using a group voting ticket (GVT) essentially simplified the process of following the party’s HTV card (see discussion on page 5). If it was true that most voters wanted to allocate preferences according to their preferred party’s ordering in 1984, however, an analysis of the allocated preferences from 2016 shows that this is no longer the case.

The analysis in this appendix uses the full preference file published by the AEC for the 2016 election, which lists the preferences of all 14.4 million formal Senate ballot papers, both above and below the line. It is included in order to indicate where preferences from certain parties come from—and who they are going to. The analysis reveals that voter behaviour is much more complex than all left-leaning or all right-leaning voters preferencing the same way or allocating preferences to the same parties.

The analysis focuses on first, second and third preferences for the sake of simplicity, but also because it is likely that voters’ first few preferences are more reflective of their sympathy with a party than their later preferences. In some cases, voters may not have heard of six parties on the ballot paper, but it is very likely that they would know at least three.

Note that this analysis examines the preferences as allocated by voters, and not the preference flows of the count. It is an analysis of voter intention, as opposed to electoral effect. For example, the analysis shows who received the second and third preferences of the Coalition and Labor voters—even though in the count most of those preferences will count toward electing the Coalition or Labor candidate, and very few will be transferred to count against electing the second or later preferenced candidates.

Charts in this section only include parties that received at least 0.1 per cent of the state Senate vote. All charts and tables in this appendix are based on Parliamentary Library analysis of Senate preferences published by the AEC on its website.[86] Senate votes for the territories were not included in this analysis as the two territories each only elect two senators and preferences play a much smaller role in determining the result.

Note: Political party abbreviations used in this appendix are official AEC abbreviations, a list of which is at Appendix B.

Preferencing of established parties

The decrease in the primary vote of the major parties over time is discussed in the main body of the paper above. The preference data allows an examination of how many of voters who did not use their first preference for an established or major party (here defined as the Coalition, the ALP or the Greens) instead use a second or third preference for one of the established parties.

The ten largest (in terms of first preference vote) of the non-major parties are included in this analysis to see what proportion of minor-party voters gave their second or third preference to an established party (see Table A1). While the proportion varied, between 28 per cent and almost 70 per cent of the minor-party vote across the states potentially came back to the established parties as preferences.[87]

Table A1: Percentage of votes for top ten minor parties for which second or third preferences went to established parties, by state

Note: Top ten parties by number of Australia-wide first preference votes, excluding the Coalition, the ALP and the Greens.

This data suggests that many voters for minor parties are retaining one or more major parties as a ‘backup’ vote. However, it is impossible to know the motivation behind the preferencing on the basis of this analysis. These votes could constitute a protest against the major parties—denying them the benefits of a first preference vote (such as public funding) while ensuring that the major party does eventually get their preference—or could have been cast for some other reason (by virtue of name recognition of the major parties, for example).

Yet these voters are, in fact, using the preference system in the manner that it was designed to be used. Their first preference may go to the most appealing party—regardless of that party’s prospects of election—but their vote may ultimately count for a party with a chance of actually electing a candidate if their preferred party does not poll well enough to do so.

Graphic analysis of above-the-line preference flows, state by state

The graphs for each state below depict the flows between the first, second and third above-the-line preferences in that state. The strength of flows between parties are according to the proportion of votes allocated as first through third preferences—that is, larger segments or thicker connections between parties represent more votes than smaller segments and thinner links. The columns of parties are in preference order: from first preference on the left side to third preference on the right. The charts include all ballot papers for the state in which there were at least six above-the-line preferences.

New South Wales

Strong preference flows eventuated between the LPNP (the joint Liberals and Nationals ticket); the Christian Democrats; and the Shooters, Fishers and Farmers. This was the recommended first three preferences for those who followed the LPNP HTV card, and equates to 32.7 per cent HTV correspondence by LPNP voters. In stark contrast, only 1.4 per cent of ALP voters followed their party’s HTV to the third preference (1: ALP; 2: GRN; 3: Renewable Energy Party (REP));[88] although it is impossible to know whether voters who allocated certain preference patterns were following the HTV card, or whether the HTV card reflected the voters’ natural preferences.

Figure A1: New South Wales voters’ first three preferences

NSW elected two Senators from minor parties: the Liberal Democrats (LDP) and One Nation (ON). The preference flows to and from these two parties are set out in Table A2 and Table A3 below. Both have very different preference flows. The LPNP make up almost 80 per cent of first preferences of the votes that had the LDP as second preference. Whereas ON second preference votes came almost equally from ALP and LPNP first preferences, almost half of the preferences from the LDP went to the LPNP as second preferences. In neither of these cases did the ALP, LPNP or LDP HTV cards recommend those preference orders.

Table A2: Preferences to and from the Liberal Democrats (LDP) in New South Wales

Table A3: Preferences to and from One Nation (ON) in New South Wales

Victoria

The Victorian Senate election is notable mainly for the election to the Senate of Derryn Hinch’s Justice Party (DHJP) candidate Derryn Hinch. It is clear from Figure A2 that DHJP received reasonably strong votes in first, second and third preference positions. While the DHJP did not direct preferences on its HTV,[89] it received the first column on the Victorian Senate ballot paper; and 3.8 per cent of the DHJP votes were ‘donkey votes’, in that voters filled out the first three groups on the ballot paper in order (1: DHJP; 2: Group B; 3: AJP).[90]

Figure A2: Victorian voters’ first three preferences

The Liberals and Nationals again got reasonably strong correspondence with their HTV card, with 43.5 per cent of LPNP voters following the HTV card for the first three preferences (1: LPNP; 2: Family First; 3: Australian Christians).

Almost as many ALP voters gave their second preference to the Greens (16.7 per cent) as followed the ALP’s HTV card (19.5 per cent) (1: ALP; 2: Australian Sex Party; 3: DHJP).

While the DHJP received strong preferencing from voters of both of the major parties—and there was also strong preferencing from the DHJP to the major parties—it was more strongly associated with the ALP than the LPNP. There were also reasonably high amounts of preferencing from DHJP voters to the Animal Justice Party, but relatively little to the Greens (see Table A4). 

Table A4: Preferences to and from Derryn Hinch’s Justice Party (DHJP) in Victoria

Queensland

The LNP in Queensland managed to receive 23.0 per cent correspondence with its HTV card, allocating its second preference to Family First (FFP) and third to the Christian Democratic Party (Fred Nile Group/CDP). The ALP had 13.7 per cent correspondence with its HTV recommendation of second preference to the Greens and third to the Australian Sex Party/Marijuana (Hemp) Party (ASXP/HEMP) joint ticket.[91]

Figure A3: Queensland voters’ first three preferences

Queensland was the only state that elected two ON candidates to the Senate. As can be seen in Figure A3, ON polled reasonably well in terms of first, second and third preferences coming from voters from a variety of other parties.[92] Over a third of the votes for ON as a second preference came from those who put the LNP first; however, only one of the parties that ON voters preferenced second received as much as 10 per cent of the ON preferences, and that party, the Australian Liberty Alliance (ALA) was second on the ON HTV card (see Table A5).

Table A5: Preferences to and from One Nation (ON) in Queensland

South Australia

South Australian Senate preferencing at the 2016 election was characterised by the strong performance of the Nick Xenophon Team (XEN) in both first and later preferences, and the Family First Party (FFP), particularly in terms of second preferences.

The Liberal Party (LP) in South Australia fared particularly poorly in terms of correspondence with its HTV card compared to other states, with only 12.6 per cent of LP voters following its recommended preferencing (1: LP; 2: FFP; 3: LDP). In contrast, 15.4 per cent of Labor voters followed the Labor HTV card (1: ALP, 2: GRN; 3: AJP).[93]

Figure A4: South Australian voters’ first three preferences

Both FFP and XEN are minor parties with strong levels of support in South Australia. However, the two parties show different patterns of preferencing from their voters.

Over three-quarters of the votes that came to FFP as a second preference came from voters who had given their first preference to the LP, however less than 20 per cent of those who gave their first preference to FFP gave their second to the LP. Family First voters were less consistent in their preferencing, with almost a third of their preferences going to the Christian Democrats and similar proportions going to the ALP, XEN and ON.

Table A6: Preferences to and from Family First (FFP) in South Australia

Nick Xenophon Team voters were much more strongly associated with the major parties. Over four-fifths of the voters who put XEN second gave their first preference to either the ALP or LP, and were about equally split between them. Those who voted 1 XEN were more divided, with about one quarter each voting ALP or LP second. In terms of ideological groupings, slightly more XEN voters gave their second preference to the ALP and left-of-centre parties (GRN) than to the LP and right-of-centre parties (ON and FFP). This suggests that XEN does occupy a centrist position in the minds of voters, but leans slightly to the left.

Table A7: Senate preferences to and from Nick Xenophon Team (XEN) in South Australia

Western Australia

The ALP received 19.0 per cent correspondence with its HTV recommended preferencing (1: ALP; 2: GRN; 3: ASXP/HEMP), and the LP received 36.9 per cent HTV correspondence with its preferencing (1: LP; 2: NP; 3: Australian Christians (AUC)).[94]

The strong Liberal preferencing largely accounts for the significant representation of AUC as the most popular third preference in Western Australia. Also of note is that the Liberals and Nationals did not run joint tickets in the WA Senate election—although there are strong preference flows between the two parties, with 89 per cent of Nationals second preference votes coming from the Liberals.

Figure A5: Western Australian voters’ first three preferences

One of the four seats ON won was in WA, which occurred at the very end of the count of the votes through preferences from a primary vote of 3.99 per cent (or 0.5 quotas). One Nation received reasonably strong preferencing from both LP and ALP voters, with roughly a third of its second preference votes coming from the LP and one quarter from the ALP.

Preferences from ON voters were much less tightly clustered, with only the LP and the Sex Party (ASXP)/HEMP joint ticket receiving more than 10 per cent of ON voters’ second preferences. Small parties—like the Shooters, Fishers and Farmers Party; Derryn Hinch’s Justice Party; Rise Up Australia; and the Australian Liberty Alliance—did relatively well in terms of preferences from ON voters.

Table A8: Senate preferences to and from One Nation (ON) in Western Australia

Tasmania

Tasmania elected one senator from the Jacqui Lambie Network (JLN) and, as Figure A6 and Table A9 show, JLN had both a reasonably strong primary vote and relatively strong second and third preferences from a variety of different parties. It is notable, however, that almost half of the preferences to the JLN came from the ALP, whereas roughly equal numbers of JLN preferences went to the ALP and ON.

The strong ALP preferencing for JLN may account for some of the low correspondence of ALP voters with the ALP HTV card (only 6.3 per cent), which put JLN as third preference behind the Greens as second. Correspondence with the Liberal HTV was better, but also relatively low (at 13.2 per cent). The Liberal HTV card preferenced the CDP and the Shooters, Fishers and Farmers Party (ASP).[95] Another effect may be the large ALP below-the-line vote for Lisa Singh.

Figure A6: Tasmanian voters’ first three preferences

Table A9: Preferences to and from the Jacqui Lambie Network (JLN) in Tasmania

Appendix B: Political party abbreviations

Source: AEC[96]

• References

  [1].     P Norris, ‘Choosing electoral systems: Proportional, majoritarian and mixed systems’, International Journal of Political Science, 18(3), July 1997, pp. 297–312.

  [2].     DM Farrell and I McAllister, The Australian electoral system: Origins, variations and consequences, UNSW Press, Sydney, 2006.

  [3].     The number of seats is referred to as the ‘district magnitude’ by those who study electoral systems.

  [4].     The requirement to put full preferences on the Senate ballot paper actual preceded the move to STV by 14 years.

  [5].     Australian Electoral Commission, ‘Informality (%) House of Representatives and Senate’, Australian Electoral Commission website.

  [6].     Joint Select Committee on Electoral Reform, First Report, Australian Government Publishing Service, Canberra, September 1983.

  [7].     Ibid., p. 64.

  [8].     Commonwealth Electoral Legislation Amendment Act 1983.

  [9].     M Maley, ‘Senate electoral reform’, Australian Public Law, blog, 29 September 2015.

  [10].    Ibid.

  [11].    J Lundell and ID Hill, ‘Notes on the Droop Quota’, Voting Matters, 24, October 2007, pp. 3–6.

  [12].    The two territories elected two senators, each with terms tied to the House of Representatives parliamentary term, and so the quota for the territory senators does not change at a double dissolution election and is always one third of the vote.

  [13].    The CEA also allows multiple parties to combine and run as a single group on the Senate ballot paper. The Liberal and National parties often run as a single group in Senate elections.

  [14].    AEC, ‘First preferences by candidate – Vic’, Virtual Tally Room – 2004 Federal Election, AEC website, 9 November 2005.

  [15].  B Brown, Outrageous Senate outcome – Brown, media release, 3 November 2004.

  [16].    E Colman, ‘Green with frustration over secret preference deals’, Weekend Australian, 6 November 2004, p. 21.

  [17].    A Green, ‘Reform is needed of the Senate voting system run by party bosses’, The Sydney Morning Herald, 29 October 2004, p. 17.

  [18].  B Brown, ‘Second reading speech: Senate Voters’ Choice (Preference Allocation) Bill 2004’, Senate, Debates, 9 December 2004, p. 4.

  [19].  AEC, ‘First preferences by candidate – Vic’, Virtual Tally Room – 2010 Federal Election, AEC website. 15 September 2010.

  [20].    Dropulich was not successful in the 2014 re-run of the WA Senate election so ultimately was not elected to the Senate.

  [21].  M Maley, op. cit.

  [22].    M Ronaldson (Special Minister of State), Ministerial reference to the Joint Standing Committee on Electoral Matters, media release, 9 December 2013.

  [23].    Joint Standing Committee on Electoral Matters, Interim report on the inquiry into the conduct of the 2013 federal election: Senate voting practices, Commonwealth of Australia, Canberra, May 2014, p. v.

  [24].    Ibid.

  [25].    Ibid, p. vii.

  [26].  M Grattan, ‘Proposed Senate voting reforms would curb micro parties’, The Conversation, 12 May 2014.

  [27].  G Gray, Gary Gray endorses optional preferential voting in the Senate, media release, 9 May 2014.

  [28].    M Grattan, ‘Will Senate voting reform end up in the too hard basket?’, The Conversation, 22 September 2015.

  [29].    H Aston, ‘Brough’s attach riles vote-wary crossbenchers’, The Sydney Morning Herald, 23 September 2016, p. 5.

  [30].    P Coorey, ‘Early poll threats as Senate reform looms’, Australian Financial Review, 3 February 2016, p. 3.

  [31].    D Crowe and R Lewis, ‘Greens take aim at micro-party deals’, The Australian, 4 February 2016, p. 5.

  [32].    Parliament of Australia, ‘Commonwealth Electoral Amendment Bill 2016 homepage’, Australian Parliament website. 

  [33].    P Hendy, ‘Consideration in detail: Commonwealth electoral amendment Bill 2016’, House of Representatives, Debates, 24 February 2016, p. 2116.

  [34].    S Morrison, ‘Reference to Committee: Commonwealth Electoral Amendment Bill 2016’, House of Representatives, Debates, 22 February, 2016, p. 1570.

  [35].    D Muller, Commonwealth Electoral Amendment Bill 2016, Bills Digest, 96, 2015–16, Parliamentary Library, Canberra, 1 March 2016.

  [36].    Joint Standing Committee on Electoral Matters, Advisory Report on the Commonwealth Electoral Amendment Bill 2016, Commonwealth of Australia, Canberra, 2 March 2016.

  [37].    Ibid., p. 67.

  [38].    Ibid.

  [39].    ‘Occasional Note: Prorogation and a new session of Parliament’, Procedural Informal Bulletin no. 303, Senate, Canberra, 23 March 2016.

  [40].    ‘Crossbenchers to launch High Court challenge to senator voting changes’, Sunday Canberra Times, 20 March 2016, p. 5.

  [41].    P Karp, ‘Senate voting changes: David Leyonhjelm turn to crowdfunding to pay for legal challenge’, Guardian, 24 March 2016.

  [42].    P Wearne, ‘Labor rejects legal bid’, West Australian, 21 March 2016, p. 14.

  [43].    http://www.austlii.edu.au/au/cases/cth/HCATrans/2016/85.html

  [44].    Day v Australian Electoral Officer for the State of South Australia & Anor [2016] HCA 20.                                                                                                                                                          

  [45].    Votes exhaust when the ballot paper has no more preferences that can be allocated to candidates who are still in the count.

  [46].  Ibid. at [37].

  [47].    A Twomey, ‘Before the High Court: Day v Australian Electoral Officer (SA): Senate Voting Reforms under Challenge’, Sydney Law Review, 38, 231–41, p. 241.

  [48].    Day v. Australian Electoral Officer for the State of South Australia & Anor, op.cit.

  [49].    See, for example, DM Farrell, Electoral Systems: A Comparative Introduction, 2nd edn, Pangrave Macmillan, Houndmills, Basingstoke, Hampshire, UK, 2011.

  [50].    None of the specific variants of STV used for Senate elections have ever had an official name, making a distinction between them somewhat difficult. No electoral systems with the essential characteristics of the Senate voting system (either pre- or post-2016) exist elsewhere in the world, meaning there is little in the way of precedent. The main distinguishing characteristics of the system as it operated between 1984 and 2014 was the use of GVTs, hence ‘the GVT system’. Lacking a more correct alternative, this paper refers to the post-2016 Senate voting system as the ‘new Senate voting system’.

  [51].    DM Farrell and I McAllister, ‘The 1983 change in surplus vote transfer procedures for the Australian Senate and its consequence for the single transferable vote’, Australian Journal of Political Science, 38(3), 2003, pp. 479–91.

  [52].    The writ is an order, under section 12 of the Constitution, issued by the Governor of a state to hold an election for vacancies in the Senate and that nominates a date at which the election must be declared and the writs returned. The return of the writ is accompanied by a certification of the results of the election and signifies the conclusion of the election under the CEA.

  [53].    I Brightwell, ‘Can we trust that our Senate votes were counted correctly?’, Crikey, 5 August 2016.

  [54].    As it has for recent Senate elections, the AEC has made the preferences recorded for each formal vote in the 2016 Senate election available for download on its website. This is sufficient detail for a third party to independently verify the accuracy of the count.

  [55].    Both the ACT and Tasmania also tend to feature lower numbers of candidates and therefore a less imposing ballot paper. The NT, which like those two jurisdictions has a small population and a relatively small Senate ballot paper, also features much higher rates of above-the-line voting—suggesting that it is Hare-Clark which is responsible for the high rate of below-the-line voting, not just the size of the ballot paper.

  [56].    H Aston, ‘Tasmania gives ALP factions a lesson’, The Sydney Morning Herald, 20 July 2016, p. 7

  [57].    Joint Standing Committee on Electoral Matters, Inquiry into Commonwealth Electoral Amendment Bill 2016, 1 March 2016, p. 18.

  [58].    K Bonham, ‘Senate reform performance review part 1’, Dr Kevin Bonham, blog, 6 August 2016.

  [59].    GJ Bukley, Submission to the Joint Standing Committee on Electoral Matters, Inquiry into the Commonwealth Electoral Amendment Bill 2016, 29 February 2016.

  [60].    B Day, ‘Second Reading: Commonwealth Electoral Amendment Bill 2016’, Senate, Debates, 16 March 2016, p. 2041–2086.

  [61].    D O’Neill, ‘Second Reading: Commonwealth Electoral Amendment Bill 2016’, Senate, Debates, 16 March 2016, p. 2222.

  [62].    AEC, Exhaustion: Senate ballot paper study 2016, Senate Ballot Paper Study papers, AEC, n.d.

  [63].    R Day, ‘Written submissions of Plaintiff’, Submission in Day v. Australian Electoral Officer for the State of South Australia & Anor, [2016] HCA 20.

  [64].    A Green, ‘How voters reacted to the Senate’s new electoral system’, Antony Green’s Election Blog, 11 October 2016.

  [65].    K Bonham, ‘Senate performance review part 1’, op.cit.

  [66].    A Green, ‘How voters reacted to the Senate’s new electoral system’, op. cit.

  [67].    AEC, Exhaustion: Senate ballot paper study 2016, op.cit.

  [68].    J Ireland, ‘Labor warns of voter “gerrymander”’, The Sydney Morning Herald, 22 February 2016, p. 9.

  [69].    G Hutchens, ‘Reports of confusing advice about new Australian Senate voting rules’, The Guardian (online), 2 July 2016.

  [70].    AEC, Submission to the Joint Standing Committee on Electoral Matters, Inquiry into the conduct of the 2016 federal election and matters related thereto, 1 November 2016, p. 40.

  [71].    Ibid.

  [72].    ALP, Dissenting report, Joint Standing Committee on Electoral Matters, Advisory Report on the Commonwealth Electoral Amendment Bill 2016, Commonwealth of Australia, Canberra, 2 March 2016.

  [73].    H Aston, ‘Reforms may give Coalition rare majority’, The Sydney Morning Herald, 17 February 2016, p. 5.

  [74].    G Druery, Evidence to Joint Standing Committee on Electoral Matters, Inquiry into Commonwealth Electoral Amendment Bill 2016, 1 March 2016.

  [75].    Commonwealth Electoral Act 1918, section 282.

  [76].    Australia, Senate, Re-counts of Senate votes pursuant to section 282 of the Commonwealth Electoral Act—Letters from the Australian Electoral Officers for New South Wales, Queensland, South Australia, Tasmania, Victoria and Western Australia, Canberra, 30 August 2016.

  [77].    B Shorten (Leader of the Opposition), Transcript of doorstop: Sydney, media release, 4 August 2016.

  [78].    K Bonham, ‘Senate performance review part 2’, Dr Kevin Bonham, 8 August 2016

  [79].    See, for example, G Druery, Evidence to Joint Standing Committee on Electoral Matters Conduct of the 2013 federal election and matters related thereto, 1 May 2014.

  [80].    JSCEM, Interim report on the inquiry into the conduct of the 2013 federal election: Senate voting practices, op.cit.

  [81].    L Rhiannon, ‘Second reading speech: Commonwealth Electoral Amendment Bill 2016’, Senate, Debates, 2 March 2016, p. 1578. The Australian Greens did not lodge a dissenting report to the JSCEM interim report.

  [82].    K Bonham, ‘Senate performance review part 1’, op.cit.

  [83].    Essential Research, The Essential Report, 19 July 2016.

  [84].    Ibid.

  [85].    Essential Research, The Essential Report, 1 March 2016.

  [86].    AEC, ‘Senate downloads’, 2016 Federal Election Tally Room, AEC website.

  [87].    This analysis uses the voters’ expressed preferences, rather than preferences as counted. A ballot paper that preferenced a minor party above a major party may have flowed to that major party when the minor party candidate was excluded, or may have elected the minor party candidate and therefore not counted against any of the later preferences. At the time the voter casts their vote they do not know which of these will be the case, and for any given vote it may be impossible to determine which candidate it voted against once the count has run.

  [88].    Australian Broadcasting Corporation (ABC), ‘Senate How to Votes – New South Wales’, Federal Election 2016, ABC website.

  [89].    Derryn Hinch’s Justice Party ran an open ticket, advising voters to ‘vote 2–6 your own preference’ on its HTV card.

  [90].    ABC, ‘Senate How to Votes – Victoria’, Federal Election 2016, ABC website.

  [91].    ABC, ‘Senate How to Votes – Queensland’, Federal Election 2016, ABC website.

  [92].    One of whom was subsequently found to have been invalidly elected by the Court of Disputed Returns due to holding dual citizenship. The replacement selected in a recount chose not to sit as a One Nation senator.

  [93].    ‘Senate How to Votes - South Australia’, ABC Federal Election 2016 website.

  [94].    ABC, ‘Senate How to Votes - Western Australia’, Federal Election 2016, ABC website.

  [95].    ‘Senate How to Votes – Tasmania’, ABC Federal Election 2016 website.

  [96].    AEC, ‘2016 Federal Election Political Parties’, 2016 Federal Election Tally Room, AEC website.

• Licence

  © Commonwealth of Australia

  

  Creative Commons

  With the exception of the Commonwealth Coat of Arms, and to the extent that copyright subsists in a third party, this publication, its logo and front page design are licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Australia licence.

  In essence, you are free to copy and communicate this work in its current form for all non-commercial purposes, as long as you attribute the work to the author and abide by the other licence terms. The work cannot be adapted or modified in any way. Content from this publication should be attributed in the following way: Author(s), Title of publication, Series Name and No, Publisher, Date.

  To the extent that copyright subsists in third party quotes it remains with the original owner and permission may be required to reuse the material.

  Inquiries regarding the licence and any use of the publication are welcome to webmanager@aph.gov.au.

• Disclaimer

  This work has been prepared to support the work of the Australian Parliament using information available at the time of production. The views expressed do not reflect an official position of the Parliamentary Library, nor do they constitute professional legal opinion.

  For copyright reasons some linked items are only available to members of Parliament. All links are correct at the time of publication and are not subsequently updated.

  Any concerns or complaints should be directed to the Parliamentary Librarian at webmanager@aph.gov.au. Parliamentary Library staff are available to discuss the contents of publications with Senators and Members and their staff. To access this service, clients may contact the author or the Library‘s Central Entry Point for referral.

Related

FlagPost

Political party defections

By Cathy Madden

3 min read

FlagPost

Highlights of the 2024-25 Mid-year economic and fiscal update

By Dr Kate Wagner

3 min read

FlagPost

Status update: local content quotas for streaming services

By Nell Fraser

5 min read