EMC Submission No. 103 Received 9 September 2019

Submission to the Victorian Parliament's Electoral Matters Committee Inquiry into the Conduct of the 2018 Victorian State Election

By Antony Green Election Analyst

Contents Introduction ...... 2 Background on Preferential Voting ...... 3 The Introduction of PR-STV in ...... 5 Further Developments with Hare-Clark ...... 5 The Development of the Senate's Electoral System ...... 6 Electoral Changes at the 2016 Election ...... 9 How Voters Reacted to the New Senate System in 2016 ...... 10 Changes to Nominations under the Senate's New Electoral System ...... 12 Turning Votes into Seats Under the Senate's New Electoral System ...... 13 Lessons for Victorian Legislative Council Elections from the Senate Experience ...... 16 A Note on Victorian Informal Voting ...... 19 Discussion on Options ...... 20 Recommendations ...... 22 Appendix 1: Hare-Clark Ballot Paper Examples – ACT ...... 23 Appendix 2: Sample Senate Ballot Paper (Post 2016 changes) ...... 24 Appendix 3: Ireland Ballot Paper Example ...... 25 Appendix 4: Sample PR-STV How-to-Vote recommendation from 1922 NSW Election ...... 26 Appendix 5: 1974 NSW Senate Election – Labor How-to-Vote...... 27 Appendix 6: 1999 NSW Legislative Council Ballot Paper (the 'tablecloth') ...... 28 Introduction

When introduced proportional representation for Legislative Council elections in 2006, it adopted most of the features of the electoral system then used to elect the Senate and all other mainland Legislative Councils.

As shown in Table 1, there has been a pattern for Australian state parliaments to follow the Commonwealth's path in upper house electoral reform, first with the use of proportional representation, then with the adoption of a divided ballot paper and group voting tickets.

Table 1 - Following the Senate Example

Year of Election where rules first applied Introduction of Introduction of Divided Ballot Paper Proportional and Group Voting Abolition of Group Jurisdiction Representation Tickets Voting Tickets Senate 1949 1984 2016 1978 1988 2003 Victoria 2006 2006 .. 1989 1989 .. 1975 1985 2018 Note: The Queensland Legislative Council was abolished in 1922. The Tasmanian Parliament uses proportional representation in the lower house and single member electorates in the Legislative Council. The ACT uses proportional representation for its one chamber while the uses single member electorates.

Majoritarian voting in the Senate was abandoned for proportional representation in 1949. The upper house proportional model was followed by South Australia in 1975 (it used a form of D'Hondt proportional representation initially), and by NSW when popular election for the Legislative Council was introduced in 1978. Western Australia abandoned majoritarian voting in favour of proportional representation in 1989, followed by Victoria in 2006.

As will be explained later, a divided ballot paper with group voting tickets was adopted for the Senate in 1984, largely as a solution to high rate of informal voting. South Australia in 1985 and NSW in 1988 incorporated the Senate changes into their existing proportional systems, while Western Australia and Victoria introduced proportional representation and the Senate electoral system at the same time.

Introduced at a time when there were fewer political parties, group voting tickets essentially institutionalised party how-to-votes. It was not anticipated that new parties would form entirely to use group voting tickets. Over three decades, parties learnt to 'game' the system with tickets, to engage in exotic preference swaps that became known as 'preference harvesting'. The number of groups on state Senate ballots rose from an average 7.7 in 1984 to 33.7 in 2013. The proportionality of party representation was increasingly distorted compared to the proportionality of party vote.

The problems created by group voting tickets were exposed by the 1999 NSW Legislative Council election. Voters were presented with what has become known as the 'tablecloth', a triple-decked ballot paper one metre wide by 700mm deep with 81 columns, 27 per row, and 264 candidates. (See Appendix 6). South Australia experienced similar problems after 2002.

2 Antony Green – Victorian EMC Submission After the 1999 fiasco, NSW abolished group voting tickets for the 2003 Legislative Council election. The divided ballot paper was retained, voters given a new voting option to indicate preferences for parties above the line. The system was first used at the 2003 NSW election, and along with reform of party registration rules, it reduced the number of groups on the ballot paper from 81 to 16.

The NSW problems have been repeated in other jurisdictions, but none adopted the NSW solution until the Commonwealth abandoned Senate group voting tickets in 2016. South Australia followed in 2018. Only Victoria and Western Australia retain group voting tickets as part of their electoral systems.

As this submission explains, Victorian Legislative Council results are following the same path to electoral distortion played out previously at Senate elections. In particular, the 2018 Victorian Legislative Council election displayed the same electoral distortions as the 2013 Senate election, a result that led to reform of the Senate's electoral system.

In this submission I will outline the problems that developed at Senate elections in the three decades after the introduction of group voting tickets in 1984. I will then compare results of the 2013 Senate election under group voting tickets with 2019 election results under the new electoral system. The comparison shows how the new system produces a more proportional result with fewer distortions produced by exotic preference deals.

It is best to start by outlining some basic features of preferential voting in multi-member electorates. This will categorise types of preferences in multi-members systems, and how ballot paper designs have enhanced or diminished the ability of voters and parties to control preferences.

Background on Preferential Voting

Preferential voting is the dominant method of voting in Australia, used for federal, state, territory and most local government elections. Preferential voting is used to elect representatives from both single member and multi-member districts. Very few countries other than Australia use preferential voting.

Preferential voting in single-member electorates is usually referred to in academic literature as the alternative vote. In recent years it has seen a revival in the United States where it is often called instant run-off voting.

In Australia single member preferential voting has three variants. By far the most common form is full preferential voting where voters must number all squares. Optional preferential voting is used in New South Wales, until recently in Queensland and the Northern Territory, and is the dominant form in other countries where preferential voting is used. In Tasmania and the ACT, limited preferential voting is used, voters being asked to indicate a minimum number of preferences.

Multi-member preferential voting is known in the academic literature as proportional representation by single transferrable vote, shortened to PR-STV or just STV. In Australia, PR-STV can be divided into two broad categories based on ballot paper structure.

• Hare-Clark ballots - used to elect the governing lower house in Tasmania and the ACT. Hare- Clark uses a ballot paper where votes must be cast for candidates. Sample Hare-Clark ballot papers are provided in Appendix 1.

• Divided ballots - used to elect upper houses at elections for the Commonwealth, New South Wales, Victorian, Western Australian and South Australian parliaments. Voters have the

Antony Green – Victorian EMC Submission 3 option to vote for parties 'above the line' or for individual candidates 'below the line'. An example of a divided ballot paper is provided in Appendix 2.

Recent developments in Australian electoral systems suggest that the Divided Ballot category should be sub-divided into two further categories based on how much control over preferences voters have when using 'above the line' party voting squares.

• Divided Ballots with Group Voting Tickets (GVTs), as used for the Victorian and Western Australian Legislative Councils. These systems allow voters to choose only one party above the line, after which the ballot paper is imputed to adopt that party's full list of preferences.

• Divided Ballots with Party Preference Choice, as used in the Senate and for the New South Wales and South Australian Legislative Councils. Voters can give preferences to parties by numbering above the line boxes. Preferences are imputed to be for candidates of the first chosen party, then the second party, and so on. Preferences above the line are optional (NSW, SA), limited (Senate) and limited below the line (NSW, SA, Senate).

Each of these ballot paper types and associated rules for completion have an impact on how voters indicate their preferences, and on the overall flows of preferences produced. The different ballot paper structures influence how voters indicate preferences, and alter the balance of control over preferences between voters and parties.

The academic literature on PR-STV identifies two types of preferences as key to explaining the party outcomes produced by PR-STV. These are -

• Intra-party or within-party preferences, how preferences flow between candidates of the same party

• Inter-party or between-party preferences, how preferences flow between candidates of different parties

Much of the international literature on PR-STV has been written based on analysis of Irish elections. An example Irish ballot paper is shown in Appendix 3. Candidates are listed in alphabetic order and party affiliation is shown. Preferences are fully optional, so a voter only needs to number one square to cast a formal ballot. The number of members elected from each constituency varies between three and five.

The Irish ballot paper, combined with optional preferential voting, means it does not fit into any of the Australian PR-STV ballot structures. Votes are more evenly split between candidates of the same party, and parties usually nominate only as many candidates per seat as have a chance of being elected. Parties encourage voters to number preferences for all candidates of the party, to maximise within-party preferences. They also encourage voters to preference candidates of potential partners, to maximise between-party preferences. Parties also nominate high profile individuals in the hope of luring votes and preferences from other parties.

As explained below, Australian PR-STV elections began with a version of the Irish ballot paper. Australian PR-STV ballot papers have evolved to influence the flow of both within and between party preferences. The adoption of party groupings on ballot papers, combined with a required minimum number of preferences, has maximised within-party preference flows. These flows have increased further with the use of divided ballot papers, while group voting tickets deal with between-party preferences in a way that takes choice away from voters.

4 Antony Green – Victorian EMC Submission The Introduction of PR-STV in Australia

When first introduced in Tasmania, PR-STV had a ballot paper very similar to Ireland's, but without party affiliations. From 1909 to the 1940s, Tasmanian elections had a single list of candidates in alphabetic name order with no party affiliations. Voters were required to give three preferences, a number that stayed in place until the 1970s when it was increased to equal the number of vacancies to be filled.

New South Wales used the same electoral system and ballot paper for three elections between 1920 and 1925. A sample Coalition how-to-vote card from 1922 election is reproduced in Appendix 4. The how-to-vote identifies the Coalition candidates without suggesting an order of preferences, and then suggests a further preference sequence for candidates of other parties.

This sort of how-to-vote encouraged strong within-party preference flows by indicating the party's candidates, and then tried to influence between-party preferences with the recommendation of further preferences. Some Tasmanian parties still recommended within-party preference orderings.

Further Developments with Hare-Clark

When PR-STV was introduced for the Senate in 1949, a Senate ballot paper was very similar to the Hare-Clark ballot paper used in Tasmania since 1941. The development of the Senate ballot paper after 1949 is more relevant to the Victorian Legislative Council's electoral system, but it is worth explaining the different development path taken by the Hare-Clark ballot.

In 1941 Tasmania adopted the Senate's newly introduced horizontal ballot paper, but continued to list each group's candidates in alphabetic rather than party determined order. A ballot draw was used to determine the order of columns. Party names were not printed on the ballot paper until the 1970s.

Concerns about linear voting, where voters completed a preference sequence straight down a party's column, encouraged calls for the order of candidates in each group to be randomised. Known as Robson Rotation after its chief advocate, Liberal MP Neil Robson, it distributed linear votes across a party's candidates by giving each candidate an equal number of ballot papers where their name was printed first in the party's column. The number of variations was later increased to further randomise preferences.

Robson rotation was first used in Tasmania in 1980, implemented by Labor MPs trying to undermine a decision of the Labor state executive to issue a how-to-vote with a recommendation of preferences. In later years, the distribution of how-to-votes was also banned. The ACT adopted Hare-Clark in 1995.

Hare-Clark with Robson rotation has proved effective for Tasmanian and ACT lower house elections. It is a system that maximises choice between candidates of a party, and party candidates actively campaign against each other. The quota for election in both Tasmania and the ACT is about 10,000 votes, and as the only chamber elected on polling day, there is plenty of opportunity for candidates to attract the attention of voters.

It remains doubtful that such a system would be nearly as effective if used for the much larger electorates used for state and federal upper house elections. The quota for election to the Victorian Legislative Council is more than 70,000, and it is difficult to see how candidates can attract attention in a campaign with so much focus on higher profile contests in lower house seats.

Antony Green – Victorian EMC Submission 5

The Development of the Senate's Electoral System

Senators have always been elected from multi-member state-wide electorates, but the form of the ballot paper and method of voting have undergone major change. Numbering preferences was introduced in 1919 to align with the preferential voting for House elections, and became full preferential voting in 1934. Candidates were first grouped by party in 1922 on a vertical ballot paper, switching to a horizontal ballot paper in 1940. While candidates were grouped by party, party names were not printed. The 1940 changes also allowed a party to determine the order its candidates were listed, a decision that became important after 1949. The switch to PR-STV in 1949 involved no change to the ballot paper, merely a switch from a majoritarian to a proportional counting system. PR-STV for the Senate was adopted with full preferential voting, despite objections from Tasmanian Senators, and despite evidence of high informal voting when used in NSW in 1920.

While Australia's persistent high rates of informal voting have in part been due to , full preferential voting has also played a role, especially at Senate elections. The first graph on the next page plots the rate of informal voting at Senate elections since the adoption of numbered ballot papers in 1919, and shows the impact of changes to the ballot paper structure and method of voting in 1984. Between 1919 and 1983 the rate of informal voting averaged 9.1 per cent per election. This decreased to an average of 3.5 per cent after the new ballot paper was introduced in 1984. (Why informal voting was so high pre-1984 can be understood by examining the Labor how-to-vote for the 73-candidate NSW Senate race in 1974 in Appendix 5.)

The new Senate system introduced in 1984 split the ballot paper horizontally, a thick black line dividing the ballot paper into an area for party votes ‘above the line’ and votes for candidates ‘below the line’ (See example in Appendix 2). Full preferential voting was retained by requiring electors voting ‘below the line’ to number every square, and also by parties having to number all squares on their lodged group voting tickets. If voting 'above the line', a elector could only choose a single party square, the vote being implied to have that party's full list of preferences.

While the new 1-only option reduced the level of Senate informal voting, it created confusion for voters. The 1984 election saw a surge in House informal voting as many voters repeated their Senate vote with an informal 1-only House vote. This pattern of an increase in lower house informal voting following the introduction of a divided ballot paper and group voting tickets was repeated in each state. Avoiding a return to high rates of Senate informal voting played a role in the design of the changes to the Senate’s electoral system introduced in 2016, particularly in the provision of generous savings provisions. (See the second graph on the next page for House and Senate informal voting since 1949.)

6 Antony Green – Victorian EMC Submission

Antony Green – Victorian EMC Submission 7 Using the new Senate system, around 98 per cent of major party voters and 90 per cent of minor party voters voted above the line. The asymmetry in effort between one number above the line and full preferences below herded voters above the line, a problem that worsened as the number of candidates and parties contesting Senate elections increased. The average number of groups per state rose from 7.7 in 1984 to 33.7 in 2013. Printing technology restricted Senate ballot papers to 1.04 metres wide, requiring the Australian Electoral Commission (AEC) to shrink font sizes in response to the increasing number of groups. In 2013 the AEC began issuing magnifying sheets with ballot papers.

The division of the ballot paper was clearly an effective solution to informal voting, but the new system had political consequences by delivering parties greater control over between-party preferences. Before group voting tickets, parties sought to influence between-party preference by distributing how-to-vote material outside polling places. Group voting tickets granted parties control over the distribution of preferences for all votes completed above the line, resulting in parties recommending only an ‘above the line’ vote. The system also allowed minor and even micro parties to control between-party preferences for the first time, opening a new market for trading preference, inducing talk of the system being ‘gamed’.

To assess the impact of between-party preferences, it is best to compare PR-STV results with likely outcomes under a non-preferential system of proportional representation system that assigns final seats based entirely on party first preference vote share.

Table 2 on the next page compares the outcome of Senate elections since 1977 with outcomes had the same vote shares been used to allocate seats to parties using a List Proportional Representation system (List-PR). The analysis uses the same Droop quota used for the Senate, but instead of using preferences to allocate the final seats in each contest, it allocates final seats non-preferentially using a highest remainder method.

The difference between PR-STV and a List-PR model can be measured by the number of parties that elected seats from trailing positions, that is by preferences allowing the party to pass the party with the highest remainder. It is also possible to count the number of parties passed during the process of electing trailing winners.

As an observation, the stronger the flows of between-party preferences, the more that PR-STV will diverge from a List-PR for an election with the same party vote shares. The weaker the flows, the more PR-STV outcomes will match List-PR outcomes.

The columns in Table 2 on the next page are explained as follows.

• Filled quotas – the number of Senators elected through complete quotas filled on the total of first preference votes by party;

• Highest remainder – the number of Senators elected from the highest remainder or leading partial quota on first preference votes;

• Trailing wins – the number of Senators elected after having trailed the party with the highest remainder on the initial counts; and

• Parties passed – the number of higher polling parties passed by trailing winners.

8 Antony Green – Victorian EMC Submission Table 2 - Senate Results Compared to List-PR with Highest Remainder Allocation

Highest Election (Seats) Filled Quotas Remainder Trailing Wins Parties Passed 1977 (30) 24 5 1 1 1980 (30) 25 4 1 1 1983 (60) 54 3 3 2 1984 (42) 33 7 2 2 1987 (72) 62 6 4 6 1990 (36) 28 6 2 2 1993 (36) 30 4 2 3 1996 (36) 28 7 1 1 1998 (36) 24 7 5 5 2001 (36) 25 6 5 6 2004 (36) 29 6 1 3 2007 (36) 27 8 1 2 2010 (36) 27 7 2 4 2013 (36) 21 6 9 33 2016 (72) 52 18 2 7 2019 (26) 22 14 .. .. Source: Calculations by author. Excludes Territory Senate races.

Under group voting tickets at the 2013 election, a quarter of the Senators elected were from trailing positions, and the ratio of parties passed to trailing wins was much higher than at any previous election. In Western Australia, Wayne Dropulich of the Australian Sports Party was elected despite the party polling just 0.23 percent or 0.016 quotas. The Sports Party finished 21st of 27 parties on first preferences, but received ticket preferences from 20 different parties to achieve a quota, 15 of those parties having polled a higher first preference vote. Preferences allowed Dropulich to leap frog parties and defeat a Labor candidate who began the count with a remainder of 0.86 quotas. In Victoria, and the Australian Motoring Enthusiasts Party began the count with 0.51 percent or 0.035 quotas, receiving preferences from 22 other parties including nine with more votes, and passed a Liberal candidate who began the count with 0.81 quotas.

Over three decades, candidates and parties had learnt the rules of the game. A so-called micro-party alliance used the tactic of ‘preference harvesting’, where minor and ‘micro’ parties ignored ideology and instead strategically directing preferences to each other ahead of all larger parties. With preferences accumulated into the pool by group voting tickets, victories for candidates such as Muir and Dropulich became possible. Comparing tickets to below the line votes for the same party revealed very different preference decisions.

Electoral Changes at the 2016 Election

Immediately after the 2013 Federal election, the parliament’s Joint Standing Committee on Electoral Matters (JSCEM) launched an inquiry into the Senate’s electoral system. The Coalition, Labor and Green members came to a unanimous conclusion on reform, recommending the system adopted for the NSW Legislative Council following the 1999 NSW Legislative Council election. The proposals languished until early 2016 when they were revived as part of the ’s planning for a . On 22 February 2016 the Prime Minister announced the government’s intent to move on reform. After negotiations the final legislation included the following features:

Antony Green – Victorian EMC Submission 9 • The divided Senate ballot paper was retained but group-voting tickets were abolished;

• JSCEM’s recommendation for optional preferential voting above the line was replaced by instructions that voters should complete at least six preferences;

• After first proposing to retain full preferential voting below the line, JSCEM’s original limited preferential voting proposal was adopted with instructions that voters should number at least 12 squares;

• Generous savings provisions were included. Any above the line vote with at least a valid first preference was to be formal, while a below the line vote required at least six valid preferences; and

• JSCEM’s proposals to toughen party registration were not pursued, but a proposal to print party logos on ballot papers was adopted.

Having supported the original JSCEM recommendations, the Labor Party opposed the legislation. Former JSCEM member and Labor’s spokesman on electoral reform, , spoke in the House in support of the changes before voting against the bill with his party. In the Senate the Greens would not permit the government to use ‘guillotine’ procedures to close debate, allowing Labor to engage in an overnight filibuster before the bill was finally passed.

Criticism of the bill centred on the number of exhausted preferences that the reforms would produce and how this would disadvantage smaller parties. It was argued the legislation was an example of cartel party behaviour, incumbent parties combining to disadvantage new entrants, and that vote exhaustion would be the new way electors could be disenfranchised. Leading Labor’s to the bill, Senator Penny Wong spoke of three million votes exhausting, based on an assumption that most voters would continue with the ‘1’ only option in use for three decades. A constitutional challenge against the changes was launched, but the High Court dismissed the case.

The changes created substantial logistical difficulties for the AEC. The new system required all ballot papers to be optically scanned and data entered to create electronic versions, and the preference distribution software was upgraded to cope with a twenty-fold increase in records. Confusion arose over the difference between the ballot paper instructions for six preferences above the line against the savings provision allowing a single ‘1’ to count. The AEC engaged in an extensive publicity campaign on the new rules and ballot paper issuing officers were trained to state the new instructions. The only negative was reports of longer queues at polling places caused by voters taking longer to complete their Senate ballot papers.

How Voters Reacted to the New Senate System in 2016

Following the election, the AEC released an electronic dataset of all formal ballot papers. Table 3 categorises this data based on whether ballot papers were completed above or below the line, and based on how many valid above the line preferences were completed. (Similar data for the 2019 election is not yet available.)

10 Antony Green – Victorian EMC Submission Table 3 - 2016 Senate election - Formal Ballot Papers Categorised by Method of Completion and Number of Preferences

Percentage of all Ballot Papers Marked as Above the Line by Number of Preferences Below State 1 2-5 6 7-12 >12 the Line NSW 4.7 4.1 80.9 4.3 0.6 5.4 VIC 2.4 3.6 83.5 4.5 0.8 5.3 QLD 2.0 3.3 83.2 4.5 0.8 6.1 WA 2.2 3.4 83.5 4.2 1.2 5.5 SA 2.3 3.0 79.2 5.2 1.7 8.5 TAS 1.1 2.2 61.1 5.0 2.5 28.1 ACT 1.3 1.8 70.6 11.1 .. 15.2 NT 2.3 2.8 50.8 35.5 .. 8.6 National 3.0 3.6 81.2 4.8 0.8 6.5 Note: Calculated by author from ballot paper data released by AEC. The number of groups was substantially smaller in the two Territories, 10 groups on the ACT ballot paper and seven in the Northern Territory.

By party, Labor at 81.8 percent and the Coalition 86.4 percent had a higher rate of 1-6 above the line voting compared to 75.6 percent for other parties. The reverse was true for below the line voting, respectively 5.3, 4.1 and 10.1 percent. The standout entries in Table 6 are for Tasmania and the ACT, where familiarity with the Hare-Clark variant of PR-STV to elect local parliaments encouraged more voters to delve below the line. That both jurisdictions are considerably smaller meant that voters had more knowledge of individual candidates.

The rate of informal voting rose from 3.1 percent to 3.9 percent. Of the informal votes, 64.1 per cent were blank ballots, 17.4 percent were informal for having multiple first preferences above the line, and another 13.5 percent informal for not having at least six preferences below the line. A total of 908,305 ballot papers were marked above the line with fewer than six preferences. Were it not for the generous savings provisions, the rate of informal voting would have reached the high levels seen before the 1984 ballot paper reforms.

Experience with Hare-Clark at state elections has always encouraged a higher proportion of Tasmanian voters to vote below the line. With the 2016 reforms removing the requirement to number every square below the line, many more Tasmanians decided to pick and choose candidates.

While intra-party preference flows remained strong in Tasmania, both Labor and Liberal voters below the line were discerning in their choice of candidate. Both parties had chosen to demote sitting Senators, and voters responded by re-arranging the party tickets. Twelve percent of Liberal voters first preferenced the demoted Richard Colbeck, and 18 percent of Labor voters did the same for the demoted Lisa Singh. Richard Colbeck was unable to overcome the advantage received by higher placed Liberal candidates from ticket preferences, but Lisa Singh was re-elected.

Lisa Singh had been demoted to the unwinnable sixth spot, but attracted 0.795 of a quota in her own right. She not only defeated her replacement in fifth position, but was elected before the fourth ranked Labor candidate. In an example of friends and neighbours voting, Lisa Singh polled 32.2 percent of the Labor vote in her former state seat of Denison, and 24.1 percent in neighbouring Franklin. For the first time in half a century voters elected a Senator out of order from the party ticket, a warning to party power brokers to pay more attention to the views of voters in the smaller states.

Antony Green – Victorian EMC Submission 11 Changes to Nominations under the Senate's New Electoral System

The 2016 election was a double dissolution election where 12 Senators per state were elected rather than six. The quota for election was reduced from 14.3% to 7.7%. There had not been a double dissolution election since 1987, so it was difficult to compare the 2016 result under the new voting system with previous elections.

The real test came with the 2019 election, a half-Senate election that would allow the results to be compared with the last half-Senate election under group voting tickets in 2013.

Between 1984 and 2013, parties had learnt how to use the system to their advantage. Over time more and more minor parties nominated for election, secure in their ability to exchange preferences and aggregate their votes together via group voting tickets. The rapid increase in ballot paper groups after 2007 was a sign of how minor parties were using the group voting ticket system. The new Senate system ended guaranteed preference flows and added a disincentive for like- minded parties to compete against each other. The average number of groups per state declined from 33.7 in 2013 to 24.5 in 2019. The number of groups nominating per state and territory since 2007 is shown in Table 4.

Table 4 - Groups contesting Senate election, by State

YEAR NSW VIC QLD WA SA TAS ACT NT TOTAL 2007 25 23 24 21 19 11 8 5 136 2010 32 21 23 22 18 10 4 6 136 2013 44 39 36 27 33 23 13 12 227 2016 41 38 38 28 23 21 10 7 206 2019 35 31 26 23 16 16 7 9 163 Source: AEC nominations.

Some of the decline in groups nominating is caused by the decline in the number of parties contesting every State. Previously, minor parties might have concentrated on one State, but nominated candidates in other States to increase the pool of minor party votes. With the end of group voting tickets, the significant decline in groups at the 2019 election is due to fewer parties contesting every state. Table 5 tallies how many States parties contested.

Table 5 - Number of smaller parties contesting one or more States, 2013 and 2019

Number of Smaller Parties contesting States Contested 2013 2019 One State 9 15 Two States 5 6 Three States 4 9 Four States 6 1 Five States 7 1 Six States 14 10 Source: AEC nominations, calculations by author. Totals do not include Labor, the Coalition or Greens.

Where in 2013 there were 27 minor parties that contested four or more States, in 2016 there were only 12. This withdrawal partly explains the decline in groups contesting. Where previously small parties were encouraged to contest every State to build the pool of minor party vote for preference harvesting, the tactic was self-defeating with the abolition of group voting tickets.

12 Antony Green – Victorian EMC Submission Turning Votes into Seats Under the Senate's New Electoral System

The impact of preferences on Senate results was demonstrated in Table 2 that compared past outcomes under Senate PR-STV with an alternative non-preferential system. The new Senate system weakened party control of between-party preferences by abolishing group voting tickets and increased the proportion of ballot papers that exhausted their preferences through limited preferential voting. Taken together these changes weighted the system in favour of parties that started the count with a highest remainder over parties that relied on preferences to win from a trailing position.

The hypothesis that the Senate's new electoral system weights the allocation of final seats in favour of highest remainders can be tested using the results of the 2013 and 2019 half-Senate elections. The results of both elections are compared with outcomes had a List-PR with a highest remainder method been used, the approach used earlier in compiling Table 2.

A summary of this comparisons is provided in Table 6. It shows that nine Senators were elected from trailing partial quotas in 2013, but no trailing parties were elected under the new system in 2019. With the new electoral system having weaker party control over preferences and a greater number of exhausted preferences, senate voting with voter control over preferences behaves more like List-PR than the Senate system using group voting tickets.

Table 6 - Comparing PR-STV and simulated List-PR results

2013 Senators Elected 2019 Senators Elected Filled Quotas 21 22 Highest Remainder 6 14 Trailing Wins 9 .. Source: Calculations by author. Excludes the four Territory Senators.

More detail on the 14 Senators elected from leading partial quotas in 2019 is shown in Table 7. On the left are the partial quotas and parties for successful Senators, while on the right are the highest polling unsuccessful candidates and parties.

Table 7 - 2019 election – success from partial quotas State Successful Party and Partial Quota Unsuccessful Party and Partial Quota NSW 0.70 LIB 0.61 GRN 0.35 ONP VIC 0.74 GRN 0.51 LIB 0.20 ONP 0.20 DHJP QLD 0.72 LNP 0.72 ONP 0.70 GRN 0.58 ALP WA 0.93 ALP 0.86 LIB 0.83 GRN 0.41 ONP SA 0.76 GRN 0.65 LIB 0.34 ONP TAS 0.88 GRN 0.62 JLN 0.24 ONP Source: AEC election results, calculations by author. Minor party codes are DHJP – Derryn Hinch Justice Party, JLN – , ONP – 's One Nation.

Antony Green – Victorian EMC Submission 13 Of the 15 parties with a partial quota above 0.5 at the start of the count, only Labor in Queensland failed to win a seat. That was a contest where four parties started with more than 0.5 partial quotas in a race for three seats. In the other five States, the partial quota for the next party in order, in each case One Nation, ranged from 0.20 quotas to 0.41 quotas, all well short of the partial quota of the successful sixth party.

It was a very different pattern in 2013 where every State except Queensland saw candidates elected from trailing partial quotas, shown in Table 8 by underlined text. In Victoria the ratio of the lowest elected party to the highest defeated candidate was 0.04 quotas to 0.81, and in Western Australia 0.02 quotas to 0.86.

Table 8 - 2013 election – success from partial quotas State Successful Party and Partial Quota Highest Unsuccessful Parties and Quotas NSW 0.67 LDP 0.39 L/NP 0.55 GRN VIC 0.76 GRN 0.04 AMEP 0.81 LIB QLD 0.90 LNP 0.69 PUP 0.42 GRN WA 0.74 LIB 0.66 GRN 0.02 SPRT 0.86 ALP 0.35 PUP 0.35 NAT SA 0.92 LIB 0.50 GRN 0.26 FFP 0.74 XEN 0.59 ALP TAS 0.82 GRN 0.46 PUP 0.63 LIB Source: AEC election results, calculations by author. Underlined text indicates parties successful from trailing partial quotas. WA result based on the original 2013 Senate result, not on the 2014 re-election. Minor party codes are, LDP – Liberal Democrats, AMEP – Australian Motoring Enthusiasts Party, PUP – Palmer United Party, FFP – Family First, XEN – team, SPRT – Australian Sports Party.

Table 9 below categorises all successful and unsuccessful groups contesting the 2013 election under group voting tickets against the new system at the 2019 election, categorising all parties based on their initial partial quota.

Table 9 - Successful and Unsuccessful Groups based on Initial Partial quotas – 2013 and 2019 Senate elections Number of Successful and Unsuccessful Groups 2013 Election (Old Senate) 2019 Election (New Senate) Initial Partial Quota Successful Unsuccessful Successful Unsuccessful 0.9 to < 1.0 (14.3%) 2 .. 1 .. 0.8 to < 0.9 (12.9%) 2 2 3 .. 0.7 to < 0.8 (11.4%) 2 1 4 .. 0.6 to < 0.7 (10.0%) 3 1 5 .. 0.5 to < 0.6 (8.6%) .. 2 1 1 0.4 to < 0.5 (7.1%) 2 1 .. 1 0.3 to < 0.4 (5.7%) 1 2 .. 2 0.2 to < 0.3 (4.3%) 1 8 .. 4 0.1 to < 0.2 (2.9%) .. 9 .. 27 < 0.1 (1.5%) 2 159 .. 100 Source: AEC results, calculations by author. Based on the original 2013 senate election in WA, not the 2014 re- election.

14 Antony Green – Victorian EMC Submission As noted earlier, 14 of the 15 parties that began the 2019 count with a partial quota above 0.50 were elected. This contrasts starkly with the 2013 result under group voting tickets. In 2013 there were six parties that were unsuccessful having started the count with more than 0.50 partial quotas. There were six parties successful after starting with a partial quota under 0.50 quotas. The most notorious cases were Ricky Muir in Victoria and Wayne Dropulich in Western Australia who leap-frogged to victory from less than 0.1 quotas.

None of this analysis means that a party can't win from a trailing position. At the 2016 Senate election, two parties were successful from trailing positions, Family First in South Australia and a second One Nation candidate in Queensland. The last three NSW Legislative Council elections have seen trailing parties win the final seat despite high rates of preference exhaustion.

However, the weaker flow of between-party preferences under the new Senate system means that the gap between leading and trailing parties must be narrower for a trailing party to win. From the data in Table 8, the third Liberal candidate in Victoria at the 2013 Senate election started the count with 20 times the vote of the Motoring Enthusiasts Party, and in WA, the second Labor candidate had 40 times the vote of the Sports Party. A comparable result in a lower house single member seat would be a candidate on 43% being passed on preferences by a candidate that began the count on 2%.

This analysis confirms that the new Senate system weights the allocation of final seats in favour of parties with the highest remainders on first preferences. This is not the same as advantaging parties with the highest first preference vote. After the allocation of seats to filled quotas, the remaining partial quotas of parties polling above a quota must compete with the initial partial quotas of parties that polled less than a quota on first preferences. Who wins the final seats is then a battle of preferences.

Group voting tickets permitted parties to trade preferences in the race to fill final seats. The new Senate system hands the power over between-party preferences back to voters. At the first two elections under the new Senate system, voters made very different decisions on preferences compared to the complex group voting tickets previously lodged by parties.

The clearest between-party trend in preferences in the new system is that voters are more likely to direct preferences to parties they know over parties they don't know. A party that polls poorly on first preferences tends to attract fewer preferences, while parties with a profile high enough to attract a significant first preference vote, also tend to attract more preferences from excluded parties.

Antony Green – Victorian EMC Submission 15 Lessons for Victorian Legislative Council Elections from the Senate Experience

When proportional representation was adopted for Victorian Legislative Council, staggered terms were abandoned, and the state was split into eight regions, each region electing five MLCs. The Senate's group voting ticket system was adopted with only one significant difference. For a formal below the line vote, a voter only needed to complete five preferences, not the full sequence of preferences required at Senate elections.

In Table 2 (page 9), data was provided comparing results under PR-STV versus possible results under a non-preferential List-PR system. Table 10 repeats the analysis using Victorian Legislative Council results 2006. An extra column has been to include the average number of groups contesting regions.

Table 10 - Legislative Council Results Compared to List-PR with Highest Remainder Allocation

Average Groups per Highest Parties Election Region Filled Quotas Remainder Trailing Wins Passed 2006 8.6 30 8 2 5 2010 7.1 30 8 2 2 2014 16.5 27 5 8 27 2018 18.3 26 4 10 47 Source: Calculations by author based on VEC results.

As was the case with Senate elections under group voting tickets, the later elections in 2014 and 2018 show a significant increase in groups contesting. The 2014 and 2018 elections took place after the tactic of preference harvesting came to public attention at the 2013 Senate election.

The column labelled "Trailing Wins" follows the same trend as in Table 2, with an increase in Trailing Wins and Parties Passed as the number of contesting groups increased. At the 2013 Senate election, one quarter of Senate seats (9 of 36) were won by trailing candidates, as was the case at the 2018 Victorian Legislative Council election (10 of 40). There were 27 groups passed by trailing candidates at the 2014 election, 47 in 2018, comparable with 33 at the 2013 Senate election..

Table 11 repeats the Senate analysis in Table 8 (page 14) comparing the partial quotas of all successful candidates against the partial quotas of the highest polling unsuccessful candidates. It displays the same pattern evident at the 2013 Senate election with low polling parties leap-frogging much higher polling competitors to win finals seats.

In Eastern Metropolitan Region, the began the count with 0.04 quotas and defeated the Greens with 0.54, more than 13 times as many votes. In South-Eastern Metropolitan Region, the Liberal Democrats started with 0.05 quotas and defeated the second Liberal candidate with 0.74 quotas, fifteen times as many votes. In Southern Metropolitan region, the Party with 0.08 quotas defeated the Greens who began the count with 10 times as many votes on 0.81.

16 Antony Green – Victorian EMC Submission Table 11 - 2018 Victorian Legislative Council election – success from partial quotas State Successful Party and Partial Quota Highest Unsuccessful Parties and Quotas Eastern Metro 0.04 Transport Matters 0.54 Greens Eastern Vic 0.30 Shooters Fishers Farmers 0.40 Greens Nrthn Metro 0.99 Liberal 0.56 Labor 0.20 Reason Nrthn Vic 0.91 Labor 0.87 Liberal/National 0.29 Justice Party 0.47 Shooters Fishers Farmers 0.23 Liberal Democrats 0.39 Greens S-E Metro 0.99 Labor 0.74 Liberal 0.05 Liberal Democrats Sthn Metro 0.08 Sustainable Australia 0.81 Greens Western Metro 0.77 Labor 0.52 Greens 0.41 Justice Party Western Vic 0.27 Justice Party 0.79 Liberal/National 0.17 Animal Justice 0.45 Greens Source: VEC election results, calculations by author. Underlined text indicates parties successful from trailing partial quotas.

Table 12 repeats the analysis of Senate results used in Table 9 (page 14) across all four Legislative Council elections under the current system.

Table 12 - Successful and Unsuccessful Groups based on Initial Partial quotas – 2006 to 2018 Victorian Legislative Council Elections Number of Successful and Unsuccessful Groups 2006 2010 2014 2018 Initial Partial Quota Win Lose Win Lose Win Lose Win Lose 0.9 to < 1.0 (16.7%) 3 .. 2 .. 1 .. 3 .. 0.8 to < 0.9 (15.0%) 2 .. 1 1 ...... 2 0.7 to < 0.8 (13.3%) 1 1 3 .. 2 1 1 2 0.6 to < 0.7 (11.7%) 1 1 4 3 2 1 .. .. 0.5 to < 0.6 (10.0%) 2 4 .. 3 2 3 .. 3 0.4 to < 0.5 (8.3%) .. 4 .. 3 .. 6 1 3 0.3 to < 0.4 (6.7%) .. 4 .. .. 1 2 1 2 0.2 to < 0.3 (5.0%) .. 6 .. 5 1 3 4 11 0.1 to < 0.2 (3.3%) 1 8 .. 19 3 35 1 19 < 0.1 (1.7%) .. 31 .. 13 1 68 3 90 Source: VEC results, calculations by author.

At the 2006 election, one MLC was elected from under 0.3 quotas, the DLP winning the final seat in Western Victoria. No MLC was elected from under 0.6 quotas in 2010. As the number of groups increased and the flows of preferences became more labyrinthine, there were five MLCs elected from under 0.3 quotas in 2014. The results were even more remarkable in 2018, with eight MLCs elected from less than 0.3 quotas, three from under 0.1 quotas, and seven candidates with more than half a quota on first preferences being defeated.

Antony Green – Victorian EMC Submission 17 The 2018 Victorian Legislative Council election result displays all the same preference distortions of the 2013 Senate election. The 2013 result undermined the legitimacy of the electoral system and led to the abolition of group voting tickets. This was also the situation that led to reform of the NSW Legislative Council's electoral system ahead of the 2003 election, and the South Australian Legislative Council's electoral system ahead of the 2018 election.

Under the current rules, there is nothing to stop more and more parties nominating for Victorian Legislative Council elections. Group voting tickets hand almost total control over preferences to political parties, allowing them to manipulate preferences in a way they would never achieve if voters had to be persuaded to complete their own preferences. Based on the last two elections, the 2022 Victorian election will see another increase in parties nominating, and ever more complex preference deals.

The best solution is to adopt what has already been adopted for Legislative Council elections in NSW and South Australia, and for Senate elections. That is to abolish group voting tickets, to retain the existing divided ballot paper, and to hand control over between-party preferences back to voters.

18 Antony Green – Victorian EMC Submission A Note on Victorian Informal Voting

As noted earlier, the introduction of above-the-line voting for Senate elections caused an increase in lower house informal voting. As the graph below shows, exactly the same trend has been seen at Victorian elections, with the introduction of the new Legislative Council ballot paper in 2006 coinciding with lower house informal voting rising above the rate in the upper house.

Much of the increase in lower house informal voting has been caused by voters using the upper house 1-only option on their lower house ballot paper. That the upper house instructions are responsible has been shown in the past by the decline in 1-only voting at by-elections, and also by the decline in 1-only voting at House of Representatives elections with the introduction of the 1-6 instructions on Senate ballot papers.

Currently there are too many rules that define why a vote can't count. The savings provisions in the Victorian Electoral Act for lower house elections should be expanded. If a ballot paper has a valid first preference for a candidate who will not be excluded during the distribution of preferences, it seems ridiculous to reject that ballot paper for lacking preferences that will never need to be counted.

This is not advocating for optional preferential voting, merely suggesting that more effort be put into trying to avoid the unnecessary exclusion of ballot papers. We should be allowing more votes to remain in the count for the preferences they have rather than exclude ballot paper because they lack a preference that will not be needed to finalise the count.

Antony Green – Victorian EMC Submission 19 Discussion on Options

What are the advantages of switching to use the new Senate Electoral System for Legislative Council Elections?

The new Senate system puts the power over preferences back into the hands of voters in the same way that voters have control over preferences in lower house contests. As was shown in the analysis of the 2019 Senate election, abolishing group voting tickets rewards parties that campaign for votes over parties that trade in backroom preference deals.

How Many Above the Line Preferences Should be Required?

Instructions on the NSW and South Australian Legislative Council ballot papers instruct voters to mark 1 in a square above the line, and if the voter wishes, to mark further above the line preferences.

The new Senate system instructs voters to mark at least six preferences above the line, but a savings provisions permits any vote with a first preference above the line to remain formal. This allows any vote that was formal under the old electoral system to continue as formal under the new system.

How Many Below the Line Preferences Should be Required?

This number varies depending on the number of vacancies. It is 15 in NSW for 21 vacancies, 12 in the Senate for 6 vacancies (12 at a double dissolution election), and 12 in South Australia for 11 vacancies. Strong savings provisions should be included for whatever number is chosen. Victorian ballot papers currently suggest five preferences and I see no good reason to abandon the current number.

What are the Implications for Conducting the Count?

Ballot papers are easy to count under optional preferential voting. Single '1' ballots can be tallied and entered as a bulk number, as in the current system.

As a guide, at the 2019 NSW Legislative Council election, 69.7% of ballot papers were single ones, 27.7% were above the line votes with preferences, and 2.6% were below the line votes. The 2019 election was the first held after the new Senate system was introduced in 2016. This probably explains why it was the first NSW Legislative Council election where 1-only above the line votes fell below 80%.

The Senate 1-6 system means that all ballot papers have to be optical scanned or data entered, a 15- fold increase on the current data entry rates at Victorian Legislative Council elections.

Should Hare-Clark ballots be adopted?

Tasmania and the ACT use Hare-Clark ballots to elect one chamber, the chamber of government, and there is only one chamber elected on polling day. Candidates actively campaign against each other, and how-to-votes are banned outside polling places.

20 Antony Green – Victorian EMC Submission The Victorian Legislative Council is a second chamber elected on the same day as the much higher profile contest in the lower house. You could not sensibly ban upper house how-to-votes if they are still permitted for the lower house.

Victorian Legislative Council regions have more voters than the state of Tasmania. Tasmania elects five members each from five well known and long-established regional seats with a quota one-seventh what is required for a Legislative Council region. Victorian Legislative Council regions are not well- known.

The idea of introducing Robson Rotation to randomise candidate order is pointless in a system where candidates of the same party do not compete against each other. As well as greatly complicating the printing and counting process, this is just an attempt to randomise votes that by choice are being given to the party ticket rather than individual candidates.

Robson Rotation has been introduced in situations where most votes are for candidates, not where most votes are for the party. Distributing party votes across all candidates on a ticket can have perverse outcomes unless all parties are required to nominate the same number of candidates. Even then, small numbers of voters voting for candidates over parties can have perverse outcomes.

There are many functioning democracies around the world using closed list systems that do not allow choice of candidate. Much of the advocacy of Hare-Clark concerns wanting parties to nominate more candidates than they can elect so that voters can make the choice over who to elect. Even then, parties can undermine the intent of Hare-Clark voting by nominating enough high-profile candidates to win possible vacancies, and lower-profile candidates to pad out the ticket.

Hare-Clark is a candidate-based system for small elections. It does not scale-up in size because it is heavily dependent on voters knowing candidates. It especially won't scale-up well for large upper house electorates elected at joint elections.

For better or for worse, Legislative Council elections are battles between parties, not candidates. There seems no good reason to impose a candidate-based ballot paper structure on an electoral contest that is overwhelmingly party based.

Should the Method of Distributing Preferences be Changed?

Abolishing group voting tickets is the major issue that the Electoral Matters Committee should address. However, as the changes will increase the number of ballot papers that exhaust their preferences, the Committee should engage expert advice on changing several matters concerning the distribution of preferences from surplus to quota votes.

The two most important questions are –

• Should ballot papers that exhaust at the next preference be excluded from the calculation of transfer value? This means that votes that exhaust at the elected candidate will remain with the elected candidate, and only ballot papers with further preferences be distributed. • Should the current Inclusive Gregory method of distributing surpluses be replaced by Weighted Inclusive Gregory, as has already been done for the WA Legislative Council and in NSW Local Government?

I recommend these matters be carefully considered rather than make my own recommendations.

Antony Green – Victorian EMC Submission 21 Recommendations

The following are my recommendations for changes to the Victorian Legislative Council's electoral system. They are based on experience with the reformed Senate system. My recommendations follow on from my discussion of options in the previous section.

Recommendation 1

Victoria should follow New South wales, South Australia and the Commonwealth and abolish group voting tickets for Legislative Council elections.

Recommendation 2

Victoria should retain the current divided ballot paper used for Legislative Council elections.

Recommendation 3

Victoria should permit voters to express preferences for parties above the line, as is now allowed in New South Wales, South Australia and for the Senate.

Recommendation 4

Victoria should adopt the New South Wales and South Australian models of fully optional preferential voting, with only one above the line preference being required, all further preferences being optional.

Recommendation 5

If Victoria chooses to follow the Senate example in recommending a minimum number of preferences, then the saving provision should allow a single '1' above the line to be formal. The principle should be that no vote that has been formal since 2006 should become informal as a result of changing the structure of the ballot paper.

Recommendation 6

The rules for formality in the Legislative Assembly should be examined to allow more ballot papers with incomplete preference numbering to be accepted into the count.

Recommendation 7

The Committee should engage expert advice on changing the formulas for distributing preferences from surplus to quota votes of elected candidates.

22 Antony Green – Victorian EMC Submission Appendix 1: Hare-Clark Ballot Paper Examples – ACT

Below – Explanation of the ACT ballot paper from a voter guide.

Below – Sample ACT Ballot Paper

Antony Green – Victorian EMC Submission 23 Appendix 2: Sample Senate Ballot Paper (Post 2016 changes)

Below – Sample layout of the current Senate ballot paper. The 2016 changes added logos and changed the instructions compared to the pre-2016 ballot paper.

24 Antony Green – Victorian EMC Submission Appendix 3: Ireland Ballot Paper Example

Below – Example Irish ballot paper. Note the alphabetic ordering of candidates, the use of party names and logos, and the inclusion of address and photos.

Antony Green – Victorian EMC Submission 25 Appendix 4: Sample PR-STV How-to-Vote recommendation from 1922 NSW Election

Below – Example of a Coalition how-to-vote from the 1922 NSW election conducted under PR-STV rules. The how-to-vote highlighted Coalition candidates, leaving the voter to choose an order between them, and then suggested an ordering of preferences for other candidates.

Note: When NSW used PR-STV for the first time in 1920, preferences were compulsory and the informal vote rose above 9%. In 1922 voters only needed to number as many preferences as there were vacancies to fill and informal voting fell to 3%.

26 Antony Green – Victorian EMC Submission Appendix 5: 1974 NSW Senate Election – Labor How-to-Vote

Antony Green – Victorian EMC Submission 27 Appendix 6: 1999 NSW Legislative Council Ballot Paper (the 'tablecloth')

Below: A photo of the author holding a copy of the 1999 'tablecloth' ballot paper.

A voter engaging with their ballot paper.

28 Antony Green – Victorian EMC Submission