South Australian

PARLIAMENT RESEARCH LIBRARY

REPORT ON THE RESULTS OF THE SOUTH AUSTRALIAN STATE ELECTION 18 MARCH 2006

by Jenni Newton-Farrelly

Research Paper No 1 of 2007

© 2007 SOUTH AUSTRALIAN PARLIAMENT RESEARCH LIBRARY

Not to be reproduced in part or whole without the written permission of, or acknowledgement to, the South Australian Parliament Research Library.

This Information Paper has been prepared by the Research Service of the South Australian Parliament Research Library. While all care has been taken to ensure that the material is both accurate and clearly presented, the responsibility for any errors remains with the author.

ISSN 0816-4282

TABLE OF CONTENTS

The State election of 18 March 2006

Introduction 1 Overview 2 The election date 3 Voters, turnout and informality 4 The Legislative Council election 4. 1 How the Upper House count works 4. 2 Above-the-line votes and tickets 4. 3 Below-the-line votes 4. 4. The Australian Labor Party 4. 5 The 4. 6 The 4. 7 The Greens 4. 8 Family First 4. 9 IND No Pokies 4.10 The National Party 4.11 Dignity for the Disabled 4.12 One Nation 4.13 The Shooters Party 5 The House of Assembly election 5. 1 Overview and swings 5. 2 The Australian Labor Party 5. 3 The Liberal Party 5. 4 The Australian Democrats 5. 5 The Greens 5. 6 Family First 5. 7 The National Party 5. 8 Dignity for the Disabled 5. 9 One Nation 5.10 Independents 5.11 Summary: the effectiveness of How To Vote cards. 6 Country and city seats 7 Are the major parties till losing market share? 8 Ticket votes in the House of Assembly ballot 9 Proposed changes to the Legislative Council 10 Implications of the State election for the coming Federal election

Attachment 1: How does the Legislative Council voting system work? Attachment 2: Distribution of ALP preferences, 2006 Distribution of Liberal Party preferences, 2006 Distribution of National Party preferences to ALP and LIB, 2006 Attachment 3: Distribution of Democrat preferences to ALP and LIB, 2006 Attachment 4: Distribution of Greens preferences to ALP and LIB, 2006 Attachment 5: Distribution of Family First preferences to ALP and LIB, 2006 Attachment 6: Distribution of Dignity For the Disabled preferences to ALP and LIB, 2006 Distribution of One Nation preferences to ALP and LIB, 2006

Bibliography i

Introduction

The State election of 2006 was interesting for all sorts of reasons. On election night the Tally Room was galvanized by the level of support shown for Nick Xenophon in the Legislative Council ballot, interested to see that Family First would win a second seat and quiet witness to the decline of the Democrats. As the results for the House of Assembly firmed during the night it became apparent that the ALP would not only be returned to government in its own right but would have a comfortable majority. Then it became clear that the ALP would take not only most of the Liberal marginal seats but also several that commentators had not considered to be vulnerable. It was a Rann-slide, they said.

This paper summarises the results of the election. It covers the results in both the Legislative Council and the House of Assembly, for each of the parties. It also looks in some detail at the flow of preferences, and how the parties tried to control this flow through their tickets and How To Vote cards. Parties trade preferences in the Legislative Council election for preferences in that ballot and also in the House of Assembly election, and even when a party is too small to expect to win seats it is useful if it can claim to have delivered the seat to someone else. So this paper looks at who received the preferences, and how much control parties were able to exercise over them.

Then the paper looks at what those results can tell us about elections in South .

In this second section the paper considers several continuing questions: is it still the case that country seats swing less than city seats? Are the major parties still losing market share? And what should we do about ticket votes in the Assembly?

Finally, the paper addresses two very current questions, one State and one Federal. The State question relates to the Premier’s proposal to reduce the size of the Legislative Council and reduce the term of an MLC to 4 years: if we had elected 16 MLCs in 2006, who would have won those seats? The Federal question relates to opinion polls which currently show levels of voter support for the Federal ALP that are comparable with voter support for the State ALP. Which seats would change hands if South Australian electors voted at the coming Federal election in the same way that they did at the State election? Where are the voters who have supported the ALP at State elections but not at recent Federal elections? And what would our Senate vote look like?

i Table 1: Legislative Council and House of Assembly results, 1997 to 2006. Legislative Council House of Assembly First preferences Quotas Seats First preferences 2PP Seats won won won won vote won (No.) (%) (No.) (No.) (No.) (%) (%) (No.)

State election of 11 October 1997 ALP 274,098 30.6 3.67 4 312,929 35.2 48.5 21 LIB 339,064 37.8 4.54 4 359,509 40.4 51.5 23 DEM 149,660 16.7 2.00 2 146,374 16.4 NAT 9,233 1.0 0.12 15,488 1.7 1 Greens 15,377 1.7 0.21 1,910 0.2 UAP 11,920 1.3 0.16 13,569 1.5 IND No Pokies 25,630 2.9 0.34 1 All Others 71,521 8.0 0.96 40,135 4.5 1 IND, 1 IND LIB TOTAL 896,503 100.0 12.00 11 889,914 100.0 100.0 47

State election of 9 February 2002 ALP 306,450 32.9 3.95 4 344,559 36.3 49.1 23 LIB 373,102 40.1 4.81 5 378,929 40.0 50.9 20 DEM 68,317 7.3 0.88 1 71,026 7.5 NAT 4,412 0.5 0.06 13,748 1.5 1 Greens 25,725 2.8 0.33 22,332 2.4 One Nation 16,829 1.8 0.22 22,833 2.4 Family First 37,443 4.0 0.48 1 25,025 2.6 IND No Pokies 11,984 1.3 0.15 SA First 9,567 1.0 0.12 16,902 1.8 All Others 76,633 8.2 0.99 52,678 5.6 2 IND, 1 CLIC TOTAL 930,462 100.0 12.00 11 948,032 100.0 100.0 47

State election of 18 March 2006 ALP 340,632 36.6 4.39 4 424,715 45.2 56.8 28 LIB 241,740 26.0 3.12 3 319,041 34.0 43.2 15 DEM 16,412 1.8 0.21 0 27,179 2.9 NAT 6,237 0.7 0.08 19,636 2.1 1 GREENS 39,852 4.3 0.51 1 60,949 6.5 One Nation 7,559 0.8 0.10 2,591 0.3 Family First 46,328 5.0 0.60 1 55,192 5.9 IND No Pokies 190,958 20.5 2.46 2 D4D 5,615 0.6 0.07 3,974 0.4 Shooters 5,991 0.6 0.08 All Others 29,545 3.2 25,884 2.8 3 IND TOTAL 930,869 100.0 11.62 11 939,161 100.0 100.0 47

SOURCE: my calculations based on data from: SA. State Electoral Office, 1998, Statistical Returns: General Elections 11 October 1997, SEO, . SA. State Electoral Office, 2003, Statistical Returns for the South Australian Elections 9 February 2002, SEO, Adelaide. SA. State Electoral Office, 2006, Results and Outcomes: 3, State Election 18 March 2006, SEO, Adelaide.

REPORT ON THE RESULTS OF THE SOUTH AUSTRALIAN STATE ELECTION, 18 March 2006.

The State election of 18 March 2006

1 Overview For the first time in decades, the election results for the Legislative Council were at least as exciting as the results in the House of Assembly.

In the Legislative Council:

• Nick Xenophon was re-elected. Back in 1997 when he had originally been elected to the Legislative Council, it was the first time that an Independent candidate had been elected to the Legislative Council in his or her own right. Nick Xenophon was a candidate again in 2006, unsure of whether he would be re-elected. On the day, his support was so strong that he was re-elected, the second person on his ticket was elected and it was a real possibility that a third person from his group would have been elected. • Meanwhile the Democrat vote in the Legislative Council was greatly reduced – neither of the two Democrats up for re-election was returned. • ’s first-ever Greens MLC was elected. • A second Family First MLC was elected. • The ALP won 4 seats in the Council, which re-elected or replaced the 3 ALP Members whose terms had expired and also returned to the ALP the position which Terry Cameron had occupied as an Independent since 1998. • The LIB vote in the Council was so low that only 3 LIB MLCs were elected – the lowest number since 1975 when the Liberal Party won only 3 seats (and even then the Liberal Movement won another 2). • The composition of the Legislative Council after this election is ALP 8; LIB 8; IND Nick Xenophon No Pokies 2; Family First 2; DEM 1; Greens 1.

In the House of Assembly:

• There was a swing towards the ALP of landslide proportions (7.7% on a 2PP basis). Of the 7 marginal LIB seats going into the election, 6 are now held by the ALP. The swing was widespread and consistent in direction across the State - 431 of the 47 Assembly seats swung towards the ALP. • With a two party preferred result of 56.8% ALP to 43.2% LIB across the State, the ALP won the majority of the 2PP vote and won the majority of seats, and formed government. The ALP went into the election holding 22 seats and governing with the support of IND and NAT members, and emerged with 28 seats and the ability to govern in its own right. • The Liberal Party went into the election with 21 seats, and emerged with 15, losing 6 seats to the ALP. The Liberal Party now has its smallest cohort in the House of Assembly since 1970. • The NAT Member representing the seat of Chaffey was re-elected, the IND Members representing Mt Gambier and Fisher were both re-elected, and the Member for Mitchell was re-elected as an IND Member. The IND Member representing Hammond did not contest that seat and it returned to LIB hands. • The composition of the Assembly after the election is ALP 28; LIB 15; IND 3; NAT 1.

Table 1 is a comparative summary of the election results in both Houses, since 1997.

1 Chaffey, Flinders, Mackillop and Mt Gambier swung to LIB.

2 The Election Date After several unsuccessful attempts,2 the Constitution Act 1934 was changed in 2001 to fix both the term and the date of election for the House of Assembly. The State election of the 18th of March 2006 was the first election to be held under this new legislation; House of Assembly and Legislative Council elections will now be held on the third Saturday of March every 4 years3 and as a result the next State election will be held on the 20th of March 2010.

3 Voters, turnout and informality There have been more than a million electors on the South Australian roll since the early 1990s; at the 2006 State election 1,055,347 electors were enrolled.4

In 2006 these electors had a lot of choices about where they could vote. There were 696 polling booths across the State, and mobile polling teams visited another 48 locations in remote areas. Then there were teams who visited 242 hospitals and other declared institutions. There was no shortage of places to register a vote, and in addition, people who knew that they would be interstate on the day could lodge a pre-poll vote and people who were unable to attend a booth could lodge a postal vote.

Nonetheless, nearly 8% of the electors on the roll did not attend a polling booth, and another 3% (in the House of Assembly election) or 5% (in the Legislative Council election) attended at the polling booth but either didn’t lodge a ballot paper at all or lodged one that was full of messages or mistakes, or quite blank. This means that about 11-12%, or 1 in every 8 or 9 people on the electoral roll, did not have their choice of candidate recorded as part of the election process.

Given that the whole point of an election is to find out how everyone feels about the candidates, the aim of any electoral authority is to reduce this figure as much as possible, so is this figure acceptable? Table 2 below shows the proportion of electors interstate whose votes have not counted at recent elections.5

Table 2 shows that voters are least likely to waste an opportunity to vote by staying away from the polling booth, at federal elections (and also at the Tasmanian Upper House elections). Whether that is a function of the Australian Electoral Commission’s arrangements and advertising, or perhaps of increased media publicity about the importance of a federal election, is not something we can see from the figures, but it is apparently the case in Victoria as well.6

Table 2 shows that South Australia does relatively well at getting people who are on the electoral roll, to attend at a booth and accept a ballot paper. South Australia also does very well to limit the loss of votes due to informality in the House of Assembly election, where we lost the votes of just 3.3% of electors due to informality at the last election, compared to roughly 5% at the federal election. New South Wales loses a smaller proportion of their votes through informality, and so do Queensland and Tasmania, but these States have variations on an optional preferential voting system so voters do not need to mark a preference for every candidate on the ballot paper, and that will inevitably reduce the number of votes that are informal by virtue of being incomplete (and not according with voting tickets).

2 Gilfillan’s Constitution Act Amendment Bills of 1990; M.Evans’ Constitution (Parliamentary Terms) Amendment Bill 1991; Gilfillan’s Constitution (Parliamentary Terms) Amendment Bill 1993; Elliott’s Constitution (Parliamentary Terms) Amendment Bill 1997; Hanna’s Constitution (Parliamentary Terms) Amendment Bill 1999; and finally Hanna’s Constitution (Parliamentary Terms) Amendment Bill 2001 which was passed after LC amendments were accepted. 3 There is a proviso that the State election can be delayed for up to 3 weeks if that day would follow Good Friday, or if a Federal election is held in March or if there is a State disaster (Constitution Act 1934 s.28(3)). 4 State Electoral Office 2006, Results and Outcomes 3: State Election 18 March 2006, SEO, Adelaide, at p.3. 5 In Tasmania the Lower House is elected using a proportional representation count across the whole of the State so I have used the figures for the 2007 elections of the Legislative Council seats of Nelson, Pembroke, Montgomery, Rowallan and Pemberton, which are STV ballots like our House of Assembly ballot. 6 See Macdonald, C., G. Gardiner and B. Costar, 2007, “The 2006 Victorian State Election”, Victorian Parliamentary Library Research Service Research Paper No. 1 of 2007, at p.34: “Victorians turn out to vote in House of Representatives elections in consistently higher numbers than they do to vote for the Legislative Assembly.”

Table 2: Proportion of electors in each State or Territory whose votes did not count through non-attendance or informality. Proportion of electors whose votes did not count In the Lower House (STV) In the Upper House (PR) through through through through non- Total non- Total attendance informality attendance informality (%) (%) (%) (%) (%) (%)

New South Wales, 2007 7.4 2.6 9.9 7.4 5.7 13.0 Victoria, 2006 7.3 4.0 11.2 7.3 3.3 10.6 Queensland, 2006 9.5 1.9 11.4 South Australia, 2006 7.7 3.3 11.0 7.7 4.8 12.5 Western Australia, 2005 10.2 4.7 14.9 10.0 2.9 12.9 Tasmania (a), 2007 16.9 3.2 20.1 5.1 4.2 9.3 Northern Territory, 2005 19.9 3.0 22.9 ACT 2005 (b) 7.2 2.5 9.7 Federal (in SA), 2004 5.2 5.3 10.5 5.2 3.4 8.6 Federal (across Australia), 2004 5.7 4.9 10.6 5.7 3.6 9.2

NOTES: (a) Tasmania uses proportional representation for its Lower House and Single Transferable Vote for the Upper House, so its LC figures are in the Lower House columns and represent the combined total of votes in the seats of Nelson, Pembroke, Montgomery, Rowallan and Pemberton. Tasmania’s LA figures are in the Upper House columns. (b) The ACT uses proportional representation for its Lower House elections. SOURCES: my calculations from: ACT Electoral Commission, 2006, Electoral Compendium 1989 to 2006, ACT Electoral Commission, Canberra. A Green, 2007, 2007 New South Wales Election; Preliminary analysis, NSW Parliamentary Library Research Service Background Paper No 2/07 C Macdonald, G Gardiner and B Costar, 2007, “The 2006 Victorian State Election”, Victorian Parliamentary Library Research Service Research Paper No. 1 of 2007 SA. State Electoral Office, 2006, Results and Outcomes: 3, State Election 18 March 2006, SEO, Adelaide. www.ecq.qld.gov.au www.electionswa.com.au www.electoral.tas.gov.au www.ntec.nt.gov.au

Our Upper House election appears to have lost the votes of 4.8% of electors due to informality, which seems high compared to other jurisdictions. This is surprising because 95% of Legislative Council votes are above-the-line votes which are almost impossible to make mistakes about, but the State Electoral Office has surveyed ballot papers after recent elections and has found that roughly half of the informal votes in both the Legislative Council and the House of Assembly elections are deliberately informal – they are left blank.7 Table 3 below shows the proportion of informal ballot papers that have been simply blank, over the past 3 elections.

7 SA. State Electoral Office, 1999, Election Report : General Elections 1997, SEO, Rose Park; SA. State Electoral Office, 2003, Election Report for the South Australian Elections 9 February 2002, SEO, Rose Park, at pp. 44-45; SA. State Electoral Office, 2007, Election Report, SEO, Rose Park, at pp. 48-49.

Table 3: Informal ballot papers: proportion that were blank, 1997 to 2006 Legislative Council House of Assembly (%) (%)

1997 n.a. 45.9 2002 44.3 42.3 2006 43.2 50.1

SOURCE:: my calculations from: SA. State Electoral Office, 1999, Election Report : General Elections 1997, SEO, Rose Park. SA. State Electoral Office, 2003, Election Report for the South Australian Elections 9 February 2002, SEO, Rose Park. SA. State Electoral Office, 2007, Election Report, SEO, Rose Park.

The figures in Table 3 are from a survey of informal ballot papers conducted after each recent election by the State Electoral Office. I can only find a similar informality survey in one other jurisdiction – the Northern Territory – and at their 2005 election just 15% of their informal votes were blank. So the number of blank ballot papers in the South Australian election seems startlingly high. It is probable that the cause of such a high rate here in South Australia is a statement on our ballot papers that does not appear in any other State or Territory. Fairly prominently placed on the face of each ballot paper for the Legislative Council or for the House of Assembly is a notice that says: “You are not legally obliged to mark the ballot paper.”

This notice first appeared on ballot papers at the 1985 State election.8 At that same election, the ticket vote provision was introduced to re-incorporate incomplete ballot papers back into the formal count if they accorded with a ticket lodged by a candidate, and this provision will have acted to reduce the informality rate at the same time as the new notice has acted to increase it.

In addition to the voters who lodge blank ballot papers, there is another group of voters who write “marks or messages”9 on the ballot paper but do not mark any preferences. These people account for 12% of the informal LC ballot papers and almost 18% of the Assembly informals, so no preferences at all are marked on 55% of the informal LC ballot papers and nearly 68% of informal Assembly ballot papers.

There is one area where informality is often a result of people trying but failing, and this is below-the-line voting in the Legislative Council. Above-the-line voters only have to mark one preference so it is not surprising that in 2006 99.4% of above-the-line voters could manage to do it properly and their votes counted towards the result. Below-the-line voters attempted a more difficult task and only 88% managed it10 – one in 8 below-the-line voters made a mistake and their choices were not counted towards the result. An informality rate of 12% seems far too high to be acceptable, yet it is a real improvement on 2002 when one in 5 below-the-line voters made a mistake. The improvement seems to be due to the fact that in 2006 voters only had to number 54 candidates (rather than 76 in 2002).

Below-the-line voting is democracy at its most accessible, where voters can choose to determine the order in which they want their preferences to be distributed to every candidate, rather than accepting a party’s list and order of preference distribution. But such a high error rate shows that many people are failing to achieve what they set out to do. It is important that the below-the-line provision not be abandoned simply because some voters make mistakes. Could the voting system be changed to help include their votes in the count? Requiring everyone to vote above-the-line is one solution to the problem, but it would take away the one

8 See Schedules 1 and 2 of the Electoral Act Regulations 1985, South Australian Government Gazette 22 August 1985, pp.525-539. The provision was inserted into the new Electoral Act 1985 by the Democrats, as an attempt to reconcile the LIB proposal to introduce voluntary voting into both houses and the ALP proposal to make voting for the Upper House compulsory for anyone on the roll (voting had formerly been voluntary for the Council) SAPD 2 April 1985: 3739. 9 SA. State Electoral Office, 2007, Election Report, SEO, Rose Park, at pp.48-49. 10 SA. State Electoral Office, 2007, Election Report, SEO, Rose Park, at p.48.

avenue by which voters have an opportunity to express their own preferences in the Legislative Council count, and people who normally vote below-the-line would no doubt object to the loss. Another way of reducing the informality rate of these below-the-line votes might be to allow below-the-line voters to stop after they register a vote for, say, 11 candidates. (This is the system that operates now in the Legislative Councils of New South Wales11 and Victoria12 and in both of Tasmania’s Houses.13) Admittedly, this would effectively be introducing optional preferential voting for one group of people, or one part of the ballot, while officially insisting on compulsory preferential voting, but it would be difficult to object to such a change when our ticket vote provisions14 in the House of Assembly count, also re- incorporate incomplete votes into the count and were introduced for the same reason – to lower the informal vote.

4. The Legislative Council election 4. 1 How the Upper House count works The Legislative Council is elected on the basis of a proportional representation count. Half of the 22 members of the Council are elected every 4 years, and each MLC serves an 8-year term.

Attachment 1 describes, in detail, the two stage count in the Legislative Council election. In the first stage, the first preference of each formal ballot paper is counted, a quota is determined and some Members are elected; during the second stage the parties and candidates with fewest votes are progressively excluded from the count, their votes are distributed to surviving candidates and the final few Members are elected.

Proportional Representation systems are designed to allocate seats to parties or IND candidates in a way which represents voter support as closely as possible.

4. 2 Above-the-line votes and tickets Table 4 below shows that the vast majority of voters - about 96% of voters - vote above-the- line for a party or a group. They mark their Legislative Council ballot paper just once.

Table 4: Above-the-line and below-the-line votes, Legislative Council ballot, 1985 to 2006 Above-the-line votes Below-the-line votes Total (%) (%) (%)

1985 96.2 3.8 100.0 1989 95.7 4.3 100.0 1993 96.2 3.8 100.0 1997 95.7 4.3 100.0 2002 97.1 2.9 100.0 2006 94.6 5.4 100.0 Average 95.9 4.1

SOURCE: my calculations from: SA. State Electoral Office, 2006, Results and Outcomes: 3, State Election 18 March 2006, SEO, Adelaide.

11 See clause 2(1) of Schedule six to the Constitution Act 1902 (NSW) which requires below-the-line voters to mark a preference for15 candidates. 12 Victoria only requires 5 preferences because they only elect 5 MLCs for each region. See s.93A of the Electoral Act 2002 (Vic). 13 Tasmania requires 5 preferences for the Legislative Assembly as 5 Members need to be elected for each region. For the Legislative Council where only 1 Member needs to be elected, the Electoral Act (s.102) requires that voters mark 3 preferences if there are more than 3 candidates, 2 preferences if there are only 3 candidates, 1 preference if there are only 2 candidates. That is, no voter needs to mark more than 3 preferences in the Legislative Council ballot, or more than 5 in the Legislative Assembly ballot. 14 See sections 63 and 93 of the Electoral Act 1985, which allow an incomplete House of Assembly ballot paper to be considered to be a formal vote if the voter’s preferred candidate has lodged a ticket with the Electoral Commissioner, to show that candidate’s own choice of preferences.

Above-the-line voters cannot influence the flow of their preferences during the count – they must accept that if their party of choice is not successful in winning seats (or indeed if some votes are left over after a seat has been won) then their votes will be distributed to other parties and IND candidates in an order which is determined by their party of choice. The mechanism for this is the ticket. Before the election the State Electoral Commissioner requires each party or IND candidate to lodge a ticket which specifies the order in which that party’s above-the-line votes will be distributed during the count.

The ticket must specify the order of preference for every candidate on the Legislative Council ballot paper. So even though parties might dislike the idea of their votes being transferred to a particular group or candidate, they must include these people at some stage in their order of preferences.

Over the years, smaller parties and Independents have evolved a practice of arranging their Legislative Council tickets to first exchange preferences with other minor parties and IND candidates before they will finally send their preferences to the ALP or LIB. The effect of this practice is that in order to win the final seat, the ALP and LIB now need to have a larger proportion of a quota than a smaller party does at the start of the second stage of the Legislative Council count (during which smaller parties are excluded and their votes are distributed to continuing parties or IND candidates). The reason is that the larger parties do not receive preferences during the count, whereas the group of smaller parties do receive preferences from other smaller parties as they are gradually excluded during the count. For example, in 1997 Nick Xenophon won a seat having begun the second stage of the count with only 34% of a quota, because the preferences of many candidates and parties were transferred to him as these smaller groups were excluded, whereas at the same election the Liberal Party was not able to win enough preferences from excluded candidates to convert 54% of a quota into another seat.

The arrangements that the smaller parties and Independents have made in relation to their preferences are not illegal, nor are they operating against the intention of the Electoral Act. After all, the whole point of above-the-line voting is that parties lodge tickets to direct the flow of preferences. The only way for the major parties to break into this situation is for them to win a higher (second or third) place on the tickets of the Greens, Democrats, Family First and No Pokies, but it is not immediately apparent what the major parties might trade in order to win this support. Certainly a promise of ALP or LIB preferences in the House of Assembly ballot would not be seen by these smaller parties as particularly valuable because they would need to have a 20-30% share of the first preferences in an Assembly seat to have any chance of the ALP or LIB vote being counted out.

On this reasoning I think it is likely to remain the case that a smaller party will be able to win a seat with under 40% of a quota but the ALP and the Liberal Party will continue to need a larger portion to win a final seat. It should be made clear that the size of the quota will remain the same for all parties, and anyone with a quota of votes will always win a seat. What is different, is the ability of the smaller and the larger parties to build onto a portion of a quota using preferences from excluded candidates, in order to win one of the final seats.

4. 3 Below-the-line votes To voters, the attraction of making a below-the-line vote is that they can choose the order in which parties or IND candidates should receive their preferences. A below-the-line voter is free to reject a party’s list (the order in which a party lists its candidates), and also a party’s ticket (the order in which a party specifies its preferences will be distributed to other parties and IND candidates).

Table 4 showed that at recent elections, only about 4% of the formal votes in the Legislative Council ballot have been below-the-line votes. These votes actually have little effect on the outcome of the ballot. This is shown by the fact that if we construct a model in which we assume that every vote is in fact an above-the-line vote and preferences flow strictly

according to the tickets that each party or IND candidate lodges with the State Electoral Office, then the model elects the same candidates as were elected in the actual count.15

Although below the line votes have not had much of an effect on the outcome of the ballot, they are interesting for several reasons.

The number of below-the-line votes was about 35,000 from 1985 until 1997 but in 2002 it was much lower – it fell to 27,428. This was largely due to the fact that there were 76 candidates in 2002 and many below-the-line voters were unable to completely number all of the candidates over the full ballot paper and ended up lodging an informal vote. The State Electoral Office did a sample survey of informal Legislative Council ballot papers after the 2002 election and found that one in five below-the-line votes were informal, mostly because of mistaken numbering.16 After the enormity of the 2002 sheet, in 2006 the Legislative Council ballot paper was a reasonable size again. This time below-the-line voters had only to number 54 candidates without making a mistake, and many more below-the-line voters succeeded in making a valid formal vote. What is different about the below-the-line votes in 2006 was that there were so many of them - 50,474, compared to only 27,428 in 2002.

This sudden increase in 2006 was due to an enormous number of people who lodged below- the-line votes for Nick Xenophon. In fact, there were so many below-the-line votes for Nick Xenophon that in 2006 every second below-the-line vote was for Nick Xenophon.

It is very difficult to see why such a lot of Nick Xenophon’s voters chose below-the-line voting. It is examined in more detail in section 4.9 (below), but I suspect that it was partly because these voters disagreed with the list he had chosen, and partly also that many voters think that if they are not voting for a major party they have to vote below-the-line. This last idea is not something that I can show directly from the figures but it is supported by the fact that the Greens, the Democrats, Dignity for the Disabled and IND No Pokies, all received 12 to 15% of their votes from below-the-line votes, whereas below-the-line votes make up only about 2% of the total votes for the ALP and the Liberal Party.

4. 4 The Australian Labor Party The Australian Labor Party (ALP) went into the election with 7 Legislative Council Members, of whom 4 had been elected only in 2002 and still had another 4 years to serve, and 3 had been elected in 1997 and were at the very end of their terms. The ALP aimed to win at least these three seats again, plus another ALP seat that had been won in 1997 by the ALP and had been taken up by their candidate Terry Cameron, but had been alienated from the ALP in 1998 when Cameron left and formed his own party, SA First. The ALP did indeed win 4 seats, and Cameron was not returned.17

Table 5 below shows that the ALP won almost half of all first preference votes for the Legislative Council back in 1985 when the current count was first introduced, lost support in 1989 and then again - badly - at the 1993 State Bank election and has been gradually re- building support in the Legislative Council since then. In 2006 the ALP won 4.39 quotas and might have hoped to convert that 39% of a quota into another seat by winning preferences from excluded candidates; after all Nick Xenophon had won his seat in 1997 with 34% of a quota. But as noted above18 in fact the preferences went to the smaller parties first, and the ALP was restricted to 4 seats.

15 See Attachment 1 “How Does the Legislative Council Voting System Work?” especially the last section: “Can we model the Legislative Council voting syste?” 16 State Electoral Office, 2003, Election Report for the South Australian Elections 9 February 2002 , SEO, Adelaide, at pp.44-45. 17 The relationship between the ALP and Cameron was not close so the ALP preferenced Cameron almost last on their ticket, ahead only of One Nation, and in return Cameron sent his preferences directly to Family First and then, if they should be excluded, to the Democrats, the Greens and the Liberal Party, well before the ALP. 18 See 1.4.2 Above the line votes and tickets

Table 5: ALP share of first preferences and seats won, Legislative Council, 1985 to 2006. ALP share of all Quotas Seats first preferences won won (%) (No.) (No.)

1985 48.0 5.76 5 1989 39.7 4.77 5 1993 27.4 3.29 4 1997 30.6 3.67 4 2002 32.9 3.96 4 2006 36.6 4.39 4

SOURCE: my calculations from: SA. State Electoral Department, 1986, General Elections 1985: Statistical Returns, SA PP146 of 1986-87 SA. State Electoral Department, 1991, Periodical and General Elections 1989: Statistical Returns, SA PP161 of 1991-92 SA. State Electoral Office, 1996, Statistical Returns for General Elections 1993 and By-elections 1994, SEO, Adelaide. SA. State Electoral Office, 1998, Statistical Returns: General Elections 11 October 1997, SEO, Adelaide. SA. State Electoral Office, 2003, Statistical Returns for the South Australian Elections 9 February 2002, SEO, Adelaide. SA. State Electoral Office, 2006, Results and Outcomes: 3, State Election 18 March 2006, SEO, Adelaide.

As noted above the ALP won 4.39 quotas - 4 seats and another 30,338 votes. This bundle of 30,338 votes was not big enough to win the ALP a fifth seat, so eventually it19 had to be transferred to surviving candidates. At that point it was an immensely powerful factor in the count, because it was big enough to determine which other party would win the final seat.

The ALP preferences gave the final seat to the Greens. In some ways that was no surprise because over the past 20 years the ALP tickets have generally favoured the Greens and the Democrats. In 1989 and 1993, for example, the ALP ticket favoured the Democrats before the Greens, but then for the 1997 and 2002 elections the ALP lodged two tickets which differed only in one respect – one sent excess ALP preferences to DEM before the Greens and the other sent excess ALP preferences to the Greens before the Democrats. Having two ALP tickets meant that in practice half of any excess ALP votes would be transferred to the DEM candidate and half to the Greens candidate. But in 2006 the ALP returned to its original strategy and lodged just one ticket for the Legislative Council ballot, and that ticket directed ALP preferences in the following order: Greens, Family First, Dignity for the Disabled, IND Xenophon, DEM, LIB, One Nation. This ticket meant that the Democrats would not receive any surplus ALP votes until the Greens, Family First, D4D and finally IND Xenophon candidates had been either elected or excluded. This 2006 ALP ticket indicates a real change in the relationship between the ALP and the Democrats, and this new policy of giving an advantage to the Greens over the Democrats extended to the House of Assembly ballot too, where the ALP consistently sent their preferences in the various House of Assembly contests to the Greens before the Democrats.

As it happened, in the Legislative Council ballot the Greens won more than twice as many votes as the Democrats, so that by the time the excess ALP votes were distributed the Greens were much closer than the Democrats to winning the final seat. They were so far ahead that even if the ALP had followed its strategy of 1997 and 2002 and lodged two tickets

19 By the time it was distributed to other candidates this bundle had increased to 30,519 votes.

(thereby giving only half their preferences to the Greens), the Greens would still have won the final seat.

And as it turned out, the order on the ALPs Upper House ticket was absolutely crucial for the Greens and the Democrats, and also for IND Xenophon No Pokies, who were all in line to win a third seat in the Upper House. If the ALP had lodged just one ticket and it had preferenced to the Democrats they would have won the final seat, and similarly if the ALP votes had all gone to No Pokies the final seat would have been theirs.20

In summary, in 2006 the ALP’s decision to lodge just one ticket and to preference towards the Greens, brought the Greens into the Legislative Council and meant that the Democrats did not win a seat and the Nick Xenophon’s IND No Pokies group did not win a third seat.

The ALP’s 2006 Legislative Council ticket not only put the Democrats at a disadvantage compared to the Greens, but also put the Democrats at a disadvantage compared to Family First, Dignity for the Disabled and IND Xenophon. This disadvantage was not reflected in the ALP’s tickets and How To Vote cards for the concurrent ballot in the House of Assembly – in that contest the ALP consistently directed their preferences to DEM before Family First, so it seems that the ALP intended to give Family First an advantage just in the Legislative Council ballot. Bennett21 accepts Kelton and Anderson’s view22 that Family First

“made a Legislative Council preference deal with the ALP, wherein it gave that party its upper house second preferences in return for a conscience vote on the same-sex rights Bill that had stalled in the Parliament in the previous year.” Perhaps another part of such an agreement may have been that the ALP would put Family First higher than the Democrats on its Legislative Council ticket. In the end, that proved to be irrelevant as Family First won enough votes in its own right to win a quota before the ALP’s excess votes were distributed. But it provided a safety margin for Family First – my estimate is that Family First could have won half the number of first preference votes it did actually receive, and still could have won a seat based on the ALP preferences23.

About 2% of ALP votes were below-the-line votes. This is quite a small percentage24 but it does represent about 6,500 ALP votes, and below-the-line voters must, almost be definition, be in disagreement with their party over one of two things: the party’s list or its ticket. So parties could pay some attention to these voters.

The order in which each party lists its candidates on the ballot paper will determine the order in which its candidates are elected, and so the list will determine whether an individual candidate is elected or not. Parties also determine the order in which their preferences will flow to other parties or IND candidates, through the party ticket. Above-the-line voters need to accept both the list and the ticket, but below-the-line voters do not – they can determine which candidate will received their vote at each point along the way. So can we see whether ALP below-the-line voters particularly disliked the party’s list or the party’s ticket?

If an ALP voter particularly disliked the ALP list – if the voter’s favourite candidate had been placed low on the list, for example, and looked likely to miss out on a seat – the voter might use the below-the-line option and give a first preference vote to that favourite candidate. That doesn’t seem to have been a major cause of ALP below-the-line voting in 2006. The first candidate on the ALP’s list was Carmel Zollo and she received 89% of the below-the-line votes for the ALP. The male ALP candidates further down the list each received between 1%

20 The result would have needed one more step than in the case of the Greens - if the Democrats had won all of the ALP excess votes, then the DEM total would have been boosted above the IND No Pokies and Greens totals and the lowest of these – IND No Pokies – would have been excluded. The IND No Pokies ticket would have sent half of their votes to the Democrats and half to the Greens and so the Democrats would have won a quota. The same thing would have happened if the excess ALP votes had been transferred to IND No Pokies. 21 Scott Bennett, 2006, “Understanding State Elections: South Australia and Tasmania 2006”, (Commonwealth) Parliamentary Library Research Brief No 17 of 2005-6, p.6 22 “Hard Ball on ‘Bible Belt’ deals” The Advertiser, 2 March 2006. 23 In making this calculation I have just taken half of the Family First votes out of the ballot, i.e. I have not assumed that these “lost” Family First votes would have been won by any other party. 24 By comparison 17.6% of DEM votes were lodged as below-the-line votes.

and 4% of the remaining below-the-line votes; and the only other female candidate received 5%. A second indication that ALP below-the-line voters were fairly happy with the party’s list of candidates, is that figures in the election reports25 show that 89% of the below-the-line votes for Zollo, then preferenced straight down the ALP list to the next ALP candidates (Sneath and then to Wortley). So the party list does not seem to be the reason why the ALP voters chose to vote below-the line.

If it was not the party’s list that below-the-line voters disagreed with, it can only be the ALPs’ ticket that these voters disliked. The only way to prove this would be to compare the ALP below-the-line voters’ ballot papers with the ALP ticket (which means having access to ballot papers and actually finding the right ones now that the count has concluded). But even without concrete evidence it seems irrefutable that about 6,500 voters voted ALP in the Upper House but would have sent their preferences to parties or IND candidates in an order that did not match the ALP ticket. Given that it was the distribution of the ALP excess votes that decided who won the final seat in the Upper House it would be interesting to see who these below-the-line voters preferred. The question for the ALP – indeed for any party – is whether it is a problem to have lost control over the preference distributions of a group of their voters (in this case 2% of them) or whether perhaps this is such a small group that the party is willing to see it as pointing out the remarkable flexibility of the Legislative Council electoral system to allow voters various ways of expressing their choices.

4. 5 Liberal Party Table 6 below shows the LIB first preference vote at each Legislative Council election since 1985. Table 6: Liberal Party share of first preferences and seats won, Legislative Council, 1985 to 2006. LIB share of all Quotas Seats first preferences won won (%) (No.) (No.)

1985 39.3 4.72 5 1989 41.1 4.93 5 1993 51.8 6.22 6 1997 37.8 4.54 4 2002 40.1 4.83 5 2006 26.0 3.12 3

SOURCE: as for Table 5 above

The Liberal Party went into the Legislative Council election with 9 MLCs, having won 4 in 1997 and then 5 in 2002; the terms of the 1997 cohort expired on election day and the Liberal Party hoped at least to win those seats again. But there was a solid swing against the Liberal Party in the Legislative Council ballot and their share of first preferences declined from 40% in 2002 to just 26% in 2006, which translated into 3.12 quotas. No party or IND candidate could win an extra quota on the basis of 12% of a quota so that even with the allocation of preferences from excluded candidates during the count the Liberal Party won only 3 seats. This is the lowest level of support for the Liberal Party since the current count was introduced for the Legislative Council at the 1985 election, and as Table 5 shows above, it is even lower than the ALP’s share at their disastrous election of 1993. Three seats is the lowest number of seats the Liberal Party has won in the Legislative Council since 1975 when the Liberal vote was split between Liberal Party and Liberal Movement candidates.

On the first day of the 2006 session the Liberal Party returned to the Council with just 8 MLCs, 5 elected in 2002 and 3 elected in 2006.

25 My calculations from figures in the unpublished tables showing the Result of the Transfer and Distribution of Preferences, from the State Electoral Office, and Table 2.4 of the SEO’s Results and Outcomes publication for 2006.

Like the ALP, the Liberal Party is not good at winning large bundles of preference votes during the count. In 2006 the Liberal Upper House candidates won 3 quotas plus 9,020 votes that continued into the count, and while this bundle of 9,020 votes was big enough to keep the Liberals in the count until many of the smaller parties and IND candidates were excluded, the LIB group did not pick up any additional votes through these parties’ tickets26 because they had all put each other higher than the Liberal Party on their tickets. So eventually the candidate who was fourth on the LIB list was excluded, and at this point the bundle of LIB above-the-line preferences went to Family First, who won a seat.27

At first glance, it seems silly to say that the Liberal Party gave the seat to Family First, because Family First was in such a strong position having won 60% of a quota just from first preferences. Indeed by the time the LIB preferences were transferred, Family First was only 4000 votes short of their quota. But actually it is quite true that the other parties remaining in the count at that stage (the ALP, the Greens, IND Nick Xenophon and the Democrats) would all have transferred their preferences amongst themselves before any above-the-line votes would have gone to Family First. So it is quite true that if the LIB preferences had not been transferred to Family First, then Family First is unlikely to have won a seat.

The pattern of below-the-line voting by LIB supporters was almost exactly the same as for ALP voters - the vast majority of Liberal voters use the above-the-line voting provisions, only 2% of LIB votes came from below-the-line votes, and about 90% of the LIB below-the-line voters sent their preferences to the second and third candidates on the LIB list. So once again it seems that it was the LIB ticket that the majority of below-the-line voters disliked, not the LIB list of candidates.

4. 6 The Australian Democrats

Table 7: Democrat share of first preferences and seats won, Legislative Council, 1985 to 2006. DEM share of all Quotas Seats first preferences won won (%) (No.) (No.)

1985 5.5 0.66 1 1989 10.7 1.28 1 1993 8.0 0.96 1 1997 16.7 2.00 2 2002 7.3 0.88 1 2006 1.8 0.21 0

SOURCE: as for Table 5 above.

The 2006 Legislative Council election was devastating for the Democrats who seemed vulnerable from the start, entering the election with poor polls and 2 of their 3 MLCs up for re- election. No Democrat candidate won a seat, and the party is now represented by just one Member who will be up for re-election in 2010.

While it is true that the ALP ticket denied the Democrats a seat in the Upper House,28 it is indisputable that the number of votes the Democrats won in their own right was very low in 2006. The Democrats first contested Upper House seats in South Australia in 1979, and in the almost-30 years since then they have averaged just under 9% of first preferences, and never below 5%. In 2006 they polled just 1.8%. This translated into just over a fifth of a

26 The LIB group did of course pick up below-the-line votes throughout the count – and lost some. Net gain for the LIB group was 1,930 through the count. 27 Interestingly, had the LIB votes not been transferred to Family First they would still have won a seat, but on the basis of votes transferred from the ALP rather than from LIB. 28 See section 4.4.

quota, which historically has simply been too low a base from which to win a seat.29 Indeed if the ALP ticket had preferenced to the Democrats before the Greens and the Democrats had then won the final seat, they would have won a seat with the smallest first preference vote on record – 21% of a quota compared to Nick Xenophon’s 34% in 1997.

Following their traditional policy of not aligning with either of the major parties, the Democrats lodged two tickets for the Upper House preferencing to either the ALP or LIB but only after various Independents and then the Greens, and IND No Pokies. Given that the Democrats have always lodged two tickets and have therefore directed half of their preferences to the ALP and half to LIB, they are not in a strong position to bid for a high spot on the ticket of either the ALP or the Liberal Party. The result in 2006 seems to be that the major parties reciprocated and put the Democrats low on their own tickets,30 and the Greens, Family First and even IND No Pokies received preferences before the Democrats. So with a poor start, and receiving little assistance along the way, the Democrats were left behind in the race for the final seat.

Below-the-line voting is much more important for Democrat voters than for supporters of the major parties. About 15% of DEM first preference votes came from below-the-line votes. This is in line with the supporters of the Greens, Nick Xenophon’s No Pokies group and Dignity for the Disabled – each of these groups received between 11 and 15% of their votes from below- the-line votes.

4. 7 The Although Kris Hanna had represented the Greens in the House of Assembly from 2003 to 2006, the first Greens representative in the Legislative Council was elected in 2006. Table 8 below shows the vote for the various Greens parties which have contested the Legislative Council ballot since 1993.

Table 8: Greens share of first preferences and seats won, Legislative Council, 1993 to 2006. Greens share of Quotas Seats all first preferences won won (%) (No.) (No.)

1993 Green Party SA 1.3 0.16 1993 Green Alliance 0.4 0.05

1997 Australian Greens 1.7 0.21 2002 Australian Greens 2.8 0.33 2006 Australian Greens 4.3 0.51 1

SOURCE: as for Table 5 above.

Two Greens parties have contested the Legislative Council election at various times since 1993, winning 1 to 2% of first preferences. The Australian Greens have campaigned since 1997, and finally in 2006 they won 4.3% of first preferences, or just over half a quota. This gave the top Greens candidate a good chance of staying in the count long enough to accrue votes from Independents and minor parties as they were excluded. In fact the Greens were not particularly effective at picking up votes in this stage of the count and Family First and the Democrats received a lot more votes from this process. But when the ALP’s excess votes were distributed they went to the Greens, and in 2006 this was a large bundle of votes, large enough for the Greens to win the final seat.

The Greens won a seat in 2006 because they had enough first preferences, they were high on the ALP ticket and the ALP had a big bundle of excess votes.

29 The lowest successful vote so far has been Nick Xenophon who won a seat in 1997 with 0.34% of a quota. 30 The Democrats were 8th on the ALP ticket and 6th on the LIB ticket in 2006.

Interestingly, although the Greens were high on the ALP ticket, and the ALP ticket was effectively responsible for putting the Greens into the Legislative Council, the Greens’ own ticket for the Legislative Council was not particularly oriented towards the ALP. The Greens LC ticket preferenced to the Democrats first and only then to the ALP. If the Greens want to keep their privileged position on the ALP Upper House ticket it is possible that there might be some pressure from the ALP for a higher position on the Greens ticket. Still, the ALP accrues a more important and more direct benefit from receiving Greens preferences in the Lower House ballot, where the Greens are always likely to be excluded, rather than Greens preferences in the Upper House ballot where Greens preferences are actually not likely to be distributed now that they have greater support and are likely to stay in the count until the end. On this reasoning, as long as the Greens do reasonably well in the House of Assembly ballot, and preference to the ALP in that ballot, they are likely to be in a strong position to occupy a very high spot on the ALP Legislative Council ticket.

Below-the-line votes made up about 14% of the total number of Greens votes in 2006. It is difficult to know why such a lot of Greens (and DEM and D4D and No Pokies) voters use the below-the-line voting option. We can’t see the order these below-the-line voters would have preferred for distribution of their votes because they were not, in the end, distributed.

4. 8 Family First This party first contested the Legislative Council election in 2002 and won 4% of the first preference vote and enough preferences from excluded candidates to win a seat in the Legislative Council. The feat was repeated in 2006, with a small increase in their first preference vote to 5%, which translated into 60% of a quota and enough preferences from excluded candidates to again win a seat in the Legislative Council. There are now two Family First (FF) MLCs.

Table 9: Family First share of first preferences and seats won, Legislative Council, 2002 and 2006. Family First share of Quotas Seats all first preferences won won (%) (No.) (No.)

2002 4.0 0.48 1 2006 5.0 0.60 1

SOURCE: as for Table 5 above.

Family First won over 46,000 first preference votes but was a big winner of preferences at the 2006 election. Family First won 5,600 votes from the Shooters Party, 6,000 from the Nationals, 7,000 from One Nation and finally 9,500 from the Liberal Party. With the transfer from the Liberal Party, Family First reached the quota (77,573) and won a seat.

Interestingly, Family First was also quite high on the ALP ticket – third, after the Greens. It is not immediately apparent why the ALP would have had Family First so high on their Legislative Council ticket, when Family First showed no orientation towards the ALP at all. Family First votes would only have been transferred to the ALP if all of the conservative parties had already been excluded, as well as Dignity for the Disabled, IND Nick Xenophon, and 6 IND candidates. This does seem to confirm the arrangement noted by Kelton and Anderson31 and referred to above, namely that the ALP gave Family First its preferences, “in return for a conscience vote on the same-sex rights Bill that had stalled in the Parliament in the previous year.”

The Liberal Party ticket for the Legislative Council sent their preferences straight to Family First, before any other party or IND candidate. It is strange to see that in return, the Family First ticket ranked the Liberal Party only 12th, but to be fair this was higher than the other

31 Kelton, G., and L.Anderson, 2006, “Hard Ball on ‘Bible Belt’ deals” The Advertiser, 2 March 2006, p.25.

parties which stayed in the count – the ALP, the Greens, Democrats and IND Nick Xenophon No Pokies – so had Family First votes actually been distributed according to the FF ticket, they would have gone to the Liberal Party.

Unlike supporters of the Democrats and Greens, Family First voters did not favour the below- the-line voting provisions. Only 6.1% of Family First votes came from below-the-line voters (and the average for all parties was 5.7%).

4. 9 IND Nick Xenophon No Pokies In 1997 Nick Xenophon stunned commentators by being elected to the Legislative Council as an Independent, campaigning for a reduction in the number of poker machines. In the past, people who had already been elected to the Council had sometimes left their parties and remained in the Council as Independents but no-one had previously won a seat as an Independent. In 2002 it seemed possible that the campaign might win another seat in the Legislative Council but voter support was not high enough to elect either of the two candidates endorsed by the campaign.32 So in 2006 when Nick Xenophon was himself up for re-election there was some doubt about whether he would have the support to return to the Parliament. The Advertiser reported that “his survival is largely dependent on favourable preference deals from the major parties”.33 On the day, he stunned the commentators again with the strength of his support. In 1997 he had won a seat on the basis of 34% of a quota – in 2006 he won 2.46 quotas. Not only did he win a seat, he won two, and then also enough votes to continue in the count to pick up more votes, and until quite late in the count it looked like he might win three seats. Indeed if the ALP ticket had preferenced to IND Nick Xenophon rather than the Greens, he would have won three seats.

Table 10: IND Nick Xenophon’s share of first preferences and seats won, Legislative Council, 1997 to 2006. Share of Quotas Seats all first preferences won won (%) (No.) (No.)

1997 Inependent No Pokies Campaign 2.9 0.34 1 2002 Inependent Nick Xenophon's No Pokies Campaign 1.3 0.15 0 2006 Independent Nick Xenophon – No Pokies 20.5 2.46 2

SOURCE: as for Table 5 above.

The IND Nick Xenophon No Pokies group lodged two tickets, which differed at the 6th and 7th preferences (DEM then Greens, or else Greens then DEM) and also at the 13th and 14th preferences (ALP then LIB, and LIB then ALP). The tickets directed No Pokies preferences to Ralph Clarke (standing as an IND on a platform of buying back ETSA), Dignity for the Disabled, an IND standing on a social justice platform, an IND standing on a no-drugs platform, and then to the Democrats and the Greens, before reaching back into the ranks of the Independents. The ALP and LIB were higher on his ticket than were Family First and One Nation. But on the day, the third IND No Pokies candidate remained in the count until the end, so No Pokies preferences were not distributed.

We have seen that voters for smaller parties are more likely to use the below-the-line voting option and voters for No Pokies were no exception – about 12% of all votes for IND Nick Xenophon No Pokies were below-the-line votes. But because there were so many voters

32 It may also be true that the vote was split by another candidate who ran as “IND No Pokies “ but that candidate only gained 1946 votes. 33 Anderson, L., 2006, “It's envy, says maverick; Knives out as Mr X's tactics anger rivals” The Advertiser, 20 February 2006, at p.3

who supported the No Pokies campaign, this meant that 46% – nearly one in every 2 – below-the-line votes was for Nick Xenophon. We know that many below-the-line voters make mistakes in their ballot papers and their votes become informal, and in 2006 about 46% of these unintentionally informal below-the-line votes would have been for IND Nick Xenophon. If somehow they had been able to be accepted into the count34 he would have won a third seat.

Below-the-line votes do not follow a party’s ticket, so it is probably not surprising that when Nick Xenophon himself was declared elected and his excess votes rolled down to elect Ann Bressington, nearly half of Nick Xenophon’s below-the-line votes did not actually transfer to Bressington but instead went to other candidates – 8% went to the ALP, 10% to LIB, 7% to the Greens and 6% to the Democrats. 35 This is a tantalising hint about where the support for Nick Xenophon came from in the first place, but only a hint, because it only refers to a group of below-the-line voters, not all voters. Still, when it was apparent how well Nick Xenophon had performed and how poorly the Democrats had performed, it was tempting to think that the support for Nick Xenophon might have been drawn from the group of people who would otherwise have voted Democrat. If that had been the case then I would have expected any leakage of below-the-line votes to have strongly favoured the Democrats. These figures show that it did not.

Why did over 23,000 voters choose to give their first preference to Nick Xenophon by voting below-the-line? Were they people who would have voted above-the-line for Nick Xenophon but disliked his list or perhaps his ticket? The IND Nick Xenophon No Pokies ticket would have sent his preferences within his group to Ann Bressington and and then to a series of IND candidates and minor parties, starting with Ralph Clarke. Only 52% of the below-the-line votes for Nick Xenophon went to Bressington and of the 6,68236 below-the-line votes that would have been transferred only 171 went to Ralph Clarke. So it does look like at least half of the below-the-line voters preferred not to send their preferences in the same direction as Nick Xenophon’s ticket.

4. 10 The National Party While two people have represented the National Party (NAT) in the House of Assembly, the National Party has never won a seat in the Legislative Council. Nonetheless, the party contests the Legislative Council election each time and receives between 1/2 and 2% of first preference votes. In 2006 the National Party won 6,237 first preferences, acquired another 181 votes from excluded candidates and was then excluded.

Table 11: National Party share of first preferences and seats won, Legislative Council, 1985 to 2006. NAT share of all Quotas Seats first preferences won won (%) (No.) (No.)

1985 1.6 0.20 0 1989 0.8 0.09 0 1993 0.7 0.09 0 1997 1.0 0.12 0 2002 0.5 0.05 0 2006 0.7 0.08 0

SOURCE: as for Table 5 above.

34 e.g. if the ballot paper could have been included into the count for as far as the vote was formal. 35 Because Nick Xenophon won more than a quota of votes, his excess votes were distributed to other candidates, and the way that this is done is to actually distribute all of his votes at a fraction of their face value. So we can see the second preference for all of his above-the-line and formal below-the-line votes. 36 My calculation from the Result of Transfer and Distribution of Preferences table from the SEO.

There is an interesting relationship between the National Party vote in the House of Assembly and their vote in the Legislative Council: only about one in every three voters who gave a National Party candidate their first preference in the House of Assembly also then gave their Council vote to the Nationals.37 In 2006 the National Party won 19,636 first preference votes in 4 House of Assembly seats, but only 6,237 votes in the Legislative Council ballot across the whole State.

The National Party does not have anywhere near enough support in the Legislative Council to win a seat, but the votes it does win in that contest can be promised in return for support in the House of Assembly ballot. It looks like this happened in 2006: the National Party Legislative Council ticket preferenced to Family First before anyone else, and Family First received 6,015 above-the-line votes from NAT. Then in the House of Assembly election, Family First preferenced to NAT before LIB in Chaffey, and Karlene Maywald received 678 preference votes from Family First voters. Neither of these bundles of transferred votes made the difference between winning a seat and losing it, but both could be seen as useful insurance.

Voters who gave the National Party their support in the Legislative Council ballot were not likely to vote below-the-line – only 4% used these provisions.

4. 11 Dignity for the Disabled Dignity for the Disabled (D4D) contested both the Legislative Council election and also 10 seats in the House of Assembly elections; 2006 was the first election in which they have participated. In the Legislative Council they won 5,615 votes across the State which put them into the ranks of smaller parties which can potentially deliver a useful number of votes to another party – the Shooters Party won only about 400 more votes and the Nationals only about 600 more.

Table 12: Dignity for the Disabled share of first preferences and seats won, Legislative Council, 2006. D4D share of all Quotas Seats first preferences won won (%) (No.) (No.)

2006 0.6 0.07 0

SOURCE: as for Table 5 above.

Almost every party38 placed D4D higher on their ticket than the Democrats, and in most cases also higher than IND Nick Xenophon No Pokies. This is a remarkable degree of accommodation for a group which was new to the scene. In the end it did not have any effect on the count though, as D4D was the smallest of the parties and so could not stay in the count long enough to receive votes from excluded candidates. D4D only received preferences from three IND candidates (and a few below-the-line votes) before being excluded.

When the D4D candidates were excluded, their above-the-line votes were transferred to the Democrats. After the Democrats the D4D ticket preferenced to a range of IND candidates and minor parties before returning to Family First, then Xenophon, then the Greens. The ALP and LIB would not have received any preferences from D4D above-the-line voters while these minor parties were still in the count.

37 In Chaffey, which was the seat with the strongest support for the National Party at the Legislative Council level, Karlene Maywald received 10,358 first preference votes whereas the National Party candidates for the Legislative Council received only 2,627 in total. 38 except the Nationals, the Greens and the Democrats themselves

The below-the-line voting provisions were taken up by about 12% of voters who supported Dignity for the Disabled, roughly the same level as voters who supported the Greens and IND No Pokies.

4. 12 One Nation The 2006 election was not a good one for One Nation. In the Legislative Council they won 7,559 first preferences, less than half the support they were given in 2002, and although they stayed in the count until quite a late stage they picked up very few preference votes. This is because One Nation was high on only a few tickets – Family First, the Shooters Party and an IND – and in each case Family First was even higher, so the votes were intercepted by Family First. The few preferences One Nation did receive were mostly below-the-line votes.

Table 13: One Nation share of first preferences and seats won, Legislative Council, 2002 to 2006. One Nation share of Quotas Seats all first preferences won won (%) (No.) (No.)

2002 1.81 0.21 0 2006 0.81 0.10 0

SOURCE: as for Table 5 above.

Meanwhile One Nation lodged two tickets – both preferenced to the Shooters Party, then to Family First and then much later one preferenced to the ALP before LIB and the second to LIB before the ALP. In fact, by the time that the One Nation votes were distributed, the Shooters Party candidates had already been excluded from the count so the One Nation votes went to Family First.

The vast majority of One Nation voters seem to have been quite happy to accept the One Nation ticket and party list – only 6% of One Nation votes in the Legislative Council were below-the-line votes.

4. 13 The Shooters Party The Shooters Party last contested the Legislative Council elections in 1993 so it was a surprise to see them again in 2006. But their candidates have both contested elections more recently, not only on behalf of the Shooters Party but also for the United : Robert Low contested the Senate election for the Shooters Party in 1996, Elder for the UAP in 1996 and the Legislative Council for the Shooters Party in 2006. His colleague Michael T Hudson contested the Legislative Council election in 1997 for the UAP and the Legislative Council again in 2006 for the Shooters Party. Both the Shooters Party and the UAP seem to be broad churches – Ellis Wayland contested the Legislative Council in 1997 for the UAP and then went on to contest the Senate election for the Nationalist Party in 1998.

Table 14: Shooters Party share of first preferences and seats won, Legislative Council, 1993 to 2006. Shooters share of Quotas Seats all first preferences won won (%) (No.) (No.)

1993 1.17 0.14 0 1997 Did not contest 2002 Did not contest 2006 0.81 0.10 0

SOURCE: as for Table 5 above.

The Shooters Party may have been as disappointed as the Democrats with the size of their vote in 2006. They won 5,991 votes but I suspect they would have anticipated a vote twice that size: in 1993 the Shooters Party won 10,622 votes in the Legislative Council ballot and 8,973 in 1996 for the Senate (and in 1997 the UAP won 11,920 in the Legislative Council ballot.). The Shooters Party did not contest the House of Assembly election.

The Shooters Party preferences to the conservative, small party, side of politics. In 2006 their ticket sent their preferences to Family First, where they stayed. In return, the Shooters Party had a high place on the tickets of One Nation (where they were second) and Family First (where they were ranked higher than DEM and the Greens). The larger conservative parties are not so keen – the Liberal Party and also the National Party would have sent their preferences to the Democrats before the Shooters Party.

In 2006 only 6% of Shooters Party voters used the below-the-line voting provisions in the Legislative Council ballot; similarly 6% of One Nation and 6% of Family First voters voted below-the-line (but 15% of Democrats).

5 The House of Assembly election

5 1. Overview and swings

Like the Legislative Council ballot, the House of Assembly election took commentators by surprise with big swings – the ALP won 6 seats from the Liberal Party, some with swings above 10%. In addition there were solid swings to the ALP in their own seats, probably most appreciated in the ALP marginals.

The pendula attached show the relative strength of the House of Assembly seats going into the 2006 election and then as a result of the election. (The pendulum going into the election uses the 2002 results adjusted to take account of the electorate boundary changes ordered by the EDBC in 2003.) Taken together the pendula illustrate the swings across all seats, and in particular the movement of the LIB seats towards the ALP. Six of the seven most marginal seats held by the Liberal Party before the election are now Labor seats. Of this most- marginal group, only Stuart, the Liberals’ second-most-vulnerable seat, did not change hands.

I have calculated the average swing in various sectors of the pendulum, and these figures are shown next to the first pendulum. I have excluded the 6 seats that did not have a 2PP actual result at both elections.

Marginality is not an agreed theoretical concept, so it is not always clear just which are the marginal seats. Twenty years ago it was common to draw lines across a pendulum at 5% and 10% and say that seats which would change hands with a swing of under 5% were marginal, seats which needed a swing of 10% or more to change hands were safe, and everything else was likely to be interesting to watch on election night. Then the line drawn at 5% changed to 6%. The Australian Electoral Commission adopted that change, and the State Electoral Office and the Electoral Districts Boundaries Commission have generally followed the AEC’s example, although the Commission seems to use both 5% and 6%.39 Meanwhile the parties still work on 5% and 10%. Academics are split. Antony Green and the ABC use 6%, and have also moved the 10% line to 12%.40 My feeling is that I should be consistent with the

39 The Commission’s report of 2003 stated that “The redistribution creates two more seats within the 6 per cent swing-to-lose figure” (EDBC, 2003, Report, the Commission, Adelaide, at para 41.) but the tables at Appendix 13 of that Report and Appendix 11 of the 2007 Report divide the seats according to 5% bands. 40 See A.Green, 2007, “2007 New South Wales Election; Preliminary analysis”, NSW Parliamentary Library Research Service Background Paper No 2/07 at p.30: “Seats with a marghin (sic) less than 6% are classified as marginal. Seats with a margin between 6% and 12% are classified as safe. Seats with margins above 12% are classified as Very safe.

PENDULUM LEADING INTO THE STATE ELECTION OF 2006 PENDULUM AFTER THE STATE ELECTION OF 18 March 2006 (EDBC estimates based on the actual results at the State election of 2002 and the boundaries ordered by the EDBC in 2003) (actual results at the State election of 2006 on the boundaries ordered by the EDBC in 2003)

ALP SEATSSWING TO LIB SEATS ALP SEATSSWING TO LIB SEATS 2PPLOSE 2PP 2PPLOSE 2PP

31 31 31 31 80.9 Chaffey (NAT vs ALP 2CP rethrow) 30 30 30 30

29 29 29 29 Ramsay 78.5 28 28 78.1 Mt Gambier (IND vs ALP 2CP rethrow) 28 28 78.5 Flinders (ALP vs LIB rethrow)

27 27 Taylor 77.4 27 27 Mt Gambier (IND vs LIB 2CP actual result) 76.6 76.7 Flinders 26 26 Croydon 76.0 26 26 Playford 75.8 Port Adelaide 75.7 25 25 Cheltenham 75.4 25 25 Enfield 74.5 Safe ALP seats going into the 24 24 Safe LIB seats going into the 2006 Napier 74.3 24 24 2006 election election swing to ALP 23 23 73.2 Hammond (CLIC vs ALP 2CP rethrow) swing to ALP 23 23 (%) (%) Ramsay 7.8 22 22 Mackillop -1.9 Kaurna 72.0 22 22 72.2 Mackillop Croydon 7 71.4 Fisher (IND vs ALP 2CP rethrow) Bragg 6.8 71.8 Chaffey (ALP vs LIB rethrow) Port Adelaide 6.9 21 21 71.3 Chaffey (LIB vs ALP 2PP rethrow) Goyder 6.9 21 21 Taylor 9.6 Ramsay 70.7 Finniss 9.3 Cheltenham 8.7 20 20 70.3 Mackillop Schubert 7.1 20 20 Enfield 8.7 69.6 Bragg Kavel 3.3 Lee 69.3 Napier 10.2 Croydon 69.0 19 19 Waite 8 Torrens 69.1 19 19 Playford 12.8 Port Adelaide 68.8 Davenport 5.2 Kaurna 11.3 18 18 Frome 8 West Torrens 68.3 18 18 Average swing 9.2 Taylor 67.8 Morphett 4.9 Reynell 67.6 17 17 Average swing 5.8 Chaffey (NAT vs LIB actual result) 67.2 17 17 Cheltenham 66.7 Little Para 66.7 66.7 Fisher (IND vs ALP actual result) 66.2 Hammond (LIB vs ALP 2PP rethrow) Colton 66.3 16 16 66.0 Goyder Ashford 66.1 16 16 Enfield 65.8 65.8 Finniss Wright 65.3 15 15 Mitchell (ALP vs LIB rethrow) 65.2 15 15 Napier 64.1 Elder 64.9 Chaffey (NAT vs LIB actual result) 64.0 14 14 Giles 64.4 14 14 63.5 Schubert Playford 63.0 13 13 13 13 62.7 Kavel 62.8 Bragg Fisher (IND vs LIB actual result) 62.1 12 12 62.0 Waite Florey 62.1 12 12 62.0 Hammond 61.6 Davenport 11 11 61.4 Frome 11 11 Kaurna 60.7 Fairly safe ALP seats going into 10 10 60.3 Morphett Adelaide 60.2 10 10 60.1 Flinders (LIB vs NAT actual result) the 2006 election Lee 59.6 59.8 Heysen Fairly safe LIB seats going into 59.4 Kavel swing to ALP 9 9 59.0 Unley the 2006 election Bright, Fisher (ALP vs LIB rethrow) 59.4 9 9 59.1 Goyder (%) West Torrens 58.8 swing to ALP Lee 9.7 Torrens 58.3 8 8 (%) 88 West Torrens 9.5 Heysen 6.8 Morialta 57.9 Torrens 10.8 Little Para (formerly Elizabeth) 57.0 7 7 Unley 7.9 77 Little Para 9.7 Average swing 7.4 Newland 56.8 56.5 Finniss Average swing 9.9 6 6 Mt Gambier (IND vs LIB actual result) 56.2 6 6 56.4 Davenport, Schubert Reynell 55.9 55.7 Fisher (LIB vs ALP 2PP rethrow ) 55.6 Mt Gambier (ALP vs LIB rethrow) Giles 55.2 5 5 55.4 Newland 5 5 55.4 Morphett Mitchell 54.7 Hartley 54.6 Marginal ALP seats going into the Colton 54.0 4 4 54.5 Bright Marginal LIB seats going into the Norwood 54.2 4 4 54.0 Waite 2006 election Ashford 53.7 2006 election swing to ALP Florey & Elder 53.5 53.5 Morialta swing to ALP 53.4 Frome (%) Wright 53.1 3 3 53.4 Mawson (%) 3 3 53.0 Heysen Reynell 11.7 52.5 Light Newland 12.2 Giles 9.2 52.2 Stuart Bright 13.9 Colton 12.3 52.1 Mt Gambier (LIB vs ALP 2PP rethrow) Morialta 11.4 Mawson 52.2 Ashford 12.4 Hammond (CLIC vs LIB 2CP actual result) 52.1 2 2 52.0 Hartley Mawson 5.6 Light 52.1 2 2 Florey 8.6 Light 4.6 Elder 11.4 Adelaide 51.0 1 1 Stuart 1.6 1 1 51.1 Unley Wright 12.2 V Hartley 6.6 V 50.6 Stuart, Mitchell (IND vs ALP actual result) Adelaide 9.2 Norwood 50.4 Average swing 8.0 Norwood 3.8 Average swing 10.1 SOURCES: Electoral Districts Boundaries Commission, 2003, Report 2003, The Commission, Adelaide, at Appendix 13; and State Electoral Office, 2003, Statistical Returns for the South Australian Elections, 9 February 2002, SEO, Rose Park. SOURCE: State Electoral Office 2006, Results and Outcomes: 3; State Election 18 March 2006, SEO, Adelaide. electoral administrations and the Boundaries Commission, so I have used 6% and 10%. It probably doesn’t matter.

The calculations show that the group of seats which went into the election as safe LIB seats, experienced an average swing to the ALP of just under 6%; in the fairly safe LIB seats it was higher at 7.4% and then in the LIB marginals the swing towards the ALP averaged 8%. The swings in the LIB marginal seats are particularly interesting – Stuart, Hartley and Light are all seats that went into the election as LIB marginals that had been contested strongly and continuously by both parties for many years and it is probably not surprising that these seats did not show big swings, but there were very big swings in the next group of seats. In Newland, Bright and Morialta, the swings towards the ALP were all well over 10%. Until the 2002 State election these seats had been fairly safe LIB seats (in the region of 56% to 60% LIB) but at that election the ALP drew them down the pendulum into more marginal territory. The size of the swings in Newland and Bright may also be at least partly attributable to the loss of their sitting members to retirement.

The other interesting thing about the figures surrounding the pendulum is how consistent the swings were across the State. In 2002 there had been seats that swung towards the ALP and other seats that swung towards the Liberal Party, but in 2006 the swings reverted to a much more consistent pattern. Almost every seat did swing towards the ALP (of the seats with an ALP:LIB 2PP result in both 2002 and 2006, only Mackillop did not swing towards the ALP) and swings were consistently high, even in the safest ALP seats where we might have expected very little unrealised support for the ALP.

The following sections look at the outcome of the 2006 election for the ALP and the Liberal Party and also for each of the smaller parties. One question which often arises in relation to smaller parties is why they contest Assembly seats at all, given that they are much more likely to win a seat in the Upper House and that contesting both Houses must divide scarce resources. The answer seems to be that smaller parties contest Assembly seats in order to win Council seats. Smaller parties contesting the Legislative Council elections receive less press coverage than the major parties do, even during an election campaign, so they do need to canvass electors personally as much as possible, and a candidate door-knocking for a House of Assembly seat can be given a more manageable task than someone asked to canvass electors throughout the State.

Winning more first preference votes from individual electors is the main aim of course but these votes can be supplemented with preference votes as other parties are excluded during the Legislative Council count, and those preferences arrive as a result of agreements made before the election, between the parties and the IND candidates contesting each time. Preference agreements are made partly on the basis of compatible ideological positions and partly on the grounds of a party’s ability to deliver votes, not only in the Legislative Council ballot but also in the House of Assembly contest. For example, a party which can demonstrate an ability to deliver votes in a marginal Lower House seat may receive a higher position on the Legislative Council ticket of the major parties. This all means that for smaller parties, the ability to demonstrate their voters’ compliance with a How To Vote card in the House of Assembly elections, is an important part of securing support in the Legislative Council election. For this reason the following analysis pays particular attention to the distribution of each party’s preferences and whether this demonstrates either compliance or an underlying orientation towards one party.

5. 2 The Australian Labor Party The ALP went into the election holding 23 of the 47 House of Assembly seats and with the status of an incumbent government, albeit a minority government reliant upon the support of IND and NAT MPs. It emerged from the election with a clear majority of 28 seats and the ability to govern in its own right.

The ALP contested every House of Assembly seat,41 and won 45.2% of all formal first preference votes cast across the State. When preferences were allocated from excluded

41 In 1977 the ALP did not contest Mitcham, Goyder and Flinders, which were won by DEM, IND LIB and NAT Members respectively.

candidates the ALP won 56.8% of the two party preferred vote. The Electoral Districts Boundaries Commission noted that the ALP had won the majority of seats with the majority of two party preferred vote, and concluded that for its purposes the 2003 redistribution had produced boundaries which were fair.42

Unlike the 2002 campaign and election, this was not a marginal seats campaign and the result was that almost every seat swung towards the ALP, many by large amounts. Referring again to the pendula, it is apparent that the ALP achieved strong positive swings in every sector of the pendulum.

The ALP now holds 28 seats, of which 21 are in the “safe” part of the pendulum – i.e. they would require a swing of 10% or more to change hands to the Liberal Party. Indeed the ALP won 21 seats on first preferences alone.43 While this may look like a comfortable position for the ALP to have achieved, it is only 10 years since the 1997 State election showed us what havoc can be caused when Independents and minor party candidates split safe margins in a large group of safe seats. I have previously called seats with a big 2PP margin “dangerously safe”44 reasoning that where a party has more than 2/3rds of the first preferences, that majority can be split. Actually, given the number of first preference votes that other smaller parties hold, it would be more correct to say that where a party has twice as many first preferences as the next candidate, there is a real risk of that large body of votes being split.

On this reasoning, the ALP margins in 15 ALP seats45 are vulnerable to splitting.

ALP candidates received the following preference flows:

• ALP candidates received a minority of Family First preferences (39% of FF preferences went to ALP candidates overall). • ALP candidates received the majority of Greens preferences (77% of Greens preferences went to ALP candidates overall). • ALP candidates received the majority of DEM preferences (61% of DEM preferences went to ALP candidates overall). • ALP candidates received the majority of Dignity for the Disabled preferences (55% of D4D preferences went to ALP candidates overall).

Nonetheless, the ALP is not obliged to any other party for its House of Assembly seats. As noted above, it won 21 seats on first preferences alone, and only 7 with the assistance of preferences from other parties or IND candidates. These 7 seats were Light, Mawson, Norwood, Hartley, Newland, Morialta and Adelaide. None of these seats was delivered to the ALP by smaller parties or IND candidates:

• Family First preferenced away from the ALP in Norwood, Hartley and Adelaide and issued a split ticket in the other seats so Family First support did not win these seats for the ALP. • The Greens Party may feel that they delivered the seats to the ALP via Greens preferences, but it seems clear from the analysis which follows later in this paper, that Greens voters will send their preferences to the ALP regardless of the direction that a Greens How To Vote card might urge. • The Democrats and Dignity for the Disabled did not direct their preferences but instead issued split tickets. • The various IND candidates did not have enough votes in their control to make a difference in these seats anyway.

42 Electoral Districts Boundaries Commission, 2007, 2007 Report of the Electoral Districts Boundaries Commission, EDBC, Adelaide, at p.14 ”…the requirements of section 83(1) were satisfied, in that Labor attracted more than 505 of the popular vote and was able to form government”. 43 These two groups are not identical: Adelaide is a safe ALP seat but the ALP candidate required preferences; Bright was won without requiring preferences but needs a swing of just under 10% to change hands. 44 See Newton-Farrelly, J, 2002, “Report on the Results of the South Australian State Election, 9 February 2002”, Parliamentary Library Information Paper No. 19, Adelaide, at p.12. 45 Ramsay, Taylor, Croydon, Playford, Pt Adelaide, Cheltenham, Enfield, Napier, Kaurna, Lee, Torrens, West Torrens, Reynell, Little Para and Wright.

Smaller parties derive (or try to derive) influence through being able to deliver preferences to competing parties via their How To Vote cards. By contrast ALP (and LIB) How To Vote cards are not powerful tools in this way, partly because their candidates are rarely excluded so their preferences are rarely distributed, and partly because the ALP and the Liberal Party are unlikely to ever preference towards each other, so the direction of their preferences is really not subject to negotiation. It is only in the Legislative Council that the ALP and the Liberal Party can derive some negotiating benefit from the order in which their preferences will be allocated to parties or candidates.

Each of the 47 ALP candidates for the House of Assembly seats lodged How To Vote cards for their seat, but the cards do not come into effect unless the ALP votes are distributed. This happened only in 3 seats – Chaffey, Flinders and Mt Gambier. In each of these seats the ALP How To Vote card put the LIB candidate last, which meant that the ALP preferenced to the IND or NAT candidate. Table 15 below shows where the ALP preferences actually went.

Table 15 shows that in Chaffey almost 83% of ALP votes went in the direction suggested by the ALP How To Vote Card, i.e. towards the NAT candidate (Maywald) rather than the LIB candidate. In Flinders 76.6% of ALP votes went to the NAT candidate rather than to Penfold (LIB), and in Mt Gambier almost the same percentage – 76.4% went to McEwen (IND) rather than to the LIB candidate. But is this evidence of ALP voters’ compliance with their How To Vote card or evidence of their own underlying orientation? To see whether ALP voters would preference towards a LIB candidate if their ALP How To Vote card directed them this way, we would need to have at least one or two cases where the ALP has issued that kind of card, but it has never happened.

Table 15: Distribution of ALP preferences, House of Assembly seats, 2006 ALP ALP Actual Distribution of ALP Preferences first HOW TO VOTE card went to: went to: preferences preferenced to: NAT / IND LIB NAT / IND LIB (No.) (No.) (No.) (No.) (%) (%)

Chaffey 1914 NAT 1587 327 82.9 17.1 Flinders 2,380 NAT 1823 557 76.6 23.4 Mt Gambier 4,582 IND 3,502 1,080 76.4 23.6

ALL 3 SEATS 8876 6912 1964 77.9 22.1

SOURCE: my calculations from unpublished worksheets from the State Electoral Department. See Attachments.

5. 3 The Liberal Party As it has done since 1979, the Liberal Party contested each House of Assembly seat. The 2006 election result was a poor one for the Liberal Party. It emerged from the election holding 15 seats, the lowest number of LIB Members in the House of Assembly since the size of the House was increased to 47 in 1970.

The small number of seats was a function of the overall drop in support for the Liberal Party across the State. On a two party preferred basis, the LIB result was certainly the poorest since 1982 when the State Electoral Department started to count out a two-party-preferred result for each seat, and probably the weakest result for the Liberals since the 1944 election (Jaensch has estimated 2PP figures for the postwar period.) Table 16 shows the 2PP vote for the Liberal Party and the ALP, at each State election since 1944.

Table 16: Two Party Preferred Vote, State Elections 1944 to 2006. ALP LIB (%) (%)

1944 (est.) 53.8 46.2 1947 (est.) 51.0 49.0 1950 (est.) 48.4 51.6 1953 (est.) 52.9 47.1 1956 (est.) 49.6 50.4 1959 (est.) 50.4 49.6 1962 (est.) 54.9 45.1 1965 (est.) 54.4 45.6 1968 (est.) 52.0 48.0 1970 (est.) 51.6 48.4 1973 (est.) 51.5 48.5 1975 (est.) 49.4 50.6 1977 (est.) 53.4 46.6 1979 (est.) 45.0 55.0 1982 (est.) 50.9 49.1 1985 53.1 46.9 1989 47.9 52.1 1993 39.1 60.9 1997 48.5 51.5 2002 49.1 50.9 2006 56.8 43.2

SOURCES: Jaensch, DH, 1971, “Under-Representation and the ‘Gerrymander’ in the Playford Era”, Australian Journal of Politics and History, Vol. 17 (1), April 1971, pp82-95. Jaensch, DH, 1975, “Don’s Party? Labor in South Australia”, Current Affairs Bulletin, Vol. 52 (3), August 1975, pp4-14. Jaensch, D, 1977, The 1977 Elections in South Australia; A Statistical Analysis, Politics Department Occasional Monograph No. 6, Flinders University. Jaensch, D, 1979, The 1979 Elections in South Australia; A Statistical Analysis, Politics Department Occasional Monograph No. 7, Flinders University. Jaensch, D, 1983, The 1982 Elections in South Australia; A Statistical Analysis, Politics Department Occasional Monograph No. 8, Flinders University. For 1982 and 1985, the two party preferred figures are official but unpublished State Electoral Department figures obtained by rethrowing ballot papers. SA. State Electoral Department, 1991, Periodical and General Elections 1989: Statistical Returns, SA PP161 of 1991-92. SA. State Electoral Office, 1996, Statistical Returns for General Elections 1993 and By-elections 1994, SEO, Adelaide. SA. State Electoral Office, 1998, Statistical Returns: General Elections 11 October 1997, SEO, Adelaide. SA. State Electoral Office, 2003, Statistical Returns for the South Australian Elections 9 February 2002, SEO, Adelaide. SA. State Electoral Office, 2006, Results and Outcomes, 3: State Election 18 March 2006, SEO, Adelaide.

The 2006 election was remarkable for the size of the swings that occurred and also for the fact that large swings towards the ALP occurred in every area of the pendulum. While the Liberal Party would have been disappointed with the overall result, it is nonetheless true that seats that were held by LIB Members going into the election did not swing towards the ALP quite as much as seats that were already held by the ALP. Overall, seats that went into the election as LIB seats swung towards the ALP by 6.8 percentage points, whereas seats that were held by ALP Members going into the election swung towards the ALP by 9.7 percentage

points. Six LIB seats did change hands but if the LIB seats had swung as much as the ALP seats, then the Liberal Party might also have lost Heysen and Unley.

LIB candidates received the following preference flows:

• LIB candidates received the majority of Family First preferences (61% of FF preferences went to LIB candidates overall). • LIB candidates received a minority of Greens preferences (23% of Greens preferences went to LIB candidates overall). • LIB candidates received a minority of DEM preferences (39% of DEM preferences went to LIB candidates overall). • LIB candidates received a minority of Dignity for the Disabled preferences (45% of D4D preferences went to LIB candidates overall).

My calculations show that there is one seat where the Liberal Party may be obliged to another party or IND candidate for the seat, i.e. where the ALP may have won if preference flows had been different. In Stuart, the outcome was quite close and the LIB sitting member received only about 30% of preferences that could have been transferred from DEM, the Greens, and the IND No Rodeo candidate46, but 67% of Family First preferences. Family First preferenced to LIB in Stuart, and my estimate is that if Family First had preferenced to the ALP in Stuart then at least 60% of Family First preferences would have flowed to the ALP47 and the seat would have changed hands to the ALP.48

Unley was another very close result for the Liberal candidate, and again Family First preferenced to LIB, so Unley might appear to be another seat just like Stuart. In fact the number of Family First votes was quite low in Unley (only 2.5% of first preferences) so it seems clear that even if Family First had preferenced to Labor in Unley the Liberal Party would still have won that seat.

The Liberal Party issued How To Vote cards in each seat, but it is hard to know whether LIB voters were influenced by these cards. For a start, LIB candidates are nearly always in the final two in each seat so LIB votes are rarely distributed to other candidates, which mean that we don’t have many opportunities to see whether the LIB voters have followed their How To Vote cards. In addition, LIB How To Vote cards will never preference to the ALP ahead of an IND or NAT candidate, so we have had no opportunities at all to see whether it is the LIB How To Vote card or the underlying orientation of LIB voters that is the most important determinant of how many LIB votes are transferred to a given candidate. All of which makes LIB How To Vote cards seem largely irrelevant, but they seem to be very relevant in a few seats - the seats where the LIB candidate is actually excluded.

In the House of Assembly election, LIB candidates were only excluded in 2 seats – Mitchell and Fisher – so it was only in these two seats that LIB voters had their preferences distributed. The Liberal Party might claim that it delivered both seats to the IND candidates via large numbers of preferences, because the LIB vote was substantial (21% of all first preferences in Mitchell and 19% in Fisher) and the vast majority of LIB voters did preference in line with the LIB How To Vote cards (86% in Mitchell and 89% in Fisher). But can the Liberal Party claim to have delivered these seats to the IND Members? I think they might be able to in Mitchell, but not in Fisher.

The result was very close in Mitchell - if 115 voters had sent their preferences to the ALP candidate rather than to Hanna, the seat would now be represented by the ALP. There were 4,094 LIB votes in that seat, and when the LIB candidate was excluded 86% of them went to IND (Hanna) rather than to the ALP. If slightly fewer LIB votes - 83.2% or fewer – had gone to Hanna, then the ALP candidate would have won the seat. So the Liberal Party could claim that they (or their voters) delivered the seat to Hanna, and it would be a difficult argument for anyone to refute. No other party could have made the difference – the Democrats also preferenced to Hanna before the ALP, but their vote was so low that they could not claim to

46 The DEM candidate issued a split How To Vote card; the Greens and IND No Rodeo candidates issued cards that preferenced to the ALP before LIB. 47 See the Family First section later in this paper. 48 A compliance rate of only 43% would have been required. See the Family First section below.

have given Hanna the edge, and the other candidates all preferenced to the ALP or issued split How To Vote cards, so they cannot make the claim at all.

The situation was different in Fisher. The LIB How To Vote card preferenced to the IND candidate (Such) before the ALP candidate, and 89% of LIB voters did send their preferences that way, but even if the LIB preferences had not been so strongly in Dr Such’s favour, he probably would still have won the seat. If LIB preferences had been divided equally between Such and the ALP, Such would still have won the seat; indeed the ALP candidate would have required 90% of LIB voters’ preferences in order to win the seat. Certainly that could only have happened if the Liberal How To Vote card preferenced to the ALP rather than Such49 and even then my feeling is that LIB voters would not have complied at such a high rate. So if the Liberal Party could not have delivered the seat to the ALP, could it argue that it delivered the seat to Dr Such? It is true of course that Dr Such won the seat with the help of LIB preferences but wouldn’t more than 10% of LIB voters have sent their preferences in his direction rather than to the ALP, even without a How To Vote card?

In the end, this may be all academic. It is hard to imagine a circumstance where the Liberal Party and indeed LIB voters, would rather have a seat represented by the ALP than by an IND Member, so perhaps it will always be the case that, at least in the House of Assembly, LIB How To Vote cards are essentially either irrelevant or predictable.

5. 4 The Australian Democrats For the first time since 1985,50 the Democrats did not stand a candidate in every seat – Giles had no DEM candidate. Whether this was a strategic move or a symptom of a weakening in the party itself, the Democrats must have been just as disappointed with their vote across the State as the Liberals would have been with theirs.

Across the State, Democrat candidates won 2.9% of the formal first preference votes in the House of Assembly ballot. This is the lowest share of the first preference vote they have ever won. Table 17 below shows the history of the DEM vote in the House of Assembly.

Table 17: DEM share of the first preference vote, House of Assembly, 1977 to 2006. DEM ALL DEM Seats Seats first first share of contested won preferences preferences first preferences (No.) (No.) (%) (No.) (No.)

1977 26,098 743,329 3.5 3 1 1979 60,969 734,966 8.3 36 1 1982 54,457 764,862 7.1 43 0 1985 34,732 816,888 4.3 43 0 1989 88,720 863,751 10.3 47 0 1993 82,939 912,086 9.1 47 0 1997 146,374 889,914 16.4 47 0 2002 71,026 948,032 7.5 47 0 2006 27,179 939,161 2.9 46 0

SOURCE: as for Table 5 above

49 The Greens and Family First issued How To Vote cards that preferenced to the ALP, and the Democrats issued a split card. 50 The Democrats stood candidates in just some seats in 1977, 1979, 1982 and 1985 and then first stood candidates in every seat in 1989.

There are several safe LIB seats, where the Democrats have historically done so well that they have had a good chance of splitting the LIB vote and winning the seat. The DEM vote in these seats – Kavel, Heysen and Waite – has been consistently stronger than in other seats, but in 2006 the Democrat vote was so low across the State that this sort of result now looks like a thing of the past. The strongest DEM result in a House of Assembly seat in 2006 was 5.3% in Heysen, a seat where in 1997 the DEM candidate won 29.2% of first preferences.

Table 18: Distribution of DEM preferences, House of Assembly, 2006 Democrat Actual Distribution of Preferences first went to: went to: preferences ALP LIB ALP LIB (No.) (No.) (No.) (%) (%) Adelaide 558 383 175 68.6 31.4 Ashford 928 553 375 59.6 40.4 Bragg 864 472 392 54.6 45.4 Bright 395 245 150 62.0 38.0 Chaffey (a) 271 115 156 42.4 57.6 Cheltenham 842 592 250 70.3 29.7 Colton 560 363 197 64.8 35.2 Croydon 781 426 355 54.5 45.5 Davenport 927 570 357 61.5 38.5 Elder 564 347 217 61.5 38.5 Enfield 673 456 217 67.8 32.2 Finniss 544 330 214 60.7 39.3 Fisher (a) 291 181 110 62.2 37.8 Flinders (a) 316 178 138 56.3 43.7 Florey 636 423 213 66.5 33.5 Frome 393 170 223 43.3 56.7 Giles no DEMOCRAT candidate Goyder 523 263 260 50.3 49.7 Hammond 739 405 334 54.8 45.2 Hartley 562 366 196 65.1 34.9 Heysen 1,044 657 387 62.9 37.1 Kaurna 423 275 148 65.0 35.0 Kavel 709 412 297 58.1 41.9 Lee 603 377 226 62.5 37.5 Light 370 215 155 58.1 41.9 Little Para 486 307 179 63.2 36.8 Mackillop(b) 425 178 247 41.9 58.1 Mawson 476 251 225 52.7 47.3 Mitchell (a) 344 236 108 68.6 31.4 Morialta 606 377 229 62.2 37.8 Morphett 649 362 287 55.8 44.2 Mt Gambier (a) 347 229 118 66.0 34.0 Napier 736 407 329 55.3 44.7 Newland 646 373 273 57.7 42.3 Norwood 545 364 181 66.8 33.2 Playford 759 547 212 72.1 27.9 Port Adelaide 499 319 180 63.9 36.1 Ramsay 693 471 222 68.0 32.0 Reynell 472 314 158 66.5 33.5 Schubert 754 413 341 54.8 45.2 Stuart 353 219 134 62.0 38.0 Taylor 475 298 177 62.7 37.3 Torrens 672 367 305 54.6 45.4 Unley 609 402 207 66.0 34.0 Waite 957 584 373 61.0 39.0 West Torrens 660 402 258 60.9 39.1 Wright 428 254 174 59.3 40.7 ALL SEATS 27107 16448 10659 60.7 39.3

(a) rethrow figures. (b) the figure for Mackillop does not include declaration votes. SOURCE: my calculations from unpublished worksheets from the State Electoral Department. See Attachments.

In the House of Assembly elections, Democrat candidates issue How To Vote cards which direct their voters’ preferences in relation to smaller parties and IND candidates, but which allow the voters themselves to choose whether their preferences should then go to either the ALP or to LIB. In most seats the smaller party and IND candidates are excluded from the count anyway, so effectively the Democrats do not direct their preferences in the House of Assembly, and whether a House of Assembly vote will end up with the Labor or LIB candidate depends on each individual DEM voter.51 In 2006 when the votes were scrutinised it became clear that DEM voters had given about 61% of their preferences to the ALP and 39% to LIB candidates in the various House of Assembly ballots (see Table 18 above). This is an interesting ratio, given that in the Legislative Council ballot the party’s split tickets would send 50% of above-the-line DEM votes to the ALP and 50% to LIB, if the DEM candidates were excluded from the count.

Democrat voters’ 60:40 orientation towards the ALP in the House of Assembly ballot in 2006 is no surprise. Democrat preferences have traditionally been more likely to go to the ALP than to LIB in the Lower House, and at roughly this rate – for example at the 2002 election Democrat voters sent 62.8% of their preferences to the ALP. But an interesting factor is that this pattern has persisted even though the Democrat vote was very weak in 2006. The similarity of the preference choices of the 2002 and 2006 DEM voters suggests that the voters who moved away from the Democrats in 2006 were probably not associated with a particular Labor-aligned or LIB-aligned group within that party but were simply a cross section of DEM voters.

Table 19 below shows that DEM voters’ orientation towards the ALP is stronger in safe ALP seats than in safe LIB seats. Almost 64% of DEM voters in safe ALP seats sent their preferences to the ALP, whereas DEM voters in safe LIB seats were just as likely to give their preferences to LIB as to Labor.

Table 19: Direction of Democrat voters’ preferences in various sectors of the pendulum, 2006 % of DEM voters who sent their ALP seats preferences to the ALP LIB seats 2PP (%) (%) 2PP

safe ALP seats 60% + 63.6 50.0 60% + safe LIB seats fairly safe ALP seats 56-60% 61.0 57.1 56-60% fairly safe LIB seats marginal ALP seats 50-56% 60.7 59.6 50-56% marginal LIB seats

SOURCE: My calculations from unpublished worksheets from the State Electoral Office.

Would the underlying orientation towards the ALP shown in Table 19 above, change if a Democrat candidate issued a How To Vote card that was favourable to a given party? Are DEM voters likely to be influenced by a How To Vote card?

In 2006 there were five seats where the final candidates were not ALP and LIB candidates. In four of these five seats the DEM How To Vote cards directed voters clearly to one of the final candidates. • In Flinders, the two versions of the DEM How To Vote card both preferenced to NAT before LIB.52 In the count, the final two candidates were NAT and LIB, and the majority of DEM voters (70%) did send their preferences to the NAT candidate. • In Chaffey, the two versions of the DEM How To Vote card both preferenced to LIB before NAT.53 In the count, the final two candidates were NAT and LIB, and the

51 Where a DEM voter makes a mistake which is rescued by the operation of a ticket lodged by the DEM candidate, the tickets would operate as for the Legislative Council i.e. half of the preferences to the ALP and half to LIB. 52 One preferenced to the ALP, then NAT then LIB, and the other preferenced to NAT, then LIB then the ALP.

majority of DEM voters (55%) sent their preferences to the NAT candidate, against both versions of the DEM How To Vote card. • In Mitchell, the two versions of the DEM How To Vote card both preferenced to the IND (Hanna) before the major parties. In the count, the final two candidates were ALP and IND (Hanna), and the majority of DEM voters (62%) did send their preferences to Kris Hanna. • In Mount Gambier, the two versions of the DEM How To Vote card both preferenced to the LIB candidate before the IND candidate (McEwen).54 In the count, the final two candidates were LIB and IND (McEwen), and the majority of DEM voters (59%) sent their preferences to the LIB candidate. In the fifth seat - Fisher - there were two DEM How To Vote cards; the first put IND (Such) before the ALP and the second put the ALP before IND (Such). The majority of DEM voters (62%) sent their preferences to Dr Such rather than the ALP candidate.

If Flinders, Chaffey, Mitchell and Fisher were the only cases, it might be arguable that what we are seeing is not so much compliance with a DEM How To Vote card, as an underlying orientation of DEM voters who seem to consistently prefer to send their preferences to an IND candidate or minor party candidate rather than a major party candidate, and when faced with the choice of an ALP candidate or LIB, preferring the ALP. But the result in Mt Gambier contradicts this. In Mt Gambier 59% of DEM voters sent their preferences in line with the DEM How To Vote card, to the LIB candidate rather than the IND candidate McEwen, so at least in Mt Gambier we do seem to be looking at a case of compliance rather than orientation. But if compliance becomes a governing principle in understanding how DEM preferences flow, and the flow of DEM preferences to a LIB candidate (in Mt Gambier) becomes understandable, we must then see Chaffey as a case of non-compliance with the DEM How To Vote.

Is compliance perhaps not a very useful concept in relation to the Democrats? Should we dismiss the 2006 Mt Gambier result, perhaps as evidence of the strength of a local campaign more than anything else? If there were more seats involved a pattern might be clearer, so we should look back at the four seats that had a two-candidate preferred result in 2002. In Mt Gambier in 2002, as well as in Chaffey, Hammond and Fisher in 2002, the distribution of DEM preferences is consistent with these voters having a preference for the IND or NAT candidate, then the ALP candidate and finally the LIB candidate.55 In Mt Gambier and Fisher this order of preferences was compliant with the DEM How To Vote cards, but in Chaffey and Hammond it was not56.

Combining the 2002 and 2006 results gives us 9 cases and a general pattern not of compliance but rather of DEM voters preferring an IND or NAT candidate, then an ALP one, then a LIB candidate, generally without regard to the DEM How To Vote card. Mt Gambier in 2006 is an exception, and we can really only guess at the cause.

5. 5 The Greens Greens parties have contested House of Assembly seats at State elections since 1993, but have never won a seat. However, they have actually had a Member on the floor of the house. Kris Hanna resigned from the ALP in 2003 and represented his seat – Mitchell – for Greens SA until just prior to the 2006 election. He then contested the election as an Independent and won it, so the Greens now have no representation in the Assembly.

For the first few times that the Greens contested House of Assembly elections, they stood candidates in only 2 or 3 seats, but in 2002 the Australian Greens contested 22 seats, and finally in 2006 they contested all 47 of the House of Assembly seats and won 6.5% of the first preference vote. They won a bigger share of the vote than Family First (5.9% of first preferences) and much more than the Democrats (2.9%), making the Greens in some ways

53 One preferenced to the LIB, then NAT then ALP, and the other preferenced to ALP, then LIB then the NAT. 54 One preferenced to the ALP, then LIB then IND, and the other preferenced to LIB, then IND then the ALP. 55 Unpublished worksheets from the State Electoral Office. 56 In Chaffey, DEM voters complying with theDEM How To Vote card would have sent their preferences to LIB before NAT but only 31% did; in Hammond DEM voters complying with the theDEM How To Vote card would have sent their preferences to LIB before IND Lewis, but only 47% did.

the invisible third force in the House of Assembly, though not actually holding a seat in that House.

Table 20: Average Greens vote in various sectors of the pendulum, 2006 ALP seats GREENS as % of formal votes LIB seats 2PP (%) (%) 2PP

safe ALP seats 60% + 6.2 5.4 60% + safe LIB seats fairly safe ALP seats 56-60% 5.4 8.2 56-60% Fairly safe LIB seats marginal ALP seats 50-56% 6.3 7.8 50-56% marginal LIB seats

SOURCE: my calculations from: SA. State Electoral Office, 2006, Results and Outcomes: 3, State Election 18 March 2006, SEO, Adelaide.

In individual seats the Greens vote was quite variable – it ranged from 1.8% of formal first preferences in Mt. Gambier to 17.7% in Heysen. Unlike the case with the Family First vote, it is not at all clear that the Greens vote was stronger in safe (or indeed in marginal) seats. Table 20 above shows the average Greens vote in seats within the different sectors of the pendulum.

Table 20 does not show that support for the Greens is likely to be stronger in safer (or indeed in more marginal) seats. Actually, in 2006 the seats where the Greens vote was strongest were Kavel, Davenport, Waite and Heysen – seats where the Australian Democrats vote had, in the past, been quite strong. In addition, the seats where the Greens vote was lowest in 2006 were all won by IND or NAT MPs. The combination of these two factors makes the Greens look like the new home of the protest vote component that had previously seemed to belong to the Democrats.

We know that the DEM vote collapsed in 2006; is it possible that support for the Greens and for the Democrats comes from the same voters, and the Greens picked up a lot of the voters who would previously have voted Democrat? Table 21 shows the share of first preference votes gained by the DEM candidates and also by the Greens candidates in 2002 and 2006. It does indicate fairly clearly that the Greens did best in 2006 in seats where the Democrats had done well in 2002 but where they collapsed in 2006. What makes it seem possible to think that the DEM and Greens voters are a single group of voters, is that for most seats the number of people who voted either DEM or Greens (i.e. the size of the DEM +Greens vote) was roughly the same in 2002 and in 2006.

Table 21: DEM and Greens share of the first preference vote, 2002 and 2006 2002 2002 2002 2006 2006 2006 DEM Greens DEM +Greens DEM Greens DEM +Greens (%) (%) (%) (%) (%) (%)

Adelaide 5.9 5.2 11.1 2.9 9.3 12.2 Ashford 9.6 5.4 15.0 4.6 9.1 13.8 Bragg 10.6 10.6 4.3 9.3 13.6 Bright 8.2 4.7 12.9 2.0 6.5 8.5 Chaffey 4.0 4.0 1.4 2.4 3.8 Cheltenham 6.8 3.0 9.8 4.4 6.0 10.3 Colton 4.4 5.5 9.8 2.7 5.8 8.6 Croydon 7.9 7.9 4.0 9.6 13.6 Davenport 16.9 16.9 4.7 11.4 16.1 Elder 6.7 3.9 10.7 2.9 7.3 10.2 Enfield 6.1 6.1 3.5 6.2 9.7 Finniss 9.1 5.4 14.5 2.6 7.7 10.3 Fisher 6.6 6.6 1.4 3.4 4.9 Flinders 6.1 6.1 1.7 4.1 5.8 Florey 6.9 3.7 10.6 3.4 5.7 9.1 Frome 4.2 4.2 2.0 3.7 5.7 Giles 9.4 9.4 0.0 4.5 4.5 Goyder 6.3 6.3 2.5 4.5 7.1 Hammond 5.6 5.6 3.6 7.2 10.8 Hartley 6.6 4.2 10.8 2.9 6.7 9.6 Heysen 16.3 8.8 25.1 5.3 17.7 23.0 Kaurna 5.7 5.4 11.2 2.2 7.3 9.5 Kavel 9.1 5.3 14.3 3.6 10.0 13.6 Lee 5.5 5.3 10.9 3.0 7.6 10.6 Light 5.5 2.9 8.3 1.8 5.0 6.8 Little Para (a) 7.0 7.0 2.4 4.6 7.0 Mackillop 2.0 2.0 2.5 4.2 6.6 Mawson 8.5 8.5 2.3 5.0 7.4 Mitchell 8.1 8.1 1.7 3.4 5.1 Morialta 10.4 10.4 3.0 6.3 9.3 Morphett 7.2 5.2 12.4 3.1 7.6 10.6 Mt Gambier 2.2 2.2 1.7 1.8 3.5 Napier 8.6 8.6 3.7 5.6 9.4 Newland 10.3 10.3 3.2 5.2 8.4 Norwood 8.0 6.0 14.1 2.8 8.5 11.2 Playford 8.5 8.5 3.7 5.0 8.7 Port Adelaide 6.1 4.2 10.3 2.5 6.6 9.1 Ramsay 7.0 7.0 3.6 4.9 8.5 Reynell 6.7 4.2 10.9 2.4 4.7 7.0 Schubert 8.9 4.5 13.3 3.6 7.5 11.1 Stuart 3.9 3.9 1.7 3.6 5.3 Taylor 7.1 7.1 2.4 3.9 6.2 Torrens 8.7 8.7 3.4 6.5 9.9 Unley 7.2 6.1 13.3 3.1 9.8 12.9 Waite 12.3 5.3 17.6 4.6 10.6 15.1 West Torrens 6.6 6.2 12.8 3.3 8.3 11.6 Wright 7.3 7.3 2.0 3.9 5.9 ALL SEATS 7.5 2.4 9.8 2.9 6.5 9.4

(a) Little Para was called Elizabeth in 2002 SOURCE: my calculations from SA. State Electoral Office, 2003, Statistical Returns for the South Australian Elections 9 February 2002, SEO, Adelaide, and SA. State Electoral Office, 2006, Results and Outcomes: 3, State Election 18 March 2006, SEO, Adelaide.

Chart 1 which is derived from Table 21, shows the relationship between the DEM and Greens votes more clearly, and also shows that the combined DEM + Greens vote in 2006 was generally a couple of points lower than in 2002, which might indicate that there is still scope for the Greens vote to increase (or perhaps for the DEM vote to recover) by a couple of percentage points. (We might also expect the Greens vote to increase in 2010 because of an incumbency advantage that this first Member will bring to the party.)

Chart 1: Share of all formal first preferences won by the Democrats or the Greens, House of Assembly, 2002 and 2006

30.0

25.0

20.0

15.0

10.0

% of allvotes formal first preference 5.0

0.0 Lee Light Giles Elder Kavel Unley Waite Bright Bragg Stuart Florey Taylor Fisher Colton Frome Napier Wright Enfield Hartley Finniss Kaurna Goyder Reynell Heysen Mitchell Chaffey Torrens Ashford Morialta Flinders Playford Ramsay Mawson Croydon Newland Adelaide Morphett Norwood Schubert Mackillop Davenport Hammond Little Para Mt Gambier Mt Cheltenham Port Adelaide West Torrens

2002 2006

A solid majority of Greens preferences flow to the ALP. Greens candidates stood in all 47 House of Assembly seats in 2006, and issued How To Vote cards that preferenced to the ALP in 46 of them. On average 75.8% of Greens voters sent their preferences to the ALP.57 In only one seat (Mt Gambier) the Greens How To Vote Card preferenced to LIB, and even then 56% of Greens voters still sent their preferences to the ALP. This resistance to a How to Vote card that directs preferences to a LIB candidate indicates that Greens voters in the other seats were not so much complying with a How To Vote card as expressing their own underlying orientation.

What is remarkable about the strong flow of percentage of Greens voters’ preferences to the ALP in 2006 is that it occurred even in LIB seats. Table 22 below shows that even in safe LIB seats 2 in every 3 Greens voters sent their preferences to the ALP and in most seats 3 in every 4 went to the ALP.

Table 22: Direction of Greens voters’ preferences in various sectors of the pendulum, 2006 % of Greens voters who sent ALP seats their preferences to the ALP LIB seats 2PP (%) (%) 2PP

safe ALP seats 60% + 77.1 66.6 60% + safe LIB seats fairly safe ALP seats 56-60% 78.4 75.2 56-60% fairly safe LIB seats marginal ALP seats 50-56% 76.0 74.1 50-56% marginal LIB seats

SOURCE: My calculations from unpublished worksheets from the State Electoral Office.

57 This is consistent with the 2002 result when 78% of Greens voters sent their preferences to the ALP.

Taken together, these figures indicate that the flow of Greens preferences to the ALP is very strong and is a function of Greens voters’ orientation towards the ALP rather than disciplined compliance with Greens How To Vote cards. Indeed it seems quite clear that even if the Greens did not issue How To Vote cards at all, about 75% of Greens preferences would still flow to the ALP candidate in each seat.

5. 6 Family First Family First contested House of Assembly seats for the first time in 2002 and returned in 2006. Their support was stronger in 2006, partly perhaps because of the benefit of having a Family First Member already in the Parliament (a FF Member was elected to the Legislative Council in 2002). In the House of Assembly, Family First’s support increased from 4.6% of formal first preferences in the 27 seats they contested in 2002 to 6.1% of first preferences in the 45 seats they contested in 2006. Overall, in 2006 Family First won 5.9% of all first preference votes, nearly twice as big a share as the Democrats (2.9%) but not quite as big as the Greens (6.5%).

The FF vote was quite variable in size – it ranged from 2.5% of formal first preferences in Unley to 15.7% in Kavel. While there were exceptions, in general Family First won a higher proportion of their votes in safer seats, both Labor and Liberal seats. Table 23 below shows that Family First candidates won a greater share of the vote in safer seats than in marginal seats, and generally also did better in ALP seats than in LIB seats.

Table 23: Average Family First vote in various sectors of the pendulum, 2006 ALP seats Family First as % of formal votes LIB seats 2PP (%) (%) 2PP

safe ALP seats 60% + 6.7 5.7 60% + safe LIB seats 6.4 without Kavel, fairly safe ALP seats 56-60% 5.6 8.3 including Kavel 56-60% fairly safe LIB seats marginal ALP seats 50-55% 5.0 4.2 50-55% marginal LIB seats

SOURCE: My calculations from unpublished worksheets from the State Electoral Office. Excludes Croydon and Ramsay (both safe ALP) where FF did not run candidates.

No FF candidate won a seat in the House of Assembly election, so it happened that every Family First vote actually helped another party’s candidate, and in most cases it was a LIB candidate. Across the State, 61% of FF preferences went to LIB and the remaining 39% of FF first preference votes went to the ALP.58 This is interesting: a majority of the Family First voters live in ALP seats, but they sent their preferences to LIB rather than the ALP.

Family First policies are socially conservative and the majority of their preferences go to the LIB candidates, so is it the case that there is an underlying orientation towards the Liberal Party amongst FF voters or was this mostly a case of compliance with Family First How To vote cards? To look at orientation we need to see where Family First voters send their preferences if the party does not direct them towards either of the major parties, and for the first time Family First did issue split tickets in 2006, in 6 seats. These seats were: Light, Mawson, Morialta and Newland (all went into the election as LIB marginals) and Torrens (fairly safe ALP) and Mitchell (marginal ALP but held by IND Hanna who had resigned from the ALP). Table 24 below shows that in 2006 when the Family First How To Vote card did not direct preferences to either major party, an average of 60% of Family First preferences across these seats went to LIB.

58 In some seats, one of the final two candidates was an IND or NAT candidate, but the polling officials have conducted rethrows of the ballot papers in these electorates, so we know whether the Family First votes would have ended up with the ALP or the LIB candidates in every seat.

Table 24: Flow of Family First Preferences in 6 seats, 2002 and 2006.

2002 FF How To Vote card preferenced to LIB 2006 FF How To Vote card was split FF first Preferences distributed to: FF first Preferences distributed to: preferences ALP LIB preferences ALP LIB (No.) (%) (%) (No.) (%) (%)

Light 933 27.5 72.5 1403 32.4 67.6 Mawson 1,391 34.5 65.5 1330 33.6 66.4 Mitchell 922 35.4 64.6 1062 40.4 59.6 Morialta 1,054 42.2 57.8 1,174 42.2 57.8 Newland 1,348 27.8 72.2 1438 42.8 57.2 Torrens 1,245 33.7 66.3 1439 50.4 49.6 Average 1,149 33.4 66.6 1308 40.4 59.6

SOURCE: My calculations from unpublished worksheets from the State Electoral Office. See Attachments.

As it happened, each of these seats had also been contested by a Family First candidate in 2002 and in each case a Family First How To Vote card had directed preferences to the LIB candidate. Table 24 shows that in 2002 67% of Family First voters in these same seats sent their preferences to LIB.

The figures in Table 24 indicate that Family First voters have an underlying orientation towards the Liberal Party that would result in about 60% of Family First voters sending their preferences to a LIB candidate even without direction from the , but a Family First How To Vote card directing preferences to LIB might increase this proportion to about 67%. There is some variation between the six seats: in Mawson and Morialta there seems to be an underlying orientation towards LIB but the existence of the How To Vote card made no difference at all, whereas in Torrens there seems to be no underlying orientation to LIB but a How To vote card preferencing to LIB did have a marked effect. So it seems that there is another factor in understanding the flow of Family First votes – perhaps the party affiliation of the sitting Member plays a part.

Table 25 below summarises the proportion of FF preferences that went to the ALP and LIB candidates according to whether the seat was an ALP or LIB seat going into the election. It shows several things. Firstly, Table 25 shows that compliance with the Family First How To Vote card was stronger where the card directed preferences to flow towards the sitting member (or the retiring member’s party). If Family First directed preferences to the sitting Member (or the retiring Member’s party) 60-70% of Family First voters sent their preferences in that direction.

Table 25: Compliance with Family First How To Vote cards, 2006. FF preferences went to: ALP LIB (%) (%)

Seats which were FF issued a split HTV card and ticket (Torrens) 50 50 ALP going into the FF HTV card and ticket preferenced to ALP (7 cases) 63 37 election FF HTV card and ticket preferenced to LIB (12 cases) 37 63

Seats which were LIB FF issued a split HTV card and ticket (4 cases) 38 62 going into the FF HTV card and ticket preferenced to ALP (no cases) n.a. n.a. election FF HTV card and ticket preferenced to LIB (16 cases) 29 71

NOTE: Excludes Croydon and Ramsay (no FF candidate) and Chaffey, Fisher, Hammond, Mitchell and Mt Gambier, which were not held by sitting MPs who belonged to the ALP or the Liberal Party. SOURCE: My calculations from unpublished data obtained from the State Electoral Office.

Secondly, it is clear from Table 25 that Family First voters were actually complying with How To Vote cards rather than simply expressing an underlying orientation towards one party or the other. In ALP seats, if Family First directed their preferences to the ALP 63% of Family First voters complied with the card, and if the FF How To Vote card preferenced to the Liberal Party, the same proportion – 63% – complied.

The figures in Table 25 that relate to split tickets are based on just a few seats; the pool cannot be increased by looking at 2002 because Family First did not issue split tickets at all at that election, so we will need to wait for 2010 figures for confirmation. Still, at present it looks to be the case that Family First voters in seats with an ALP sitting member have no orientation towards the ALP or to LIB, but that they will comply with a Family First How To Vote card directing their preferences to either of the major parties, and that this How To Vote card will deliver about 63% of their vote. At the same time, Family First voters in seats with a LIB sitting Member do appear to have an orientation towards the Liberal Party (worth about 10% of FF preferences) and, adding to this general orientation, a Family First How To Vote card preferencing to LIB will deliver the votes of another 10% of Family First voters.

In summary, it seems quite clear that Family First can not only reliably deliver 60-70% of their preferences to a party but can withhold them as well.

In 2006, did Family First preferences affect the outcome of the election in any seat? There are two seats – Stuart and Mitchell – where I suspect that a different Family First How To Vote card would have changed the result.

In Stuart, Family First preferenced to the LIB sitting Member Graham Gunn. If Family First had issued a How To Vote card directing preferences to the ALP, a compliance rate of 45.3% or higher would have sent enough preferences to the ALP candidate to have made the seat change hands to ALP. As shown in Table 25, in other seats59 where Family First preferenced to the ALP, compliance was 63%, so I think that even allowing for the fact that Stuart had a LIB sitting MP, 45% would have been well within the bounds of possibility.

In Mitchell, my sense is that Family First voters, not the party itself, made it possible for the sitting member to retain the seat. The final contest in this seat was between the ALP candidate and the sitting MP Kris Hanna, standing as an IND candidate. Family First had issued a split ticket: the top segment preferenced to IND then ALP and the bottom segment preferenced to ALP then IND. So the Family First party did not direct voters towards either the ALP or Kris Hanna. Family First voters, on the other hand, showed a clear preference for Hanna, and 66.5% of Family First preferences went to IND (Hanna) compared to 33.5% to the ALP candidate.

But I think that Family First could have delivered Mitchell to the ALP. In Mitchell in 2006, if Family First had issued a single How To Vote card that preferenced to the ALP candidate before Kris Hanna then even if only 45% of Family First voters complied the ALP candidate would have won the seat. Table 25 above showed that even where the Family First How To Vote card preferenced away from the sitting Member, about 63% of Family First voters did comply with a Family First How To Vote card, so a compliance rate of 45% would have been quite probable.

Stuart and Mitchell seem to be quite clear cases where Family First votes did influence the result in 2006, and if Family First were to claim to have delivered Stuart to the Liberal Party, it would be hard to argue against the proposition. In Mitchell, Family First supporters, not their party, were responsible.

5. 7 The National Party National Party candidates contested Flinders, Finniss, Chaffey and Mackillop and won 25% of first preferences in these seats overall. Table 26 below shows that in each case the NAT How To Vote card preferenced to LIB, and a clear majority of NAT preferences did go to the LIB candidate. Because there were no seats where the NAT card preferenced away from

59 Admittedly all ALP seats.

LIB, it is not clear how much of this was due to the How To Vote card and how much to a pre- existing orientation towards the Liberal Party.

The strongest support for the National Party was in Chaffey where the NAT sitting Member (Maywald) won the seat outright with 53% of all formal first preference votes.

In Flinders the seat was won outright by the LIB candidate (53% of first preferences) but the NAT candidate won 24% of first preferences, twice the first preference vote of the ALP candidate in that seat. Apart from the LIB How To Vote card, all of the How To Vote cards in the Flinders contest preferenced to NAT before LIB, so the NAT candidate accrued 69% of all of the preferences from excluded candidates and the contest finished as a LIB: NAT contest. The final two candidate preferred result in that seat was LIB 60% to NAT 40%.

Of course Flinders was represented by a National Party Member for 20 years from 1973 until 1993, so a solid level of support for the National Party is not unexpected in this seat, but the interesting thing about this particular result is that it indicates that support can be revived. After Peter Blacker lost the seat in 1993, the NAT vote declined until in 2002 the NAT candidate won just 9.4% of first preferences, but this 2006 result of 24% puts the Nationals back into a position from which they could mount a serious challenge in 2010.

In Finniss the NAT candidate did not win enough support to be a challenger for the seat, but nonetheless did make a difference to the outcome through preferences. The National Party candidate won 17% of first preferences, the NAT How To Vote card preferenced to the LIB candidate, and compliance was high (71%). The LIB candidate won the seat. The National Party might claim to have delivered Finniss to the Liberals, but my guess is that this is another case of orientation rather than compliance. If the National Party have directed their preferences to the ALP, and 65% or more of the NAT voters had sent their preferences to the ALP candidate, the seat would have changed hands to the ALP. I have no record of a NAT How To Vote card ever preferencing to the ALP, so there is no information about what level of compliance the National Party might expect of its voters if it did ask them to preference to the ALP, but my guess is that 65% would be very unlikely.

Finally, in Mackillop the NAT candidate won only 5% of first preferences, too few to make a difference in this safe seat.

Table 26: Distribution of NATIONAL PARTY preferences to ALP and LIB, 2006 NAT NAT share NAT Actual Distribution of Preferences first of all first Vote Card went to: went to: preferences preferences preferenced ALP LIB ALP LIB (No.) (%) to: (No.) (No.) (%) (%)

Chaffey (a) 10358 53.2 LIB 2920 7438 28.2 71.8 Finniss 3,622 17.5 LIB 1,034 2,588 28.5 71.5 Flinders (a) 4,596 24.3 LIB 930 3666 20.2 79.8 Mackillop 917 5.3 LIB 211 706 23.0 77.0

All contested seats 19493 24.6 ALL SEATS 19493 2.1 5095 14398 26.1 73.9

(a) The final 2 candidates in these seats were not ALP and LIB but a rethrow of ballot papers was conducted to find a 2PP result. (b) figures for Mackillop are just for votes lodged at booths. SOURCE: my calculations from unpublished worksheets from the State Electoral Office. SeeAttachments

Table 26 shows that where the NAT How To Vote card preferenced to LIB, a clear majority of NAT preferences did go to the LIB candidate. Because there were no seats where the NAT card preferenced away from LIB, it is not clear how much of this was due to the How To Vote card and how much to a pre-existing orientation towards the Liberal Party.

5. 8 Dignity for the Disabled Dignity For The Disabled contested the House of Assembly election for the first time in 2006. They ran candidates in 10 seats and won 2% of the first preference vote in those seats, and 0.4% of the first preference vote overall.

Dignity for the Disabled targeted marginal seats (3 ALP marginals, 5 LIB marginals and Mitchell held by IND Kris Hanna) and also ran in Unley (safe LIB going into the election). Although Unley remained in LIB hands the 5 other LIB seats all changed hands but D4D cannot claim to have made the difference in these seats, because it did not direct its preferences to either party.

D4D issued split How To Vote cards in each seat, but on average across the 10 seats, 55% of D4D preferences went to the ALP. This small but regular orientation towards the ALP is not explained by ballot paper position in these seats.

Table 27: Distribution of Dignity for the Disabled Preferences, House of Assembly, 2006. DIGNITY FOR THE DISABLED HOW TO Actual Distribution of Preferences first share of ALL first VOTE card went to: went to: preferences preferences preferenced ALP LIB ALP LIB (No.) (%) to: (No.) (No.) (%) (%)

Wright 506 2.4 split 261 245 51.6 48.4 Bright 492 2.4 split 261 231 53.0 47.0 Mawson 469 2.3 split 234 235 49.9 50.1 Adelaide 441 2.3 split 220 221 49.9 50.1 Mitchell 438 2.2 split 296 142 67.6 32.4 Morialta 418 2.0 split 260 158 62.2 37.8 Unley 362 1.8 split 193 169 53.3 46.7 Hartley 323 1.7 split 191 132 59.1 40.9 Norwood 274 1.4 split 148 126 54.0 46.0 Newland 251 1.2 split 131 120 52.2 47.8

ALL SEATS 848 2.0 470 378 55.4 44.6

SOURCE: my calculations from unpublished worksheets from the State Electoral Office. See Attachments.

5.9 One Nation One Nation first contested House of Assembly seats in 2002 when they stood candidates in all 47 House of Assembly seats and won 2.4% of the first preference vote – nearly 23,000 first preference votes. Just 4 years later almost 90% of these One Nation voters chose another party: in 2006 One Nation won fewer than 3,000 votes. In 2006 One Nation candidates contested just 6 seats (Ashford, Goyder, Hammond, Kavel, Newland and Port Adelaide) but did not choose the seats that had been strongest for the party in 2002, and they attracted just under 2% of the first preference votes in those six seats. Support in 2006 was stronger in country seats than in the city – One Nation candidates in Goyder and Hammond reached 2.7% and 4.1% of first preferences, respectively.

One Nation How To Vote cards sent their second preferences to Family First in each case, and then to LIB in the country seats. In the city seats they preferenced away from the sitting member or the retiring sitting member’s party (to LIB before ALP in Ashford and Pt Adelaide, and to ALP before LIB in Newland). The preference flows shown in Table 28 below indicate that a majority of One Nation voters allocated their preferences to LIB rather than ALP in all 6 seats, even in Newland.

Table 28: Compliance with One Nation How To Vote cards, 2006. ONE NATION HOW TO Distribution of Preferences First share of all first VOTE CARD went to: went to: preference votes preferences preferenced to: ALP LIB ALP LIB (No.) (%) (No.) (No.) (%) (%)

Hammond 845 4.1 LIB 290 555 34.3 65.7 Goyder 552 2.7 LIB 246 306 44.6 55.4 Kavel 373 1.9 LIB 183 190 49.1 50.9 Pt Adelaide 310 1.6 LIB 141 169 45.5 54.5 Newland 306 1.5 ALP 137 169 44.8 55.2 Ashford 205 1.0 LIB 94 111 45.9 54.1

ALL 6 SEATS 2281 2.1 950 1331 41.6 58.4

SOURCE: my calculations from unpublished worksheets from the State Electoral Office.

Table 28 shows that overall, about 58% of One Nation voters sent their preferences to LIB, but the figures are affected by the larger vote and higher compliance rate in Hammond. It was more common for about 55% of One Nation preferences to go to LIB. Table 28 also indicates that in Kavel and Newland, the existence of a One Nation How To Vote card seems to have made little difference to the behaviour of One Nation voters. In Kavel just as many One Nation preferences flowed through to the ALP as went to LIB, even though the How To Vote card preferenced to LIB, and in Newland the card preferenced to the ALP but only 45% of preferences flowed in that way.

The preference flows in both Kavel and Newland are consistent with ballot paper position. In both of these seats the One Nation candidate was at the top of the ballot paper and a donkey vote straight down the paper would have sent One Nation preferences to the ALP in Kavel and to LIB in Newland. Given that compliance or overall orientation in most of these seats was only weak, and the level of support for One Nation vote was so low, One Nation cannot claim to have made a decisive difference to the outcome in any seat.

5. 10 Independents There are currently 3 IND members in the House of Assembly, and they were the only IND candidates with substantial support in any seat in 2006. Apart from the fact that each was re- elected in 2006, their situations are all quite different.

In Fisher, Dr Bob Such won the seat from the ALP in 1989 and retained it in 1993 and 1997, each time as the Liberal candidate. He resigned from the Liberal Party in 2000 and then in both 2002 and 2006 he contested– and won - the seat as an Independent. In both 2002 and 2006 he won more first preference votes than any other candidate and went on to win the seat with a strong majority (with a 2candidate preferred result of 62.1% against the ALP in 2002 and then 66.7% against LIB in 2006).

In 1993, Dr Such won the seat outright, winning more than 50% of the first preferences, but at each election since then he has needed preferences from ALP or LIB voters. As long as the ALP and LIB candidates refuse to direct preferences to each other, Dr Such looks to be safe in Fisher. Meanwhile in both 2002 and 2006 Dr Such issued a split How to Vote card preferencing to all of the minor party candidates and then to either the ALP or LIB.

Table 29: Summary of the results in Fisher, 1989 to 2006 2 candidate preferred 2 party preferred Share of first preference votes result result IND IND ALP LIB (Such) Other Total ALP LIB (Such) ALP LIB (%) (%) (%) (%) (%) (%) (%) (%) (%)

1989 40.8 46.4 12.8 100.0 46.9 53.1 46.9 53.1 1993 22.3 64.3 13.5 100.0 29.0 71.0 29.0 71.0 1997 28.3 49.5 22.2 100.0 40.2 59.8 40.2 59.8 2002 22.1 31.3 33.5 13.2 100.0 37.9 62.1 44.3 (a) 55.7 (a) 2006 26.4 18.5 45.2 9.9 100.0 33.3 66.7 59.4 (a) 40.6 (a)

(a) results from a rethrow of ballot papers. NOTE: Figures in bold are Dr Such’s results. SOURCE: as for Table 5.

Interestingly, the character of the seat changed underneath Dr Such between 2002 and 2006. The two party preferred vote in Fisher made the seat LIB in 1989, 1993 and 1997, and in 2002 when the ballot papers were rethrown to find a two party preferred result it was not a surprise to see the seat would still have been a LIB seat if Dr Such had not been a candidate (the rethrow showed ALP 38% to LIB 62%). But in 2006 the rethrown ballot papers produced a two party preferred result of 59% ALP and 41% LIB, indicating that if Dr Such had not been a candidate the seat would have changed hands to the ALP. By the time of the next election in 2010, Dr Such will be 66 and will have served in the Parliament for over 20 years, and if he should decide then not to contest the seat again, we should probably think of the seat going into that next election as a fairly safe Labor seat.

Another seat represented in the House of Assembly by an Independent Member, is Mt Gambier, held by Rory McEwen. He won the seat in 199760 as an Independent, and was re- elected in 2002 and again in 2006. Mr McEwen has already stated that he will not contest the seat again in 2010, in which case we will probably treat Mt Gambier as a marginal LIB seat going into the State election of 2010, simply on the basis of the consistent orientation towards the Liberal Party in the rethrown ballot papers after the last three elections (see Table 30 below). Another reason why the seat seems to be a Liberal one is that the ALP first preference vote in the seat has not changed very much at all over the period so that where Mr McEwen has managed to win a larger share of the first preference vote (in 2002 for example) it seems to have been very much at the expense of the LIB candidate and (to a lesser extent ) candidates from other smaller parties. For this reason it seems that the LIB + IND(McEwen) voters might make up one group, and if McEwen is no longer a candidate most of his votes might move back to a LIB candidate. There is a complicating factor: in the rethrow exercises61 half of McEwen’s ballot papers have sent their preferences to the ALP, indicating only a rather loose allegiance to the Liberal Party.

As an Independent Member but nonetheless a Minister in the ALP government from 2002 until now, Rory McEwen seems to have given his LIB-oriented electorate the best opportunity for access to the ALP government. An interesting question is whether, after being represented by a strong local IND Member for 13 years,62 voters might be see an advantage to being valued by both major parties, in which case another IND candidate (or even a National Party candidate) might do well in this seat in 2010. While there is little in the history of the Mt Gambier seat itself to indicate that the Nationals would do well there, people from

60 At that election the seat was called Gordon. 61 I don’t have figures for 1997 but in 2002 48.5% of McEwen’s preferences went to Labor and in 2006 the proportion was 52.4%. 62 The seat was also represented by the IND John Fletcher for 20 years from 1938 until 1958.

this part of the State feel quite strong links with Victorian border areas, and National Party Members hold two of the three seats on the Victorian side of the border (Lowan and Mildura; South-West Coast is held by a LIB).63

Table 30: Summary of the results in Mt Gambier, 1997 to 2006 2 candidate 2 party preferred Share of first preference votes preferred result result IND IND ALP LIB (McEwen) Other Total LIB (McEwen) ALP LIB (%) (%) (%) (%) (%) (%) (%) (%) (%)

1997 21.6 41.5 22.5 14.4 100.0 49.9 50.1 37.7 (a) 62.3 (a) 2002 17.3 20.3 58.4 4.1 100.0 23.4 76.6 47.9 (a) 52.1 (a) 2006 22.2 33.9 35.6 8.4 100.0 43.8 56.2 44.4 (a) 55.6 (a)

(a) results from a rethrow of ballot papers. NOTE: Figures in bold are Rory McEwen’s results. SOURCE: as for Table 5.

The third State seat held by an Independent member is Mitchell. This seat has traditionally been strongly Labor; the State Bank election of 1993 delivered it to the Liberal Party for one term and Kris Hanna won it back for the ALP in 1997. He was re-elected as the ALP Member for Mitchell in 2002 but resigned from the ALP in 2003 and represented the seat as a Member of Greens SA for three years.64 Just prior to the 2006 State election he became an Independent.

Table 31: Summary of the results in Mitchell, 1997 to 2006 2 candidate 2 party preferred Share of first preference votes preferred result result IND IND ALP LIB (Hanna) Other Total ALP LIB (Hanna) ALP LIB (%) (%) (%) (%) (%) (%) (%) (%) (%)

1997 38.8 40.2 21.0 100.0 50.9 49.1 50.9 (a) 49.1 (a) 2002 43.1 38.1 18.9 100.0 54.7 45.3 54.7 (a) 45.3 (a) 2006 41.0 20.7 24.6 13.6 100.0 49.4 50.6 65.2 (a) 34.8 (a)

(a) results from a rethrow of ballot papers. NOTE: Figures in bold are Kris Hanna’s results. SOURCE: as for Table 5.

Table 31 above makes it clear that the ALP vote in Mitchell did not change markedly when Hanna won the seat as an IND candidate in 2006. This indicates that Hanna has split the LIB vote or the share of the vote that would otherwise have gone to other smaller parties. My guess is that the biggest threat to Hanna’s position is a strengthening of support in this electorate for the Greens or Family First, which could result in preferences being sent to the ALP or LIB candidates rather than to Hanna. Certainly the final result was so close in 2006, with only 228 votes separating the ALP and IND candidates, that if Family First had issued a How To Vote card that preferenced to the ALP before Hanna (rather than a split card)65 the seat would have gone to the ALP.

63 Claire Macdonald, Greg Gardiner and Brian Costar, 2007, “The 2006 Victorian State Election”, Parliamentary Library of Victoria Research Paper No 1, 2007. 64 January 2003 until February 2006. 65 The Family First How To Vote card was split: the top option preferenced to LIB, D4D, IND(Hanna) then ALP. The bottom option preferenced to ALP, D4D, LIB then IND (Hanna).

Mr Hanna is a relatively young member and in 2010 will still only be in his 40s, so there is no reason to expect that he will not contest the seat in 2010, and we should see it as a seat which is marginal in every dimension

5. 11 Summary: the effectiveness of How To Vote cards. Parties live in hope that they can make arrangements to gain other parties’ preferences. This certainly works in the Legislative Council, but in the House of Assembly ballot it is not at all clear that parties can deliver preference votes because, in the end, it seems that few voters will change their personal preferences in order to comply with a How To Vote card. Sections 5.2 to 5.9 above show that the Democrats and Dignity for the Disabled did not direct their preferences, ALP and LIB How To Vote cards always put the other major party last, the majority of Greens voters will always send their preferences to the ALP rather than LIB, and a small majority of One Nation voters will always send their preferences to LIB rather than the ALP. The only group which can reasonably claim to deliver votes to either the ALP or the Liberal Party is Family First who managed (admittedly only in ALP seats) to deliver 63% of their preferences to ALP candidates or to LIB candidates.

How To Vote cards do, obviously, influence some voters, and that is important in very tight contests. There are three seats which might have had a different outcome in 2006 if parties had issued different How To Vote cards:

• In Stuart, if Family First had preferenced to the ALP, the ALP may have won Stuart. • In Mitchell, if Family First had issued a How To Vote card that preferenced to the ALP (rather than a split card), then the ALP may have won Mitchell. • Also in Mitchell, in the unlikely event that the Liberal Party had preferenced to the ALP before IND Hanna, the ALP may have won Mitchell. • In Mt Gambier, if the ALP had issued a split How To Vote card, preferencing to either LIB or IND McEwen, my guess is that Mt Gambier would have changed hands from McEwen to LIB.

Apart from these three cases, smaller parties – the Greens, Family First and Dignity for the Disabled, for example – really do not have the power to deliver seats. So why do the parties continue to do preference trades before each election? Probably because occasionally it does makes the difference between winning and losing a seat. For Family First, for example, receiving a good spot on the Legislative Council ticket of the Liberal Party, might seem likely to be worth a lot more than an extra 10% of Family First preferences will be worth to the Liberal Party in the Assembly competition. But in a tight competition for a marginal seat even a few votes will be valuable, and both parties did benefit in 2006 - the Liberal candidate won Stuart with Family First preferences and Family First won its Legislative Council seat with LIB preferences. If the Liberal Party had won more votes in the Legislative Council and not been excluded during the count, the arrangement would not have worked very well for Family First at all in 2006, but there will always be that risk in the Council ballot. Parties make these arrangements for the long term, and we have evidence of that in the ALP’s gradual move away from the Democrats towards the Greens (see section 4.4 above).

6 Country and city seats The Electoral Districts Boundaries Commission looks at the probability, at a future election, of seats changing hands with a given State-wide swing, and so it has looked at swings in some detail. In its 1998 Report, the Commission stated that “rural seats are usually much less volatile than metropolitan seats”66 and that “there seems to be a pretty clear pattern in the six elections held since 1979, showing that country seats at each election swing less on average than the metropolitan seats, although the degree of drag in the country varies from election to election.”67

66 Electoral Districts Boundaries Commission, 1998, 1998 Report of the Electoral Districts Boundaries Commission, The Commission, Adelaide, at para. 4.10. 67 as above.

The 2002 election was difficult to analyse in these terms because the swings were not at all consistent even in direction, but the 2006 election allows us another chance to check this generalisation. This election does seem to have been another, more predictable, election and country seats did swing less than city seats. If we look again at the pendula, it seems clear that in 2006 each of the country seats swung by less than those metropolitan seats with equivalent margins going into the election.

• The group of seats which went into the election as safe LIB seats was made up of 6 country68 and 4 city seats – 10 seats in all. The country seats swung, on average, 5.6 percentage points towards the ALP, and the city seats swung, on average, 6.2 percentage points towards the ALP; • In the group of three seats which went into the election as fairly safe LIB seats the only country seat was Heysen, which swung less than Unley and Newland; • In the group of 6 seats that went into the election as marginal LIB seats, the only country seat was Stuart, which swung by much less than the rest of the group. Stuart swung only 1.6 percentage points towards the ALP (and stayed in LIB hands) but the rest of the LIB marginals swung on average by 8.4 percentage points to the ALP (and all changed hands to the ALP). • Giles, the only ALP country seat, swung to the ALP by 9.2 percentage points while the other fairly safe ALP seats swung towards the ALP by 10.3 percentage points.

The 2006 election results confirm that country seats do swing more sedately than city seats.

7 Are the major parties still losing market share? In House of Assembly elections, the two party preferred vote describes the outcome of each election quite well. For many years, the only parties that could have formed government have been the ALP or the Liberal Party, and neither the Nationals nor the Democrats have developed as major forces in the Assembly. The only problem with the 2PP count is that it tends to obscure the existence of the other parties in the ballot, as if every voter really does prefer either Labor or Liberal and the other parties are inconsequential. The two party preferred count is not used to analyse or describe the results in the Legislative Council. It is quite possible to work out a 2PP figure for the Council result, using each party’s ticket to see whether their preferences would have gone to the ALP or to LIB, but in fact the 2PP vote is an awkward fit with the Legislative Council election simply because (at least since 1985) other parties do win enough support in the Legislative Council ballot to win seats.

Because the 2PP vote does not describe the Legislative Council ballot very well, there is no simple statistic to compare how the major parties have performed in the two Houses. One thing we can look at, is how many voters give their first preference to the major parties and how many prefer someone other than a major party candidate. Table 32below shows that voter support for the smaller parties and IND candidates has grown very strongly over the past 20 years, at both the Assembly and the Council levels.

Table 32: ALP and LIB share of first preference votes in the House of Assembly and Legislative Council, 1985 to 2006. State Legislative Council House of Assembly election ALP LIB All others Total ALP LIB All others Total (%) (%) (%) (%) (%) (%) (%) (%)

1985 48.0 39.3 12.7 100.0 48.2 42.2 9.6 100.0 1989 39.7 41.1 19.1 100.0 40.1 44.2 15.7 100.0 1993 27.4 51.8 20.8 100.0 30.4 52.8 16.8 100.0 1997 30.6 37.8 31.6 100.0 35.2 40.4 24.4 100.0 2002 32.9 40.1 27.0 100.0 36.3 40.0 23.7 100.0 2006 36.6 26.0 37.4 100.0 45.2 34.0 20.8 100.0

SOURCE: as for Table 5 above.

68 The LIB safe country seats were Mackillop, Goyder, Finniss, Schubert, Kavel and Frome.

Table 32 shows that in the Legislative Council ballot in 1985, almost 90% of voters gave the ALP or the Liberal Party their first preference vote and only about 13% of voters chose a party other than the majors. Support for the smaller parties and IND candidates grew very strongly so that by 1997 voters were almost equally likely to choose Labor, or LIB, or someone else. Still, what seems to be happening in the Legislative Council is that since the State Bank election of 1993 the ALP has been steadily re-building its share of the first preference vote at every election, while the Liberals’ share has been weakening (in 2006 the Liberals won only half the support they had in 1993). Meanwhile, almost in the background, other parties or IND candidates have been steadily building their own support, apparently at the expense of the Liberals.

Given that support for smaller parties and IND candidates in the Legislative Council is now heading towards 40%, it is strange to see that their support at elections for the House of Assembly seats seems to have stabilised at 20-25% over the last three elections. Table 32 indicates that the ALP has been re-building its share of first preference votes even more strongly in the Assembly than in the Council and it seems that this growth has been at the expense of the smaller parties and IND candidates as much as the Liberal Party. But perhaps we should see this from the point of view of the voters as much as the parties. Perhaps the voters will vote for people who they think will have a reasonable chance of winning.

In summary, the major parties have indeed been losing ‘market share’ in both Houses, over the last 20 years, but at least since 1993 the situation really appears to be that the Liberal Party has been losing voters to both the ALP and the group of smaller parties or IND candidates.

8 Ticket votes in the House of Assembly ballot In the Legislative Council ballot almost everyone votes above the line by marking a “1” against one party or group of IND candidates. That vote is a formal vote and if the preference needs to be distributed to another party or group, it travels in a path that has been specified by the group, not by the voter.

Meanwhile in the House of Assembly ballot, a voter who lodges an incomplete vote is actually lodging an informal vote. But in South Australia, these informal votes are sometimes rescued an re-included into the count through a savings provision known as the “ticket vote”.69 • A ballot paper which has all of the squares marked in order, except one, is informal but will be re-included into the count as formal on the grounds that the blank square is assumed to be the voter’s last choice, so effectively “the voter’s intention is clear”. • A ballot paper which has just one preference marked, is informal but will be re- included in the count if the preferred candidate has lodged a ticket with the State Electoral Commissioner before the election to show how those preferences should be treated. (This is called the candidate’s ticket, and it is usually the same as his or her How To Vote card.) • A ballot paper which has more than one preference marked on it, but is incomplete, is informal but may be re-included into the count if the preferences coincide with the order of preferences on the preferred candidate’s ticket.

The ticket vote provision is not publicised, because it is simply a savings provision designed to re-integrate votes that are assumed to be mistakes and that would otherwise be informal. And there are a lot of these votes: in 2006 there were 43,553 incomplete House of Assembly ballot papers that were re-incorporated into the count through the ticket vote provisions. Given that the average number of formal votes in a House of Assembly electorate was just under 20,000 in 2006, the ticket vote re-incorporates about two electorates-worth of votes back into the count.

The ticket vote provision sits uneasily within the rest of the electoral system because it allows the parties to decide how that vote should be distributed during the count, even though it is

69 See s. 63 of the Electoral Act 1985 (SA).

quite possible that the voter may have a different series of preferences. By comparison in the Council ballot, above-the-line voters at least tacitly agree to accept the order of preferences that each party imposes. For this reason the ticket vote provision is sometimes criticised by electoral observers, but the problem really is that there is no obvious, useful alternative.

One alternative could be to revoke the ticket voting provision for the House of Assembly and simply exclude these incomplete votes. That would require us to ignore a large number of voters and accept a much higher informal count: in 2006 the ticket vote provision reduced the informal vote from 8.1% to just 3.6%.

Another alternative could be that we could allow these votes back into the count but only as far as they are formal. At the point on the ballot paper where it is unclear which candidate the vote should be transferred to next, the vote would be excluded from the count. This system was actually used in Federal elections for many years but was changed after the 1996 Federal election on the grounds that it allowed voters to register a valid vote that included no preferences and therefore introduced a form of optional preferential voting within a full preferential voting system.70

Apart from the fact that allowing an optional preferential component in a full preferential system seems clumsy, allowing votes to exhaust is likely to be a problem for the Boundaries Commission, because we would no longer have a complete figure for the two party preferred vote across the State. The Commission requires a 2PP figure for the State in order to make a judgement about whether the party which won a majority of seats did so with the support of a majority of voters across the State. Without a State-wide 2PP figure the Commission would lose its main test of the fairness of a previous redistribution.

Finally, allowing votes to exhaust would give the major parties an advantage over smaller parties or IND candidates. It would only be a small advantage, but votes just marked “1” for ALP or LIB candidates would stay in the count until the end, because ALP and LIB candidates usually stay in the count until the end, and votes marked with a “1” for, say, a Greens candidate, would exhaust once that candidate was excluded. Because the option of allowing votes to exhaust is likely to introduce a small bias in favour of the major parties, it is unlikely to fit into the current electoral system which has such a strong emphasis on fairness.

Perhaps it would be easier for commentators to accept the ticket vote provisions if they knew how benign ticket votes are.

• Although there are a lot of ticket votes across the State, they only make up about 5% of the votes in any one seat. • Most ticket votes are for major party candidates who stay in the count, so most ticket votes are never actually distributed to another candidate: in 2006, 80% of ticket votes were not transferred.71 • The only parties which do have a chance to determine the order of preference of the 20% of ticket votes that are distributed, are the smaller parties and the IND candidates whose candidates are excluded during the count.

For these reasons it is clear that the ticket vote provision in the House of Assembly ballot is not a mechanism by which the major parties gain an advantage or can unfairly influence the electoral system.

Still, given that several seats were won on quite small margins in 2006, it is possible that there were seats where ticket votes were so important – so different from normal votes – that they changed who won the seat. In fact this was not the case, in any seat. If the ticket vote

70 Albert Langer recommended that voters mark a ballot paper as 1,2,3,3; a vote in this form would be informal but would be re-included in the scrutiny under s.270(2) of the Commonwealth Electoral Act 1918. The Joint Standing Committee on Electoral Matters recommended in its Report of the Inquiry into the conduct of the 1996 Federal Election and matters related thereto that s.270(2) be repealed. See Recommendation 13 at p.32. 71 My calculations from Table 2.13 of SEO, 2006, House of Assembly Elections Data: District and Statewide Summary Figures, State Election 18 March 2006, compiled by the State Electoral Office for the Electoral Districts Boundaries Commission.

provision had not been operating at the 2006 House of Assembly elections, and all of the 43,553 votes re-introduced into the count through the ticket vote provision across the State had been treated as informal votes, then the 2PP result in the three closest seats would barely have changed. Stuart would have been LIB 50.7% (it was actually LIB 50.3%), the result in Unley would have been LIB 51.0% (it was actually LIB 51.1%) and in Light the result would not have changed (it was ALP 52.1%, with or without ticket votes).72

So if there is to be a debate about whether the ticket vote provision is useful or appropriate, any argument that the ticket vote provision distorts the result or provides an advantage to the major parties is unlikely to be convincing.

On the other hand, the number of votes re-included in the count due to the ticket vote provisions might be a reason to look again at electoral education programmes. In an election where almost all of the electoral material stated that voters must fill in all of the squares on the House of Assembly ballot paper, 43,553 voters lodged incomplete ballot papers. Ticket votes are associated more strongly with particular seats, more than particular parties, and they are all seats with generally low educational levels.73 While One Nation voters are more likely to be ticket voters (7.4% compared to 4.6% of all voters), in general it looks like ticket votes correlate with educational background or perhaps language proficiency, rather than political allegiance. In 2006, 4.6% of voters all across the State lodged ticket votes, but in Enfield, Napier, Cheltenham, Ramsay, Croydon and Taylor more than 6% of voters lodged ticket votes. In Waite, by contrast, only 2.8% of votes were ticket votes. In all of these seats, ticket votes are just as common amongst LIB voters as among ALP voters.

9 Proposed changes to the Legislative Council In November 2005 Premier Mike Rann foreshadowed that in 2010 he would hold a referendum on the future of the South Australian Legislative Council.74 He has since repeated his commitment to the referendum and to change. Voters would be asked whether they would prefer to keep the Legislative Council electoral system as it stands, abolish it, or reform it. The reform options include changing MLCs’ terms from 8 years to 4, and reducing the number of MLCs from 22 “to 16 or maybe more”.75

Various aspects of the Premier’s proposal will be addressed in another research paper, but a summary of the 2006 results under different hypothetical electoral systems is shown in Table 33 below.

Table 33: Composition of the Legislative Council in 2006 under hypothetical electoral systems. ALP LIB DEM Greens Family IND TOTAL First No Pokies Council of 22 MLCs: Actual current 22-Member Legislative Council (11 MLCs elected in 2002 and 11 in 2006): 8 8 1 1 2 2 22 Hypothetical 22-Member Council all elected at the same time in 2006: 8 6 1 2 5 22 Council of 16 MLCs: Hypothetical 16-Member Legislative Council (8 MLCs elected in 2002 and 8 in 2006): 6 6 1 1 2 16 Hypothetical 16-Member Legislative Council all elected at the same time in 2006: 6 4 1 1 1 3 16

SOURCE: my calculations applying a ticket vote model to data from SA. State Electoral Office, 2003, Statistical Returns for the South Australian Elections 9 February 2002, SEO, Rose Park; and SA. State Electoral Office, 2006, Results and Outcomes, 3: State Election 18 March 2006, SEO, Rose Park.

72 My calculations based on Tables 2.14 and 2.15 of SEO, 2006, House of Assembly Elections Data: District and Statewide Summary Figures, State Election 18 March 2006, compiled by the State Electoral Office for the Electoral Districts Boundaries Commission. 73 ABS SEIFA Indexes at www.abs.gov.au 74 “Rann to call referendum for 2010; Abolish the Upper House”, The Advertiser, 24 November 2005, p.1. 75 Hon MD Rann, Premier, “Ministerial Statement”, House of Assembly Hansard 24 November 2005, pp.4130-4131.

Table 33 shows that if we had elected 16 MLCs in 2006, then neither of the major parties would have the majority of seats in the chamber, and the big beneficiaries of the change would have been the IND No Pokies group who would have 3 seats in a much smaller House.

It is commonly thought that reducing the number of MLCs would increase the size of the quota and make it harder for small parties to win seats. But Table 33 above shows that if we had a 16-Member Council and had elected all 16 MLCs in 2006, then the Democrats would have won a seat, whereas if we had elected 22 MLCs in 2006, the Democrats would not have won a seat. While this seems to be counter-intuitive the answer lies in the fact that each time we change the size of the quota, there will also be changes to the size of the bundle of votes available for transfer from parties which have already won seats. In the hypothetical example where we elect 16 MLCs, the Democrats would receive a large bundle of votes available for transfer from the Liberal Party, but in the hypothetical example of electing 22 MLCs, there would be no LIB votes available for transfer because the LIB candidate would remain in the count. This example shows that who wins Council seats in a Council of a given size, has a lot to do with how many votes each party receives and then also the pattern of preference distribution during the count.

In 2006, the party which would have benefited significantly from the sorts of changes the Premier has proposed would have been the IND Nick Xenophon No Pokies group, who would have had 5 MLCs elected in 2006 if we had been electing all of the 22 Members of the Council at that election.

10 Implications of the State election for the coming Federal election A Federal election is likely to be held later in 2007, and throughout 2007 the major polls have been showing a 2PP result of between 56% and 60% for Labor.76

At the last Federal election, in 2004, the ALP won only 45.6% of the 2PP vote in South Australia, so the current level of support for the ALP shown by recent polls would mean a swing of 10 to 15% across the State. We know that swings are not uniform and that individual seats will swing less or more than the State average, so if a swing of this magnitude actually were to take place in South Australia, which seats would change hands?

I think that the results of the 2006 State election can be a good guide. We need to recognise that it is essentially the same group of people who voted at the last Federal election and at the last State election, so if we are looking for the distribution across South Australia of an ALP 2PP vote of between 55 and 60%, results of the last State election (ALP 2PP 56.8%) should be useful. A comparison of the results of the 2004 Federal election and the 2006 State election should show us two things: firstly, which Federal seats would change hands with an ALP 2PP result of 56.8% across South Australia at the coming Federal election, and secondly, where are the voters who voted LIB at the Federal level but ALP at the State level?

I have re-distributed the State results to the current Federal electoral boundaries, and the outcome is shown in Table 34 below. If each South Australian voter cast his or her vote at the coming Federal election in the same way as at the State election of 2006, then the ALP would win 5 extra seats in South Australia – Boothby, Kingston, Makin, Sturt and Wakefield. Overall, Labor would win 8 seats and Liberal would retain 3 - Mayo, Grey and Barker.

76 Newspolls showed the ALP with a 2PP vote of 56% in February 2007; 57% in both March and May 2007; 56% in July 2007; 59% in September 2007. Shanahan, D., 2007, “ALP in strongest position since '01”, The Australian, 6 February 2007, p.1 Shanahan, D., 2007, “Labor's lead hits record high - Opposition surges 14 points ahead despite slip in Rudd's popularity” The Australian, 6 March 2007, p.1 Lewis, S., 2007, “We face annihilation: Howard” The Australian, 23 May 2007, p.1. Shanahan, D., 2007 “Howard checks Rudd's march - Labor holds election-winning lead but voters embrace indigenous plan” The Australian, 10 July 2007, p.1 Shanahan, D., 2007 , “Howard humbled in polls as world leaders fly in for APEC summit - Rudd opens up crushing lead”, The Australian, 4 September 2007, p.1

Table 34: State 2006 results applied to House of Representatives seats ALP LIB outcome (%) (%)

Adelaide Federal 2004 51.3 48.7 State 2006 61.6 38.4 remains ALP swing to ALP 10.3 -10.3

Barker Federal 2004 30.1 69.9 State 2006 35.8 64.2 remains LIB swing to ALP 5.7 -5.7

Boothby Federal 2004 44.6 55.4 State 2006 51.9 48.1 changes hands to the ALP swing to ALP 7.3 -7.3

Grey Federal 2004 36.2 63.8 State 2006 45.5 54.5 remains LIB swing to ALP 9.3 -9.3

Hindmarsh Federal 2004 50.1 49.9 State 2006 61.8 38.2 remains ALP swing to ALP 11.7 -11.7

Kingston Federal 2004 49.9 50.1 State 2006 64.0 46.0 changes hands to the ALP swing to ALP 14.1 -14.1

Makin Federal 2004 49.1 50.9 State 2006 65.3 34.7 changes hands to the ALP swing to ALP 16.2 -16.2

Mayo Federal 2004 36.4 63.6 State 2006 45.9 54.1 remains LIB swing to ALP 9.5 -9.5

Port Adelaide Federal 2004 62.9 37.1 State 2006 77.4 22.6 remains ALP swing to ALP 14.5 -14.5

Sturt Federal 2004 43.2 56.8 State 2006 54.2 45.8 changes hands to the ALP swing to ALP 11.0 -11.0

Wakefield Federal 2004 49.3 50.7 State 2006 62.7 37.3 changes hands to the ALP swing to ALP 13.4 -13.4

WHOLE STATE Federal 2004 45.6 54.4 State 2006 56.8 43.2 swing to ALP 11.2 -11.2

SOURCE: my calculations based on applying the State 2006 results to Federal boundaries.

Table 34 also shows the size of the swing that would be registered in each Federal electorate if each South Australian voter did cast his or her vote at the coming Federal election in the same way as he or she did at the State election of 2006. There would be a swing of 11.2% to Labor across the State, on a two party preferred basis, but that swing would be amplified in Kingston (which shows a swing of 16.2% to the ALP) and Wakefield (13.4% towards the ALP). The swing would be dampened in Barker (which shows a swing of only 5.7%) and also in Boothby (which swings by only 7.3%).

Another way of understanding these figures is to see the difference between the Federal and the State results – which I have labelled the “swing” – as a measure of the size of the group of voters who are least “rusted on” – they have voted ALP at the State election, but not Federally. In the Federal seat of Adelaide, for example, about 11% of the voters preferred the LIB candidate at the last Federal election but nonetheless voted ALP at the State election of 2006. In Makin, Kingston, Port Adelaide and Wakefield, about 13 to 15% of the voters did not vote ALP at the last Federal election but nonetheless have some recent experience of voting ALP.

This group of voters who did not vote ALP in 2004 but who nonetheless can vote Labor, is who will decide whether the South Australian seats change hands. What proportion of this group would need to vote ALP at the coming election for the Federal seats to change hands to Labor?

• In Kingston, fewer than 2 in 100 (1.4%) of that group of voters who did not vote Labor at the last Federal election but did vote ALP at the State election of 2006, would need to vote Labor, for the seats to change hands from LIB to the ALP. • In both Makin and Wakefield, only about 1 in every 16 (6%) of that group would need to vote Labor for the seat to change hands from LIB to the ALP. • In Sturt, 2 out of 3 (63%) of that group would need to vote Labor for the seat to change hands from LIB to the ALP. • In Boothby, 3 in every 4 (75%) of this group would need to vote Labor at the coming election for the seat to change hands to Labor.

What about the Senate vote? If voters in South Australia did choose to vote at the coming Federal election in much the same way as they voted at the recent State election (and the same parties contested the Senate election), then the result in the Senate would be that South Australians would return 2 LIB, 2 ALP, 1 Greens and 1 IND Nick Xenophon Members to the Senate. But the No Pokies team will not contest the Federal election, so who would win that last seat?

The assumptions are important to this calculation. The Nick Xenophon No Pokies ticket preferenced to Ralph Clarke, but I could not just transfer all of the Nick Xenophon No Pokies group’s votes to Ralph Clarke. Ralph Clarke did not win many first preference votes at all in his own right and it would be too much to think that if Nick Xenophon had not contested the election then that enormous group of voters - 21% of all first preferences - would have all preferred Ralph Clarke to any other party. So I first recalculated the outcome assuming that the votes that No Pokies actually won in 2006 were won by all of the other parties in the same proportion as Nick Xenophon’s below-the-line votes were allocated i.e. 20% to the ALP, 17% to LIB, 14% to the Greens, etc, and the result then was 3 ALP, 2 LIB 1 DEM.

Secondly, I recalculated the result allocating all of the votes for the Nick Xenophon IND No Pokies group to each other party or IND candidate in accordance with their share of the first preference vote overall (i.e. on a pro rata basis), and the result was 3 ALP, 2 LIB, 1 Family First.

So if South Australians vote for the Senate in a similar way to our vote for the Legislative Council, it seems fairly clear that we would return 3 ALP and 2 LIB Senators and then one from either the Democrats, Family First or perhaps even the Greens.

Conclusion

What does this all mean for the Parliament and the State? Is it possible to assess an election?

I think we can try, by asking questions including these:

Did voters understand what they were required to do? Did the outcome reflect what the voters wanted? Did the election produce a government with a workable majority? Have we learned something from this election?

Voters’ knowledge. To play a meaningful part in the elector process, voters have two tasks: to attend, and to lodge a valid vote. South Australians are happier to attend a polling booth at a Federal election than a State election. It is not clear just why this is the case, although it happens in Victoria as well. There is scope here for research on voters’ attitudes.

Lodging a valid vote is partly a matter of knowledge and partly a matter of attitude. Each House of Assembly ballot paper says

“Number the squares from 1 to (the number of candidates) in the order of your choice.”

Voters are not aware of the House of Assembly ticket vote provisions, and it is illegal to recommend that people rely on them, so the vast majority of the 43,500 voters who marked a ballot paper with an incomplete vote did it by mistake. There is an over-representation of these votes in some urban electorates, so useful research could look at whether the cause might be lower educational levels, lack of English language or other factors. Once more is known, voter education programmes might be tailored for these areas.

Each ballot paper (Assembly and Council) also carries the notice

“You are not legally obliged to mark the ballot paper.”

In 2006 nearly 40,000 voters lodged ballot papers which were quite blank. It is true that no- one is legally required to vote, just to attend at a polling booth and to take a set of ballot papers. But if we like to characterise our system as “compulsory voting” when it is actually one of “voluntary voting with compulsory-attendance”, that is because there is higher aim, even more important than compulsory voting, that prevents us from implementing a compulsory voting system. That higher aim is the secrecy of the ballot. Our system does not allow a ballot paper to be numbered or otherwise marked to identify who the voter is. So if no ballot paper can be linked to a particular voter, it is not possible to tell whether an elector has voted or not, compulsory voting cannot be implemented and the best we can do is to approximate compulsory voting through compulsory attendance. To require electors to attend and then tell them that they need not vote subverts the intention of the system and makes a mockery of the effort required to encourage voters to turn out. No other State or Territory has considered such a notice. If we are serious about wanting to hear every voter’s voice, it is not helpful to say that silence is acceptable.

So there is scope for the Parliament to consider the contradictory effects of the ticket vote which rescues incomplete votes in secret, and the voluntary voting notice on each ballot paper which condones completely blank ballot papers.

Voters’ wishes. Did the election give voters what they wanted?

Just over 50% of voters are now represented in the House of Assembly by the candidate who was their first choice, and over 60% are represented by a candidate who was either their first choice or the more acceptable of the final two candidates. 77 On this basis, we can say that the outcome of the House of Assembly election does reflect most voters’ intentions.

In the Legislative Council, 92% of voters elected at least one candidate from the party of their first choice, to represent them in the Upper House. The outcome of the Legislative Council election therefore also clearly reflects voters’ intentions very well, but it is actually harder to say that the composition of the Council represents the wishes of the voters. Certainly the members who were elected in 2006 reflected the wishes of the voters in 2006, but the current composition of the Council includes 11 Members who were elected as a result of a different election, 4 years before. From this point of view, the Premier’s proposal to elect all of the Members of the Council at the same time would have the effect of more closely representing voters’ intentions, as long as we retain the State-wide electorate and proportional representation count.

A government with a workable majority. South Australia has had minority governments for 12 of the past 18 years. The Rann ALP government went into the 2006 election as a minority government, holding office with the support of two Independents and a NAT Member, and emerged from the election with a clear majority of seats. In the 47-seat House of Assembly the ALP now hold 28 seats, and the Rann ALP government can act to implement its programme with confidence that it will remain in government until at least the next election.

The government does not hold the balance of power in the Legislative Council, and indeed no government has had this advantage since 1970. This is partly because voters choose to vote differently for the Upper and Lower Houses, and partly because voters can only elect half of the Members of the Legislative Council at each State election. While successive governments have worked hard over the past 35 years or so to provide a fairer electoral system for the Council, introducing full franchise, proportional representation and a Statewide electorate for the Council, it is nearly 100 years since all of the seats in the Council have been filled at the same time.

It is nonetheless true that even if we had elected all 22 MLCs at the 2006 State election, the ALP would not have won a majority of the seats in the Legislative Council, and the government would not have a majority in both Houses. While this situation may be frustrating to governments eager to implement the programmes on which they have been elected, the fact that successive governments have managed their legislative programmes without Council majorities since the 1970s78 indicates that having a majority in both Houses is an aim rather than a necessity, and should not in itself be a reason for any government to change the electoral system.

Learning from this election. Democracy is well served in this State by the South Australian State Electoral Office’s policy to provide comprehensive and informative election data and post-election survey information. This information is vital to any analysis of the operation of the current system, as well as to an analysis of the possible effects of changes to the current electoral system.

The calculations which I have made for the various segments of this paper have relied upon the SEO figures. The calculations show that How To Vote cards have less effect than generally thought, although Family First How To Vote cards did affect the outcome in two seats – Mitchell and Stuart. The paper has shown that ticket votes in the Assembly are more common in electorates with lower educational levels (but that they have almost no effect on the outcome anyway), and that voters turn out more readily at Federal than at State elections. It has shown that if we had elected all 22 Members of the Legislative Council at the State

77 Both of these figures are slight underestimates, as I should really exclude those people who lodged blank ballot papers, but the figures in relation to blank ballot papers figures are all post-election survey figures, not actual counts by electorate. 78 Actually no postwar ALP government has had a majority in the Upper House.

election of 2006 the current Council would now have 5 members of the IND No Pokies group (and two fewer LIB MLCs and no Democrats).

The paper has also confirmed that country seats do swing less than city seats, and that the major parties – and in particular the Liberal Party – have been losing market share over the past 20 years. It ahs also shown that between 6% and 16% of voters in federal seats preferred LIB at the Federal election of 2004 but have voted Labor since, and that if the current polls are right the ALP is likely to win 8 of the 11 federal seats in South Australia.

Looking towards the next State election, in 2010, the paper has noted that the Labor Party will need to defend not only its own marginal seats but also 15 Labor seats which are so safe that they are vulnerable to splitting by IND or IND Labor candidates, and that as well as the marginal ALP and LIB seats, Mt Gambier could be a very interesting seat to watch on election night.

Attachment 1

The following section updates material originally published in Parliamentary Library Information Paper No. 17: The South Australian Legislative Council: possible changes to its electoral system produced in 2000.

HOW DOES THE LEGISLATIVE COUNCIL VOTING SYSTEM WORK?

Voting The Legislative Council ballot paper enables electors to vote for one party (an above-the-line vote, also known as a ticket vote) or to vote for every individual candidate (a below-the-line vote, also known as a preferential vote). If the elector chooses to vote above-the-line he or she is only required to mark one box, but voting below-the-line requires a mark in every box in order of preference.

Before election-day, each party (and each independent candidate) informs the State Electoral Office of the order in which it wants its first preference votes to be distributed to every other candidate contesting the election. This order of preferences is called the ticket, and is used to distribute the party’s votes in the event that the party does not gain enough votes to continue in the count.

It is permissible to lodge two tickets. The Democrats have always lodged two tickets, usually differing only at the point where their preferences would be transferred to the ALP or the Liberal Party. The effect of a party lodging two tickets is that half of that party’s preferences follow the first ticket and half follow the second.

Large posters showing the order of preferences for each ticket lodged by each party or Independent candidate, are displayed at polling stations on election day, and they are also published on the State Electoral Office’s website before the election.

When a voter chooses to give his vote to a party or Independent using an above-the-line vote, he is accepting that party’s (or Independent’s) choice in relation to the order of preference distribution. A voter who does not accept that choice, can only specify the order in which he prefers his vote to be distributed by voting below-the-line and numbering each candidate.

On average since 1989, 96% of formal ballot papers have been marked with above-the-line votes, and only 4% of ballot papers have specified each candidate.

Counting - The quota In order to be elected to the Legislative Council, a candidate must receive a quota of votes. The quota is calculated as:

(total formal votes / number of seats available, plus one) plus one. This number is rounded down to a whole number.

When we elect 11 Members to the Legislative Council, the quota is about 8.3% of total formal votes. If the number of Members was to grow, then the quota would be lower (and vice versa).

A candidate does not need to receive a whole quota of first preference votes to win a seat; additional preference votes can accrue to the candidate as other candidates with fewer votes are excluded during the count. This means that it is quite possible for a candidate with only a fraction of a quota of first preferences to still win a seat. Under the current system where we elect 11 candidates, it is possible for a candidate with only about a third of a quota (25 to 26,000 votes in 2006) to win enough preference votes from other candidates to finally win a whole quota (and hence a seat).

- the process of the count The Legislative Council count uses a ‘top-down bottom-up’ method.

Each party nominates its candidates in an order which reflects the importance which the party places on having that candidate win a seat – the most important candidate is at the top of the party’s list. All ticket votes for a party are allocated to the top candidate on the party’s list, as well as any first preferences gained by that candidate from below-the-line votes. Second and subsequent candidates on the party’s list are shown at this stage as having only those first preference votes which they gained from below-the-line votes.

Once all formal first preference votes have been counted and allocated to individual candidates, the quota can be calculated, and candidates who have gained a quota or more can be declared elected. If a party or IND candidate wins one quota it wins one seat and the first person on the party’s list of candidates is declared elected.

Their excess first preference votes are then distributed to another candidate. The above-the-line votes flow to the party’s next candidate in the order dictated by the party list, or if the party has no more candidates then to the next candidate on the party’s ticket. Where a party has lodged two tickets, every second preference vote is distributed according to the second ticket. Below-the-line votes flow in the order specified by each voter on each ballot paper.

When the excess votes of those candidates who have been elected, have all been distributed to other candidates, the count moves to the second stage, which is often referred to as the “bottom-up” stage. At this point the candidate with the lowest number of votes is determined to be the least-preferred candidate, and is excluded from the remainder of the count. Once again these votes are allocated to continuing candidates either according to the candidate’s ticket or (in the case of below-the-line ballot papers) according to the order specified on each ballot paper. This process of excluding the candidates with the least support continues until a continuing candidate has received enough preference votes to win a quota (at which stage that candidate is declared elected) and then resumes until all seats have been won.

When a candidate has achieved a quota and there are excess votes that need to be distributed to other candidates, it could be distorting to the count if the votes that are transferred are not representative of all of the votes that are held by the candidate. For this reason, all of the candidate’s votes are transferred, but at a fraction of their value, and this fraction is called the transfer value. The transfer value is calculated as: surplus first preference votes/total first preference votes.

While the proportional representation count does not produce results that are exactly proportional to the first preference vote, it does produce results that are closer to strict proportionality than any other commonly-used voting system.

How many votes are required for a candidate to be elected? A candidate needs to win a quota of votes (about 8.3% if we elect 11 Members) to win a seat. Not all of these votes need to be first preference votes.

Under the current arrangement where 11 Members are elected at each Legislative Council election, a successful candidate requires at least enough first preference votes to win about a third of a quota. In order to continue in the count for long enough to receive preference votes from other candidates as they are gradually excluded, the successful candidate cannot be the least- preferred candidate at the start of the count (to be fair, this will usually be guaranteed by winning a third of a quota). Finally, the successful candidate must receive the remaining portion of his or her quota as preferences from other candidates, so must be placed higher on their tickets than other candidates who continue in the count. In order to continue in the count long enough to receive extra preference votes, a candidate needs to be placed high on the tickets of those candidates who are excluded early in the count (generally the minor IND candidates) as well

as candidates with larger parcels of votes that may be transferred (generally the smaller parties such as the Democrats, Greens, Family First).

Each of these requirements will be played out differently at each election.

Winning more than a third of a quota is no guarantee of winning a seat if preferences do not flow towards you. In 2006 the ALP received only about 1200 preference votes during the second stage of the count because preferences from Independents and minor parties consistently flowed to other IND or minor party candidates rather than to the ALP.

Similarly, being placed high on the ticket of other candidates or parties is no guarantee of receiving their preference votes, because they may not be excluded before you are. For example, in 2006 the Greens and IND No Pokies were not excluded so anyone hoping to gain from the transfer of their votes was disappointed.

How many votes are required for a party to win the majority of seats? Members of the Legislative Council each serve an eight-year term; 11 of the 22 Members are elected each 4 years at the same time as the election for House of Assembly Members.

The lagged election cycle for the Legislative Council means that a party needs to win 6 Council seats at two consecutive elections in order to have control of the House.

To win 6 of the 11 available seats, a party needs to win 6 quotas, which works out at slightly over 50% of the total number of formal votes. A party could win these 6 quotas by winning 50% (or more) of the first preferences or, at a minimum about 44.2% of all formal first preference votes and then making up the additional 5.8% of votes in second or subsequent preferences of excluded candidates.79

Winning 50% of the vote in order to win a majority of the seats is not an unreasonable task – we expect that in the House of Assembly elections. But in the Legislative Council a party must do this for two elections in a row to win the majority of seats. No party has ever managed to do this – the ALP won 6 seats in 1975, LIB won 6 in 1979 and again in 1993 but these wins required large swings towards the parties at the time and the swings could not be repeated at a subsequent election.

On the evidence of recent history, no government can expect to have a majority in the Legislative Council during its first term.

Can we model the Legislative Council voting system? It has already been pointed out that over the past six State elections for the Legislative Council an average of 96% of formal ballot papers were marked with above-the-line votes, and only 4% of ballot papers specified each candidate. Therefore it should be possible to model the Legislative Council vote as if every voter’s preferences followed his or her chosen party’s ticket.

I have calculated what the outcome would have been if all of the first preference votes had been above-the-line votes, and therefore all preference flows followed each party or IND candidate’s ticket. The exercise produces the same successful candidates as the actual count, in each election from 1985 to 2006.

Therefore I think this very simple ticket-vote model can be used as the basis for an analysis of hypothetical changes to the electoral system.

79 5 quotas (= 41.7% of formal first preferences) plus about 0.3 of a quota (= 2.5% of first preferences) to remain in the count, totals 44.2% of formal first preference votes.

Attachment 2

Distribution of ALP preferences, 2006

ALP ALP share ALP Actual Distribution of ALP Preferences first of all first Vote Card went to: went to: preferences preferences preferenced NAT / IND LIB NAT / IND LIB (No.) (%) to: (No.) (No.) (%) (%)

Chaffey 1,914 9.8 NAT 1,587 327 82.9 17.1 Flinders 2,380 12.6 NAT 1,823 557 76.6 23.4 Mt Gambier 4,582 22.2 IND 3,502 1,080 76.4 23.6

These 3 seats 8,876 15.0 6,912 1,964 77.9 22.1 ALL HA SEATS 424,715 45.2

SOURCE: unpublished worksheets from the State Electoral Department.

Distribution of LIBERAL PARTY preferences, 2006

LIB LIB share LIB Actual Distribution of LIB Preferences first of all first Vote Card went to: went to: preferences preferences preferenced ALP IND ALP IND (No.) (%) to: (No.) (No.) (%) (%)

Fisher 3,774 18.5 IND 416 3,358 11.0 89.0 Mitchell 4,094 20.7 IND 573 3,521 14.0 86.0

These 2 seats 7,868 19.6 989 6,879 12.6 87.4 ALL HA SEATS 319,041 34.0

SOURCE: unpublished worksheets from the State Electoral Department.

Distribution of NATIONAL PARTY preferences to ALP and LIB, 2006

NAT NAT share NAT Actual Distribution of Preferences first of all first Vote Card went to: went to: preferences preferences preferenced ALP LIB ALP LIB (No.) (%) to: (No.) (No.) (%) (%)

Chaffey (a) 10358 53.2 LIB 2920 7438 28.2 71.8 Finniss 3,622 17.5 LIB 1,034 2,588 28.5 71.5 Flinders (a) 4,596 24.3 LIB 930 3666 20.2 79.8 Mackillop (a) 917 5.3 LIB 211 706 23.0 77.0

All contested 19493 24.6 ALL SEATS 19493 2.1 5095 14398 26.1 73.9

(a) The final 2 candidates in these seats were not ALP and LIB but a rethrow of ballot papers was conducted to find a 2PP result. (b) figures for Mackillop are just for votes lodged at booths. SOURCE: unpublished worksheets from the State Electoral Office.

Attachment 3 Distribution of DEMOCRAT preferences to ALP and LIB, 2006

DEM DEM share DEM How To Actual Distribution of DEM Preferences first of all first Vote Card went to: went to: preferences preferences preferenced ALP LIB ALP LIB (No.) (%) to: (No.) (No.) (%) (%)

Adelaide 558 2.9 split 383 175 68.6 31.4 Ashford 928 4.6 split 553 375 59.6 40.4 Bragg 864 4.3 split 472 392 54.6 45.4 Bright 395 2.0 split 245 150 62.0 38.0 Chaffey (a) 271 1.4 split 115 156 42.4 57.6 Cheltenham 842 4.4 split 592 250 70.3 29.7 Colton 560 2.7 split 363 197 64.8 35.2 Croydon 781 4.0 split 426 355 54.5 45.5 Davenport 927 4.7 split 570 357 61.5 38.5 Elder 564 2.9 split 347 217 61.5 38.5 Enfield 673 3.5 split 456 217 67.8 32.2 Finniss 544 2.6 split 330 214 60.7 39.3 Fisher (a) 291 1.4 split 181 110 62.2 37.8 Flinders (a) 316 1.7 split 178 138 56.3 43.7 Florey 636 3.4 split 423 213 66.5 33.5 Frome 393 2.0 split 170 223 43.3 56.7 Giles no DEMOCRAT candidate Goyder 523 2.5 split 263 260 50.3 49.7 Hammond 739 3.6 split 405 334 54.8 45.2 Hartley 562 2.9 split 366 196 65.1 34.9 Heysen 1,044 5.3 split 657 387 62.9 37.1 Kaurna 423 2.2 split 275 148 65.0 35.0 Kavel 709 3.6 split 412 297 58.1 41.9 Lee 603 3.0 split 377 226 62.5 37.5 Light 370 1.8 split 215 155 58.1 41.9 Little Para 486 2.4 split 307 179 63.2 36.8 Mackillop (b) 425 2.5 split 178 247 41.9 58.1 Mawson 476 2.3 split 251 225 52.7 47.3 Mitchell (a) 344 1.7 split 236 108 68.6 31.4 Morialta 606 3.0 split 377 229 62.2 37.8 Morphett 649 3.1 split 362 287 55.8 44.2 Mt Gambier (a) 347 1.7 split 229 118 66.0 34.0 Napier 736 3.7 split 407 329 55.3 44.7 Newland 646 3.2 split 373 273 57.7 42.3 Norwood 545 2.8 split 364 181 66.8 33.2 Playford 759 3.7 split 547 212 72.1 27.9 Port Adelaide 499 2.5 split 319 180 63.9 36.1 Ramsay 693 3.6 split 471 222 68.0 32.0 Reynell 472 2.4 split 314 158 66.5 33.5 Schubert 754 3.6 split 413 341 54.8 45.2 Stuart 353 1.7 split 219 134 62.0 38.0 Taylor 475 2.4 split 298 177 62.7 37.3 Torrens 672 3.4 split 367 305 54.6 45.4 Unley 609 3.1 split 402 207 66.0 34.0 Waite 957 4.6 split 584 373 61.0 39.0 West Torrens 660 3.3 split 402 258 60.9 39.1 Wright 428 2.0 split 254 174 59.3 40.7

All contested seats 27,107 2.9 16,448 10,659 60.7 39.3 ALL HA SEATS 27,107 2.9 16,448 10,659 60.7 39.3

(a) The final 2 candidates in these seats were not ALP and LIB so a rethrow of ballot papers was conducted to find a 2PP result. (b) figures for Mackillop are just for votes lodged at booths. SOURCE: unpublished worksheets from the State Electoral Office.

Attachment 4 Distribution of GREENS preferences to ALP and LIB, 2006

GREENS Greens Greens How Actual Distribution of Greens Preferences first of all first Vote Card went to: went to: preferences preferences preferenced ALP LIB ALP LIB (No.) (%) to: (No.) (No.) (%) (%)

Adelaide 1,797 9.3 ALP 1,515 282 84.3 15.7 Ashford 1,828 9.1 ALP 1,531 297 83.8 16.2 Bragg 1,854 9.3 ALP 1,429 425 77.1 22.9 Bright 1,315 6.5 ALP 1,033 282 78.6 21.4 Chaffey (a) 475 2.4 ALP 303 172 63.8 36.2 Cheltenham 1,140 6.0 ALP 913 227 80.1 19.9 Colton 1,191 5.8 ALP 976 215 81.9 18.1 Croydon 1,869 9.6 ALP 1,332 537 71.3 28.7 Davenport 2,250 11.4 ALP 1,726 524 76.7 23.3 Elder 1,396 7.3 ALP 1,088 308 77.9 22.1 Enfield 1,190 6.2 ALP 956 234 80.3 19.7 Finniss 1,586 7.7 ALP 1,210 376 76.3 23.7 Fisher (a) 699 3.4 ALP 591 108 84.5 15.5 Flinders (a) 773 4.1 ALP 460 313 59.5 40.5 Florey 1,083 5.7 ALP 836 247 77.2 22.8 Frome 750 3.7 ALP 543 207 72.4 27.6 Giles 854 4.5 ALP 569 285 66.6 33.4 Goyder 928 4.5 ALP 633 295 68.2 31.8 Hammond 1,476 7.2 ALP 1,019 457 69.0 31.0 Hartley 1,289 6.7 ALP 1,025 264 79.5 20.5 Heysen 3,508 17.7 ALP 2,811 697 80.1 19.9 Kaurna 1,444 7.3 ALP 1,172 272 81.2 18.8 Kavel 1,951 10.0 ALP 1,587 364 81.3 18.7 Lee 1,537 7.6 ALP 1,239 298 80.6 19.4 Light 1,027 5.0 ALP 706 321 68.7 31.3 Little Para 912 4.6 ALP 705 207 77.3 22.7 Mackillop (b) 693 4.2 ALP 439 254 63.3 36.7 Mawson 1,019 5.0 ALP 750 269 73.6 26.4 Mitchell (a) 665 3.4 ALP 525 140 78.9 21.1 Morialta 1,296 6.3 ALP 982 314 75.8 24.2 Morphett 1,594 7.6 ALP 1,188 406 74.5 25.5 Mt Gambier (a) 381 1.8 LIB 215 166 56.4 43.6 Napier 1,114 5.6 ALP 858 256 77.0 23.0 Newland 1,075 5.2 ALP 804 271 74.8 25.2 Norwood 1,679 8.5 ALP 1,378 301 82.1 17.9 Playford 1,039 5.0 ALP 781 258 75.2 24.8 Port Adelaide 1,289 6.6 ALP 1,015 274 78.7 21.3 Ramsay 938 4.9 ALP 690 248 73.6 26.4 Reynell 942 4.7 ALP 708 234 75.2 24.8 Schubert 1,551 7.5 ALP 1,136 415 73.2 26.8 Stuart 737 3.6 ALP 555 182 75.3 24.7 Taylor 776 3.9 ALP 531 245 68.4 31.6 Torrens 1,274 6.5 ALP 951 323 74.6 25.4 Unley 1,937 9.8 ALP 1,569 368 81.0 19.0 Waite 2,222 10.6 ALP 1,758 464 79.1 20.9 West Torrens 1,644 8.3 ALP 1,324 320 80.5 19.5 Wright 817 3.9 ALP 585 232 71.6 28.4

All contested seats 60,804 6.5 46,650 14,154 76.7 23.3 ALL HA SEATS 60,804 6.5 46,650 14,154 76.7 23.3

(a) The final 2 candidates in these seats were not ALP and LIB but a rethrow of ballot papers was conducted to find a 2PP result. (b) figures for Mackillop are just for votes lodged at booths. SOURCE: unpublished worksheets from the State Electoral Office.

Attachment 5 Distribution of FAMILY FIRST preferences to ALP and LIB, 2006

FF FF share FF How To Actual Distribution of Preferences first of all first Vote Card went to: went to: preferences preferences preferenced ALP LIB ALP LIB (No.) (%) to: (No.) (No.) (%) (%)

Adelaide 517 2.7 LIB 154 363 29.8 70.2 Ashford 1,002 5.0 LIB 386 616 38.5 61.5 Bragg 769 3.9 LIB 201 568 26.1 73.9 Bright 905 4.5 LIB 309 596 34.1 65.9 Chaffey (a) 970 5.0 LIB 240 730 24.7 75.3 Cheltenham 1,895 9.9 LIB 932 963 49.2 50.8 Colton 820 4.0 LIB 321 499 39.1 60.9 Croydon no Family First candidate Davenport 897 4.5 LIB 219 678 24.4 75.6 Elder 1,081 5.6 LIB 411 670 38.0 62.0 Enfield 1,094 5.7 ALP 718 376 65.6 34.4 Finniss 1,168 5.6 LIB 343 825 29.4 70.6 Fisher (a) 1,037 5.1 LIB 386 651 37.2 62.8 Flinders (a) 785 4.2 LIB 115 670 14.6 85.4 Florey 1,410 7.4 LIB 490 920 34.8 65.2 Frome 1,038 5.2 LIB 398 640 38.3 61.7 Giles 1,106 5.8 LIB 468 638 42.3 57.7 Goyder 1,558 7.6 LIB 537 1,021 34.5 65.5 Hammond 1,907 9.3 LIB 578 1,329 30.3 69.7 Hartley 814 4.2 LIB 267 547 32.8 67.2 Heysen 996 5.0 LIB 271 725 27.2 72.8 Kaurna 1,328 6.8 LIB 491 837 37.0 63.0 Kavel 3,081 15.7 LIB 915 2,166 29.7 70.3 Lee 920 4.6 ALP 537 383 58.4 41.6 Light 1,403 6.8 split 455 948 32.4 67.6 Little Para 1,759 8.8 LIB 628 1,131 35.7 64.3 Mackillop (b) 1,214 7.0 LIB 258 956 21.3 78.7 Mawson 1,330 6.6 split 447 883 33.6 66.4 Mitchell (a) 1,062 5.4 split 429 633 40.4 59.6 Morialta 1,174 5.7 split 496 678 42.2 57.8 Morphett 868 4.1 LIB 262 606 30.2 69.8 Mt Gambier (a) 842 4.1 LIB 220 622 26.1 73.9 Napier 2,122 10.7 ALP 1,350 772 63.6 36.4 Newland 1,438 7.0 split 616 822 42.8 57.2 Norwood 508 2.6 LIB 188 320 37.0 63.0 Playford 1,686 8.2 ALP 1,101 585 65.3 34.7 Port Adelaide 1,120 5.7 ALP 714 406 63.8 36.3 Ramsay no Family First candidate Reynell 1,755 8.7 LIB 672 1,083 38.3 61.7 Schubert 1,630 7.9 LIB 489 1,141 30.0 70.0 Stuart 943 4.6 LIB 310 633 32.9 67.1 Taylor 1,694 8.5 ALP 1,114 580 65.8 34.2 Torrens 1,439 7.3 split 725 714 50.4 49.6 Unley 495 2.5 LIB 140 355 28.3 71.7 Waite 895 4.3 LIB 233 662 26.0 74.0 West Torrens 988 5.0 ALP 588 400 59.5 40.5 Wright 1,541 7.3 LIB 521 1,020 33.8 66.2

All contested seats 55,004 6.1 21,643 33,361 39.3 60.7 ALL HA SEATS 55,004 5.9 21,643 33,361 39.3 60.7

(a) The final 2 candidates in these seats were not ALP and LIB but a rethrow of ballot papers was conducted to find a 2PP result. (b) figures for Mackillop are just for votes lodged at booths. SOURCE: unpublished worksheets from the State Electoral Office.

Attachment 6

Distribution of DIGNITY FOR DISABLED preferences to ALP and LIB, 2006

D4D D4D share D4D Actual Distribution of Preferences first of all first Vote Card went to: went to: preferences preferences preferenced ALP LIB ALP LIB (No.) (%) to: (No.) (No.) (%) (%)

Adelaide 441 2.3 split 220 221 49.9 50.1 Bright 492 2.4 split 261 231 53.0 47.0 Hartley 323 1.7 split 191 132 59.1 40.9 Mawson 469 2.3 split 234 235 49.9 50.1 Mitchell (a) 438 2.2 split 296 142 67.6 32.4 Morialta 418 2.0 split 260 158 62.2 37.8 Newland 251 1.2 split 131 120 52.2 47.8 Norwood 274 1.4 split 148 126 54.0 46.0 Unley 362 1.8 split 193 169 53.3 46.7 Wright 506 2.4 split 261 245 51.6 48.4

All contested seats 3,974 2.0 2,195 1,779 55.2 44.8 ALL HA SEATS 3,974 0.4 2,195 1,779 55.2 44.8

(a) The final 2 candidates in this seat were not ALP and LIB but a rethrow of ballot papers was conducted to find a 2PP result. SOURCE: unpublished worksheets from the State Electoral Office.

Distribution of ONE NATION preferences to ALP and LIB, 2006

ON ON share ON Actual Distribution of ON Preferences first of all first Vote Card went to: went to: preferences preferences preferenced ALP LIB ALP LIB (No.) (%) to: (No.) (No.) (%) (%)

Ashford 205 1.0 LIB 94 111 45.9 54.1 Goyder 552 2.7 LIB 246 306 44.6 55.4 Hammond 845 4.1 LIB 290 555 34.3 65.7 Kavel 373 1.9 LIB 183 190 49.1 50.9 Newland 306 1.5 ALP 137 169 44.8 55.2 Port Adelaide 310 1.6 LIB 141 169 45.5 54.5

All contested seats 2591 2.1 1091 1500 42.1 57.9 ALL HA SEATS 2591 0.3 1091 1500 42.1 57.9

SOURCE: unpublished worksheets from the State Electoral Office.

Bibliography

Statutes Constitution Act 1934 (SA) Constitution (Parliamentary Terms) Amendment Act 2001 (SA) Constitution Act 1902 (NSW) Electoral Act 1985 (SA) Electoral Act Regulations 1985 (SA) Electoral Act 2002 (Vic) Electoral Act 2004 (Tas) Electoral Act 1992 (Qld)

Bills Constitution Act Amendment Bill 1990 (Gilfillan) Constitution (Parliamentary Terms) Amendment Bill 1991 (M.Evans) Constitution (Parliamentary Terms) Amendment Bill 1993 (Gilfillan) Constitution (Parliamentary Terms) Amendment Bill 1997 (Elliott) Constitution (Parliamentary Terms) Amendment Bill 1999 (Hanna) Constitution (Parliamentary Terms) Amendment Bill 2001 (Hanna)

Websites www.aec.gov.au www.ecq.qld.gov.au www.elections.act.gov.au www.elections.nsw.gov.au www.electionswa.com.au www.electoral.tas.gov.au www.ntec.nt.gov.au www.seo.sa.gov.au www.vec.vic.gov.au

Other ACT Electoral Commission, 2006, Electoral Compendium 1989 to 2006, ACT Electoral Commission, Canberra.

Anderson, L., 2006, “It's envy, says maverick; Knives out as Mr X's tactics anger rivals” The Advertiser, 20 February 2006, at p.3.

Australia. Joint Standing Committee on Electoral Matters, 1997, The 1996 Federal Election: Report of the Inquiry into the conduct of the 1996 Federal Election and matters related thereto, AGPS, Canberra.

Australian Bureau of Statistics, SEIFA Indexes at www.abs.gov.au

Bennett, S., 2006, “Understanding State elections: South Australia and Tasmania 2006”, Commonwealth Parliamentary Library Research Brief No. 17 of 2005-6.

Green, A., 2006, “Western Australian State Election 2005”, Western Australian Parliamentary library Election Papers Series, 2/2005.

Green, A., 2007, “2007 New South Wales Election; Preliminary analysis”, NSW Parliamentary Library Research Service Background Paper No 2/07.

Jaensch, DH, 1971, “Under-Representation and the ‘Gerrymander’ in the Playford Era”, Australian Journal of Politics and History, Vol. 17 (1), April 1971, pp82-95.

Jaensch, DH, 1975, “Don’s Party? Labor in South Australia”, Current Affairs Bulletin, Vol. 52 (3), August 1975, pp4-14.

Jaensch, D, 1977, “The 1977 Elections in South Australia; A Statistical Analysis”, Politics Department Occasional Monograph No. 6, Flinders University.

Jaensch, D, 1979, “The 1979 Elections in South Australia; A Statistical Analysis”, Politics Department Occasional Monograph No. 7, Flinders University.

Jaensch, D, 1983, “The 1982 Elections in South Australia; A Statistical Analysis”, Politics Department Occasional Monograph No. 8, Flinders University.

Kelton, G., 2005, “Rann to call referendum for 2010; Abolish the Upper House”, The Advertiser, 24 November 2005, p.1.

Kelton, G., and L. Anderson, 2006, “Hard Ball on ‘Bible Belt’ deals” The Advertiser, 2 March 2006, p25.

Lewis, S., 2007, “We face annihilation: Howard” The Australian, 23 May 2007, p.1.

Macdonald, C., G. Gardiner and B. Costar, 2007, “The 2006 Victorian State Election”, Victorian Parliamentary Library Research Service Research Paper No. 1 of 2007.

Newton-Farrelly, J, 2002, “Report on the Results of the South Australian State Election, 9 February 2002”, Parliamentary Library Information Paper No. 19, Adelaide.

Northern Territory Electoral Commission, 2006?, 2005 Legislative Assembly General Election Report, Results and Statistics, NTEC, Darwin.

Rann, Hon M.D., “Ministerial Statement”, SA House of Assembly Hansard 24 November 2005, pp. 4130-4131.

Shanahan, D., 2007, “ALP in strongest position since '01”, The Australian, 6 February 2007, p.1

Shanahan, D., 2007, “Labor's lead hits record high - Opposition surges 14 points ahead despite slip in Rudd's popularity” The Australian, 6 March 2007, p.1

Shanahan, D., 2007 “Howard checks Rudd's march - Labor holds election-winning lead but voters embrace indigenous plan” The Australian, 10 July 2007, p.1

Shanahan, D., 2007 , “Howard humbled in polls as world leaders fly in for APEC summit - Rudd opens up crushing lead”, The Australian, 4 September 2007, p.1

SA. Electoral Districts Boundaries Commission, 1998, 1998 Report of the Electoral Districts Boundaries Commission, EDBC, Adelaide.

SA. Electoral Districts Boundaries Commission, 2007, 2007 Report of the Electoral Districts Boundaries Commission, EDBC, Adelaide.

SA. State Electoral Department, 1986, “General Elections 1985: Statistical Returns”, SA Parliamentary Paper 146 of 1986-87.

SA. State Electoral Department, 1991, “Periodical and General Elections 1989: Statistical Returns”, SA Parliamentary Paper 161 of 1991-92.

SA. State Electoral Office, 1996, Statistical Returns for General Elections 1993 and By- elections 1994, SEO, Adelaide.

SA. State Electoral Office, 1998, Statistical Returns: General Elections 11 October 1997, SEO, Adelaide.

SA. State Electoral Office, 1999, Election Report: General Elections 1997, SEO, Rose Park.

SA. State Electoral Office, 2002, How To Vote cards lodged by the candidates contesting elections for House of Assembly and Legislative Council, polling day 09/02/2002, SEO, Rose Park.

SA. State Electoral Office, 2003, Statistical Returns for the South Australian Elections 9 February 2002, SEO, Rose Park.

SA. State Electoral Office, 2003, Election Report for the South Australian Elections 9 February 2002, SEO, Rose Park.

SA. State Electoral Office, 2006, How To Vote cards lodged by the candidates contesting elections for House of Assembly and Legislative Council, polling day 18/03/2006, SEO, Rose Park. SA. State Electoral Office, 2006, House of Assembly Elections data; District and Statewide Summary figures, State Election 18 March 2006, (prepared for the Electoral Districts Boundaries Commission), SEO, Rose Park.

SA. State Electoral Office, 2006, Results and Outcomes, 3: State Election 18 March 2006, SEO, Rose Park.

SA. State Electoral Office, 2007, Election Report, SEO, Rose Park.

SA. State Electoral Department, unpublished worksheets from the 2002 and 2006 State elections showing the distribution of preferences.

SA. State Electoral Department, unpublished tables from the 2006 election entitled “Result of the Transfer and Distribution of Preferences”.