Newsletter of the Proportional Representation Society of Australia
December 1996, QN1996D

"Proposition H": San Francisco Narrowly Loses its PR Opportunity

The result of San Francisco's municipal referendum on "Proposition H" on 5th November 1996 (see QN1996C) was that a small majority of voters (54%) voted to replace the current winner take all, plurality, at-large system with a previously-used system of single-member electorates. A large minority of voters (46%) preferred instead quota-preferential proportional representation, which had never been used before in San Francisco and which its proponents called preference voting.

The result is not totally disheartening for the PR campaigners. Nearly half the voters were more impressed with PR, despite its never having been used in San Francisco, than they were with the familiar single-member plurality system - that unfortunate mainstay of nearly all US elections. It is also important that this support is for quota-preferential rather than party list PR.

The appeal in QN1996C enabled the PRSA to donate US$625 to the San Francisco campaign, for which we received their thanks. Major donors were the NSW Branch (A$200) and the Victorian Branch (A$100). Most of the balance came from the National Society's funds.

PRSA Submission to Joint Standing Committee on Electoral Matters

In October, the Proportional Representation Society of Australia finalized its submission to the Federal Parliament's Joint Standing Committee on Electoral Matters on matters related to the conduct of the 1996 federal elections. After consultation among National Office-bearers, a 54-page submission, The General Elections of 1996: Grounds for Re-examining some of the Fundamentals of our Electoral System, was forwarded by the PRSA President, Mr Bogey Musidlak.

The submission, full copies of which are available upon request, included a summary of each of its six major sections. An abridgement follows:


    Albert Langer's imprisonment during the 1996 federal election campaign ought to result Section 329A of the Commonwealth Electoral Act being repealed, and the definition of a formal vote being thoroughly reviewed.

    Electors have a single transferable vote in all our major public elections, but the proliferation of formality rules arbitrarily disregarding voters' intentions is inexcusable.

    Australian Electoral Commission research after the 1987 election showed that 1 in 5 attempts to vote below-the-line failed. In the trade-off between informal and exhausted votes, the PRSA comes down firmly on the side of respecting voters' expressed wishes, while strongly encouraging the marking of preferences to maximize the likelihood of achieving full vote effectiveness. If party boxes remain, there will be no risk of high levels of vote exhaustion under significantly-liberalized below-the-line requirements.


    Over 27% of Tasmanian voters marked their Senate ballot-papers below-the-line - a week after they had used Hare-Clark at their State polls. Over 16% of Senate voters in both the Territories voted below-the-line: in both those cases there were relatively few candidates. Australia-wide eleven elected senators obtained more than 10% of their first preference votes below-the-line.

    Altogether, 117 Group Voting Tickets were registered by 58 separate entities. In about a third of cases, parties or groups lodged two or three tickets. Sections 272 (4)-(5) of the Commonwealth Electoral Act anticipate constitutional problems with multiple tickets in the absence of individuals having voted directly for a single stream of candidates. It is unclear how the Court of Disputed Returns would deal with this contingency plan in the event of a challenge to the validity of registering multiple tickets. Western Australia, which has similar constitutional restrictions, allows only a single ticket to be registered in its PR Legislative Council elections.

    The PRSA calls upon the Committee to recommend remedial action before a future election is thrown into chaos by last-moment uncertainty arising from a legal challenge. Suggestions are also made about how to improve the quality of information about numbering in registered Group Voting Tickets that is available to voters before polling day.


    Proposals to divide Senate electorates into smaller regions will rightly attract strong criticism as they invite attempts at boundary manipulation that is presently unavailable to would-be manipulators of Senate elections. Such divisions would produce State and national outcomes inaccurately reflecting voters' wishes.

    Ideas of imposing an initial exclusion or qualifying threshold for eligibility to receive preferences from other candidates are incompatible with the current quota-preferential counting methods. Past experience with the now discontinued Dunstan version of PR in South Australia and Modified d'Hondt procedure in the ACT shows such inherently unstable schemes to be entirely without merit. On the 1996 election figures, both a 50% and 80% threshold would have provoked a major outcry about their arbitrary unfairness.


    Taking examples from various 1996 scrutinies, the PRSA shows how the definition of the transfer value is improved, to avoid unnecessary exhaustion of votes and the prospect of anomalies as awkward as finding some electors get more than one vote. The practical desirability of reintroducing provisions for the deferment of some surpluses is demonstrated.


    The Society suggests that an indicative countback be held to discourage State Parliaments from making an opportunistic choice of replacement in relation to Senate casual vacancies involving Independents or parties that have merged or become defunct.


    Use of Robson Rotation to rotate names within party columns would remove lottery effects from close House of Representatives contests, and would always lead to the candidate with the conscious support of more voters prevailing at the end. Robson Rotation is used for the election of Tasmanian MLCs in single-member districts.

First New Zealand MMP Elections

New Zealand's first MMP elections for its House of Representatives, on 12th October 1996, gave little comfort to the supporters of that country's previously exclusively single-member electorate plurality system, who had, at the 1994 referendum (QN72), lost their campaign for the retention of that system. The election results were as follows:

Party Constituency Votes Party Votes Constituency Seats Party Seats Total Seats Party Votes (%) Total Seats (%)
Nationals 698,920 700,687 30 14 44 33.82 36.67
Labour 640,796 584,113 26 11 37 28.19 30.83
NZ First 278,021 276,842 6 11 17 13.36 14.17
Alliance 231,913 209,319 1 12 13 10.10 10.83
ACT 77,334 126,421 1 7 8 6.10 6.67
United NZ 42,666 18,243 1 1 2 0.88 0.83
SUB-TOTAL (Seats won) 1,969,650 1,915,625 65 55 120 92.45 100.00
Christian Coalition 31,989 89,704

Legalise Cannabis 3,407 34,394

McGillicuddy Serious 12,178 5,989

Progressive Greens 7,438 5,287

Mana Maori 4,793 4,070

Animals First

Natural Law 5,383 3,189

Ethnic Minority

Green Society 1,140 2,363

Conservative 4,377 1,482

Superannuants 686 1,395

Advance NZ 650 949

Libertarians 553 671

Asia Pacific United 293 478

Te Tawharau 818 404

12 OTHER PARTIES 1,657 16,434

INDEPENDENTS SUB-TOTAL (No seats won) 75,362 156,431

TOTAL 2,045,012 2,072,056 65 55 120 100.00 100.00

The results for the 65 constituency seats (a majority of the seats in the House), which continue to be filled by that plurality system, failed to show a clear majority party as that system's supporters might have hoped. The largest number of constituency seats for a single party was the 30 of the 65 available obtained by the National Party. That suggests that the result under the old system would have been no different from MMP in that, to govern, the Nationals would still need a coalition partner.

The 4.33% voting for the Christian Coalition formed the largest single group of victims of the arbitrary 5% threshold below which party votes suddenly change from being able to ensure filling of at least 6 seats for a party to filling none. Contrary to quota-preferential PR practice, MMP's threshold prevented voters for the Christian Coalition from having any effect on which of the larger groupings would gain priority in being elected, thus unnecessarily narrowing the ultimate support base of those larger groupings and their legitimacy in gaining seats. That contrasts with the 6.10% voting for the Association of Consumers and Taxpayers (ACT) being rewarded with seven party seats. Had the Christian Coalition rather than ACT won a constituency seat, a bonus of four party seats would have been reaped, with each of the five biggest parties dropping a party seat.

ATSIC Regional Council Elections

When nominations for the elections in 35 ATSIC Regions held on 12th October 1996 closed one month earlier, there were 1269 candidates for 375 vacancies, including 420 females. Last-minute amendments had simplified the relationship between the number of elected representatives in a Region and its estimated population of Aborigines: the maximum was set at 12, instead of the previous limit of 20, where that population exceeded 10,000; while the minimum was made 8, rather than 10, for populations of below 1,000. The 92 ATSIC electoral districts created consisted of Regions of which ten were undivided, with each having from 9 to 12 representatives, while the remaining Regions were divided into Wards each Ward having from 1 to 8 representatives.

The incidence of female nomination ranged from 23% in the Northern Territory to 38-39% in New South Wales and Victoria. Average nominations per vacancy varied from 2.8 in the Northern Territory, South Australia and Tasmania to 4.7 in Queensland. Voters in the larger Wards and undivided Regions generally had the widest range of candidates to choose from, as shown below:

Classes of ATSIC electoral district in terms of no. of vacancies per district No. of ATSIC electoral districts with that no. of vacancies No. of vacancies per class of ATSIC electoral district Mean No. of candidates per ATSIC electoral district
1 17 17 3.6
2 14 28 6.6
3 17 51 8.6
4 15 60 12.9
5 10 50 15.2
6 4 24 26.0
7 4 28 25.3
8 1 8 16.0
9* 1 9 23.0
10* 3 30 30.3
11* 2 22 50.5
12* 4 48 36.3
TOTAL 92 375

* These are the ten undivided Regions.
The 1st, 2nd & 3rd quartile classes are shown in bold type.

Optional preferential voting is specified in the Aboriginal and Torres Strait Islander Commission Act 1989, and a quota-preferential counting method is set out in a Schedule. The PRSA's submission to the independent review panel in August 1994 examined the problem of the last few elected candidates sometimes succeeding with significantly less than a full quota.

As quotas are typically low (often less than 100), giving each ballot-paper a value of 10 or 100 would stop the quota being adjusted too far upwards, and also help avoid most losses by fractions. Abandoning the flawed Senate transfer value approach that has sometimes led to ballot-papers rising in value during the course of a scrutiny, would greatly lower the level of exhausted votes.

The PRSA showed that in tandem, these measures would normally result in the last elected candidates obtaining a quota of votes, or close to it. As our suggestions were not acted upon this time, we will analyse a selection of the most recent elections to illustrate what improvements would have accrued from applying our approach.

Telephone Voting for South Australian Municipal Elections - by Key Entry

The Secretary of the PRSA's South Australian Branch, Mr Deane Crabb, has drawn attention to a front page report in The Advertiser of 14th October 1996 that reveals State Government plans to introduce electronic telephone voting for municipal and union polls next year. The State's Electoral Commissioner, Mr Andy Becker, is quoted as saying that the move was part of SA Government moves to make the entire electoral system electronic. The report also quoted him as saying that SA would use the system for its next Upper House poll, due from next March.

It appears that voters would cast their ballot by a toll-free call with secure identification by means of a personal identification number. Presumably some sort of cross-check would apply to prevent people dialling with randomly invented numbers hoping they might be taken as valid voters. A recorded message will explain the procedure and voters will then key in numbers for their preferences.

A major reason for implementing the system is the large cost saving possible. At present a major election costs at least $3.1 million. The main cost of the telephone alternative would be $0.5 million, which would be for the total telephone charges involved. Mr Becker said that he would examine a voting system being developed by a consortium of Telstra and an Auckland-based company.

The SA Attorney-General said that counting would be much quicker than the several weeks it took manually. That would remove one of the major arguments of quota-preferential PR opponents - that the time taken by such a scrupulously fair system is too great to be acceptable.

A great advantage of electronic voting is the efficient retention of records of every vote, and the ability to rapidly access those records to conduct recounts, whether to deal with errors subsequently detected, to resolve disputes, or to fill casual vacancies by Tasmanian-style countback.

Japan's Hybrid Electoral System

Under its first use, on 20th October 1996, of a mixture of 300 single-member electorates (first-past-the-post) and 200 seats determined by eleven regional lists (d'Hondt highest average method), Japan experienced a record low voter turnout of 59.7%. Previously, 511 vacancies had been filled in electorates returning three, four or five members without preferential voting: uncertainty about how many votes would be needed for election used to lead to fierce competition among candidates of different factions of the Liberal Democratic Party (LDP), prompting the worst excesses of money politics. The LDP governed between 1955 and 1993, usually obtaining 6-9% more of the seats than its voter support, which ranged from 42 to 49% over the period 1967-90.

As distinct from the hybrid New Zealand system where national party list votes determine the final outcome in terms of the Lower House's composition by parties, in Japan the two votes cast by each person are treated independently. Consequently a party can obtain a huge benefit from the distortions of first-past-the-post methods, and not have this tempered very much by the proportional representation component of the system. That candidates were allowed to nominate for both individual and list seats has aroused continuing public controversy.

As is to be expected with any list system, the fragmentation and rearrangement of parties that began in the early 1990s continued. In particular, the Social Democratic Party of Japan (SDPJ), traditionally the major opposition party, but supplying a Prime Minister during part of the previous Diet, lost about half of its sitting members to the newly-formed Democratic Party (DP) shortly before the election. With much trade union support switching to this grouping, the former Socialists have now been reduced to just 15 members, compared with 70 immediately following the 1993 elections.

The second-largest grouping became the New Frontier Party (NFP), melded together in stages by Ichiro Ozawa, two-party-system proponent and one-time LDP heavyweight. Surpassed on the Left by both the DP and Japanese Communist Party (JCP), the SDPJ retains some clout because of its 35 prefectural and national seats won in the 252-member Upper House, the House of Councillors, in 1992 and 1995 (the LDP has 110). Every 3 years, half of the members of the Upper House are elected for a 6-year term by single non-transferable voting determining 76 seats in the 47 prefectures (from 1 to 4 vacancies in each), and a d'Hondt allocation applied to national lists determining the remaining 50.

The LDP's overall vote stayed around 36%, but its Lower House numbers rose from 44% to 48% because of large single-member bonuses gained in 7 of the 11 regions. Exit polls indicated that up to 30% of electors split their votes in some areas, at times because their most preferred party did not nominate a candidate in their single-member district. Results for the 5 largest parties are shown in the table below.



Single-member districts Vote (%) 38.6 28.0 10.6 12.6 2.2
Seats (%) 56.3 32.0 5.7 0.7 1.3
Seats 169 96 17 2 4
Regional List Allocation Vote (%) 32.8 28.1 16.1 13.1 6.4
Seats (%) 35.0 30.0 17.5 12.0 5.5
Seats 70 60 35 24 11
Total seats 239 156 52 26 15

Our 10% of Unelected Senators

The depth of Australians' democratic instincts and feelings might seem to be judged by the apparent popular wish, if we are to no longer be a Monarchy, to have a Head of State directly elected by the people.

Those feelings have not yet become deep enough for widespread concern over the large fraction of the Senate that is often no longer directly elected. That fraction was usually much smaller prior to the major 1977 alteration to Section 15 of the Constitution, which removed all Senate casual vacancy poll provisions. The mooted possible breaking of the Coalition's promise to have half of a 1997 Constitutional Convention directly elected appears not to have aroused an immediate vociferous backlash.

Vacating Senator (Elected by the People) Replacement Senator (Unelected by the People) State Party Date Sworn in as a Senator (Unelected by the People) Years Able to Sit in the Senate Unelected
M.Tate K.Denman Tas. ALP 30AUG1993 5.8
B.Archer E.Abetz Tas. Lib. 23FEB1994 5.4
K.Sibraa B.Neal NSW ALP 14MAR1994 5.3
G.Richardson M.Forshaw NSW ALP 11MAY1994 5.1
O.Zakharov J.Collins Vic. ALP 9MAY1995 4.1
G.Evans S.Conroy Vic. ALP 30APR1996 3.2
J.Coates K.O'Brien Tas. ALP 9SEP1996 2.8
M.Baume W.Heffernan NSW Lib. 2OCT1996 2.7


National President:  Bogey Musidlak
14 Strzelecki Crescent  NARRABUNDAH ACT 2604

National Secretary:  Deane Crabb
11 Yapinga Street PLYMPTON SA 5038

Telephone:     (08) 8297 6441, (06) 295 8137
Facsimile:     (03) 9589 1802
Internet: http://www.cs.mu.oz.au/~lee/prsa

97 Pirie Street ADELAIDE SA 5000