of the Proportional Representation Society of
On the initiative of ACT Greens MLA Kerrie Tucker, the ACT Legislative Assembly, in December 2001 referred, to its Standing Committee on Legal Affairs for early investigation and report, "the appropriateness of the size of the Legislative Assembly for the ACT and options for changing the number of members, electorates and any other related matter". Till May 2002, the Committee received several dozen submissions and oral evidence.
The Australian Democrats said an increase to three seven-member electorates would improve the quality of representation, the functioning of the Committee system, and Ministerial performance. The ACT Electoral Commission endorsed that arrangement as the most desirable option, commenting favourably that odd numbers no less than five were entrenched, and advocating that the same number of vacancies occur in an odd number of electorates. Academics from both ACT universities also favoured three seven-member electorates.
The Liberal Party's ACT Division supported an increase, in principle. It rejected five 5-member electorates as giving insufficient proportionality between votes and seats, and resulting in boundaries that divided too many natural communities of interest. Three seven-member electorates were its preferred option, but alarmingly, despite the party's role in having had odd numbers per electorate entrenched, the submission indicated that three six-member electorates and one five-member electorate, based largely on existing town centres, constituted a 'model worthy of examination' that 'may be of interest to other political parties, as a compromise model to consider'.
The ACT's Young Liberal Movement rejected an Assembly of 21 members as having electorates that were too large. It suggested an increase to 23 MLAs in at least four seats.
In May submissions, the ACT Branch of the Labor Party and the Stanhope Government both advocated five 5-member electorates, to provide an adequate buffer of representatives for some time into the future and to enable more local representation. The ACT Branch indicated that 'electorates of six or four may also successfully achieve these democratic goals' but recognized these were outside the entrenched provisions. The ALP Government pointed to changes in the ratio of MLAs to electors since the introduction of self-government, and the Assembly's workload. Three years earlier, the Chief Minister was part of a Committee majority concluding, on balance, to retain a 17-member house for the 2001 elections (see QN1999B).
On 29th June 2002, a majority of the Standing Committee on Legal Affairs reported in favour of an increase to three 7-member electorates. However Labor MLA John Hargreaves issued a minority report supporting 23 MLAs, most probably in the arrangement raised in the Liberal submission rather than in electorates returning seven or nine MLAs. He said the governing party's back bench needed to outnumber the executive for effective democracy and that was not possible with even 21 MLAs.
As the threat to the entrenched principle of odd numbers per electorate appeared to increase in succeeding months, the PRSA's ACT Branch was impelled to devise a strategy to minimize the risk of any backward step being taken. The federal Territories Minister might refuse to act on a request for a huge increase in MLA numbers, but if more than two-thirds of the Assembly showed a preparedness to accept electorates returning an even number of MLAs the Hare-Clark system would be significantly weakened.
Below is an edited version of the May 2002 written submission, "Enhancing Voter Empowerment under the Hare-Clark System", that the PRSA's ACT Branch made to the Standing Committee on Legal Affairs:
The Branch shows how much better representation is achieved in the 7-member electorate than in those returning 5 MLAs, both in terms of more closely reflecting voters' expressed wishes and in electing female MLAs. It stresses the need to present any proposed increase in the size of the Assembly as part of a package to improve voter participation in legislative and accountability processes.
Efforts to achieve 3 electorates of 7 members each could most readily be supported by the Branch. A 19-member Assembly with one 5-member and two 7-member electorates might be acceptable. It is not calamitous if the Assembly size stays at 17. Voters have accepted it so far.
Setting the scene for discussion
After three elections, each fairly translating voters' wishes into levels of Assembly representation, the Hare-Clark system is an accepted part of the political landscape. Some earlier organized antipathy has officially been set aside.
In part this rising acceptance can be attributed to the populace having no reason to be alarmed by developments since the first election in 1995. The predictions of greater campaign and ongoing contact by candidates, electorate-boundary stability and manageable ballot-papers have been borne out. Minority governments that have formed without surprise in the wake of voting and seat patterns have been able to implement programs announced prior to elections, sometimes after Assembly discussion and amendment.
Improvements have been made progressively following thorough Elections ACT reviews of the operation of electoral legislation, and vigorous public discussion of other suggestions. Last-moment thrusting of materials at voters entering polling places came to an end before the 1998 elections, and has generally been welcomed. The splendid investigatory work of Dr Ken Brewer and the ACT Electoral Commissioner, Mr Phillip Green, led to an inspired extension of Robson Rotation at the last election that removed lingering party-linear effects as a factor in determining the order of candidate exclusion or election.
The 2001 election was important in debunking several myths. One was that majorities are impossible to achieve in practice in individual electorates. Another was that the supposed mammoth advantages of incumbency will always make female representation difficult to achieve.
First, the Australian Labor Party won a majority of seats in Brindabella after starting with nearly 44% of first preferences. After polling just over 1% less in Ginninderra, it was not far from obtaining a majority there also, and might have done so if as few as 110 voters had changed their minds about the relative merits of the final continuing candidates from the Liberal and Labor parties.
There were six new faces in the Assembly after the election. Only one of them replaced an MLA retiring voluntarily. This maintained a pattern of around one-third of an outgoing Assembly not coming back. Six women were elected, emphasizing the aberrant nature of the election of just two women in 1998. The defeat then of two women by just 23 and 45 votes respectively greatly affected the Assembly's gender balance. A third, who fell some 600 votes short, narrowly became an MLA in 2001.
Consistency in the past 10 years - stability in practice
The PRSA's ACT Branch was one of the four original members of the Hare-Clark Campaign Committee formed in 1991. We campaigned actively for the current three-electorate 5-5-7 model for the Legislative Assembly once supporters of proportional representation agreed that this was the most workable local-member option available in the light of the Federal Parliament's decision to legislate for a 17-member Assembly under self-government. Hare-Clark supporters were therefore astounded to see a last-ditch single-member 1992 election-day poster screaming 'Hare-Clark means more politicians!', but not surprised that it had little apparent effect. After that the ACT Branch has promoted the voter-empowering features of our electoral system wherever possible, and been wary of proposals for change with obvious flaws or that carried a risk of antagonizing numerous voters.
We are proud to have been involved in:
· promoting the need to entrench the key
principles of the Hare-Clark system (including
the desirability of ensuring that each
electorate return an odd number of members no
fewer than five) following the attempted
white-anting of Robson Rotation
in late 1993;
An odd number of vacancies in each electorate is important, as majorities of votes translate into majorities of seats in that case. Seeking a majority of votes becomes the ultimate pursuit for parties with relatively strong support, whereas with even numbers of vacancies, getting half the seats can be guaranteed with a bit under half the votes, but a majority is usually beyond practical reach.
Review of the Governance of the ACT
We opposed the 1998 ideas of the joint
Commonwealth-ACT Working Party on the Review of
the Governance of the
· the Assembly automatically grow in size with
population increases; and
Any process to automatically increase the Assembly's size at regular intervals would immediately draw public hostility. Apart from introducing near-certain chronic boundary instability, such ongoing growth in Assembly numbers would probably unleash in the minds of various political contestants the thought of attempting to place the smallest and largest electorates strategically to maximize potential benefits to themselves. Such preoccupations elsewhere have been absent so far in the ACT. Anything that could cloud the current clear focus on the need for parties to develop attractive policies and endorse credible candidates to promote them is to be regretted.
The Pettit Review thought Independents would continue to be elected, so the prospect was slight of larger parties with disparate levels of support emerging with the same number of representatives when the number of vacancies was even.
A perfunctory look at half-Senate elections for 6 vacancies since 1990 (the size of the Parliament was increased in 1984, and there was a double dissolution in 1987) shows avoidable difficulties when even numbers of vacancies are filled. With a quota just over one-seventh of votes, 42.9% is enough for 3 seats, but 4 are usually beyond reach.
The key questions that need investigation
The key questions are:
· Is there a representation problem at the
The ACT Branch has noted the stability of electoral boundaries since 1995 and the evolution and enunciation of Redistribution Committee principles that boundaries should remain intact except when adjustments are unavoidable on account of statutory criteria about acceptable tolerances regarding the ratio of electors to members in any electorate. Voters have widely welcomed this feature of electoral arrangements.
Under the entrenched principle that electorates return an odd number of members no fewer than five, once an element of local representation is specified, 17 MLAs can only be elected from one 7-member and two 5-member electorates. The respective quotas are the first integers greater than one-eighth and one-sixth of the formal votes.
It is worth considering the differences that can be expected when 12.5% or 16.7% respectively guarantees election to the Assembly, before examining ACT outcomes at the past three elections. In both cases, a majority of the votes translates into a majority of the seats.
Representation in 7-member electorates should be better, as the finer granularity leaves greater uncertainty about outcomes, and thus minimizes the chances of parties or candidates just treading water, thinking there is nothing to lose or gain. Swings of just 2-3% can often change outcomes, keeping all involved much more on their toes. Independents and parties realistically expecting at most one candidate to be elected know the importance of getting as many first preferences as possible to stave off exclusion in the middle of the scrutiny. The higher quota for five-member electorates makes election significantly harder for independents and smaller parties or groupings.
As 25% of the vote forms two quotas in a seven-member electorate and 37.5% three quotas, larger parties cannot take their level of representation for granted, and will work hard to ensure they achieve a third seat at least, knowing their prospects are slight unless they start with at least 31-32% of first preferences.
In 5-member electorates, 33.3% secures two places and 28-30% or so often suffices. Larger parties aim at two seats, but might be modest and focus on other electorates if 2 are comfortable and a third is out of reach.
Although Robson Rotation prevents larger parties from minimizing their numbers elected from a given level of support, it will generally be harder to achieve a majority when seven vacancies exist. A party with strong support in a 5-member electorate knows that at worst every vote above 33.3% contributes to electing a third candidate, but in a 7-member electorate 37.5% may be used to elect three candidates before a fourth place is really expected.
It is important to note that at each election there has been significant difference in support for the two largest parties. Had single-member electorates applied, on each occasion there would have been at best just a token Opposition presence in the Assembly. The most noticeable feature has been the large seat bonus achieved by the second party in 5-member electorates. This is because support remained just strong enough to achieve a second quota through the flow of preferences and thereby translate even slightly under 30% of first preferences into 40% of seats.
Instead of a single-seat advantage being held each time by the party with strongest support, had the votes achieved always been in 7-member electorates, the difference between the parties in the next (21-member) Assembly would instead have been two or three seats overall. This is because support levels below 30% make it extremely unlikely that three quotas could be achieved, and no certainty attaches to levels in the low 30s.
Overall, 7-member electorates have clearly provided better matches of votes and representation than the 5-member ones, and have facilitated a higher percentage of women being elected. The ACT Branch could thus not oppose more of them, as long as that was done in a way that did not undermine current broad respect for our Hare-Clark electoral system.
Are there numbers other than five or seven for
the size of an electorate that would also be
satisfactory? The experience of the Senate in
its five elections between 1990 and 2001 is a
perfect example of why six-member electorates
would be unsatisfactory. The typical outcome has
been a 3-2-1 division of places, a minor party
or independent managing to build (usually,
rather than getting it on first preferences
alone) a quota earlier than the weaker of the
large parties could assemble a third quota.
However a 3-3 split between Labor and the
Coalition occurred in both NSW (despite an 8%
advantage in votes for Labor) and Victoria in
1993, and both the Australian Democrats and One
Nation were successful as part of a 2-2-1-1
Having the real target for larger parties at 42.9% for three seats and leaving then no hope of obtaining a majority in any but extraordinary circumstances has produced major frustrations. Regular menaces muttered by governments in the direction of those holding the balance of power in the Senate have been interspersed with acknowledgements that there is a systemic problem arising from the filling of an even number of vacancies at half-Senate elections. A similar gradually-diminishing problem that arises for higher even numbers is always avoided if a neighbouring odd number is chosen instead. There are no grounds for electorates with an even number of vacancies above five.
Should entrenched principles be overturned to
create electorates of only two, three or four
members, with ineffective votes up to 33.3%, 25%
or 20% respectively - far greater than for a
7-member electorate? In a 4-member electorate,
20% suffices for one seat and 40% secures two.
Securing three would be impracticable. There
might be attempts to insert electorates of 3, 4
or 5 MPs, as in
A party confident of getting 35% over a particular area might seek to put in a 4-member electorate in pursuit of half the seats there. If it did better elsewhere, there would be some prospect of picking up a majority in seats with an odd number of vacancies and turning overall voter support well short of 50% into a possible majority of seats.
Three-member electorates, where the quota is just over 25% of formal votes, offer the strongest party some chance of achieving two seats once it gets a bit more than 35% of first preferences, especially if none of the smaller parties or independents manage to get to 10%. Such gaming opportunities are not currently available, nor should the door be opened for them when 5- and 7-member electorates give a more open political environment, and outcomes the community sees as fair.
The ACT Branch believes that any serious suggestion that electorates return fewer than five members would constitute a fundamental assault on the integrity of the Hare-Clark system, signalling an attempt to achieve through artifice what proponents were simply not confident of obtaining more naturally through the attractiveness of their policies and endorsed candidates.
By contrast, ACT experience under the unlamented modified d'Hondt system and, later, the need for tablecloth ballot-papers for the 1999 NSW Legislative Council poll (for 21 vacancies Statewide) and the 2001 SA Legislative Council poll (for 11 vacancies Statewide) show that while ineffective vote percentages can be reduced to under 10%, the size of the ballot-paper can become so unwieldy as to draw widespread public derision. If the area of an electorate gets too large, candidates are less likely to undertake extensive door knocking and small-group contact, which puts them into close touch with electors.
The 2002 polls to elect the President of France starkly revealed to the whole world the unsatisfactory nature of this important public election not providing for fully preferential and transferable voting. The President is, unlike the U.S. President, elected by direct and popular vote, but the rules do not give electors a preferential vote that is transferable in accordance with the voters' instructions, as marked by the voters on their ballot-papers.
At the first stage the full number of candidates is put to a national vote, but the only way a valid vote can be cast is for the voter to vote for one candidate only, with no indications of preferences being sought or permitted. If one candidate gains an absolute majority of votes that person is then declared elected. If that is not the case, the two candidates with the highest number of votes then become the only two candidates in a "run-off" poll some days later. Such elections could be held in one poll by letting voters mark their preferences among all candidates, which is the standard in Australian single-vacancy polls.
Instead, the election wastefully has two stages. The arbitrary rationale used to justify this awkward and restrictive procedure is that the winner of the "run-off" poll needs an absolute majority of votes to claim an incontrovertible mandate, lacking at the first poll. Of course it could have been achieved there with preferential voting. Non-preferential voting can guarantee that only if there are fewer than three candidates.
the crude procedure of instantly "guillotining"
all but the top two candidates is how the
"run-off" is staged. That conveniently avoids
recognizing that the exclusion of all but the
two highest-polling candidates after the first
poll could still leave many voters, even an
absolute majority of voters, having voted for
somebody other than the two "front runners". In
the first round of the French presidential
polls, in April 2002, only 36.74% of French
voters voted for EITHER of the two "front
runners". That left 63.26% of the nation's
voters voting solely for somebody other than the
two final contestants, as the following table of
the results shows. The marking 'R' or 'L' denotes
candidates known to have a general leaning to
the Right or Left politically.
Normally these elections mitigate public dissatisfaction with the system as the two candidates in the "run-off" poll usually have one right-wing candidate and one left-wing candidate, giving those two major blocs a crudely based chance for a final test of strength, with other blocs forced to choose between those two only. At the "run-off" poll on 5th May 2002, the major blocs, who were represented at the initial poll by the outgoing President, Jacques Chirac, and the outgoing Prime Minister, Lionel Jospin, delivered nearly all their votes to Chirac, and away from Le Pen, whose stance was further to the right than either of them.
The "run-off" poll gave 82.21% of the vote to
Chirac, with only 17.79% for Le Pen, who thus
gained only 1.13 percentage points from the
voters that did not vote for him in the first
round. If Jospin, whose first round vote was
less than one percentage point below Le Pen's,
had been in the "run-off" instead of Le Pen, he
would have gained much more of the vote that
went to Chirac. Jospin would probably not have
defeated Chirac, but a preferential ballot would
have certainly provided a more realistic picture
of the relative strengths of
Proportional Representation Society of
National President: Bogey Musidlak 14 Strzelecki Cr. NARRABUNDAH 2604
National Secretary: Deane Crabb 11 Yapinga St. SOUTH PLYMPTON 5038
Tel: (08) 8297 6441, (02) 6295 8137 firstname.lastname@example.org
Printed by Prestige Copying &
Printing, 97 Pirie Street ADELAIDE SA 5000