Current Issues Brief 10 1998-99

Scott Bennett Politics and Public Administration Group
20 April 1999

Contents

Major Issues

Introduction

The Senate-perceived problems

What changes might be made?

House of Representatives-perceived problems

Should changes be made to the electoral system?

Can Australia establish a perfect electoral system?

Endnotes

Appendix: Non-major party House of Representatives victories 1946-98

Table 1: First-past-the-post in 1998

Table 2: The 'winner's bonus'-House of Representatives elections

Major Issues

National electoral systems are often the subject of debate and controversy-as has been the case in New Zealand and the UK throughout the 1990s. In Australia, possible changes to the electoral system in use for Commonwealth elections have also come under discussion. Senators Helen Coonan (Lib, NSW) and Meg Lees (AD, SA) have recently made recommendations for reform, prompting a debate which has discussed wide-ranging issues such as the role of the Senate and the principles of representation, as well as the narrower issue of what is the most appropriate voting method for either house of the national parliament. In addition to Coonan's and Lees' considered contributions, there have been briefer comments by One Nation leader, Pauline Hanson, and the journalist, D. D. McNicoll. This Current Issues Brief outlines these views, as well as discussing the possible problems involved in making significant changes to the Commonwealth Electoral Act 1918.

Introduction

Electoral systems provide the practical means for translating votes into parliamentary seats. In a political system like the Australian, this has two major outcomes. The electoral system has a direct impact upon who shall govern-the party (or parties) that win a majority of lower house seats is commissioned to form the government. The second consequence of the operation of the Australian electoral system is that it also allocates seats in an upper house, a body that is usually seen as having some functions that are different from that which is sometimes called the 'house of government'.

How well does the Australian system meet Australia's needs? There is no doubt that in regard to the central function of providing the nation with stable government, the system has worked satisfactorily, but recent public criticism suggests that some people, at least, would like to see various changes made.

Senate power

As a backdrop for the discussion that follows, it is relevant to talk first about Senate power.

In his work on comparative government, Alan Ball of the University of Portsmouth has observed that 'most second chambers are politically unimportant'.(1) Of all upper houses in existence, the Australian Senate is in fact one of the most powerful, a point that must be appreciated in any discussion of the voting method used for our upper house. It has been noted that its:

It is because the Senate is so powerful and, hence, so much of a prize for the parties, that there is any serious discussion of its voting system. If it were to be as relatively insignificant as most other upper houses, such as the French or the Japanese, there would be little to contest. The power to block Supply legislation that was a matter of controversy in 1974-5, is clearly central to this question of Senate power. Despite the Constitutional Commission's recommendation in 1988 that this be removed by constitutional amendment,(3) there is unlikely to be any agreement between the parties over the question, despite this being a change that would probably suit both the Coalition's and Labor's long-term interests. Laurie Oakes has described Howard Government frustration with the Senate as '[reinforcing] the argument for Senate reform by demonstrating how the Upper House has frustrated both major parties.(4)

An even more drastic way of tackling this question would be to amend the Constitution to strip back all Senate power to a position where the upper house had little more than a delaying power, much like the House of Lords since the Parliament Acts of 1911 and 1949. In Britain this matter is currently being reconsidered by the Royal Commission set up to investigate proposals to alter the place of the House of Lords after the removal of the hereditary peers. The British have realised that there is a basic dilemma involving the relationship between upper and lower houses in any legislature. If there is too much pruning of upper house power it can leave that body as largely redundant. This is the old dilemma posed about two hundred years ago by the French politician, the Abbe Sieyes, when he asserted that 'if a second chamber dissents from the first, it is mischievous; if it agrees with it, it is superfluous'.(5) Australians are well aware that this is a conundrum not easily solved.

The Senate-perceived problems

An unrepresentative upper house? (1)

It has been argued that the Senate is unrepresentative in at least two ways. The first is part and parcel of the original federal compact, namely the equal number of Senate seats given each original state, described in a recent paper by Liberal Senator Helen Coonan as making 'a nonsense of the notion of one vote, one value'.(6) Coonan has also stated that nearly one hundred years after the creation of the Commonwealth of Australia:

It is a measure of the frustration of the major parties that a Liberal Senator should express a view that would once have only been heard from a Labor politician.

An unrepresentative upper house? (2)

The other aspect of this matter of Senate representativeness relates to the fact that the balance of power is so often held, and is exploited, by minority party or independent Senators. To win a Senate seat at a half-Senate election, a 'quota' of 14.3 per cent of first and later preferences and surplus votes needs to be gained. It is this low quota (when compared with the 50 per cent + one vote for House of Representatives seats), that makes it relatively easy for a party or an independent with a moderate amount of electorate support to win a Senate seat.

Coonan has asked whether it is in the best interests of the broader community and good government, if 'for practical purposes effective power [in the Senate] is concentrated in the hands of the few, especially disproportionate to their electoral support'.(8) This has been a constant complaint heard about the Senate for many years, regardless of whether Labor or the Coalition have been in power.

What changes might be made?

Changing Senate numbers?

The simple answer is for the major parties to increase their popularity with voters. For both Labor and the Coalition a vote of 42.9 per cent would almost certainly win three seats per State, thus excluding minor party and independent victories. This is a target that should be well within their capabilities, yet in the most recent Commonwealth election in 1998 both sides were well below the 40 per cent mark nationally, and only managed to share the six seats in Victoria. This latter result has become quite rare, due to the marked decline in major party Senate votes in each decade from the 1950s to the present.(9)

There have been some suggestions that the reduction of the size of the Senate to ten Senators per State would, by increasing the size of the quota to 16.7 per cent, usually exclude even the strongest of minor parties. An illustration of the effect of increasing a quota can be seen in the 1998 reduction in the size of the Tasmanian House of Assembly. The quota for the five seats was increased from 12.5 per cent to 16.7 per cent, and despite the Tasmanian Greens' 1998 vote being similar to that in 1995, their representation fell from four members to one.(10) While the major party vote remains stagnant, though, half-Senate elections for five Senators would probably see the major parties still well below the 50 per cent needed to win three of a State's five seats. The chances of a prominent minor party, such as the Democrats, winning the final seat in each State would thus still be high.(11)

Coonan's criticism about the equality of Senate numbers for the Original States is unlikely to be supported. There would be little chance of any alteration of such a fundamental part of the Constitution, even were the major parties to support it in the national Parliament. There would be very strong opposition-that would probably be bipartisan-in the smaller States. In any case, it is likely that many federal Liberals and Nationals would also oppose such a change. Overall, then, there would be sufficient opposition to make the double majority of s.128 unlikely to be achieved.

Altering Parliamentary terms

Another way of changing the Senate would be to revisit the 1988 effort to amend the Constitution to alter Parliamentary terms. If all Senate members were elected at the one time as occurs in Western Australian Legislative Council elections, this would increase the chance of one party gaining control of the body after an election. There would be no Senators who had been elected 2-3 years before a general election, as is the case now. A restoration of major electoral strength would make this scenario quite possible, particularly if each State had an uneven number of Senators (e.g. 11 or 13).

Attacking the proportional representation 'problem'

A key target for critics is the proportional representation (PR) voting method itself, which is blamed for the continued failure of the major parties to gain control of the upper house.(12) Suggested reforms seem to follow three main approaches: replacement of proportional representation by preferential voting, alteration of electorates, or imposition of electoral thresholds.

Replacement by preferential voting?

In 1994, press reports stated that Prime Minister Keating was looking into changing Senate voting arrangements so that there would be separate, single-member electorates, with preferential voting the method employed.(13) No detail was given but it seemed to involve a change similar to that later suggested by former Liberal Party Federal Director, Andrew Robb. Robb proposed a system similar to the Victorian Legislative Council arrangements, whereby each State would be divided into six regions, each represented by two Senators, one of whom would retire every three years.(14) Malcolm Mackerras of the Australian Defence Force Academy has claimed that s.7 of the Constitution only allows the people of a State to vote for Senators 'as one electorate', but he ignores the fact that in using the words 'until the Parliament otherwise provides', the section also provides for Parliament to make other arrangements at any time in the future.(15) The importance of the Parliament having such a power was certainly the view of some of the founders, who argued the case for making the section flexible enough so that should the Commonwealth Parliament 'find that this method is unworkable, they will be able to alter it'.(16) This would suggest that the introduction of single-member Senate electorates would be quite feasible.

Because of the high quota of votes needed to win a seat (50 per cent + one vote), this change to our electoral arrangements would be very likely to work to the detriment of minor party and independent candidates, particularly as the electorates in each mainland State would be larger than they are in House of Representatives elections. The larger the size of the electorate, the more reliant would party candidates be on their party organisation, a factor that would be of particular benefit to Coalition and Labor Party candidates-New South Wales seats would contain nearly 680 000 voters, for example, compared with the average of 81 500 for the State's 50 House of Representatives seats in 1998. According to Coonan, the Coalition would have gained a Senate majority of between four and six seats in 1996 had this system been in place.(17)

Alteration of electorates?

Senator Coonan has suggested another option, namely the establishment of three State electorates, each being represented by four Senators, two of whom would retire every three years. Such a two-seat contest would mirror the arrangements currently in place for the Territory Senate elections.(18) The quota of votes to win one seat would be 33.3 per cent, but double that to win two seats, so that such an arrangement would virtually guarantee that the seats would be shared by the major players, to the exclusion of all other parties and candidates. Despite the near-defeat of ACT Liberal Senator Margaret Reid in 1998, the twenty ACT and Northern Territory Senate contests since 1975 have seen half the seats won by the Coalition and half by the ALP. The problem, though, is that if an even number of Senators is to be elected in this way at each election, it is highly unlikely that either major side could secure a Senate majority.

Electoral thresholds?

A third method for tackling proportional representation involves the imposition of an electoral threshold, a device commonly used by major parties in other political systems to protect their position from attack from minor parties:

Typically, the electoral arrangements that include an electoral threshold will specify a minimum percentage of the vote that a party or individual must achieve to be able to participate in the allocation of seats. Coonan's paper canvasses four threshold levels: five per cent (the German figure), 7.14 per cent (50 per cent of a Senate quota), ten per cent (no rationale is given for this figure) and 11.43 per cent (80 per cent of a Senate quota).(20) Her paper shows that of the sixteen minor party and independent Senators elected in 1993, 1996 and 1998, all would have survived a five per cent threshold. Three would have been defeated under a 7.14 per cent threshold, ten under a ten per cent figure, and the 11.43 per cent threshold would have seen only four gaining election (Stott Despoja and Kernot in 1996, Lees and Hill in 1998). With the remaining dozen seats likely to be pretty evenly shared due to the evenness of party strengths, even the use of the highest threshold on the Coonan list would be quite likely to see the balance of power in the Senate held by minority Senators.(21)

The simple answer might be that a higher threshold than 11.43 per cent should be considered, but this would have to be looked at very carefully, for a figure much higher than the 11.43 per cent might have blocked the Liberal Party's David McGibbon in 1987 (party vote 14.2 per cent), and the Nationals' Ron Boswell (1990, 12.8 per cent) and Bill O'Chee (1993, 13.6 per cent). Even the ten per cent threshold would have removed the Queensland Nationals from the count in 1998. In all cases the preferences of these parties would have flowed on to the Coalition partner, but as Coonan herself notes about thresholds:

House of Representatives-perceived problems

Preferential voting has long seemed a settled part of the Australian electoral system. Since the 1998 election, however, there have been three public criticisms made of this voting method.

The height of the hurdle

An extremely difficult hurdle for minor parties in Australia is the requirement for a candidate to gain 50 per cent + one vote to win a seat, something that non-major party candidates find difficult to achieve. Only 14 such victories have been achieved in House of Representatives seats since 1946.(23) The hurdle is undoubtedly high, but it is also relevant to mention that for most of the preferential voting elections since 1919 the sound electoral health of the major parties has seen them able to get the lion's share of the vote. If the 1998 House of Representatives national votes for the ALP, Liberal Party and National Party are tallied, we find that these parties secured nearly 80 per cent. Such a figure does not leave many votes for non-major party candidates. The major parties' reply to criticism of the height of the hurdle is simply to reply that all candidates have an equal chance to achieve it-the problem, in other words, is not with the voting method but with the minor parties.

Despite this, the Australian Democrat leader, Senator Meg Lees has recently criticised this aspect of Australia's electoral system. According to her it denies representation to one in five voters (based on 1998 figures), because it effectively 'minimises voters' effective choices to just two similar parties'.(24) Her solution is an adaptation of the German and New Zealand electoral systems that has been investigated by the Jenkins Commission in its discussion of possible electoral changes in the UK.(25) Lees has proposed the election of 80-85 per cent of MHRs by preferential voting, where voters choose between candidates, with the remaining 15-20 per cent of MHRs being elected by a second, State-wide, vote where electors would choose between parties. This latter, or 'top-up' vote, would be allocated in each State to the parties that were the most under-represented according to their popular vote. Lees' claim is that if a 15 per cent top-up formula had been in operation in New South Wales in 1998, 43 of the 50 seats in that State would have been elected in single-member electorates using preferential voting, and the remaining seven would have been top-up seats based on voters' second votes.

Lees has described this model as fair compromise between a desire to keep single-member electorates, and 'the need to ensure that the growing number of Australians opting for third parties is afforded representation in the Parliament.'(26)

A return to first-past-the-post?

In the wake of her defeat in Blair (Qld), Pauline Hanson described the voting method as unfair because, as she had led comfortably on first preferences (36.0 per cent), she believed she should have won the seat. The journalist, D. D. McNicoll, echoed this criticism and called for the installation of first-past-the-post for House of Representatives elections. He pointed out that first-past-the-post is easier to understand, something that is surely of great importance in a democracy, and is much quicker to count.(27) There is no doubt about McNicoll's claims, for the first-past-the-post method is simplicity itself, requiring just the single vote against one name, and in Britain, for example, the results in a great many seats are known by the end of polling day.

The problem with first-past-the-post, though, is that it is far more prone than preferential voting to throw up anomalous results. In Britain, for instance, the Labour Party won 49 per cent of the vote in 1951 yet failed to win office; in 1997, by contrast, Labour won 44 per cent of the vote and gained a huge majority of 179 seats. Labour twice won the highest share of the vote in the 1950s yet lost both elections.(28) Preferential voting in Australia is also unlikely to produce the startling Canadian general election result of 1993, when the Conservative Government's numbers in the House of Commons fell from 155 to just two members. Anomalous results like these have seen New Zealand replace first-past-the-post with multi-member proportional,(29) and Britain set up the Jenkins Commission (see above), to investigate the possibility of replacing the British system.

If a particular Australian electorate cannot produce a candidate with an absolute majority, then the allocation of preferences is designed to find the most preferred candidate. One can argue that in Blair the Liberal candidate, Cameron Thompson, was preferred by more voters than was Hanson who had, after all, been well below a 50 per cent vote on first preferences. Certainly the system treated other candidates more harshly than her: five of the seven candidates who led on first preferences but lost after the distribution of preferences, received a higher vote than did she. Warwick Smith (LP, Bass) actually received 45.7 per cent while Eoin Cameron (LP, Stirling) gained 41.7 per cent.(30) The major parties have said little about such results in the past, presumably because they see these as part and parcel of preferential voting, and a recognition that every voting method produces strange results from time to time. If first-past-the-post had been used in 1998, the result would have been as follows:

Table 1: First-past-the-post in 1998

The 'wrong' winner

Lees has also criticised the voting method for allowing the 'wrong' party to win office. She has noted that in five post-1945 elections (1954, 1961, 1969, 1990, 1998) the party with the lower two-party preferred vote has been able to form a government on the basis of having won a majority of House of Representatives seats.(31) The problem Lees is pointing to is a consequence of using what is generally called a 'majority' voting method in single-member electorates.(32) In such an arrangement it is impossible to guarantee that the party with a majority of votes will actually win a majority of seats.

In fact, it is theoretically possible for a party to gain 49.9 per cent of the first preference vote in every House of Representatives electorate yet fail to return any members to the House. Although most majority systems do manage to allocate parliamentary seats to parties in rough proportion to votes won by each, nothing actually guarantees this.(33) This helps explain why it is possible for a party to receive a majority of votes yet fail to win government, as occurred in 1954 when the ALP won 50 per cent of first preferences, yet won only 57 of 121 electorates and remained in Opposition. It also explains how the Coalition retained office in 1998 with just 39.5 per cent of first preferences and only 49.0 per cent of the two-party preferred vote. To an important extent, such results depend on which side of electorate boundaries that votes fall, rather than the total number of votes a party receives.(34)

The lack of proportionality

Lees' other criticism of preferential voting has been the tendency for the voting method to produce 'disproportionate representation of the Government in terms of seats won'. This was especially obvious in the landslide elections of 1966, 1975, 1977 and 1996 when the winning party gained a much higher proportion of seats than votes.(35) In 1996, for instance, the Coalition won 63.5 per cent of House seats on a first preference vote of just 47.2 per cent.

It might be supposed that an electoral system should allocate half of the parliamentary seats to the party that wins half of the vote-or as political scientists would put it, there should be a high degree of 'proportionality' in an electoral system. A number of studies have shown, however, that electoral systems are very crude instruments for ensuring proportionality:

In his famous study of the impact of electoral systems, Douglas Rae observed that a high degree of proportionality is hardest to achieve in single-member electorates, of the type in use for House of Representatives elections.(37) This phenomenon is exaggerated by what has been noted about the way in which all voting methods-and not only majority systems-

The largest seat-winner tends to benefit most from this phenomenon, often gaining what has been called the 'winner's bonus'. This is usually seen most clearly in single-member electorates where majority methods are in use:(39)

Table 2: The 'winner's bonus'-House of Representatives elections

Source: Scott Bennett, Winning and Losing. Australian National Elections, Melbourne University Press, Melbourne, 1996, p. 60; Scott Bennett, Andrew Kopras and Gerard Newman, 'Federal Elections 1998', Research Paper No. 9 1998-99, Department of the Parliamentary Library, 1999, p. 14.

There is actually no solution to this which will satisfy all people. Each individual's judgement of what is 'best' depends upon the results one wants to achieve. If the aim is a locality-based system that is highly likely to produce a lower house majority and, therefore, stable government, there is no doubt that since 1919 preferential voting has produced clear results in all elections except that of 1940. If one is concerned to maintain some relationship between voter and local MHR, single-member electorates should be retained. On the other hand, if one seeks a voting method that guarantees some representation of the ideas of significant pockets of the people-such as supporters of environmental issues or the followers of Pauline Hanson-then the aim should be to establish a proportional representation system.

Should changes be made to the electoral system?

The answer to this question is a matter of political judgment. Typically, such changes are made because parties and politicians judge that they will be of direct benefit to them, and because of this, sometimes mistakes are made. Some of the more obvious pitfalls are discussed briefly here.

Lack of understanding of the method being altered

Altering voting methods can be an uncertain business, often because of a lack of understanding by those making the changes. We have already seen this three times in relation to Senate elections in this country.

Senate elections-first-past-the-post (1902)(40)

The first method chosen for Senate elections was first-past-the-post, a method that only works satisfactorily in multi-member electorates when candidates stand as independents. With parties contesting an election, many multi-member results are likely to be highly distorted, as occurred in some early Senate results. After three uncontroversial elections in the first decade, in the 1910 election all 18 Senate seats were won by Labor, and in 1917 the eighteen were all won by the Nationalists. In the double dissolution election of 1914 the ALP won 31 of the 36 seats.

Senate elections-preferential voting (1919)(41)

The change to preferential voting for the 1919 Senate election indicated that people had not realised that this method, also, was not suited to multi-member electorates. In that first election the Nationalists (46.4 per cent of the vote) won 18 of the 19 seats, while Labor's 42.8 per cent of the vote returned a solitary seat. More often than not Senate results between 1919-1946 gave bizarre results: Labor was unable to win any of the 18 seats in 1934, for instance, while the opposition parties failed to win a single seat in 1943.

Senate elections-proportional representation(42)

The change to the Single Transferable Vote variant of proportional representation for the 1949 election also showed a lack of familiarity with what was being proposed. As with the earlier use of first-past-the-post, the consequences were not immediately apparent, but from the time the Democratic Labor Party gained the balance of power following the 1964 Senate election, the game had changed. By the time the Australian Democrats, the Greens and Brian Harradine had entered the picture, the major parties had come to resent the impact of proportional representation on party representation in the Senate.

The increase in Senate numbers

Another example of a lack of understanding occurred when the Representation Act 1983 increased the number of each State's Senators from ten to twelve, a change that added to the difficulty for parties attempting to gain control of the upper house. Before this change, the fact that five were to be elected in half-Senate elections meant that it was always possible that one party could gain a majority of Senate seats decided in a particular year-in 1955, for instance the Coalition won 17 of 30 seats across the nation. At the extreme, if the one party were successful in all six States in two consecutive elections, then that party would have a 36-24 majority. Of course, a majority might not be achieved due to differing results in each State, plus the fact that only half of the Senate was elected on the one day, but a majority was still quite possible. The key to such a result, clearly, was having an uneven number of seats decided in each State.

The increase in Senate seats to 12 left the position where Senate results (assuming the absence of minor parties) was highly likely to be 3-3, as occurred in New South Wales in 1993 and Victoria in 1993 and 1998.(43) For a 4-2 result to be achieved in any State Labor or the Coalition would need to secure an almost unprecedented 57.1 per cent of the vote-since 1949 this figure has been achieved only once, in 1975 when the Coalition's Queensland Senate vote was 57.3 per cent. Clearly, even if minor parties could be pushed from the scene (barring a major collapse on one side of politics), no party could expect to win more than half of the Senate seats over the period of two half-Senate elections. A Senate in which the Coalition and the Labor Party each had one-half of the members, would probably be more rigid than when the government of the day is able, by negotiation, to construct majorities on particular pieces of legislation. Indeed, it might be argued that a government is actually better off in that position than if the Senate seats were neatly halved between the Coalition and Labor. At least it has the option of negotiating with a party or individual Senator(s) who control the balance of power. As Gerard Henderson has noted,

It can be argued that introducing changes to the current Senate voting method should be handled with care lest unforeseen consequences produce something worse than the arrangements that have now been in place for fifty years.

Voter unpredictability

One uncertainty associated with voting method alteration is that voters are unpredictable-and prone to change their minds. Politicians can make alterations, but they cannot be certain just how voters will behave when it comes to the act of casting a ballot. One unanticipated consequence of the 1948 change to proportional representation for Senate elections, has been the propensity of many voters to vote for one party in lower house elections and for another for the upper house. Typically, the major parties win fewer votes in Senate than in House elections, while the reverse applies for the Australian Democrats. Since 1949, the major party Senate vote has on average been 4.4 per cent lower than for House elections, with the gap climbing to 7.8 per cent in the four elections of the eighties. The nineties figure has been 3.9 per cent.(45)

Reasons for this have been discussed in the literature. One theory is that voters do not want to 'waste' votes. We have seen how the method of voting used for House elections makes it very difficult for non-major party candidates to win House seats. Many people who vote Democrat in the Senate may follow such a vote with a vote for a major party in the concurrent House of Representatives election. It has been theorised that such voters are seeking not to 'waste' their House votes.(46)

A second possibility is that there are voters who actually seek to produce an upper house with a political balance different from the lower house. It has been suggested that these voters see merit in the existence of a strong upper house, on the grounds that it 'represents a strategic way' for voters 'to express issue-based grievances or to put a brake on major party dominance'-even if they support a major party in the lower house election. According to Bean and Wattenberg such strategic voting that seeks to curb executive power, 'is a significant factor in explaining different patterns of party strength' between the two houses in Australia.(47)

Whatever the reasons for this, vote-splitting of this type plays a part in reducing the chances of a major party gaining control of the Senate, but was not predicted by those instituting the change in 1948.

Interfering with the logic of an established voting method

Many voting methods have been created by mathematicians, concerned to produce a method that has an internal logic, and which produces a planned type of result-the creators of preferential voting saw it as having particular benefits not possessed by the first-past-the-post method, while those who developed different methods of proportional voting sought to achieve a certain type of proportionality in election results. An Australian example of a piecemeal change that ignored a system's internal logic occurred in regard to the first ACT elections after the achievement of self-government.

The first two elections for the ACT Legislative Assembly used a modified form of the d'Hondt method of proportional representation. An electoral threshold, such as those discussed earlier in this paper, was inserted into the legislation, as was an arrangement concerning the distribution of preferences which was uniquely Australian in its concept. The threshold was designed to eliminate all parties and candidates from the count if they failed to reach one quota (5.6 per cent) on first preferences. The 'modified d'Hondt', as it was called, was in fact the result of deals done between parties in the national parliament, rather than a carefully-applied modification of this well-tried European voting method. The Australian Electoral Commission was very critical of this distortion of the d'Hondt system, and the Electoral Commissioner concluded that 'any plausible justification for it in terms of underlying principles is now invisible behind the veil of compromises and trade-offs'.(48) The altered method certainly pushed many candidates out of the count, for in the 1989 election 49 per cent of the candidates (57 of 117) were eliminated after the count of first preferences. Quite unexpectedly, however, the proportionality of the final result was lessened in an unfortunate fashion. The Australian Electoral Commission noted that three of the excluded parties and one excluded independent

Another threshold problem relates to an apparent lack of understanding in reform suggestions of the way thresholds typically are used elsewhere. Most operate in systems where election quotas are very low and there is a great likelihood of a massive splintering of the party system. Australian Senate quotas are much higher, and parties need a significant share of the popular vote to get a candidate elected. It has been claimed that the purpose of the imposition of thresholds for Senate elections would be 'to skew the system in favour of candidates from larger parties', rather than to achieve the degree of representation that proportional representation can achieve.(50)

The message might be to leave properly functioning voting methods well alone-or else introduce a completely different method.

Can Australia establish a perfect electoral system?

A discussion such as this implicitly raises a simple question that is extremely difficult to answer: 'What is the best voting method?' Despite the claims that are made by advocates of different methods, there is, in fact, no 'best' method, because the functions of particular elections may well be different for different legislative bodies.

Despite controversies that often surround national electoral systems, it is a fact that they are rarely changed in liberal democracies. When changes do occur they tend to be minor technical adjustments such as registration questions, refinements to the provisions for the drawing of boundaries, or alteration of polling-day arrangements, rather than significant alterations of the type recently spoken of in Australia. Even incoming governments rarely break the pattern: 'Particular electoral systems are maintained even when the elites forming the government change'.(51) A major factor in this seems to be the recognition of a point made earlier in this paper, that changes to electoral systems are often unpredictable in their consequences. An obvious example has been the unstable parliamentary situation in New Zealand that is the direct consequence of the establishment of a new electoral system. Since the first MMP election in 1996, no party has been able to gain an absolute majority in the national parliament, and there has been a high degree of parliamentary instability. It is already clear that the difficulty of predicting the consequences of voting method change for the British House of Lords is likely to be the problem that is most likely to wreck any sort of consensus on the question of reform of the Lords.(52)

When the question of altering the nation's electoral system arises, is it best to leave well alone-particularly if appears to be working adequately, and is accepted by the great mass of the population? University of Western Australia political scientist, Bruce Stone, has noted that:

As Professor Don Aitkin has argued, Australia may have the balance of party and public needs just about right, claiming that he is yet to see another national electoral process that is 'as fairly conducted, as universally supported and as decisive'.(54) A British commentator on voting methods has asserted that:

It so happens that in the past fifty years the Commonwealth electoral system, as well as the systems in New South Wales, Western Australia and South Australia, have adopted the pattern of a majority system for the lower house and a proportional method for the upper (Tasmania has this in reverse). Has Australia stumbled on a reasonable compromise?

Endnotes

2. Alan R. Ball, Modern Politics and Government, Macmillan, Basingstoke, 5th ed 1993, p. 160.

3. Report of the Advisory Committee on Executive Government June 1987, Canberra Publishing and Printing, Canberra, 1987, p. 22.

4. Final Report of the Constitutional Commission, AGPS, Canberra, 1998, vol. 1, pp. 218-47.

5. Laurie Oakes, 'Case for Senate reform crosses the political divide', Bulletin, 9 February 1999, p. 32. For a Labor view that makes the same point, see John Faulkner, 'How to really fix up the Senate', Age, 9 February 1999.

6. Quoted in Eric Edgar Howitt, Cameral Government. Unicameral and bicameral legislatures, Viridia Books, Melbourne, 1992, p. 187.

7. Helen Coonan, 'The Senate. Safeguard or Handbrake on Democracy?', address to Sydney Institute, February 1999, p. 23.

8. Ibid.

9. Ibid., p. 20.

10. Scott Bennett, 'The Decline in Support for the Major Parties and the Prospect of Minority Government', Research Paper No. 10 1998-99, Department of the Parliamentary Library, 1999, p. 4.

11. See Department of the Parliamentary Library Research Notes, 'The Reduction in the Size of the Tasmanian Parliament' (no. 2, 1998-99) and 'Tasmanian Election 1998' (no. 6, 1998-99).

12. This change would also require a complementary reduction in the size of the House of Representatives. s.24 of the Constitution says that the number of MHRs 'shall be, as nearly as practicable, twice the number of the senators'.

13. Scott Bennett, Winning and Losing. Australian National Elections, Melbourne University Press, Melbourne, 1996, pp. 48-51.

14. Ibid., p. 82.

15. Australian Financial Review, 12 June 1997.

16. Australian, 25 February 1999.

17. Mr Higgins (Vic), Official Record of the Debates of the Australasian Federal Convention Sydney, 2nd to 24th September 1897, Government Printer, Sydney, 1897, p. 371; see also Mr McMillan (NSW), p. 369.

18. Coonan, op. cit., p. 24.

19. Ibid.

20. Martin Harrop and William L. Miller, Elections and Voters. A Comparative Introduction, Macmillan, Basingstoke, 1987, p. 65.

21. Coonan, op. cit., pp. 25-7.

22. For more on thresholds for the Senate, see Margaret Healy and Gerard Newman, 'An Electoral Threshold for the Senate?', Department of the Parliamentary Library, Research Note 1998-99, forthcoming.

23. Ibid., p. 27.

24. See Appendix.

25. John Cherry, discussion paper prepared for Senator Meg Lees, January 1999, p. 5.

26. See John Warhurst, 'Changing the British Voting System? The Jenkins Commission Report', Research Note 1998-99, Department of the Parliamentary Library, forthcoming.

27. Cherry, op. cit., p. 8, see also table pp. 12-13.

28. Australian, 8 December 1998.

29. Bill Coxall and Lynton Robins, Contemporary British Politics, Macmillan, Basingstoke, 3rd ed 1998, p. 138.

30. Geoffrey Palmer and Matthew Palmer, Bridled Power. New Zealand Government under MMP, OUP, Auckland, 1997, pp. 22-4.

31. Scott Bennett, Andrew Kopras and Gerard Newman, 'Federal Elections 1998', Research Paper No. 9 1998-99, Department of the Parliamentary Library, 1999, p. 10.

32. Cherry, op. cit., pp. 2-3.

33. The elections of 1954, 1961 and 1969 were also affected by an electoral inequality that is no longer permissible under the Commonwealth Electoral Act. There was no requirement for regular redistributions, and the quota variation was 20 per cent (today ten per cent).

34. Martin Harrop and William L. Miller, Elections and Voters. A comparative introduction, Macmillan, Basingstoke, 1987, p. 51.

35. Bennett, Winning and Losing, p. 59.

36. Cherry, op. cit., p. 3.

37. David Solomon, 'Elect the government!', in Michael Coper and George Williams (ed), Power, Parliament and the People, Federation Press, Sydney, 1997, p. 51.

38. Douglas W. Rae, The Political Consequences of Electoral Laws, Yale University Press, New Haven, rev. ed. 1971, pp. 19-21.

39. Ibid, p. 70.

40. Bennett, Winning and Losing, pp. 59-60.

41. Commonwealth Electoral Act 1902.

42. Commonwealth Electoral Act 1918-1919.

43. Commonwealth Electoral Act 1948.

44. The Tasmanians had long been aware of this. After the 1955 and 1956 elections for five six-member electorates produced 15-15 results in the House of Assembly, the parliament was increased so that each electorate had seven members. The 1998 reduction was to five-member electorates.

45. Courier Mail, 4 February 1999.

46. Figures calculated from Gerard Newman, 'Federal Election Results 1949-1998', Research Paper No. 8 1998-99, Information and Research Services, Department of the Parliamentary Library, 1999.

47. Bennett, Winning and Losing, pp. 174-5.

48. Shaun Bowler and David Denemark, 'Split Ticket Voting in Australia: Dealignment and Inconsistent Votes Considered', Australian Journal of Political Science, vol. 28, no. 1, March 1993, p. 30; Clive S. Bean and Martin P. Wattenberg, 'Attitudes Towards Divided Government and Ticket-splitting in Australia and the United States', Australian Journal of Political Science, vol. 33, no. 1, March 1998, p. 35.

49. Scott Bennett, The 1989 ACT Election, Political Science Department, The Faculties, Australian National University, Canberra, 1989, pp. 6-7.

50. Commonwealth Parliament, Inquiry into the ACT Election and Electoral System, Government Printer, Canberra, 1989, pp. 81-2.

51. Geoffrey Goode, Proportional Representation Society, letter to Age, 9 February 1999.

52. Andrew Reeve and Alan Ware, Electoral Systems. A comparative and theoretical introduction, Routledge, London, 1992, p. 4.

53. Modernising Parliament-Reforming the House of Lords', Cmd 4183, http://www.official-documents.co.uk/document/cm41/4183/ref-01.htm, 12 February 1999, ch. 7.

54. Bruce Stone, 'Small Parties and the Senate Revisited: The Consequences of the Enlargement of the Senate in 1984', Australian Journal of Political Science, vol. 33, no. 2, July 1998, p. 211.

55. Canberra Times, 4 March 1993.

56. Vernon Bogdanor, What is Proportional Representation? A guide to the issues, Martin Robertson, Oxford, 1984, p. 157.

Appendix: Non-major party House of Representatives victories 1946-98

Source: Australian Electoral Commission

Note: Hanson nominated as Liberal but lost party endorsement prior to polling day, but too late to have the party label removed from ballot papers. The Australian Electoral Commission includes her votes in the Liberal total, but counts her as an independent in the tally of seats won.