Home | Databases | WorldLII | Search | Feedback Victoria University of Wellington Law Review

Beever, Graham --- "The New Game with the Old Rules: Boundary Determination under MMP" [2003] VUWLawRw 5; (2003) 34(1) Victoria University of Wellington Law Review 135

The New Game with the Old Rules: Boundary Determination Under MMP

Graham Beever[*]

I INTRODUCTION

A The Political Context

In 1993, New Zealanders voted to adopt a new electoral system, Mixed Member Proportional Representation (MMP). MMP is a system of proportional representation in which voters elect members to represent each electoral district, and also cast a party vote to determine the final composition of Parliament. Surveys of public opinion at the time of the 1993 referendum indicated a high level of dissatisfaction with the previous First Past the Post (FPP) system, partly because of a feeling of "disenchantment with the two major parties".[1] According to a 1993 survey, only 14% of respondents considered the two main parties to be trustworthy.[2] It was felt that by introducing a system of proportional representation, voters would be able to weaken the power of government by necessitating the formation of coalition governments in most circumstances.[3] Voters also perceived that their votes were "more meaningful" because the party vote meant that "voters in safe electorates would have as much incentive to vote as those in closely contested electorates".[4] MMP meant that votes for smaller parties were not "wasted", because such parties were more likely to obtain seats in Parliament. Thus the MMP system achieved "greater fairness between parties" because it provided a more accurate representation of voters' choices in Parliament.[5]

The MMP system that was finally implemented at the 1996 general election had its genesis in the 1986 Report of the Royal Commission on the Electoral System. MMP was largely based on the Personalised Proportional System used in Germany since 1949. The Royal Commission's report recommended a number of important elements to be included in any implementation of the new electoral system in order to ensure its effectiveness in the New Zealand context. However the MMP system finally offered to New Zealanders in the Referendum of 1993 differed from that recommended by the Royal Commission in a number of key respects. The Commission recommended, amongst other things, that:

• the size of Parliament should be increased to 120 MPs irrespective of any decision to change the electoral system.[6]
• the electoral "tolerance", which determines the extent to which electorates can differ in population, should be increased from five per cent to 10 per cent.[7]
• the "threshold" for gaining representation in Parliament should be four percent of the party vote, or victory in one constituency seat.[8]
• this threshold should be waived for parties "primarily representing Maori interests".[9]
• the separate Maori seats should be abolished if MMP were adopted, on the basis that the Commission had "no doubt the Maori people would use the flexibility and opportunities of MMP to ensure that their interests were adequately represented".[10]
• each of the parties in the House of Representatives should have its own representative on the Representation Commission which determines electoral boundaries.[11]

These features were nowhere to be seen in the provisions of the Electoral Act 1993, which set out the mechanics of the MMP system that would apply in the event of an affirmative referendum vote. Voters faced a choice between an MMP system with 120 MPs, or the retention of FPP with 99 MPs.[12] The electoral tolerance was not increased to ten percent.[13] The threshold was set at five percent of the party vote or one constituency seat, and this requirement was not waived for Maori parties.[14] The Maori seats were not abolished, and the Representation Commission continued to contain only two political representatives: one to represent government parties, the other opposition parties.[15]

It is not entirely clear why Parliament chose to disregard so many of the Royal Commission's recommendations. The Electoral Law Select Committee's report on the Electoral Reform Bill which implemented the proposed system provides little illumination. Although the Committee specifically addressed the question of the number of MPs and the retention of the Maori seats, none of the other differences were considered in any depth in their report to Parliament.[16] In relation to some of the omissions, a cynical interpretation seems compelling. The leadership of the two major parties had been "vocal in their opposition to MMP in the 1992 Referendum campaign", where voters chose MMP as the system to compete against FPP in the 1993 referendum.[17] The major parties had much to fear: the likely end of majority single party government, the possible disintegration of their parties into separate publicly competing factions, and the disruption to the two party equilibrium, which might give new parties an opportunity to displace them. There was also a risk of displacement for incumbent MPs, as the parties re-selected their existing members for new constituencies and the party list. Although Labour and National were less vocal in their opposition to MMP in the course of the 1993 referendum campaign, this can be attributed to a variety of factors. First, it was becoming increasingly apparent that they were perceived as having a "vested interest" in retaining the status quo, and that the more vociferously they opposed MMP, the more popular it was likely to become. In addition, by 1993 there was a well organised and very well funded anti-MMP lobby group, the Campaign for Better Government, which diminished the need for intervention by the major parties.[18] It should also be borne in mind that, unlike the 1992 referendum, the vote in 1993 coincided with a general election, a major distraction for the parties.

The decision to retain 99 MPs if FPP prevailed invites the conclusion that Parliament was attempting to 'sabotage' MMP, given voters' known distaste for politicians.[19] Indeed the Electoral Reform Coalition, the major pro-MMP lobby group, complained of "a parliamentary plan to prejudice the MMP vote by presenting an uneven playing field".[20] The decision to retain the Maori seats and not provide for a waiver of the threshold for Maori parties could perhaps be put down to a reluctance by Parliament to tackle such a potentially controversial and divisive issue.[21] However the adoption of a higher threshold, and the failure to include representatives from all parties on the Representation Commission, both seem to soften the impact of MMP for the two major parties, even if only slightly. Although the failure to increase the tolerance from five to ten percent seems less clearly to favour the incumbent parties, it is perhaps an indication of a reluctance to change the existing electoral law framework more than was absolutely necessary.

It therefore seems probable that a number of aspects of our current MMP system were adopted, not because of their contribution to the workability of MMP which was the Royal Commission's primary consideration, but because they made MMP less desirable and served to entrench the positions of the two major parties. From such a starting point it is hardly surprising that political players, and to a lesser extent voters, are sometimes accused of "not having adjusted to MMP".[22] New Zealanders are playing a new political "game", but some of the old rules still apply. This paper will examine the extent to which our electoral laws are out of tune with MMP in the area of electoral boundaries, and in particular the boundary determination process. Two specific aspects of the process will be examined: the composition of the Representation Commission, and the retention of the electoral tolerance level at plus or minus five percent.

B The Role and Origins of Electoral Boundaries

The use of electoral boundaries to define geographical constituencies is a feature of most electoral systems. By linking elected representatives to a specific place, electorates provide a bond between citizens and their representatives. This bond enhances access by citizens to their representatives, and in turn assists representatives to remain in touch with the people and communities they represent. To the extent that social or economic interests coincide with geographical location, such a system can also provide an effective way to provide representation for a diverse range of interests.

The precise mechanism by which representatives are chosen within a constituency based system can vary. In many jurisdictions several members are chosen to represent each constituency, as in the case of the "Single Transferable Vote" (STV) system used in Ireland and elsewhere.[23] Such an approach might result in overly large electorates in a sparsely populated nation like New Zealand, whose MMP system operates on the basis of single-member constituencies. Electoral systems can combine single-member and multi-member elements. This is the case in Japan where members of the two Houses of the Diet (Parliament) are elected to represent a variety of single-member and multi-member regions.[24]

New Zealand's electorate based system of representation has its origins in medieval England. Members of Parliament up until the eighteenth century were chosen to represent particular boroughs or counties. MPs were "conceived of as representing places (or property) rather than people".[25] This accorded with the property based voter franchise of the time. However, as the modern democratic principle of "one person – one vote" began to emerge, the extraordinary disparity between the sizes of different electorates increasingly became an issue.[26] Such disparities meant, in effect, that some voters were getting a much greater "share" of representation in Parliament than others.

The first of a series of major revisions of electoral boundaries in Britain occurred in 1832.[27] Whilst Britain has never formally adopted a requirement for electorates to fall within a certain percentage range (or tolerance) of the average electorate size, equality of population is now established as a criterion of considerable importance in the determination process.[28] In New Zealand, population equality has been given even greater legal recognition through the implementation of a binding tolerance of plus or minus five percent deviation from the average electorate size.[29] The level of this tolerance, particularly in the new context of MMP, is discussed in detail in section IV.

C The Importance of Electoral Boundaries

A parliament that reflects the will of the people is the defining characteristic of democracy. One measure of the extent to which this has been achieved under a given electoral system can be obtained by considering the degree to which the percentage of votes cast for each party has been accurately translated into its share of seats. In the most pure possibility, a proportional system with no "threshold" for representation, this translation from votes to seats is likely to be very accurate. This is the case under the purely list-based proportional system used in Israel.[30] The MMP systems of Germany and New Zealand come close to this ideal, but they are compromised by the existence of the five percent threshold, and the possibility of "overhang" seats.[31] Of course these measures have been introduced for valid reasons, the former to minimise the risk of a "proliferation [of parties which] could threaten the stability and effectiveness of government",[32] and the latter to enable the highly valued role of electorate based MPs, whilst limiting the possibility of an overly large Parliament.

Concerns about the workability of Parliament can also give rise to calls for an electoral system that departs in more extreme ways from the principle of proportionality, such as New Zealand's former FPP system. Under such a system there is no expectation that parties will obtain seats in proportion to their share of the vote. Instead the electoral contest is broken down into a series of separate races in each constituency with the party receiving the most votes in a given electorate winning that seat. Within each electorate the success of the winning party is exaggerated – by obtaining the largest share of votes the winner receives 100 percent of the "seat". Although this distorts the principle of proportionality, it remains "fair" within each electorate because the winner of each seat does have the highest level of support. However when these individual results are aggregated to provide the overall composition of Parliament there is a risk that this "fairness" will evaporate. There is no guarantee, particularly in a close election, that the party that receives the highest share of the vote will receive the largest proportion of seats.

This is not merely an abstract possibility. In the New Zealand elections of 1978 and 1981, National won a majority of seats, and thus remained in government, on the basis of fewer votes than Labour.[33] A similar effect occurred in the 2000 Presidential election in the United States. Democratic candidate Al Gore obtained a greater share of the nationwide popular vote, but Republican George Bush was (eventually) successful on the basis that he had won a greater share of the electoral college vote, a system whereby each State casts all of its votes in favour of the candidate who received the highest level of support within the State.[34]

In close votes such as these, the exact contours of electoral boundaries (or the less arbitrary state boundaries in the United States example) can be determinative of the eventual outcome. This is because the support for different parties is not distributed uniformly throughout the country – indeed if it were the party which received the most votes would win all the seats. Two factors have the potential to produce a distortion in the translation of votes into seats: the precise location of electorate boundaries, and the relative population of electorates.

To see how the location of electorate boundaries can be so influential, suppose there are two adjacent electorates. Seat 1 is a "safe" seat for Party A, which has an equally high level of support throughout that electorate. Seat 2 is a marginal seat with a narrow majority of support for Party B ahead of Party A. If the boundary between the two electorates were moved sufficiently far within the present boundaries of Seat 1, Party A could secure victory in both seats. If deliberately contrived this process is known as "gerrymandering" although it may of course occur as a result of legitimate determination or revision processes. This effect was at play in the 1978 and 1981 elections discussed above. National was able to secure victory because Labour's votes were more concentrated in their "safe" seats than National's were.

Gerrymandering, be it deliberate or unintended, is not the only way that the contours of electoral boundaries can influence the outcome of an election. Variation in relative electorate size, or malapportionment, can also have major implications. The "unfairness" inherent in unequal electorates that was noted earlier has the potential to substantially influence the overall result, particularly if the size disparities are correlated with some social or economic factor. In New Zealand one such factor of particular importance is the distinction between the primary and secondary economies.[35] Before 1945, a "country quota" operated in New Zealand which permitted rural constituencies to have significantly smaller populations than average, ostensibly to ensure that the electorates did not become unmanageably large in geographical terms.[36] Any party that had particularly strong support from rural voters would naturally benefit from this as they would receive a higher ratio of seats per voter than a party whose support was drawn largely from urban voters. Perhaps unsurprisingly, the country quota was abolished by the First Labour Government.

D Boundaries Under MMP

The importance of electoral boundaries is greatly diminished under MMP because the makeup of Parliament is determined by the party vote. Therefore the number of electorate seats won by a party will generally not affect the number of seats won overall. However electorate seats do still provide a number of strategic advantages.

For example, holding a significant number of electorate seats provides an opportunity for a party to increase its share of seats above its share of the party vote. Under the Electoral Act 1993, electorate seats won by a party in excess of its party vote entitlement are retained, giving it a greater proportion of seats in Parliament - often described as an "overhang". This in turn could be crucial to the overall outcome. Such overhang seats are a regular occurrence under MMP in Germany where each of the 16 federal states has separate party lists, however to date there have been no overhang seats in New Zealand.[37]

Perhaps more importantly, in the case of smaller parties, the winning of one or more electorate seats may be crucial to securing any representation in Parliament. As was noted earlier, the Electoral Act requires a party to win either five percent of the party vote or at least one electorate seat to receive its full entitlement of seats. This "threshold waiver" was invoked in 1999 when New Zealand First narrowly won Tauranga thus securing a total of five seats in Parliament despite falling short of the five percent threshold.[38]

There may also be other more subtle advantages of holding electorate seats. Arguably, they are still seen by the public as more prestigious, and possibly offer advantages in campaigning by providing a more clearly defined focus for local political activity. From the point of view of an individual MP, holding an electorate seat gives much greater control over one's electoral fate. A good performance as a local MP is likely to translate directly into electorate votes, whereas a high performing list MP is still at the mercy of their party's overall electoral performance to secure re-election.[39]

II THE BOUNDARY DETERMINATION PROCESS

As noted above, the precise way in which boundaries are drawn can have potentially major political consequences, even going so far as to determine the outcome of an election. It would seem therefore to be of fundamental importance that the process by which such boundary determinations occur must be as fair as possible. This suggests a number of requirements:

First, the body charged with making determinations must be independent, and be seen to be independent, from Parliament. Such independence ensures that even if the exact boundary determinations do favour a certain party or group of parties, they are no more likely to favour one than any other. Thus it is preferable for such a task to be delegated to an independent commission comprising experts of established integrity and independence. If political involvement is required it should involve representation from all significant political groupings so that any partisan interests can be expected to "cancel each other out".

Secondly, it is important that such a commission be delegated full authority to make decisions, and not be required to recommend determinations to Parliament with the possibility they will be rejected. To allow this would be to render the independence of the commission somewhat redundant.

Finally, it is desirable that the commission be given clear guidelines on which to base their determinations. By constraining the commission in this way the risk of politically motivated decisions will be limited further still. Preferably these guidelines will serve to limit the risk of gerrymandering by providing non-arbitrary criteria to determine boundary location, such as topographical features, and communities of interest. The guidelines could also be expected to limit the risk of malapportionment by placing some constraint on the permissible variation between the population of electorates. The risk of malapportionment arising as a result of population shifts can also be minimised by insisting on regular boundary reviews.

By international standards, New Zealand's boundary determination process measures up favourably against these criteria.[40] Section 28 of the Electoral Act provides for a Representation Commission made up primarily of politically independent appointees. The Commission's decisions take effect upon publication in the Gazette and notification to the Governor-General.[41] The Commission's determinations are tightly constrained by a detailed formula including a number of criteria governing the positioning of boundaries.[42] The Commission is also constrained by a narrow tolerance for variation in electorate populations.[43] The Commission is required to review the boundaries every five years, based on the data from each periodical census.[44] The risk of malapportionment arising between reviews is further diminished by providing that one of the criteria for determining the boundaries of electoral districts shall be "any projected variation in the ... electoral population of those districts during their life".[45] The provisions relating to the Representation Commission are "reserved" or entrenched, under section 268 of the Act.[46] This means that they can only be amended or repealed by a three quarters majority of the House, or by a simple majority in a referendum.[47] This emphasises the constitutional importance of the provisions and their central role in the electoral process.

However the conformity of the electoral process with the criteria for fairness proposed above is not absolute. In particular, the continued presence of political representatives on the Representation Commission has been controversial, especially considering there are only two such representatives, not one for each party. This apparent unfairness in the composition of the Representation Commission, given the multi-party nature of proportional representation, is one important way in which the electoral laws underpinning MMP are inappropriate for the new electoral system.

III THE COMPOSITION OF THE REPRESENTATION COMMISSION

A The Problem

The composition of the Representation Commission is set out in section 28 of the Electoral Act 1993. Although the formula that the Commission must apply in its decision making was amended from the earlier Act to allow for the smaller number of constituencies under MMP, and a new decision making criterion was added,[48] the membership of the Commission itself was unchanged.[49]

As noted, the Representation Commission consists primarily of politically neutral appointees. These include ex officio the Surveyor-General, the Government Statistician, the Chief Electoral Officer, and the Chairperson of the Local Government Commission.[50] There is also a separately appointed Chairperson, who in practice is always a judge.[51] However in addition to these neutral members of the Commission there are two persons appointed "on the nomination of the House of Representatives ... one of those members being nominated to represent the Government and one to represent the Opposition".[52]

This provision was relatively uncontroversial under FPP, where Parliament was overwhelmingly dominated by the two major parties. However, under MMP where there are currently seven parties represented in Parliament, the justification for this state of affairs is unclear. As the appointments are made on the nomination of the House they are still essentially controlled by Labour and National, to the exclusion of the smaller parties.[53] It is no coincidence that both of the political representatives during the last boundary revision were former MPs for the respective major parties in government and opposition.[54]

The risk of an unfair outcome arising from such arrangements is heightened in the case of a party such as the Greens who, though allied to the Labour-led Government, are not part of the governing Coalition and are therefore expected to be represented by the Opposition appointee. The notion that this person can fairly represent the interests of both National and the Greens is plainly absurd – these two parties sit on opposite sides of the political spectrum. This clear conflict of interest would be especially sensitive if the Representation Commission were considering the boundaries of the Coromandel electorate, a seat which was hotly contested between the Greens and National in the elections of both 1999 and 2002. The seat was potentially of huge importance to the Greens as it guaranteed them a "threshold waiver" if they fell short of five percent of the party vote, under the rule noted earlier.[55] Even if the Opposition appointee was neutral as between National and the Greens, this situation would amount to an irreconcilable conflict of obligations.

B International Comparisons

The MMP Review Committee considered the issue of whether or not to retain the political appointees on the Representation Commission. Perhaps not surprisingly they were unable to reach the near unanimity required under their terms of reference to lead to a recommendation.[56] Whilst the two major parties supported the status quo, all other parties participating in the Committee "considered the current situation to be unsatisfactory".[57] The ACT, Alliance, Green and United Parties all supported the complete removal of political representatives from the Commission, and as a second option generally supported the inclusion of representatives from all parliamentary parties.[58] Labour and National argued that the "value and necessity" of political representatives justified their retention, and that the inclusion of further representatives would render the Commission "too big and unwieldy".[59]

One of the arguments put forward by the smaller parties on the MMP Review Committee for the removal of political appointees was the fact that "of the major democracies that use single member districts in their electoral systems", entirely neutral commissions are invariably used.[60] Cited in support of this point are the systems used in Australia, Canada, Germany, India and the United Kingdom. A notable exception to this pattern is the United States, where partisan or bi-partisan committees generally control the boundary determination process. Whilst there is some evidence to suggest that such "conflictual redistricting" may actually have some advantages for the democratic process,[61] the United States system is characterised by extensive litigation and provides a poor model for New Zealand.[62]

The case of Germany provides an especially relevant example, as it is the German model of proportional representation that inspired New Zealand's MMP system. In Germany, a neutral commission, the Wahlkreiskommission, or Electoral Districts Commission (EDC), draws up electoral boundaries in accordance with similar statutory criteria to New Zealand.[63] The EDC is made up of the President of the Federal Statistical Office, a judge of the Federal Administrative Court, and five other (neutral) members.[64] Unfortunately however, the approach of the German boundary determination system leaves much to be desired when assessed against the requirements for fairness proposed earlier. In particular, the EDC does not have delegated authority to determine boundaries, but must report to the Bundestag or Parliament who may accept or reject their recommendations.[65] Whether the German EDC would be likely to retain a completely neutral membership if it were ever given full delegated authority is a matter for conjecture. However New Zealand could certainly learn from the current composition of Germany's EDC.

A better example may be provided by a much closer jurisdiction, Australia. Boundary determinations for the Federal House of Representatives in Australia are determined separately in each State through a two stage process. A set of proposed boundaries is drawn up by a "Redistribution Committee" which consists of the Federal Electoral Commissioner, the State Electoral Officer, the State Surveyor-General, and the State Auditor-General.[66] A period of consultation, and a final determination, is made by the "augmented Electoral Commission" a body consisting of the Redistribution Committee for the State plus the Federal Electoral Commission. The Federal Electoral Commission in turn consists of three members, all politically neutral, and the Federal Electoral Commissioner is one of these members, so the total size of the augmented Electoral Commission is six. These individuals are all senior public servants, or members of the judiciary. Neither the Australian Parliament, nor the relevant State Parliament, has the power to reject or amend the recommendations of the augmented Electoral Commission.[67]

The two-stage process adopted in Australia ensures that whilst State officials with local knowledge are dominant in the provisional boundary determination process, the final decision is made under the supervision of a Commission evenly split between State and Federal officials with a Federal official, a member of the Federal Electoral Commission, in the Chair. There is no such need to balance the roles of State and Federation in New Zealand and it seems that there is little reason to change the fundamental structure of the Representation Commission which has worked well for nearly 120 years.[68] However the entirely neutral commission, coupled with full delegated authority, provides an excellent model for other jurisdictions.

C Conclusion on the Political Appointees

Given the importance of ensuring a fair and independent boundary determination process, it is submitted that the positions for political representatives on the Representation Commission should be abolished. By failing to acknowledge the fact that New Zealand now operates a multi-party electoral system, the current arrangements are clearly inappropriate for the new MMP environment.

Although the adoption of political representatives for all parliamentary parties might be preferable to the status quo in principle, it would result in a Commission dominated by partisan interests, even if it could be hoped that these would "cancel out". It would also mean that parties that have not yet been elected to Parliament and which are seeking to establish themselves would not be represented on the Commission, whilst the interests of current parliamentary parties would. Whatever legitimate reasons there may be for partisan input during the determination process, these can be adequately addressed through the submission processes established under sections 34 and 38 of the Electoral Act which provide opportunities for parties to have input at an early stage in proceedings, as well as prior to the final determination.[69]

It should be noted that an identical criticism could be made regarding the composition of the Electoral Commission, when it performs the politically sensitive task of allocating broadcasting time and funding for party political broadcasts.[70] Although the Electoral Commission is not directly involved in the issue under consideration, the determination of electoral boundaries, the situation is clearly comparable to that of the Representation Commission. Both Commissions include two political appointees, appointed on essentially the same basis,[71] and both deal with matters of considerable political sensitivity. The Electoral Commission itself has vigorously opposed the presence of political representatives on both bodies, citing the need for impartiality, and noting the fact that the present arrangements appear to favour the two major parties.[72]

The "unfairness" of the current composition of the Representation Commission is not the only way in which New Zealand's boundary determination process has failed to adjust to MMP. Another important aspect is the level of the electoral tolerance which has not been adjusted to reflect the diminished importance of electorate seats under MMP. Whilst this may not result in any great injustice for individual parties, it does lead to a number of unfortunate and, it is submitted, unnecessary consequences.

IV THE ELECTORAL TOLERANCE

A The Royal Commission and Select Committee Reviews

As already noted, the Royal Commission on the Electoral System which considered the possible implementation of MMP in New Zealand, recommended that the tolerance be increased to plus or minus 10 percent if MMP was to be implemented. In doing so they emphasised that "the tolerance can be increased without affecting the ultimate fairness or proportionality of representation in the House".[73] This preference for a 10 percent tolerance was reflected in the original Electoral Reform Bill 1992.[74] However it was decided during consideration of the Bill by the Electoral Law Select Committee that this would allow for too great a variation in the size of electorates, and the Act as it was finally passed contained a tolerance of just five percent.[75]

In their inquiry into the 1996 election, the first under MMP, the Electoral Law Select Committee revisited the question of the size of the tolerance. They also concluded that a tolerance of 10 percent "had the potential to allow too great a variation in the size of electorates".[76] However a substantial minority of the Committee, consisting of the Labour and Alliance members, supported an increase in the tolerance. In 2001, the MMP Review Committee examined this issue yet again.[77] A substantial majority of submissions on the issue advocated an increased tolerance.[78] However once again the Committee was unable to reach agreement on the issue. The ACT, Green, and National Parties all favoured an increase in the tolerance to 10 percent. Labour and United Future supported retention of the current five percent tolerance, whilst the Alliance conditionally supported retention. In its report the committee emphasised the point that "an increase to 10 percent could see electorates differ from each other by as much as 20 percent" which was thought to be too great a divergence.[79] A 20 percent divergence is possible if the smallest electorate is 10 percent below quota, and the largest is 10 percent above quota.

One interesting point which emerges from the above narrative is the near complete lack of consistency from individual parties on the appropriate size for the tolerance. Even in the brief period from the Electoral Law Select Committee Inquiry into the 1996 Election, which was released in 1998, and the MMP Review Committee's report of 2001, virtually all the parties involved had changed their position. In 1998, National and ACT both supported retention of the status quo, while Labour and the Alliance preferred an increase in the tolerance.[80] By 2001, it was Labour and the Alliance who supported the status quo, while ACT and National favoured an increase![81] Only United (now called United Future) remained consistently opposed to an increase, while New Zealand First and the Greens were both present at only one of the two inquiries.[82] This lack of consistency suggests that support on either side of the issue is very "soft", perhaps because the issue is seen as relatively trivial and certainly not a matter of principle. This suggests a search for consensus on the issue may be far from futile.

B The Problems of a Low Tolerance

1 Communities of Interest

Perhaps the most significant problem caused by a low tolerance is the inevitable difficulty that the Representation Commission has in drawing boundaries that reflect "community of interest" as required under the Act.[83] This criterion is closely related to the requirement to consider "facilities of communications" and also "topographical features". As the Commission noted in the introduction to its 2001 report, "[t]he splitting of small communities has been avoided where possible. The Commission has endeavoured ... to place communities in the same electorate as the adjoining area with which [they have] the most interaction. It is conscious that this has not been achieved in all instances because of the requirement to balance factors of quota limits and other statutory criteria constraints".[84]

Examples of such difficulties are not hard to find. For instance, the small Canterbury towns of Amberley and Sefton, just north of Christchurch, have close social, economic, and cultural links with the city of Christchurch and the rural population of the Canterbury plains. As one objector suggested in response to the Representation Commission's 2001 report "their community of interest lies naturally with Rangiora and Christchurch".[85] Nevertheless these communities are included in the Kaikoura electorate, whose main centre is Blenheim, nearly 300 kilometres north.[86] This has major implications for access to the local MP, and makes the MP's job harder if they are required to represent such diverse interests. If the tolerance were larger, these communities could be included in the nearby electorate of Waimakariri which includes the northern outskirts of Christchurch, and a significant portion of the northern Canterbury plains.

2 Frequent Boundary Changes

A lower tolerance also leads to more frequent boundary changes. In New Zealand, this problem is largely confined to the North Island general seats and the Maori seats. Since the number of electorates in the South Island is fixed, the South Island boundaries need only occasional minor changes to reflect shifts in population distribution. In the North Island, however, on both occasions that the boundaries have been reviewed since the advent of MMP, the Representation Commission has had to confront a large increase in the relative population, especially in Auckland. They have addressed this on both occasions by creating an extra electorate in Auckland, which has led to major boundary changes across much of Auckland and beyond.[87] Equally drastic changes have been required in the Maori seats.

Frequent boundary changes disrupt the democratic process in several ways. The relationship built up between local MPs and their constituents is disrupted each time a particular community is transferred from one electorate to another. This undermines what the Royal Commission described as a "highly desirable" feature of our electoral system, namely the ongoing link between constituents and their local MP.[88] There is also significant administrative disruption to the Electoral Commission and the Registrar of Electors who have to modify their systems and structures, as well as political parties who have to form new electorate organisations.

3 Large electorates

Another problem related to a low tolerance is the unavoidability of having some physically very large electorates. This has always been a problem for the Maori seats but the reduction in the number of electorates under MMP has seen the emergence of some particularly large general electorates. The most extreme example of this is the West Coast-Tasman electorate which extends over 500 kilometres along much of the length of the South Island. Such large electorates make it more difficult for MPs to service their constituents' needs, a point which was emphasised in several submissions to the MMP Review Committee.[89]

However the problem of large electorates may be overstated when compared to the difficulties of grouping communities of interest, and the risk of frequent boundary changes. Modern communications and transport facilities should go some way to mitigating the problems of servicing such electorates. Another factor to consider is that parties are free to assign list MPs to work in specific electorates, and if a particular MP is having difficulty dealing with a large electorate it would be in their party's interest to assign a list MP to work with them. Other parties are also likely to want to establish a similar presence in electorates held by rival parties. It is also worth remembering that even New Zealand's largest electorates are dwarfed by comparison to some of the larger electorates in other jurisdictions. The electorate of Kalgoorlie in Australia covers over 200 million square kilometres and includes most of the land area of the State of Western Australia. This entire area is represented by one MP in the Federal House of Representatives.[90] Nevertheless, if very large electorates can be avoided without any overall unfairness it could only enhance the ability of MPs to perform their democratic functions.

C International Comparisons

New Zealand's five percent tolerance is low by international standards. As noted by the Review Committee, Australia employs a tolerance of 10 percent, and Canada, France, and Germany all have tolerances between 20 and 25 percent.[91] The United Kingdom has no specific tolerance, but its rather ad hoc approach to reviewing boundaries has led to substantial divergences from quota, often greater than 10 percent.[92] This is the case even though many of these countries operate FPP type systems where divergences in electorate population can have much greater political consequences. The United States provides the only exception to this trend, with tolerances often as low as one percent. However as already noted the United States system involves partisan determination of boundaries, often coupled with extensive litigation, and provides a poor model for New Zealand. The United States situation is also heavily influenced by the Fourteenth Amendment to the Constitution which guarantees "equal protection of the law", so a direct comparison with New Zealand is less appropriate.[93]

Once again it is worthwhile to examine the situation in Germany, given the close similarity between the German electoral system and that of New Zealand. Under the Federal Electoral Law, the Electoral Districts Commission was formerly required to draw up boundaries within a tolerance of 25 percent. However the Bundestag was free to disregard the EDC's recommendations provided no electorate exceeded a tolerance of 33 percent. In general the Bundestag was very reluctant to accept changes, often for partisan reasons, so Germany essentially had a de facto tolerance of 33 percent.[94] The Federal Electoral Law has recently been amended to reduce the target for the EDC to 15 percent, and the absolute limit to 25 percent, largely in response to an ongoing proliferation of overhang seats, brought about in part by malapportionment.[95]

It may well be that the huge difference in electoral tolerances between Germany and New Zealand is as much a result of the prevailing political cultures as anything else. The West German MMP system was set up in something of a political vacuum shortly after World War Two, meaning that it was essentially starting "from scratch". It seems that the German people have embraced MMP, with all its characteristics, in a way that New Zealanders are yet to do. It may be that the entrenched culture of FPP with its strong emphasis on local MPs has led to a perception in New Zealand that electorate seats "still matter" to a much larger extent than the electoral reality justifies.

D Conclusion on the Electoral Tolerance

The Royal Commission clearly identified that an "added advantage" of MMP would be "the lessening of problems associated with the boundary-fixing process".[96] This was partly because of the much diminished importance of winning electorate seats, but was also based on the assumption that there would be a relaxation of the tolerance as recommended by the Commission. The eventual decision to retain a tolerance of five percent under MMP was misguided because it failed to recognise that the fundamental problem that gave rise to the need for such a low tolerance, that is, the risk of partisan advantage arising through the boundary determination process, was vastly diminished. At the same time the substantial problems caused by a low tolerance, such as the difficulty in recognising communities of interest, were just as serious as before. The fact that MMP would require fewer electorates also meant that the problem of geographically large electorates, one symptom of a low tolerance, would be worsened under the new regime. The current tolerance level therefore fails to reflect the realities of the new electoral system.

For these reasons, it is recommended that the electoral tolerance be increased to plus or minus 10 percent. This would recognise the importance still placed on electorate seats, and would not lead to any overwhelming advantage to one or more parties. It would, however, provide the Representation Commission with a greater degree of flexibility, thus better allowing it to fulfil its statutory task of recognising communities of interest in its determinations. It would also reduce the frequency of boundary changes, as well as providing an opportunity for some of the largest electorates to be diminished to a slightly more manageable size.

V CONCLUSION

It has been argued that New Zealand's electoral law has failed to adjust fully to the new MMP electoral system in its approach to electoral boundary determination. This may partly be a result of a deliberate attempt by the two major parties to ensure that the new system would be as advantageous to them as possible. The presence of two political representatives on the Representation Commission fails to reflect the multi-party nature of MMP. In addition, the retention of the electoral tolerance at the relatively low level of five percent fails to recognise the opportunities that the new system provides, given the greatly diminished importance of electorate seats.

As already noted, the provisions relating to these two aspects of the electoral system are "reserved" under the Electoral Act. They therefore cannot be changed without a three quarters majority of Parliament, or a simple majority in a referendum. The latter option seems unlikely given the relatively minor public importance attached to these issues. In the case of the political representatives on the Commission, the consistent position of the two major parties in support of the status quo makes any change unlikely in the foreseeable future. However in relation to the tolerance level, no party other than United Future has held a consistent position in the two select committee inquiries to recently consider the matter.

Although the United Future Party plays a pivotal parliamentary role subsequent to the 2002 general election, a consensus amongst all other parties would provide the necessary majority. In any case, United Future's position may well be negotiable given the drastic change to the composition of the party's caucus.[97] Although this is clearly an issue of minor importance to most parties, it has considerable significance to the citizens of Amberley and Sefton who find themselves represented by an MP based 300 kilometres away, even though they clearly identify with the nearby electorate of Waimakariri. Likewise, it is an issue of importance to residents of the West Coast who find access to their local MP greatly inhibited by the very large size of their electorate. It is to be hoped that in their inquiry into the 2002 General Election, the Justice and Electoral Select Committee will finally be able to build a consensus in favour of reform.

* * *

APPENDIX: THE BOUNDARY DETERMINATION PROCESS IN DETAIL

Under section 35 of the Electoral Act the Representation Commission is required to review the electoral boundaries after each five-yearly census. The formula that the Representation Commission is required to apply in determining the boundaries is set out in sections 35 and 36 for general electorates, and section 45 for Maori electorates. Sections 35 and 36, but not section 45, are "reserved provisions" under section 268 of the Act meaning that they cannot be changed without a three quarters majority in the House of Representatives or a simple majority in a referendum. The key elements are as follows:

(1) There are to be 16 general electorates in the South Island.[98]
(2) The "electoral quota" for the South Island is determined by dividing the South Island's general electoral population by 16.[99] This figure represents the average size of general electorates for the South Island, and is currently 54,308.[100]
(3) The general electoral population of the North Island is divided by this figure which is then rounded to determine the number of general electorates for the North Island, which as of 2002 will be 46.[101]
(4) The North Island's electoral quota is determined by dividing the general electoral population of the North Island by the number of electorates obtained in step 3.[102] This will be close to the South Island quota. It is currently 54,296.[103]
(5) Steps 3 and 4 are repeated in relation to the Maori electoral population to determine the number of Maori electorates (seven after the 2002 election) and the Maori electoral quota (53,099).[104]
(6) The Commission is then required to determine the electoral boundaries in such a way as to ensure that the population of each electorate is within a "tolerance" of plus or minus five percent of the relevant electoral quota.[105]

In addition to this requirement the Commission must consider the following criteria when it is determining the boundaries:[106]

(1) The existing boundaries of electoral districts.
(2) Community of interest, or in the case of Maori electorates, community of interest among the Maori people generally and members of Maori tribes.
(3) Facilities of communications.
(4) Topographical features.
(5) Any projected variation in the electoral population of those districts during their life.

These final criteria are not ranked in any order of importance, but "each ... must be taken into account by the Commission which must gather information concerning them from such sources as are available to it."[107]

[*] Submitted as part of the LLM programme at Victoria University of Wellington.

[1] Jonathon Boston and others New Zealand Under MMP: A New Politics? (Auckland University Press, Auckland, 1996) 3.

[2] Jack Vowles and others Towards Consensus? The 1993 Election in New Zealand and the Transition to Proportional Representation (Auckland University Press, Auckland, 1995) 132.

[3] Jack Vowles and others Voters' Victory? New Zealand's First Election Under Proportional Representation (Auckland University Press, Auckland, 1998) 4.

[4] Voters' Victory? above, 3.

[5] Voters' Victory? above, 2.

[6] Royal Commission on the Electoral System Report of the Royal Commission on the Electoral System: Towards a Better Democracy (Wellington, 1986) 43.

[7] Royal Commission on the Electoral System, above, 43.

[8] Royal Commission on the Electoral System, above, 44.

[9] Royal Commission on the Electoral System, above, 44.

[10] Royal Commission on the Electoral System, above, 103.

[11] Royal Commission on the Electoral System, above, 295.

[12] The referendum consisted of a straight choice between the former FPP system as provided in the Electoral Act 1956, and the proposed MMP system as provided in the Electoral Act 1993.

[13] Electoral Act 1993, s 36.

[14] Electoral Act 1993, s 191(4).

[15] Electoral Act 1993, s 28.

[16] Electoral Law Select Committee "Report on the Electoral Reform Bill 1992" [1991-93] XXIV AJHR I 17C.

[17] Jack Vowles and others Towards Consensus? The 1993 Election in New Zealand and the Transition to Proportional Representation (Auckland University Press, Auckland, 1995) 178.

[18] Towards Consensus? above, 180.

[19] Towards Consensus? above, 157.

[20] Towards Consensus? above, 183.

[21] In relation to the decision to retain separate Maori seats the Report of the Electoral Law Select Committee on the Electoral Reform Bill provides some insight. The Committee held a number of hui to discuss the topic and concluded that "in light of ... the overwhelming amount of comment contained in the submissions favouring the retention of separate Maori representation ... the committee agreed the provision ... should be retained under the proposed MMP system".

[22] See for example, Colin Espiner "Is MMP the Villain?" (12 April 2002) The Press 7.

[23] Basil Chubb The Government and Politics of Ireland (2 ed, Standford University Press, Stanford, 1982) 146.

[24] Kichimoto Asaka "Electoral Reform in Japan: A Comparative Constitutional Law Perspective" (1997) 27 VUWLR 25, 29.

[25] David Butler and Iain McLean "The Redrawing of Parliamentary Boundaries in Britain" in David Butler and Iain McLean (eds) Fixing the Boundaries: Defining and Redefining Single-member Electoral Districts (Dartmouth, Aldershot, 1996) 1, 4.

[26] Butler and McLean, above, 4 onwards.

[27] Butler and McLean, above, 4.

[28] Butler and McLean, above, 14 onwards.

[29] Electoral Act 1993, s 36.

[30] In fact Israel now employs a very low threshold (1.5 percent).

[31] For instance the five percent threshold means that a party may gain four percent of the vote but receive zero percent of the seats in Parliament. Conversely if a party is the beneficiary of an overhang, which will be explained shortly, it may receive a greater share of seats than its share of the vote.

[32] Royal Commission on the Electoral System Report of the Royal Commission on the Electoral System: Towards a Better Democracy (Wellington, 1986) 45.

[33] Chief Electoral Office Official Election Results (Ministry of Justice, Wellington, 1978-1981).

[34] John R Vile A Companion to the United States Constitution and its Amendments (Praeger, Westport, Conn, 1993) 61. It is perhaps worth noting that the entire Florida vote count fiasco could have been avoided if the electoral college system had not been used: the outcome of the popular vote was clear.

[35] R J Johnston "Electoral Geography" in Martin Holland (ed) Electoral Behaviour in New Zealand (Oxford University Press, Auckland, 1992) 25, 26.

[36] Johnston, above, 41.

[37] Helga A Welsh "Four Years and Several Elections Later: The Eastern German Political Landscape After Unification" in David P Conradt and other (eds) Germany's New Politics: Parties and Issues in the 1990s (Berghahn, Providence, 1995) 43, 53.

[38] Chief Electoral Office General Election 1999: Official Results (Ministry of Justice, Wellington, 1999).

[39] In a survey commissioned by the MMP Review Select Committee, 61 percent of respondents agreed with the statement that "list MPs are not as accountable to voters as electorate MPs", while only 15 percent disagreed.

[40] These provisions are summarised in more detail in the Appendix.

[41] Electoral Act 1993, s 38(9).

[42] Electoral Act 1993, s 35(3)(f).

[43] Electoral Act 1993, s 36.

[44] Electoral Act 1993, s 35(2)(c).

[45] Electoral Act 1993, s 35(3)(f)(v).

[46] However s 45 of the Electoral Act 1993, which applies to the determination of Maori electorate boundaries, is not reserved.

[47] The provisions are only singly entrenched. Therefore it would be possible in theory for Parliament to repeal the entrenching provision (s 268) by a simple majority, and then proceed to amend or repeal the entrenched provision, also by a simple majority. The risk of adverse criticism, and perhaps a sense of moral obligation, has prevented this to date.

[48] The new criterion was the consideration of any projected population variation within the lifetime of the boundaries.

[49] Beth Bowden and Lloyd Falck "Redistribution and Representation: New Zealand's New Electoral System and the Role of the Political Commissioners" in David Butler and Iain McLean (eds) Fixing the Boundaries: Defining and Redefining Single-member Electoral Districts (Dartmouth, Aldershot, 1996) 147.

[50] Electoral Act, s 28(2).

[51] Electoral Act, s 28(2)(f).

[52] Electoral Act, s 28(2)(e).

[53] See "Manipulation Claim on Electorate Boundaries" (25 July 2001) The Press 6.

[54] Hon David Caygill, former Labour MP, representing government parties; Ian McLean, former National MP, representing opposition parties.

[55] In fact the Greens obtained five percent of the party vote in both elections.

[56] MMP Review Committee "Inquiry into the Review of MMP" [2001] AJHR I 23A, 10.

[57] MMP Review Committee, above, 53.

[58] MMP Review Committee, above, 53.

[59] MMP Review Committee, above, 53.

[60] MMP Review Committee, above, 54.

[61] See Andrew Gelman and Gary King "Advantages of Conflictual Redistricting" in David Butler and Iain McLean (eds) Fixing the Boundaries: Defining and Redefining Single-member Electoral Districts (Dartmouth, Aldershot, 1996) 207.

[62] See for example Davis v Bandemer [1986] USSC 160; (1986) 478 US 109.

[63] German Federal Electoral Law, Art 3(2).

[64] German Federal Electoral Law, Art 3(1).

[65] German Federal Electoral Law, Art 3(3).

[66] Michael Maley, Trevor Morling and Robin Bell "Alternative Ways of Redistircting with Single-member Seats: The Case of Australia" in David Butler and Iain McLean (eds) Fixing the Boundaries: Defining and Redefining Single-member Electoral Districts (Dartmouth, Aldershot, 1996) 119, 127.

[67] Maley, Morling and Bell, above, 127.

[68] The Representation Commission was first established in 1887.

[69] Electoral Act 1993, ss 34, 38.

[70] Electoral Act 1993, s 4; Broadcasting Act 1989, s 71.

[71] Electoral Act 1993, s 8.

[72] See Electoral Commission "Political Representation on the Representation Commission" (MMP/ECOM/3).

[73] Royal Commission on the Electoral System Report of the Royal Commission on the Electoral System: Towards a Better Democracy (Wellington, 1986) 151.

[74] Electoral Reform Bill 1992.

[75] Electoral Law Select Committee "Report on the Electoral Reform Bill 1992" [1991-93] XXIV AJHR I 17C.

[76] Electoral Law Select Committee "Interim Report on the Inquiry into the 1996 General Election" [1998] AJHR I 17A, 86.

[77] MMP Review Committee Inquiry into the Review of MMP [2001] AJHR I 23A, 11.

[78] MMP Review Committee, above, 11.

[79] MMP Review Committee, above, 13.

[80] Electoral Law Select Committee "Interim Report on the Inquiry into the 1996 General Election" [1998] AJHR I 17A, 86.

[81] MMP Review Committee, above, 14.

[82] The Greens were a constituent party of the Alliance in 1998, and New Zealand First refused to participate in the MMP review.

[83] As noted, the criteria which the Commission must consider are set out in the Appendix.

[84] Representation Commission Report of the Representation Commission (Wellington, 2001) 9.

[85] Representation Commission Summary of Objections and Counter-objections (Wellington, 2002) <http://www.elections.org.nz/elections/general/repcomm.html> (last accessed 10 October 2002).

[86] Report of the Representation Commission, above, 52.

[87] Report of the Representation Commission, above, 52.

[88] Royal Commission on the Electoral System Report of the Royal Commission on the Electoral System: Towards a Better Democracy (Wellington, 1986) 20.

[89] MMP Review Committee Inquiry into the Review of MMP [2001] AJHR I 23A, 11.

[90] Australian Electoral Commission <http://www.aec.gov.au/> (last accessed 10 October 2002).

[91] MMP Review Committee, above, 12.

[92] David Butler and Iain McLean "The Redrawing of Parliamentary Boundaries in Britain" in David Butler and Iain McLean (eds) Fixing the Boundaries: Defining and Redefining Single-member Electoral Districts (Dartmouth, Aldershot, 1996) 1, 9.

[93] The United States Constitution, amendment XIV.

[94] See Butler and McLean, above, 12.

[95] German Federal Electoral Law, amendment XIII.

[96] Royal Commission on the Electoral System Report of the Royal Commission on the Electoral System: Towards a Better Democracy (Wellington, 1986) 45.

[97] The 2002 election saw the United Future Party's representation increase from one seat to eight.

[98] Electoral Act 1993, s 35(3)(a).

[99] Electoral Act 1993, s 35(3)(b). The general electoral population is a measure of the total population after subtraction of: (a) Maori electors who have opted to register on the Maori roll, and (b) a proportional component of Maori who are ineligible to vote.

[100] Representation Commission Report of the Representation Commission (Wellington, 2001) 2.

[101] Representation Commission, above, 2.

[102] Electoral Act 1993, s 35(3)(d).

[103] Representation Commission, above, 2.

[104] Electoral Act 1993, s 45. Representation Commission, above, 2.

[105] Electoral Act 1993, s 36.

[106] Electoral Act 1993, s 35(3)(f).

[107] Timmins v Governor-General [1984] 2 NZLR 298, 305 (HC).