Keith Rankin [lecturer, political economist]
3/37 Trafalgar St.
Onehunga, Auckland, New Zealand
email: firstname.lastname@example.org or email@example.com
30 July 2000
Should there be another referendum on the electoral system?
The electoral system is a core part of New Zealand's constitutional arrangements. Those arrangements form the written and unwritten bedrock around which the political, social and economic life of our nation revolves.
A stable and prosperous nation has constitutional arrangements that are perceived as only one remove away from permanent. The obvious contrast at present is Fiji, where constitutional arrangements rest on quicksand. Another example of the pitfalls of constitutional expediency is Peru, where the President has altered to constitution to prolong his own power.
Overturning the electoral system is a major constitutional event of the kind that no person should experience more than once in a normal lifespan. Fine-tuning, though, should be more frequent. But even that should be done sparsely, with due care, and only after public education and discussion.
The former "First-Past-the-Post" (FPP) electoral system was rejected comprehensively in 1992 and again in 1993, under circumstances which favoured a vote for the status quo.
To even consider restoring FPP after those results would be to create a Fiji scenario, in which constitutional arrangements would be seen as flexible and dispensable if they block the route to power of some autocrat. Despite the complexity of the issues, barely 200,000 New Zealanders could bring themselves to go to a polling booth in 1992 to register support for an electoral system which was the only one which New Zealanders had ever known.
A rejected electoral system should not be reconsidered for at least two generations. After two generations, a past system has the same novelty value as do new alternatives. The politics of nostalgia are not good politics. The past - eg the 1960s - was never quite the golden weather that we often paint it as. The political marketplace of the present is vastly more democratic than the closed shop duopoly of our recent past.
A new electoral system should be allowed to run for at least one generation. That's how long it lakes to build a new electoral culture, and to separate the electoral system from the governments of the day. MMP will not have really arrived until the journalists, public servants and politicians who learned their political craft in the MMP era have all passed through the system. The transition to an MMP parliament will only be complete when we no longer have any MPs who first entered parliament in 1990 or before.
There may be a case for replacing MMP with STV ("single transferable voting"). But not until the 2020s or 2030s. And probably not even then. STV is a recipe for the personality politics that New Zealanders have repeatedly shown a distaste for.
Opposition to FPP runs throughout New Zealand's political history. FPP created a political duopoly that has only once been broken . It took the two established parties (Reform and United/Liberal) to combine, during the Great Depression, before another party (Labour) could be admitted to that sanctum of duopoly power. In an unfortunate act of short-term political expediency, in 1934 Labour chose to support FPP and to stop campaigning for proportional representation. New Zealand had to wait another life-time before the two-party stranglehold was broken. Thanks to that short-sighted 1934 about-face, Labour has spent most of that life-time in impotent opposition. The events of 1992 and 1993 represent the culmination of over 100 years of struggle to achieve genuine representative democracy in New Zealand.
The danger is not only a return to FPP. Equally dangerous is a move to switch to the "Supplementary Member" SM system, which is FPP combined with token representation for "minor" parties. SM contains all of the faults of FPP, plus all of the alleged faults of FPP.
Referenda do not, in general, enhance the democratic process. They are required to legitimise changes to a nation's constitutional arrangements. But that's about all they are able to do.
The result of a referendum is particularly sensitive to the phrasing of the question put. A binding referendum should only ever be of the form that accepts or rejects a piece of legislation.
In general, referenda convey little more than sentiment. Scientific questions cannot be addressed by referendum. They can only be addressed by knowledge. Sentiment is best ascertained through scientifically-designed opinion polls. A range of questions must be asked.
If only a single question is offered, as in a referendum, people are obliged to fit a whole range of very different opinions - informed and uninformed - into a yes/no answer. The answer can then be interpreted in about as many ways as there are shades of opinion.
Even informed members of the public hold inconsistent views. There is some evidence that, in 1993, while MMP was well understood to be a proportional system, that nevertheless many voters thought that the mechanism of allocating seats was that of SM. Some people in 1996 seem to have thought that the 'candidate vote' was the most important party vote, and that the 'party vote' was a vote for a coalition partner. This misunderstanding about MMP seems to have dissipated in 1999.
What mattered in the 1993 referendum is that people voted not so much for MMP, but for the principle of proportional representation. Beyond such basic principles, decisions on electoral matters are essentially technical. Decisions on how to give the public what they want need to be made by a combination of informed representatives (who are answerable to their voters), public servants and independent psephologists and public policy academics.
Political representatives represent their voters' interests and values, not their opinions which are always likely to be, to some extent, uninformed and inconsistent. Members of the public neither the time nor the economic incentive to be informed about everything. We pay our representatives to be informed and to make wise judgements on our behalf.
If the results of social surveys are to be believed (I have in mind the work of Alan Webster and Paul Perry from Massey University, published in New Zealand Politics at the Turn of the Millennium), there is a substantial minority of public opinion in favour of non-democratic systems of governance.
If people with antidemocratic agendas form alliances with the cynics and the nostalgic supporters of FPP, it is possible that democratic means can be used to overturn democracy itself. Something like that happened in Germany in 1933 when voters turned en masse to the two parties on the extremes (Communists and National Socialists), both of which openly repudiated democracy.
If there is one occasion in which the wish of the people must be rejected, it is when people vote to dispense with democracy. One way to avoid this dilemma is to not offer it. One way to avoid a return to the political power-trips of Sir Robert Muldoon, Sir Roger Douglas and Ruth Richardson is to not offer the opportunity to give up the hard-won struggle for proportional representation.
Should there be a reduction in the size of Parliament?
The 1999 referendum was conducted in an information vacuum. It's timing ensured that there could be no real public debate on the various explicit and implicit issues and agendas.
One group openly campaigned for fewer MPs as a back-door means of reversing the 1992 and 1993 electoral referenda.
I heard the referendum's instigator, Margaret Robinson, say in a radio interview that she was promoting this issue because she wanted a smaller Cabinet and she wanted politicians to be better behaved. Neither of these issues was addressed in the referendum question, but a reading of the letters to the editor in 1999 clearly indicated that the number of MPs was not actually the central issue. Rather, it was an attempt to censure politicians, whom the public see as being too easily seduced by the expense account lifestyle and as indulging in too much destructive bickering.
It is not clear from the limited public discussion whether the intent of the referendum question was to bring a return to the pre-1996 system of determining the size of Parliament, or whether 99 really was intended as some kind of magic number for a permanent size for our parliament. The principle, adopted in the 1960s, of maintaining the South Island quota would be seriously threatened by the adoption of a permanent 99-seat parliament.
Also largely absent from the limited discussion was the awareness that, before 1996, the size of Parliament had been growing by about 7% per decade, while the MMP Parliament was fixed at 120 MPs. Thus it can be argued that, given we are making decisions for at least the whole of the 21st century, the present MMP system is a recipe for a small parliament. Under the old system, we could expect to have more than 200 MPs by the year 2100.
We could have asked any of the following, and been given an equally (or more) emphatic vote as we did in the 99-MP referendum:
We would get answers: yes to [a], no to [b], no to [c], no to [d]) that cannot all be satisfied. The referendum last year is no more worthy of a literal implementation than are these alternative referenda.
Fine-tuning options for MMP
While plebiscites on the electoral system itself should be separated by generations, an electoral system must be amenable to fine-tuning. Fine-tuning is a serious constitutional matter however. The bias should be in favour of the maxim: "if it ain't broke, don't fix it".
Fine-tuning may be done by referendum or by a substantial majority (I think 75%) of MPs. Referenda are a waste of time if the issues have not been subject to robust public debate.
Our present MMP legislation retains one of the major faults of the unrepresentative FPP electoral system. Most electorate MPs "win" with less than 50% of the electorate vote. Their success or otherwise is due to the distribution of votes for other candidates. That lack of majorities in most electorates is of course how, prior to the introduction of a correcting party list system, "majority" governments used to be elected with less than 40% of votes cast.
The belief that a person can win an election despite being outvoted is ingrained in New Zealand. Indeed, media representatives persist in calling a winning candidate's margin a "majority" when it is no such thing.
The simple solution is to adopt preferential voting (PV) for the electoral vote. The most immediate impact of that would be that there would be no need for major parties to stand-down candidates in order to prevent a split-vote giving a seat to "the enemy". (In 1999, we saw a double stand-down in Wellington Central.) Single-member electorate non-transferable voting (FPP) only makes sense if there are just two candidates per electorate.
The French system, which New Zealand adopted in 1908 and 1911, is an alternative to preferential voting. It allows for a second ballot if no candidate gains a majority in the first ballot. But the combination of a two-ballot electorate system with a one-ballot party vote would create frustration all around.
In Australia, where the preferential system is used, coalition partners stand candidates against each other with no problems.
Preferential voting within MMP has additional advantages. PV is a transferable voting system like the multi-member electorate "Single Transferable Voting" (STV) system. STV is the only viable alternative for MMP, given the public's distaste for closed shop politics. A combination of PV in electorates and STV in local body elections will enable a referendum on MMP versus STV in or after the 2020s to be conducted on the basis that both systems are understood.
The Coromandel electorate has revealed that a de facto form of preferential voting already exists. Many Labour voters voted for their second-choice candidate, the present Green MP Jeanette Fitzsimons.
Parties qualify for list seats if the total party vote is 5 percent, or if the party wins an electorate.
There has been much disquiet about the latter rule. It appears to many that tactical voting is more widespread under MMP than under FPP. Further, parties like Labour and the Alliance are seen to be doing deals, whereas under FPP they were at each others' throats, conceding power to National in 1993.
The public is ambivalent. They clearly prefer parties to cooperate in the House, yet feel uncomfortable with cooperation on the hustings.
The "electorate seat" rule is misunderstood. It is not a backdoor way of getting 5 extra MPs into Parliament on the basis of the vote for just one. Rather, it's a means by which established parties can avoid the arbitrariness of the "five percent" rule.
We know that electors whose expectations are often unrealistic or inconsistent can unfairly punish small parties in government. The "electorate MP" rule enables minor coalition partners to remain in parliament on a proportional basis despite a fall in support to below 5%. This rule eliminates the need for nationwide tactical voting of the kind that bolstered the Free Democrats in Germany for so long.
The "electorate MP" rule should be retained and the 5% threshold reduced to 4%. That will make Parliament more like a marketplace and less of the political oligopoly that it is today. While entry and survival would become easier, entry into this market would still be far from easy.
Unlike FPP, SM, STV and PV, electoral boundaries have no party political significance under MMP.
Hence we have an opportunity to use these boundaries in a new way. It has been argued by people in the South Island, and by the MP for Gisborne, that MMP leads to an under-representation of their regions. We have an opportunity to correct that by making rural electorates smaller and urban electorates bigger.
I would like to see electoral boundaries permanently match established provincial and city boundaries. The South Island should have at least 25% of all the electorate MPs.
The total number of electorates should not be allowed to exceed 75, which would allocate 19 to the South Island. If the Auckland and Wellington urban electorates become quite large, population-wise, that will be offset by the inevitability that those two centres are over-represented in party lists.
If we come to have fewer than 45 list MPs, it is likely that overhang MPs will become a frequent occurrence. That issue will become politically sensitive, especially if a party with overhang MPs appears to have the "balance of power".
We should look to adopting the system common in Europe of having fixed term Parliaments. Elements of instability, although exaggerated, did exist in the 1996-99 parliament. The main reason was the tactic used by some of the political parties, especially when the coalition folded in 1998, to try and force an early election rather than to try to form a new coalition government.
If we had had a fixed term, then Labour would have "bitten the bullet" and formed a coalition with either New Zealand First or National. Further, Labour would have maintained cordial relations with New Zealand First following the formation of the National-NZF coalition. The negative parliamentary behaviour that upsets many voters would have been minimised if the environment was one in which each party had an incentive to maintain a constructive working relationship with each other party.
To use rugby parlance, the option of an early election leads some parties to try to "collapse the scrum" rather than to "play the game".
We do not need to replace MPs at all. Parliament can operate with 119 or even 115 MPs. A list MP can be assigned to an electorate without an MP, if an electorate MP dies, resigns, or (as Whangarei's John Banks did) goes walkabout.
At present, the parties are a bit like Super-12 rugby teams, with the leader as team manager. At present there is a de facto substitution rule. Helen Clark clearly wants to replace Dover Samuels.
I believe that if a party loses an MP, then it should simply lose one vote. If it is a party of government, that may affect the government's ability to gain confidence. But then bye-elections are usually opportunities to replace a government MP. The non-replacement of an electorate MP could upset the proportionality of Parliament less than having a bye-election.
Parties should only choose candidates who are committed to at least three years service as MPs. Don McKinnon should not have been on National's list, and should not have been replaced.
Another problem with replacing list MPs occurred in 1999, when Deborah Morris resigned. When a party fractures, as New Zealand First did, it is not clear why a resignation from the B faction should be replaced by someone who is obliged to vote with the A faction. The government's margin of confidence should never fall by more than 1, when an MP leaves or dies.
There should be a simple no-replacement rule, for electorate and list MPs.
Looking towards the long-run, I would like to see the whole concept of nationwide general elections replaced by a 4-year cycle of annual regional elections. Each MP would be elected for 4 years.
There would be a South Island election every 4 years. North Island elections would basically follow the rugby Super-12 boundaries, although Manukau, Papakura and Franklin would belong to the northern region, and Taranaki might belong to the central North Island region. Each of these "Super-12" regions would hold full MMP elections every four years. Thus there would be a regional election every year.
With every year an election year, the election-year economic cycle would come to an end. While there would be a greater sense of continuity and stability in government, governments would be more frequently accountable for any failure to represent the interests of their constituents.
With a fully regional form of MMP, there would be no need for a party threshold. With no more than 35 MPs being elected in any one election, there would be a built-in threshold of around 3%.
Maori electorates would be treated as a 5th region. Maori elections could coincide with regional elections in the central North Island region. In the Maori region there would be at least 10 seats, including 4 list seats, up for grabs. A Maori party with 20% of support of electors on the Maori role would gain entry into Parliament.
Maori would also gain representation through the regional party lists, and the general electorates.
If we do have the referendum that we should not have...
The constitutional system becomes a bad joke if a twice-rejected electoral system is restored within a decade.
Constitutional change is too important to rush. There should be no referendum before 2004. A referendum in 2004 should be structured just like the 1992 exercise. Part (a) would be a question "Do you wish to retain the MMP system?" Part (b) would be a head-head runoff between STV and SM.
Part (b) should have only two choices, ensuring that there is no split vote. For reasons already discussed, FPP should not be a choice. But if it is there, then it should be there instead of SM, which is a version of FPP. It would be a travesty if FPP came through the middle with a 34% vote in a referendum itself conducted as a three-way FPP poll.
Such a referendum should not be a postal vote. There are two reasons for this.
First, many postal votes are cast well before the official election date. This means many votes cast will be ill-informed by the campaign debate. Such voters will not be able to change their minds as the campaign unfolds.
Second, the turnout in a referendum of this type is as important as the result. People who choose not to participate in the exercise should not be given the opportunity to make a throw-away postal vote.
A preliminary referendum in 2004 should not proceed to a final referendum in 2005 if:
If, on account of an early general election in 2003 or 2004, then the referendums should be postponed rather than brought forward.
There should be no referendum on MMP until at least the 2020s. There should be no referendum that includes the twice-rejected FPP system until at least 2050 (2 generations).
There should be a "fine-tuning" exercise in 2005. Either a ratifying referendum or a 75% vote in Parliament. The most important issues that it should address are whether to adopt preferential voting for electorates, to relax the present rules on electoral boundaries, to fix the term of Parliament (3 years so long as we continue to have nationwide general elections), and to disallow replacement MPs.