Keith Rankin's Thursday Column
27 July 2000
Submissions for the MMP Review are due in by Monday. The really important issue is the first part of part (c) of the
terms of reference: "whether there should be a further referendum on changes to the electoral system".
For me, the answer is 'no'. Referenda generally fail to achieve democratic outcomes. With reference to the MMP review,
constitutional changes of the magnitude of changing an electoral system should never be considered more than once in a
generation. An electoral system cannot be properly assessed while it's in transition. An electoral system remains in
transition while the culture associated with the previous system survives through those politicians and journalists
whose political education took place under that system.
The constitutional framework is the bedrock of a society. A society's legal foundations should never be regarded as
impermanent. (They are of course impermanent, but should not be regarded as such.) It should not be an option to change
one's constitutional arrangements as a response to getting a government that some people - especially media people or
people with access to the media - show cynicism towards. Cynicism towards what we have only begets objects deserving of
cynicism; it's self-fulfilling.
No broad constitutional provision introduced by binding referendum should be subject to rejection for a generation
(25-30 years). If we are to consider replacing MMP by say STV (single transferable voting), the referendum should happen
around the year 2020. A decade's experience of STV in local body elections would minimise the chance of a far-reaching
decision being made by an uninformed public.
Any provision rejected by binding referendum, as the "first-past-the-post" (FPP) voting system was, should never be
placed before the people again before two generations have passed (50-60 years).
There are always problems in the present - in today's present, in yesterday's present, and in tomorrow's present - that
will be or will have been blamed on some aspect of our constitutional arrangements. The MMP electoral system remains the
easiest constitutional target at present, because it is new. On the other hand, it's easy to become nostalgic towards
our past. Robert, Roger and Ruth don't seem quite as menacing when they are no longer daily presences on our TV screens.
It takes two generations to eliminate the nostalgia factor.
FPP was rejected by referendum twice in the 1990s; an indicative referendum in 1992 and a binding referendum in 1993.
That system should never be again considered until it is able to be presented as something quite new. I am sure that FPP
would be nonsensical to a population in 2050 that had become thoroughly accustomed to proportional representation.
Both the 1992 and 1993 referendums were, in essence, about FPP. They were not about the exact mechanics of MMP. What
mattered is that MMP gave everyone's vote equal status and that the resulting Parliament would reflect that equality of
suffrage. FPP was rejected for deep and fundamental reasons. Whether or not we saw the 1993 Electoral Act as cast in
concrete, we rejected FPP permanently. As the education campaign progressed, most of us came to see that FPP was
In 1992, only 200,000 New Zealanders could be bothered voting for FPP. The higher vote for FPP in 1993 did not represent
a change of mind from 1992 to 1993. Rather, FPP was supported by two groups of people: those who valued decisiveness
ahead of democracy, and those who feared change. The growth in support for FPP from 1992 to 1993 was a "better the devil
that you know than the devil you don't know" kind of vote. Many of these negative votes for FPP in 1993 came from people
who would not have voted at all had there not been a concurrent general election. Yet, despite the conservatism that
normally prevails in this kind of plebiscite, several hundred thousand more people voted for MMP than voted for the
1993-96 FPP Government.
Fine-tuning an electoral system is of course quite different to replacing an electoral system. Binding referenda once or
twice a decade do have a place, but only if the public is informed though processes of public education and public
Referenda should only be on constitutional matters, and 80-90% of such referenda should be of the "fine-tuning" variety.
All referenda should take the two-ballot form that we experienced in 1992 and 1993. There should be no short cuts.
(The experiment with Citizens Initiated Referenda has proved a disaster. The results of the two 1999 referendums were
exactly as any observer could have predicted, and have told us precisely nothing that can be converted into meaningful
policy. There was little public debate about either issue. Many people did not know about either referendum until they
were handed their ballot papers. Referenda such as these, conducted on the cheap, have the potential to create very
costly policy mistakes. Leading questions - especially the "motherhood and apple pie" variety - give predictable
answers. 1000 other leading questions would have given equally predictable answers. Some answers would contradict
The process of constitutional change should start with an official inquiry. The next step would be an indicative
referendum that is accompanied by an educative process. Voters must become exposed to the issues before they are able to
cast their votes. Therefore such a referendum should NOT be a postal referendum. In postal referenda, many of us vote
well before the official polling date, meaning that much of the educating takes place after rather than before votes are
Further, a good indicative referendum requires us to make at least a small effort to go out and vote. It is the turnout
figures that tell us about the depth of our preference. The size of the turnout is as important as the vote count for
the winning option. As a suggested rule of thumb, if the turnout is more than 40% and more than 60% of votes cast reject
the status quo option, then the second referendum should be proceeded with.
The second referendum needs to be a binding referendum held in conjunction with a general election. This time around,
although the constitutional debate becomes diluted by the general election campaign and the multiplicity of issues that
surround that, a general election turnout is necessary to ratify a vote for constitutional change. It's because of the
debate surrounding the first referendum that we can be sure that the voting public is adequately informed when they make
the final decision.
The process for changing the electoral system that we followed, from 1986 to 1993, was a model process. We should accept
nothing less in the future.
Short-cutting constitutional change is a recipe for disaster.
Consider the indicative referendum about the number of MPs. We knew in December 1999, as we did in October 1999, that
people do not trust "politicians" (in the collective form). The election however actually told us that, in most
electorates, we liked our own individual MPs, many of who were reelected with astonishing margins. Only a handful of MPs
were actually rejected by the public last November. Of those, many were former New Zealand First MPs who faced
conflicting loyalties while in government. In enabling the parliament to run for three years, those MPs were fulfilling
the wishes of the majority of New Zealanders who, when polled, indicated that they did not want an early election.
Voters, who cannot be expected to think through all their opinions to their logical conclusions, frequently hold
contradictory opinions simultaneously.
(Indeed, in the MMP referenda of 1992 and 1993, it appears that, while just about everyone understood that MMP was
proportional, a large number of MMP supporters confused the mechanics of MMP with the mechanics of one of the other
options, the "Supplementary Member" [SM] system. The binding 1993 vote was valid, however, because it was ultimately
about outcomes - ie proportional representation - and not about the mechanics of achieving that outcome.)
Referenda can do two things. First, they can resolve a constitutional issue. But only when the process is correct and a
substantial public debate takes place before voting. Second, they can convey public sentiment. In this second case,
sample polling tells us as much about sentiment as a referendum can, at a fraction of the cost.
Direct democracy - ie rule by national referenda or by MPs taking instructions from their constituents through local
referenda - doesn't work. Most citizens abstain from the process because they can never be adequately informed on the
issues. By way of contrast, in a representative democracy, MPs represent their constituents' interests rather than their
constituents' prejudices. MPs are paid to be informed. All MPs have equal access to information. In a representative
democracy, my interests are as well represented as are your interests, whether or not you are better informed than me,
or I am better informed than you. Unlike society, the House of Representatives is a level playing field.
We vote for political parties and for local candidates to represent our interests. The dual vote is crucial. Under FPP
we voted for candidates we didn't like because they stood for the party that we disliked the least. We now have two
positive votes rather than a single negative vote. The party vote is much more important than the local candidate vote,
because each electorate holds a diversity of interests, and therefore cannot be accurately represented by a single MP.
Accurate representation arises from political parties gaining seats in proportion to their voter support. Geographically
diverse representation is an important though secondary consideration.
Good democratic decisions are made by informed members of a representative parliament. If scientific research shows
that, for example, we are safer if prisoners are civilised rather than degraded, then informed MPs will go for that
option, even if the public incorrectly believes that "getting tough" will better achieve public safety. If, through
California-style referenda, sound scientific decisions are overturned by the uninformed preferences of citizens, then we
will sometimes end up with the exact opposite of what we really wanted. To take another example, we don't need to cull
our Parliament in order to make politicians more trustworthy and better behaved. An MP cull may actually make MPs on
average less trustworthy and more rowdy.
Democracy is not about having MPs vote according to constituents' instructions. Democracy is about having our interests
proportionately and honestly represented in that arena we call Parliament.
We feel distrust for politicians in the main because they openly disagree. If only, we think, the politicians who
represent others' interests would agree with the ones who represent our interests. In our own lives, we get on by
suppressing open conflict. Despite this phobia, we like to prevail; to get our own way.
Politicians disagree because they represent different interests. Their disagreements are our disagreements. We could
always suppress political disagreement by making sure that the majority of our politicians at any one time represent a
single interest, and that all other interests are impotent at that time. Yes, we could succumb to nostalgia, and
reintroduce FPP. Under FPP the three Rs - Robert, Roger, and Ruth - did not need to argue their cases scientifically.
For a few years, each had the power to impose their own brands of witch-doctory. The only way to oppose them was to
introduce an electoral system that denied them the power to rule untrammelled.
It took many years of rational argument, passionate debate and consciousness raising to overturn the electoral system
that gave so much power to so few. We could lose it all if, through an inappropriate referendum in 2002 or 2005, we
allow relatively trivial concerns about MMP to override the monumental concerns that in the past led us to reject FPP
not once but twice.
© 2000 Keith Rankin