Exploring The 2011 Referendum on New Zealand's Voting System
Analysis - By Keith Rankin.
The referendum accompanying the 2011 general election should be understood as a decision between two broad choices: a 'multi-party
proportional representation' versus a 'two-party, winner bonus' system. MMP (mixed member proportional) and STV (single
transferable voting) are examples of multi-party proportional representation; FPP (first past the post), SM
(supplementary member) and PV (preferential voting) are examples of 'two-party, winner bonus' systems.
Voting should be very easy for supporters of multi-party proportional representation: one tick for MMP, and the other
tick for STV. Supporters of two-party winner bonus systems will, on the other hand, vote against MMP and choose between
FPP, SM or PV.
Because the choice of alternative systems will be carried out using FPP voting, the winning post for the alternative
system vote will probably be 40%, as it typically is most electorate and mayoral elections carried out under the FPP
system. Thus, even if only 40% of people vote for MMP in the first question, if they all also vote STV in the second
question, then the 2014 referendum will be a simple run-off between MMP and STV.
Clearly the vote MMP campaign should feature a simple "two tick" message: a tick for MMP and a tick for STV.
I am a supporter of multi-party proportional representation, for four main reasons.
First, I believe that legislation should be passed only when supported by a combination of parties that together
represent more than 50% of voters. Further, political deals should be done in the open, between parties, rather than in
the back-rooms between factions of parties that govern without majority support. (It is very rare in the history of any
open democracy for a single party to gain more than 50% popular support. The United States has the appearance of
majority support in each of its houses only because it has such a tight two-party system.)
Second, the median voter should determine the outcome; not the median electorate, nor the boundaries commission. In New
Zealand in the two-party FPP eras (1890-1905; 1931-1993), in most cases the median voter did not vote for the
conservative option. Yet commonly – at least in the 1949-93 era, the median electorate returned a conservative MP. This
is why most National and Act activists favour two-party electorate-based systems over proportional systems. The
elections of 1954, 1966, 1969, 1978, 1981 and 1993 would almost certainly have yielded a centre-left government had a
proportional system been in place. There is probably no case of the reverse, with the possible if unlikely exception of
1957.
In the period from 1954 to 1993, the median voter never voted National, yet the median electorate did in every election
except 1957, 1972, 1984 and 1987. The reason for this is that the average winning margin ("margin", not "majority"!) for
National MPs was always lower than the average winning margin for Labour MPs. While this was usually a result of
National supporters being more evenly spread through the country than Labour supporters, it also created an opportunity
for the electoral boundaries to be drawn in such a way as to make one result or another more likely.
The worst case of boundary distortion was in 1978. In Wellington city, with four electorates, more voters supported
Labour, but three electorates were won by National. The Wellington boundaries that year were odd to say the least.
Similarly, in Gisborne and Hunua for example, large clusters of urban Labour voters were joined with slightly larger
numbers of rural or semi-rural National voters. Another egregious case was, in the 1980s. The Rangitikei boundary was
drawn so as to exclude Marton, the major town in the Rangitikei, with the predictable consequence of unseating Social
Credit MP Bruce Beetham.
My third reason for opposing two-party winner bonus systems is that they are all based on single-member electorates. (SM
has a token nationwide list, but is designed so that a substantial majority of the list MPs will be from the same two
parties that prevail in all – or almost all – of the electorates.) Such systems disenfranchise voters who live in
"safe-seat" electorates. In New Zealand's history of FPP, the elections were determined by voters in about 10 marginal
electorates. Most electorates had the same party represent them through the entire post-war FPP era.
STV is an electorate-based system, but it relies on multi-member electorates to achieve a near-proportional multi-party
result. PV is the single-member-electorate version of STV, which is a two-party winner bonus system. Australia uses PV
in its lower house, and STV in its senate. Tasmania and Ireland use STV in their lower houses.
My final reason for favouring either MMP and STV is that each gives voters the opportunity to make a personal vote. A
good example of the personal-vote effect in New Zealand under MMP is that of Nick Smith, in Nelson. He has been MP for
Nelson throughout the MMP era, though more often than not, Labour has won the party vote in Nelson. Both MMP and STV
give voters more opportunities to override the parties' choices of candidates. Nick Smith will almost certainly win
Nelson again, even if he's given a low place on National's party list. Under any of the two-party winner bonus systems,
Labour voters in Nelson will feel obliged to vote for the Labour candidate, despite the strong personal liking many of
them hold for Nick Smith. Nelson will be one of the few marginal electorates – under any of the two-party winner bonus
systems – that would determine which of National or Labour could govern.
We might note that, with smaller electorates, a part of Nelson would have to be included with a new Tasman electorate.
The result of the 2017 election, if conducted under a two-party winner bonus system, could depend on which part of
Nelson the Boundaries Commission chooses to exclude from the Nelson electorate.
A vote this year for MMP and STV should see off the threat from political conservatives to re-establish a two-party
winner bonus system. Such a system, such as the FPP system that was comprehensively rejected in the 1992 referendum,
enables conservative parties to rule for most of the time, even if most voters do not support such parties.
*************
Keith Rankin teaches at Unitec's Department of Accounting and Finance.