The MMP Files: Under Alternative Voting Systems
The MMP Files : 17 September 2005 under Alternative Voting Systems.
Keith Rankin, 26 September 2005
Our MMP voting system has come under much recent attack for doing exactly what it was designed to do; namely delivering a Parliament with representation by parties in proportion to votes cast.
Critics have yet to state which parties they believe should have had their votes discounted. Much of the "letters-to-the-editor" criticism of MMP appears to be of the "I want my preferred party to rule regardless of how many votes it actually got" ilk. The belief is that an election is like an FA Cup Final, that there are in reality just two contestants, that there must always be an outright winner (the "dog"), and that it doesn't really matter if the winner scores less than 40% of the vote, or, for that matter, fewer votes than the loser. For people who see elections in this way, the toss of a coin between the two possible dogs (a coin with no tails, lest the dogs be wagged) would be superior to MMP. (Indeed MMP is more like an old-fashioned cricket test match, in which the usual result is a draw.)
What would have happened on September 17 if the election had been held under any of three other systems: FPP ("nearest the post"); SM ("supplementary member"); 1-2-3 ("preferential MMP")? [STV – "single-transferable-vote" or "multi-member-preferential" – requires wards rather than electorates, so no STV result can be inferred from available data. See my www.scoop.co.nz/stories/HL0509/S00158.htm and www.keithrankin.co.nz/kra05VotingMethodsSubmission.doc for more on 1-2-3.]
Under FPP you only have one vote, an electorate vote. But it is a party vote, not the personal vote that it is today. Thus, we must take the party votes for each electorate to determine the FPP outcome. Doing that, we get a parliament of 69 MPs: 36 National, 33 Labour. Pretty much a repeat of 1981; National would have got enough seats to form a majority single party government, despite getting both less than 40% of the vote, and getting less than the other dog (Labour).
Probably we should give a seat to Peter Dunne, another to Jim Anderton, and a third to Tariana Turia. That would leave a result of 35 National; 34 all others. Still a majority of seats to National. Maybe the Maori Party would have won other seats off Labour. More likely, if we had stuck with FPP in 1993, the Maori party (or one like it) would have suffered the same fate as Mana Motuhake did in its first election. However, if the 2008 election were to be held under FPP, the Maori Party would almost certainly retain its four seats. Despite our claimed aversion to politicians, we definitely favour sitting electorate MPs over their less familiar rivals.
FPP is biased in favour of National. While the median (ie middle) voter almost never votes National, the median electorate typically votes National, especially in a close election. 2005 is no exception.
The middle electorate votes National because: (i) National voters are more evenly spread throughout the country than are Labour voters, and (ii) because the Maori electorates creamed sufficient Maori Labour voters from provincial city electorates to make them National in any close election. In other words, Labour voters were clustered in seats like Manurewa, Christchurch East and Southern Maori. Under FPP, the average Labour "majority" (ie winning margin) was almost always larger than the average National majority.
SM is a system which, with few compensating benefits, combines all the faults of FPP with the alleged faults of MMP. [SM is favoured by Don Brash, and was also advocated by Jenny Shipley; see www.listener.co.nz/default,3265.sm (John Armstrong) and www.greens.org.nz/searchdocs/PR3577.html (Rod Donald).] While it gives the National and Labour "dogs" a significant over-representation, it also gives sufficient token seats to the smaller "tail" parties. In a close contest between the dogs, the not-quite-amputated tails still wag. (We should also notice that, in early 1984, anyone in National's caucus who crossed Muldoon was a tail wagging the Muldoon dog: Marilyn Waring, Mike Minogue and Derek Quigley brought Christmas early for the Muldoon turkey. Also Roger Douglas, with his alternative budget in 1981, was very much a tail that successfully wagged the Rowling dog.)
Under SM, the FPP electorate vote and the supplementary party vote are, in reality, both party votes. (Many people still think that's true of MMP. A poll in Epsom got it badly wrong by asking voters which "party" they would give their electorate vote to!?)
For a 120-member Parliament under SM, the 2005 election would have delivered – on election night – 54 to National, 49 to Labour, 5 to the Maori Party, 3 to NZ First, 3 to Green, 2 to United-Future, 2 to Act, and 2 to the Progressives. (Maybe Rodney Hide would not have won Epsom.) Either way, the result would have been a more intriguing deadlock than the actual MMP result, with National getting five more seats than Labour despite getting fewer votes. The "letters to the editor" would have been full of the stupidity of a system that clearly favoured National and the Maori Party over all others.
Finally lets consider my preferred 123-preferential version of MMP. A preferential version means that a vote need never be wasted, because it transfers to the 2nd or 3rd choice if a voter's first choice of party or candidate is eliminated from contention. 123-preferential (1-2-3) makes the contest into a true match-race between the right-leaning parties and the left-leaning parties. If New Zealand First is classed as "centre-right", then the popular vote on the night favoured the right.
Under 1-2-3, Act supporters would have been free to vote for Act, knowing that if Act did not get 5% (or if Rodney Hide did not win Epsom) then they would still be casting a vote for the right. Most Act voters would have chosen National or Libertarianz as their 2nd choice. Those choosing Libertarianz would have chosen National as their 3rd choice.
Under 1-2-3, there would have been no side-show in Epsom. Labour voters would have made National their 2nd or 3rd choice, given that it was a 3-way contest in which left-voters liked Hide the least. Winston Peters would have won Tauranga with the transfer of Labour votes.
Under 1-2-3, the centre-right (including New Zealand First) would have just prevailed over the centre-left (62:60). Votes for the likes of Destiny and Christian Heritage would have counted for the right. Preferences from the Legalise Cannabis Party would have helped the Greens, getting Nandor Tanczos back into Parliament.
Under 1-2-3, there would have still been a Maori Party overhang. A likely result would have been: Progressive 1, United-Future 3, Maori 4, Act 6, Green 7, NZ First 7, National 46, Labour 48.
Any debate about the merits and demerits of different voting systems needs to take place in a comparative context. Critics of MMP should walk-the-walk, as I have done, by demonstrating how, by some criteria, their preferred alternative to MMP might have yielded results as good as or better than the actual MMP outcome.