by Keith Rankin
12 September 2005
One of the most important effects of New Zealand's form of proportional representation (MMP) is the significance of the 5% threshold for the "minor parties". A vote of 4.99% for a party means a very large number of wasted votes; of votes that do not count.
As a result of the 5% electoral cliff, the most important unknown at election time may be whether Part A or Party B will fall off the cliff-edge. Commonly, such unknowns must be resolved before we know whether we will have a government of the right or of the left.
As a consequence of that, the presence of an electorate MP has become significant as a way of avoiding the distractions that arise from the 5% rule. However, that usually creates a further set of distractions. Maybe parties which have made it into Parliament and which have therefore played a constructive role in the governance of the country should be exempted the threshold. Indeed, that's the exact purpose of the "electorate MP" rule overriding the 5% threshold.
Do-It-Yourself Preferential Voting.
The "electorate MP" distraction is rooted in our retention of non-preferential "first-past-the-post" voting. It means that voters have to do what they often did in pre-MMP days; that is, adopt do-it-yourself (DIY) preferential voting.
Preferential voting means ranking at least 2 candidates (or parties) in order of preference. If a voter's first choice is eliminated then it's the voter's second preference that counts. Voters can therefore afford to vote for a candidate who stands a high chance of being eliminated from the contest.
In DIY preferential voting, voters use the polls to judge whether their most favoured candidate will be either a "contestant" or an "also-ran". (Each electorate has two contestants; the two "finalists", to use a sporting analogy. The remaining candidates are also-rans.) If voters conclude that their preferred candidate will be an also-ran (ie a non-contestant), then they (tactically) vote for their preferred contestant (who most likely will be their second or third choice candidate).
DIY preferential voting is inefficient, however, in that not all voters are willing or able to play the game. As a result, in pre-MMP days most MPs were "elected" with less than 50% of the votes cast. (We even had the gall to call the margin between the contestants the "majority", even when a candidate had no majority. A majority – by definition – is more than 50%.)
Of course voting should not be a game. Voters should be able to feel that if they vote for the party they like the most – and for the person they like the most – then the election outcome will faithfully reflect that preference.
In the absence of true preferential voting, we find that the parties themselves take some steps to help voters get preferential outcomes in their electorates by pointing out who the contestants are likely to be. In some cases, parties withdraw (or do not stand) candidates that they know will be also-rans. These withdrawals are not "deals" in any negative sense of the word. They are simply ways of helping voters to distinguish the contestants from the also-rans. Withdrawals help us to know who is the preferred contestant.
(The name for our electorate voting system – first-past-the-post – is actually a misnomer. The 50% "post" is often achieved by none of the candidates. If it is achieved, only one can candidate can be "past the post". If it is not achieved, the candidate closest to the post when all votes are counted is declared the winner. Our electorate and local authority system is really NP: "nearest-to-the-post". Preferential voting – as in Australia – is true first-past-the-post voting; the winner must get past the 50% post. STV is also first-past-the post, except that, if say 5 candidates are to be elected, then it is "first-five-past-the-post".)
Dealing with the 5% Threshold Distraction.
It is not easy for "major" parties (defined as parties not at risk of falling below 5%) to prevent votes for supporting parties from being wasted. Their best bet is to not stand electorate candidates against contestant candidates from minor parties. Then each electorate would become a contest of the strongest candidate from the left-wing bloc against the strongest candidate from the right-wing bloc. This can be tricky re centre parties such as New Zealand First and United-Future. Both major parties would probably be best not to stand electorate candidates against Winston Peters and Peter Dunne. Rather, they should place strong local candidates at electable positions on their party lists.
National's foolish decision to contest Epsom against Act could cost the Right bloc this election. Similarly, Labour's decision to contest Coromandel could cost the Left bloc the 2005 election.
The problem is that not standing electorate candidates is seen as "doing a deal", and "doing deals" has pejorative connotations in New Zealand. However, it's obvious to most voters that when friends fight each other, their mutual enemy usually wins. In practice we make common-sense accommodations throughout our lives.
A Better Solution.
There is a better way to solve the problem of wasted votes of parties that do not reach the 5% threshold. It is to formally adopt preferential voting for both electorate candidates and party lists.
In my "Submission to Select Committee Inquiry into the 2004 Local Authority Elections" (http://www.keithrankin.co.nz/kra05VotingMethodsSubmission.doc) I advocated a form of preferential voting called "123" or "trifecta" voting. (A trifecta is a popular racing bet, in which punters are rewarded if they choose the first three horses in the correct order.)
In 123 voting, voters would be asked to vote for up to three candidates or parties, in order of preference. In a party vote, in most cases only the first preference would count. In the event however of a party not qualifying (because it gets less than 5% of the total vote or because it does not get an electorate MP), then the vote will count for the most favoured qualifying party. Thus the vote would only be wasted if none of the 3 preferred parties qualify to be represented in Parliament. All voters who really want to participate in the election will manage to choose one qualifying party. Voters who distrust major parties would typically choose one only as their third preference.
For electorates (or mayoral elections), the moment a candidate is declared to be an also-ran, then the second preference votes are counted. Where second-preferred candidates are also also-rans, then third preferences are counted.
For multi-member contests – eg city councils – preferential voting is known as STV (single-transferable voting). As already noted, STV is simply first-past-the-post with multiple winners. (The post is the "quota".) 123-STV is simple to count because it restricts voters to three preferences, no matter how many names appear on the ballot paper or how many candidates will be elected. It achieves essentially the same result as more complex variations of STV.
Strictly-speaking, STV should be called "multi-member-preferential". As that stands for MMP, then it cannot be called that. The important thing to know is that it is a preferential system in which winners do get past the post, and that tactical voting serves no purpose under a preferential system.
Strategic versus Tactical Voting.
Tactical voting is the avoidance of casting a wasted vote. Strategic voters, however, influence the make-up of their chosen government. There will always be opportunities for strategic voting in any multi-party electoral system.
A typical scenario in 2005 might be that a voter wants to secure National's promised tax cuts, but does not want the Maori electorates abolished. Such a voter will want National (or National/Act) to govern as a minority government. If we had 123 preferential voting this year, a good strategic option would be United-Future 1, NZ First 2, National 3. If Peter Dunne wins Ohariu-Belmont, then the party vote counts for United-Future. Otherwise it would count for NZ First or National.
In the absence of 123 voting, such strategic voters would have to choose United-Future. So long as Peter Dunne is assured of winning Ohariu-Belmont, then there is little risk that it will be a wasted vote.
A strategic scenario for right-wing voters might be to choose Act to inject a bit of philosophy into an otherwise pragmatic National government. That choice would be straight-forward under 123 voting. It would be risky though under present voting arrangements. At present the correct strategy for such a voter is conditional on the likelihood that Rodney Hide wins Epsom.
Under 123-MMP, a typical Parliament would contain (at least) a left, centre-left, centre, centre-right and right party. In some elections, one of those minor parties would miss out and have its votes reallocated. But, come the next election, the party that missed out last time would have a good chance of making it back in. Contrast that with the position the Alliance faces today. Laila Harre's inability to win Waitakere in 2002 consigned the Alliance to electoral oblivion.
This time it looks like Act's turn to face oblivion, especially if Labour voters tactically vote for the National candidate in Epsom. (The Epsom contest is unambiguously between Act's Hide and National's Worth.) I – as a voter who has more often than not voted left of centre – think that the loss of Act would be a tragedy. If Act goes, the far-right will re-infiltrate National. We are more likely to get extremist right policies in the future if Act goes than if Act stays. If I lived in Epsom, I would certainly vote Hide, but not Act.
In pre-MMP days we had a non-first-past-the-post electorate voting system we called FPP in which we nominally voted for people but really voted for political parties. (Indeed, the last time a new-entrant MP was elected into Parliament without the affiliation of a party was a life-time ago; 1935 I think.)
MMP ("mixed-member-proportional") explicitly brought about two changes: (i) a direct party vote; and (ii) a genuine personal – ie non-party – vote in electorates. Under MMP it's easier for popular candidates who get off-side with their parties to stay in Parliament.
MMP can be improved through the adoption of preferential voting. This would ensure that there are no wasted votes; that no-one will feel obliged to not vote for their preferred candidate or party.
The best form of preferential voting for all elections – party, electorate and local authority (single-member and multi-member) – is 123-trifecta voting. The great irony is that, if we adopt preferential voting, we will be moving, for the first time, to an actual first-past-the-post voting system.
Today's critics of MMP need to: (i) appreciate that the problems they cite are mainly due to our retention of non-preferential voting; and (ii) explain how their preferred alternative (eg the former NP "nearest-to-the-post system", or the modified NP "supplementary member" SM system) solves the problems they are concerned about. It is meaningless to criticise one electoral system without offering a better one.