The Electoral Commission has to review six aspects of MMP (plus whatever else the public puts before it). Here's my thoughts on the first issue: the thresholds for representation in Parliament.

Of the issues that the Electoral Commission has to look at in its review of MMP, I predict the question of what threshold a party should have to attain before getting proportional representation in Parliament will be the most fraught. In part, that's because I suspect that even though most people think the current rules aren't working perfectly, they will disagree over what a better rule would look like. But it's also the issue that has the most potential to mess with the future electoral prospects of the parties presently in Parliament, hence they'll care a lot about what is recommended on this point ... and in the final analysis, any change to the law relies on them passing it. 

So obviously there's a need for a sensible, authoritative voice to speak out on this issue and put forward a cogent, coherent and convincing picture of what should be done. I guess that's why God invented David Farrar. But as you're here, you're going to have to put up with my half baked thoughts, which you can then pull apart in the comments thread like the vicious, mean-spirited, pitiless animals that you are. Not like those very polite, thoughtful and friendly readers over at Kiwipolitico. Could someone tell me how we can get an audience like that here at Pundit?

But I digress. Currently under MMP, parties can get proportional representation (i.e. a share of the list seats to bring their numbers in Parliament into alignment with their share of the party vote) in one of two ways. Either they must gain 5% or more of the Party Vote, or else have one of their candidates win an electorate seat. The following is a part primer/part reform suggestion on that state of affairs. 

Why have two different thresholds?

The short answer is, because West Germany did and the 1986 Royal Commission on the Electoral System just copied that country when recommending an MMP system for New Zealand. And as far as I know (based on not very thorough research, I admit) the reason why West Germany had an electorate threshold alongside a party vote threshold was to accommodate regional parties that may be strong in an area but not strong enough nationally to cross the party vote threshold. (West Germany is a federal system, where individual states have an identity somewhat separate from the centralised nation - hence there is reason why regional-based political movements exist there.)

So, West Germany had an alternative 3 electorate seat/5 percent party vote threshold in place. The Royal Commission then recommended that NZ adopt a 1 electorate seat/4 percent party vote threshold ... without really giving any reasons for why the electorate seat threshold should be carried over to here, and choosing 4% as its preferred balance point between allowing small parties into Parliament and preventing undue fragmentation of Parliament that might hamper the creation and operation of effective government.

Parliament then took those recommendations, kept the electorate seat threshold, bumped the party vote threshold up to 5% (why? Well ... why do you think MPs from parties already in Parliament might want to make it harder for other parties to get in there as well?), and passed the Electoral Bill 1993. Then, when the country voted in favour of MMP at the 1993 referendum, this Bill became law ... and we got MMP as we know it.

How have these thresholds worked in practice?

I think it's fair to say, not as expected. First up, the 5% of the party vote threshold has been tougher to crack than probably was anticipated (or hoped for) back in the early 1990s. In the six MMP-era elections, only 7 parties have got over 5% of the party vote at least once. Of these, only Labour and National have done so in all six elections (the Greens only ran as an independent party from 1999 - they too have managed (just) to stay above 5% since then). This difficulty largely can be traced to Labour and National's recent success in dominating the party vote: their combined share of that vote in the past 3 elections has been 80%, 79% and 74%, which leaves precious little for the smaller parties to divide between them. 

Consequently, the electorate seat threshold has become an important means through which small parties have been able to retain proportional representation in Parliament (hence its becoming known as the "electorate lifeboat" option). Four parties (or maybe six, if you want to count the Mana and Maori Parties here, but I wouldn't) have made use of it since 1996 - although three of those parties have on occasion returned only the single electorate MP due to a low party vote. While this outcome has meant that Parliament has remained more proportional than it otherwise would have, the way in which the electorate lifeboat has operated is not without its problems.

First of all, it's been a bit hit-and-miss in its maintainance of proportionality. Most notoriously, Rodney Hide's victory in Epsom in 2008 meant ACT's 3.65% of the party vote brought 4 other MPs into Parliament ... while Winston Peter's loss in Tauranga meant NZ First's 4.07% of the party vote was wasted. Equally, in 2002 Jim Anderton brought Matt Robson into Parliament with him by winning Wigram and getting 1.7% of the party vote. But in 1999, the Christian Heritage Party's 2.38% of the party vote was wasted - not to mention the Christian Coalition's wasted 4.33% in 1996. So at best, electorate lifeboats retain proportionality for only some lucky players. 

Second, the electorate lifeboat option only helps retain proportionality by (ironically) giving the voters in particular electorates a disproportionate say via their electorate vote. So, voters in places like Epsom, Ohariu, Wigram etc are able to "double dip", in that their electorate vote helps ensure one party remains in Parliament whilst their party vote helps elect list MPs for another party. And make no mistake, voters in these areas are conciously double dipping - take a look, for instance, at the split electorate/party votes in Epsom in 2005. In contrast, it doesn't really matter who voters in non-electorate lifeboat seats cast their electorate votes for - this won't alter the overall shape of Parliament one iota.

Third, smaller party reliance on the electorate lifeboat opens the way for inter-party deals over the outcome of the vote. Not that there's anything illegal about the two Johns' cup of tea in Newmarket (or Helen Clark indicating to Labour voters that Peter Dunne was a good candidate for Ohariu-Belmont). However, it seems fair to say that such spectacles have not gone down well with the voting public; indeed, they have been one of the main grounds for griping about MMP.

What to do about these thresholds? 

As may already be clear, I think the electorate seat threshold ("electorate lifeboat") ought to go. Whatever its original purpose was (and this isn't clear in the New Zealand context), it has turned into a de-facto means of evading a party vote threshold that has proven too demanding in practice. And it makes little sense to have such an evasion option available to some parties but not others, depending in large part upon whether other parties are willing to collude in the evasion. So I'd get rid of it.

However, that then brings us to the issue of the existing party vote threshold - which is where the rubber really hits the road. A few preliminaries on this, before I say what I think should be done.

First up, for me the only good reasons for having any party vote threshold at all must stem from claims about the functioning of the legislature or government. In other words, I don't put much stock in arguments that without a party vote threshold you might get "silly" or "extremist" people elected to Parliament. Using a threshold to save the voters from their own (allegedly) poor choices seems overly paternalistic to me.

That said, there still are a couple of claims that I think need taken seriously. The first is the risk pointed to by the Royal Commission back in 1986 - without a threshold, Parliament could become so fragmented amongst small political parties that the difficulties of putting together (and maintaining) a governing coalition will impair effective government. And effective government is a good thing that we ought to seek to preserve.

The second is a bug-bear of my own. MMP is designed to enable parties to place representatives in Parliament to engage in lawmaking/governing. But to be able to effectively engage in lawmaking/governing, you have to be a part of a team of like-minded people and not a lone gunman. Of course, a single MP can act as an occasional flag-waver for his supporters on the floor of the house (Hone Harawira), or may even be able to trade his vote into a measure of executive power (Peter Dunne/John Banks). But their parties are, far all intents and purposes, non contributors to the day-to-day functioning of Parliament.

Now, there is no objective metric for deciding how important these matters are vis a vis the issue of proportionality (which is, after all, what MMP is meant to deliver). And the relative importance of each issue also depends on what you think is likely to happen if the rules on the party vote threshold are changed. Will parliamentary representation really fragment to a significant degree once voters realise that a vote for (say) the Aotearoa Legalise Cannabis Party likely won't be wasted? Or has politics in New Zealand now so stabilised that we've pretty much got the only parties we're likely to see?

Having said all that, my own view (at this stage, subject to revision) is that a threshold of 2.5% is warranted. I think 5% has proven too high a threshold in practice, and that it inhibits the emergence and establishment of new political movements. However, I think that (in the interests of creating a legislature that consists of actual parties (in the sense of teams of MPs) and is likely to continue to provide a stable basis for government) requiring parties to achieve a level of party vote support that will return at least 3 MPs to Parliament is still a good thing to do. So that's what I'm minded to submit to the Electoral Commission.

Before throwing it over to your comments, one last point. Any form of representation of the voters is going to fail to fully include all groups (much less individuals) vying for a share of parliamentary representation. That's because "representation" is inherently a compromise (some would say compromised) concept. So, for example, even adopting a no threshold rule for MMP wouldn't really result in a "no threshold" system - it just sets the level at which a particular group of voters get represented at about 1/120th of the party vote (actually, whatever the Sainte Lague formula dictates, but that's too hard for my head). But why 1/120th of the party vote ... why not 1/150th (by increasing Parliament to 150 MPs)? Or why not 1/2000th (in a 2000 seat Parliament)? What is it about 1/120th of the party vote that just happens to be the right level at which a group of voters ought to be entitled to a representative in Parliament? 

Indeed, if we follow the matter to its logical conclusion, the argument that proportional representation ought to be the overriding imperiative leads us down the path of Borges' cartographers: 

In that Empire, the Art of Cartography attained such Perfection that the map of a single Province occupied the entirety of a City, and the map of the Empire, the entirety of a Province. In time, those Unconscionable Maps no longer satisfied, and the Cartographers Guilds struck a Map of the Empire whose size was that of the Empire, and which coincided point for point with it.

It is worth, then, remembering the fate suffered by this monument to a mania:

The following Generations, who were not so fond of the Study of Cartography as their Forebears had been, saw that that vast map was Useless, and not without some Pitilessness was it, that they delivered it up to the Inclemencies of Sun and Winters. In the Deserts of the West, still today, there are Tattered Ruins of that Map, inhabited by Animals and Beggars; in all the Land there is no other Relic of the Disciplines of Geography.

Comments (24)

by Chris de Lisle on February 29, 2012
Chris de Lisle

Well, you could always do away with seats and make a party's vote in the house equivalent to the vote they received in the general election- so a party that got 13.674% of the party vote would get 13.674% of the vote on any given motion in Parliament. But I suppose, even then you'd have to round it off at some number of decimal places... So I take your point. 

A vaguely related question; if proportionality to the views of the public is so important, why do we only seem to care about it at all on election day?

by Andrew Geddis on February 29, 2012
Andrew Geddis


You could. It would be the final death of Burkean representation, of course ... MPs would have no right to independent judgment, only deemed to channel the Party line all matters. And hence there'd be no more conscience votes. Plus, doing away with seats doesn't address the threshold question - before a party's share of the party vote is counted in Parliament, what level of support must it attain?

A vaguely related question; if proportionality to the views of the public is so important, why do we only seem to care about it at all on election day?

As I said, representation is a compromise(d) concept. Our reason for electing representatives to do the dirty job of governing for us - that we don't have the time, energy or motivation to do it for ourselves - inevitably means that there will be some (lots of?) situations where those representatives decide the best thing to do is something different than we ourselves would choose (not having had the time, information or inclination to fully study it). 

by Brendon Steen on February 29, 2012
Brendon Steen

My thinking on thresholds has been going largely the same way. Any principled rationale for the coat tails (electorate lifeboat) rule eludes me. Sure, if some crackpot (dedicated local MP with a passion for the electorate) gets elected then s/he gets a seat. But that's the end of it - electorate victories should equal electorate seats, and nothing more.

Proportionality is all about party votes, and that's how it should be decided.

As for the party vote threshold - I think you've hit the nail in the head by focussing on the ability of the party to make a meaningful contribution in the House (either in coalition or opposition). I defer to more knowledgable people on the question of exactly how many MPs are required to make that meaningful contribution, but three sounds sensible to me.

In order to avoid as much mathematics as possible, I think we should express the threshold in terms of seats, rather than a percentage of the vote. We should just say "a party must receive sufficient party votes to entitle them to a minimum of three seats in Parliament, including any electorate seats they are entitled to".

Really, if our focus is on effective participation, and we determine that using numbers of seats, then we should dispense with the opacity (and, under the formula, unpredictability) of percentages and just say what we mean.

by Graeme Edgeler on February 29, 2012
Graeme Edgeler

(actually, whatever the Sainte Lague formula dictates, but that's too hard for my head)

Luckily, I have already calculated this for you :-)

approximately 0.42% of the unwasted vote

(calculated by rounding up the 120th Sainte-Laguë quotient (including every party in the calculation) to the next whole number of votes, and seeing what percentage of the vote that amounts to in respect of each election we have had). It fluctuates a little over time - sometimes closer to 0.41% would be enough.

That said, given that this seems low (less than half a full seat), I think it is presumptuous to assume that if we went with a low threshold, we wouldn't modify Sainte-Laguë at the same time (as the Royal Commission recommended). It also seems a little presumptuous of the Electoral Commission to say that if we got rid of the single seat rule, this change would result in an increased overhang.

by Graeme Edgeler on February 29, 2012
Graeme Edgeler

Any principled rationale for the coat tails (electorate lifeboat) rule eludes me.

It is a principled response that diminishes the undemocratic effect of the 5% threshold, ensuring that the votes of more voters are properly counted in the makeup of the House of Representatives.

In short, the principle is "proportionality". If one supports the principle of proportionality, then supporting the lifeboat rule is in accordance with that principle because the lifeboat rule makes the election result more proportional.

by Graeme Edgeler on February 29, 2012
Graeme Edgeler

a party that got 13.674% of the party vote would get 13.674% of the vote on any given motion in Parliament. But I suppose, even then you'd have to round it off at some number of decimal places...

Then don't use percentages, use votes. National got 1,058,636 party votes, so they get that many votes in the House. Then we could go the next step, and treat these votes as a voter's proxy, which could be moved between parties at will :-)

by Andrew Geddis on February 29, 2012
Andrew Geddis

"It is a principled response that diminishes the undemocratic effect of the 5% threshold, ensuring that the votes of more voters are properly counted in the makeup of the House of Representatives."

Accepted, in a system with an otherwise "too high" threshold (which I think 5% was). But as I note, the "proportionality enhancing" effect of the electorate lifeboat rule is a bit hit and miss (ACT vs NZ First; JAPC vs the Christians), only exists by giving a disproportionate say to some electorates over others (which is inconsistent with proportionality), and pisses people off. So ... if proportionality is what we want, the obvious answer is to dump the electorate lifeboat rule and lower the party vote threshold (whether to 0.41% (thanks for that), or 2.5%, or 4% or whatever).

by Graeme Edgeler on February 29, 2012
Graeme Edgeler

So ... if proportionality is what we want, the obvious answer is to dump the electorate lifeboat rule and lower the party vote threshold

I agree. It's just the Brendan asked for a principled rationale, so I gave one :-)

p.s. a couple of posts back, you suggested I shouldn't make a habit of being right, my apologies for doing so again (mostly). What changed you mind? I imagine it wasn't actually me, this time :-)

by Andrew Geddis on February 29, 2012
Andrew Geddis

Ha! That's hilarious ... I'd completely forgotten that exchange.

Obviously you've invaded my subconscious mind, which I find deeply disturbing. First you understand Danyl Mclauchlan better than he does himself, then you make me think what I think I'm thinking for myself ... are you the Devil?

[PS: As for "what changed my mind" ... I actually am not strongly invested one way or the other on the party vote threshold question. So given that it wasn't you - it just wasn't, OK? - we may simply have independently reasoned our way to the same conclusion.]

by Stafford on February 29, 2012

I'm generally for anything that will see Peter Dunne removed from Parliament.

by Andre Terzaghi on February 29, 2012
Andre Terzaghi

Seems to me a lot of the distrust of MMP is around minor parties having too much influence, given weight by the likes of Anderton, Peters, Hide etc being appointed ministers.

On the weight of this, I would like to see a much lower threshold for party vote to get in to Parliament (you've persuaded me to 3 MPs instead of 1), but a much higher threshold for eligibility to be appointed a minister, say an 8% party vote threshold requirement for ministers.

by Andrew Geddis on February 29, 2012
Andrew Geddis

@Stafford: "I'm generally for anything that will see Peter Dunne removed from Parliament."

You might want to have a chat to the good voters of Ohariu about that ... it's in their hands under any set of rules.


That would be a completely new development in our constitution - at present the PM can recommend that the Governor General appoint any MP as a minister, and the Governor General uses the prerogative powers to do so. So to do this we'd need to create an awful lot of new law - the "Ministers Act" or such like - to replace what has until now been a mix of convention and prerogative. And I'm always a bit wary about putting a legal straightjacket around what can be quite fluid and politically tricky arrangements.

That aside, I'd quibble a bit with the basic concept. Imagine a situation where we have a government consisting of three parties - one which got 37%, one which got 7% and one which got 5%. Should the 37% party really get all of the ministerial posts? Doesn't that start to look a bit like one-party government ... which is what we dumped First Past the Post to avoid?

by Brendon Steen on February 29, 2012
Brendon Steen

On the justification for the one-seat rule, I don't quite buy the idea that it's about improving proportionality. That might be an effect of it, but I don't think it stands up as a rationale.

As Andrew says, it's a bit selective about who deserves proportionality and who has to do without. It seems that the actual rationale has a lot more to do with rewarding electorate victories than ensuring proportionality.

by Andre Terzaghi on February 29, 2012
Andre Terzaghi

Thanks for educating me about constitutional aspects of the proposal.

As far as the basic concept being flawed, you mean kind of like our 1999 results? In that particular case, I think we would have got a better government if all the ministers had been Labour, and I must say I'm not a fan of either Labour or National. To my mind, the bigger downside might be the exclusion of the likes of the Maori Party. However, given the peculiarities of our particular version of MMP and electorates, it might be a positive thing to reduce the incentive to split the party and electorate vote and thereby try to create an overhang if the consequence might be ineligibility for ministerial positions.

As far as it being closer to an FPP government, well, yes, a bit, and I agree that's a downside, but any government would still need to negotiate confidence and supply agreements with minor parties to be able to govern. I think the Greens in the last term showed a good model of how a minor party can work with a major, without ministerial seats. And it seems to me it's always a trade-off between allowing arrangements to be fluid and politically tricky, and putting a framework in place that makes the situation less fluid and tricky but (hopefully) more transparent.

Looking at other governments around the world, such as Israel, it still seems to me allowing "extremists" into parliament doesn't do much damage, but when they get into ministerial positions they do a lot more damage, to the point that it would be beneficial to come up with some mechanism to ensure all ministers have a minimum level of support. Where that threshold might be would need a lot more thought and debate than would ever happen here.

by Kyle Matthews on February 29, 2012
Kyle Matthews

Andrew (or others),

If the electorate lifeboat was removed, what would be the likely result of that party's party vote in determining the makeup of parliament?

Would their party vote be completely excluded, and their electorate win be an overhang?

Would their seat be excluded from proportionality, and the total party vote, excluding whatever they got, be figured out across all the other seats, but not theirs?

by Graeme Edgeler on March 01, 2012
Graeme Edgeler

On the justification for the one-seat rule, I don't quite buy the idea that it's about improving proportionality.

There are other justifications: e.g. if having a percentage threshold is about ensuring we don't have a whole lot of single MP parties, then we shouldn't have a system which encourages single MP parties. If Rodney Hide is going to be there anyway, isn't it better that ACT brings in 4 extra MPs (in proportion with its party vote) than force ACT to be a single MP party, which we were trying to avoid in the first place?

by Graeme Edgeler on March 01, 2012
Graeme Edgeler

If the electorate lifeboat was removed, what would be the likely result of that party's party vote in determining the makeup of parliament?

I anticipate that the party would keep any seat to which they would be entitled in the calculation, up to the number of electorate seats they won. The seats wouldn't be automatic overhang (although that does appear to be what the Electoral Commission thinks would happen). The answer to this will also depend on what we do about overhang generally: some MMP systems don't have it.

by Brent Jackson on March 01, 2012
Brent Jackson

I agree with your views.  I would prefer a 2 seat limit but a 3 seat limit would be okay.

If the threshold is retained, I've had an idea to reduce the problem caused by votes being wasted if they're cast for a party that does not make the threshold.  Turnout is low enough as it is, without further disenfranchising voters who do actually make the effort to vote.  Each voter can be given the option to cast a second party vote which will come into effect if the first vote is for a party that does not meet the threshold.  This would allow voters to vote for the party that they wish to represent them regardless of its polling, but still ensure that their vote is not wasted by casting their second vote for a party that will get over the threshold.  This second vote could be called a "back-up vote", or a "threshold vote", and take the form of a box at the bottom of the party list in which can be written a number (or letter) corresponding to the party of choice.  (An alternative would to be allow a number 2 in your second party choice, and use a tick or cross or a 1 for your actual party vote).

I think that everyone who believes that the threshold should be lower should make sure that they go along to and do at least a quick submission.  The incumbent parties have little to gain by altering the system, so we need a strong showing to increase the chances of the threshold being lowered.


by Brent Jackson on March 01, 2012
Brent Jackson

By the way, I think a seat limit is better than a %age limit, as it is clear and easy to understand.  The calculations should be relatively straightforward.  Run the numbers and eliminate all parties with less than 1 seat.  Then continue removing parties in order of the fewest votes gained, until all parties have at least the seat limit number of seats.

by Gareth Ward on March 02, 2012
Gareth Ward

As you said Andrew, in trying to avoid the "denying of proportionality" that any threshold entails, the electorate-seat rule is in fact diminishing the broader concept of proportionality in the voting system.  So it should go.

It would be intriguing to see what happens to the higher list spots of the "lifeboat" parties if it does go though...

by Lyndon on March 02, 2012

Just FYI Umberto Eco extrapolates on the 1:1 map thing:


by John Norman on March 02, 2012
John Norman
Hoping to not appear as frivolous as the comment apprearing immediately above me (march 02, xio???) in this thread, I'd like to ask whether moving averages have merit for determining thresholds? Putting figures from Andrew's blog - 80, 79, 74 (%) - over a nine year( 3x3 Elections period) the ma comes out at 77+. That is to say in the context used above somewhat more 'space' for other parties to take up. Applied in a like manner to individual parties it would relate consistency and seriousness to such party participants.. At least these things appear the case.. hence my question..
by John Stroup on March 02, 2012
John Stroup

True proportionality, and true representation, if that is what the desired effect is indeed, would have representatives voting in accordance to the wishes and will of those that are represented. How do you find out what the wishes and will of the represented is? Ask them.

All major policy decisions should be put to a referendum. That is the only way to insure that there is not partisan party politics and cronyism going on. This does a few things, it makes politicians less “policy developers” and more administrators of policy that the electorate develops.

Unless the electorate is actually represented, you’re debating a minor point in the “mmp” representation paradigm.

The thresholds should be raised to 15%. As it is a somewhat common acknowledgement, you can get 15% of any demographic to believe almost anything.

Ron Paul has roughly 15% support, Greens got not quite 15%, and 15% [of Americans] believe in secular humanist evolution theory.

by Jack Cowie on March 11, 2012
Jack Cowie

I agree with the idea of the first comment here and a few others, who have pointed out the idea of precise proportionality where each vote in the election goes to a party (and through them to an MP) to use in Parliament.

In fact, my submission to the MMP review outlined this idea in good detail, and I strongly recommended it. One of the best aspects of it is that it naturally solves the problem of overhang MPs - to the extent that I suggest a 90-30 split of electorate and list MPs could be made to work.

Check out my submission:

Post new comment

You must be logged in to post a comment.