Thoughts on the MMP threshold in New Zealand

by kai on 25/02/2012

Note: This is a political post. It’s highly driven by my political opinions and the ongoing discussion on the preferred NZ election system. If you don’t like to read about these topics, please move on.

I was inspired to write this one up after reading a blog post of stuff by David Farrar. I don’t share his political opinion, but he’s asking a very valid question: “Why should the MMP threshold in NZ be?” (after the referendum on our voting system resulted in a majority of people wanting to keep MMP). Here are a few more thoughts on the topic – with a bit of historic explanation of why the threshold exists in the first place and what the real issue is.


New Zealand’s government is currently elected through a voting system called Mixed-member proportional representation (MMP). Essentially it’s a system where every voter has two votes – one for an electorate candidate and a second one for a party list. The party list vote determines the amount of overall seats a party has in NZ’s parliament, electorate winners will make it into parliament anyway and the remaining party seats are filled with party list candidates.

To my German friends that might sound familiar. It is indeed exactly the model that’s being used for the federal elections in Germany (“Bundestag”) and at least some of the state parliaments in Germany. Germany had a good reason for introducing MMP after World War II – the founding fathers of the Federal Republic of Germany (“West Germany”) basically wanted to avoid a scenario of the Weimar Republic’s parliament in which pretty much every man and his dog could be elected into parliament between 1920 to 1933. As we all know (we do know that, right?) – this was one of the reasons why it was so tremendeously easy for Hitler to get into power by playing the system to his advantage.

Problems with MMP

Overhang seats

There can be so called overhang seats. The most common scenario in which they occur is that a party has won more electorate seats (which are guaranteed to make it to parliament) than their share of votes through the party vote would allow them to get. Is that a bad thing? Well, yes and no. It’s not a bad thing in the grand scheme of things because there are usually not more than a small handful of overhang seats (something along the lines of 1-3 in NZ, and 1-10 in Germany – keep in mind that Germany’s parliament is five times the size of NZ’s parliament though). They are a bad thing from the point of view of “each voter’s vote should carry the same weight”. The reason is essentially that overhang seats contribute to an already existing imbalance of how powerful/strong a vote is. There are mathematical reasons for why that would always be the case in any voting system in a representative democracy (note the emphasis on “representative” here), but that shouldn’t matter for now.

Funny enough – the overhang seats issue becomes worse the more “special scenarios” politicians build into the voting system. In NZ we have Maori seats. In Germany there’s the issue of overhang seats by states (even though it’s a federal election) or seats/exceptions created for certain ethnic groups in North-Germany or near the Belgium border. I’m not saying it’s wrong to have those exceptions, but sometimes I think that politicians should just listen to mathematicians’ advise instead of tinkering with systems and make them worse in the process.


The other MMP “issue” is the magic and often discussed threshold. When the Federal Republic of Germany was founded, the founding fathers together with the western World War II allies wanted to make sure the Weimar Republic couldn’t repeat itself – ever. For a while, the “Reichstag” in the Weimar Republic had 17 parties in parliament. Forming a stable government in such an environment – impossible. Part of the solution is the MMP threshold. To avoid having again every man and his dog in parliament (as single-person fractions or tiny splinter groups), they basically decided that a party needs to win at least 5% of the party vote to become eligible for making it into parliament. That makes sense.

Repeat after me: That makes sense. Thanks!

What’s the problem then? The electorate vote is. Remember – earlier I’ve explained that candidates who win an electorate make it into parliament anyway – it doesn’t matter if the candidate’s party has made the 5% threshold or not. And now we finally arrived at the issue that’s currently being discussed.

The current legislation in New Zealand allows a party that wins at least one electorate to bring in more than this one direct candidate: all the candidates they would otherwise have got into parliament according to their share of the party vote. This has happened a few times in the past where political parties essentially campaigned heavily to get their one promising electorate candidate through to get something like 2-5 seats in parliament even though they wouldn’t have had a chance of making the 5% threshold.

In Germany there’s a similar rule: if a party wins three electorates, they can effectively ignore that threshold and bring in list candidates according to their party vote – boosting a party quite often from three to 15 or 20 seats. Interestingly enough it was just one electorate until the mid-1950s.

I do have an issue with that. Mainly because it heavily distorts the composition of the parliament and because it can contribute to difficulties forming a stable government. Note that I’m a big fan of coalition governments. It’s a difference though between governing in a coalition of two or three parties and governing in a coalition of nine or 13 parties.

If you can understand German – here’s a brilliant scientific article on rather paradox outcomes of MMP in Germany.

Now what to do with the threshold?

In his blogpost, David lists the four options:

“There are basically four options for the threshold. They are to:

  • (A) – Increase it
  • (B) – Keep it at 5%
  • (C) – Reduce it
  • (D) – Abolish it

In general terms, the higher the threshold, the fewer parties will be in Parliament, and fewer parties will be needed to form a government. The lower the threshold, the more parties there will be in Parliament, and more parties will be need to agree to form a government. Also the higher the threshold, the more wasted votes you get.”

He’s absolutely right with that and goes further into what-if scenarios based on past elections. What he doesn’t elaborate on though are the exceptions of the threshold rule. There’s no issue as such with the threshold even though New Zealand hardly had the level of historic baggage when MMP was introduced in 1996 (as opposed to Germany post World War II).

The threshold makes sense even here, because it keeps too small splinter groups out of parliament. Is it fair? No, it’s not – because it leads to wasted votes. Would a threshold of 3% or 4% be possible? Yes – it entirely would be from my point of view. It would allow for a better representation of smaller parties but still keep the total nutcases away. Increasing the threshold or abolishing it totally – rather not.

What I do strongly propose though is to abolish the exception that allows a single electorate winner to take their party list candidates into parliament. There’s no justification for it in the first place, but if we wanted to allow that, let’s at least make it a requirement that a party has to win two electorates before becoming eligible to doing that. Again – that will keep the total nutcases away but will also make sure there are less tactical games like National/ACT in Epsom going on in the future.

Adam Cameron February 26, 2012 at 8:16 am

Hi Kai
As an expat NZer, I follow what twerps^h^h^h^h^h^hpundits like Farrar say from a perspective of from the outside looking in. And I cringe an awful lot.

I cannot believe the NZ electorate is so stupid as to even have a referendum as to whether the system should continue with PR, or return to FPP. Clearly PR is intrinsically more “democratic” than FPP. I see this as a good thing, whilst not being a particular support of “democracy” as being a good way to elect a government(*). I don’t think MMP is a particularly good implementation of PR, but *any* implementation of PR is better than old-skool first-past-the-post.

Disclosure: I’m a Green, so it’s in my interests for smaller parties to get representation in parliament. Second disclosure: I think STV is the best way of implementing PR.

I’ve looked at why people wanted the referendum, and – to a person – it’s because they didn’t understand the system, or didn’t understand how they themselves actually vote. It was not a statistically-meaningful sample size though, I concede that.

I’m uncertain about your position that there’s a problem with a single-electorate-winner gets to take their party-percentage of votes to parliament. Unlike you I think, I have little time for the notion of geographically-based electorates. To me, that is an anachronistic hangover of times wherein vote counting needed to be decentralised to make it practical. When voting for *central* government, constituency-based concerns should be irrelevant. If 1% of the total (national) electorate vote for a party (rather than individual), then that party should have 1% representation in parliament. Fullstop. In NZ, that margin would be 0.8%, because there’s 120 seats to fill. People can band together in “platform blocs” (read: “parties”) if they like, but it still ought to come down to a electorate-wide vote percentage.

I think your inverse Godwin’s Law position that that sort of thing leads to the Third Reich is specious. It wasn’t who was voting for whom that caused that, it was the disastrous treatment of Germany by the victors after WWI, coupled with the economic depression of the 20s that caused that. And whether or not Hitler and his oiks made it into the Reichstag or not, they would have bullied their way into power one way or the other.

But back to the 21stC in NZ.

This leaves a real quandary in that some geographical-constituency-elected MPs (ie: they won their “electorate” vote) do actually act in the interests of their constituency (and staff an office in said constituency, and listen to their constituents and act on their behalf), as well as act on behalf of the party they are affiliated to. A lot of people see this as a good thing. I am unsure of my own position on this, because it’s good to have someone acting on behalf of their local interests, but should national parliament really be dealing with the interests of the people living in – for example – Eketahuna, simply because someone from there is in parliament? I err towards “no”. I see “constituencies” purely as an anachronism in modern national-level elections.

It’s with relief that I read that NZ’s voting population chose to stick to a PR system, but I don’t think the NZ population is prepared to engage their brains sufficiently to work out what is actually best for them. That said, they are no different from the UK population who recently vote *against* a form of PR, or the USAn population whose system is… err… well it’s a mystery to most people outwith the system, I think. And can yield some “awkward” results (like GWB being elected for not one, but two terms, I mean).


(*) I think a great percentage of people are not clued-up enough for them to be entitled to vote on things they don’t / can’t / choose-not-to understand. Basically enfranchisement should be earned, not be an automatic right once one reaches one’s late teens. IMO, obviously.

Oliver February 29, 2012 at 11:56 am

And, have you made a submission yet?

kai February 29, 2012 at 1:27 pm

Not yet, but I’m seriously considering it.

Oliver February 29, 2012 at 1:40 pm

It only takes a minute – although I’m not quite sure what they do with an individual’s submission in the end (does the sheer number of submissions with the same goal count towards anything?).

Tim Goodman November 5, 2012 at 8:45 pm

The overhang will pretty much always favour the major parties. Look at the 2011 election – without overhang, the result would have meant only one less seat – to the Nats.

I actually think it should stay too – if only to give some slim chance that the major party can actually govern. It effectively negates the effect of the minor party seat that has won at the electorate though

Kai November 5, 2012 at 8:54 pm

Yes, I agree. Historically (both in Germany and in NZ) overhang seats would usually go to the 2-3 major parties. I can see though why the MMP report would want to abolish overhang seats if the one electorate threshold is abolished. There’s clearly a chance for more overhang seats due to the small parties winning electorates here and there but not having the responding party vote to earn the overall number of seats.

To be fair – part of this situation is driven by the concept of Maori seats and the Maori and Mana parties obviously targeting those. It’s tricky to find a balance between all the objectives of a good election and seat distribution system: low disproportionality, an appropriate party vote threshold to avoid fragmentation of a parliament and catering for special scenarios like the Maori seats here in NZ or similar ethnic minority regulations in Germany. I think the overall outcome of the MMP report is very good, I just hope that parliament is actually implementing all the recommendations.

Comments on this entry are closed.

Previous post:

Next post: