Analysis by Keith Rankin.
Role: Economic historian.
Keith Rankin, 9 November 2026 – New Zealand’s MMP voting system is the old First Past the Post (FPP) system with a corrective proportional overlay. As time has passed, and fewer remember voting before 1996, it has morphed into a proportional party list system (primary vote) with a local electorate FPP underlay (second vote).
What happened – and was almost certainly going to happen in any election in which Labour and National were close to each other and both were below 35% of the total vote – was that the FPP underlay trumped what was meant to be a proportional election. This has led to an eleven-seat ‘overhang’.
This 29 October 2026 article on polling history – RNZ-Reid Research poll: View all results and charts – shows that New Zealand was in ‘significant overhang risk territory’ from late 2024, long before Finance Minister Nicola Willis made the decision to run ‘list-only’ in 2026. She made her choice in full knowledge of the polling trend. Further, it is clear that National Party support below 30% dates back to around April 2026, well before any campaign to get National Party supporters to transfer their party votes to Act or New Zealand First.
One part of the problem on Saturday was that the FPP election allocated 71 seats, leaving just 49 ‘list’ seats to make the proportional correction. In the first MMP election in 1996, it was 65 FPP seats and 55 seats to make the proportional correction. If, in 2026, there were instead 50 FPP seats and 70 proportional corrective list seats, the overhang would have been much smaller, though National would still have had few if any list MPs.
A better fix would be to go to MMMP: multi-mixed-member proportional, using multi-member electorates. I outline this below, noting that it’s easy to understand, using present voting as an example, how the overhang problems resolve under MMMP. A few other comments first, though.
The 20-Day wait for a result update!
This long wait is a complete nonsense. The election was on Saturday 7 November, and the final result is scheduled for a Friday 27 November three long weeks later. The election night tally is a provisional vote count. There is no reason why there cannot be additional provisional counts on Friday 13 November and on Friday 20 November. Such counts would facilitate the process of government formation, substantially minimising the uncertainty – the long vacuous vacuum, when mischievous narratives take hold – relating to the final count. We note that many past governments have been formed on the basis of statistically significant provisional results.
The Natural Gerrymander
One most important question to be asked today is: how can National win twice as many electorates as Labour, when both parties are equally popular? The answer is the natural gerrymander, which I discussed last week in my Gerrymandering and Dummymandering: USA and New Zealand, Scoop 2 November 2026. In that article I correctly predicted that the Wairarapa electorate could be pivotal to the final election outcome.
In the 2005 general election, the electorate vote totals for National and Labour were both about 900,000, and the electorates were split 31 each. This year is the next time after 2005 when National and Labour have been so close to each other; yet the two-party electorate lottery has swung very differently. This time has clearly been an election much more favourable to incumbent electorate MPs than was 2005.
2026 Provisional Results under Five electoral systems
One. The actual provisional result of the election, under MMP with 71 electorates, is: Right (67), Left (64); placing Opportunity as part of the Left bloc. This is made up of: N(42), L(36), G(16), A(13), NZF(12), O(7), TPM(5). Total Parliament of 131. Winning margin for the Right is 3 seats, which is 2.3% of 131. A narrow provisional win to the Right.
I have recast the provisional result under four other compatible electoral systems. The results vary from a narrow win to the Left to a ‘landslide’ to the Right, all with the same actual levels of popular support for the seven parties.
Two. Under a pure list proportional system, the result would have been: Right (60), Left (60). This is made up of: N(35), L(36), G(16), A(13), NZF(12), O(7), TPM(1). Total Parliament of 120. Winning margin to the Right is 0, which is 0% of 120. A Right-Left tie.
Three. Under an FPP single-member electorate system, the result would have been: Right (43), Left (28). This is made up of: N(42), L(21), G(2), A(1), NZF(0), O(0), TPM(5). Total Parliament of 71. Winning margin to the Right is 15, which is 21% of 71. A ‘landslide’ win to the Right.
Four. Under an SM supplementary-member electorate system – a tweaked version of FPP with some list MPs – the result would have been: Right (67), Left (43). This is made up of: N(56), L(36), G(9), A(6), NZF(5), O(3), TPM(5). Total Parliament of 120. Winning margin to the Right is 14, which is 11.7% of 120. A ‘tsunami’ win to the Right, about the same sale as the ‘blue tsunami’ in the United States last week.
Five. Under a multi-member MMP system (MMMP, see below), the result would have been: Right (60), Left (61). This is made up of: N(35), L(36), G(16), A(13), NZF(12), O(7), TPM(1). Total Parliament of 120. Winning margin to the Left is 1, which is 0.8% of 121. A tie-breaker win to the Left. (There is no Left-bias in MMMP; the Left advantage here relates specifically to Te Pati Māori.)
Description of the proposed multi-member MMP system. Call it MMMP.
Imagine a Parliament of 120 members, with 18 electorates (including two Māori electorates), normally with four-members per electorate, and 48 list MPs. There would be 72 electorate MPs and 48 list MPs; just as in a normal election today.
Parties choose up to four candidates per electorate. These candidates, 72 maximum per party, would also constitute the nationwide party list.
Each voter has a single party vote, and a single electorate vote (ie one electorate vote, for just one candidate, though ranked-choice preferential voting could be an option). The four candidates with the most votes are provisionally declared the election winners. Parliamentary representation is topped up by the party list, to ensure party proportionality.
If there is an overhang result, then each electorate only returns three MPs, and the number of list MPs grows to 66.
If there is still an overhang, then each electorate only returns two MPs, and the number of list MPs grows to 84.
If there is still an overhang, then each electorate only returns one MP, and the number of list MPs grows to 102.
If there is still an overhang, then that overhang is left in place, allowing the Parliament to be slightly above 120.
The nationwide party list would be required to be regionally balanced. There would be no list-only candidates. Each party would take one candidate from each electorate to be its top-18 on its nationwide party list. The next 18 on each party’s party list would also be one from each electorate. And so on. Within that constraint, parties would be free to rank the candidates as they please.
Application of MMMP to this month’s election.
There are two overhangs to contend with this year. The National overhang could be resolved by going down to three members per electorate, and 66 list MPs. But the TPM overhang would require going down to just one MP per electorate, and there would still be a rump overhang of one. This is because TPM’s party vote only entitles it to one MP, but it would have two out of the 18 electorate MPs. So there would be a Parliament of 121, with 18 electorate MPs, and 103 list MPs.
This would be a highly unusual result. Under MMMP, there would be every incentive to vote TPM on the party vote as well as the electorate vote. That way there could be as many as eight TPM members in Parliament – presuming TPM only stands candidates in Māori electorates – though almost always there would be less than eight, because of some non-TPM representation in Māori electorates. We also note that TPM could and should solicit general party list votes. (Indeed, during the 2010s, I party-voted TPM in every election. And I am not Māori. In 2011, I was the only TPM vote in the Avondale Primary School booth!)
On the matter of National’s senior MPs who missed out this election, this could no longer happen. First, all candidates would have to be listed in an electorate. And if they don’t make it to Parliament in that electorate, they would make it into Parliament via the party list. In general, in any situation whereby the number of MPs per electorate has to be reduced, most MPs who miss out would be returned as list MPs.
I have just described an excellent MMMP system, which tweaks MMP to remove the FPP wrinkles we saw this year, and works exactly the same from a voter perspective. One vote for one party, and one vote for one candidate. Split voting would be not only acceptable, but encouraged. The overhang loophole would be closed. And most MPs would, most of the time, be electorate MPs rather than list MPs.
About the writer:
Keith Rankin (keith at rankin dot nz), trained as an economic historian, is a retired lecturer in Economics and Statistics. He lives in Auckland, New Zealand.