Analysis by Keith Rankin.
Role: Economic historian.
Keith Rankin, 6 November 2026 – The New Zealand election took a dark turn last week with the campaign by New Zealand’s uber-elites – represented by Harry Mowbray – to try and engineer a result contrary to the public will; and, more significant, to undermine proportional democracy altogether.
New Zealand’s system of proportional representation is not perfect, and can under certain circumstances achieve a non-proportional outcome; but recognising this imperfection is not the same thing as trying to engineer a non-proportional outcome.
(Yes, Te Pati Māori has already walked this slippery slope; though we have understood from the outset that a Māori party would commonly have ‘overhang’ MPs. This was never a problem when the Māori Party was an authentic ‘centre party’. There are many who rue that party’s shift away form its former centrist position. There is nothing inherently Left about indigenous politics in New Zealand.)
Unfortunately Guyon Espiner made a hash of this interview: Harry Mowbray discusses his ‘Split Your Vote’ campaign, RNZ 30 October 2026. Mowbray tried to spill the beans on his anti-competitive anti-proportional agenda – a return to a tight elitist political duopoly – but Espiner overrode Mowbray, preventing him from fully exposing his less-than-democratic views.
Nevertheless, Mowbray made it clear that he wants people to vote Act and New Zealand First in order to make a case for New Zealand to abandon proportional representation. An end to proportional representation would of course then put an end to ‘minor’ parties such as Act and New Zealand First. He wants people to vote for these two parties – one a Labour breakaway, one a National breakaway – as part of an end-goal to collapse these two parties into National.
The problem with the excessive use by the media of the term ‘MMP’ instead of ‘proportional representation’ is that there are two reasons why some people oppose our voting system of 33 years; a ‘mixed-member proportional’ system which effectively became law on 7 November 1993. New Zealand adopted the German system. The German Federal Republic chose this particular system in 1949, with a high (five percent) representation threshold for parties; a threshold intended to create an entry barrier to anti-establishment parties on both the Right and the Left.
One reason for opposition to MMP relates to some aspects its mechanics (such as the 5% entry threshold); the other reason is that many people – typically people of a ‘ruling-class’ persuasion – dislike proportionality in principle. Mowbray’s main agenda was to undermine proportional representation, and not to create a temporary advantage for the Right. Espiner was oblivious to nuances around the white-anting of our very good but not perfect electoral system. When anyone speaks in a derogatory manner about MMP, it’s critically important to understand what they want in its place. Do they want less proportionality? Or more?
MMP in Germany
At least Guyon Espiner redeemed himself by getting a German expert on Radio New Zealand on last Tuesday: Sven Siefken on Germany’s experience with MMP overhangs, RNZ 3 November 2026. Although we should have started discussing Germany’s election result in 2017, or even earlier, more than nine years ago.
Actually, Germany had its first major-party overhang in 2005. Wikipedia says: “With polls remaining so close, speculation increased that (as in the elections of 1994, 1998 and 2002) a small number of overhang seats might significantly impact the election results.” It turned out that the 16 overhang seats then were almost-equally shared between the two major parties, in an election in which only 50% of all seats were electorate seats. Both major parties falling to around 35% of the party vote caused the overhang. An important difference with New Zealand was that each state – Germany has a federal system – had its own party list; this makes an overhang more likely.
In 2009 in Germany the overhang increased to 24, as the major parties became even less popular. So, for 2023 they fixed this by introducing ‘levelling seats’ to maintain proportionality. In 2013 the German parliament bloated by an additional 33 seats; overhang and levelling seats. In 2017, in an election held on the same weekend as the Jacindamania New Zealand election, Germany’s parliamentary bloat increased to 111 overhang and ‘corrective’ levelling seats. A red flag for New Zealand to see, if ever there was one! But nobody in New Zealand – pundits and experts alike – seemed to notice. The German overhang heavily favoured the CDU, which won 185 out of 299 electorates despite getting less than 33% of the party vote.
In 2021 the overhang increased again, with a bloat of 137 seats. The combined major party vote was less than 50%. The electorate seats were evenly distributed, because the two major parties were equally unpopular.
In 2023, the Germans decided to have a fixed 630 seat parliament with no overhang. Their solution was to disqualify some MPs who had ‘won’ their electorates; though the excluded MPs only won with a plurality of votes, not a majority. In the subsequent 2025 election, there were 23 exclusions of ‘winning’ electorate MPs. We should note that, in order to minimise electorate-winning exclusions, the number of electorate seats – or ‘mandates’ as the Germans call them – reduced to 44% of the total 630-seat parliament. In 2025, just 45% of votes went to the two major parties.
All of this should have been news in New Zealand. But it wasn’t. I find that New Zealand has adopted a somewhat more insular mindset than it had in the twentieth century.
Saving MMP in New Zealand from the ravages of its elite critics
1. Fewer Electorates
For New Zealand to have a minimum parliament-size of 120, with 40 general electorates and about 5 Māori electorates, the risk of an overhang would be small; unless, that is, New Zealand voters go to extremes to play the system. (See my third ‘extreme’ scenario in Strategy Scenarios for Voting in New Zealand in 2026, Scoop 29 October 2026.)
There need be no presumption of a fixed number of electorates in the South Island. Further, electorate sizes should become more flexible, focussing on not splitting communities. (Paraparaumu split down the middle in 2023 – half in Mana and half in Otaki.)
2. Multimember Electorates, equivalent to Wards
As an alternative or an addition to the proposal above, New Zealand could switch to multimember electorates – eg to five-member ‘wards’ – much as they use in Ireland. And these wards should be dynamic, meaning that each ward has a maximum of five MPs, not exactly five.
The Māori electorate could be one ward. Then there could be 10, 11, or 12 general wards. In the event of a zero overhang, each ward would elect its full complement of five MPs. In the event of a small overhang, wards could each lose one MP. The number of list MPs would then increase slightly. In the event of a large overhang (ie more than 11), then each ward would lose two electorate MPs, and there would be a further adjustment to the number of list MPs to ensure a total parliament of 120.
In the event of a maximum possible overhang – say an overhang of 65 MPs (noting that the present maximum possible overhang is 71 MPs) – then each electorate/ward would return just one MP. That very unlikely event would deliver a Parliament with 105 list MPs. That would create a disincentive for voters to treat any party as ‘electorate-only’.
This is an excellent reform option.
3. ‘Ranked choice’ preferential voting – Trifecta voting
To fine-tune the ward option listed above; I would advocate the used of ‘ranked-choice’ voting. I personally favour what I call ‘trifecta’ voting; each voter would number just three preferences only. (In Australia, where there is preferential voting, there is pressure to give every candidate a preference, even if they belong to the Rubbish Bin Party.)
I would favour using trifecta voting for both the party and the electorate.
For electorate voting, the result would be that each electorate (aka ward) would end up with electorate MPs from multiple parties. This process minimises electoral distortions arising from unranked voting (which we have in our local body elections in Auckland; there have been Community Boards in Auckland where every member was part of the same ticket). This would mean that electorate constituents would normally have a choice of local MPs from different parties to seek help from.
For party voting, the trifecta method would remove the problem caused by the five-percent threshold. Thus, I could vote for Te Pati Māori (1), Opportunity (2), and National (3). Because Te Pati Māori is unlikely to register 5% nationwide, my vote would then most likely transfer, in this case to Opportunity. And then if Opportunity only scores 4.99% of the total vote, in this example my vote would transfer a second time, in this case to National.
Prediction for 2026
Finally, here, my prediction for tomorrow. Last week my hunch was that the Right would get 51% percent of the party votes, and would win a greater percentage of seats, due to a greater overhang on the Right than on the Left. (And, for convenience here, I am counting Opportunity as ‘radical-centre, slightly left’). That would, therefore, have meant a Right government.
I think, now, that this has been a weak week for the Right. The Mowbray gambit might have backfired; National may get more party votes than Act or New Zealand First would like National to get. Further, I think that National will win fewer than the 40 electorates which Harry Mowbray is banking on.
My gut feeling now is that the Left will get 51% of party votes, and that there will be a larger – though not large – overhang on the Left than on the Right. That could be unwelcome news for Willie Jackson. Opportunity may get close to 10% of the party vote. Further, I think that there could be an opening for a National/Opportunity/Act government, without New Zealand First.
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.