OPINION: Ani O’Brien
I have thought long and hard about writing this. I wrote a kind of explainer about 6 weeks ago, but I wanted to be sure I was sharing the best information before throwing my support behind the strategy.
And, to be honest, I have felt really sad for National and the situation it finds itself in although, of course, much of it is self-inflicted. I have a great deal of respect for many in the party and I know they are feeling wounded by the increasing push for strategic voting.
National wheeled out Sir John Key to resist the “overhang” narrative at its recent campaign launch. Key said it was “not clever” to consider splitting our votes and that “two ticks blue” is the only strategy to ensure a strong National-led government.
Sadly, he is wrong. As former National Party minister Maurice Williamson has said repeatedly, Key is right in that two ticks was the best strategy for National voters in every other MMP election we have had so far, but he is not right in 2026.
The reality is that National and Labour are polling historically low in terms of the party vote. The latest poll had them deadlocked on 28%, meaning no matter which party ends up in the Beehive after the election, its coalition partners will wield significant power.
With Opportunity now sitting above the 5% threshold regularly, the chances of the left bloc getting over the line are looking more and more likely.
However, the centre-right has a rare lifeline – a lifeline that National is desperately telling us all to ignore. It means that if the right wing bloc is to get back into government, we may have to drag National kicking and screaming over the line. Put simply, we may have to save the party from itself.
How it works
The simplest way (it seems more complex than it actually is) to explain the situation is that under MMP, list MPs start getting seats only once a party’s party vote entitlement is bigger than the number of electorate seats it has won. If the electorate seats won outnumber the party vote seats it is entitled to, the party simply gets to keep all its electorate seats and gets no list MPs.
But because this gives them more seats than their party vote allows, it creates an overhang – extra seats that are added to Parliament because they cannot be taken away from anyone else. That is how we ended up with 123 MPs this term instead of the usual 120. Te Pāti Māori won more seats than its party vote entitled it to.
National is in the unique situation of holding 44 electorate seats while polling at 28%. So assuming it wins them all back and gets 28% of the party vote, it would end up with 44 seats, despite being entitled to only around 34, based on its party vote share. This would create an overhang, adding an extra 10 seats to Parliament.
National isn’t likely to win back all it electorate seats but it could pick up a few marginal ones it currently doesn’t have. I have seen estimates of it ending up with 36 to 42 electorates seats.
Let’s model it. Assuming National wins 40 electorates, its party vote needs to entitle it to at least 41 seats before just one person comes in from its list. So it would need somewhere around 34% to 35% of the party vote.
That means National needs to gain at least six points in the polls in order to bring in Nicola Willis who is first on the list (after removing likely electorate winners). Election day, on 7 November, is just 27 days away, with advance voting beginning on 26 October. And this assumes National doesn’t win more than 40 electorate seats.
The reality
Could a party move six points in the late stages of an election campaign? Of course, in theory. Elections can move quickly and polls are snapshots, not prophecies.
But National would need to depart totally from the trajectory we have seen so far this term. The trends are, in fact, going the wrong way. In the 1News-Verian polls National was on 34% in February, 30% in April, 29% in June, 29% in August and 28% now. Far from recording even one single upward movement in this poll series at any point in 2026, National has instead moved six points in the opposite direction.
Nor does the latest result look like an outlier when compared with other pollsters. September’s Taxpayers’ Union-Curia poll had National at 29%, down two points, while Roy Morgan had National at 31% in August, down from 32% in July.
National’s polling has not been a perfectly straight line down the slope since the 2023 election, but zoom out and the direction of travel is pretty consistent. National won 38.1% at the 2023 election and spent much of 2024 hovering around 37% or 38%. During 2025 it largely settled into the mid-30s and this year declined some more.
And that is why, as much as I sympathise with them, I struggle with National’s response to the vote splitting argument.
Its determination to push “two ticks blue!” effectively requires voters to assume National is about to stage a dramatic and unprecedented recovery in its party vote. Sure, it is possible. But is also possible that I will win the Powerball this weekend. Well, if I buy a ticket. There is a difference between something being possible and there being evidence that it is likely to happen.
The counter-argument
National has been arguing that giving your party vote to ACT or NZ First will result in experienced National ministers being replaced in cabinet by obscure small-party candidates. But National is going to lose those ministers anyway.
As I said, it will need to gain around six points in the polls in order to save just one of them. They would need to surge up the polls even further in the next four weeks to save Chris Bishop, Paul Goldsmith and Gerry Brownlee.
This is why I think it is a circumstance that is in many ways self-inflicted. The party made choice after choice throughout this term of government that lost it favour with its voters. But what National chose not to do probably hurt it the most.
It also made the choice not to have Nicola Willis contest an electorate – it could even have parachuted her into Chris Penk’s vacated seat. Would it have looked desperate? Yeah, probably, but she would likely have won the safe seat and not be looking down the barrel of a sabbatical or early retirement.
In any case, those details are a distraction. National is in all likelihood going to lose Willis, Bishop, Goldsmith and Brownlee. A massive loss of experience and institutional knowledge, no doubt. But it is reprehensible to send centre-right voters on a fool’s errand to save them, all but guaranteeing that the coalition ends up in Opposition.
New Zealanders who want the centre-right coalition back in, or at least want to keep Labour/Greens/Te Pāti Māori/Opportunity out, deserve better than to be gaslit into thinking it is realistic that National will achieve such a massive turnaround against the trend.
There is a lifeline to reach for and the opportunity to potentially save National from itself. The centre-right can split its vote.
Because it is most likely to win (except in Epsom and Tāmaki where ACT is incumbent), National can snap up as many of the electorate seats as possible for the centre-right.
Party vote ACT or NZ First
Moving a party vote from National to ACT or New Zealand First does not manufacture another centre-right MP out of thin air. But because National will be winning so many electorate seats and won’t be getting list MPs, there is no point in giving it the party vote.
Reducing its party vote further does not make its electorate MPs disappear, whereas giving it to NZ First or ACT actually results in more list MPs because their party vote entitlement will be higher than the number of electorate seats they win.
In broad terms, this increases the size of the centre-right bloc and – depending on how much of an overhang is created – will likely get all three parties back into government.
Can it go wrong?
Yes. Human beings do not tend to vote mathematically or behave rationally.
To get the full benefit of this strategy would require a lot of centre-right voters to split their votes. The model works by holding National’s electorate victories steady while reducing its party vote. Real elections do not politely hold variables constant for the convenience of Microsoft Excel. But as things are now in the polls, a small percentage shift in the party vote toward ACT and NZ First could make enough of a difference.
If National’s party vote collapses, some fear that its electorate vote may collapse with it. That is why the strategy relies on NZ First and ACT voters giving their candidate vote to National (except in Epsom and Tāmaki) too. National needs to win the electorates and the two smaller parties need to sweep up the party votes.
The complexities and risks are why I was reluctant to declare this as some kind of genius foolproof strategy several months ago. The principle is real and the maths add up, but I wanted to see if the party votes would move at all in the polls. With four weeks to go, I think it is safe to say they simply will not move enough.
Vote splitting is a much needed lifeline for the centre-right.
The interests of an individual party and the arithmetic of a governing bloc are not always identical and this is the source of the suddenly rather heated argument on the Right. National wants to save National. People promoting the overhang strategy are looking at the coalition as a whole.
MMP was designed to make Parliament proportional. Thirty years later we have arrived at the point where people with spreadsheets are trying to determine how many extra MPs can be produced by deliberately making one party less proportional. Democracy is magnificent.
None of this is a loophole invented by Harry Mowbray or Maurice Williamson or anyone else. It is a quirk of MMP and it has produced overhangs before.
Elections NZ describes the phenomenon explicitly on its website. It is totally legal and fair game for Te Pāti Māori to ask for electorate votes while encouraging party votes to go to the Greens or Labour. And it is totally fine for the centre-right to use the mechanism too.
All is fair in love, war and politics.
Ani O’Brien is a political columnist and analyst. The above first appeared on her Substack and is republished with permission
