History says Christopher Luxon can’t win – but it isn’t really backing Chris Hipkins either
Saturday, 18 July 2026
Henry Cooke is The Post’s political editor.
OPINION: Here are some patterns.
No party since the World War II has won Government with its support level in the 20s. At every election in the MMP era the main party of “the Government” has lost support in the lead-up to the election, with its showing on election day averaging 4.1 percentage points lower than its poll rating four months prior. No main governing party has ever been this low at this point in the cycle. The previous floor was Labour's 32.5% average in mid-2008 – and Labour lost that election. National is currently averaging around 29%.
But here are some others!
Labour have never won back office after a single term out of power, and there have been no single-term governments under MMP. Labour have not won an election with a man in charge since 1987. The last Prime Minister to lose office and then win re-election did it 66 years ago – two others have tried and failed, and both of them were Labour MPs who gained the premiership partway through the term.
We can keep going.
Read more:
RNZ poll: National hits new low under Luxon, Opportunity nears Parliament
Opportunity has had ‘positive, informal’ discussions about appearance in election debates
No party in the MMP era has entered Parliament without splitting off from another – sorry Opportunity! The last Wellingtonian to win an election and become Prime Minister was Walter Nash in 1957 – sorry Chris Hipkins! No party since the 1940s has replaced a prime minister mid-term then gone on to win that election – sorry Chris Bishop, or Nicola Willis, or Erica Stanford.
It would be churlish to suggest that past patterns are not useful tools for predicting the future.
After all, they are patterns – things that have happened over and over. Our ability to work out that the sun is going to come up again every day and it is going to get warmer then colder throughout the year is what allows us to have civilisation. So some refuge in the “well this has never happened before” is enlightening.
But it pays to remember that we are working with a fairly small sample size. There have only been 10 MMP elections so far. And they are far from controlled experiments in a lab, each one bringing a whole new series of exogenous circumstances to the table. New things happen all the time! No woman had ever won an election before 1999, and Helen Clark had actually lost one already. Christopher Luxon is the first person to become prime minister after just one term in Parliament.
We are also talking about a very protean electorate. Between 2023 and election day around 115,000 Kiwis will have died and 190,000 aged into enrolment – and around 100,000 adults will have become permanent residents, and generally eligible to vote. Many of those who died might have stopped voting years ago, and many young people and new immigrants will not end up enrolled, but we are talking about a dramatically different electorate this election, as we are most elections. That’s before you get into the overseas vote, which the Electoral Commission expects to be higher than ever this time around.
It is changeable not just in its composition but in its mores. Winston Peters, always a bit of a rogue but enough of a gentleman to attract older voters, is now happy using terms like “gripper” and “wanker” in public, language that would have verged on disqualifying in days gone by. His vote seems fine – indeed, he is heading into his largest-ever party conference this weekend. And few voters appear rusted on anywhere – the RNZ-Reid Research poll found 54.4% said they would seriously consider voting for a different party, although this is the kind of thing that people tell themselves without necessarily ever acting on.
All this change is starting to make old nostrums seem very old indeed.
The biggest departure to the election that most predicted is the real possibility that Opportunity (formerly TOP) enter Parliament.
The party have now scored north of 4% in two reputable polls – TVNZ’s Verian and RNZ’s Reid Research. And it isn’t the result of some technical quirk where they started asking about Opportunity where they hadn’t before – both firms confirmed to The Post this week that no change to their methodology had taken place.
The possibility itself upends the election.
If Opportunity win seats in Parliament it becomes extremely difficult to see a path to Government for National. Take the RNZ-Reid Research results, which projected re-elected coalition with National holding 36 seats, NZ First, 15, and ACT 10 for a total of 61 seats – just over the line.
Add just 10,000 or so votes to Opportunity – a bit over 0.3% – and that path to power becomes impossible. National dip down to 35 seats, NZ First to 14, and ACT to 9 – leaving the coalition three seats short of a majority.
It also leaves Labour short, even with the Greens and Te Pāti Māori – further behind in fact. But it is quite reasonable to assume that Opportunity would go with Labour in this scenario. Party leader Qiulae Wong has said repeatedly she wishes to talk to the party who gets the most votes first – that would be Labour. Her party’s main policy platform of a land tax used to pay for a Universal Basic Income is dramatically to the left of Labour and the Greens, making it very hard to imagine serious talks with National, ACT, and NZ First ever really getting traction. And the same RNZ/Reid Research polling this week found more of its supporters wanted it to go with Labour than National.
On the other hand, Opportunity being incredibly close to 5% but not clearing the hurdle makes it very hard to imagine Chris Hipkins retaking the ninth floor. It appears that the party’s vote is largely (if not entirely) coming from the left. As NZ First does all it can to mop up the wasted vote on the right, recruiting former leaders of minor parties such as Alfred Ngaro and Elliot Ikilei, a substantial wasted vote on the left could easily lock Hipkins out of power.
This kind of intricate seat projection work is a bit technical and nerdy – but it is exactly what is going through the minds of nervous party strategists right now.
Opportunity is clearly not for everyone. But it seems there are a large number of voters who are displeased with simply being offered a left bloc and a right bloc, with either the current Government in a slightly new configuration or Labour supported by the Greens and Te Pāti Māori. It would take very few of them to actually lend Opportunity their votes for the election to be changed completely.
Or it could also be a flash in the pan. Columns like this could look hilarious in hindsight come election night, when Opportunity finish well below the line.
Other black swan events for the election are entirely possible. What if the RBNZ hikes interest rates twice again ahead of election day to get on top of Iran-driven inflation, including a “Halloween Hike” as early voting has begun? What if the Government’s controversial changes to voting laws really do see dramatically lower turnout, scrambling the turnout assumption in each pollster’s model, and seeing the Coalition returned easily? What if there is some level of scandal on either side of the fence and the vote is materially impacted by that?
As ever when trying to work out how millions of individual people make up their mind, we are groping around in the dark here. The electorate is never quite as legible as a poll or focus group might make you believe, and it does not respond to events in a predictable way. Hence the need for everyone who thinks about politics to rely on patterns from the past to do so, alongside the brief glimpses of statistical clarity one receives from a good opinion poll.
Using those patterns and those polls, some elections really are knowable this far out. This one isn’t.