The poll that may end Luxon’s leadership matters. But the conversation is so narrow that what we leave out is as important and deserves a serious debate

Natalia Albert*

The Taxpayers’ Union-Curia poll last week put National at 28.4%, its lowest result since Christopher Luxon became leader. The political media did what it always does with a number like that: it went straight to the only question it knows how to ask. Will he stay or will he go?

It’s not a bad question. Whether Luxon leaves matters. The poll matters. But I want to argue that every time we stop there, at the number, at the leadership drama, we miss something. We miss quite a lot, actually.

So let’s unpack it.

Before I go further, I want to flag that Bryce Edwards has done the forensics on who is actually behind these numbers, who commissioned what, who funds whom, and why the media’s reporting standards are embarrassingly inconsistent. That piece is worth your time. What I want to do here is something different. I want to look at what happens after the poll lands and why the conversation we have in response to it is so much smaller than it should be.

Polls shape political reality

It’s called the bandwagon effect: the tendency for people to update their own views in the direction of whatever the polls say the majority thinks. Research identifies three ways polls induce conformity: people want to be on the winning team, people assume the crowd knows something they don’t, and people resolve cognitive dissonance by switching to whoever looks like they’re winning. A poll that triggers a leadership crisis may contribute to the very collapse in confidence it was measuring.

When voters see 28.4%, they absorb it and make it gospel. The sociologist Pierre Bourdieu went further, arguing that polls don’t discover public opinion at all — they produce it. The act of asking millions of people whether they approve of a leader manufactures a discrete, measurable opinion on a question most of them had never consciously formed before.

On this reading, the poll doesn’t just reflect the political landscape. It actively shapes it. Which means when we ask whether Luxon should go because the polls say so, we are in part responding to a political reality the polls themselves helped construct.

And what has that constructed reality produced? One question, repeated across every outlet, every panel, every dinner table conversation this week. Will he go?

Even if Luxon goes, less changes than you think

Let’s say the caucus moves. Let’s say there’s a new face at the podium by April. What actually changes? Really? Let’s say it’s Erica, or Nicola, or Chris. Will that transform anything substantial or structural?

A National leadership change does not reset the coalition architecture. Luxon leads a three-party government. ACT and NZ First are independent power centres with their own electorates, their own brand, and their own survival imperatives going into November. A new National leader inherits those relationships and doesn’t get to renegotiate them from scratch. If anything, a mid-term leadership change destabilises them. Peters and Seymour would immediately start recalibrating their leverage.

Then there’s everything else that doesn’t change. The existing coalition agreements. The budget cycle already in motion. The policy commitments already locked in, the regulatory reforms, the public service restructuring, the Roads of National Significance programme. The institutional machinery of this government keeps running regardless of who stands at the podium. What changes is just a name and a face. What doesn’t change is the direction, or the institutions that drive the direction.

This is also worth remembering: in New Zealand we don’t vote for individuals at the national level. We vote for parties. The leader is not the government. The party and its coalition commitments are. Our fixation on individual politicians as the primary unit of political analysis is analytically shallow, and it actively obscures where power actually sits and how it actually operates. Leadership drama is compelling. It is also one of the smallest parts of the political picture. And the more space it takes up, the less room there is for the questions that actually matter.

The deeper problem is what we’re not even imagining

When a bad poll drops, the entire political conversation collapses into one question. The leadership question. And while everyone is watching that drama, nobody is asking the harder one. What would a genuinely different political offer actually look like? What would it take to govern differently?

Last week, I published a piece on what a Labour and Opportunity Party pre-election coalition could look like, the structural complementarity between a capital gains tax and a land value tax, the MMP arithmetic that makes it viable, and the case for Labour signalling coalition intent before the election rather than after. The comments I got were revealing. Some readers couldn’t hold the hypothetical at all. The framing kept collapsing back into “but TOP isn’t polling well enough” or “Labour would never do that.” Which are answers to a different question. The question I was asking is whether the architecture makes sense. Whether the political imagination exists to even try.

It doesn’t, right now. And I think poll obsession is part of why.

When we train ourselves to think in terms of who’s up, who’s down, who’s in, who’s out, we lose the capacity to think structurally, to think of the broader system that matters so much. We lose the ability to ask what a creative, coalition-aware opposition strategy would actually look like. We let everyone off the hook. National waits out the storm. Labour watches the numbers without building anything. Minor parties stay in permanent limbo. The media refreshes the leadership drama. Nobody has to do the harder work of imagining what change requires because the poll cycle keeps supplying a simpler story.

This is not just analytically incomplete. It has a real cost. Votes fall below threshold and disappear. Coalition possibilities that could shift the arithmetic never get tested because nobody dares name them before election night. The conversation we need to have about what governing differently actually means gets crowded out by a question that, even when it resolves, resolves into very little.

So watch the polls. But watch what they’re doing to the conversation too

Whether Luxon survives the next month matters. I’m not dismissing it. But the leadership question is the smallest possible version of the change question, and we keep treating it like it’s the whole thing.

The number tells us the public has lost confidence in this government’s direction. That’s significant. But it doesn’t tell us whether the opposition has a plan, whether the coalition architecture exists to do something different, or whether any of the actors with the power to change things are thinking beyond the next poll.

Those are the questions that actually determine what happens after 7 November. Right now, nobody is being asked to answer them. And that is a problem the polls will never show us.

Reference: Pierre Bourdieu, “Public Opinion Does Not Exist,” Les Temps Modernes, No. 318, January 1973. Translated into English in 1979 and published in Communication and Class Struggle, Vol. 1, International General, New York, pp. 124–130.

*Natalia Albert is a political scientist living in Wellington exploring how to govern divided societies in diverse, liberal democracies, with a focus on New Zealand politics. She writes weekly on her Substack, Less Certain. Albert stood as a TOP candidate in the 2023 election.