The Government's council shake-up is on pause until after the election. Who knew institutions were so institutional?

By Natalia Albert*

Last week Simon Watts pressed pause on Head Start, the Government's fast track for merging councils. Mayors were fuming. Porirua's Anita Baker called ministers cowards. Andrew Little said the Government didn't have the numbers around the Cabinet table. Labour, which had already called it a "false start", got it’s I-told-you-so moment.

The easy read is that the Government got cold feet six weeks out from an election. It’s not that simple. What I’m interested in is what was the Government thinking when they set out to redraw 37 years of institutional plumbing in three months and then seem genuinely surprised that it couldn't. Like, in what world? Make it make sense? 

First, what actually got paused?

In November 2025, Simon Watts and Chris Bishop announced they would abolish regional councillors and replace them with boards of mayors, followed by regional reorganisation. Bishop called it the most significant change to local government since 1989. So why then give them three months? That doesn’t make sense. In what world? 

Consultation closed in February with 1,153 submissions. The official summary found broad support for change, but overall opposition to the actual proposal. A "final proposal" was meant to land in March so drafting could begin. It didn't, and no bill has ever been introduced.

Then in May came Head Start. Councils had three months to put up merger plans, with Bishop's warning: lead your own reform, or we will do it for you. By 9 August, 18 proposals had arrived from about 44 councils. Cabinet promised decisions in September. On 24 September, it announced it wouldn't make any until early 2027.

In August, Bishop said past reforms failed because doing nothing is always an easy option. I get the frustration. But history is also right: these reforms are hard, very hard. Much harder than maybe politicians understand. Institutions like regional councils sit on years and years of legacy, formal rules, informal rules and layers up layers of decisions standing on years of legacy. 

Political scientists describe institutions as formal rules plus the informal practices that grow around them. The formal part of a regional council is the legislation, the boundaries, the rates. The informal part is everything that has settled on top: the staff who know the rivers, the consent regimes, the iwi relationships, the flood schemes, the ports. 

What came back from Head Start

Southland sent two competing proposals, with Gore in both. Otago sent two rival models. The Western Waikato proposal included Hamilton and Taupō, neither of which signed up; Hamilton voted 12–3 not to take part. In Wellington, Upper Hutt opted out and Kāpiti wants its own authority.

Watts's own explanation for the pause was that none of the proposals came with the unanimous view of its region, and that more work was needed on representation and on how regional functions would be transferred. What did he expect? Seriously? In three months, that a few A3s were going to be developed and he was going to get a simple set of options? Really? Never. 

They're the settled answers to who decides and who does what. Unsettle them, and every one of those questions reopens at once.

So how did Bassett do it in 1989?

This is the obvious objection, and the Government invited it by invoking 1989. Labour's Michael Bassett turned roughly 850 local bodies into 86 councils, effective 1 November 1989. If it was possible then, why not now? Because almost nothing about 1989 applies in 2026.

Labour had campaigned on local government reform back in 1984, so there was a mandate and years of lead-in. The Local Government Commission's findings were binding. There were no polls. It was a single-party government under first past the post, with no upper house and no coalition partners: Geoffrey Palmer's "unbridled power". And the whole state was being torn down and rebuilt under Rogernomics. 

But the biggest difference is that in 1989 there was nothing to unpick. Regional councils didn't exist yet. Bassett built them. Their boundaries were largely drawn around water catchments, because they were designed to manage rivers and land, and two years later the Resource Management Act built the whole environmental regime on top of them. Thirty-seven years of regional plans, consents, monitoring data and relationships now sit on those lines.

The rules have also changed since. MMP arrived in 1996. Local communities killed merger after merger: the Commission dropped Northland's and Wellington's proposals in 2015 for lack of public support, Hawke's Bay voted 66 percent against in the same year, and Wairarapa's poll failed in 2017. A 2019 amendment made polls mandatory for major Commission-led reorganisations.

Head Start's answer was to go around all of that with bespoke legislation. Which tells me the Government knew the formal machinery was slow.

For comparison, the Auckland supercity took a Royal Commission from October 2007 to March 2009, then legislation, before the council started in November 2010. Three years, for one region, with the groundwork done by a commission. Head Start gave the whole country three months.

Politics runs on a different clock

Paul Pierson's Politics in Time is the book I kept thinking about while reading all this.[3] Two of his ideas do most of the work.

The first is increasing returns. Every year an institution survives, more gets built on top of it, and the cost of changing it grows. Regional councils have had 37 years of compounding.

The second is time horizons. Politicians are rewarded inside a three-year cycle. Institutional change pays off over decades. Pierson sorts political processes by how long the cause and the outcome take to unfold: some are like a tornado, fast in and fast out; others are like global warming, slow to build and slow to show. 

Incompetence. I don't mean Bishop and Watts are stupid. They're capable politicians. I mean institutional competence, which is knowing what the machinery takes to move. It's a specific skill, and on this evidence, it was missing.

Isn't this just about the election?

Watts himself said they didn't want to be making decisions this close to the election. Little says they didn't have the numbers. On the numbers, I honestly don't know. The only person claiming the coalition lacked them is Little, who is hardly neutral. ACT backed the process on the record. I couldn't find NZ First's position on the mergers at all.

The three-year term is an institution too. The Government launched in May a process it needed decided by September, ahead of a November election. It built a plan that could only fit one of two clocks, and it picked the wrong one.

Redrawing boundaries is engineering work

I've written before about the public service, and the argument is the same one: transforming large systems is closer to engineering than to policy or politics.

Redrawing councils means re-plumbing. Who owns the port? Whose plan governs the river? Who runs civil defence in the next flood? What happens to the statutory Māori seats on Bay of Plenty and Canterbury's regional councils, or to Rotorua's Māori ward? 

Having worked inside a minister's office, I understand why announcements get ahead of the plumbing. The announcement is the part politics rewards.

So, what now?

If the coalition is returned, Cabinet looks at the proposals in early 2027, with media reporting a legislation target of mid-2027. If not, the whole programme is open to review. Councils, meanwhile, have spent an estimated $9 million on proposals that are now sitting in a drawer.

The case for simpler local government hasn't gone anywhere. But if the next government wants it, it should start by asking the engineers how long it takes and then believes them.

*Natalia Albert is a political scientist specialising in New Zealand politics. She is a regular contributor on RNZ Nights with Emile Donovan, Afternoons with Jesse Mulligan and writes for interest.co.nz. Less Certain is her weekly newsletter. Albert stood as a TOP candidate in the 2023 election.