Build things again. Why New Zealand should finish what Muldoon started.

An argument for industrial ambition, generation at scale, and the second smelter we never built.

The country that forgot how to build

In 1980, Robert Muldoon stood in Parliament and announced that New Zealand would build a second aluminium smelter at Aramoana, near Dunedin. It would be powered by a series of new hydroelectric dams on the Clutha and lower Waitaki. The full package would have added roughly 1,200 megawatts of generation to the national grid, supported around 2,000 direct jobs in Otago, and anchored a southern industrial economy built on the one competitive advantage this country has always had in abundance — clean, firm, renewable electricity.

It never happened. The smelter was cancelled in October 1982 after its Swiss partner Alusuisse walked away, aluminium prices collapsed, and a determined grassroots campaign — complete with a mock declaration of independence — turned Aramoana into a political liability. Only the Clyde Dam survived, downgraded from 612 to 432 megawatts, completed eleven years behind schedule, and feeding a national grid that never got the industrial customer it was designed for.

Forty-four years on, it is worth asking whether we were right. Not about Aramoana specifically — the site itself, a salt-marsh wildlife reserve at the mouth of Otago Harbour, was genuinely the wrong place — but about the instinct behind the project. The instinct that a small country at the bottom of the world, with abundant renewable energy and a shrinking industrial base, should use that energy to make things. The instinct that the government should think in forty-year horizons, not three-year electoral cycles. The instinct that we are capable of building, at scale, what we need.

Because if you look honestly at what New Zealand has become in the four decades since Think Big was abandoned, the answer is uncomfortable. We stopped building power stations. We stopped building factories. We stopped building the things that build things. What we built instead was a housing market, a tourism industry and a banking sector, and we told ourselves this was sophistication. It was not. It was a country slowly running down its industrial capital and replacing it with mortgage debt.

This essay argues for reversing that. Not by resurrecting every Think Big project, and not by building a smelter at Aramoana specifically, but by reclaiming the underlying proposition — that the state should use its unique ability to take forty-year risks to build the generation capacity and industrial anchors that private capital alone will never finance. The smelter is a case study, not the point. The point is that we should be building things again.

Where we actually are

Begin with the facts on the ground, because any argument about the future has to start with an honest accounting of the present.

In August 2024, the Prime Minister declared an energy security crisis. Wholesale electricity prices spiked above $800 per megawatt-hour during winter. Major industrial users — Oji Fibre Solutions, Winstone Pulp, Methanex — curtailed production or shut down outright. Households faced the largest power bill increases in a generation. The Tiwai Point aluminium smelter voluntarily cut its electricity consumption by roughly 30% under a demand-response agreement with Meridian, reducing national demand by an estimated 330 gigawatt-hours over four months just to keep the rest of the grid viable.

This is not the situation of a country with surplus electricity. This is the situation of a country that has spent twenty years failing to build generation while demand quietly grew.

The numbers are clear. New Zealand currently generates around 43,000 gigawatt-hours of electricity per year. The Ministry of Business, Innovation and Employment forecasts demand to grow somewhere between 35% and 82% by 2050, driven by electrification of transport, industrial heat, and — increasingly — data centres. The Electricity Authority’s 2025 generation pipeline shows around 44 gigawatts of proposed new capacity, but 82% of it is intermittent (wind and solar), which means even if all of it got built, it would not provide the firm baseload power that industrial New Zealand runs on. Meanwhile, the country’s gas fields are in terminal decline, Huntly’s coal stockpile is what stands between us and winter blackouts, and every gentailer has quietly concluded that the rational strategy is to manage scarcity rather than build capacity.

This is the context in which the smelter debate has to be held. The question is not whether New Zealand can afford to add industrial load to the grid. The question is whether New Zealand can afford to keep pretending that a 21st-century economy can be built on a generation fleet designed in the 1970s and a political culture that treats every new dam as a moral failure.

The case against, stated fairly

Let me state the strongest version of the case against building a second smelter, because the argument that follows is not worth making unless the opposition is taken seriously.

The case against goes like this. Aluminium smelting is a commodity business with thin margins, dominated globally by Chinese producers with access to subsidised coal power. The Tiwai Point smelter has threatened closure four times in fifteen years, each time extracting a lower electricity price from Meridian, which means the existing smelter is already being subsidised by every other electricity consumer in the country. A second smelter would double that exposure and double the hostage-taking. Globally, primary aluminium capacity is shrinking outside China for structural reasons — smelters need electricity under US$40 per megawatt-hour on 10-to-20-year contracts to be viable, and in most Western markets, including New Zealand’s, wholesale prices now routinely exceed that threshold. Data centres are willing to pay US$115 per MWh for the same power. In a free market, the electricity will go where it earns the highest return, and that is not a smelter.

Further, the case continues, the dams required to power a second smelter are politically and environmentally impossible. The Kawarau Gorge scheme from the 1980 plan would flood one of the most iconic river gorges in the world and destroy the adventure tourism industry that Queenstown was built on. Lower Waitaki dams would drown braided riverbeds that are the last refuge of half a dozen endangered species. The Resource Management Act, Treaty of Waitangi obligations, and simple democratic opposition make these projects functionally un-buildable, and any government that tried to force them through would be voted out.

And finally, the jobs numbers are smaller than advocates claim. Tiwai Point employs roughly 1,000 people on 572 megawatts of firm power. A second smelter of similar scale would add maybe 2,000 to 2,500 direct jobs, including multipliers. That is not a transformative number for a country of five million people. For the same electricity, you could run twenty hyperscale data centres, or a dozen green hydrogen plants, or electrify a third of the national vehicle fleet.

This is a serious case. It is, broadly, the consensus case in Wellington and in the executive suites of the gentailers. It is also, I will argue, wrong — not because its facts are wrong, but because it answers the wrong question.

The right question

The right question is not “what generates the highest financial return per megawatt-hour in 2026?” The right question is “what kind of country do we want New Zealand to be in 2065?”

Those are very different questions, and the honest reason we keep getting the wrong answer is that we keep asking the first one.

A data centre, to use the fair comparison, employs roughly one person per megawatt once built. A 40 MW facility typically runs with around 45 staff; hyperscale facilities over 100 MW operate with fewer people per megawatt because of automation. The jobs are mostly technical, mostly based in Auckland, mostly short-contract or contractor positions, and mostly filled by people who could earn similar money in Sydney or San Francisco — and often do. The building phase lasts maybe two years and then the facility runs for its 15-to-25 year design life with a skeleton crew before being superseded by whatever Nvidia releases next.

An aluminium smelter employs roughly 1.75 people per megawatt during steady-state operation — Tiwai’s 1,000 staff on 572 MW of draw — and a further multiplier of two to three times that in supporting industries — smelter pot reliners, cathode suppliers, freight operators, dry dock services, engineering firms, ship chandlers, trucking companies, the whole industrial ecosystem that grows up around a facility that has to run continuously for forty years. These are unionised jobs on collective agreements, paying well above the national median wage, concentrated in regional economies that would otherwise depopulate. Tiwai alone is estimated to generate around 6.5% of Southland’s GDP.

For the same electricity as Tiwai uses, you could build roughly five hyperscale data centres employing perhaps 250 people between them. Or you could have the smelter and its 2,500+ direct and indirect jobs. The arithmetic is not complicated; it is just unfamiliar to a policy class that has spent thirty years being told that the future is in services.

It is worth noting what the alternative uses of this electricity have in common. A subsea cable to Australia — occasionally floated as a way to “monetise” surplus New Zealand generation — would send the electrons to Newcastle so that Australian smelters, Australian data centres, and Australian industrial users could employ Australian workers to turn them into economic value. New Zealand would collect a transit fee. It is the electron equivalent of exporting logs instead of milled timber, or raw milk powder instead of finished cheese — the value-added activity, the skilled employment, and the industrial capability all happen at the other end of the supply chain. This is a mistake New Zealand has been making with one commodity after another for 150 years, and each time the country has got poorer relative to the buyers doing the processing. The proposition here is to stop making that mistake with electricity. If we build the generation, we should also build the industry that consumes it, and we should capture the jobs.

And even that comparison understates the difference, because it counts only direct jobs and misses the single most important thing a smelter does for a regional economy — it creates a deep supply chain. Economists call this backward linkage depth — the number of tiers of domestic suppliers an industrial facility sustains. A smelter has three to five tiers of them. It imports around 600,000 tonnes of alumina a year, plus petroleum coke and cryolite and pitch, all of it moving through a port, a rail line, a trucking fleet, a warehousing system and a customs operation. It produces 600,000 tonnes of finished metal that has to be cast, packaged, and shipped. It needs pot reliners on a rotating schedule — a specialist trade that exists in New Zealand essentially because Tiwai exists. It needs refractory specialists, cathode suppliers, high-amperage electricians, process control engineers, industrial gas suppliers, specialty chemicals operators, cranes, riggers, boilermakers, welders and fabricators. Each of these firms then services the rest of the regional industry as a byproduct, which is why Southland still has heavy engineering capacity that most New Zealand regions have lost. NZAS estimates its total employment footprint at around 3,200 jobs in Southland, against roughly 1,000 direct staff. That is a 3.2x multiplier, which is normal for heavy industry.

A data centre has one tier of backward linkage. It buys electricity. Every other input — the servers, the chips, the memory, the switches, the cooling equipment — is manufactured in Taiwan or Vietnam or China or Korea and arrives as finished goods in a small number of containers. The capital expenditure is enormous, but almost all of it flows offshore to Nvidia and TSMC and the Asian assemblers. There is no ongoing industrial supply chain because the facility does not consume anything physical that has to be made locally. Academic estimates of data centre employment multipliers typically run 1.2–1.5x, and most of that multiplier is 18-to-24 months of construction employment that disappears once the facility is operational. In steady state, a data centre is effectively a one-to-one employment proposition. Five hyperscale data centres employing 250 people directly will support perhaps another 100 jobs in the surrounding economy. A smelter employing 1,000 people directly supports another 2,200.

So the real comparison is not five data centres at 250 jobs versus a smelter at 1,000. The real comparison is 350 total jobs versus 3,200 total jobs for the same electricity. That is a nearly 10x difference in employment per megawatt, and the smelter jobs are distributed across a regional economy — the trucking company in Invercargill, the engineering firm in Bluff, the refractory installer in Dunedin — rather than concentrated in a single windowless building next to a motorway.

And this is the honest mistake at the heart of the case against. We have convinced ourselves that making things is somehow less sophisticated than serving things, that an economy of cafés and consultants and cloud migrations is more advanced than an economy that makes metal and pulp and steel. It is not. It is just more fragile. Every country that has tried to replace its industrial base with services alone has ended up with the same set of problems — regional decline, wage stagnation, a trade deficit it cannot close, a population that cannot afford houses in the cities where the service jobs are — and New Zealand is a decade or two further down that road than most.

The smelter, in other words, is not an economic curiosity. It is one of the few industrial anchors we have left, and it is the reason a town like Bluff still exists as anything other than a ferry terminal.

Why governments, not markets

The reason this argument has to be made to government rather than to gentailers is simple — the kind of investment required to build a second smelter and the generation to power it is not the kind of investment any private actor will finance in the current market structure.

A private investor looking at a new smelter in 2026 is looking at a $5–8 billion capital commitment, a construction period of five to seven years, and a 40-year operating horizon in which the first decade will probably lose money before the asset begins to pay back. The electricity contract alone would need to lock in around $40/MWh for twenty years to make the economics work — which means someone, somewhere, has to take the price risk on two decades of wholesale electricity. In New Zealand’s current market, no gentailer will write that contract voluntarily, because they are being rewarded by shareholders for managing scarcity, not building abundance. The Electricity Authority and Commerce Commission’s Energy Competition Task Force, established in direct response to the August 2024 price spike, has acknowledged that ownership of stored hydro is concentrated in a small number of gentailers, and that the incentive to manage stored water for price rather than volume is a structural feature of the current market design. This is not a conspiracy theory — it is rational behaviour in a market where the players own both the generation and the retail, and where scarcity rents exceed volume revenue.

The market, in short, is working exactly as designed. It is producing the outcome that makes sense for the companies in it. That outcome is not the outcome that makes sense for the country.

This is the classic case for state action, and it is the case Muldoon — for all his faults — understood better than any prime minister since. The state can take forty-year risks. The state can underwrite electricity prices over periods longer than any private investor will touch. The state can issue bonds at sovereign rates to build infrastructure whose returns show up as regional GDP, tax revenue, and trade balance improvements over decades rather than as shareholder distributions next quarter. The state can, in effect, act as the buyer of last resort for the industrial demand that the private market will never commission on its own because the coordination problem is too large.

The Clyde Dam, for everything that was wrong with it, is a case in point. It was built against intense opposition, through emergency legislation, by a government that was widely hated by the time it was finished. It cost three times its initial estimate. It has now been generating roughly 2,100 gigawatt-hours a year of renewable electricity for thirty-three years, with at least another fifty years of operating life ahead of it. The cumulative wholesale value of the electricity it has produced is somewhere between $4 and $6 billion, against a real construction cost of around $2 billion in today’s money. The Muldoon government got essentially every detail wrong, and the dam is still, measured over its actual lifetime, one of the best investments the New Zealand state has ever made. That is the nature of physical infrastructure. It endures, and its returns compound across generations in a way that no service-sector investment ever will.

A second smelter plus the generation to power it would be the same kind of investment. Expensive, controversial, plagued by delays, subject to cost overruns, hated by roughly half the country for the duration of its construction. And, forty years from now, the single largest employer in the lower South Island, the anchor of a regional economy, the reason a steel-fabrication sector and a specialty-chemicals sector and a machining sector exist within New Zealand at all. The returns do not show up in next year’s budget. They show up in 2065, when your children are trying to build lives in a country that still has an industrial spine.

What we would actually build

If we take seriously the proposition that New Zealand should build a second smelter and the generation to power it, the question becomes — what, specifically, would we build?

Not Aramoana. The original site is a salt-marsh wildlife reserve at the mouth of Otago Harbour, and the Save Aramoana campaigners of 1980 were right that it should never be industrialised. The smelter plan that called for the destruction of Te Ngaru village and the displacement of the oyster-catcher colonies was a site selection failure, and nothing in 2026 makes it a better idea than it was in 1980. A serious second-smelter proposal would look at sites with existing industrial zoning, adequate port access, and minimal ecological conflict — candidates include brownfield sites around Marsden Point, an expansion corridor near the existing Tiwai facility, or a new site near Port Taranaki. Each has its own political complications; none involves flooding a wildlife reserve.

Not the Kawarau Gorge dam. The 1980 Otago hydro scheme was drawn up by engineers who had not yet absorbed that some landscapes are economically more valuable unflooded than flooded. Queenstown’s adventure tourism industry now generates several billion dollars a year and employs more people than the entire 1980 hydro-and-smelter package would have. A rational modernisation drops the Kawarau immediately.

What is left, though, is substantial. The 2004 study Waters of National Importance, commissioned by the Ministry of Economic Development and prepared by East Harbour Management Services, identified around 2,460 megawatts of technically viable undeveloped hydroelectric capacity. The larger candidates included Waitaki (up to 590 MW), Clutha (up to 410 MW), Grey (350 MW), Waiau (235 MW), and Ngaruroro (135 MW). Meridian’s North Bank Tunnel concept on the Waitaki, which was consented in 2008 before being shelved when demand growth stalled, would add 200 to 285 MW without a new dam, using a 34-kilometre diversion tunnel on the north bank. The Lake Onslow pumped-hydro scheme, cancelled by the coalition government in 2023, would add 1,500 MW of dispatchable capacity and 5,500 gigawatt-hours of storage — more than the nominal full storage of all existing New Zealand hydro lakes combined. A private consortium including former Meridian and Transpower chief executives has since revived the scheme for fast-track consideration, which tells you something about what the industry actually thinks when it does not have to posture for shareholders.

Add to this — a geothermal build-out in the central North Island, where the resource is abundant and the environmental footprint modest (geothermal now provides around 18% of national electricity and could realistically double); a serious offshore wind programme in the South Taranaki Bight, where the resource is world-class and the grid connection is short; and a pumped-hydro storage programme that uses old mine pits (the Macraes pit alone could provide around 0.5 GWh of storage) and existing lakes as upper reservoirs to firm the intermittent build.

The programme, in total, could realistically add 4,000 to 5,000 megawatts of new firm generation by 2040, roughly doubling existing hydro capacity, with the smelter providing 600 to 700 megawatts of anchor demand and the remainder absorbed by electrification of transport, industrial process heat, and — yes — a managed amount of data centre growth. The point is not that the smelter takes all the electricity. The point is that the smelter provides the demand certainty that makes the generation economics work for everybody else.

The consent problem

This is where the argument gets hardest, and where I want to be most honest. The Resource Management Act, Treaty of Waitangi obligations, and four decades of accumulated case law have made it essentially impossible to build large infrastructure in New Zealand at a reasonable cost or within a reasonable timeframe. The Clyde Dam took eleven years and cost three times its estimate. The Waterview Tunnel took seven years. The City Rail Link — a six-kilometre tunnel through central Auckland — is now running to more than a decade of construction and north of $5 billion. A country that builds this slowly cannot build its way out of anything.

There are two honest responses to this. The first is to say that the consent system is working exactly as intended — it protects landscapes, cultural sites, and communities from overreach by powerful actors, and the cost of that protection is that we build less — and that the right trade-off is to accept we will build less. This is the current consensus position, and it is coherent, but it means accepting a country of gradual industrial decline.

The second response is to say that the costs of not building are now higher than the costs of building, and that the system has to be reformed. This is uncomfortable because it involves trade-offs that cost real people real things, and because the history of fast-track legislation in New Zealand is not pretty. The Clutha Development (Clyde Dam) Empowering Act 1982 is the precedent nobody wants to invoke, and with reason — it overrode a High Court decision, bulldozed landowner rights, and was repealed as soon as the National government that followed had a chance.

And yet. The coalition government’s 2024 Fast-Track Approvals Act, for all its faults, is an attempt to acknowledge that the cost of the status quo has become unsustainable. A serious infrastructure programme would go further. It would establish a permanent pathway for projects of national significance, with mandatory engagement with iwi, binding environmental mitigation, and transparent cost-benefit thresholds — but with hard timeframes and a bias toward building rather than not building. It would fund the Environmental Protection Authority properly so that consents could be assessed quickly rather than slowly. It would settle the ongoing Treaty claims over water rights rather than litigating them project by project for another forty years. And it would accept that some consent outcomes will be wrong in individual cases, because a country that never makes consent mistakes is a country that never builds anything.

Iwi partnership in particular is not optional, and should not be framed as an obstacle. The most successful infrastructure projects in New Zealand over the last twenty years — Ngāi Tahu’s commercial infrastructure investments across the South Island, Tainui Group Holdings’ Ruakura inland port in Hamilton, Ngāti Whātua Ōrākei’s role in Auckland waterfront development — are all projects where iwi participated as commercial partners, not just consultees. A second-smelter-plus-generation programme in the South Island cannot be built without Ngāi Tahu, Ngāti Whātua, Kāi Tahu ki Otago, and the iwi of Southland as partners in the revenue, not just the consent process. Done properly, this is not a delay mechanism; it is a stability mechanism. Projects with iwi as commercial partners tend to survive changes of government in a way that projects without them do not.

The forty-year ledger

Here is the calculation that I suspect we have been avoiding for forty years.

A second-smelter-and-generation programme, built over fifteen years from a 2026 decision, would cost somewhere in the order of $25 to $35 billion in today’s money. That is a staggering sum. It is also roughly what the current government is spending on road and rail projects over the same period, with substantially less to show for it at the end.

Against that cost, what does the forty-year ledger look like?

On the electricity side, 4,000 to 5,000 MW of new firm capacity would generate roughly 20,000 to 25,000 gigawatt-hours per year. At a conservative long-run wholesale price of $100/MWh, that is $2–2.5 billion per year in electricity value, or $80–100 billion over forty years. Even if half of that goes to the smelter at a below-market industrial rate, the other half flows into the general grid at market prices, substantially reducing wholesale costs for every other consumer in the country. The hidden subsidy that electricity users currently pay to sustain gas peaking and coal backup disappears, because there is enough firm renewable capacity to displace the thermal fleet.

On the smelter side, a 600,000-tonne-per-year facility at current aluminium prices ($2,500–2,800/tonne) produces $1.5–1.7 billion per year in export revenue. That is roughly equivalent to the entire wine industry, from a single facility. Over forty years, that is $60–70 billion in exports, essentially all of it in hard currency, most of it flowing into a regional economy that is currently among the poorest in the country.

On employment, 2,000–2,500 direct jobs plus a multiplier of 2.5–3x gives you 5,000–7,500 total jobs sustained for forty years. At an average salary of $90,000 (reasonable for skilled industrial work), that is $450–675 million per year in wages, the vast majority of it in regional New Zealand. Over forty years, that is $20–25 billion in wage income, most of which is spent locally, most of which flows back to the tax system, and all of which represents people who are not on the unemployment benefit, not depending on Working For Families, not being forced to move to Australia for work.

On tax revenue, a facility of this scale, combined with its supporting industries, would contribute somewhere in the order of $400–600 million per year in combined corporate tax, PAYE, and GST. Over forty years, that is $16–24 billion in direct fiscal contribution. The Crown’s own equity stake in the generation assets would add another revenue stream on top of that.

Against the capital cost of $25–35 billion, the ledger shows forty-year gross benefits in the order of $180–220 billion, of which perhaps $60–80 billion flows directly to the Crown through dividends, taxation, and avoided welfare spending. These are rough numbers, and any serious proposal would need them stress-tested. But the order of magnitude is not in doubt. This is not a marginal investment. It is, on a forty-year view, one of the best investments any government in the OECD could currently make.

And every year we do not make it, the opportunity cost compounds. Every year, another tranche of school leavers finds there is no manufacturing career available in the region where they grew up. Every year, another cohort of engineers and electricians moves to Queensland because Queensland is building things and we are not. Every year, another industrial supplier shuts down because the volume is not there to sustain it. The consequence of not building is not a stable state. The consequence of not building is a slow, compounding erosion of the industrial capability required to build anything at all.

The climate case

There is one more argument, and it is the one that was not available to Muldoon in 1980. New Zealand has some of the cleanest electricity in the world. Tiwai Point, on Manapouri hydropower, produces aluminium with roughly one-tenth the carbon intensity of Chinese coal-powered aluminium. As carbon border adjustments come into force across the EU, UK, and eventually the US, low-carbon aluminium is commanding premiums of $20–$150 per tonne over commodity aluminium, and that premium is forecast to rise sharply over the next decade as scope-3 emissions accounting becomes universal.

If the world is going to make 100 million tonnes of aluminium a year — and it is, because aluminium is the structural metal of decarbonisation, essential for solar frames, wind turbine nacelles, EV bodies, and transmission conductor — then the question is where that aluminium is made. Every tonne made in New Zealand on hydropower is a tonne not made in China on coal, and the global emissions difference is roughly 15 tonnes of CO2 per tonne of aluminium produced. A 600,000-tonne smelter on renewable New Zealand electricity would therefore avoid around 9 million tonnes of CO2 emissions globally per year — roughly 12% of New Zealand’s entire annual gross emissions — compared to the counterfactual of that aluminium being produced in China.

This is a genuinely unusual situation. It is rare for a New Zealand industrial investment to be a net benefit to global emissions at the scale of 12% of the country’s total output. But aluminium is one of the few industries where this is true, because the emissions are overwhelmingly determined by the electricity source, and because global production currently averages roughly four times the emissions intensity of New Zealand production. The climate case for building a second smelter in New Zealand, powered by new renewable generation, is therefore considerably stronger than the climate case for most industrial projects, and is substantially stronger than the climate case for almost any available alternative use of the same electricity.

What this actually requires

If we are serious, the policy programme writes itself. It is not the programme of any currently existing political party, but it is coherent and achievable on a fifteen-to-twenty-year timeframe.

First, a National Generation Build-Out Authority, sitting above the existing state-owned enterprises, with a statutory mandate to commission 4,000 to 5,000 MW of new firm capacity by 2045. This body would hold Crown equity in new generation, issue long-duration bonds at sovereign rates, and have the statutory power to commission generation directly where the private market fails to do so. Its operating model would be closer to Australia’s Snowy Hydro or the Canada Infrastructure Bank than to the gentailer model that has produced twenty years of under-investment. Its performance would be measured in megawatt-hours delivered, not in dividend yield.

Second, a Second Smelter Development Authority, empowered to negotiate a forty-year power purchase agreement with the generation authority, select a site, seek binding iwi partnership, and manage the capital raise in partnership with a global smelting operator — likely Rio Tinto, Alcoa, or Emirates Global Aluminium, all of whom are currently searching for low-carbon smelter sites. Crown equity would be retained at around 25–30% to ensure that the jobs, the technology transfer, and the supply-chain benefits are locked to New Zealand and do not evaporate at the first offshore ownership change.

Third, a Projects of National Significance Act, replacing the current fast-track regime with a permanent pathway that guarantees consent decisions within 24 months of application for any project meeting defined thresholds, subject to mandatory iwi partnership, mandatory environmental mitigation, and full judicial review within the statutory timeframe. This is the single most difficult piece of the programme, and the one that will attract the most opposition, and it is also the one without which none of the rest can happen at pace.

Fourth, a Sovereign Industrial Bond programme, modelled roughly on the Canadian Pension Plan Investment Board, to finance the build-out over fifteen to twenty years at sovereign rates. The Crown’s ability to borrow at around 4–5% nominal on thirty-year maturities is one of the most underused fiscal tools in the country, and it should be deployed for infrastructure whose returns stretch across generations rather than for operating expenses that should be funded from taxation.

Fifth, and perhaps most importantly, a National Industrial Apprenticeship Programme, starting five years before commissioning, training the pot reliners, high-voltage electricians, refractory specialists, port engineers, and process-control technicians that the smelter and its supporting industries will require. We do not currently have enough of any of these skills in New Zealand. The programme would partner with polytechnics and Ngāi Tahu training arms to ensure that the employment benefit flows to people who grew up in the region rather than to fly-in contractors.

The objection I want to take seriously

The strongest objection to everything I have argued is not economic. It is historical. It goes like this — Think Big was tried, and it did not work. The Motunui synthetic petrol plant was a fiasco. The New Zealand Steel expansion was assessed as value-destroying at the time of approval and the assessment turned out to be correct. The national debt that Muldoon accumulated financing these projects took fifteen years of painful reform to unwind. What makes you think we can do it better this time?

The answer has two parts. The first is that Think Big was not a single undifferentiated policy — it was a portfolio of projects, some of which worked and some of which did not, and the aggregate return on the portfolio is considerably better than the standard political narrative admits. The Clyde Dam, the Motunui methanol plant, the Kapuni ammonia-urea plant, the Marsden Point expansion, the third Tiwai potline, and the New Zealand Steel expansion together have generated cumulative economic activity in the order of $50–100 billion over their operating lives, against capital costs of perhaps $20–30 billion in today’s money. The failures were real; the successes were larger. Think Big was not a disaster. It was an uneven success that has been retrospectively reframed as a unified failure for reasons that have more to do with the politics of the 1984 reforms than with the actual economic record.

The second part of the answer is that we have learned things in the intervening forty years. We know that governments should not pick specific technology winners (Motunui’s synthetic petrol was obsolete the day it opened). We know that forcing through consents with emergency legislation produces legal fragility and political backlash. We know that ownership structures matter — the Crown should hold equity, not just guarantee debt. We know that iwi partnership is essential, not optional. We know that skills programmes have to precede construction, not follow it. We know that long-duration financing needs to match long-duration assets. We know that energy markets need to be designed to reward abundance, not scarcity. None of this was well understood in 1980; most of it is well understood now. A Think Big programme executed with 2026 institutional knowledge is a different proposition from Think Big executed with 1980 institutional knowledge.

The question is not whether we are capable of doing this better than Muldoon did. Of course we are. The question is whether we are capable of doing it at all.

The country we choose

Two visions of New Zealand in 2065.

In the first, the trajectory of the last twenty years continues. Tiwai Point closes around 2044 when its current contract expires, because Rio Tinto’s global strategy no longer has room for a 1970s-era smelter in a high-cost jurisdiction. Southland enters structural decline; Invercargill’s population halves; Bluff becomes a summer-tourist village and a commuter suburb for a shrinking aluminium workforce that no longer exists. The national electricity system is 85% renewable by capacity but periodically blacks out in winter because nobody has built enough firm generation to cover dry years, and the gas field is empty. The industrial base is gone. Manufacturing is 5% of GDP and falling. Per-capita income has fallen to 65% of Australian levels, and the migration outflow that started in the 2020s has continued uninterrupted for a generation. The country is a pleasant retirement destination for wealthy retirees, a pleasant holiday destination for Australians, and a place where young people leave. The things we still make, we make in small volumes for niche export markets. The big decisions about New Zealand’s future are made offshore, in boardrooms in Sydney and Singapore and Shanghai.

In the second, a government somewhere between 2026 and 2030 decides that this is not good enough. It commits to a twenty-year industrial programme on the scale of Think Big but executed with the institutional lessons of the intervening forty years. It builds 4,500 megawatts of new firm generation. It commissions a second smelter, not at Aramoana, in partnership with Ngāi Tahu and a global smelting operator. It trains a generation of skilled industrial workers. It accepts that some projects will be controversial and some will be imperfect, and it builds them anyway. By 2045, the programme is delivering. By 2055, Southland and Otago are among the wealthiest regions in the country, with median incomes above the national average and net inward migration. By 2065, New Zealand is the world’s largest producer of low-carbon primary aluminium, exports $8 billion a year in basic metals, has rebuilt a machining and fabrication sector off the back of the smelter’s supply chain, and has closed the per-capita income gap with Australia for the first time since 1973. The country makes things again. Young people stay.

The case for the first vision is that it is the path we are already on and it is easier to continue than to change direction. The case for the second vision is that it is the country most New Zealanders say they want when you ask them, and nothing in the physical, economic, or political reality of New Zealand makes it impossible. What makes it hard is that it requires the political class to think on a forty-year horizon rather than a three-year horizon, and it requires the voting public to accept short-term disruption for long-term benefit. Neither of those things is easy. Neither is impossible.

The Aramoana smelter never got built. It should have been built somewhere else, in a different form, on a different schedule. The mistake of 1982 was not that the smelter was cancelled; the mistake was that we took its cancellation as a lesson that ambition itself was the problem, and we have been living with the consequences of that lesson for forty years. The consequences are visible in every declining regional town, every young engineer’s Brisbane departure lounge, every winter electricity price spike.

We can keep learning the wrong lesson. Or we can build things again.

Your turn

Is there an argument for building things again in New Zealand — or against it? Where did this essay get the call right, and where did it get it wrong? Add your view in the comments.