A plan for New Zealand · September 2026
26,000 came home; a net 37,700 did not. Most went to Australia, where the same work pays about half as much again. New Zealand has grown by adding people, not by making each person more productive. Wages stalled, houses cost ten times income, and a quarter of adults can’t get a doctor’s appointment in time. Six moves fix the causes, not the symptoms. Every one has a number, a cost and a trade-off.
Why we’re behind
The OECD, the IMF, Treasury and the Reserve Bank have said the same thing in different words this year. Nobody in Parliament has turned it into a plan. This is the diagnosis in three sentences.
Since 2013 the population has grown fast on migration while output per hour has barely moved. More people in the same houses, hospitals and roads feels like growth in the headline and like decline at the kitchen table.
The OECD calls ours one of the most housing-biased tax systems in the developed world. Buying an existing house moves money from one New Zealander to another; the damage is that bidding up scarce land has paid more reliably than building homes, plant or firms. Capital per hour worked is about half Australia’s. KiwiSaver is 32% of GDP; Australian super is 133%.
We spend more of GDP on infrastructure than most rich countries and rank near the bottom for what we get. Gas output has halved since 2016. Councils now face a $48bn water bill and rates rose 14% in a year.
The result
Each dot is 100 people. The pale ones are the 6,000 a year we used to lose before Covid. Every orange dot is on top of that. Two in five of them are aged 18 to 30, and the record was 55,300 in 2024.
The six moves
Each move starts with the number that proves the problem, then three things we would actually do, why they work, what they cost, and what they cost you. Where the other parties stand is listed so you can compare.
New Zealand’s grocery, banking, building-supply and electricity markets are shared out between two to four companies each. Four years of reports have not changed that. Competition law with teeth will.
Where a market study finds entrenched concentration (two firms over 70%, or four over 85%), entry has failed, and the Commission shows the remedy’s benefits exceed the lost scale and the cost of doing it, it can order divestment on a published three-year clock. Concentration triggers the study; harm and a net benefit justify the order. Australia is legislating the same power. The threat is what makes the wholesale-access rules work.
Independent retailers and factories can then buy electricity at the price the gentailers charge themselves. Bespoke contracts that finance new generation stay allowed, but disclosed. By the end of the first term the Authority and the Commission report on four tests: whether independents get hedges on the same terms, gentailer retail margins against independents’, forward-market liquidity, and the independents’ share. If the tests show harm and separation passes a net-benefit test, it is ordered under the power legislated in year one. A raw wholesale-to-retail spread is not the trigger: about 40% of a household bill is generation and 35% is networks, so the spread alone proves nothing.
Full open banking, account numbers you can take with you like a phone number, and a properly capitalised Kiwibank as a challenger, funded by KiwiSaver providers rather than the taxpayer.
Prices fall when a new entrant can actually get in. The Commission has said for four years that the grocery duopoly is unchanged and that banking is “a stable oligopoly.” Reports have not moved prices. A credible break-up power does, even when it is never used.
$30m a yearFor Commission capacity and the hedge exchange. No state supermarket, no buy-back of a bank.
Forced divestment can lose some scale efficiency and takes years to litigate. Pre-announcing the first-term tests keeps investment flowing while the market is given a real chance to fix itself.
Greens: buy 120 supermarkets for a state chain, $2.8bnNZ First: split the gentailers now; buy back BNZNational: keep the current rulesLabour: a split is “not off the table”; details pending
The 2023 Grocery Industry Competition Act created a regulator and wholesale-access rules. The Commission’s June 2026 report found the duopoly share unchanged at 82% and that 21 overseas chains it approached declined to enter. The missing piece is a credible structural remedy.
Building supplies: automatic recognition of Australian-certified products so a Bunnings or a builder can import at Australian prices without a second approval.
Zoning reform has started to work; Auckland’s upzoning cut rents relative to the rest of the country. What stops the next 100,000 homes is not planning permission. It is that nobody will pay for the water, roads and schools up front.
Seeded with $2bn of Crown capital, lending up to $20bn over ten years against targeted rates and developer contributions. The Local Government Funding Agency already lends councils cheap money, but against the council’s own balance sheet, inside debt limits most growth councils have reached. The bank lends to a ring-fenced vehicle against a targeted rate on the new lots, off the council’s books, with the Crown buffer as first loss and one standard contract instead of a bespoke deal each time. The people who get the new suburb pay for it over 40 years, instead of every ratepayer paying for it now.
As of right, no hearing. This locks in and extends Going for Housing Growth so the rules stop changing with every government.
We have done it for three: Tauriko West, Rotokauri and eastern Porirua, 1,000 homes each, every pipe and school place estimated at stated unit rates. It is a concept estimate on our own assumptions, not a tender; the model is below. Every council gets the same template, plus standard designs for local water and streets, repeat procurement across councils so contractors price a pipeline rather than a one-off, and a published unit-cost comparison for every council’s growth works. Cheaper finance for expensive infrastructure is only half a fix.
When supply can respond, scarcity rents fall. The peer-reviewed Auckland estimate is that rents were about 23% lower after eight years than they would otherwise have been: a modelled counterfactual, not a fall in rents, but substantive evidence that widespread upzoning works. The binding constraint now is infrastructure finance: councils near their debt limits cannot fund pipes, so land that is zoned still does not get built.
$2bn capital, onceOn the Crown balance sheet, repaid by targeted rates. Plus about $182m of Crown capital for the three proof-of-method suburbs over 2030–34. Bank operating costs and any calls on the guarantee are not yet costed; the register says so.
Targeted rates raise the cost of a new home in a new suburb. That is honest pricing; it is still far cheaper than the housing shortage. Density brings construction disruption to existing streets.
National: 5% deposits for households earning up to $300,000; rates cap from 2029Labour: restore KiwiBuild-style programmes; CGT on investment propertyOpportunity: 1.75% land value tax to cut land prices 10–15%Greens: public housing build and rent controls
How the bank works: it issues Crown-guaranteed bonds and on-lends to councils or water entities against a targeted rate on the new lots. The $2bn is a first-loss buffer, not the lending pool. If development stalls, the rate still attaches to the land; what the land cannot carry falls on the buffer before it falls on the Crown. Construction overruns stay with the council and developer under the contract, not the bank. The cost model below shows what the council share works out to per new home.
Councils face a combined $48bn bill for water under Local Water Done Well. Putting that debt into long-dated, ring-fenced vehicles with a Crown-backed lender is the difference between 9% and 3% rates rises.
The Infrastructure Funding and Financing Act 2020 already allows targeted-rate financing. It has been used a handful of times. The bank makes it the default.
You cannot fix hospitals while the front door is shut. Free visits to a doctor you cannot get an appointment with are worth nothing. We fund the people first, then the price.
Multi-year contracts for teams of GPs, nurse practitioners, pharmacists and mental-health workers, paid on access and continuity, not on visit counts. First 40 areas by the end of year one.
Five years in New Zealand publicly funded health, public or private practice, and the loan is gone. A retention payment priced against the pay gap, worth tens of thousands of dollars to a new graduate. It is not yet proven the best one: for two years it runs alongside targeted retention payments and supervision support, and we keep whichever retains more clinical years per dollar. Plus 200 extra supervised training places a year and faster recognition of overseas qualifications.
Your referrals, results and medicines in one place you can see. Health NZ stays as it is for a full term; every dollar goes to clinicians, not org charts.
Early primary care keeps people out of emergency departments; Wellington’s ED met its six-hour target 58% of the time this year. Team-based care uses scarce clinicians better. Retention is cheaper than recruitment: a bonded graduate costs a fraction of an overseas hire.
$450m a year$350m for access and workforce, about $100m for loan write-offs at expected uptake. The largest single item in our operating plan.
Training takes years, so the seven-day standard starts in 40 areas and scales. Access targets must be contracted alongside quality so nobody is rushed through.
Labour: three free GP visits a year, free prescriptions, funded by a CGTNational: $5.8bn new health operating funding over the forecast; regional devolutionACT: more private provision
Capacity and price are not rivals. The Health Survey finds 25.5% of adults faced a time barrier to a GP and 14.9% a cost barrier. Labour’s Medicard subsidises visits into a system that already cannot see a quarter of adults in time; without more clinicians it lengthens the queue. We keep every subsidy that exists and add capacity first.
The Medical Council’s 2024 survey shows 73% of New Zealand graduates still practising here ten years after registration, 82% for the most recent cohort with ten years’ data, while three in four overseas-trained doctors have left within ten years. The write-off is aimed at the quarter who go and the overseas doctors who stay only briefly. It will spend money on some who would have stayed anyway, which is why it is trialled against alternatives rather than assumed.
A data centre is the new aluminium smelter: a great deal if it brings its own power, a disaster if it takes yours. We are not against them. We are against your bill paying for them.
A new load contracts new firmed generation at least equal to its peak, or a mix of new supply and demand it can shed at peak and in a dry year, assessed together on location, timing, firming and network cost. It pays its own grid connection and water. In return: a consent decision within 12 months and no moratorium. Same rule for a smelter, a data centre or a hydrogen plant. Self-sufficiency is not the goal; not raising everyone else’s bill is.
Statutory 12-month decisions, national grid corridors designated once, and a standing dry-year reserve: Huntly kept ready and imported LNG as a bridge until enough geothermal and storage is built.
The public hedge exchange from our competition plan, bills that show what the power actually cost, and insulation, solar, hot-water storage and batteries financed where the full cost pays back. Exported power is paid what it is worth at the time and place it is supplied, and the Electricity Authority is directed to enforce that, not just recommend it.
Generation, lines and firming are complements; building any one without the others does nothing for bills. A data centre that funds additional firmed generation in the right place can lower system costs for everyone; one that does not raises them for everyone. Location, timing and transmission decide which, so the rule is written around all three. Ireland let data centres reach a quarter of its electricity use before writing rules. Australia is legislating an underwriting rule now.
$40m a yearFor consenting and grid planning. Private capital builds the generation. The LNG facility is already in the Government’s programme.
New lines and wind farms affect real communities; the corridor designation trades local objection for national certainty. Firming costs show up in bills before the savings do.
Greens: one-year pause on all data centres over 50MWNational: an “additionality test”, no numbers or date yetLabour: must fund own renewable supply and connectionACT: no new rules
Government figures put data centres at about 0.6% of national consumption now, rising to about 3% by 2030. Datagrid alone at full load is about 4–6%. Datagrid and the proposed Stratford centre together would use about 13% of current supply. Mercury’s chief executive has said the two “would probably consume our ability to build power stations over the next five to ten years.”
Eight consented South Island renewable projects total 857MW. The rule turns that pipeline into contracts.
We do not want more of your money. We want to change what we tax. Tax the gains people make on property and shares, and use every dollar to cut the tax on working and on building a business, in step, as the revenue arrives.
Businesses get rollover relief when they reinvest. This is what Treasury, the OECD and the IMF have all recommended. It replaces the bright-line test.
The cuts are staged behind the revenue by a rule, not a promise. No cut is made until two years of receipts have been certified. Each year’s step is sized to a sustainable revenue line, 90% of the three-year average of certified receipts, less the recurring cost of every cut already made. Receipts above that line fill a stabilisation account until it holds a full year of the cuts’ cost; only then does the next step go ahead. If receipts fall short, the account pays the difference; if it would run dry, threshold indexation pauses until it refills and the company rate holds. The independent fiscal council certifies the numbers and publishes the downside case each year.
Everyone in from our first Budget, contributions rising to 5% plus 5% within three years. Funds may hold unlisted New Zealand infrastructure; nothing is directed, and no fund is asked to bail out anything. Keep Investment Boost. Keep the 2025 overseas-investment liberalisation.
Wages follow capital per worker. Countries that tax land and capital gains and lightly tax work and business investment have deeper capital markets and higher wages. Australia’s super system is why its firms can raise money at home. The switch changes what pays: tax reform has to make productive investment more attractive, not only make the alternatives less attractive, which is why the company cut, rollover relief and Investment Boost travel with the gains tax.
$0 net, by ruleThe Tax Working Group’s 2019 projection for a comparable design was $0.4bn in year one rising to $5.9bn by year ten (Table 5.2), 1.2% of GDP long run; it is indicative, seven years old and, in the Group’s words, “volatile”. A company-rate cut to 25% costs roughly $2bn a year and wage indexation about $1bn, growing. On that path the first step is in 2033 and the full company cut lands in 2036, not on day one. Because the rule sizes cuts to a trailing average, about 58% of the decade’s receipts come back as cuts by 2039 and the rest fills the buffer and reduces debt; in the long run about nine dollars in ten come back. Neutral is the design and the rule, certified each year, not a guarantee in a sentence; the year-by-year profile and two downside cases are in the register.
A CGT adds compliance for investors and starts from a valuation day, and its revenue is slow and lumpy in the early years, so the tax cuts arrive gradually. The company-tax cut benefits foreign shareholders too. Exempting the family home keeps a real housing preference in the system, which the Working Group itself flagged; we chose the version that can be legislated. Treasury’s 2025 paper treats a gains tax as one part of a wider capital-income problem, not the whole answer. Compulsory KiwiSaver reduces take-home pay in the short run for people who were not saving.
Labour: 28% CGT on investment property only, spent on healthGreens: 2.5% wealth tax, 33% inheritance tax, corporate rate up to 33%Opportunity: land value tax funding a $19,400 citizen’s incomeNational: no new taxes; compulsory KiwiSaverACT: spending cuts, no new taxes
Why not a land value tax? It is the economists’ favourite and Opportunity’s policy. It hits cash-poor farmers and retirees every year whether or not they sell, which is why it needs deferral schemes and has never passed. A gains tax on realisation is what Treasury, the OECD and the IMF actually recommended, and it is the version that can be legislated.
Treasury, November 2025: it is “more likely a question of when, not if” the tax system responds, and the long-run answer “lies in taxing capital income.” The IMF listed a comprehensive CGT among revenue options in August 2026.
Australia’s company rate for small business is 25%. Matching it removes one reason for firms and founders to shift across the Tasman.
The reforms of the last three years are the right ones. The danger is that they get reversed at the next change of government the way the last lot were. We would keep them, fund the teachers to deliver them, and catch every child who falls behind.
A knowledge-rich curriculum, explicit teaching, and protected planning and coaching time for teachers. No new restructure of NCEA’s replacement until the first cohort has been through it.
40,000 places a year: groups of no more than four, at least three sessions a week for a term or more, in school hours, trained tutors and a paid lead teacher in each school, assessed termly, with learning-support specialists and attendance caseworkers attached. The meta-analysis behind this (Nickow, Oreopoulos and Quan: a pooled effect of about 0.29 standard deviations, larger for trained tutors, younger children, three or more sessions a week and in-school delivery) is for programmes that look like that, so the funded programme has to look like that. The best-evidenced intervention in education, and the one we do least.
Employer co-funding with completion bonuses, and a published employment outcome for every vocational course.
Foundational skills compound: a child who reads at eight learns everything else faster. High-dosage tutoring has the strongest experimental evidence of any school intervention. The binding constraint is teachers and time, not curriculum documents.
$200m a year40,000 tutoring places at about $3,000 each is $120m, our assumption from overseas programme costs since no New Zealand unit cost exists yet; the rest for coaching, learning support and evaluation.
Tutoring must not strip core classes of teachers. Protect the arts, science and te reo Māori in a broad curriculum.
National: structured literacy, NCEA replacement, attendance targetsLabour: keep structured literacy; Apprenticeship Boost expansion
We would not reverse the 2024–26 reforms. Reversal is the single most damaging thing a new government could do to a ten-year-old.
Straight answers
The hard ones, in one line each, with the reasoning and where everyone else stands. Some of these are unpopular. We would rather say so now.
Do you want more immigration or less?
Grow per person, not just population.
Migration is good for New Zealand. Unplanned migration into a housing shortage is not. Each year we would publish a three-year capacity plan: expected population and household growth from every flow, temporary visas, residence and New Zealanders coming and going, set against homes being completed and the Infrastructure Commission’s capacity assessment. The residence band for the year is set from that plan, forward-looking, not by a formula on last year’s building figures. A residence approval for someone who has already worked here for years, which is what the skilled category now requires, is not a new arrival and is not counted as one. Construction and health workers keep an open pathway whatever the band, because tightening the door on the people who build the houses is how the loop tightens on itself. The 2002–2025 average net inflow of about 30,000 a year is the centre of the band, not a cap and not a target. Skills list reviewed yearly; single-employer work visas ended; the Labour Inspectorate funded to enforce.
Others: NZ First: a values pledge and citizens-only super. ACT: five-year welfare stand-down and a $6-a-day visa levy. Labour: worker protection, no numbers. National: “quality over quantity.” Voters: 36% say numbers are too high, 43% about right.
Should we let the AI data centres in?
Yes, if they pay their own way.
Any load over 50MW brings new firmed generation equal to its peak, or a mix of new supply and demand it can shed at peak and in a dry year, judged together on location, timing, firming and network cost; and it pays its own connection and water. Meet that and you get a decision in 12 months. Refuse it and you do not connect. No moratorium, no subsidy, no secret consents: large loads are publicly notified. Self-sufficiency is not the test; not raising everyone else’s bill is.
Others: Greens: a one-year pause. National and Labour: an “additionality” rule with no numbers yet. ACT: no rules.
Do you have a solar policy?
Yes. Finance it where it pays, and pay exports what they’re worth.
Rooftop solar is real generation, and EECA finds it pays for itself in most of the country for households that use power in the daytime: returns of roughly 6–14% a year on a north-facing array, more if an EV is charged in daylight, and usually better from a hot-water diverter on the cylinder you already have than from a battery, which EECA says is still too expensive for most households. It does not fix our real shortage, which is winter evenings and dry years, and a household battery shifts lunchtime to dinner, not autumn to spring. So we will not pretend panels replace the 3 gigawatts. Three rules instead. One: exported power is paid what it is worth at the time and place it is supplied. The Authority’s July 2026 rules already require time-varying export plans; we make them binding on every retailer and publish the rates side by side, no minimum price and no pretence every unit is worth the retail rate. Two: finance, not grants. A loan repaid through the rates bill over 15 years, on the model all three big parties now back, for solar, hot-water storage, insulation or a battery, but only where an independent assessment shows the full cost pays back. The Crown’s exposure, interest subsidy, defaults and administration, is published every year; it is not free money. Three: renters and people who cannot borrow get the plug-in kits and community batteries funded from the same line, and the tariff rules protect them as much as anyone with a roof.
Others: Labour: $160m SolarSaver, grants up to $3,000, rates-linked finance, $30m community batteries. National: Home Energy Fund, rates-repaid loans, $7m Crown equity, 80,000 homes over 15 years. Greens: publicly owned Kiwipower and zero-interest clean energy loans.
Will you raise the super age?
Yes. Slowly, and not for anyone over 55.
65 stays for everyone born before 1971. From 2036 the age rises two months a year, reaching 67 in 2047, and tracks life expectancy after that. Super stays universal. Be clear about who this lands on: someone born in 1971 waits two extra months and gets at least six years’ notice from a 2029 election; the full two years falls on people born after 1982, who get more than 20 years of compulsory KiwiSaver first. Super already costs more than the entire education budget and reaches $31bn by 2029/30.
Others: Labour: 65 stays. NZ First: 65, citizens only from 2029. National: “more to say” on 67. ACT: 67. 58% of voters want 65 kept. We know.
What about Te Tiriti?
Settled, honoured, and not relitigated.
No Treaty-principles referendum. No new power for any tribunal to veto Cabinet. Settlements are kept in full and honoured. Services are co-designed with iwi where that gets better results, and we measure Māori outcomes and fund what closes the gaps. The country has spent three years arguing about this; we would spend the next three delivering.
Others: ACT: strip “principles of the Treaty” from legislation, “one law for all.” Te Pāti Māori: entrench Te Tiriti, make Tribunal findings binding. National and Labour: status quo.
Crime, and the boot camps?
Fund what works. Close what doesn’t.
Hotspot policing with independent safeguards, addiction treatment that is actually evaluated, and courts that resolve cases in months, not years. The military academies close unless the published evaluation shows lower reoffending. Victims get timely support either way.
Others: Labour: scrap the boot camps now. National: keep them.
Climate, farming and the ETS?
Keep net zero. Measure methane first.
The Emissions Trading Scheme stays, with a credible cap. Farms get measurement and co-funded trials now, and a pre-announced backstop from 2030 if verified progress falls short. Adaptation gets a national cost-sharing rule so we stop subsidising new building in known flood paths.
Others: ACT: scrap net zero, repeal the Zero Carbon Act. Greens: revoke fast-track mining consents. National: removed on-farm pricing by 2030.
The MMP referendum and a four-year term?
Four years yes. Dumping MMP no.
A four-year term with a stronger select-committee system, as the Electoral Commission recommended. Keep MMP; lower the threshold to 4% and remove the coat-tail rule, as the same 2012 review advised. That is the reform the big parties have refused for 14 years, because it lets new parties in.
Others: National: referendums on both. NZ First: “a power grab.”
What about families and childcare?
Make the next hour of work pay.
The $100m families line in our ledger buys three things: simpler access to the childcare help that already exists, with new places funded in the areas where there are none; a pilot that smooths the benefit abatements that can take 70 cents or more from each extra dollar earned, with the effective tax rate on extra earnings published; and targeted insulation and disability access. No universal benefit increase is promised, and FamilyBoost stays as it is.
Others: National: FamilyBoost at 40% of costs, $1,500 Baby Boost. Labour: free prescriptions, $20 fare cap. Greens: $10,000 tax-free threshold.
Pay equity?
Restore the claims process, with the bill published.
The 2025 Act cancelled 33 claims covering more than 150,000 workers, mostly women in care, education support and health. We would restore a fair, evidence-based claims process, and require Treasury to publish the fiscal impact of each settlement before it is signed, so the argument is about the number, not about whether the process exists. The union estimate of $6bn and Treasury’s objection to it are both published; a settlement schedule is not a blank cheque.
Others: Labour: reinstate, blueprint not yet costed. National: keep the 2025 Act. PSA: restoring it lifts 91,000 families by about $300 a week.
Roads, rail and the planning law?
Maintenance first. Pass the planning bills, then leave them alone.
Transport: maintenance and resilience before low-return expansion; every project ranked by published, independently reviewed cost-benefit across all modes; the Cook Strait ferry order stands; Auckland’s time-of-use charge proceeds with real bus alternatives in place first. Planning: the Planning Bill and Natural Environment Bill replacing the RMA are broadly right; we would pass them, keep build-by-right zoning and national grid corridors inside them, and then not rewrite the planning law again for a decade. Fast-track approvals stay, with public notification for any load over 50MW.
Others: Labour: amend but not scrap fast-track; keep granted consents. Greens: revoke fast-track mining consents. National: $1.8bn Cambridge–Piarere expressway; RMA replacement bills before the House.
Crime: gangs, sentencing, police numbers?
Certain and fast beats long and slow.
What actually cuts crime is the near-certainty of being caught and dealt with quickly, not the length of the maximum sentence. So the money goes to the parts of the system that are slowest. The $60m justice line buys about twenty more District Court judges with the registry staff and courtrooms to use them, so the median time from charge to disposal comes down from years to months, with the figure published quarterly; visible policing in the places where offending concentrates, with an independent audit of who gets stopped; and an evaluation budget so every treatment programme is judged on reoffending. Prison stays for violent and serious offenders. Faster courts alone do not guarantee a given fall in crime, and we do not claim one: the measures are days to disposal, reoffending within two years, and the victimisation rate in the Crime and Victims Survey. The gang laws stay if the published evaluation shows they cut harm; if it does not, they go. We would be judged on the victimisation rate, not on how tough a press release sounds.
Others: National: military academies, three strikes, gang patch ban, 500 more police. ACT: harder sentencing. Labour: scrap boot camps. Te Pāti Māori: prison abolition over time.
Defence, AUKUS and the world?
Fund the plan. Choose allies by capability, trade with everyone.
The 2025 Defence Capability Plan, $12bn over four years on the way to 2% of GDP, is the right size for a country that cannot defend itself alone and depends on sea lanes it does not control. It is already in the Government’s books, so we add nothing to it; we would fund it in full and judge it on deployable, maintained equipment, not announcements. Australia and the Five Eyes are the core relationship. AUKUS Pillar 2 technology projects are taken one by one where they add capability without touching the nuclear-free law; Pillar 1 is not on the table. Trade: China stays our largest customer and we keep reducing how much depends on any one market. The Hormuz shock showed fuel security is a defence question. The Government’s Marsden Point diesel reserve holds 93 million litres, about nine days, for $1.2m a month; the $40m fuel-security line in our ledger holds nearly three times that at the same rate: enough to take diesel from nine days to about a month, or to split the extra between diesel, petrol and jet. The fuel itself is bought once, held as a Crown asset and rotated, not spent.
Others: National: the Capability Plan, exploring Pillar 2. NZ First: foreign affairs and Pacific focus. Greens: opposed to AUKUS. Labour: broadly the same plan, cautious on Pillar 2.
The public service: cut it or rebuild it?
Neither. Measure what it delivers.
Headcount is the wrong number in both directions. 8,700 fewer jobs is not a result and 8,700 more would not be either. Every agency publishes what it delivered against what it was funded for, and gets funded on that. Back-office and digital functions are merged where duplication is real, but no restructure happens without a published business case, the same rule we apply to Health NZ. The consultant and contractor bill is cut by rebuilding in-house policy and digital capability where the whole-life cost, recruitment and training included, is lower than contracting; genuinely specialist work stays contracted. Wellington gets the same growth plan as everyone else, not a promise that the state will always be its largest employer.
Others: ACT: cut 43 agencies to 19 and 28 ministers to 18. National: 8,700 jobs cut, agencies merged. Labour: reverse some cuts.
What do I get, this week?
Bracket creep back, and cheaper power and groceries. That’s it.
Every other party has a sweetener: a fare cap, free prescriptions, a baby bonus, a tax-free threshold. We do not, because most are paid to everyone whether they need them or not, and they leave the causes untouched. Help that is targeted and works is a different thing: our doctor loan write-offs, childcare places where there are none, and smoother benefit abatement are all in the ledger, and each has a published test of whether it works and what it costs. The only universal cash we put back in your pocket is the tax you would otherwise lose to bracket creep, indexed to wages permanently, plus what competition takes off your power bill and your grocery bill. It is less exciting than a cheque. It compounds.
Others: Labour: $20 fare cap, free prescriptions, three free GP visits. National: $1,500 Baby Boost, FamilyBoost. Greens: $10,000 tax-free. Te Pāti Māori: $30,000 tax-free.
Where does the growth actually come from?
Cheaper inputs are the floor. Firms that scale are the plan, and it isn’t written yet.
Housing, power, health and competition are constraints. Removing them lets growth happen; it does not make it happen. The 1.5% productivity target on our scorecard is a target we ask to be judged on, not a forecast of what the six moves deliver. The part that turns lower costs into higher output is firms adopting better technology, scaling, and selling more valuable things abroad, and we have not designed it. Three first-term commitments, each in the register as a gap with no money booked until it has a design, a cost and a measure: close the growth-capital gap the OECD identifies for firms raising $5m to $50m, starting with what compulsory KiwiSaver and the tax switch already move; a technology-adoption programme for firms and public services judged on measured productivity, not activity counts; and one procurement, research and talent pipeline for commercialisation. We would rather publish the hole than a paragraph pretending to fill it.
Others: National: Investment Boost, science-system reforms. Labour: industry policy to come. ACT: cut regulation. Greens: a public investment company.
Who is behind this, and who pays for it?
No one yet. Every dollar will be published.
This is a founding concept with no MPs, no members and no money, published to test whether the argument holds. If it becomes a party, every donation over $1,000 is disclosed within a week of receipt, not once a year; no anonymous or foreign donations; and the same disclosure rules we would legislate for everyone apply to us first. Until then, the only thing we are asking for is an argument.
Others: Current law: donations over $5,000 disclosed annually; over $20,000 within ten working days. Every party in Parliament has faced a donations story in the last two terms.
Council rates and water?
Take the pipes off the rates bill.
The National Infrastructure Bank lends 40-year money against targeted rates, so a $48bn water programme is not paid for through 9% annual rates rises. A rates cap without a way to pay for pipes just means broken pipes.
Others: National: cap rates rises at 2–4% from 2029. Labour: opposed the cap.
Red lines and tradeables
Under MMP nobody governs alone. A platform that will not say which parts are negotiable is asking you to guess. Here is the list, in advance, so nobody can call a deal a betrayal.
Gains tax and the cuts travel together, staged behind receipts. We will not pass one half without the other.
The 50MW rule, additional firmed supply or flexibility worth the same, is the price of our support for any energy or data-centre policy.
Structured literacy, housing growth rules, Investment Boost and the Infrastructure Plan stay. We will not join a government that reverses them.
We will not sign a fiscal plan that books savings nobody has found.
The start year and the pace are negotiable; the direction is not.
We would take a four-year term without the threshold change, or the reverse.
Thresholds, the clock and which markets go first can move; the existence of the power cannot.
Operating lines can be delayed or scaled if the allowance will not carry them. We will say which.
The money
New operating spending takes just under half of one year’s $2.4bn Budget allowance at full rollout, 46% to be exact. That is not spare money; it is space other new spending loses, and the surplus path holds only if the rest of the allowance is not also spent. The tax switch is separate and phased behind its own receipts. Capital sits on the Crown balance sheet and is repaid by the people who benefit.
A broad capital gains tax, family home and KiwiSaver excluded, pays for a company-tax cut to 25% and wage-indexed thresholds. The revenue ramps slowly (the Tax Working Group put a comparable design at $0.4bn in year one and $5.9bn by year ten), so cuts are staged behind certified receipts by rule: no cut until two years are certified, each step sized to 90% of the three-year average less cuts already made, a stabilisation account holding a year of the cuts before the next step, indexation paused if it would run dry. Neutral by design, not by promise: the year-by-year profile and two downside cases are in the register.
$2bn seeds the National Infrastructure Bank; it lends against targeted rates and is repaid over 40 years. $182.3m of Crown capital over 2030–34 for the three proof-of-method suburbs, alongside council and developer shares.
Treasury: Budget 2026 fiscal outlook and operating allowances
Proof of method
Before we ask any council to sign up, we show the bill. 1,000 net homes in each of Tauriko West, Rotokauri and eastern Porirua: local water, streets, a school’s worth of places and a shared hub, with contingency and escalation on top. Move the sliders to stress it.
A 1,000-net-home serviced neighbourhood within the initial growth area, with local streets, water connections and shared community facilities.
Independent concept estimate at September 2026 prices, excluding GST. Quantities and unit rates are published assumptions, not a tender. No council or landowner has yet agreed to take part.
Defaults: 30% contingency and 3% escalation, on top of 15% for design, consents and programme. These are alternative assumptions, not confidence intervals.
| Works | Quantity | Unit cost | Base |
|---|---|---|---|
| Local drinking-water mainsWater · Assumed installed reticulation including reinstatement; excludes trunk upgrades. | 5 km | $1,500,000 | $7.5m |
| Local wastewater networkWater · Assumed network and reinstatement. | 5 km | $2,000,000 | $10.0m |
| Local pumping and connectionsWater · Allowance for connections and pumping; capacity not hydraulically verified. | 1 allowance | $5,000,000 | $5.0m |
| Local drainage and detentionWater · Local works only; avoids adding the existing trunk programme. | 1 allowance | $8,000,000 | $8.0m |
| Neighbourhood streetsTransport · Pavement, footpaths, lighting and utilities corridor; water pipes priced above. | 4 km | $6,000,000 | $24.0m |
| Additional walking/cycling linksTransport · Links outside street cross-sections; no duplicate footpath charge. | 3 km | $800,000 | $2.4m |
| Bus stops and priority worksTransport · Local access package; buses funded in annual operations. | 1 allowance | $2,000,000 | $2.0m |
| Parks: land and establishmentCommunity · Assumed three hectares; land and landscaping included, not a quoted valuation. | 3 ha | $1,500,000 | $4.5m |
| Shared community hubCommunity · Building and fit-out; use shared facilities instead of a separate library, hall and centre. | 1,000 m² | $6,500 | $6.5m |
| Hub and clinic landCommunity · Purchase allowance; no free land assumed. Excludes school land below. | 1 allowance | $2,000,000 | $2.0m |
| School capacityCommunity · All-in allowance for buildings, share of land and fit-out; Ministry network assessment could alter need. | 200 places | $80,000 | $16.0m |
| Primary-care premisesCommunity · Premises and fit-out only; staffing included separately in annual operating model. | 500 m² | $6,500 | $3.3m |
| Direct works | $91.2m | ||
| Design, consents, programme · 15% | $13.7m | ||
| Contingency · 30% | $31.4m | ||
| Total, 2026 dollars | $136.3m | ||
| Escalation to spend dates | $27.3m | ||
| Total cash, 2030–34 | $163.5m | ||
Excludes GST, finance, private home construction, residential land, lot earthworks, power and telecom connections, and existing trunk programmes, which are listed under dependencies.
Water 70/20/10, transport 50/40/10, community 20/60/20 (developer/Crown/council). Council share financed over 30 years at 5.5% costs $1.5m a year: about $1,523 a year per new home if it were recovered entirely through a targeted rate on the new lots. The developer share is recovered through section prices, not rates.
Undiscounted 30-year envelope: $370.5m, real capital plus 30 years of full operation. Conservative; not a net-present-value appraisal.
Depends on the separately funded SH29 enabling works and available bulk water/wastewater capacity. This model excludes those trunk works and the later SH29 corridor project; it is not a budget for all Tauriko West.
NZTA lists $264m for the enabling works. That is a wider shared project estimate, not money newly committed by Next or a unit rate for this tranche.
It gives a reproducible budget for the defined tranche and a funding proposal. It cannot give surveyed quantities, ground conditions, remaining trunk costs or secured grants. 2030: design and capacity checks. 2031–32: local networks. 2033–34: facilities. Contracts are released only after validated design. Do not multiply these three across the country; every site needs its own appraisal.
Scorecard
Every measure has a public source and a date. We report against all eight every year, including the ones going the wrong way.
Our output per hour is roughly 30% below the top half of the OECD and well below Australia’s. If we lift productivity growth from 0.2% to 1.5% a year and Australia keeps growing at about 1%, the gap narrows by roughly five points a decade. That does not close it in ten years. It stops it widening and starts it closing, which no government has managed since 2008. Closing it fully is a generation’s work, and we would rather say so than promise 2036.
None of the money in the ledgers assumes this growth arrives, and the target is not an estimate of what the six moves deliver; the part of the plan that turns lower costs into firms that scale is a stated gap. If growth does arrive, the extra revenue goes to thresholds and debt, in that order, and the fiscal council reports it.
| Measure | Now | Target | By |
|---|---|---|---|
| Net loss of New Zealanders to Australia specifically, per yearStats NZ, December 2025 year, Australia only; averaged 3,000 in 2014–19. The hero chart is net loss to all countries (6,000 in 2014–19). | 28,500 | under 10,000 | 2035 |
| Adults unable to get a GP appointment in timeNZ Health Survey; back to the 2021/22 level. Cost barrier (14.9%) reported alongside. | 25.5% | 12% | 2033 |
| Grocery market share of the two big chainsCommerce Commission annual grocery report. Share is a proxy; the report’s gross-margin series is published with it, because a smaller share at the same margin is not a result. | 82% | under 70% | 2034 |
| Average wholesale electricity priceElectricity Authority; 2019 level | ~$160/MWh | under $100 real | 2034 |
| New generation consented and commissionedTranspower and EPA registers. Consents are permission; commissioned megawatts are power. | — | 3 GW consented, 1.5 GW commissioned | 2033 / 2035 |
| Winter security of supplyTranspower security of supply assessment. Firm supply through winter evenings and dry years, not just nameplate capacity. | “fragile”; below standard by 2031 on current loads | above Transpower’s standard every winter | from 2033 |
| New homes consented and completed per yearStats NZ, year to July 2026, up 21%; the 2022 peak was about 51,000. Completions (code compliance certificates) lag consents by a year or more and are the measure that matters; we report both. | 40,900 consented | 55,000 consented; completions published against them | 2033 |
| PISA maths scoreOECD PISA; the 2012 level | 480 | 500 | 2034 cycle |
| Labour productivity growthStats NZ productivity statistics. A target we ask to be judged on, not an estimate of what these moves deliver, and not revenue we have spent; nothing in the ledgers assumes it. | 0.2% a year | 1.5% a year | 2030–35 average |
What we won’t do
No “efficiency dividend.” No growth dividend spent in advance. If it is not in the ledger, it is not funded.
The state is a bad grocer. Competition law with teeth is cheaper and works.
Structured literacy, housing growth rules, Investment Boost, the infrastructure plan: kept, whoever passed them. Continuity protects delivery while it matures; it does not make a policy immune to evidence, and the same evaluations apply to them as to ours.
Election day is 7 November
Party registration for 2026 closed on 6 August. So this is a plan and an argument, not a ballot line. Take it, test it, send it to someone who is packing for Brisbane. If the argument holds, 2029 is the election it is built for.
Founding concept, not a registered party. This page collects no personal information and sends nothing.
The one sentence
Tax gains, not work. Pay your own way. Break the duopolies. Build the pipes, then the homes. A doctor this week. Every child reading by eight.
Sources
Government and official statistics establish the facts. The recommendations and cost assumptions are ours. Research window: 2020 to 9 September 2026.