Money · Credit · New Zealand · Part 15
How should New Zealand pay for the maintenance of everything it builds?
Building an asset is only the first financing decision. The harder commitment lasts for decades.
A new hospital opens.
A road is completed.
A water network is extended.
A school is rebuilt.
A digital system goes live.
Those moments are visible.
Maintenance is not.
There is no ribbon-cutting ceremony for replacing a pump before it fails.
No campaign photograph for resurfacing a road at the right time.
No dramatic announcement for patching a roof while it is still watertight.
But this is where national wealth is either protected or quietly consumed.
Every new asset is also a promise to fund its maintenance, renewal and eventual replacement.
New Zealand's largest future infrastructure challenge may be looking after what already exists
Te Waihanga's 2026 National Infrastructure Plan says as much as 60 cents in every dollar of future infrastructure spending may need to go toward maintenance and renewals.
It calls this New Zealand's biggest long-term infrastructure investment challenge.
Te Waihanga — Looking after what we've got ↗
The Commission's recommendation is direct:
maintenance and renewals should be the first investment priority.
Te Waihanga — Making maintenance and renewals the first investment priority ↗
That changes the way we should think about capital budgets.
The first call on future capital is not necessarily the next new project.
It may be preserving the assets already carrying the country.
Maintenance, renewal and replacement are different
We should separate three concepts.
Maintenance is the work required to keep an asset operating as intended.
Renewal replaces components that have reached the end of their useful life while maintaining broadly the same service capability.
Replacement or enhancement may involve replacing the whole asset or materially increasing its capacity or service level.
Those distinctions matter because they can justify different funding approaches.
Routine maintenance should normally be treated as a recurring cost of providing today's service.
A major renewal may be lumpy and happen only once every twenty or forty years.
An expansion may partly serve future population growth and therefore raise different intergenerational funding questions.
Depreciation is a warning signal — not a savings account
This is one of the easiest concepts to misunderstand.
Depreciation is an accounting expense representing the consumption of an asset's service potential over its useful life.
It is not automatically cash sitting in a bank account waiting for the asset to be replaced.
The Office of the Auditor-General compares councils' renewal expenditure with depreciation because depreciation provides a useful indicator of how much of the asset base is being used up.
But the Auditor-General also warns that it is only an indicator because infrastructure has long life cycles and actual renewal requirements can be uneven.
Office of the Auditor-General — Councils' financial and infrastructure strategies ↗
Depreciation tells us that assets are being consumed. It does not, by itself, tell us exactly how much cash should be spent on renewals in any one year.
The council numbers show why this matters
In their 2024–34 long-term plans, 58 councils forecast total capital expenditure of about $91.9 billion.
Of that, about $39.5 billion was planned for renewing existing assets.
Across the decade, forecast renewals were expected to average about 85% of forecast depreciation.
The Auditor-General says that is an improvement on earlier plans, but still indicates that some councils may not be planning enough reinvestment to maintain service levels.
Office of the Auditor-General — 2024–34 long-term plans ↗
Actual 2023/24 council renewal expenditure was about 84% of depreciation.
Office of the Auditor-General — Councils' financial performance 2023/24 ↗
Again, neither ratio proves that every council is underfunding every asset.
But sustained gaps should force questions about asset condition, future renewal peaks and service risk.
Water shows how dangerous the averages can be
The national average can conceal very different asset classes.
For councils' 2024–34 plans, forecast water-supply renewals were equivalent to about 113% of depreciation.
Wastewater renewals were about 87%.
Stormwater and drainage were only about 43%.
The Auditor-General specifically noted concern about stormwater given increasing severe-weather risks.
Office of the Auditor-General — Three-waters renewal forecasts ↗
This is why renewal planning cannot be managed through one national ratio.
The condition and criticality of each network matter.
Local government already has an intergenerational principle in law
Section 100 of the Local Government Act 2002 generally requires a council to set projected operating revenue at a level sufficient to meet projected operating expenses, subject to a prudence exception.
When using that exception, councils must consider matters including the cost of maintaining service capacity and asset integrity throughout useful life, available funding, and the equitable allocation of responsibility for maintaining assets over that life.
Local Government Act 2002 — section 100 ↗
That gets to the heart of the issue.
Who should pay for an asset that is consumed over several generations?
My first rule: routine maintenance should normally be funded from current revenue
If today's users consume today's service, the normal starting point should be that today's revenue pays for ordinary maintenance.
For councils, that can mean rates and user charges.
For Crown social infrastructure, it can mean operating appropriations.
For commercial networks, it may be built into prices charged to customers.
Borrowing every year to mow the grass, patch routine leaks or perform scheduled servicing would simply shift recurring operating costs into the future.
That is rarely good capital discipline.
My second rule: predictable renewals need a long-term funding path before the asset fails
Renewals are different because they can be large and irregular.
A pipe may last eighty years.
A roof may last thirty.
A major road surface may require renewal on a much shorter cycle.
Those costs should not arrive as financial surprises.
Asset-management plans should forecast when components are likely to require renewal, what they will cost and how that expenditure will be funded.
Te Waihanga is currently calling for capital-intensive central-government agencies to prepare and publish long-term asset-management and investment plans, including multiple funding scenarios.
Te Waihanga — Long-term asset-management planning ↗
Should we literally set aside depreciation as cash?
Sometimes a dedicated renewal fund or reserve can be useful.
But I would not make a rule that every dollar of accounting depreciation must automatically sit idle as cash.
That can be inefficient and may not match the actual renewal profile.
The better principle is:
the owner must demonstrate a credible funding path for expected lifecycle costs.
That might involve:
annual cash contributions to a dedicated renewal fund,
retained earnings,
future rates or user charges,
multi-year appropriations,
or planned borrowing for a large renewal when spreading the cost across future beneficiaries is justified.
What matters is that the renewal obligation is visible and financeable.
Borrowing can smooth a lumpy renewal — but it cannot replace lifecycle discipline
Imagine a water-treatment plant that lasts forty years and requires a major replacement costing hundreds of millions of dollars.
It may not be sensible to accumulate every dollar in cash decades in advance.
Borrowing at replacement time can spread part of the cost over the users who will benefit from the renewed asset.
But there is a difference between smoothing a large lifecycle cost and repeatedly borrowing because the owner failed to fund maintenance for thirty years.
The first can be an intergenerational financing decision.
The second is deferred maintenance being converted into debt.
Borrowing should smooth a renewal cycle, not hide the absence of one.
Growth and renewal costs should be separated
Suppose a council replaces an old pipe with one twice the capacity because a city is growing.
Part of the project is renewal.
Part is growth.
Those components may have different beneficiaries and therefore different funding sources.
The renewal component may appropriately fall on existing network users.
The additional capacity may justify development contributions, targeted levies or other growth-related funding where legally available and economically appropriate.
If those costs are mixed together, existing households can end up paying for growth they did not create, or new development can escape costs it imposes on the network.
User-funded networks should price the lifecycle, not only today's operating cost
For infrastructure such as water, electricity, ports, airports and other network services, prices can help recover the full efficient cost of maintaining and renewing assets.
A tariff that covers today's staff and electricity bill but ignores tomorrow's pipe replacement is not genuinely cost-reflective.
Likewise, artificially low prices can look affordable today while creating a much larger replacement bill later.
Where affordability support is justified, I would rather make that subsidy visible than underprice the entire network and starve it of renewal capital.
Central government needs longer funding horizons
Annual Budgets are a difficult environment for long-lived asset management.
A hospital network cannot sensibly plan thirty years of renewals if agencies only have confidence about the next short funding window.
Te Waihanga therefore recommends more predictable government funding signals and multi-year budgeting for capable infrastructure agencies.
Te Waihanga — Predictable funding and multi-year budgeting ↗
This does not mean agencies receive blank cheques.
It means high-quality asset plans can be matched to credible multi-year funding envelopes so procurement and renewals can be sequenced efficiently.
The current system is being strengthened because the information is not good enough
Te Waihanga reports that, as of June 2025, four of eight capital-intensive central-government agencies said their asset registers did not meet required standards, and four lacked asset-management plans that met expectations.
The Government agreed in 2026 to strengthen assurance over asset management and long-term investment planning.
Treasury — Strengthening Investor Assurance ↗
That is fundamental.
You cannot fund a renewal plan intelligently if you do not know exactly what assets you own, what condition they are in or when they are likely to fail.
Every new project should arrive with a lifecycle invoice
I would add a simple rule to the National Investment Test from Part 12.
No major new public asset should be approved without a published lifecycle funding plan.
That plan should show:
expected useful life,
annual maintenance requirements,
major renewal dates,
replacement assumptions,
operating costs,
resilience requirements,
the proposed funding source,
and which generation of users is expected to pay.
The capital cost on opening day is therefore only one line in the real price of the asset.
I would also publish a renewal coverage dashboard
For each major asset-owning public organisation, I would want to see:
asset condition
maintenance completed versus planned
renewals completed versus planned
renewal expenditure relative to depreciation
forecast five-, ten- and twenty-year renewal requirements
unfunded renewal obligations
service failures attributable to asset condition
and deferred-maintenance risk.
The renewals-to-depreciation ratio would be useful, but never treated as a pass/fail target on its own.
Condition and service risk should decide whether reinvestment is adequate.
The political incentive has to change
New infrastructure is visible.
Maintenance is mostly invisible when it works.
That creates a structural political bias.
A government can announce a new building.
It gets less recognition for preventing the old one from deteriorating.
But economically, the second decision can be more valuable.
If a $50 million renewal extends the useful life of a $500 million asset by twenty years, that may create more value than a much more visible new project.
The absence of a breakdown is not evidence that maintenance spending achieved nothing. It may be evidence that it worked.
Deferred maintenance is a form of hidden borrowing
This is how I would describe the problem conceptually.
When an owner skips necessary maintenance to make today's budget look cheaper, it creates an obligation for the future.
The liability may not appear as a conventional bond.
But somebody later will have to pay for:
the postponed maintenance,
a more expensive repair,
reduced service,
or premature replacement.
In that sense, deferred maintenance behaves like hidden borrowing from the future condition of the asset.
And the interest rate can be brutal.
A national maintenance rule
If I reduced the article to one policy framework, it would be this:
1. Know the assets. Complete asset registers and condition data.
2. Price the lifecycle. Maintenance, renewal, operating and replacement costs before approval.
3. Fund routine maintenance from recurring revenue.
4. Establish a credible long-term funding path for renewals.
5. Separate renewal from growth expenditure.
6. Use borrowing only where it improves intergenerational allocation or smooths genuinely lumpy capital needs.
7. Use multi-year funding where asset plans justify it.
8. Report deferred maintenance openly.
9. Do not approve a new asset without its lifecycle funding plan.
10. Make maintenance performance as visible as new capital announcements.
My conclusion
New Zealand does not have a shortage of things it wants to build.
The harder test is whether we are willing to pay to own them properly after the opening ceremony.
A mature investment system would treat maintenance and renewals as part of the original investment decision, not as somebody else's problem twenty years later.
It would distinguish accounting depreciation from actual renewal need.
It would match funding to beneficiaries and asset life.
And it would force every asset owner to show how today's infrastructure will remain tomorrow's asset rather than tomorrow's liability.
The cheapest year to own infrastructure is often the year you postpone maintaining it. The most expensive years come later.
Next
This creates another question that sounds simple but is not:
Should infrastructure pay for itself?
Part 16 will separate financial return from economic and public return and examine when user charges make sense, when taxpayers should fund an asset, when cross-subsidy is justified, and why a project can be valuable even if it never produces a commercial profit.
Primary sources
Te Waihanga — Making maintenance and renewals the first investment priority ↗
Office of the Auditor-General — Councils' financial and infrastructure strategies ↗
Office of the Auditor-General — Councils' financial performance 2023/24 ↗