The cut they never tested

What NZTA’s own modelling says about the Total Mobility subsidy cut — and the questions it never asked

From 1 July 2026, the Total Mobility subsidy falls from 75 percent to 65 percent of a fare.

The scheme exists to give disabled people with long-term impairments subsidised door-to-door transport where ordinary public transport will not serve them. The official justification for the cut is funding pressure.

The remedy chosen is a blanket reduction in the subsidy rate — a cut that lands on every user equally, whether they take one trip a fortnight to a medical appointment or several a day.

I have now read the modelling and the options analysis that sit behind this decision, released to me under the Official Information Act. They do not make the case the Government says they make.

They show officials reaching for the lever that saves the most money, justified on data the agency itself will not stand behind, having never once asked what the cut costs the disabled people who rely on the scheme.

This post sets out why.

First: there was never a cost-benefit analysis.

This is the single most important fact in the release, and it is admitted in plain words.

The Options paper states that no analysis was conducted to determine the benefit-cost ratio of the current scheme, the impact on users, or whether the benefits outweigh the increasing costs.

The modelling measures one thing only: how much money each lever saves. It does not measure what those savings cost in lost mobility, lost independence, or lost community participation for disabled New Zealanders.

This matters because the entire exercise is framed as fiscal responsibility. But you cannot responsibly cut a support scheme when you have deliberately declined to measure the benefit you are cutting.

Treasury would not accept a spending proposal built this way; it should not accept a savings proposal built this way either. The Government is asking the public to accept that the cut is worth it, while holding a set of documents that explicitly never tested whether it is.

Second: the cut works by deterring disabled people from travelling — and the modelling says so openly.

The savings in this policy do not come only from paying a smaller share of each fare. They come substantially from people taking fewer trips because the trips now cost them more.

The modelling is built on an assumed price elasticity of demand of around 0.7, which means a rise in price is expected to produce a corresponding fall in the number of trips taken.

In Plain English, the scheme saves money in part because disabled people, faced with a higher out-of-pocket cost, will travel less.

Decreased demand is not an unfortunate side effect of this policy that officials failed to foresee.

It is one of the two ways the policy is designed to work, and it is written into the model that produced the savings figures.

A scheme whose purpose is to enable community participation is being adjusted, by design, to reduce the amount of participation it pays for.

Third: the cut is justified on data the agency itself describes as unreliable.

Throughout the analysis, NZTA is candid that its evidence base is weak.

The paper states that the data came from two sources, that neither is a perfectly accurate depiction of reality, and that this made modelling the impacts difficult.

The results are repeatedly described as indicative rather than definitive, with further work required to rectify errors and inconsistencies in the data.

One region’s results are flagged as needing to be treated with caution because of data quality issues.

There is nothing wrong with officials being honest about the limits of their data. The problem is what is being built on top of it. A ten-percentage-point cut to the transport subsidy of a disabled population is a real-world consequence with real-world costs, and it is being justified by modelling that its own authors will not call definitive.

Caveats that would ordinarily prompt more work before a decision have instead been carried straight through to a decision.

Fourth: the funding “crisis” is largely a product of an earlier policy choice — and costs are not running away.

The demand growth that underpins the pressure narrative is not a mystery.

The subsidy rose from 50 percent to 75 percent in April 2022 as part of the half-price fares scheme.

Demand grew at roughly 8 percent per annum afterwards, against roughly 2 percent per annum in the years before. The increase in demand corresponds directly to the increase in subsidy.

The pressure now being used to justify the cut was created by a deliberate decision to make the scheme more generous — a decision that did exactly what such a decision does.

The earliest model in the release, prepared in August 2024 — seven months before any formal review was publicly announced projects total scheme costs rising on a blunt assumption of 5 percent demand growth a year.

But the agency’s own per-kilometre cost data is essentially flat. The pressure is volume, not runaway operating costs.

And volume assumptions can be tested. The Horizons region, for which Ministry of Transport data has already been released, shows a 33 percent structural drop in trip volume — evidence pointing in the opposite direction to the growth assumption that drives the forecast. No sensitivity analysis was released.

A forecast resting on a single untested demand assumption is an assertion, not a projection.

I should be precise about the figures, because precision matters when these documents will be scrutinised.

The Options paper works from a current-year shortfall of around $10.36 million and a three-year pressure of around $60 million. The larger multi-year Crown figures that have circulated rest on a different model with a different baseline, and I treat the headline “gap” as contested rather than established until the two can be reconciled.

The case against the cut does not need an inflated number; the Government’s own documents are damaging enough.

Fifth: a more targeted lever existed, and the blunter one was preferred because it saved more.

The Options paper modelled three levers, not one. Alongside the subsidy cut it tested trip caps and fare caps.

I want to be fair about trip caps, because the document is clear: capping the number of trips, even at the strictest level modelled, reduces national subsidy by only around 7 percent and does not close the funding gap.

Trip caps are not the silver bullet they might first appear to be, and it would be wrong to claim NZTA simply ignored them.

But fare caps are a different matter. A fare cap limits the subsidy paid on any single trip, which means it targets the longest and most expensive journeys rather than reducing the subsidy on every trip a person takes.

The modelling shows a ten percent reduction in fare caps would take the shortfall from $10.36 million to around $0.9 million — close to balancing the scheme. It is feasible quickly and at no development cost.

Yet it was set aside on the basis that its impact would not be “as great as a subsidy change.”

That is the choice laid bare. Faced with a lever that nearly balances the scheme by trimming the most expensive outlier trips, and a lever that balances it by reducing what every disabled person receives on every journey, officials preferred the second because it saved more. Saving more is not the same as choosing well.

The concentration of use in the scheme — where the highest-using fifth of clients account for around three-quarters of the subsidy, very likely because they have the highest needs — means a blanket cut falls hardest on exactly the people the scheme exists to serve.

Where this leaves us.

Strip away the language of fiscal necessity and the official case reduces to a series of decisions the Crown’s own documents expose.

A cut was chosen without any analysis of what it costs the people it affects. It is designed, in part, to work by deterring disabled people from travelling.

It rests on data the agency will not call reliable, and on a demand forecast it never stress-tested against its own regional evidence.

A fairer, more targeted lever was available and was passed over because it saved less. And the pressure being used to justify all of this was substantially created by an earlier choice to expand the scheme.

Disabled New Zealanders are being asked to give up a tenth of their transport subsidy on this basis. They are entitled to a decision built on evidence of what they stand to lose. The documents prove that decision was never made.

---

The documents referred to in this post were released under the Official Information Act. I am continuing to seek the underlying sensitivity analysis and the reconciliation of the competing cost forecasts, and will publish them when received.