Report
ACC is approving fewer planned operations, and the levy on your wages is set in law to rise twice more
Source: ACC, Quarter Three Performance Report 2025/26 (March 2026)
New decision tools are lowering approval rates for operations like knee and shoulder repairs, ACC's own quarterly report says, and challenging a decision now takes 121 days. Behind it sits a $66.0 billion claims bill that inflation is adding to faster than the turnaround can cut it.
By Henry Smith · ACC · 2026-07-08
What this means for you The levy taken out of your wages funds the ACC account in the worst shape, and that levy is already set in law to rise each April through 2027. This report is the starting position for whatever rate gets set after that. ACC is also approving fewer planned operations, which pushes more treatment toward the private system.
If you are waiting on a knee or a shoulder repair, ACC has become more likely to say no, and it says so itself. New decision tools for planned operations are "lowering approval rates from historical baselines", according to ACC's latest quarterly report. The same report describes staff going back over home help and care support already granted, "reducing or stopping entitlements where clinical need was no longer evidenced". If you want to argue with a decision, the average challenge now takes 121 days to settle, almost 14 days longer than a year earlier (ACC, 8 July 2026, pages 6 and 14).
At the same time, the levy taken out of your wages is set in law to rise twice more. It goes from $1.67 per $100 you earn in 2025/26 to $1.75 in 2026/27, then $1.83 in 2027/28, GST included (Inland Revenue).
So what does that mean for a working household? Not a new bill in the post. It means paying more into the scheme while the scheme approves less, and waiting four months to contest the difference. The cover is narrowing at the same time as the price of it is going up.
Why is ACC tightening?
ACC expects to pay out $66.0 billion on injuries that have already happened. That is not a guess about future accidents. It is the lifetime cost of every claim already on its books, a figure ACC calls its outstanding claims liability, and it grew $2.4 billion in the nine months to 31 March 2026 (ACC).
The turnaround is working on its own terms. The same report has ACC now expecting a full-year loss of $2,158 million, which is $726 million better than the $2,884 million it budgeted, helped by claims payments running $380 million under budget and investment income $574 million above it.
What the turnaround cannot outrun is inflation. Over those nine months ACC's Turnaround Plan, its programme for getting injured people back to work sooner, took $652 million off the claims bill. Higher assumed wage and price growth put $1,457 million back on (ACC, page 5).
That is $2.23 back on the bill for every dollar the plan got off it, and the division is ours.
| What moved the claims liability, nine months to 31 March 2026 | $m |
|---|---|
| Expected increase (growth built in every year, the way interest builds on a mortgage) | +1,901 |
| Inflation assumptions (higher future wage and price growth) | +1,457 |
| Cost growth ACC cannot control (mostly Sensitive Claims, its cover for mental injury caused by sexual violence) | +214 |
| Discount rate rising to 5.34% from 5.15% | -513 |
| Turnaround Plan savings (ACC's "influenceable release") | -652 |
| Net increase ($63,556m to $65,964m) | +2,408 |
Components sum to +$2,407m; ACC's printed figures carry $1m of rounding. All figures from ACC's Quarter Three Performance Report 2025/26, pages 5 and 21. The division and the 2.23-to-1 ratio are ours.
That $652 million is a net figure, and the gross tells you more. ACC saved roughly $1,110 million, most of it on weekly compensation, the income it pays you while an injury keeps you off work. Sensitive Claims then took $441 million of that straight back, on pages 6 to 8 of the same report. Four dollars in every ten ACC saved were gone before inflation was counted.
Whose levies are falling short?
Yours, most likely. ACC keeps a separate pot for each group that pays it: employers, vehicle owners, and wage earners. The Earners' Account holds the levy taken out of your wages, and it is carrying almost all of the shortfall.
Levy income across the scheme is $259 million short of budget, a 7% miss ACC blames on "weaker labour market conditions". Of that, $224 million sits in the Earners' Account, about 87 cents in every shortfall dollar (ACC, page 19; the account split is our arithmetic).
That same account is in the worst shape of the levied accounts. Its funding ratio, meaning how much of its future bill it already holds in assets, is 79.1% against a 100% target, down from 83.5% in June 2025 (the same report, page 21).
| Account | Funding ratio at March 2026 (target 100%) |
|---|---|
| Earners' (levy on wages) | 79.1% |
| Motor Vehicle | 131.5% |
| Work (employers) | 135.1% |
| Non-Earners' (government-funded, not levied) | 49.9% |
All ratios from ACC, page 21.
There is a squeeze inside that account the report does not spell out. The wages ACC collects its levy on are barely moving. Unemployment sat at 5.3% in the March 2026 quarter, and pay across the economy rose just 2.0% over the year (Westpac IQ, 6 May 2026).
The wages ACC pays compensation against are doing the opposite. They are climbing at about 6.5%, where ACC had budgeted about 2.5%, on page 7 of the same report. Its income rides the weak number. Its costs ride the strong one.
The legislated levy path stops on 31 March 2028, and ACC has to consult on the rate that follows before then. It goes into that consultation with the wage earner account below target, levy income short, and a bill inflation keeps pushing up. Nothing in the report says what will be proposed. Nothing in it points to a smaller ask, either.
Is $32,000 a household a real bill?
Spread $66.0 billion across the country's 2,064,700 households and it comes to just under $32,000 each. Across the 2.89 million people in work, about $22,800 (Stats NZ, 7 April 2026, and Stats NZ labour market, 6 May 2026, both).
No invoice is coming. That $66.0 billion is a promise stretched over decades, and ACC holds $55.6 billion of investments to meet it. Count what the scheme owns against what it owes and it is $13.9 billion short (ACC, page 20).
Divided by households, that shortfall is about $6,700, and the year's forecast loss adds roughly another $1,045 each (the per-household sums are ours). The $32,000 is the size of the promise. The $6,700 is the part nobody has paid for yet, and it is what sets the pressure on your levy and on how freely ACC approves the next knee repair.
What does this mean for my cover?
ACC covers injury and never illness, and the injury half is tightening at the edges: fewer planned operations approved, support already granted withdrawn, four months on average to challenge it. For a household that treats ACC as its safety net, the net has not been removed. It has been drawn tighter, and the gap now sits with the private system. Health insurance covers that gap from the other side.
A scheme cutting back on support for injured people is not about to start covering illness. Where a household budget assumes pay keeps arriving when someone cannot work, that assumption holds for accidents only. Income protection extends it to illness. No levy rise buys that.
What could we not check?
The report contradicts itself on one figure: what it saved on home help and care support, which it calls social rehabilitation. Page 5 says about $130 million, page 6 says about $180 million, page 7 nets to $179 million. We used page 7, the one that reconciles (ACC, pages 5 to 7).
The "$574 million higher than budget" investment income is two things added: $104 million of interest, dividend and rent above budget, plus $470 million of gains. Both are our sums of the five account columns on page 19 of the same report, which never shows that split itself. The report prints no publication date either, so our 8 July 2026 date comes from the PDF file's metadata.
Stats NZ's own pages returned only their titles on 20 August 2026. We confirmed the 2,064,700 household figure through Scoop's republication of the Stats NZ release of 7 April 2026, and the 2.89 million employment level through Trading Economics' copy of the official employment survey, which records 2,892,000 employed in the March 2026 quarter. Both are second-hand copies of Stats NZ data and we cite them as that.
The reading that the next levy consultation inherits this arithmetic is ours. The claims bill itself is an estimate made by actuaries, the people who price future claims, and it moves when their assumptions move. The per-household figures show scale, not a bill.
What this means for your cover
ACC pays for accidents, not illness, and it stops at a published ceiling. Income protection is the cover that answers both. Income protection in New Zealand
Sources
Every source below was read and checked on 21 August 2026.
- Quarter Three Performance Report 2025/26 (March 2026), ACC, 8 July 2026
- Dwelling and household estimates: March 2026 quarter, Stats NZ, 7 April 2026
- Labour market statistics: March 2026 quarter, Stats NZ, 6 May 2026
- Dwelling and household estimates: March 2026 quarter, Scoop Business, 7 April 2026
- New Zealand employed persons, Trading Economics, retrieved 20 August 2026
- ACC earners' levy rates, Inland Revenue, retrieved 20 August 2026
- First impressions: NZ labour market surveys, March quarter 2026, Westpac IQ, 6 May 2026
Talk to a licensed adviser about what this means for you · free, no obligation.
Explore related pages: Life Insurance, Income Protection, Health Insurance, Trauma Insurance, For Brokers.