The 4% Cap Does Not Cap Your Bill. It Caps the Average, and the Council Already Knows Which Doors Are Shut.

The Government is selling its rates cap as a brake on household bills. That is not what the law does. The Local Government (Rates Capping) Amendment Bill, which passed its first reading on 2 September 2026, caps a narrower thing called the rates price. That is the council's total general rates, uniform annual general charges, targeted rates and penalties, divided by the number of properties it rates. The first band is 2 to 4 percent until 30 June 2030. Councils must start planning around it from 1 July 2027. They must obey it from 1 July 2029, unless a regulator gives them a temporary exemption. Water charges are outside it. Fees are outside it. Your own rates bill is not the thing being capped.

Auckland's long-term plan already aims to keep rises at or under 3.5 percent from 2027/28. The hard year is 2026/27. The average residential rise is 7.9 percent, about $321 a year, taking a typical $1.28 million house from about $4,057 to $4,378. City Rail Link is the reason. Owning and running it costs about $235 million a year. About $167 million of that is interest, $42 million is depreciation, and $26 million is the extra cost of running the stations and trains. The council has said the gap between its 3.5 percent path and a 4 percent ceiling is only about $15 million a year. That is the whole buffer. In 2026/27 the group plans to spend $7.35 billion running the city and $3.64 billion building things, against $3.34 billion of rates.

Where the rates dollar goes

Transport takes about 37 cents of every rates dollar. Community services take about 28 cents. Water takes about 8 cents. Economic and cultural spending takes about 6 cents. Depreciation is $1.84 billion. Interest is $729 million. Staff cost $1.37 billion. Group debt was $15.5 billion at June 2026 and is heading toward $21.3 billion. Moody's rates the council Aa2 with a negative outlook, and has already said a rates cap is bad for that rating because it removes the freedom to raise revenue if things go wrong. S&P rates it AA, stable. Auckland is the biggest guarantor of the Local Government Funding Agency, so a downgrade here lifts borrowing costs for councils across the country.

The Government's own officials said the policy did not match the problem and did not meet their quality test. Mayor Wayne Brown's warning, that a cap can end up raising rates, is about interest. If the credit rating slips, the extra interest on this debt wipes out that $15 million buffer several times over. The bottom of the band matters too. A council cannot set the average rise below 2 percent without permission. That floor exists so a council cannot skip renewals and debt payments just to look good before an election.

What households actually get

What households actually get is a slower average path and a long-term plan that is easier to predict. They do not get protection from a revaluation, a shift in who pays, or a Watercare bill. Drinking water, wastewater and stormwater sit outside the cap, including the stormwater share currently hidden inside general rates. Fees and charges are also outside it. The Prime Minister's office and Local Government Minister Simon Watts confirmed that after a messy public mix-up. A single bill can still rise by more or less than 4 percent, because property values and rating categories decide how a capped pot is shared. In 2026/27 the average business property is already up 10.47 percent, against 7.9 percent for homes, because business values grew more slowly and policy says business must keep paying 31 percent of general rates plus the water quality, natural environment and climate action transport rates.

The council loses the ability to go over 4 percent when interest, a big renewal year, or a shared project blows the budget. Growth is the awkward part. New homes and businesses lift total rates income without breaking the per-property test, but the pipes and roads have to be paid for before those properties exist. The Bill's answer is a special exemption, plus tools that make growth pay for growth. If the council keeps undercharging developers, the band gets tighter, not looser. Local boards will feel it before the city centre does. Once transport contracts and interest are treated as fixed, community spending is the flexible part. Grants, events and facility hours get cut first. Road renewals get protected in the speeches and neglected in the programme.

The doors that are already shut

People will look for ways around it. Most of the obvious ones are already shut. The council still sets rates under the Local Government (Rating) Act 2002, and still has to justify who pays under the Local Government Act 2002. The regulator can order a new rates decision if the council, or its long-term plan, breaks the band. From 2030 the test is a three-year average. The year 2029/30 has no average. That is the first year the cap bites cleanly.

A new targeted rate does not beat the cap. Official guidance says existing and new targeted rates, including voluntary ones, count toward the rates price. The water quality rate, the natural environment rate, the climate action transport rate, the waste rate, a new local rate and a business improvement district rate all count. Giving a general rates rise a new name does not change the total.

A higher business differential does not beat the cap either. It only changes who is angry. Auckland already charges urban business about 2.42 times the residential rate in the dollar, and holds business to 31 percent of general rates plus those three targeted rates. Putting more of a fixed sum on business makes a home bill rise by less than the average. The sum divided by the number of properties does not change. The council cannot collect an extra dollar by rating Newmarket harder to spare Mt Eden. Any big shift still has to be justified by who benefits and who causes the cost. Without that, it can be challenged. It is not new money.

Heavier late-payment penalties do not work either, because penalties are inside the rates price. Renaming a rate as a fee fails if it is still a compulsory charge on land for a general service. The regulator looks at the rates decision, not the label on the invoice.

What the council can still do

What the council can still do, legally, is change who pays. Differentials, the uniform annual charge and revaluations all do that. Property value decides your share of the pot. It does not decide the size of the pot. A suburb whose values rise faster than the rest of the city will think the cap has been broken when the bill moves 8 percent. A targeted rate on one area still goes into the citywide total. It can make one group pay for one service. It cannot create a second pot of rates.

The open doors are the ones that are not rates. Fees and charges are the clean one. A regulatory fee cannot be more than the cost of the service, and the council still has to show who benefits. Full-cost consents, commercial rents, venue hire, parking, marina and landfill charges, and public transport fares are allowed. A fee that is really a general rate will be treated as a rate.

Charges on new development are also outside rates. They are meant to work with the cap. If the council keeps discounting them and funding the gap from rates, the band punishes that choice.

The sharpest existing bypass is a levy under the Infrastructure Funding and Financing Act 2020. That levy is not a rate. It sits on the land, gets collected with the rates, and pays for a separate entity that borrows for one named project. Milldale is the working example. It does not count toward the cap. It also does not vanish from the household bill. Used for a real growth asset, it is legitimate. Used to fund ordinary running costs, lenders will not back it, and it will be treated as avoidance.

Water is a real exclusion and a political trap. Moving stormwater into a separate water charge only helps the cap to the extent those dollars were not already being taken out of the sum. Households still pay. Water has its own regulator. Counting the same exclusion twice is exactly what the direction power is there to stop.

Real growth, averages and exemptions

Real growth in the number of rated properties does raise total income inside the rules. The count used is the council's own forecast from the year before. Real new titles count. A padded forecast will not survive the first compliance check.

The three-year average lets one year go over 4 percent if the three years together stay inside the band. A renewal spike can sit in the middle year, with the other two years at 2 percent. It cannot be 6 percent three times in a row. Exemptions are the official way through. One type covers disasters the council could not have planned for, and the Minister decides. The other covers a justified need on top of careful financial management, including growth pipes built before the houses exist, and the regulator decides after the public has been asked. Shared projects that need extra rates need that second type. An application that saves grants while letting renewals slip will fail.

Debt, asset sales, returns from the Auckland Future Fund and transport subsidies are outside the rates price. The interest on the debt is not. Every dollar borrowed to dodge the band comes back later as rates.

The accounting tricks that backfire

The oldest accounting trick is to stop funding depreciation. The law expects a balanced budget unless the council formally decides otherwise, and it must think about maintaining assets and not dumping costs on the next generation. Leaving $1.84 billion of depreciation out of this year's rates cuts the bill now and builds a renewal backlog. The auditor will mark it. The rating agencies will treat it as a credit problem. It can be done. It is how capped councils end up broke. Labelling maintenance as capital spending just brings the cost back as depreciation and interest. Auditors reject the label. Debt inside a council company is still group debt. A lease that behaves like a loan is a loan. Shuffling revenue between columns in the annual report does not change the rates decision the regulator looks at.

The order the pressure moves

The pressure will move in a set order. First to fees and other income, because they sit outside the band. Then to the $106 million of operating savings already promised, which cannot be counted twice. Then to grants and economic development subsidies, about $229 million in the 2026/27 plan. Then to facility hours and local board discretionary money. Then to bus and train service above the contract minimum, and after that to delayed building work, with renewals the real loss. Last comes the credit rating. That turns a clever workaround into a higher interest bill, and then into either an exemption application or a deeper cut.

The path that actually works

The path that actually works is dull. Do not add new facilities unless the full lifetime cost fits inside the 3.5 percent path and leaves the $15 million buffer alone. Charge the real cost for fees, leases and development. Sell or transfer buildings the city does not need, so the depreciation and the renewals shrink for good rather than for one year. Hold transport to a rates-funded budget that stays inside the band. Keep funding depreciation. Use an exemption only for a growth project with a date to return to the band.

Differentials and targeted rates pick the loser. They do not beat the cap. Skipping depreciation and hiding debt are how a 4 percent cap becomes a ratings downgrade. The only lawful way to raise more money for the same services is to charge users where there is a user, charge growth where there is growth, and use a project levy where one asset can carry its own debt. Everything else comes back as interest.