Your Tax Dollar, Keep The Receipt

By ExtraStrength

·6 July 2026·13 min read

You earn $80,000 a year in New Zealand.

PAYE takes around $17,000 off you before you see it.

GST takes another few thousand from everything you buy.

By the time the dust settles, you've handed about $20,000 to the government.

You get a summary from IRD. Neat little dashboard, pretty. Income tax, deductions, balance.

After that: nothing.

You have no idea what happened to your money.

None.

Not a clue.

But somehow a country of 1.3 million people on the edge of the Baltic Sea figured this problem out in 2001. Estonia. About the size of Auckland. They can tell you exactly which civil servant accessed your file and why. Can show you, in real time, where the tax dollar went after it left your account.

Where's my money, Brian?

Let's start with what New Zealand has, because it's not nothing, especially compared to other countries.

myIR is genuinely good. IRD's transformation programme cost the better part of $1.5 billion and produced one of the better tax collection systems in the OECD. Returns are pre-filled. Refunds drop in days. They know when to email you and ask where's your money. As far as collection goes, IRD has it dialled.

Then there's RealMe, launched around 2013, to a bit of a fizz. The government's single-login service. One username and password gets you into 163 government services across 56 agencies.

"Self service."

"Banks accept it."

"Real estate agents use it."

It "runs on Azure."

Sounds great, John. But look at the uptake.

Fewer than 16 percent of New Zealanders have actually registered for it. Internal reviews describe "lower than forecast adoption." Most Kiwis interact with RealMe once every five years, forget their password, and reset it, if you can. I once forgot a password and had to phone up because the reset link did not work.

RealMe is an authentication layer. It tells the government who you are. It does not connect what the government knows about you. Inland Revenue still doesn't talk to MSD. MSD doesn't talk to ACC. ACC doesn't talk to Health NZ. Each agency runs its own data island and asks you to verify yourself separately every time you cross the bridge.

Then there's Hira.

Hira (formerly nHIP, the national Health Information Platform) is the closest thing NZ has built to X-Road. The framing is almost identical to Estonia's. "Islands in the stream." Federated data sources. An API marketplace. HL7 FHIR standards. Real-time access to a citizen's health record assembled on demand from multiple trusted sources. The Ministry of Health spent four years scoping it. They rejected the single electronic health record approach in favour of an interoperable data exchange layer.

They started building.

For health only.

We can see the architecture clearly when it is one sector at a time. We have never built it across sectors. IRD still does not talk to MSD. Hira will not talk to ACC. Health data is being integrated on FHIR while welfare still runs on COBOL. The technology is sound but the overarching architecture is timid.

myIR collects. RealMe authenticates. Hira federates health data.

We built a front door with a single room and no roof.

Estonia built the house.

The House Estonia Built

Estonia is the most advanced digital state on earth. Millions agree.

In December 2024 the country declared 100 percent of its government services digital. Every single one. You can register a business in under an hour. File your taxes in five minutes. Vote online. Get divorced with a mouse click. The only things you still have to do in person are get married, sell property, and pick up an ID card.

The plumbing underneath all of this is called X-Road.

X-Road is a decentralised data exchange layer (no, not crypto, sorry bro) that lets government agencies and private companies share data securely without anyone keeping a central database. Each agency holds its own data. X-Road is the protocol that lets them ask each other questions. Every query is logged. Every transaction is digitally signed and time-stamped.

You give the government your data once.

After that, every agency that needs it asks for it from the agency that already has it. You never re-enter your address. You never upload your birth certificate twice. When a child is born, the system automatically initiates child support benefits and pre-fills the parental leave application before the parents even ask.

There is also a citizen-facing audit log. You can see, in your own portal, who accessed your records and when. If they didn't have a reason, you can complain and they get fined.

X-Road is open source. Free. Available on GitHub. Already adopted by Finland, Iceland, the Faroe Islands, Greenland, Japan, the Mexican state of Quintana Roo, and around 25 others.

Estonia had no advantages we don't have. Smaller population, a fresh start right after the Soviet Union collapsed, no legacy systems to protect. They built it on a shoestring in the early 2000s while we were still mailing tax receipts.

The technology to do this in New Zealand has existed for 25 years. It costs almost nothing and it is sitting in a public repository.

X Marks The Road

When IRD collects your $20,000 and sends it to Treasury, Treasury allocates it across votes.

Health gets $30 billion. Education gets $19 billion. Welfare gets $50 billion. Backbenchers get an allowance for housing. Winnie gets his durry allocation.

From there, the money flows into department budgets, then into programmes, then into contracts, then into vendor payments, then into salaries and consultants and contractors and procurement.

At every step the audit trail gets thinner.

By the time your dollar lands as a payment to a third-tier subcontractor on a Health NZ IT project, there is no real way to trace it back to your PAYE deduction. Nobody can. The system was never designed for it.

The Government released its formal response to the National Infrastructure Plan at the time of writing. The numbers are not looking good. NZ spends 5.8 percent of GDP on infrastructure, one of the highest rates in the OECD. We rank toward the bottom for efficiency. Fourth to last in the OECD for asset management. Infrastructure Minister Chris Bishop put it plainly: "Many central government agencies do not properly understand what they own or have long-term investment plans." That should be on the next election billboard.

The proposed fix is legislation in 2027 to amend the Public Finance Act so agencies have to publish their plans. Better data "over time." A "professional standard for public sector leadership." Twenty-five years after Estonia built X-Road, the official solution to NZ government not knowing what it owns is to ask agencies, more firmly, please write it down.

This is where programmable money should walk in, and I don't mean the to-the-moon speculative kind.

Imagine for a second with me. Every time you get taxed, your tax gets tagged as it enters the system. As an agricultural country, this should not be hard to imagine, just a little note on the ear of every dollar. Every dollar carries metadata: source, vote allocation, programme, contract, vendor. Smart contracts release procurement payments only when delivery milestones are met. Citizens can pull up a public dashboard and follow the money from PAYE deduction to end spend.

Put simply, it's just accounting software with a public ledger.

I know, easier said than done, but this is not some radical idea.

RFK Jr. ran on it in the US. "300 million eyeballs on our budget," he said. "If somebody is spending $16,000 for a toilet seat, everybody will know about it." The much-lamented DOGE explored blockchain spending tracking in early 2025. Did it work? Probably not. But the direction is not wrong. Wyoming launched a state-backed stable token in 2024. The Government Blockchain Association in the US is running a four-phase pilot to put county budgets on chain.

The question is no longer whether the technology works.

It works*

*Nobody wants to be scrutinized so lets just, not do that yeah?

Take KiwiBuild. $2.1 billion allocated. 100,000 homes promised in 10 years. 258 delivered. Overhead per home produced: $300,000. A private developer working at scale runs at around $5,500 in overhead per dwelling.

If every dollar of that $2.1 billion had been on a public ledger, with smart contracts pausing disbursement at missed milestones, the failure surfaces in month three. Not year two. But nobody in government wants that. That is the actual problem.

The Real UBI Problem

UBI gets shot down the same way every time.

Higher taxes. Can't afford it. Where does the money come from. Money for druggies on the doll.

That reaction misses what the serious proposals actually say.

The Opportunity Party, AKA TOP, AKA Gareth Morgan yelling at the sky, is heading into the 2026 election on a Citizen's Income of $370 a week for almost all adults. It's funded by a 1.75 percent land value tax.

For most working people, income tax goes down. The tax burden shifts from labour to land.

Some say its a more tax proposal, another way to look at it is a who-pays-and-what-you-get-back proposal.

The Greens' version, their 2023 income guarantee, was $385 a week minimum for anyone out of work or studying, funded through a wealth tax and higher rates on top incomes.

Different mechanism, similar argument: take less from people who work, take more from accumulated assets, make sure the floor is real.

Neither party can tell you how they'd actually administer it cleanly.

Even people sympathetic to a citizen's income assume any government program will leak, bureaucratically multiply, and produce a department of 2,000 staff testing applications. The whole point of these proposals is that there is no test. No eligibility check. No application.

Every qualifying adult gets it. Every fortnight. Done.

But you cannot run a clean citizen's income without the infrastructure underneath.

You need verified identity at population scale. An authoritative record of who is a citizen. An automated payment rail that does not need MSD to review each transfer. Traceable tax revenue on one side and traceable dividend payment on the other. An audit log a citizen can read to verify their own contribution and their own receipt.

That is X-Road plus a population-scale identity layer plus programmable money. Exactly the architecture we have not built.

Without it, any citizen's income recreates every bureaucratic problem it was meant to dissolve. MSD reviewing applications. Eligibility disputes. Fraud detection teams. Appeals processes. The welfare state it was supposed to simplify becomes the thing that administers it.

With it, it's the simplest government programme ever designed.

Tax in, payment out, every transaction visible to the person who funded it.

Nothing to administer.

The political fight stops being about whether you trust the government to handle the money.

On My Desk, Monday.

Opacity benefits the people who design the system.

Politicians do not want millions of eyeballs on their budget.

Its ‘complex’ by design.

Agencies do not want their procurement on a public ledger.

Consultants on long-term government contracts do not want milestone-based smart contracts that pause payment when deliverables slip. And the administrative state that runs welfare, grants, and disbursement does not want a payment rail that makes most of its job redundant.

NZ has decades of Oracle, SAP, mainframe COBOL, custom Microsoft stacks, and bespoke departmental systems welded together with API duct tape.

Replacing it is hard. Replacing it without breaking the country is harder. The procurement contracts to maintain it are written by the vendors who profit from it staying complex.

The numbers on this are not small.

Catalyst, the Wellington-based open source firm that has been building NZ government tech for 25 years, calculated that NZ government spends over $12 billion a year on IT in siloed, ad hoc fashion.

The Government's own new digital direction projects up to 30 percent savings on a $13 billion public sector tech spend. That's $3.9 billion. Roughly half the entire defence budget. Achievable by switching to open standards and local expertise, the same model Estonia used.

The savings have sat there for years.

Australia has the same problem at bigger scale. myGov has over 25 million users. Backend interoperability is still fragmented. Only 11 percent of Australian public servants report being able to easily access data across agencies. The Digital Transformation Agency keeps publishing strategy documents about whole of government while Centrelink and Medicare and the ATO continue to operate as separate data fiefdoms.

The US is worse. The Department of Defense has failed its audit six years running. It cannot account for trillions in assets. USAID lost the bulk of its programmes in 2025 partly because nobody could trace what the money was actually doing.

The technology exists everywhere it's needed. The political incentive doesn't.

Estonia built X-Road because they had no choice. Post-Soviet, broke, sparsely populated, and one major Russian cyberattack away from collapse. Necessity moved them. Comfort is what's stopping everyone else.

Feel Helpless?

Sometimes it's just easier to not wait for government to fix itself.

Treasury publishes the Financial Statements of the Government every year.

They are unreadable.

Hundreds of pages of accounting language with no plain-English breakdown of where the money actually went.

The Government Electronic Tenders Service publishes every contract awarded, structured badly, going stale fast, and is a all round terrible experience.

Some legend literally just decided fuck it and built the better version: https://bettergets.nz/

But this approach of build first should be applied elsewhere.

Identity, for example. The Trust Framework Authority launched in 2024, which means RealMe is no longer the only way to verify who someone is.

Banks, real estate, education providers, and gambling operators all need that verification, and the plumbing between the old agency systems and the new third-party identity providers is aching for an update.

AML and KYC compliance requirements give you the regulatory tailwind.

This is a B2B SaaS that government itself will eventually have to fund.

So What.

Estonia built X-Road because they had nothing else.

NZ has different conditions but a similar opportunity.

Small enough to move fast. Digitally literate population. Functional collection infrastructure. A health platform that has already proved the architecture works. A government that admitted it doesn't know what it owns. A quiet groundswell of people who know those voted in are not spending correctly.

The technology to do this exists.

It's open source.

It's free.

It's running in 25 countries.

The business cases are obvious. Anyone reading this could pick three of the ideas above and have a working prototype by Christmas.

The only missing piece is political will. And political will tends to arrive in the same window where the public starts demanding receipts.

We pay the bill every year, but we've never been shown what we bought.

Shout out to all the legends I've ripped info from for this piece:

Key takeaways

New Zealand's IRD collects tax efficiently through myIR but provides no way for citizens to trace spending after collection. Estonia's X-Road, launched in 2001, is a free, open source, decentralised data exchange layer now used by 25+ countries that lets citizens audit which government agency accessed their data and why. NZ has built equivalent pieces in isolation, myIR (tax), RealMe (authentication, under 16% adoption), and Hira (health data federation), but none connect to each other or across agencies. This missing connective infrastructure is also the primary technical obstacle to running a low-bureaucracy citizen's income, as proposed by TOP (Citizen's Income) and the Greens (income guarantee), and represents unbuilt white space in civic tech and government-adjacent SaaS, including public spending transparency tools and cross-agency identity verification.