Two Hundred and Twenty-Six Million

The Luxembourg Inland Revenue, the ACD, is running on infrastructure from the 1970s and 1980s. The government's solution: a €226 million digital overhaul, spread over five years, now before the Chamber of Deputies as a draft law.[1]

€226 million for a tax administration modernisation. MPs called it an enormous sum "for a computer programme." ACD Director Jean-Paul Olinger's response was blunt: the project covers not just the digital solution but also the processing of tax returns, subsequent controls, and ensuring taxpayer compliance. The ACD processes 880,000 tax cards and 580,000 tax returns annually, managing €16 billion in tax revenue. It comprises around 70 different services, handles taxes for 380,000 households and 125,000 businesses, and exchanges four million documents per year. It is not a computer programme. It is the operating system for a significant chunk of the state.

Paper

The most striking detail in the whole story: only a small fraction of private tax returns are completed digitally via MyGuichet. The majority still arrive on paper. In 2026. In one of the wealthiest countries in the world, the tax system that processes €16 billion a year still runs on paper forms and 1970s infrastructure.

Olinger acknowledged this directly. "We are trying to become modern, accessible, and close to the citizens," he said. "Even with AI and the new developments that are coming, we can no longer build upon it." The "it" being the existing system. When the director of the tax administration says the foundation is beyond repair, that is not a suggestion. It is a diagnosis.

The interface problem

The planned new platform is supposed to simplify work for citizens, businesses, and ACD staff. Olinger specifically mentioned wanting "an interface that allows the citizen and the tax official to exchange information with each other." This sounds basic. It is basic. The fact that it needs to be stated as an aspiration rather than an existing feature tells you how far behind the system has fallen.

A tax system that cannot communicate with taxpayers except through paper forms and post is a system that creates friction at every point of contact. Every question requires a letter. Every clarification requires a phone call during limited office hours. Every error requires a new paper form. The cost of this friction is not just inconvenience. It is delay, it is error, it is compliance failure, and it is a measurable drag on the economy.

Why €226 million is both too much and not enough

The MPs who questioned the price tag are right to question it. Large government IT projects have a track record of coming in over budget, underperforming, and sometimes failing entirely. The history of digital transformation in European public administrations is littered with expensive disappointments. The UK's NHS IT programme. Germany's electronic tax system. France's multiple attempts at digitalising social security. The pattern is consistent: ambitious scope, rising costs, delayed delivery.

But the alternative, doing nothing, is not free either. Running a €16 billion revenue operation on 1970s infrastructure has its own costs: inefficiency, errors, the inability to implement policy changes quickly, the impossibility of using modern analytics to detect fraud or improve compliance. The question is not whether to spend money. It is whether this particular €226 million, spent in this particular way, will produce the transformation that is needed.

Olinger at least framed the project honestly. He did not promise savings. He did not promise that the system would pay for itself. He said they can no longer build on what they have. That is a necessary condition for investment. It is not a sufficient one. The sufficient condition is a plan that delivers the system on time, on budget, and with measurable improvements for the people who actually use it: the 380,000 households and 125,000 businesses who currently interact with their tax administration through paper forms and 50-year-old software.

€226 million. Five years. The number is either the beginning of a modern tax administration or the price tag on another expensive lesson about why government IT projects fail. There is no middle ground worth celebrating.