Part I — Situation overview

On 25 September, Zoltán Tanács, Minister of Science and Technology, announced that his ministry, the Ministry of Science and Technology, would file a criminal complaint in the matter of the electronic land registry system (E-ING). According to the ministry’s statement, the complaint requests an investigation on suspicion of budget fraud, breach of trust, use of a false private document, and possible failure to fulfil supervisory or control obligations. The contract for application development was concluded with Tigra Kft. on 13 June 2019. The project received more than HUF 15.9 billion, entirely from EU funding. Sources report the value of the contract differently; according to the ministry’s data, the amended contract was worth HUF 10.73 billion gross. According to the documentation, in May 2022 the final development milestone was also certified as completed, payment was made, and in 2023 the project was officially closed. After the change of government, in summer 2026, the new leadership of the state IT group reviewed the system and assessed its overall completion level at 52 per cent. The back-office interface works, but falters. 10–15 per cent of the functions intended for clients have been completed, and the completion level of the land transactions and agricultural module is zero. On the same day Finance Minister András Kármán appointed Béla Slánicz as president of the Hungarian State Treasury (MÁK). The minister listed among the new president’s tasks the renewal of the outdated pension payment system and the elimination of the institution’s “IT dependency”.

The case is not new. Former independent MP Ákos Hadházy had disputed the project’s accounts for years, and the European Anti-Fraud Office (OLAF) opened an investigation in January 2026. In July MIAK wrote about the procurement model error and vendor dependency in the Gondosóra and KRÉTA cases (Gondosóra and KRÉTA), and in August about the lessons of the Treasury cyberattack (The real lesson of the Treasury cyberattack). There is no final decision in the E-ING case. The named companies and decision-makers are entitled to the presumption of innocence.

MIAK’s reading: the E-ING is instructive not because it was late — that is common in state IT. But because the state certified a half-finished system as complete, paid for it, and for three years nobody knew about it. In this case, independent acceptance before payment, public status reporting on projects and named responsibility were all missing at the same time.

Part II — Foundations in the literature

Three sources provide the interpretative framework. Daniel Kahneman, the Israeli–American psychologist awarded the Nobel Memorial Prize in Economic Sciences in 2002 for his research on decision-making biases, describes the planning fallacy in Thinking, Fast and Slow: planners systematically underestimate cost and time. As a remedy he recommends the “outside view”, that is, using the actual outcomes of similar earlier projects. In the same book he discusses the sunk cost fallacy, because of which organisations keep financing even failing projects. Peter F. Drucker (Austrian–American management thinker, who laid the foundations of the methodology for measuring organisational effectiveness) writes in The Effective Executive that feedback must be built into the decision, and that reports are no substitute for on-site verification. The Criminal Code (Btk., Act C of 2012) also makes failure to fulfil supervisory and control obligations punishable. However, this answers the question of who is responsible for a faulty acceptance only after the fact and in a narrow range of cases. The detailed discussion of the literature — by author, with quotations — can be found in section 6.4 Literature in detail.

Part III — MIAK’s concrete proposal

MIAK proposes three measurable measures. Their common goal is that a state IT system should only count as finished if an independent review has confirmed that it works — and that this status should be publicly trackable by anyone.

3.1 Independent handover-acceptance review as a condition of payment (government decree by 31 March 2027)

MIAK proposes that the Government prescribe by decree that for every state IT development above HUF 1 billion, and for every one financed from EU funds, the final milestone may only be certified after an independent handover-acceptance review. The review should be carried out by an organisation that has no connection either with the developer or with the project organisation certifying performance. The review should examine not the documentation but the system: the client-side functions must be tested on real case types, under a load similar to live conditions. A separate, numerical completion value must be recorded for every module undertaken in the contract. In the E-ING case it was precisely this method that brought to light the 10–15 per cent client-side completion level. 10 per cent of the contractor’s fee must be retained after successful acceptance until six months of stable live operation, or covered by a performance guarantee. The summary of the acceptance report should be public. This is the institutional form of Drucker’s principle of on-site feedback (see 6.4.2). The rule should also apply to the new government’s own commitment: when the ministry reports the E-ING as completed by the end of the year, this too should be confirmed by an independent acceptance review. The proposal applies the KI8 results-based efficiency measurement and the G20 ex-post impact assessment programme points to state IT.

3.2 Public state IT project portfolio (first publication by 31 December 2026, quarterly thereafter)

MIAK proposes that the Government publish a list of all ongoing state IT projects above HUF 1 billion. For each project, the list should include the original and current budget, the original and current completion deadline, the source of funding, the main contractor and known subcontractors, the independent completion value by module, and the names of the project owner and of the organisational unit certifying performance. Every project should receive a simple, three-level (red, amber or green) delivery-confidence rating, given not by the project organisation but by the independent unit overseeing the portfolio. The data should appear in machine-readable form, so that researchers and journalists can also analyse them. For new projects, the cost and deadline plan must be compared with the actual overruns of earlier projects in the portfolio. This is the practical application of Kahneman’s outside view (see 6.4.1): within a few years the list itself becomes the reference class for Hungarian state IT projects. In a functioning portfolio report, the E-ING problem would have been visible by 2023 at the latest, not in 2026. The proposal builds on the A1 public money dashboard, the G1 data-driven budget and the D2 open data programme points.

3.3 Named chain of responsibility and a continuation decision on inherited projects (inventory by 31 December 2026, Treasury action plan within 60 days)

MIAK proposes that for every project in the portfolio it be recorded who the project owner is, who certifies performance and who authorises payment. These should be three different persons, and their names should appear in the internal register for the entire lifetime of the project. That way an ex-post investigation does not have to spend months searching for who signed the 2022 milestone certificate. The supervisory liability rule of the Criminal Code (see 6.4.3) gives an answer only after the fact and only if a crime has been committed; this rule sets out responsibility in advance, for every project. For every IT project above HUF 1 billion launched before the change of government, an independent completion inventory should be prepared using the same method by which the E-ING’s 52 per cent value was established. Only in this way will it become clear whether the E-ING is an exception or a pattern. On the basis of the inventory, a public, reasoned decision should be made on every project: continuation, restructuring or termination. The decision should be based solely on the remaining cost and the expected benefit, not on the amount already spent. According to Kahneman’s sunk cost thesis (see 6.4.1), it is precisely the 16 billion already spent that is the strongest temptation towards a bad decision. The new president of the Hungarian State Treasury should publish an 18-month action plan within 60 days of taking office (28 September). It should set out with what schedule, budget and independent acceptance condition the renewal of the pension payment system will be included in the portfolio, and what the nine-month mid-term checkpoint is. The proposal extends the recovery principle of the A8 cohesion accountability programme point to domestic project management as well.

The common principle of the three proposals is that the word “finished” in state IT should be a verifiable statement, not an administrative act. Criminal proceedings examine whether someone committed a crime. The rule, however, decides whether at the next project a half-finished system can pass acceptance in the same way. Drucker and Kahneman reach the same conclusion: a report and a plan are only worth something if they are regularly compared with reality, from the outside.

Part IV — Expected effects and risks

Dimension Expected effect Risk
Public administration The handover of land registry, Treasury and other back-end systems is based on real operation; a half-finished system does not go live Independent acceptance may delay introduction by weeks or months; reviewer capacity may become a bottleneck
Budget and EU funds The 10 per cent retention and independent acceptance reduce the fees paid for work not performed, and the risk of losing EU funding Suppliers may price in the retention and the stricter acceptance; bid prices may rise
Market and competition The public portfolio and performance histories make visible who delivers reliably For smaller companies, the retention is a liquidity burden; flexibility in the forms of guarantee is therefore important
Transparency Delays and cost overruns come to light quarterly, not years later, and not from an MP’s revelations A red rating may become a target for political attack, which may create an incentive to “prettify” the rating
Services to the public Continuation decisions are made on the basis of the remaining benefit; public-facing functions do not remain on paper A termination decision may temporarily keep case handling in the old, outdated system

The main trade-off is the relationship between speed and certainty. An independent acceptance slows down the handover, and short-term political interest always pulls towards announcing “finished” as soon as possible. The E-ING case shows, however, that a fast but false report of completion is much more expensive in the long run: according to the press, the switch of land registry offices to the incomplete system caused disruption lasting months. The second question is the independence of the rating. If the colour rating in the portfolio report is given by the project organisation itself, the report quickly becomes a cosmetic document. That is why the rating must be assigned to a unit independent of the project, and its methodology made public. The third risk is the political use of the inventory of inherited projects. The inventory does not serve to condemn the previous government, but to underpin continuation decisions. That is why the same method must also be applied to projects launched under the new government.

Part V — Measurability and summary

5.1 What is worth following? (suggested KPIs)

MIAK proposes the following performance indicators (KPIs: Key Performance Indicators) for monitoring:

  • Portfolio coverage: the percentage of ongoing state IT projects above HUF 1 billion that appear on the public list with a budget, a deadline, an independent completion value and a named project owner. Suggested target: 100% by 30 June 2027.
  • Independent acceptance rate: the percentage of final milestones certified from 2027 that were preceded by an independent handover-acceptance review. Suggested target: 100%.
  • Cost and deadline overrun: the median for portfolio projects compared with the original plan, published annually. The first year is the baseline; the suggested target is a reduction in overruns over the following three years.
  • E-ING delivery: the system’s overall and client-side completion level, confirmed by an independent review, at the year-end deadline undertaken by the ministry. Suggested target: every contracted module works, including the land transactions and agricultural module.
  • Inventory of inherited projects: the number of projects for which an independent completion value and a public continuation decision have been produced by 31 December 2026. Suggested target: every project above HUF 1 billion launched before the change of government.

5.2 Summary

MIAK’s request is simple: in state IT, only an independent review should be able to declare something “finished”. The Government should tie the final payment by decree to an independent handover-acceptance review and partial retention. By the end of the year it should publish the public portfolio of IT projects above HUF 1 billion, and name the project owner and the performance certifier for every project. An inventory of inherited projects should be prepared using the same method that revealed the real state of the E-ING. And the new president of the State Treasury should publish a measurable action plan within 60 days. The fate of the criminal complaint will be decided by the investigating authority, the prosecution service and ultimately the court. The rules, however, can be decided on now.

Two of MIAK’s foundational values are at stake here. Accountability, because in the E-ING case performance certification and payment carried no personal, traceable responsibility, so a half-finished system could count as finished for three years. And data-drivenness, because without the actual cost and deadline data of the portfolio, every new project falls into the same optimistic planning trap. Kahneman’s outside view can only be applied if the results of earlier projects are measured and publicly available.

Part VI — Reasoning and further sources

6.1 Press framing by spectrum

In the left-liberal band, Telex, building on the ministry’s written statement, gave the most detailed, numbers-based report. It published the offences named in the complaint, the changes in the contract values, the subcontractor role of 4iG and the completion level by module. As background, it also recalled the OLAF investigation and Ákos Hadházy’s earlier revelations. 444.hu, by contrast, put the sharpest sentence of the minister’s social media video — that “this was probably stolen too” — at the centre of its framing. The paper also quoted Ákos Hadházy’s claims that the contractor had billed for staff who had never been seen working on the project. These are claims by a politician, not established by any authority. Népszava gave a factual summary of the contractual milestones and the setting up of the internal IT team, but mistakenly named the ministry as the Ministry for Innovation and Technology. On the same day, in an opinion piece, Népszava criticised the 2022–2024 school laptop programme. There too the recurring motif is that a project was formally “fulfilled” while its actual usability remained questionable.

In the public-affairs band, ATV carried the news citing the Telex article and highlighting the minister’s video. On the appointment of the new president of the State Treasury, 24.hu listed the finance minister’s three expectations — faster payments, the elimination of IT dependency and an end to overpriced procurement. Telex linked the same news to the summer cyberattack. In the economic band, Portfolio covered both stories. In the E-ING case it highlighted the “false performance certificates”, the ending of vendor exposure and the takeover of the source code. For the Treasury appointment, it emphasised the date of taking office and the widening of the circle of external service providers. The paper indicated that an artificial-intelligence-based assistant had contributed to the preparation of its articles. The pro-government conservative band (Magyar Nemzet, Mandiner) did not put the topic in its top focus on the monitored day.

A common feature across the whole spectrum is that the reports are based on announcements by the current government. The position of the main contractor and of the former project managers does not appear in the downloaded articles. This is a gap from the perspective of balanced reporting, and it is precisely why it would be important for independent, public data on the status of projects to be available.

6.2 Facts and data

Data Value Source
Start of the digitalisation of the land registry 2015 Telex, 25 September 2026
Conclusion of the application development contract 13 June 2019, Tigra Kft. (main contractor) Népszava, 25 September 2026
EU funding more than HUF 15.9 billion, 100% EU funding; according to the minister the project cost more than HUF 16 billion in total Telex, ATV, Portfolio, 25 September 2026
Contract value (ministry statement) originally HUF 8.99 billion net, amended to HUF 8.449 billion net, HUF 10.73 billion gross Telex, 25 September 2026
Contract value (minister’s video) originally around HUF 5.9 billion net, later around HUF 8.5 billion net; more than HUF 10 billion gross paid out Népszava, Portfolio, 444.hu, 25 September 2026
4iG’s subcontracted sub-task more than HUF 2 billion net (according to the company’s 2019 annual report) Telex, 25 September 2026
Certification of the final milestone; closure of the project May 2022; 2023 Telex, Portfolio, Népszava, 25 September 2026
Overall completion level according to the July 2026 technical review 52% Telex, Portfolio, Népszava, 25 September 2026
Completion level of the client-side application 10–15% Telex, Portfolio, 25 September 2026
Completion level of the land transactions and agricultural module 0% (development of the map-based procedures had essentially not even started) Telex, Portfolio, Népszava, 25 September 2026
Committed completion deadline end of 2026, with an internal state IT team; the operation contract with Tigra terminated 444.hu, Portfolio, 25 September 2026
OLAF investigation January 2026 Telex, ATV, 25 September 2026
New Treasury president takes office 28 September 2026 Portfolio, 25 September 2026

Two remarks belong to the figures. First, the original value of the contract is reported differently by the ministry’s written statement (HUF 8.99 billion net) and the minister’s video (HUF 5.9 billion net). It is likely that the two figures relate to different parts or stages of the contract, but the public sources do not clarify this. This in itself is an argument for the portfolio proposed in point 3.2, where the original and current budget would appear in one place, using a uniform method. Second, the 52 per cent completion value was established by the new leadership. It is not a fact established by a court, and the main contractor’s position is not known. This is precisely why it is important that in the future completion be measured by a review independent of the contracting parties.

6.3 Policy dimensions

  • Public administration and e-government (programme points) — the handover-acceptance regime, results-based efficiency measurement and the land registry as a base register for one-stop-shop administration (programme point ID: KI8, KI1);
  • Digitalisation and AI regulation (programme points) — machine-readable publication of project portfolio data (programme point ID: D2);
  • Transparency and anti-corruption policy (programme points) — tracking public money, accountability of EU-funded projects and recovery (programme point ID: A1, A8, A2);
  • Economy (programme points) — data-driven budgeting and ex-post impact assessment for large state investments (programme point ID: G1, G20).

6.4 Literature in detail

6.4.1 Daniel Kahneman: Thinking, Fast and Slow

Kahneman describes the planning fallacy as a feature of plans and forecasts that are close to the best-case scenario and do not take into account the actual outcomes of similar earlier cases. According to an international study he cites, in rail projects launched between 1969 and 1998 planners systematically overestimated passenger numbers, and the projects ended with an average cost overrun of 45 per cent. Over thirty years the lesson did not improve the estimates. Kahneman emphasises that a faulty budget is not always accidental. The authors of plans are often driven by the incentive to get their plan approved, and they know that a programme is rarely stopped merely because it is more expensive or slower than expected. In his view, therefore, responsibility lies primarily with those who approve the plan and do not ask for an outside view. The remedy is reference-class forecasting, associated with the name of the Danish planning expert Bent Flyvbjerg: the cost and time requirement of a new project must be measured against the statistical distribution of similar projects.

Another chapter of the book discusses the sunk cost fallacy. Organisations tend to invest further money in a delayed, increasingly unpromising project, because stopping it would mean admitting earlier failure. Kahneman specifically points out that from the perspective of the manager responsible for the project this error can even be rational, since termination leaves a mark on their career. That is why the manager facing such a decision is often replaced: the new manager does not carry their predecessor’s mental accounts. Applied to the Hungarian case: the public portfolio in point 3.2 would create the domestic reference class without which every new state IT project is planned from the best-case scenario. And the continuation decision in point 3.3 makes use of the moment when a new leadership can weigh things up without the burden of earlier decisions.

📖 Source: Daniel Kahneman: Thinking, Fast and Slow

6.4.2 Peter F. Drucker: The Effective Executive

Of the five elements of effective decision-making, Drucker considers the last to be feedback. It must be built into every decision that the expectations behind it are continuously compared with actual events. The author takes an example from military practice: the officer who gives an order goes himself, or sends one of his own people, to see whether the order has been carried out. Not because he does not trust his subordinate, but because experience shows that reports cannot be relied on. According to Drucker, reports and figures are necessary, but if feedback is not based on direct contact with reality, the decision-maker becomes detached from reality. He also warns that the computer increases this danger, because it takes the decision-maker even further away from the ground.

The book’s introduction also gives a practical rule on the action plan. The executive must set out in advance what results they commit to over the next 18–24 months, and with what deadlines. Two checkpoints must be built into the action plan: one at the mid-point, for example in the ninth month, and one at the end. In the E-ING case, the 2022 milestone certificate was exactly the kind of report that nobody compared with the actual operation of the system. The 2026 review, by contrast, did exactly what Drucker recommends: it looked at what actually works. The independent acceptance review in point 3.1 makes this step a precondition for payment, and the Treasury action plan in point 3.3 applies Drucker’s 18-month plan with a mid-term checkpoint to the new leadership of the State Treasury.

📖 Source: Peter F. Drucker: The Effective Executive

6.4.3 Act C of 2012 on the Criminal Code

Of the offences named in the criminal complaint, the Criminal Code defines the basic case of breach of trust as follows:

“Any person who is entrusted with managing the assets of another and causes financial loss by violating the obligations arising from this entrustment commits breach of trust.” (Section 376 (1))

On failure to fulfil supervisory or control obligations, the text available in the knowledge base, reflecting the law as in force in 2017, states:

“The manager, or a member or employee authorised to exercise control or supervision, of an economic operator who fails to fulfil his supervisory or control obligations and thereby enables a member or employee of the economic operator to commit budget fraud within the scope of the economic operator’s activities, is guilty of a felony punishable by imprisonment for up to three years.” (Section 397)

The two provisions show that criminal law recognises supervisory liability, but under narrow conditions and after the fact. For liability, a crime must have been committed, it must be proven, and the supervisory omission must have enabled it. Criminal law gives no answer to a negligent performance certification that does not constitute a crime. Nor to the question of who in the project organisation signed the milestone certificate on the basis of what evidence. Following the complaint, this must now be uncovered after the fact by the investigating authority. The named chain of responsibility in point 3.3 fills this gap with an administrative rule: the roles of project owner, performance certifier and payment authoriser are separated, and their names are recorded in advance and documented. Criminal liability remains exclusively for the court to decide.

📖 Source: Act C of 2012 on the Criminal Code

6.5 International comparison

Several European and Anglo-Saxon countries have institutionalised portfolio-level, independent control of state IT projects. In the United Kingdom, the government has published the portfolio of major government projects (Government Major Projects Portfolio) annually since 2013. In it, every project receives an independent delivery-confidence rating (Delivery Confidence Assessment, that is, an assessment of the likelihood of successful completion), ranging from green to red. Major projects undergo independent expert review at key decision points. In the Netherlands, since 2015 an independent board has given opinions on state IT projects larger than EUR 5 million. Parliament also receives its opinions, and ministries are obliged to respond to them. In Denmark, the state IT project council has prepared risk assessments of larger state IT developments since 2011, and a regular, public summary is produced on the status of projects. In the United States, since 2009 a public platform (IT Dashboard) has provided information on federal IT investments, on which agencies’ chief information officers give a risk rating to every major investment.

The common element of the models is that the authoritative judgement on the status of a project is not the project organisation’s own report, but that of a unit independent of it, and this judgement is public or at least accessible to parliament. The British experience also shows that a red rating is politically uncomfortable, which is why the independence of the rating and the publicity of its methodology must be specifically protected (see the risk analysis in Part IV). The Estonian model of MIAK’s programme point KI1 is also relevant here: the unified state data exchange system works because the base registers, including the land registry, have actually been completed and are available. The half-finished state of the E-ING weakens one of the foundations of one-stop-shop administration.

Public administration and e-government

  • KI8 — Drucker-based efficiency measurement in public administration
  • KI1 — One-stop-shop digital administration

Digitalisation and AI regulation

  • D2 — Open data programme

Transparency and anti-corruption policy

  • A1 — Public money dashboard
  • A2 — Public procurement transparency
  • A8 — Cohesion policy accountability

Economy

  • G1 — Data-driven budget
  • G20 — Economic policy impact assessment system (Drucker audit)

Proposed new programme point: State IT project portfolio and independent handover-acceptance review — a quarterly public status report for every state IT project above HUF 1 billion, with an independent completion value and a delivery-confidence rating, tying the final payment to an independent acceptance review and partial retention, and a named chain of responsibility (project owner, performance certifier, payment authoriser) — for the Public administration and e-government area.

6.7 List of sources

Press sources (MIAK press monitor, 26 September 2026 — topic 3):

Knowledge base references (books):

  • 📖 Daniel Kahneman: Thinking, Fast and Slow
  • 📖 Peter F. Drucker: The Effective Executive
  • 📖 Act C of 2012 on the Criminal Code

MIAK internal materials:

  • MIAK policy area: Public administration and e-government (programme points; programme point ID: KI1, KI8)
  • MIAK policy area: Digitalisation and AI regulation (programme points; programme point ID: D2)
  • MIAK policy area: Transparency and anti-corruption policy (programme points; programme point ID: A1, A2, A8)
  • MIAK policy area: Economy (programme points; programme point ID: G1, G20)
  • MIAK press monitor, 26 September 2026 — topic 3, score: 85/100

Supplementary public data sources:

  • Electronic Public Procurement System (EKR) — contract data and contract amendments
  • palyazat.gov.hu — data on EU-funded projects
  • Government of the United Kingdom — Government Major Projects Portfolio annual reports
  • United States — federal IT Dashboard (public platform for federal IT investments)

Generation metadata