Part I — Situation overview

On 25 September 2026 three mutually independent accountability steps came to public attention. The Integrity Authority confirmed a report by 24.hu, which HVG then picked up: in the Authority’s asset-declaration review, more than a hundred people received the highest risk rating. They include Miklós Seszták, former KDNP development minister, and Balázs Hankó, former Fidesz minister for culture and innovation. According to the papers, the Authority developed the methodology of the review by this summer, and the rating was completed in early September. The data of the declarations, the change in wealth over time and other available data are compared. It is a risk-increasing factor if someone, as a senior political or state leader, stood at the top of the distribution system, had control over significant resources, or could override controls. The Authority stressed that the rating does not mean that someone has committed a violation, only that their declaration warrants a more thorough official review. On the same day the Ministry of Interior’s State Secretariat for Sport ordered an ex-post audit of the grant to the Fradiváros project. And the National Assembly’s Committee of Inquiry into Enforcement Abuses announced that it would oblige 65 people to give statements, including Antal Rogán and several former justice ministers, such as Judit Varga.

MIAK wrote about the reform of the asset-declaration system on 9 June: it proposed a machine-readable database, automatic cross-checking and sanctions. On the same day a MIAK analysis appeared on the independence of the Integrity Authority and the methodology of corruption loss estimates (The Integrity Authority and corruption losses), and on 17 June another on the dismissal of the Authority’s president. MIAK discussed the procedural guarantees of parliamentary committees of inquiry on 28 May, and the decision trail of the Fradiváros grant on 29 August. On 18 September it wrote about the planned investigation of politicians’ asset accumulation by the National Tax and Customs Administration (NAV) (Investigation of politicians’ asset accumulation). The proposal at that time was still a draft bill. Today’s news is different: an authority has already carried out risk-based screening, and the first names have become public through the press. The suspension of the immunity of Miklós Seszták and Balázs Hankó was requested by the prosecution service in a separate procedure; the National Assembly will decide on it on 28 September. There is no final judgment against any of the named persons, and all of them are entitled to the presumption of innocence.

MIAK’s reading: risk-based screening is the right tool, and precisely the kind of data-driven approach that MIAK itself proposes. A filter is credible, however, only if its criteria are public, it applies equally to everyone, and no one treats its result as a verdict. On the same day, the grantor’s ex-post audit started weeks after the criminal complaint, and the parliamentary committee would also hear persons involved in a pending criminal case. These show that there is not yet a fixed division of labour and sequence between the institutions involved in accountability. What is at stake is not any single name, but whether the outcome of the proceedings will stand up both in court and before the public.

Part II — Foundations in the literature

Three sources provide the interpretive framework. Robert Klitgaard, American economist and author of the classic handbook of anti-corruption policy, describes a Philippine tax authority reform in his work Controlling Corruption. There, staff members were sorted into audit groups on the basis of “red flags”, but the author emphasises that the flag in itself did not identify a corrupt official; it only decided who had to be audited first and more thoroughly. Susan Rose-Ackerman, professor of law and economics at Yale University, compares politicians’ conflict-of-interest and asset-disclosure rules in her book Corruption and Government. She warns that if the response to scandals is only a search for culprits instead of a reshaping of the underlying institutional incentives, anti-corruption action looks like the persecution of political opponents. Article 6(2) of the European Convention on Human Rights (ECHR) enshrines the presumption of innocence, and Article 8 the respect for private life. It follows that an official risk flag and a finding of guilt are two separate legal categories. 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 aim is that the outcome of risk-based screening and of the accountability proceedings running in parallel be credible both legally and before the public — regardless of who happens to be in government.

3.1 Public methodology and aggregate results (within 30 days, then quarterly)

MIAK proposes that the Integrity Authority publish the methodology of the asset-declaration risk rating within 30 days. It should include the risk factors, at least an order-of-magnitude description of their weighting, the category thresholds and the data sources used. The Authority should also publish the aggregate results: how many declarations it reviewed, how many people ended up in each category, and how this breaks down by type of office (former and current member of the Government, member of the National Assembly, mayor, other). MIAK asks for public confirmation that the screening is run with the same methodology on current members of the Government and MPs as well. If, owing to the design of the first round, it affects almost exclusively former office-holders, the Authority should justify this. The methodology should be reviewed by an independent expert panel or a university research centre, and its opinion should be published. The Authority should also report quarterly how many of the cases placed in the highest category were closed with a substantiated finding, and how many proved unfounded. This hit rate shows whether the filter works well. Klitgaard’s Philippine case (see 6.4.1) describes exactly this feedback: there, the credibility of the flags came from the fact that the results of the selected audits were measured afterwards. The proposal is a direct application of the A3 asset-declaration and the A6 checks-and-balances programme points.

3.2 A rating is not a finding: aggregate data instead of names (immediately, by internal rules; statutory codification by 31 December 2026)

MIAK proposes that the Integrity Authority, the NAV and the National Asset Recovery and Asset Protection Office (NVVH) lay down in internal rules that they do not provide information about the risk rating of individual persons and do not confirm press reports either. The Authority may only make public a closed investigative finding or a criminal complaint that has been filed. Even then, any wording that presumes guilt in advance must be avoided. The Authority should notify the rated person and give them a deadline to explain the change in their wealth in writing before any request or criminal complaint is prepared. If the person concerned makes their rating public themselves, the Authority may provide factual information on the state of the procedure. The National Assembly should raise the rule to statutory level by the end of 2026. This is the practical application of the ECHR’s presumption of innocence and protection of private life (see 6.4.3). The content of an asset declaration is data of public interest, but an official risk rating is the starting point of an investigation, not its result. MIAK’s programme point A10 proposes reversing the burden of proof in cases of unexplained asset growth. Klitgaard’s Hong Kong example (see 6.4.1), however, shows that this is only possible on a statutory basis, in a defined procedure and under judicial control. The risk rating meets none of these conditions.

3.3 A fixed institutional division of labour and sequence of procedures (procedural agreement within 60 days)

MIAK proposes that the bodies involved in accountability — the Integrity Authority, the NAV, the NVVH, the granting ministries and the prosecution service — conclude a public procedural agreement within 60 days. It should set down which body forwards which signal to whom, and in what order they act. The basic rule should be that in grant cases the grantor first carries out an ex-post audit, records its findings in writing, and only on that basis decides on recovery and a criminal complaint. If the criminal complaint precedes this, the grantor should publicly justify why. In the Fradiváros case MIAK asks the state secretariat responsible for sport to publish the grant deeds and their amendments, the club’s final reports, as well as the schedule of the ex-post audit and, on its completion, its report. For parliamentary committees of inquiry MIAK proposes three procedural limits. If the person heard is the accused in pending criminal proceedings in the same case, the committee should inform them in writing that they cannot be compelled to incriminate themselves. In such cases the committee’s questions should concern the functioning of the supervisory system, not criminal liability. The committee’s final report should not establish individual criminal liability — that is the competence of the courts. Owing to the separation of powers, a sitting constitutional judge or judge should not be compelled to appear. According to Rose-Ackerman’s warning (see 6.4.2), the credibility of accountability comes from institutional order, not from the number of named persons. The proposal builds on the KI7 official-selection and the I1 judicial transparency programme points.

The common principle of the three proposals is that screening, investigation and verdict are three separate steps, with three separate institutional responsibilities. Risk-based screening is effective if it runs quickly and broadly. That is exactly why its method must be made public and its individual results treated confidentially. Klitgaard and Rose-Ackerman arrive at the same point: a red flag is a good tool for directing attention, but if it is used as a verdict, anti-corruption action loses its credibility.

Part IV — Expected effects and risks

Dimension Expected effect Risk
Anti-corruption effectiveness A public methodology and measurement of the hit rate improve the filter’s accuracy and direct the Authority’s capacity to the highest-risk cases Publishing an overly detailed methodology could help declarants arrange their wealth so as to avoid the risk factors
Rule of law and fundamental rights Distinguishing the rating from guilt reduces the likelihood of reputational harm and of later damages claims The public may interpret the ban on confirming names as secrecy if the Authority does not regularly publish aggregate data
Political credibility Screening that also covers current members of the Government refutes the charge of selective proceedings If the first round nonetheless includes almost exclusively former office-holders, the methodology may become the subject of political dispute
Public administration The procedural agreement reduces parallel and mutually contradictory proceedings A fixed sequence may slow down a criminal complaint where there is a risk of asset stripping or destruction of evidence
Parliamentary oversight Committees of inquiry focus on the responsibility of the system, and their findings do not clash with pending trials A narrower scope of questioning may reduce the political impact of the hearings and public interest in them

The main question to weigh is the relationship between publicity and fair procedure. The public rightly expects to see who is being investigated. A risk rating, however, is a statistical probability that may also rest on incomplete and inaccurate data. If a name becomes public at this stage, a later acquittal no longer restores the person’s good reputation, and a wrongly flagged person may challenge the entire procedure through legal channels. MIAK therefore proposes full publicity of the criteria and the aggregate data, while tying the publicity of individual results to the closure of the investigation. The second question is urgency. Where there is a risk of asset stripping, a criminal complaint cannot wait for the ex-post audit to be closed. The procedural agreement must therefore contain the exception, together with a duty to justify it. The third question is the powers of the committees of inquiry. In June the National Assembly tightened the obligation to appear and to tell the truth. This strengthens parliamentary oversight, but for those accused in pending criminal trials it may run up against the prohibition of compelled self-incrimination if the committee does not clearly draw the limits of questioning.

Part V — Measurability and summary

5.1 What is worth following? (suggested KPIs)

MIAK proposes the following key performance indicators (KPIs) to follow:

  • Methodological publicity: whether the methodology of the risk rating and the aggregate category distribution are published by 26 October 2026. Suggested target: yes, followed by quarterly updates.
  • Equal coverage: what percentage of current members of the Government and members of the National Assembly have had their declarations put through the same screening. Suggested target: 100% by the end of 2026.
  • Hit rate: of the cases placed in the highest category, how many close within 12 months with a substantiated finding, how many are classified as unfounded, and how many are still pending. Worth following quarterly.
  • Number of individual name confirmations: in how many cases an authority confirmed a press report on a person’s risk rating before a closed finding. Suggested target: 0.
  • Fradiváros documents: whether the grant deeds, their amendments and the final reports have been published, and whether the ex-post audit has been closed with a public report. Suggested deadline: 31 March 2027.

5.2 Summary

MIAK’s request is simple: risk screening should be public in its method and confidential in its individual results. The Integrity Authority should publish the rating criteria and aggregate results within 30 days, and confirm that the same screening also applies to current members of the Government and MPs. The authorities should not comment on individual names before a closed finding. And the bodies involved in accountability should publicly set down who investigates what, and in what order. Parliamentary committees of inquiry uncover the responsibility of the supervisory system. Criminal liability is decided by the courts.

Two MIAK foundational values are in play here. Transparency, because an official filter can only be checked if its criteria and aggregate results are public. Without this, no one can judge whether the selection is impartial. And accountability, because accountability itself must be accountable. The authority must measure how many of its flags proved substantiated, and must answer for it if the good reputation of a wrongly flagged person is harmed.

Part VI — Reasoning and further sources

6.1 Press framing by spectrum

In the public-affairs and left-liberal band, HVG and 24.hu put the names at the centre. The headlines of both articles highlight Miklós Seszták and Balázs Hankó, while on the methodology and the composition of the “more than a hundred” people they only relay the Authority’s brief statement. Within a single paragraph, HVG linked the rating with the immunity cases of the two politicians. This easily fuses the official risk flag and the prosecution’s motion into a single story in the reader’s mind, although the two are independent of each other. At the same time the paper quoted verbatim the Authority’s warning that the rating is not a finding of a violation. The part of the 24.hu article on the methodology is behind a paywall. Telex presented the committee of inquiry’s list factually, as a list of names, and flagged a contradiction between its own sources: the document published on the parliament’s website and the recording of the committee meeting differed on whether constitutional judge Ágnes Czine would be heard. In the Fradiváros case Telex highlighted the position of the club’s president and the structure of the amounts, while ATV highlighted a journalist’s criticism. According to ATV, the state secretariat did not explain the change in its position, nor did it publish the grant deeds. Népszava covered a parallel thread, the pre-election sponsorships of Szerencsejáték Zrt. (the state lottery company). According to the paper’s own calculation, around 40 per cent of the contracts concluded between January and April went to circles close to the government.

In the economic band, Portfolio framed the hearings linked to the Schadl case as an “extraordinary development”. It published a fuller list of names than Telex, and highlighted that appearing before the committee and telling the truth are mandatory under the rules tightened in the summer. Telex’s economics section presented the legal uncertainty surrounding the NVVH’s operation in connection with a bank transaction. According to the paper, a law firm received no prior opinion from the police, the prosecution service or the NVVH on the risk involved in purchasing an affected asset. In the pro-government conservative band, Magyar Nemzet put the capacity of the asset recovery office and the additional burden falling on the courts and prosecution offices at the centre. According to the paper’s sources, the office may lure away experienced prosecutors and police officers with higher salaries. This framing approaches the question of the institutional division of labour from the side of capacity. Mandiner did not put the topic in its top focus on that day. Common to the whole spectrum is that the position of the named former office-holders does not appear in the publicly available parts of the articles.

6.2 Facts and data

Data Value Source
Persons given the highest asset-declaration risk rating more than a hundred, including two former ministers HVG, 25 September 2026 (based on the 24.hu report and the Authority’s confirmation)
Timing of the rating methodology: summer 2026; rating: early September 2026 HVG, 24.hu, 25 September 2026
Risk-increasing factors disclosed by the Authority position at the top of the distribution system, control over significant resources, the possibility of influencing decisions and overriding controls; the change in wealth over time HVG, 25 September 2026
State grant to the Fradiváros project HUF 24.947 billion (2016–2017) Telex, 25 September 2026
Use of the first phase according to the club’s final report of HUF 10 billion, more than 8.5 billion on material and operating expenses, around 1.44 billion on investment Telex, 25 September 2026
Use of the second phase according to the club’s final report of HUF 14.9 billion, more than 3.1 billion on operating expenses, 11.2 billion to FTC Labdarúgó Zrt. for acquiring player rights Telex, 25 September 2026
Procedural steps in the Fradiváros case criminal complaint to the NAV on 28 August 2026; ex-post audit ordered on 25 September 2026 MIAK, 29 August 2026; ATV, 25 September 2026
Hearing list of the committee of inquiry into enforcement abuses 65 people, in five topics, orally or in writing Telex, Portfolio, 25 September 2026
Grant and sponsorship contracts of Szerencsejáték Zrt., January–April 2026 HUF 6.25 billion net; according to the paper’s calculation around 2.3–2.4 billion to circles close to the government; no new contracts concluded from May to August Népszava, 25 September 2026
Hungary’s control of corruption score according to the World Bank’s Worldwide Governance Indicators (WGI) −0.17 World Bank WGI 2024

Two remarks belong with the figures. First, within the “more than a hundred” people the Authority did not disclose the distribution between categories, the total number of declarations reviewed, or the breakdown by type of office. It is therefore currently impossible to judge what share of the reviewed group the highest category represents. Yet this is the most important indicator of how strict the filter is. Second, the amounts in the Fradiváros case come from the club’s own final reports. Whether the accounting of the operating expenses was lawful depends on the content of the grant deeds, which were amended several times. This is exactly what the ex-post audit has to establish.

6.3 Policy dimensions

  • Transparency and anti-corruption policy (programme points) — methodological publicity of asset-declaration screening, the independence and measurable performance of the audit institutions, the procedural conditions for handling unexplained asset growth (programme point ID: A3, A6, A10, A12);
  • Justice (programme points and background material) — the presumption of innocence, the relationship between parliamentary inquiry and pending criminal trials, the burden on the courts from asset-recovery cases (programme point ID: I1);
  • Public administration and e-government (programme points) — the division of labour among the bodies involved in accountability, the grantor’s ex-post audit and proof of proportionality of wealth (programme point ID: KI7).

6.4 Literature in detail

6.4.1 Robert Klitgaard: Controlling Corruption

Klitgaard discusses the clean-up of the Philippine tax authority in the 1970s as one of the model cases of anti-corruption reform. The commissioner leading the reform first had the financial situation of the top 125 managers examined. In roughly 30 per cent of them wealth seemed disproportionate to visible income, and in the commissioner’s words he was able to “take a much closer look” at these cases. Then a three-element flagging system was built for the roughly two thousand tax collectors: late remittance, untraceable remittance, and a discrepancy between the remittance report and the taxpayer register. Here the author makes the remark most important for the lesson:

“Some of the data were unavailable and some were inaccurate, so these ‘red flags’ did not automatically identify corrupt officials.” (MIAK’s translation)

The system therefore ranked; it did not judge. Where all three flags were present, an immediate audit followed; where two, an audit according to capacity. The credibility of the method came from the fact that the results of the audits were measured afterwards: in the first two years irregularities were found in every selected case. The Hong Kong chapter also shows the limit. There, the 1971 bribery ordinance reversed the burden of proof in cases of wealth exceeding income, and made unexplained asset growth a criminal offence. Klitgaard explicitly describes this as a departure from the usual presumption of innocence. This, however, required a statute, a defined procedure and a written deadline for the suspect to give an explanation. Applied to the Hungarian case: the Integrity Authority’s rating is the equivalent of the Philippine red flag, not of the Hong Kong procedure. The hit-rate measurement of point 3.1 and the explanation deadline of point 3.2 turn this difference into an institutional rule.

📖 Source: Robert Klitgaard: Controlling Corruption

6.4.2 Susan Rose-Ackerman: Corruption and Government

Writing on politicians’ conflicts of interest, Rose-Ackerman starts from the premise that today the main danger is the use of public office for private financial ends. A conflict of interest exists even if there is no bribery, only an interest of the politician or their family in a company doing business with the state. Most mature democracies therefore require at least the disclosure of MPs’ financial interests. The author points out, however, that member states differ widely in what must be declared, whether the declaration is public, and what sanction applies. A disclosure system in itself therefore says little. What matters is how it is verified and in what procedure it is used. In the chapter on reforms the author puts it sharply:

“Otherwise, anti-corruption campaigns will be little more than witch-hunts that disproportionately target the system’s political opponents.” (MIAK’s translation)

The “otherwise” case is when the response to a scandal is exhausted in punishing the culprits, and the reshaping of the underlying incentives is omitted. Regarding independent anti-corruption offices, the author also notes that even the Hong Kong office is not free of the accusation of being used against political opponents. Where commitments are weaker, broad powers can also be abused. Applied to the Hungarian situation: in a first screening round after a change of government, those flagged are naturally mainly office-holders of the previous government. That is precisely why, under point 3.1, it must be demonstrated that the filter runs in the same way on those currently in office. And under point 3.3, the credibility of accountability comes from a fixed institutional order, not from the number of named persons.

📖 Source: Susan Rose-Ackerman: Corruption and Government — Causes, Consequences, and Reform

6.4.3 Council of Europe: European Convention on Human Rights

In its Article 6 on the right to a fair trial, the ECHR enshrines the presumption of innocence:

“Everyone charged with a criminal offence shall be presumed innocent until proved guilty according to law.” (MIAK’s translation)

Under Article 8, a public authority may interfere with the right to private life only in cases provided for by law, and only insofar as is necessary in a democratic society, for example for the prevention of crime. Together the two articles provide the legal yardstick for the risk rating. The rating is not a charge, so Article 6 does not yet apply to it in the narrow sense. If, however, an authority confirms the names of individual persons as members of a “highest-risk” group, this may create a prior appearance of guilt in the eyes of the public, and may qualify as an interference under Article 8. Such an interference must in turn show a legal basis and a justification of necessity. Article XXVIII of Hungary’s Fundamental Law sets down the same principle: no one shall be considered guilty until their criminal liability has been established by the final decision of a court. Point 3.2 therefore does not restrict publicity; rather, it separates the object of publicity. The criteria and aggregate data are fully public; the individual result is not, until the investigation is closed. The committee-of-inquiry limit in point 3.3 — namely that the accused cannot be compelled to incriminate themselves — likewise follows from the fair-trial guarantees of Article 6.

📖 Source: Council of Europe: European Convention on Human Rights (1950), Articles 6 and 8

6.5 International comparison

Risk-based verification of asset declarations has several proven European models, and each of them separates screening from findings. In France, the independent authority responsible for transparency in public life (Haute Autorité pour la transparence de la vie publique) has since 2013 verified the asset and interest declarations of members of the Government, MPs and senior officials. The interest declarations of members of the Government and MPs are public. In cases of serious doubt the authority turns to the prosecution service: the decisive step is thus a substantiated, written referral, not a risk list. In Romania, according to the background material of MIAK’s programme point A3, the National Integrity Agency (ANI) has since 2007 initiated proceedings against more than a thousand public figures over discrepancies in asset declarations. The person concerned may challenge the agency’s evaluation report before an administrative court. In Ukraine, the electronic asset-declaration system of the corruption prevention agency uses automatic, risk-based pre-screening to select which declarations undergo full verification. In the United Kingdom, an unexplained wealth order can since 2018 only be issued by a court, at the request of the authority. This is a moderate version of the Hong Kong reversal of the burden of proof, placed under judicial control.

There is also a European yardstick for extending the presumption of innocence to official statements. In 1995, in the case of Allenet de Ribemont v. France, the European Court of Human Rights held that the presumption can be violated not only by a court but also by other authorities if, in their public statements, they present someone as guilty. Directive (EU) 2016/343 of the European Union explicitly requires member states to ensure that public statements by authorities do not refer to a suspect as guilty until a final judgment. MIAK’s point 3.2 extends this yardstick to the official risk stage preceding criminal proceedings.

Transparency and anti-corruption policy

  • A3 — Publicity of asset declarations
  • A6 — Strengthening checks and balances
  • A10 — Independent Corruption Investigation Office (CPIB model)
  • A12 — Campaign-finance transparency and upper limit

Justice

  • I1 — Judicial transparency

Public administration and e-government

  • KI7 — Official selection and rotation system

Proposed new programme point: Procedural rules for risk-based integrity screening — a public methodology and aggregate results, the same screening for current and former office-holders, a ban on confirming individual names before a closed finding, an explanation deadline for the person concerned, a quarterly hit-rate report — for the Transparency and anti-corruption policy area.

Proposed new programme point: Accountability procedural agreement — a publicly fixed sequence and limits of competence for the grantor’s ex-post audit, recovery, criminal complaints and parliamentary inquiry — for the Public administration and e-government area.

6.7 List of sources

Press sources (MIAK press monitor, 26 September 2026 — topic 4):

Knowledge base references (books):

  • 📖 Robert Klitgaard: Controlling Corruption
  • 📖 Susan Rose-Ackerman: Corruption and Government — Causes, Consequences, and Reform
  • 📖 Council of Europe: European Convention on Human Rights (1950)

MIAK internal materials:

  • MIAK policy area: Transparency and anti-corruption policy (programme points; programme point ID: A3, A6, A10, A12)
  • MIAK policy area: Transparency and anti-corruption policy (background material)
  • MIAK policy area: Justice (programme points; programme point ID: I1)
  • MIAK policy area: Justice (background material)
  • MIAK policy area: Public administration and e-government (programme points; programme point ID: KI7)
  • MIAK press monitor, 26 September 2026 — topic 4, score: 85/100

Supplementary public data sources:

  • Integrity Authority — reports and announcements
  • Website of the National Assembly — documents and minutes of the committees of inquiry, MPs’ asset declarations
  • World Bank — Worldwide Governance Indicators (WGI), 2024
  • Directive (EU) 2016/343 of the European Parliament and of the Council on the strengthening of certain aspects of the presumption of innocence
  • European Court of Human Rights — Allenet de Ribemont v. France (1995)

Generation metadata