Transparency & anti-corruption

In-depth analysis

A1 — Public-spending dashboard

• Mechanism: A real-time data feed from the Hungarian State Treasury, EKR (the Electronic Public Procurement System) and NAV (the Hungarian Tax and Customs Administration) databases. The platform publishes expenditures in a machine-readable (API-queryable) format, categorised by sector, beneficiary and region. Users can build their own analyses and set anomaly alerts. Models: Slovakia’s e-Zmluvy (contract register) and Brazil’s Portal da Transparência.
• Quantified target: 95% of central-budget spending is available in real time on the dashboard by 2027; at least 50 civil-society organisations and research institutes actively use the API; the number of citizen queries exceeds 100,000 per month in the first year.
• International precedent: Brazil’s Portal da Transparência (since 2004): ~15% of corruption cases exposed have been triggered by civic data analysis. Slovakia’s e-Zmluvy: the online publication of every state contract — following its 2011 introduction, public-procurement prices fell by an average of 5–8% in the transparent categories.
• Trade-off / risk: Publishing raw data can produce a “data tsunami” effect: without proper contextualisation and visualisation, the average citizen cannot make use of the data, while professionals (e.g. lobbying interests) exploit it more effectively. The paradox of “extreme transparency”: if every internal consultation and draft is public, decision-makers may retreat to informal channels (e.g. Signal groups).

A2 — Procurement transparency

• Mechanism: An AI-based anomaly detector on EKR data: it flags single-bid procedures, atypical pricing (against price/m² benchmarks), recurring winner–contracting-authority pairs, artificially narrowed specifications, and calls for tenders modified shortly before the deadline. The system does not “judge” — it flags; investigation is a human task. Red-flag system at 3 levels (low / medium / high risk).
• Quantified target: The share of single-bid procedures falls from 30% to below 15% within 3 years; at least 20% of cases flagged by the AI system trigger a substantive investigation; public-procurement prices move on average 5–10% closer to market benchmarks.
• International precedent: Ukraine’s ProZorro system (2016): the share of single-bid procedures dropped from 40% to 18%, with annual savings of ~2 billion USD. Korea’s KONEPS: fully electronic public procurement delivered a 12% cost reduction.
• Trade-off / risk: The “cobra effect”: if the single-bid share becomes the KPI, contracting authorities can game the system by formally generating more bids (e.g. via shell companies). False positives from the AI anomaly detector can cause “alert fatigue” — reviewers eventually ignore the flags. Continuous retraining of the model and calibration of the flagging threshold are essential.

A3 — Public asset declarations

• Mechanism: A machine-readable (structured-data) asset-declaration system: politicians fill out their data on an online form, which becomes instantly comparable (time series, cross-politician comparison). Automatic cross-referencing: declared real estate is matched against the land registry, declared companies against the electronic company register, and discrepancies between the bank balance and annual income are flagged.
• Quantified target: 100% of asset declarations available in machine-readable format by 2026; at least 3 automatic cross-checks run on every declaration; an “unexplained enrichment” flagging threshold defined and applied.
• International precedent: Georgia (2017): after introducing the electronic asset-declaration system, the share of flagged discrepancies rose 300% — corruption did not increase, but detection improved. Romania’s ANI (National Integrity Agency): since 2007, proceedings have been launched against 1,000+ public figures based on asset-declaration discrepancies.
• Trade-off / risk: Concealing wealth through family members and shell-company networks can bypass the system. Genuine effectiveness requires a Ultimate Beneficial Owner (UBO) registry and international data exchange (e.g. uncovering offshore structures). The data-protection balance is also delicate: where is the line between public transparency and private life?

A4 — Lobbying register

• Mechanism: Mandatory registration for every interest-representation organisation and person engaged in lobbying activity. The register contains: who met whom, when, on what topic, and what document (position paper) they submitted. The lobbying-related entries in the diaries of public decision-makers (ministers, state secretaries, MPs) are public. The system operates on the “legislative footprint” principle: for every adopted law, it is traceable who lobbied for it.
• Quantified target: At least 500 registered lobbyists/organisations in the first year; within 3 years, a “legislative footprint” is available for 80% of draft legislation; sanctions for breaches: fines and suspension of lobbying activity.
• International precedent: Ireland’s Lobbying Act (2015): 2,200+ registered lobbyists, with GRECO rating the system as “best practice”. The EU Transparency Register: ~12,000 registered organisations, though it is voluntary and has enforcement shortcomings.
• Trade-off / risk: Crowding out into informal channels: if formal lobbying is burdened with administration, interest representation can shift to informal channels (private dinners, “chance” meetings). A “reasonable threshold” must be found: smaller civil-society organisations should not be disproportionately burdened by bureaucracy. The choice between voluntary and mandatory registration is pivotal.

A5 — Whistleblower system

• Mechanism: An end-to-end encrypted, anonymous reporting platform (technology: SecureDrop- or GlobaLeaks-based). The whistleblower receives a unique, anonymous identifier with which to track the status of their case and respond to follow-up questions without revealing their identity. Legal protection for whistleblowers: full transposition and extension of the EU Whistleblower Directive (2019/1937) — prohibition of labour-market retaliation, reversal of the burden of proof.
• Quantified target: At least 200 substantive reports in the first year; a substantive investigation is launched within 30 days for 50% of reports; 0% of whistleblowers suffer unlawful retaliation (measured through a follow-up survey).
• International precedent: USA False Claims Act (1863, expanded in 1986): whistleblowers are entitled to 15–30% of recovered public funds — this incentive results in the recovery of ~2 billion USD annually. The Netherlands’ Huis voor Klokkenluiders: a centralised whistleblower-protection institution since 2016.
• Trade-off / risk: The problem of vexatious complaints: without a filtering mechanism, the system can become overloaded and genuine reports get lost. Tension between anonymity and investigability: anonymous reports are harder to investigate because they cannot be followed up with the reporter. Introducing “qui tam” (financial incentives) is a cultural question: in Hungary, many still perceive reporting as carrying a negative connotation.

A6 — Strengthening checks and balances

• Mechanism: Creation and annual publication of an institutional-independence index: measuring the actual functioning of the Constitutional Court, the State Audit Office (ÁSZ), the prosecution service, the ombudsman and the media authority with objective indicators (e.g. the share of decisions against the government, the political composition of appointments, the degree of budgetary independence, the turnaround time of decisions). The data are also published in international comparison. The goal is not direct intervention but accountability through the power of publicity.
• Quantified target: An annual “Institutional Health Report” with 10+ indicators; the Constitutional Court’s decision turnaround time falls below 90 days; the State Audit Office (ÁSZ) conducts 50+ substantive audits per year (up from the current ~30).
• International precedent: Colombia — the Constitutional Court (Corte Constitucional) is one of the strongest in Latin America: since its introduction in 1991, ~800 substantive decisions per year, ruling against the government in ~40% of cases. Sweden: the ombudsman institution (since 1809) is the oldest in the world — continuous publicity and a strong societal norm secure its independence.
• Trade-off / risk: Measuring “institutional independence” can itself become politically charged: if the index’s methodology is contested, the measurement’s credibility suffers. Risk of “indicator hacking”: institutions may adjust their operations to improve the measured figures without actually strengthening their control function (e.g. many but superficial ÁSZ audits). Measuring the political composition of appointments is a sensitive topic but indispensable.

A7 — Media pluralism as an institutional guarantee

• Mechanism: Annual calculation of a media-market concentration index (HHI — Herfindahl–Hirschman Index) on the domestic media market: television, online news portals, radio and print media, segment by segment. If the HHI exceeds a threshold (e.g. 2,500), the media authority automatically launches an investigation. Exposure of indirect ownership (shell networks) linked to the Ultimate Beneficial Owner (UBO) registry. Public-media reform: the public-media board’s governance and financing are secured by a statutory formula (GDP-linked, automatic), independent of political bargaining.
• Quantified target: The television market’s HHI drops below 2,500 within 3 years; at least 5 independent, sustainable newsrooms operate with national coverage; the trust index of the public media (Eurobarometer) rises from 30% to above 50%.
• International precedent: Germany’s KEK (Kommission zur Ermittlung der Konzentration im Medienbereich): regular review of media-market concentration, which in 2006 blocked the Springer–ProSieben merger. Denmark: the public media (DR) was financed directly by taxpayers until 2019 (a media fee), which provided stability.
• Trade-off / risk: There is tension between regulating the media market and freedom of speech: if concentration regulation is too aggressive, it deters business investors and weakens media players financially — which can exacerbate the “news desert” effect (the emergence of news-free regions). Political resistance to public-media reform is exceptionally strong, because the current status quo has beneficiaries. Traditional HHI does not capture the dominance of online platforms (Meta, Google) — separate measurement of digital advertising-market concentration is required.

A8 — Cohesion-policy accountability

• Mechanism: (1) Full monitoring of EU cohesion and RRF fund usage: every project with a public data sheet (beneficiary, amount, purpose, deadline, delivery). (2) Mandatory cost–benefit analysis (CBA) for every infrastructure project above 500M HUF — the analysis is carried out by an independent institute, and the result is public. (3) Clawback mechanism: if the project fails to meet its target indicators, a proportional share of the amount is repaid. (4) AI-based risk assessment: anomaly screening linked to public-procurement data and the beneficiary register. The OECD’s EU 2025 Survey emphasises: the effectiveness of cohesion policy is limited by corruption and administrative capacity shortages — the solution is not less funding but better control.
• Quantified target: The irregularity rate in EU funds usage drops from 2–3% to below 0.5%; 100% of cohesion projects have a public data sheet by 2027; the delivery rate of RRF projects is 90%+ (currently most member states have delays).
• International precedent: Estonia: the irregularity rate in EU fund usage is the lowest in the EU (<0.5%), thanks to a digital registry system and strong internal audit. Poland: during the 2014–2020 cohesion cycle, introducing a “red flag” system for public procurement reduced the share of suspicious procedures by 15%.
• Trade-off / risk: Overly strict accountability can come at the expense of absorption capacity: if project owners fear sanctions, they refrain from applying and funds go unused. The administrative burden (reporting, audits) hits smaller beneficiaries disproportionately. Needed: simplification of administration for small projects (e.g. simplified cost accounting below 100M HUF), while controls remain strong for large projects. 📖 Source: OECD: EU Economic Survey 2025

A9 — Spin-dictatorship prevention index

• Mechanism: Building on Guriev and Treisman’s Spin Dictators: the most important markers of pseudo-democracies are the monopolisation of information, the co-optation of the media, the bringing of courts under control and the institutionalisation of pseudo-civic organisations. An annual “Democratic Resilience Index”: measures the manipulation resistance of the Hungarian institutional system with 15+ indicators (media pluralism, judicial independence, civic space for NGOs, funding opportunities for opposition parties).
• Quantified target: The index publicly published by 2028 with 15+ indicators; the share of “at-risk” indicators decreases year on year; within 5 years, a level above the EU average is reached. 📖 Source: Guriev–Treisman: Spin Dictators (Preface and Chapter 1)

A10 — Independent Corruption Investigation Agency (CPIB model)

• Mechanism: An independent agency modelled on Singapore’s Corrupt Practices Investigation Bureau. Lee Kuan Yew: “It is easy to start off with high moral standards, but difficult to live up to them unless the leaders are strong enough to deal with all transgressors, without exceptions.” The agency reports directly to the Prime Minister but is operationally independent; it may investigate any public official or politician; in cases of unexplained enrichment, the burden of proof is reversed.
• Quantified target: The agency established by 2028 with a staff of 100+; 500+ investigations per year; the clearance rate of corruption criminal cases rises from 30% to 60%. 📖 Source: Lee Kuan Yew: From Third World to First (Chapter 12)

A11 — Civil-society partnership programme

• Mechanism: According to Tocqueville, civil society is a “school for democracy” and a safety net against the tyranny of the majority. Introduction of mandatory civil consultation for every high-priority piece of legislation: civil-society organisations review the draft, and the list of accepted and rejected proposals is public. Annual Civil Partnership Report.
• Quantified target: 30+ laws per year with civil consultation; 200+ civil-society organisations actively participating per year; annual publication of the Civil Partnership Report. 📖 Source: Tocqueville: Democracy in America (Book II)

A12 — Campaign-finance transparency and cap

• Mechanism: Lee Kuan Yew’s experience: corruption in campaign finance is the most common entry point into systemic corruption. A statutory ceiling on campaign spending; every donation above 50,000 HUF public and available in real time; a ban on corporate donations to parties; state campaign support in the interest of a level playing field.
• Quantified target: A campaign-spending ceiling of max. 1bn HUF/party for parliamentary elections; 100% of donations publicly available within 48 hours; the share of corporate party donations drops to zero. 📖 Source: Lee Kuan Yew: From Third World to First (Chapter 12)

A13 — Disinformation resilience

• Mechanism: The key thesis of Spin Dictators: modern autocrats maintain power by manipulating the information space. Programme elements: (1) a national digital media-literacy programme in public education, (2) financing of an independent fact-checking network from public funds but with editorial independence, (3) a quarterly disinformation-campaign monitoring report.
• Quantified target: The media-literacy programme reaches 200,000 students per year by 2030; 3+ independent fact-checking newsrooms operate; the disinformation-recognition capability (measured by a media-literacy test) improves by 20 percentage points. 📖 Source: Guriev–Treisman: Spin Dictators (Preface)

A14 — International institutional participation and accountability

• Mechanism: Stiglitz documented: the international financial institutions (IMF, World Bank) “escaped the norms of democratic accountability that we expect from the public institutions of modern democracies” — the selection of leaders happens behind closed doors, and substantive participation and publicity are lacking. Hungarian programme: (1) an active Hungarian position on EU forums in favour of transparency reforms of international institutions, (2) publication of the Hungarian voting position and its justification in IMF and World Bank decision-making (analogous to the principle in KP3), (3) annual publication of an “International Institutional Transparency Assessment” — to what extent do the IMF, the World Bank, the EBRD etc. meet transparency standards?
• Quantified target: Hungarian voting justification 100% public for IMF/World Bank decisions; annual Transparency Assessment from 2028; active participation in at least 3 international institutional-reform initiatives.
• Trade-off / risk: Unilateral transparency reduces Hungary’s negotiating flexibility. EU-level coordination is more effective but slower. 📖 Source: Stiglitz: Globalization and Its Discontents