The state of public procurement remedies in Hungary: 2022-2026

2,511 cases, HUF 4.19 billion in fines: the KDB's five-year record

The Public Procurement Arbitration Board (Közbeszerzési Döntőbizottság, hereafter KDB) is the legal filter of Hungarian public procurement. Between January 2022 and August 2026 it closed 2,511 cases, and its decisions are linked to HUF 4.19 billion in fines and HUF 8.18 trillion in affected project value (based on the 2,155 cases with a known project value). This is the first case-by-case, publicly available analysis of the full period.

A KDB case is initiated either by a complainant or by an authority acting ex officio: the President of the Public Procurement Authority (Közbeszerzési Hatóság, KH, hereafter the KH President), the Integrity Authority, the State Audit Office, the National Development Centre or a ministry Managing Authority. The KDB closes the case with a decision on the merits, with an outcome of well-founded, partially well-founded or unfounded, and imposes a fine where warranted.

The consequence is concrete: a fine, and where the KDB declares lawful an exclusion for false information or undue influence, a ground for exclusion that applies for three years under Section 62(1)(h) and (j) of the Public Procurement Act (Act CXLIII of 2015, hereafter Kbt.).

What we analysed

The dataset brings together the KDB decisions and orders published on the Public Procurement Authority's portal, structured case by case. The download window covers cases closed between 1 January 2022 and 31 August 2026. The dataset contains 2,511 cases suitable for analysis, after we removed 599: 590 joined parallel case numbers, 5 cases with conflicting sources and 4 entries without a source.

The 2,511 cases break down by year as follows:

• 2022: 455 cases
• 2023: 502 cases
• 2024: 579 cases
• 2025: 601 cases
• 2026 (to 31 August): 374 cases

The number of cases is rising, but not evenly: it grew by 32 per cent between 2022 and 2025.

Merits or procedure: the win rate can only be measured on the 1,454 merits cases

A case closes along one of two lines: of the 2,511 cases, 1,454 ended in a decision on the merits (with an outcome of well-founded, partially well-founded or unfounded) and 828 in a procedural order (dismissed or terminated). Separating the two lines is the basis of every further breakdown: the win rate, the fine and judicial review are all meaningful only for cases decided on the merits.

For each case we recorded the initiator category, a sector tag, the project value (where the decision states it) and the details of the finding of infringement. Project values for fined cases were checked by manual sampling. The methodological details are summarised in the methodological note at the end of the article.

The ex officio share matches the KH summary report

The ex officio share is in line with the Public Procurement Authority's own summary report. In the 2022-2024 comparison window the analysis measures 39.2 per cent, and the summary report gives 38.6 per cent. The classification rests on the opening sentence of the decision, and the cited section of the Kbt. serves only as a secondary check.

We checked the analysis in four layers: external comparison with the KH summary reports, an independent re-reading of 99 cases, a review of the processing logic and a coverage check. The layer-by-layer reports and the full methodological documentation are available on the KDB 2022-2026 dataset page.

The KH President wins more often: 98.46 vs 62.5 per cent

By initiator, KDB cases divide into ex officio proceedings and proceedings brought on complaint. Of the 2,511 cases, 1,469 (58.5 per cent) were brought on complaint, ex officio proceedings account for 949 cases, and in 93 cases the initiator could not be classified unambiguously. Measured on the cases with an identified initiator, the ex officio share is 39.2 per cent.

The 36-point gap is not merely a reflection of legal strength. The KH President selects cases from the Authority's own contract-review experience and brings before the KDB only those where an infringement is highly likely. The complainant channel, by contrast, covers a broader portfolio of cases, in which the share of actual infringements is lower to begin with.

Half of the 949 ex officio cases trace to the KH President

The ex officio side is not a single channel. Of the 949 cases, 468 trace to the KH President (275 on other grounds, 193 based on contract review), and 481 come from other ex officio initiators: the Integrity Authority, the State Audit Office, the National Development Centre and ministry Managing Authorities.

Besides the KH President, ex officio initiations are spread across more than ten other bodies, a distribution that is substantive but not monopolistic.

The initiator is the strongest predictor of the outcome

The initiator channel is the strongest predictor of the outcome that the dataset offers. The win rate on the merits is 62.5 per cent in cases brought on complaint, and 98.46 per cent in cases initiated by the KH President. The difference is not a few percentage points, it is 36 percentage points. When the regulator initiates, almost every case ends in a finding of infringement. For private complainants, just over half do.

The complainant's institution type also splits the win rate sharply

On the complaint side, the win rate depends on the type of institution. Private complainants (tenderers and candidates) win 59 per cent of decisions on the merits, public-institution complainants (litigating as state, municipal or public-service bodies) 98.5 per cent. On the same complainant track, a win rate of close to 60 or close to 100 per cent goes with the complainant's institution type. The full reading of the institutional complainant phenomenon is in a separate article.

The ex officio share is declining year by year

The annual picture moves. The ex officio share was 50.9 per cent in 2022 and fell to 31.0 per cent by August 2026. The change is not gradual but specific to each year: the KH President's review focus, the system of ex officio notifications and the National Development Centre's initiations each move at their own pace. The full reading of the polarisation between the regulator and private complainants is in a separate article.

Construction leads in case volume, transport in fines

The KDB's sector map draws three different rankings, and no single sector tops all three. Construction leads in case volume: with 586 cases, it accounts for 23 per cent of all KDB proceedings. In the share of fined cases, culture and media leads, with a fine in 58.6 per cent of its cases. In the average fine per case, transport comes first: the average fine in a transport case is HUF 12.4 million, more than five times the construction average.

The divergence of the three rankings is no coincidence. Case volume shows where most review disputes take place. The fine rate shows how likely a dispute is to end in a sanction. The average fine speaks to the severity of the infringements found. A sector can rank differently on each of the three measures, and in many cases it does.

The distribution is structurally unconcentrated. No single sector dominates the KDB docket. Behind the top runs a long tail: ten sectors and the unidentified category together account for 54 per cent of cases.

Three different fine patterns: construction, transport, healthcare

The divergence between the case-volume, fine-rate and fine-total rankings shows most sharply in the profiles of the three largest sectors. Each follows a different logic.

Construction: small fines, many cases

Construction leads the case-volume ranking: 586 KDB cases, 23.3 per cent of the total. The share of findings of infringement in this sector is also among the highest, with fined cases making up 54.8 per cent of the sector. The weight behind the numbers is low, however. The average fine per case is HUF 2.3 million, and the sector's aggregate fine total is HUF 753 million. Construction fills the KDB docket with volume, not with heavy individual penalties. The sector's win rate in decisions on the merits is 58.4 per cent.

Transport: large fines, few cases

By case volume, transport ranks only fourth, with 195 KDB cases, 7.8 per cent of the dataset. It leads the fine ranking, however: HUF 917 million, nearly 22 per cent of all KDB fines. The average fine per case is HUF 12.4 million. That is five times the construction average and three times the dataset average. In a transport KDB case a hit is rarer, but when it comes, it is large.

The transport figure, however, cannot be treated as a finding about the sector. The HUF 917 million fine total is concentrated on three state and Budapest transport operators: the cases of the Budapesti Közlekedési Központ (BKK, Budapest's transport authority), Volánbusz and MÁV Volán. Read as the domestic transport market, the figure is misleading. The transport data in reality reflects the KDB exposure of three dominant market players. BKK's full KDB fine profile is set out in the section on the distribution of fines.

Healthcare: rare but large fines

Healthcare is the sector with the second-largest case volume (350 cases, 13.9 per cent), but its fine rate is among the lowest: a fine is imposed in 31.4 per cent of cases. When one comes, the average fine is HUF 6.7 million, the second largest after transport. The healthcare sector's win rate in decisions on the merits is 32.6 per cent, which means that in two thirds of cases the contracting healthcare institution successfully defends its procedure.

The three profiles give three different patterns. Construction is a high-traffic sector of medium severity. Transport is rare and large, driven by three market players. Healthcare is high-traffic, with a lower fine rate and large individual fines. Each of the three profiles calls for its own reading.

The ten smaller sectors that make up nearly half of the cases

Beyond the three leading sectors, ten further sector categories appear:

• IT and software: 221 cases, HUF 262 million in fines (the third-largest case volume)
• Environment and waste: 187 cases, HUF 162 million in fines
• Energy and utilities: 172 cases, HUF 209 million in fines, with a 59.4 per cent ex officio share
• Food and agriculture: 118 cases, HUF 366 million in fines
• Culture and media, defence and security, public administration and education together: 401 cases, HUF 614 million in fines

The unidentified category holds a further 182 cases, 7.2 per cent of the total. For these, the "subject of procurement" field extracted from the decisions did not allow an unambiguous sector classification.

Together the long tail makes up almost half of the cases and two fifths of the fine total. The KDB docket covers a broad range of sectors, not a system closed around one or two dominant ones.

Fine frequency by sector

Between a case and a fine lies a decision. Not every complaint leads to a finding of infringement, and not every finding of infringement is followed by a fine. The fine rate shows what share of a sector's cases ends in a fine. The highest rates:

• Culture and media: 58.6 per cent
• Education: 55.4 per cent
• Construction: 54.8 per cent
• Food and agriculture: 52.5 per cent
• Defence and security: 46.8 per cent

The lowest:

• IT and software: 24.0 per cent
• Healthcare: 31.4 per cent
• Energy and utilities: 37.8 per cent
• Transport: 37.9 per cent
• Environment and waste: 46.0 per cent

The fine rate does not speak to the severity of the infringement. It shows how likely the KDB is to find a material failing when it examines the proceedings brought in a given sector. Two factors drive it: the quality of the complaint base (how well founded the complaints filed in the sector are) and how well prepared the sector's procurement procedures are. The 24 per cent rate for IT and software is the lowest among the twelve sector categories (not counting unidentified cases). One reading is that the sector's procurement procedures are carefully prepared. The other is that a large share of the complaints filed in the IT market do not survive examination on the merits.

Utilities, education and construction are in the regulator's focus

The initiator of a KDB procedure need not be a market participant. The President of the Public Procurement Authority may launch an ex officio review under Section 152(1) of the Kbt., and the share of such cases varies strongly by sector. The sectors with the highest ex officio share:

• Culture and media: 60.9 per cent
• Energy and utilities: 59.4 per cent
• Education: 54.8 per cent
• Construction: 52.8 per cent
• Food and agriculture: 46.1 per cent

The lowest:

• IT and software: 17.7 per cent
• Defence and security: 18.7 per cent
• Healthcare: 20.7 per cent
• Transport: 28.3 per cent
• Public administration: 41.2 per cent

The ex officio figure reflects which sectors' procurement draws the most regulatory attention. It is no surprise that utilities, education and construction lead: these are the classic terrain of public assets, and the Public Procurement Authority regularly reviews the contracts concluded there. In ex officio proceedings initiated by the KH President, the KDB finds an infringement in 98.46 per cent of decisions on the merits, compared with 62.5 per cent on the complainant side. Regulatory attention is therefore denser in the sectors with a high ex officio share.

The fine-to-value ratio of the sectors is set out in the section on project size.

1,032 fined cases, HUF 4.19 billion, a median of HUF 1 million

The fine map is built from many small cases and a few very large ones at the same time. In 1,032 of the 2,511 cases the KDB imposed a fine above zero, with a combined value of HUF 4,188,945,000. The median fine is HUF 1 million and the largest individual fine HUF 200 million, imposed in two separate cases of the Budapesti Közlekedési Központ.

BKK's 12 cases produced HUF 736.5 million in fines

The largest fine payer in the full KDB dataset for 2022 to mid-2026 is a single company owned by the City of Budapest. In 12 cases in which the Budapesti Közlekedési Központ (BKK) was the contracting authority, the KDB imposed HUF 736.5 million in fines in total, HUF 715.5 million of it on BKK itself. In 2 of its cases the fine was HUF 200 million each: the D.121/2025 bus procurement and the D.362/2024 Volánbusz case. Second is the Péterfy Sándor utcai Kórház (a Budapest hospital) with HUF 160 million, from just 2 cases. Cases D.231/2024 and D.94/2025 are each worth HUF 80 million, made up of the two parallel fine bases under Section 165(6)(a) and Section 165(7a) of the Kbt. Third is ALFÖLDI NYOMDA Zrt., with 3 cases and HUF 95.4 million.

The three entities together carry HUF 991.9 million in fines, 23.7 per cent of all KDB fines. The typical KDB fine is around HUF 1 million, but a few outlier cases pull the mean up to four times that. A top 3.8 per cent of cases produces half of all fines imposed. In a third of the fined decisions on the merits, two or more parties are fined, and nearly half of the amounts imposed fall on these cases. Fines above HUF 20 million are 11.4 times more likely to reach the Budapest-Capital Regional Court than zero fines. Together, the three patterns draw the structure of the fine map.

The median is only HUF 1 million, the mean much higher

The shape of the distribution points from the median towards the mean. The median of the fine distribution is HUF 1 million. The mean is HUF 4.06 million, pulled up by a few individual fines that are orders of magnitude larger. The 90th percentile is HUF 7 million, the 99th percentile HUF 70 million. The largest individual fine is HUF 200 million, imposed on two separate occasions, both against BKK. The practical consequence of the fourfold gap is simple: a sector's fine profile has to be read with the median and the total amount together, because either one alone misleads.

Annual trajectory, 2022-2026

The annual fine total rose over the period examined. It was HUF 526 million in 2022, 637 million in 2023, 1,175 million in 2024 and 1,340 million in 2025. The partial year 2026 shows HUF 511 million up to 31 August 2026. Between 2022 and 2025 the annual fine total grew 2.5-fold, not monotonically.

Half of all fines fall on 39 cases

Breaking the total into bands shows where the HUF 4.19 billion falls. Of the 1,032 fined cases, 403 (39.1 per cent of cases) received a fine below HUF 1 million, HUF 180 million in total. That is 4.3 per cent of all fines. The second band (fines of HUF 1 to 5 million) covers 473 cases and HUF 924 million (22.1 per cent). The HUF 5 to 20 million band holds 117 cases and HUF 1,023 million (24.4 per cent). The band above HUF 20 million holds just 39 cases. These 39 cases, 3.8 per cent of fined cases, produce 49.2 per cent of all fines: HUF 2,062 million.

Concentration shows at both the band and the case level. The ten largest individual fines together come to HUF 1,034 million, 24.7 per cent of all fines. The top ten per cent (103 cases) carries 67.0 per cent of all fines. The Gini coefficient of the fine distribution is 0.743: a measure of inequality on which a value of 1 means complete concentration.

A handful of large cases and a thousand small ones, taken together. The thousand small cases are the KDB's day-to-day fine traffic. The handful of large cases is the backbone of fine policy, and read on their own they make the fine map look almost like a different institution.

The weight of the three institutions

Of the 39 cases above HUF 20 million, 11 are carried by the three largest fine payers (BKK, Péterfy, ALFÖLDI NYOMDA). Read on its own, this handful makes the concentration picture much simpler: a few institutions repeatedly end up in high-value, difficult procurement cases, and the large fines fall on them. The concentration of the top band comes partly from the fine distribution itself and partly from a few specific institutions that end up there again and again.

Fines on several parties: in a third of cases, nearly half of the fine total

More than one party was fined in 34.4 per cent of cases

A fine does not always fall on one party. In a significant share of cases the KDB holds several parties liable and imposes a separate amount on each. The decisions on the merits with a finding of infringement (well-founded and partially well-founded, with a positive fine) number 981 cases. In 337 of them, 34.4 per cent, two or more parties were fined. These 337 cases produce 49.8 per cent of the fines imposed in the 981 cases and 45.4 per cent of all KDB fines. The multi-party pattern characterises a third of the cases and makes up nearly half of the fine total.

The twin fine structure of Section 165 of the Kbt.

The typical legal basis for multi-party fines is Section 165(6)(a) and Section 165(7a) of the Kbt. Both cases of the Péterfy Sándor utcai Kórház produce their HUF 80 million each from these two bases. In the 6 cases of the Szépművészeti Múzeum (Museum of Fine Arts), every case likewise consists of two per-party entries. The fine on one institution therefore falls not once but twice within a single case. In the sample of 125 contracting-authority complainant cases, multiple fines took a different form: in two cases (D.295/2022, D.293/2024) several respondent tenderers were each fined HUF 1 million.

Fines grow with project size while their share falls: a project 10 times larger, a fine 3.3 times larger

The size of a fine is not just a matter of luck. Four measurable circumstances go with it, and together they explain 55 per cent of the differences in fine size (R² = 0.55 on a log scale, over 923 cases with project-value data). The model makes no causal claim. It is observational and describes the co-occurrences visible in the 2022-2026 dataset.

From the most important to the least:

• Project value: a procurement value 10 times larger goes with a fine 3.3 times larger on average (coefficient 0.519 on the log scale).
• Multi-party fines: fines are 2.4 times larger in multi-party cases, controlling for project value and year (coefficient 0.379).
• Year: in 2024 the fine level is 1.4 times higher than in 2022, controlling for the other factors (coefficient 0.147).
• Initiator channel: in other ex officio cases the fine is 1.29 times larger than in complainant cases (coefficient 0.110). In cases initiated by the KH President, the fine level does not differ significantly from complainant cases.

What do the four numbers say together? Fine size can partly be estimated in advance. In a contract-modification case on a two-billion-forint construction project involving several parties, the model points to a fine range of a few tens of millions of forints. The exact figure is uncertain, the order of magnitude is not.

The fine-to-value ratio as an exposure indicator

Fine size alone misleads unless it is read next to the size of the underlying procurement. A HUF 200 million fine on a HUF 20 billion contract carries a very different weight than on a HUF 100 million contract. The database holds both numbers, and their ratio is an exposure indicator that can always be computed.

The median is 1 per cent, the upper quartile 2 per cent

In the subsample of 957 fined cases with a finding of infringement and a known project value, the median fine-to-value ratio is 1.00 per cent. That is the typical exposure: the fine reaches one hundredth of the project's value. The upper quartile is 2.08 per cent, and the highest individual ratio is 27.87 per cent, in a case where the dataset recorded an original contract value much smaller than the modified contract.

Larger projects carry a lower fine ratio

Larger projects carry a lower fine ratio on average, but project size does not give a precise estimate for any single case. Across the four project-value quartiles, the median fine ratio is 2.2, 1.1, 1.0 and 0.5 per cent. Even in the highest-value quartile the ratio exceeds 2 per cent in 9.6 per cent of cases, and in the lowest-value quartile it is at most 2 per cent in 42.5 per cent of cases.

Public administration and culture: a 1.9 per cent peak

The fine-to-value ratio also differs between sectors. The median for public administration and for culture and media is approximately 1.9 per cent of project value, for construction and for defence and security approximately 1.0 per cent. The difference largely goes with typical project size: on a more expensive contract the fine is larger in absolute terms and smaller as a share.

17 per cent of cases reach court, and the KDB decides in a median of 55 days

The share of cases reaching court rises every year: from 13.8 to 21.1 per cent

More and more parties turn to the courts after a KDB decision. Of the 2,511 cases, 428 were appealed to the Budapest-Capital Regional Court, 17.0 per cent of the full dataset. The annual share is on a rising path: 13.8 per cent in 2022, 17.8 in 2024, 18.3 in 2025 and 21.1 per cent in the first eight months of 2026.

Fines above HUF 20 million reach court 11.4 times more often

The larger the fine, the more often it reaches court. The pattern shows across the 981 merits cases with a positive fine and the zero-fine reference cases together, measured by applications for judicial review before the Budapest-Capital Regional Court (N = 2,120, logistic model).

The band coefficients:

• Below HUF 1 million: no significant difference from zero fines (p > 0.1).
• HUF 1 to 5 million: 2.4 times more likely to reach judicial review.
• HUF 5 to 20 million: 3.8 times more likely.
• Above HUF 20 million: 11.4 times more likely. The magnitude is too large to put down to chance (p < 10⁻⁶).

The HUF 20 million line is not a statutory threshold but a practical one. Above it, the fine amount almost always justifies the cost of legal representation and review proceedings. Most of the BKK, Péterfy and ALFÖLDI NYOMDA cases sit in this band, and the review rate is correspondingly high. The repeat proceeding in the BKK ticket-machine case, decision D.283/2022 with a HUF 70 million fine, also sits in this band.

The 11.4-fold appeal rate does not tell the whole story on its own

Fines above HUF 20 million are 11.4 times more likely to reach judicial review. Most cases in the top band involve parties (BKK, Péterfy, ALFÖLDI NYOMDA and organisations of similar size) with their own legal departments and routine review experience. The model does not take into account how large a party is or what legal resources it brings to the hearing. Part of the 11.4-fold co-occurrence relates to the size of the fine, and another part to the fact that parties receiving large fines typically bring a legal department and an appeal routine with them.

A median of 55 days, a decision within two months

The time to a decision is short. The median time from filing to decision is 55 days. Cases closed by a procedural order (dismissed or terminated) are fast, with a median of 17 days. Decisions on the merits are slower, with a median of 73 days. The KDB thus filters out inadmissible complaints quickly and spends two and a half months on average on the rest.

Appeals come from two sides

Appeals come from two directions. Applications for review against findings of infringement on the merits come from the fined contracting authority or tenderer, while if the outcome is no infringement, the complainant itself may go to court. The rising annual share contains both directions at once, and the exact breakdown can be reconstructed case by case from the data package.

Taken together, the appeal rate and the timing suggest that parties are increasingly ready to go to court, while the KDB's pace of decision-making is not slowing down.

Methodological note

The analysis is based on 2,511 KDB cases from the period between 1 January 2022 and 31 August 2026. Of the 3,110 downloaded cases, 590 consolidated parallel case numbers, 5 cases with conflicting sources and 4 cases without a source dropped out. Of the 2,511 cases, 1,032 were fined, with a total of HUF 4.19 billion.

The individual fine metrics rest on different bases. The multi-party shares are based on the 981 cases with a positive fine and a finding of infringement on the merits, and the fine-to-value ratio on the 957 fined cases with a known project value. The fine-size model runs on 923 cases with project-value data (R² = 0.55 on a log scale), and the judicial-review model on 2,120 cases. Both models are observational, not causal.

The dataset's limits and full method are set out on the KDB 2022-2026 dataset page.

References

The data package for this article is available on the KDB 2022-2026 dataset page: the CSV download, the full methodological documentation, the processing log and the reports of the four-layer audit in one place. The decisions come from the Public Procurement Authority's portal, and the half-yearly KH summary reports are published on the KH website. The other articles in the series unpack single aspects of the dataset: the reading of the KH President's win rate and of the institutional complainant.