Courts portal data: complaints filed in 2024 at the 179 local courts, as of 13 November 2026
Complaints against fines: how many succeed and how long they take
We counted the complaint cases against fines (contravention reports) filed in 2024 that appear in the portal: 114,908. Of those decided at first instance, 24% led to the fine being cancelled or changed, for example a warning instead of a fine; 60% were rejected, and 12% were lost on procedure, for example a complaint filed late or without the court stamp fee. These are only the fines that went to court, not all fines.
What happened to the complaints
The result differs a lot by who issued the fine. “Decided” means the case ended at first instance: the complaint was granted or rejected, or the case ended another way. We chose 2024 because it is the last full year in which every complaint had at least 22 months to receive a ruling; complaints from the first half of 2025 show the same pattern.
24%of the 105,314 decided complaints led to the fine being cancelled or changed (95% confidence interval: 23.6–24.2%)
- fine cancelled
- fine changed (for example replaced by a warning, or reduced)
- rejected, the fine stands
- lost on procedure (late, inadmissible, court fee not paid, lapsed)
- ended another way (withdrawal, transfer, joinder)
Of every 100 decided complaints: 17 with the fine cancelled, 7 with the fine changed, 60 rejected, 12 lost on procedure, 4 ended another way.
Counting only the complaints decided on the merits (cancelled, changed or rejected), 29% led to cancellation or change.
Losses on procedure can be avoided: they mean a complaint filed late, inadmissible, or without the court fee paid. Check the deadline and the fee at the court or in the papers you received.
Differences between authorities also reflect which fines get contested. They do not show how well an authority works, and they do not mean the fine was wrong in fact: they are court rulings.
Another 9,594 complaints of this kind have no ruling yet and are not in these percentages.
Not sure which court covers you? Search your town: the assignment is the one in HG 1.217/2023, as on the page judecatoria-ta.
The figures cover all kinds of fines together and are not a ranking: differences between courts can reflect which fines reach each court, and the confidence interval shows how certain the number is. Only courts with at least 100 decided complaints and at least 80% of their 2024 complaints decided appear.
Map of the courts: fines issued by the police
- a higher share of complaints cancelled or changed than the national average (18.6%), beyond the margin of error
- a lower share than the national average, beyond the margin of error
- difference within the margin of error of the national average
- no police figures (too few decided police-fine complaints or too many without a ruling)
Bucharest:
The map covers only fines issued by the police, because only for those are there enough complaints at most courts. A different colour does not show that a court is more lenient or stricter: it also reflects which kinds of offences reach each court (for example speed cameras, speeding, documents), which we do not have. A colour appears only if the 95% confidence interval does not include the national average. With 134 comparisons at the 95% level, about 7 courts would be coloured by chance even if no court differed; that is why we also checked the colours against 2025. The map is a visual aid: the search above does the same job. Outlines: geo-spatial.org / ANCPI (CC BY-SA 4.0).
Outcome by who issued the fine
Of decided complaints, how many led to the fine being cancelled or changed, by issuing authority. The line shows the 95% confidence interval.
- Police (traffic, county inspectorates, Bucharest)19%77,783 decided complaints
- Local police and town hall34%10,407 decided complaints
- CNAIR (national roads)63%5,022 decided complaints
- Tax and customs (ANAF)42%1,494 decided complaints
- Labour inspectorate (ITM)33%1,282 decided complaints
- Consumer protection (ANPC)38%1,285 decided complaints
- Other public bodies36%4,744 decided complaints
Differences also reflect which fines get contested; they do not show how well an authority works. We do not know on what grounds the fines were cancelled or changed (for example evidence, formal defects or a different classification).
How long until a first ruling
We count from the day the complaint is filed to the court's first ruling. That can also be a ruling ending the case by withdrawal, annulment or transfer. The date is the date the ruling was pronounced, not the date the reasoned judgment was received. Complaints with no ruling yet are counted as pending, so the time is not understated. Every 2024 complaint could be followed for at least 22 months.
Of every 100 complaints of this kind, how many had a first-instance ruling within:
Half of the complaints had a ruling within 265 days.
A ruling at the local court can be appealed, and this data does not show what happens after that.
The full table
| Who issued the fine | Filed | Decided | Cancelled | Changed | Rejected | Lost on procedure | Ended another way | Cancelled or changed, of those ruled on the substance | Ruling within 90 days | Ruling within a year |
|---|---|---|---|---|---|---|---|---|---|---|
| All fines | 114,908 | 105,314 | 16.7% | 7.2% | 59.9% | 12.3% | 3.8% | 28.5% | 9.0% | 65.9% |
| Police (traffic, county inspectorates, Bucharest) | 84,946 | 77,783 | 13.5% | 5.2% | 66.5% | 11.2% | 3.7% | 21.9% | 8.3% | 65.0% |
| Local police and town hall | 11,180 | 10,407 | 22.5% | 11.6% | 48.5% | 15.3% | 2.1% | 41.3% | 9.1% | 70.3% |
| CNAIR (national roads) | 5,201 | 5,022 | 53.7% | 9.3% | 18.8% | 14.9% | 3.3% | 77.0% | 10.6% | 77.6% |
| Tax and customs (ANAF) | 1,719 | 1,494 | 16.4% | 25.8% | 42.2% | 12.0% | 3.5% | 50.0% | 5.8% | 60.3% |
| Labour inspectorate (ITM) | 1,555 | 1,282 | 14.2% | 18.5% | 46.3% | 15.7% | 5.4% | 41.4% | 7.0% | 51.0% |
| Consumer protection (ANPC) | 1,375 | 1,285 | 18.9% | 19.2% | 50.7% | 9.7% | 1.5% | 42.9% | 6.0% | 66.3% |
| Other public bodies | 5,392 | 4,744 | 18.0% | 18.1% | 46.0% | 13.3% | 4.6% | 43.9% | 7.2% | 59.2% |
| By who filed the complaint | ||||||||||
| Company or sole trader | 14,770 | 13,189 | 21.4% | 13.4% | 49.4% | 12.6% | 3.2% | 41.3% | 7.1% | 62.7% |
| Other petitioners (mostly individuals) | 100,138 | 92,125 | 16.0% | 6.3% | 61.4% | 12.3% | 3.9% | 26.7% | 9.3% | 66.4% |
What the data does not show
- These are only the fines that went to court. Many people pay their fine without contesting it, so the percentages are not your chance of winning.
- “Cancelled” or “changed” is the court's ruling, not proof that the issuer was wrong. “Changed” often means a warning instead of a fine.
- The data hold no fine amount, offence or evidence: similar complaints can end differently.
- The issuing authority was read from the respondent's name in the case. For 3,540 complaints the respondent is missing or not an institution, so they appear only in the total. The environment guard, public transport operators and emergency services are under “Other public bodies”.
- The 9,594 complaints without a ruling are counted as pending, not as rejected. Some of them are suspended cases.
- A ruling at the local court can be appealed, and the data do not show what follows.
- A case transferred to another court can appear twice, under different numbers. We count it as ended another way.
- Figures by court are only for the 150 courts with at least 100 decided complaints and at least 80% of their 2024 complaints decided; another 14 courts have too many complaints without a ruling to be read fairly. They show how often fines were cancelled or changed there, not how any particular judge rules, and they are not a ranking.
- We took only cases filed from 2024, so we cannot say whether complaints are more or fewer than before.
- These data have not been compared with official court statistics.
How we counted
Source
The public service of the Romanian courts portal (portalquery.just.ro): all 179 local courts, of which 175 had contravention complaints, filed between 1 January 2024 and 30 June 2025. The figures on the page cover complaints filed in 2024, so each could be followed for at least 22 months. Complaints from the first half of 2025 show the same pattern (23% cancelled or changed, against 24%).
How we read the rulings
A program reads what the court decided in each case: cancelled, changed, rejected, lost on procedure, withdrawal, or another ending. The program was checked by two AI models (Opus and Fable), each a fresh copy with no access to the program's answer. They received only the ruling text with names removed and a codebook. They are not lawyers and not human reviewers, and their agreement shows that they read alike, not that they are right.
The first check drew 375 cases, of which 360 were scored (15 were excluded because of a sampling mistake of mine, corrected before the second check). It found mistakes: some rejections were counted as “lost on procedure”, and endings by joinder or conversion of the fine into community service were not recognised; of the 49 old cases the program thought were pending, 4 were in fact finished. We corrected the rules, froze the program and checked again on 230 new cases it had never seen (227 scored). The fixed thresholds were at least 90% agreement on decided cases and at most 10% of old “pending” cases actually finished. On the 185 decided cases where the two models agreed, the program matched in 94% of cases (rare categories oversampled), and on a random sample, in 97.5%. Of the 40 old “pending” cases, none was in fact finished. Rare categories are less certain: “lost on procedure” 82% and “ended another way” 79% precision. On the group in the title, cancelled or changed, the program matched the models’ joint verdict in 97% of cases (34 predicted cases, a small sample). The two models agreed with each other in 99% of cases.
The issuing authority was read from the respondent's name (public bodies and companies, not persons). Two models checked 100 names chosen at random: the program and the models agreed in 99% of cases. The split by who filed the complaint (company or sole trader versus other petitioners) comes from a pattern on the petitioner's name (S.C., S.R.L., S.A., PFA, II) and was not checked separately; petitioners' names were not kept.
Rules fixed beforehand
Before counting results (after a pilot count of the volume), we fixed written rules, with the SHA-256 checksum of the project brief (67b0b7d35b4f): at least 60,000 complaints filed in 2024 (there were 114,908), figures by authority only for those with at least 1,000 decided complaints, figures by court only if at least 40 courts have 100 decided complaints and the split-half reliability between random halves of the same court (a rank coefficient) is at least 0.4, plus the two checks above. All the rules were met. After the page was reviewed we added stricter rules in a dated addendum (checksum 72b783b4f43d): a court appears only if at least 80% of its 2024 complaints have a ruling, and small authorities are grouped under “Other public bodies”. With these rules 150 courts remain, and the reliability was 0.75. For the map of fines issued by the police we fixed new rules in a second dated addendum (checksum 7dae7951dd92), before computing any court's rate (only the number of complaints per court and the overall reliability had been counted): only courts among the 150 shown, with at least 100 decided police-fine complaints; colour only where the 95% confidence interval excludes the national average (18.6%); at least 40 such courts and split-half reliability of at least 0.4; and, among coloured courts with at least 50 decided police-fine complaints in the first half of 2025, the same side of that half-year's average (17.1%) in at least 70% of cases. Results: 134 courts, reliability 0.74; of the 66 coloured courts, 60 had 2025 data and 51 stayed on the same side of the average (85%). Both checks passed.
Right of reply
The method and the table have not yet been sent to the Superior Council of Magistracy, the Ministry of Justice or the General Inspectorate of the Romanian Police, nor to CNAIR, ANAF, the labour inspectorate or ANPC, whose results appear on the page. If they reply, the reply appears here, with the date. No reply does not mean agreement.
Privacy
Party names were not kept, case numbers only as a code, and texts were cleaned of names, addresses and calendar dates. The page shows only totals by authority and by court, no person.
Data
Download the table above (CSV)
Source: the courts portal. Data as published on 13 November 2026. Commune outlines: geo-spatial.org / ANCPI (CC BY-SA 4.0); assignment of communes to courts: HG 1.217/2023.