Domestic violence · Romania's 179 local courts, 2024–2026

Romania's protection orders, court by court: how often local courts grant them and how fast they answer

Between January 2024 and December 2025, Romania's local courts (judecătorii) received more than 51,000 requests for a protection order, the court order that keeps an abuser away from the person they harm. When the victim applies on their own, 65% of requests decided on the merits end with an order. The average hides large gaps. At Arad local court, 80% of the 470 victim-filed requests decided on the merits ended with an order; at Brașov local court, 45% of 468. The gaps hold from one year to the next.

Find a court

A request goes to the local court for the victim's home or residence, or for the place where the violence happened. You can also pick a court on the chart or the map.

Pick a court to see how the requests filed there were decided.

Each dot is a local court

Share of requests filed by victims and decided on the merits in 2024–2025 that ended with an order, against how many there were. The grey band shows how far a court can drift from the national average by chance alone.

grants more often than chance explainswithin the chance bandgrants less oftenband (99.8%)

Map of the 179 local courts

Colour shows how often each court grants requests filed by victims, against the national average (65%); dot size, how many it decided on the merits. Tap a dot to pick the court.

grants less oftengrants more often· bigger dot: more requests

Poverty does not explain the gaps between courts

A reader asked what shows up if the map of protection orders is laid over the map of poverty. To answer, each local court was given the area it serves: the towns and communes listed against it in the government decision that sets court districts. Each area then gets its Local Human Development Index (IDUL 2018, computed by the sociologist Dumitru Sandu), which combines schooling, life expectancy, housing and links to cities. The lower it is, the poorer the place.

Ranked by development, Arad comes 11th and Brașov 8th of the 170 court areas. They are almost equally prosperous, yet they sit at opposite ends of the country on protection orders: Arad grants 80% of the requests filed by the people who were abused and decided on the merits, Brașov 45%.

The pattern holds across the country. Split Romania into three parts of roughly 6.3 million people each. In the poorest third, requests filed by victims are granted in 65% of the cases decided on the merits; in the middle third, 68%; in the most developed third, 62%. A decision within three days comes just as often: 76% of cases in the poorest third, 76% in the most developed.

Each dot is a court’s area

Poorer areas to the left, more developed to the right. Up the side, the share of requests filed by the people who were abused and decided on the merits that ended in an order, 2024–2025. If poorer areas got orders more or less often, the cloud would tilt up or down to the right. It stays flat.

grants more often than chance explainswithin the chance bandgrants less often

The one difference visible at first glance is how many people ask. Poorer areas filed somewhat more requests per 10,000 residents in 2025 (17.6, against 14.6 in the most developed). Compare courts within the same county, though, and the gap disappears (correlation −0.02), so it cannot be put down to poverty.

So what does predict how often a court grants an order? Its own record. On a scale from −1 to 1, a court’s 2024 rate matches its 2025 rate at 0.58. The poverty of the area it serves matches it at −0.06, close to zero. The gaps between courts follow the court, not the poverty of its area.

The thirds hold roughly equal numbers of people. That is why the most developed third has only 21 areas, mostly big cities and Bucharest, and the poorest has 84 areas. The comparison is between areas, not people: it says nothing about a poor woman in a rich area, or the reverse.

One country, two maps

How developed the area is

poorermore developed

How often orders are granted

grants less oftengrants more often

Each shape is one court’s area, as set by the government decision. On the left, colour shows the area’s place in the ranking of all 170, from poorest to most developed. Bucharest is one area, because the index has a single value for the whole city. Grey: fewer than 30 requests decided on the merits. Tap an area to pick that court; Bucharest’s six courts are picked from the search.

The law sets 72 hours. 76% of requests filed by victims get an answer within three days

Law 217/2003, art. 42(9): deciding a request "may not exceed 72 hours from filing". The one exception is a request that follows a provisional order issued by the police; that order is then extended by law until the court decides.

Of the requests filed by the victim (not by the prosecutor), 76% were decided within three calendar days of filing. The median is two days. 12% waited six days or more. The portal does not show whether a police provisional order came first; most provisional orders, however, reach the court through the prosecutor (see below).

31,095 requests filed by the victim in 2024–2025 and decided. The portal gives the filing time but only the day of the ruling, so the measure is "within three calendar days", an approximation that can include rulings given after 72 hours.

The gaps between courts are wide. Of the 145 courts with at least 30 such requests, 11 decided fewer than half within three days (seven of them beyond the statistical margin), and 32 more than nine in ten (14 beyond the margin). The slowest: Onești (28%), Strehaia (30%), Bucharest Sector 1 (34%), Drobeta-Turnu Severin (35%), Timișoara (37%).

When the prosecutor applies, nine in ten requests are granted

Since 2018 (Law 174/2018), a police officer who finds an imminent risk can issue a provisional protection order on the spot, valid for five days. A prosecutor confirms it and asks the court for an order on the victim's behalf. A prosecutor can also apply without an earlier provisional order.

Prosecutor-filed requests are granted in 89% of cases decided on the merits (13,282 requests), against 65% when the victim applies alone. The data cannot separate the effect of the evidence, the seriousness of the cases and the courts' practice.

In January–October 2025 the police reported 4,610 court orders that started as provisional orders; the portal gives 5,015 granted prosecutor requests in the same period. Most prosecutor requests therefore follow a provisional order, in which case that order is extended until the ruling and the 72-hour limit does not apply. Even so, 54% of prosecutor requests were decided within three days; the median is three days.

18% of decided requests end with the victim withdrawing

8,861 requests from 2024–2025 ended in withdrawal: 15% of those filed by the victim and 23% of those filed by the prosecutor. The ruling does not say why.

Law 232/2025 repealed, from 19 December 2025, article 43 of Law 217/2003, which let the victim withdraw a request made on their behalf by the prosecutor; the article is cited in 1,326 of the 8,861 withdrawal rulings of 2024–2025. In a ruling of 22 June 2026, published on 16 September 2026, the High Court of Cassation and Justice held that a victim may still withdraw a request made by the prosecutor on their behalf. Among requests filed in 2026 and decided by 23 October 2026, withdrawals are 22% of the prosecutor's requests and 13% of those filed by victims, close to the 2024–2025 shares.

What the order contains and how long it lasts

In the short text of granting rulings, the most often mentioned measure orders the abuser to keep a minimum distance from the victim (84% of summaries). A ban on any contact appears in 29%, and temporary eviction of the abuser from the home in 25%. The portal summaries are kept here up to 600 characters, and a mention does not always prove the measure was ordered; the shares are approximate.

The law allows an order of up to 12 months; since March 2026, up to 24 months if an earlier order against the same abuser expired within the previous five years. In the 18,782 summaries where the length can be read, the median is six months; 17% run for 12 months, and 34% for three months or less.The people asking for an order are women in 78% of cases, and the people it is sought against are men in 92% (read from the gendered Romanian words for "applicant" and "respondent" in the ruling: 37,362 requests where the applicant's sex can be read, 42,779 requests where the respondent's can). These three paragraphs come from the short text of the ruling and are not part of the tests fixed in advance, described below.

Requests are rising: 26,961 requests in the first nine months of 2026, against 21,506 in the same months of 2025

In January–September 2026 there were 25% more than in the same months of 2025. 2024 is left out of the comparison because the portal returns fewer 2024 cases than the police reported. The ANES helpline for domestic violence, sex discrimination and human trafficking also received almost 40% more calls in the first half of 2026 than a year earlier.

Requests filed per month at all local courts, from January 2025 to the last complete month (September 2026).

All local courts

Requests filed in 2024–2025 and decided, whoever filed them. "Granted": share of orders issued (in full or in part) among requests filed by the victim and decided on the merits. "Within three days" and "withdrawn": among all requests filed by the victim and decided. Click a column heading to sort.


Iași	Iași	2,097	76%	39%	20%
Brașov	Brașov	1,163	45%	60%	10%
Bucharest Sector 5	București	1,138	60%	84%	7%
Bacău	Bacău	1,107	61%	56%	15%
Constanța	Constanța	971	70%	85%	13%
Timișoara	Timiș	967	72%	37%	25%
Botoșani	Botoșani	941	66%	82%	5%
Focșani	Vrancea	933	68%	61%	28%
Cluj-Napoca	Cluj	923	63%	75%	23%
Bucharest Sector 2	București	908	58%	56%	20%
Ploiești	Prahova	883	64%	94%	6%
Bucharest Sector 4	București	861	45%	87%	27%
Galați	Galați	838	69%	74%	11%
Vaslui	Vaslui	834	68%	65%	29%
Bucharest Sector 6	București	828	70%	70%	21%
Bucharest Sector 3	București	825	66%	84%	14%
Piatra-Neamț	Neamț	763	68%	90%	7%
Craiova	Dolj	759	67%	70%	22%
Pitești	Argeș	729	66%	97%	10%
Buzău	Buzău	726	70%	93%	11%
Pașcani	Iași	685	64%	92%	7%
Buftea	Ilfov	646	60%	87%	8%
Arad	Arad	621	80%	81%	17%
Târgu Jiu	Gorj	620	68%	85%	24%
Giurgiu	Giurgiu	619	64%	90%	8%
Bucharest Sector 1	București	611	61%	34%	22%
Roman	Neamț	608	62%	95%	8%
Sibiu	Sibiu	544	58%	77%	15%
Bârlad	Vaslui	541	64%	98%	9%
Onești	Bacău	535	61%	28%	35%
Cornetu	Ilfov	533	49%	88%	5%
Brăila	Brăila	515	68%	84%	12%
Târgoviște	Dâmbovița	498	70%	96%	8%
Baia Mare	Maramureș	462	66%	73%	21%
Hârlău	Iași	453	83%	65%	20%
Râmnicu Vâlcea	Vâlcea	437	71%	71%	14%
Petroșani	Hunedoara	375	67%	91%	10%
Drobeta-Turnu Severin	Mehedinți	373	70%	35%	28%
Suceava	Suceava	366	71%	91%	12%
Hunedoara	Hunedoara	359	51%	95%	12%
Târgu Neamț	Neamț	354	52%	78%	14%
Rădăuți	Suceava	352	71%	94%	6%
Tecuci	Galați	348	55%	94%	5%
Moinești	Bacău	333	71%	70%	28%
Alba Iulia	Alba	327	60%	89%	12%
Bistrița	Bistrița-Năsăud	327	68%	86%	18%
Câmpina	Prahova	327	58%	100%	4%
Turda	Cluj	323	76%	81%	20%
Satu Mare	Satu Mare	318	80%	75%	12%
Tulcea	Tulcea	302	70%	40%	14%
Oltenița	Călărași	301	54%	84%	17%
Sfântu Gheorghe	Covasna	297	64%	84%	13%
Bolintin-Vale	Giurgiu	291	51%	55%	15%
Dorohoi	Botoșani	283	74%	88%	15%
Târgu Cărbunești	Gorj	283	52%	83%	7%
Fălticeni	Suceava	281	70%	80%	8%
Sighetu Marmației	Maramureș	265	68%	90%	22%
Făgăraș	Brașov	260	51%	75%	15%
Slobozia	Ialomița	244	71%	66%	8%
Călărași	Călărași	243	63%	58%	15%
Câmpulung	Argeș	241	67%	93%	8%
Găești	Dâmbovița	239	62%	99%	8%
Mizil	Prahova	235	62%	82%	10%
Alexandria	Teleorman	232	76%	83%	10%
Pătârlagele	Buzău	230	70%	76%	30%
Mediaș	Sibiu	226	79%	76%	12%
Medgidia	Constanța	221	58%	89%	8%
Motru	Gorj	216	75%	93%	13%
Vișeu de Sus	Maramureș	215	57%	94%	8%
Liești	Galați	214	59%	77%	16%
Deva	Hunedoara	211	58%	91%	9%
Panciu	Vrancea	202	63%	77%	21%
Fetești	Ialomița	201	70%	85%	7%
Făurei	Brăila	198	79%	89%	10%
Miercurea Ciuc	Harghita	191	67%	59%	16%
Mangalia	Constanța	190	73%	89%	12%
Săveni	Botoșani	190	59%	70%	9%
Costești	Argeș	189	34%	76%	7%
Râmnicu Sărat	Buzău	187	56%	93%	8%
Strehaia	Mehedinți	186	59%	30%	29%
Reșița	Caraș-Severin	184	69%	42%	19%
Huși	Vaslui	183	78%	81%	23%
Caransebeș	Caraș-Severin	181	58%	57%	13%
Turnu Măgurele	Teleorman	181	50%	100%	8%
Târgu Bujor	Galați	174	62%	63%	15%
Jibou	Sălaj	171	49%	72%	16%
Curtea de Argeș	Argeș	168	49%	91%	2%
Zalău	Sălaj	168	64%	73%	16%
Gherla	Cluj	163	69%	94%	17%
Târgu Secuiesc	Covasna	163	72%	89%	18%
Luduș	Mureș	162	63%	77%	28%
Aiud	Alba	153	64%	78%	14%
Sighișoara	Mureș	153	62%	87%	10%
Măcin	Brăila	149	65%	94%	10%
Filiași	Dolj	146	89%	88%	12%
Ineu	Arad	143	75%	55%	37%
Pogoanele	Buzău	140	53%	81%	21%
Răcari	Dâmbovița	138	77%	76%	8%
Orăștie	Hunedoara	137	67%	66%	21%
Sebeș	Alba	136	42%	92%	19%
Dej	Cluj	134	76%	83%	20%
Zărnești	Brașov	134	71%	81%	18%
Câmpulung Moldovenesc	Suceava	133	68%	93%	10%
Carei	Satu Mare	132	91%	43%	26%
Salonta	Bihor	132	58%	49%	48%
Urziceni	Ialomița	130	81%	54%	12%
Marghita	Bihor	129	77%	78%	22%
Rupea	Brașov	128	64%	49%	17%
Babadag	Tulcea	126	78%	69%	22%
Bicaz	Neamț	124	64%	59%	39%
Gura Humorului	Suceava	118	45%	79%	7%
Novaci	Gorj	118	–	–	–
Târnăveni	Mureș	118	74%	73%	32%
Vânju Mare	Mehedinți	116	76%	77%	16%
Răducăneni	Iași	115	84%	69%	15%
Topoloveni	Argeș	115	53%	100%	6%
Gheorgheni	Harghita	112	59%	80%	3%
Beclean	Bistrița-Năsăud	111	80%	75%	16%
Buhuși	Bacău	111	69%	79%	31%
Balș	Olt	108	70%	69%	16%
Sânnicolau Mare	Timiș	106	70%	66%	26%
Vatra Dornei	Suceava	106	47%	100%	10%
Băilești	Dolj	105	69%	98%	9%
Câmpeni	Alba	102	70%	55%	36%
Horezu	Vâlcea	100	57%	78%	12%
Toplița	Harghita	94	52%	90%	16%
Huedin	Cluj	93	78%	71%	39%
Odorheiu Secuiesc	Harghita	92	60%	89%	14%
Șimleu Silvaniei	Sălaj	92	71%	79%	21%
Adjud	Vrancea	89	79%	78%	8%
Bălcești	Vâlcea	89	65%	81%	33%
Brad	Hunedoara	88	55%	70%	11%
Negrești-Oaș	Satu Mare	85	33%	62%	10%
Pucioasa	Dâmbovița	84	59%	99%	9%
Chișineu-Criș	Arad	80	38%	62%	0%
Avrig	Sibiu	80	73%	89%	15%
Segarcea	Dolj	78	82%	82%	9%
Agnita	Sibiu	77	80%	75%	12%
Hațeg	Hunedoara	76	30%	78%	6%
Beiuș	Bihor	72	65%	85%	18%
Drăgășani	Vâlcea	70	68%	59%	8%
Brezoi	Vâlcea	68	60%	79%	6%
Lugoj	Timiș	68	41%	88%	9%
Podu Turcului	Bacău	65	25%	74%	11%
Hârșova	Constanța	65	50%	100%	5%
Oravița	Caraș-Severin	63	56%	59%	11%
Reghin	Mureș	63	55%	90%	6%
Dragomirești	Maramureș	62	72%	91%	3%
Săliște	Sibiu	62	77%	76%	18%
Videle	Teleorman	60	68%	77%	9%
Deta	Timiș	54	71%	74%	19%
Blaj	Alba	51	87%	76%	6%
Darabani	Botoșani	46	65%	79%	10%
Moldova Nouă	Caraș-Severin	44	73%	34%	24%
Târgu Lăpuș	Maramureș	44	36%	100%	7%
Baia de Aramă	Mehedinți	39	8%	70%	35%
Calafat	Dolj	38	83%	63%	20%
Gurahonț	Arad	37	73%	89%	11%
Orșova	Mehedinți	37	65%	62%	29%
Zimnicea	Teleorman	33	100%	90%	0%
Făget	Timiș	31	–	–	–
Vălenii de Munte	Prahova	30	68%	91%	17%
Întorsura Buzăului	Covasna	29	50%	100%	17%
Moreni	Dâmbovița	29	59%	95%	0%
Năsăud	Bistrița-Năsăud	28	60%	91%	9%
Caracal	Olt	26	–	–	–
Lipova	Arad	18	–	–	–
Aleșd	Bihor	12	–	100%	30%
Lehliu Gară	Călărași	11	–	–	–
Sinaia	Prahova	11	–	–	–
Oradea	Bihor	10	–	–	–
Slatina	Olt	8	–	–	–
Roșiori de Vede	Teleorman	6	–	–	–
Corabia	Olt	3	–	–	–
Bozovici	Caraș-Severin	0	–	–	–
Însurăței	Brăila	0	–	–	–
Murgeni	Vaslui	0	–	–	–
Târgu Mureș	Mureș	0	–	–	–
Șomcuta Mare	Maramureș	0	–	–	–

Why the gaps are not chance

The tests were written and fixed (with a digital fingerprint) before any court-level rate was computed. The stopping rule: if, for both granting and speed, the differences between courts did not exceed statistical noise or did not repeat from one year to the next, the court comparison would not be published.

The differences pass both tests. Courts that granted more often in 2024 also granted more often in 2025 (rank correlation 0.59, across 144 courts, all requests; 0.58 across 108 courts for requests filed by the victim, in the chart); courts that were fast in 2024 were fast in 2025 (0.77, across 119 courts). Variation between courts clearly exceeds noise (between-court spread on the logit scale, DerSimonian–Laird estimate: 0.34 for granting, 0.92 for speed; lower bounds of the 95% Q-profile intervals: 0.34 and 0.86).

Some courts write only "Respinge cererea" (request rejected) with no reason; here those rulings count as rejections on the merits. Counting as rejections only the rulings that say "unfounded", the national share becomes 68%, 90% at Arad local court and 48% at Brașov local court: the gap remains.

The classification was checked by hand on 300 randomly chosen requests: the outcome was read correctly in 296, and who filed it in 296.

The totals are close to those reported by the police, 3–4% above them. For the first half of 2025, IGPR reported 6,483 orders issued by courts; the portal data give 6,711 granted requests in the same period (104%). For January–October 2025: 13,029 against 12,663 (103%). For the first quarter of 2024 the portal gives only 85% of the police total (2,506 granted requests against 2,960 orders): 2024 is under-counted, probably because older cases drop out of the portal over time.

The differences between the observed rates repeat from one year to the next. The analysis does not isolate the effect of the court: courts receive different requests, and the portal does not say how serious the violence was in each case.

Each dot is a local court with at least 20 requests decided on the merits in each year. Dotted line: the same result in both years.

If you need a protection order in Romania

In immediate danger, call 112. For advice, call the free 24-hour helpline for victims of domestic violence: 0800 500 333.

The request goes to the local court for your home or residence, or for the place where the violence happened (Law 217/2003, art. 40).
There is no court fee (art. 41) and the request uses a form annexed to the law. A lawyer is required, and the court appoints one if you have none (art. 42).
A prosecutor can apply on your behalf, and the police can issue a five-day provisional order on the spot (arts 28, 31 and 40).
The law requires the court to decide within 72 hours, unless the police issued a provisional order first, which is then extended until the ruling (art. 42(9)). An order can last up to 12 months, or up to 24 months for a new order against the same abuser issued within five years of the earlier one expiring (art. 39).
Breaking an order is a crime punishable by six months to five years in prison (art. 47).

What the data do not show

The courts portal returns no cases filed before December 2023, and 2024 is incomplete (in the first quarter, 85% of the total reported by the police). Court rates for 2025 alone are close to those for 2024–2025.
The outcome is read from the short ruling text published on the portal, not from the reasoned judgment. "Granted in part" means an order was issued with fewer measures than requested.
Courts receive different requests. A lower grant rate can come from weaker requests, not only from the judges' practice. The tests show the differences repeat from one year to the next; they do not show the reason.
Only requests to issue an order under Law 217/2003 are counted. The 368 requests that cite Law 26/2024 (violence outside the family) are left out; others of that kind may remain among requests that cite no law. Requests to revoke, change or extend an order and challenges to police provisional orders are left out. Merged cases and cases sent to another court are counted once, at the court that decided them.
Requests filed in 2024–2025 still undecided on 23 October 2026 (1,583) are not in the rates.
Appeals, judges and panels are not analysed. No personal names or case numbers appear on the page.
The IDUL measures local development (schooling, health, housing, links to cities) from 2016–2018 data and set here against requests from 2024–2025. It is missing for some small communes; in every area, communes with an index hold at least 82% of residents. Court areas come from the government decision on court districts (Government Decision 337/1993), in its latest consolidated form of 2018. 243 communes created after 1993 (2% of the population) are not listed in it; each was placed with the court that most of its neighbouring communes in the same county belong to.

Method

The data come from the public query service of Romania's courts portal (portal.just.ro), queried month by month or quarter by quarter for each of the 179 local courts, for cases with the object "ordin de protecție" (91,789 cases, of which 82,373 requests to issue an order). Party names were removed at download; party roles, dates and the outcome were kept.

The outcome is the one from the first hearing in which the court ruled on the request. "Granted" = granted in full or in part; "rejected on the merits" = rejected as unfounded or with no other reason given; withdrawals and requests annulled, inadmissible or moot are counted separately. Prosecutor-filed requests were recognised from the ruling text ("filed by the Prosecutor's Office…").

Speed: calendar days from the filing date and time to the ruling date. The grey band in the chart is the 99.8% interval around the national average for a court with the same number of requests.

The analysis plan, with the tests and the stopping rule, was fixed before computing. Four deviations, all before the final run: the stability test for granting was run on all requests, while the plan asked for one per type of filer (the result for victim-filed requests is given separately); the interval method for the between-court spread, a few classification rules (requests under Law 26/2024 and procedural rulings removed, reading the prosecutor as the filer) and a repaired anonymisation error, after which the filer was re-checked on 80 requests (79 correct). Per-court data can be downloaded below.

The poverty comparison was fixed before any number was computed (24 October 2026): the Spearman correlation between an area’s index and each measure, with a bootstrap confidence interval, then the same correlation among courts of the same county. Poverty would count as going with a measure only if the national correlation passed 0.2 with an interval clear of zero and the within-county correlation had the same sign, also with an interval clear of zero. None did: grant rate −0.06 (within counties 0.01), decision within three days 0.06, requests per 10,000 residents −0.20 (within counties −0.02), share of decided requests filed by a prosecutor −0.09 (within counties −0.21, but no national effect). With town halls’ own revenue per resident in place of the index, the grant rate gives −0.02. Correlations with the grant rate and speed use areas with at least 30 requests.

Download per-court data (CSV)