National implementations of the shared European floor diverge sharply on the actual numbers. Italy tiers its maximum terms by offense severity. Germany triggers a mandatory six-month Higher Regional Court review. France's initial four months can extend to four years for the gravest offenses. The United Kingdom's statutory custody time limits cap Crown Court remand at 182 days. The United States took the most structurally distinct path of the five: the Bail Reform Act of 1984 introduced federal preventive detention for dangerousness, while cash bail, a wealth-based mechanism absent from the four European systems studied here, remains the dominant gatekeeping tool outside the federal system.
The European Court of Human Rights' own reasonable-time standard has tightened over five decades: Wemhoff v. Germany tolerated roughly three and a half years of pretrial detention in 1968, while Buzadji v. Moldova required individualized reasoning from the very first judicial review by 2016. Legislating an alternative to detention does not guarantee its use, and the European Supervision Order's documented near-total non-uptake is the clearest example. Roughly one in four European prisoners was in pretrial detention as of January 2025, a figure the Council of Europe's own SPACE I report says follows no clear regional pattern.
The ECHR Article 5 Framework: Liberty as the Rule, Detention as the Exception
Article 5 of the European Convention on Human Rights treats liberty as the default and detention as the exception requiring justification, and three of its sub-provisions do distinct doctrinal work in enforcing that default.
Article 5(1)(c) permits detention effected for the purpose of bringing a person before the competent legal authority "on reasonable suspicion of having committed an offence," or where reasonably necessary to prevent flight or the commission of an offence. This sub-provision sets the substantive grounds: reasonable suspicion is a necessary condition for any detention at all, though it stops being sufficient on its own once an initial period has passed.
Article 5(3) requires that anyone detained under 5(1)(c) "shall be brought promptly before a judge" and "shall be entitled to trial within a reasonable time or to release pending trial," with release conditional if necessary on guarantees to appear. This is where procedural promptness meets a durational limit, not a fixed number of days, but a judicially assessed standard applied case by case, a point the first FAQ question below addresses directly.
Article 5(4) sets a third, separate remedy: a speedy court determination of the lawfulness of any detention, with release ordered if the detention proves unlawful. Unlike 5(3)'s one-time promptness requirement, 5(4) functions as an ongoing, habeas-style check available throughout the detention period.
These three provisions, the grounds requirement, the durational ceiling, and the standing remedy, recur throughout the case law below.
Article 5(3)'s reasonable-time clock measures only the length of detention itself, not the length of the case as a whole. A separate reasonable-time guarantee under Article 6(1) covers how long the entire proceeding may run, a line of doctrine traced through Kudła v. Poland and Italy's own Pinto Law response to it. the companion article on fair-trial standards under Article 6 ECHR develops that distinct guarantee in full; this article's concern stays with Article 5 and the length of detention specifically.
National Regimes Compared: Italy, Germany, France, the UK, and the US
That divide, grounds-based detention against a wealth-based mechanism, organizes what follows: how each system defines its own grounds for detention, and where the United States departs from the other four.
Italy's custodia cautelare (Art. 272-315 c.p.p.) requires "serious indications of guilt" (gravi indizi di colpevolezza) before any coercive measure can be applied, plus an Article 274 esigenza cautelare, a ground equivalent to the Letellier catalogue below. Maximum terms are tiered by both offense severity and procedural stage: during the preliminary investigation, the cap is three months for offenses carrying up to six years' maximum imprisonment, six months for more serious offenses below the life-imprisonment or twenty-year threshold, and one year where the charge carries life imprisonment or a minimum twenty-year sentence. During first-instance trial, the corresponding tiers run six months, one year, and one year six months.
Article 275, as amended by Law 47/2015 in direct response to the European Court of Human Rights' Article 3 pilot judgment in Torreggiani and Others v. Italy (App. No. 43517/09, judgment 8 January 2013), which found Italy in violation over systemic prison overcrowding, imposes a least-afflictive-measure principle: custodial detention is a genuine last resort, available only where lighter measures would be inadequate. Italy separately compensates wrongful pretrial detention through its ingiusta detenzione track, a distinct process the wrongful-conviction remedies article covers in depth, including the case count and payout total.
Germany's Untersuchungshaft (Sections 112-130 StPO) requires, under Section 112 specifically, urgent suspicion, a stronger threshold than mere reasonable suspicion, plus an independent ground: flight risk, risk of collusion or evidence tampering, or, for certain serious offenses, risk of repetition. Sections 121-122 StPO supply the structurally distinctive German safeguard: once detention on the same charge passes six months, the case is automatically referred to the Higher Regional Court (Oberlandesgericht), which independently examines whether continued detention remains justified. This article does not state Germany's national pretrial-detention percentage; the statistics section below gives the population figures that are available instead.
France's détention provisoire is authorized and renewed exclusively by the juge des libertés et de la détention (JLD), a magistrate structurally independent of the investigating judge (juge d'instruction) who requests the measure, under Articles 143-1 to 148-8 of the Code de procédure pénale. Initial duration is four months, extended to six months for terrorism or drug-trafficking cases. Total duration can extend to one year for lesser offenses (délits), and up to two years for aggravated crimes or four years for the gravest category: organized crime, drug trafficking, or terrorism. At every renewal, the JLD must assess the gravity of the facts, the complexity of the investigation, and the conduct of the authorities, a test that mirrors the Strasbourg reasonableness inquiry below, applied repeatedly rather than once.
The United Kingdom's custody time limit (CTL) regime, set by the Prosecution of Offences Act 1985 and its implementing regulations, fixes maximum periods in days rather than months or years: 56 days for offenses triable only in a magistrates' court, 70 days in the magistrates' court for either-way offenses pending a venue decision, and 182 days in the Crown Court, running from the sending of the defendant for trial. These are fixed statutory numbers, unlike the Convention's reasonable-time standard, but not absolute. An extension requires the prosecution to show "good and sufficient cause" and that the case has been conducted "with all due diligence and expedition." If an extension is refused and the limit expires, the defendant is released on bail; the case itself continues rather than being dismissed.
The Bail Reform Act of 1984 (18 U.S.C. Sections 3141-3150) authorizes federal courts to detain an arrestee before trial where the government shows, by clear and convincing evidence after an adversarial hearing, that no combination of release conditions can reasonably assure the safety of any other person or the community, a standard upheld against due-process and Eighth Amendment challenge in United States v. Salerno, 481 U.S. 739 (1987). That narrow federal model governs only a minority of US cases; for the large majority, state-level prosecutions, cash bail remains the dominant gatekeeping tool.
The ECtHR Jurisprudence That Shaped the Floor: Five Cases, Five Decades
Wemhoff v. Germany, App. No. 2122/64, judgment 27 June 1968, supplies the doctrine's permissive origin point. Gert Wemhoff was detained from 9 November 1961 to 7 April 1965, roughly three years and five months, on suspicion of complex, multi-bank fraud across Berlin, West Germany, and Switzerland. The Court found, six votes to one, no violation of Article 5(3), reasoning that the investigation's genuine complexity justified the length of detention.
Letellier v. France, judgment 26 June 1991, supplies the canonical catalogue of grounds a court must identify once reasonable suspicion alone is no longer sufficient: danger of absconding, risk of pressure on witnesses or destruction of evidence, risk of reoffending, and the need to preserve public order. This catalogue underlies Italy's Article 274, Germany's Section 112, and France's Article 144, even where each jurisdiction's code expresses the categories differently.
McKay v. the United Kingdom, App. No. 543/03, (2007) 44 EHRR 41, Grand Chamber judgment 3 October 2006, establishes the architecture of automatic, prompt judicial review under Article 5(3). The officer before whom a detainee is first brought need not personally hold the power to grant bail, so long as some officer within the system does, and the review itself must be sufficiently wide to weigh the actual circumstances for and against continued detention, not merely a formal check. McKay's judicial-independence-and-scope requirements sit directly beside the fair-tribunal doctrine developed under Article 6, covered by the companion article on fair-trial standards.
Idalov v. Russia, App. No. 5826/03, Grand Chamber judgment 22 May 2012, established that separate spells of detention on the same underlying facts, interrupted by a brief release, must be aggregated for reasonable-time purposes rather than treated as independently restarting. Bagau Idalov was arrested on 11 June 1999, released on bail in July 2001, and had that bail discontinued and detention re-ordered in October 2002. The Grand Chamber found three separate Article 5 violations in his case: the overall length of detention, Russia's failure to examine his appeals speedily, and his exclusion from the hearings that decided his own detention.
Buzadji v. the Republic of Moldova, App. No. 23755/07, Grand Chamber judgment 5 July 2016, resolves an acknowledged ambiguity in the Court's own prior case law. Vladimir Buzadji, a Moldovan businessman, was investigated from July 2006 for an alleged fraud attempt and held roughly ten months on Moldovan courts' "stereotyped and abstract" reasoning. The Grand Chamber held that individualized, non-stereotyped reasons for detention are required from the very first judicial review, not after some vague "lapse of time" as earlier case law had left unclear. This is the doctrine's current, tightened standard, narrowing the permissiveness the Court itself modeled in Wemhoff nearly fifty years earlier.
Read together, these five cases trace a single arc: Wemhoff's permissive 1968 origin narrowed through Letellier's grounds catalogue, McKay's review architecture, and Idalov's aggregation rule, to Buzadji's 2016 individualized-reasoning requirement. The Court did not apply a static reasonable-time standard across five decades; it tightened its own doctrine.
Alternatives to Detention: What Gets Legislated, What Actually Gets Used
If a country creates a legal alternative to pretrial detention, does it actually get used? Not necessarily, and the clearest example is an EU-wide instrument that exists almost entirely on paper.
Release on recognizance (ROR) releases a defendant on their own promise to appear, without a financial transaction, sometimes paired with reporting or curfew conditions calibrated to assessed risk.
Active electronic monitoring (AEM) sits functionally between ROR and full detention, and the evidence on its effectiveness is genuinely mixed rather than settled in either direction. Florida's Pinellas County reported 99.6% court-appearance compliance and 96% no-new-crime rates among monitored defendants. A separate, methodologically distinct study found the monitored group's pretrial rearrest rate roughly 9 percentage points higher than a non-monitored comparison group. A third study found electronic monitoring reduced failure-to-appear by 10.6 points and new pretrial offenses by 7.4 points relative to unconditioned release, and saved roughly $8,200 per defendant compared to full detention, without significantly lowering overall system costs. These three findings do not resolve into a single verdict; the evidence base itself remains contested.
The European Supervision Order (Council Framework Decision 2009/829/JHA, 23 October 2009) lets a non-resident EU citizen charged in one Member State be supervised in their home Member State using non-custodial measures instead of remaining in pretrial detention abroad. Its real-world record is documented plainly in multiple 2020s academic sources as a near-total practical failure: the instrument is described as "hardly known and even less often used," and Italian-specific commentary calls it "completely unimplemented in Italian practice" despite its "enormous potential." This article's concern is the order's documented non-uptake; the mutual-recognition mechanics behind it are covered by the companion article on cross-border criminal justice cooperation.
The lesson generalizes beyond the European Supervision Order specifically. Legislating an alternative to detention does not, by itself, produce its use, a counter-narrative worth remembering when alternatives-to-detention reform is proposed as a straightforward fix.
Statistics and Disparities: Who Gets Held Pretrial, and Why
How many people are held pretrial in Europe and the United States, and are the figures comparable? Only with real caution.
The Council of Europe's SPACE I report (Aebi and Cocco, University of Lausanne, "Prisons and Prisoners in Europe 2025") found a pretrial detention share of roughly one in four European prisoners as of January 2025, ranging from 7 to 8 percent in Bulgaria and Czechia to 45 to 62 percent in the Netherlands, Switzerland, Armenia, Montenegro, and Albania. The Council's own report states this spread "does not follow a clear regional trend," resisting any simple East-West or common-law-versus-civil-law reading of the data.
Among the five jurisdictions this article compares, Italy holds approximately 26% of its prison population pretrial. France's prévenus (pretrial detainees) reached 26.7% of its total detained population as of 1 October 2024 (20,900 of 78,300). The United Kingdom's remand population reached approximately 17,070 people by June 2024, roughly 20% of the total prison population in England and Wales, the highest remand level in at least fifty years.
The United States reports 21.6% of its combined prison-system population as pretrial, per World Prison Brief's methodology, alongside a separately measured 69% of the local jail population unconvicted at 2024 midyear. These two US figures use different denominators (jails also hold short sentences and other holds), and neither is directly comparable to SPACE I's methodology. Germany's total prison population stood at 57,955 in 2023 and 43,746 as of 31 March 2024 (40,405 imprisoned, 2,737 in youth custody, 604 in preventive detention); a specific national pretrial-detention percentage for Germany was not confirmed and is not stated here.
Fair Trials International's EU-focused research found more than 120,000 people in pretrial detention across the EU at any time, more than one in five of all EU prisoners, at an average annual cost of 44,256 euros per detainee (36,473 euros in the UK; 24,000 euros in Germany).
Disparity data is far more developed for the United States than for the three continental European systems studied here. Black defendants are, holding similar circumstances constant, 3.6 percentage points more likely to be assigned monetary bail than white defendants, with average bail amounts $7,280 higher. In large urban areas, Black defendants accused of felonies are 25% more likely to be held pretrial than white defendants, and Black and Latino people, 30% of the national population, make up roughly 50% of all pretrial detainees. On the socioeconomic axis, roughly a third of US defendants cannot afford even a "typical" bail amount given median income, and an estimated 555,000 people, roughly a quarter of the entire US incarcerated population, are held pretrial specifically because they cannot afford bail.
Because those three systems determine detention on judicially assessed risk grounds rather than ability to pay, the specific wealth-based disparity mechanism documented in the US has no direct structural equivalent there. But no comparable disparity studies, by nationality, ethnicity, or socioeconomic status, were located for pretrial decision-making within those risk-based systems specifically, a genuine research gap rather than evidence that no such disparity exists.
Recent Reforms, 2020-2026: A Cross-Jurisdictional Roundup
Illinois's Pretrial Fairness Act, part of the broader SAFE-T Act, took effect 18 September 2023, making Illinois the first US state to eliminate cash bail entirely. Its constitutionality, argued before the Illinois Supreme Court in March 2025, remains undecided. Early data show Cook County jail populations declining without a corresponding rise in failure-to-appear rates, and a national 33-city study found no relationship between bail reform and crime rates. The countervailing trend is real too: at least 14 states saw roughly 20 Republican-sponsored bills in 2024-2025 aimed at expanding non-bailable offense categories or requiring judges to weigh criminal history more heavily.
Italy's Legge Nordio (Law 114/2024), in force 25 August 2024, requires a mandatory, fully recorded interrogation before a custodial preventive measure can be ordered, with collegiate, three-judge custody decisions phasing in from 25 August 2026. Justice Minister Carlo Nordio has publicly framed the reform's purpose as reducing recourse to pretrial custody, a specific policy goal, not a vague modernization claim.
France's Loi n. 2023-1059 of 20 November 2023, the 2023-2027 justice programming law, raised the Justice Ministry's budget 26%, from 7.6 billion euros to 9.6 billion euros, funding roughly 700 additional judges. On pretrial detention specifically, it unified the trial deadline for detained defendants requesting early trial into a single ten-week limit and created a new mechanism allowing conditional electronic home confinement for up to fifteen days pending feasibility verification.
The United Kingdom's remand population reached its highest level in at least fifty years by June 2024, with roughly 32% of remand prisoners, per a September 2022 measure, already held beyond the standard six-month Crown Court custody time limit through extensions. The government's SDS40 scheme, lowering the automatic release point for sentenced prisoners from 50% to 40% of term, is a system-wide capacity-relief measure, not a reform targeted at remand specifically, though it indirectly eases pressure on remand as well.
Across all five jurisdictions, the 2020-2026 window shows active, contested legislative attention to pretrial detention's proper limits, not a settled area of law.
Frequently Asked Questions About Pretrial Detention Limits
Does the European Convention on Human Rights set a fixed number of days a person can be held before trial?
No. Article 5(3) imposes a reasonable-time standard that a court assesses case by case rather than a specific day count. Wemhoff v. Germany, the Court's own founding judgment on the question, decided in 1968, tolerated roughly three and a half years of detention given the case's complexity. Numerical caps come from national law instead: Italy's tiered terms, France's four-month-to-four-year range, and the United Kingdom's day-based custody time limits.
Is pretrial detention in Europe mostly a wealth problem, like cash bail in the United States?
No. Italy, Germany, and France each require a judge to find a specific risk, flight, tampering, reoffending, or a threat to public order, before ordering or continuing detention; ability to pay plays no role in that assessment. A wealth-conditioned release mechanism comparable to American cash bail does not appear in any of the three continental systems; among the five jurisdictions compared here, that structural feature is distinctly American.
If a country creates a legal alternative to pretrial detention, does it get used in practice?
Not automatically. The clearest counterexample is the European Supervision Order, created in 2009 to let a non-resident EU defendant be supervised at home rather than detained abroad. Multiple 2020s academic sources describe it as barely known and rarely invoked, a description that holds specifically for Italy as well.
Are pretrial detention rates directly comparable across European countries?
Only with real caution. The Council of Europe's own SPACE I data shows detention shares ranging from single digits to over 60% of a country's prison population, and the Council's own reporting says that range tracks no clear regional pattern. The countries with the highest and lowest shares do not sort neatly along an East-West or common-law-versus-civil-law line.
Related Articles
This article describes the judicial-review stage: the grounds a court must find, and the outer time limits it must respect, once a person is already in custody awaiting trial. Two adjacent questions sit just outside its scope. The fair-trial guarantees under Article 6 of the European Convention on Human Rights, which interlock with the judicial-independence requirements McKay establishes for reviewing detention, are the subject of the companion article on fair-trial standards. The period before that first judicial review, the arrest itself and the immediate custody that follows it, is covered from an institutional-accountability angle by the companion article on police accountability and civil rights protections.
Readers specifically seeking Italy's compensation data for wrongful pretrial detention should consult the wrongful-conviction remedies article, which covers that separate track in depth.