The Unfulfilled Promise of Section 479 BNSS: Why Jails Aren't Decongesting Despite the 'One-Third' Bail Rule
Section 479 of the Bharatiya Nagarik Suraksha Sanhita, 2023 was welcomed as a rare piece of legislation that turns a constitutional aspiration into a number. A person who has been detained for up to one-half of the maximum sentence for an offence must be released on bail, and a first-time offender, someone never convicted of any offence, must be released on bond at one-third. For decades India's jails have held a population in which roughly three of every four inmates are undertrials. Parliament's formula looked like a remedy. The first full national statistics suggest the remedy has reached only a sliver of the patients.
What the Provision Says
Section 479(1) replaces Section 436A of the old Code. It covers any person detained during investigation, inquiry or trial for an offence not punishable with death or life imprisonment. For such a person, detention beyond one-half of the maximum term triggers release on bail. The first proviso reduces the threshold to one-third for a first-time offender, who is released on bond. A further proviso lets the court, after hearing the Public Prosecutor and recording reasons, order longer detention or require sureties, and the section forbids detention beyond the maximum term in any event. Delay caused by the accused is excluded from the count.
Two features temper the generosity. Section 479(2) provides that a person facing investigation, inquiry or trial in more than one offence or in multiple cases shall not be released under the section. And Section 479(3) casts a duty on the jail superintendent to make a written application to the court once the relevant fraction is complete. The statute thus assumes an administrative machine that identifies eligible prisoners, computes their entitlements and moves the court, without waiting for a lawyer.
The Supreme Court has pressed that machine. On 23 August 2024, a Bench of Justices Hima Kohli and Sandeep Mehta, in In Re: Inhuman Conditions in 1382 Prisons, held that Section 479 applies to all undertrials, including those in cases registered before 1 July 2024. On 22 October 2024, Justices Hrishikesh Roy and S.V.N. Bhatti called for proactive release and special attention to eligible women prisoners. In 2026, a Bench of Justices Vikram Nath and Sandeep Mehta directed all States and Union Territories to file affidavits giving jail-wise capacity and occupancy data as on 1 March 2026, along with decongestion measures.
The Numbers
The most reliable data come from the National Crime Records Bureau's Prison Statistics India 2024, as analysed in a recent preprint. They complicate any simple story of failure. National occupancy fell from 131.4 per cent in 2022 to 120.8 per cent in 2023 and 112.7 per cent in 2024. Prisoners at year-end fell from 573,220 to 511,542 across that period, while sanctioned capacity grew by 3.3 per cent in the last year. Undertrials numbered 371,440, down from 434,302 in 2022. Yet they still formed 72.6 per cent of all prisoners, and jails remain above capacity nationally. Delhi's occupancy was 194.6 per cent.
Section 479's own footprint is small. NCRB recorded 968 undertrials as eligible, 467 applications moved by jail superintendents, and 627 releases. Those releases amount to roughly 0.17 per cent of the undertrial population. Four jurisdictions, Uttar Pradesh (287), Rajasthan (78), Chhattisgarh (53) and Delhi (41), accounted for almost three-quarters of them. Two cautions apply. The 2024 figures capture at most six months of operation under the new law, and West Bengal's 2023 figures were substituted for its missing 2024 return. The data do not permit causal claims, but 627 releases cannot explain a fall of tens of thousands in the undertrial population. Ordinary bail accounted for the great majority of the roughly 15.9 lakh undertrial releases that year, and expanded capacity helped the occupancy figure.
Why the Mathematics Bites Less Than Expected
The first reason is simple arithmetic. A time-based threshold helps only those who stay long, and most undertrials do not. About seventy per cent have been detained for a year or less. For a three-year offence, a first-timer reaches the one-third mark at twelve months; for a seven-year offence, at two years and four months; for a ten-year offence, at three years and four months. Just over 60,000 undertrials had been confined for more than two years, and about 9,000 for more than five. These long-stayers are disproportionately accused of serious offences, which is where the exclusions and statutory bail bars operate.
The second is the exclusion of offences punishable with death or life imprisonment, a category that includes murder and several other grave crimes. The third is Section 479(2). Read literally, 'more than one offence' would defeat nearly every accused, since a single charge-sheet routinely joins several sections. A sensible construction confines the bar to separate pending proceedings, and it is the poor, who are frequently named in multiple FIRs, who stand to lose most from a loose reading. The Supreme Court should settle this expressly.
The fourth is the difficulty of proving a person is a first-time offender. 'Never convicted' requires a reliable conviction record, and jails rarely have one. The fifth is administration. The gap between 968 eligible prisoners and 467 applications shows that the superintendent's duty is honoured unevenly, and the figures say nothing of those never identified. The sixth is the exclusion of delay attributable to the accused, an open-textured exception capable of expansive use. The seventh is calculation: where several offences are charged, which maximum term governs is itself a source of dispute, and each dispute postpones release. Notably, 'first-time offender' turns on conviction, not arrest, so a person with earlier acquittals still qualifies, subject to the multiple-case bar. Finally, there is the interplay with special statutes. Where the NDPS Act, the UAPA or the PMLA imposes its own conditions on bail, courts must reconcile those with Section 479, and the answer is not uniform.
The Constitutional Frame
Section 479 is not an isolated policy choice. It carries forward a line of authority beginning with Hussainara Khatoon v. State of Bihar (1979), in which Justice P.N. Bhagwati held speedy trial to be integral to Article 21 after discovering undertrials jailed beyond the maximum sentence for their alleged offences. K.A. Najeeb (2021) held that constitutional courts may grant bail despite statutory bars where long incarceration and delayed trial make continued detention unjust. Satender Kumar Antil v. Central Bureau of Investigation (2022), by Justices S.K. Kaul and M.M. Sundresh, urged a separate Bail Act and set guidelines to keep bail the rule.
A balanced reading must also respect the opposing considerations. Release under Section 479 is not an acquittal. Victims, witnesses and the integrity of the trial deserve protection, which is why the section retains judicial discretion, a hearing for the prosecutor and bond conditions. Time-based rules are blunt, and a prisoner released at one-third may still be a flight risk or a threat to witnesses. But those concerns support targeted conditions, not systemic non-compliance with a statutory entitlement that Parliament made mandatory by saying 'shall'.
From Promise to Practice
The problem Section 479 faces is implementation. Several measures follow from the evidence. First, eligibility should be computed by software, not memory. Prison and case-management systems should flag each undertrial's one-third and one-half dates at the time of admission, and generate the superintendent's application automatically. Second, District Legal Services Authorities and Undertrial Review Committees should audit eligibility monthly and publish district-wise figures. Third, the Supreme Court should resolve the meaning of Section 479(2) and the treatment of special-statute bail restrictions, so that trial courts are not left to guess. Fourth, accountability should be built in, with courts required to decide Section 479 applications within a fixed period and record reasons for refusal. Fifth, because release on bond for first-time offenders requires no surety, courts should resist converting the entitlement into a surety requirement that the poor cannot meet.
Finally, no release formula can substitute for speedy trial. Section 479 manages the consequences of delay, not its causes. Until trials conclude within reasonable time, the one-third rule will remain an important but inadequate safeguard.
The promise of Section 479 is arithmetic, and arithmetic does not enforce itself. Two years after its commencement, the law has shown that it can release people. It has not yet shown that it will.
Author is an Advocate practicing at Delhi High Court. Views are personal.