The juvenile age debate in India is usually described as an argument about an age, and that description is the first thing wrong with it. The age never changed. A child in Indian law was a person below eighteen before the Juvenile Justice (Care and Protection of Children) Act, 2015 was passed, and a child in Indian law remains a person below eighteen after it. What Parliament built in 2015 was not a new age of juvenility but a door in the wall of the old one: a mechanism by which a child who has completed sixteen years, and who is alleged to have committed an offence of the gravest statutory grade, can be assessed by a Juvenile Justice Board and, if two successive bodies are satisfied, moved out of the juvenile system and tried as an adult by a Children’s Court. The argument was therefore never really about where childhood ends. It was about whether a single case, and the public feeling a single case generated, justified an exception to a rule that the Supreme Court had upheld, that a committee of jurists had advised keeping, and that the parliamentary committee which examined the very Bill containing the exception had recommended against enacting. This article treats the debate as what it most instructively is: a legislative process question, whose chronology can be stated exactly, whose two sides can each be put at full strength, and whose central fact is one that almost every retelling omits. The committee said no, and the House said yes, in the same season in which the young person whose case had started the argument completed his term and walked out of a special home.

Both sides of this argument can be stated in terms the other would accept, and the discipline of this article is to keep them in that form. Those who carried the provision said that a legal system which answers a planned killing or a gang rape by a person of seventeen years and eleven months with a maximum of three years in a special home has failed the victim, failed public confidence, and failed the elementary idea that consequences should bear some proportion to the act; and that an individual assessment, conducted before any transfer and repeated by a court, is a fairer instrument than either blanket immunity or a blanket reduction of the age. Those who opposed it said that the developmental evidence on which the uniform age of eighteen rests does not stop being true because a case is horrifying; that the assessment the provision demands, of a particular child’s mental capacity and understanding, is a determination the science does not support making reliably; that the children who stand before Boards are overwhelmingly the children of the poor, so a discretion of this size will be exercised on the children least able to contest it; and that the committee system of Parliament exists precisely so that a Bill of this consequence is tested against evidence before feeling becomes law, and on this occasion the test was run, the answer was given, and the answer was set aside. Between those two statements lie a set of facts that are not in dispute, a set of factual claims that are, and a residue of value judgment that no evidence can settle. The article takes them in that order.

What the juvenile age debate is about: the provision as enacted

Most public argument about this provision misdescribes the text, in both directions, and the misdescription is load bearing: the provision sounds wider when its opponents describe it and narrower when its defenders do. The statute that contains it is described in full in the Juvenile Justice Act 2015 explained, and what follows here is only the machinery of the disputed exception, stated without characterisation. Section 2(12) of the Act defines a child as a person who has not completed eighteen years of age, and Section 2(13) defines a child in conflict with law by reference to the date of the offence, so that the Act governs a person apprehended at twenty for something done at seventeen. Offences are graded by the punishment the law prescribes for them. Section 2(45) makes an offence petty where the maximum punishment is imprisonment up to three years; Section 2(54) makes it serious where the punishment runs from three to seven years; and Section 2(33) makes it heinous where the minimum punishment is imprisonment for seven years or more. Only the third grade, and only for a child who has completed or is above the age of sixteen years, opens the door.

For that class alone, Section 15 requires the Juvenile Justice Board to conduct a preliminary assessment with regard to the child’s mental and physical capacity to commit the offence, the child’s ability to understand its consequences, and the circumstances in which the offence was allegedly committed. The Board may take the assistance of experienced psychologists, psycho-social workers or other experts, and an Explanation in the section states that the preliminary assessment is not a trial. The assessment must be disposed of within three months of the child’s first production before the Board. If the Board is satisfied that it should dispose of the matter itself, it follows the procedure for summons cases and its orders are those of Section 18, whose ceiling is three years in a special home. If the Board concludes that there is a need for trial of the child as an adult, Section 18(3) empowers it to order the transfer of the trial to the Children’s Court. The Children’s Court then decides the question again, for itself, under Section 19: it may direct trial as an adult under the ordinary criminal procedure, or it may decide there is no such need and conduct the inquiry as a Board would. Transfer is therefore a conclusion two bodies must reach in sequence, and either body can end it.

What is the most a Juvenile Justice Board can order on its own?

Three years. Where a Board finds that a child has committed an offence, the heaviest order within its own power is placement in a special home for a period that may extend to three years, and the ladder below that ceiling runs through probation of good conduct, community service, group counselling, fine, and advice or admonition at its foot.

That ceiling is the fixed point around which the whole debate turned, because it was also the ceiling of the repealed Act of 2000, and it applied under that Act to every child whatever the allegation. A person of seventeen years and eleven months found to have committed a murder faced the same maximum disposition as a person of thirteen found to have committed a theft, and the only variable the Board could adjust was where on a short ladder of reformative orders the case should land. The defenders of the 2015 provision treated that uniformity as the defect; its opponents treated it as the principle. Both descriptions are accurate, and the disagreement between them is the debate.

The protections that survive a transfer are written into the same part of the statute, and they are substantial. Section 21 provides that no child in conflict with law shall be sentenced to death, or to imprisonment for life without the possibility of release, whether under this Act, the Indian Penal Code or any other law in force. A transferred child can therefore receive a long sentence, and can receive life imprisonment where the offence carries it, but only a life sentence from which release remains possible. Section 19(3) requires a child found in conflict with law by a Children’s Court to be sent to a place of safety, not to a prison, until the age of twenty one years, with reformative services provided during the stay and a progress report made every year; Section 20 then requires the Children’s Court, when the person reaches twenty one and the term is unexpired, to evaluate the record of reformative change and decide between release on conditions, with a monitoring authority appointed for the remainder of the term, and transfer to a jail to complete it. Section 23 bars joint trial of a child with an adult, so the transferred child is never tried alongside adult co-accused. The compensating loss is in Section 24: the removal of disqualification and the destruction of records that protect every other child dealt with under the Act do not apply to a child tried as an adult and found in conflict with law by the Children’s Court, so the transferred child carries the conviction and its ordinary consequences into adult life.

What protections does a child tried as an adult keep?

The child keeps two protections above all. No death sentence may be passed, and no sentence of life imprisonment without the possibility of release; custody until twenty one is in a place of safety, with a review at twenty one. But the destruction of records and the removal of disqualification that protect other children do not apply.

One feature of the mechanism attracted almost no attention in the passage debate and has attracted a great deal since, because the mechanism stands on two findings and not one. Transfer requires that the child be found to have completed sixteen years on the date of the offence, and the assessment of capacity is reached only after the gateway of age has been crossed. In a country where birth registration has historically been incomplete and school records are the first resort of the age determination the Act prescribes, the sixteen year line is in practice litigated as hard as the eighteen year line ever was, with the same documentary frailties and an additional year band inside which the outcome decides the forum. The debate of 2015 was conducted throughout as though ages were facts printed on the children concerned; the system’s daily experience is that age is a conclusion, reached on certificates that may not exist and medical opinion with margins of error on both sides of the line, and that the transfer class is therefore bounded not only by the assessment’s quality but by the proof of the birthday that admits a child to it. A provision designed to substitute judgment about a person for a rule about a date thus begins, in operation, with a dispute about the date, and the elegance of the enacted design is at its most fragile exactly where its predecessor was at its most arbitrary.

This article describes a statute and a debate; it is legal information and not advice on any individual matter, and a family facing an actual preliminary assessment should take advice on the facts of that case, because assessment orders turn on the record of the particular child and are open to challenge in appeal and revision. The description above is the whole of the disputed mechanism. Nothing in it lowers the age of a child, nothing in it makes transfer automatic on a grave charge, and nothing in it exposes a transferred child to the gallows. A reader who holds those three negations clearly has the text; the argument begins over whether the mechanism, so bounded, should exist at all.

Three designs were available, and the choice among them is the debate

Before the chronology, the design space needs mapping, because much of the argument’s confusion comes from the parties defending different instruments as though they were one. A legislature dissatisfied with a uniform age line has three distinct instruments available, and India considered all three between 2013 and 2015. The first is to move the line: reduce the age of juvenility itself, so that every person above the new age is an adult for every purpose and every offence. That was the principal public demand after December 2012, it was the demand the Verma Committee considered and rejected, and it is the instrument the enacted law is persistently but wrongly believed to have used. Its virtues are certainty and simplicity; its cost is that it sweeps the impulsive child and the calculating one into the adult system together, for a stolen bicycle as for a killing, and abandons the developmental premise for the entire age band rather than for a class of cases. The second instrument is exclusion by offence: the statute itself removes defined grave offences from the juvenile system’s jurisdiction, so that the charge, and not the child, decides the forum. Versions of this design, usually called statutory exclusion, operate in the United States alongside judicial waiver, and it has the virtue of treating like charges alike; its cost is that it makes the police officer who frames the charge the effective decision maker on the forum, and it takes no account of the difference between the child who led an offence and the child who was led. The third instrument is the one Parliament enacted: retain the age, retain the system, and interpose an individualised assessment between the charge and the forum, so that the question asked is not how old the child is, which is known, nor what the charge is, which is alleged, but what the child’s capacity and understanding were, which must be found. Its virtue is that it is the only design that even purports to track culpability; its cost is that it relocates the hardest question in the debate from the legislature, where it can be argued in public, to the hearing rooms of Boards, where it is answered child by child on evidence of a kind the science supplies only in general terms.

The migration of the question is the provision’s deepest consequence, and it explains the otherwise puzzling shape of the debate’s afterlife. A legislature that moves a line argues once, in the open, and the argument ends in a number. A legislature that builds an assessment argues once and then hands the argument to thousands of decision makers, each of whom re-enacts it in miniature: every preliminary assessment is the juvenile age debate restaged for an audience of three, with the committee’s evidence and the minister’s rhetoric replaced by a social investigation report and, where the system works as designed, a psychologist’s opinion. The appellate and revisional courts then re-enact it again on the record of the first performance. It follows that the provision’s defenders and critics were, without quite saying so, arguing about different things: the defenders about the design as enacted, which is genuinely individualised and genuinely gated, and the critics about the design as it would be operated by the district machinery that actually exists, whose capacities they had spent years documenting. Both were describing the same provision. The distance between the text and the district, which runs through this entire series as the standard measure of an Indian statute, is in this case not an implementation detail but the substance of the disagreement, because the provision is nothing but a delegation of judgment, and a delegation is worth exactly what the delegate is worth.

There is also a political reason the third design prevailed, and it should be recorded because it recurs wherever legislatures face a demand for a line they are unwilling either to draw or to refuse. An assessment mechanism permits every participant to deny that a blanket rule has been adopted. The government could say, truthfully, that it had not lowered the age, that no child would be tried as an adult on a charge alone, and that each case would turn on its own child; members uneasy about the clause could support it as a safeguard laden compromise rather than as the punitive measure the campaign had demanded; and the opposition, defeated on the clause, could point to the gates and the sentencing floor as the residue of its resistance. The design’s ambiguity was therefore not a defect of the politics but its enabling condition, and the price of the ambiguity was deferred, as deferred prices usually are, to the operators: the Boards who would have to give the abstraction a meaning in individual cases, and the appellate courts who would have to police the meanings given. When a legislature cannot agree on a rule and agrees instead on a discretion, it has not compromised between the two sides of the argument; it has decided that the argument will be continued by other people, in other rooms, without an audience. Much of what has happened to Section 15 since 2016, the narrowing, the discipline, the uneven district record, is the sound of that continuation.

The chronology, stated precisely

The episode is consistently misremembered, usually in the direction of making the law a direct answer to the case, and the correction begins with dates. In December 2012 a gang rape and murder in Delhi, committed by six persons, convulsed the country; one of the six was a few months short of eighteen years of age. The adult accused were tried in a criminal court and sentenced to death. The youngest was dealt with under the Juvenile Justice (Care and Protection of Children) Act, 2000, and received the maximum disposition that Act allowed: three years in a special home. This article states the case only as the legislative trigger it became, and narrates nothing of the offence itself, because the offence’s horror was the argument’s fuel and its details do no analytical work that its legal consequence does not do better. The legal consequence was a single visible disparity: the same act drew the gallows for the adults and a three year ceiling for the participant whose birth date fell months the other side of a line.

The state’s first institutional response was not the Bill. The government constituted a committee of jurists under Justice J. S. Verma, a former Chief Justice of India, to recommend amendments to the criminal law on sexual offences, and that committee reported on 23 January 2013. On the question that would dominate the following three years, the Verma Committee’s answer was unambiguous: the age of juvenility should not be lowered from eighteen. The committee reasoned from the developmental evidence, drawing among other material on the work of the psychologist Laurence Steinberg, that the capacities relevant to culpability continue to develop through adolescence; and Justice Verma made the practical point that a person of sixteen sentenced to a long term would emerge in his middle thirties with little assurance of having been reformed by the intervening years in custody. The wider legislative response of that season went the other way on punishment generally and is recorded in the series’ account of the Criminal Law (Amendment) Act 2013, which raised sentences and created new offences across the sexual offence law; but on the age question, the jurists’ committee had said no, and said it within weeks of the case, before any Bill existed. The Act of 2013 is itself part of the age question’s history, because it was enacted that spring substantially on the Verma Committee’s blueprint, with the age recommendation before the same Parliament that was being pressed to legislate the opposite: Parliament took the Committee’s punishments and left its age line untouched, which was, in substance if not in form, a first legislative disposition of the question, taken within months of the case and on the advice the government of the day had commissioned. The question was therefore not one that waited three years for a legislature willing to face it. It was faced, and answered in the negative, in 2013, and the answer was reversed in 2015 by a differently composed Parliament under a differently constituted government, on the same evidence in every particular except one: the imminence, in the second December, of a release date that the first December’s legislation could not have affected either.

The courts then said the same thing, twice, in the only form open to them. Petitions pressed the Supreme Court to strike down or read down the definition of a juvenile in the 2000 Act, either by holding the uniform age of eighteen unconstitutional or by carving an exception for grave offences. In Salil Bali v. Union of India, decided in July 2013, the Court upheld the age of eighteen as a constitutionally permissible classification, resting on intelligible differentia: the age had been fixed, the Court said, on the understanding of experts in child psychology and behaviour that children in conflict with law could, up to that age, still be redeemed and restored to mainstream society rather than becoming hardened criminals, and the classification drew support from the Convention on the Rights of the Child and the developmental character of childhood. In Subramanian Swamy v. Raju, decided on 28 March 2014, the Court again declined to read the definition down or to direct that a child accused of a grave offence be tried as an adult, treating the fixing of the age as a considered legislative choice and recording the submission, made on the Union’s own behalf, that the psychological, mental, intellectual and emotional maturity of a person below eighteen cannot be objectively determined on an individual or case to case basis. The reckoning rule beneath both cases, that age is counted at the date of the offence and not of the proceeding, had been settled by a Constitution Bench in Pratap Singh v. State of Jharkhand in 2005. The significance of this line of authority for the debate is structural. It established that the change, if it was to come, had to come from Parliament; no court would supply it. The legislature was therefore not ratifying a judicial direction when it acted in 2015. It was reversing, by statute, the outcome that two committees and its own courts had each declined to reach by the routes available to them.

The Bill that became the 2015 Act was introduced in the Lok Sabha on 12 August 2014 by Maneka Gandhi, the Minister of Women and Child Development in the government formed after the general election of that year, and was referred in September 2014 to the Department-related Parliamentary Standing Committee on Human Resource Development. The Bill did considerably more than create the transfer mechanism: it rebuilt the care side of the 2000 Act, made registration of child care institutions mandatory behind a penalty, created a chapter of offences against children, and gave the adoption authority a statutory basis. But its contested core was present from introduction, in clauses that permitted a child of sixteen or above accused of a heinous offence to be tried as an adult after an assessment, and its most aggressive clause went further still: clause 7 of the Bill as introduced would have subjected a person apprehended after completing twenty one years, for a serious or heinous offence committed between sixteen and eighteen, to trial as an adult, a retrospective reach that the Committee singled out for particular criticism. The sequence from here is the heart of this article, and it is worth holding in its exact order: committee report in February 2015 against the provision; official amendments in the spring of 2015 that accepted much of the report and kept the provision; passage by the Lok Sabha on 7 May 2015; a long pause in the Rajya Sabha; the release of the young person in the 2012 case on 20 December 2015, on completion of the three year term, after the Delhi High Court declined in the days before the release to intervene to extend his custody beyond what the 2000 Act allowed; passage by the Rajya Sabha on 22 December 2015, two days after the release; assent on 31 December 2015; and commencement on 15 January 2016, the statute numbered, in the Gazette’s order, Act 2 of 2016 notwithstanding the year in its title.

The committee that said no

The examination of the Bill by the Standing Committee on Human Resource Development is the best documented stage of the whole episode, and the least remembered. The Committee, chaired by Dr Satyanarayan Jatiya, a member of the ruling party, took evidence from the Ministry, from child rights organisations and from academic centres, among them the Centre for Child and the Law at the National Law School of India University, whose researchers Swagata Raha and Arlene Manoharan appeared before it to argue that the very concept of a heinous offence had no footing in a statute built for reformation and that a Board constituted to adjudicate from the child’s standpoint was being asked to perform the opposite function of expelling the child from the system. The Committee’s 264th Report was presented to Parliament on 25 February 2015, and on the transfer provision it was not equivocal. It recommended against reducing the age of juvenility and against creating a mechanism for the adult trial of children; it took the view that such a mechanism would place India in breach of its obligations under the Convention on the Rights of the Child; it found that the crime data relied on by the Ministry did not support the premise of a surge in offending by children; and it concluded that the existing framework should instead be properly implemented. Its words on the data deserve to be carried in the record rather than summarised away: an objective analysis of the data of the National Crime Records Bureau placed before it made it abundantly clear, the Committee said, that the percentage of juvenile crimes in India, 1.2 per cent, was quite low, that some incidents of juvenile crime, though a cause of serious concern, should not be the basis for introducing drastic changes in the existing juvenile justice system, and that the existing system was not only reformative and rehabilitative in nature but recognised that sixteen to eighteen years is an extremely sensitive and critical age requiring greater protection. A reader who wants the general account of what these committees are for, and what their reports can and cannot do, will find it in the series’ treatment of parliamentary committees and bill scrutiny; this article is concerned with the narrower and stranger fact that on this Bill the scrutiny happened, was thorough, reached a clear conclusion on the Bill’s central provision, and was published in time to matter.

The evidence the Committee gathered is itself part of the record, because it shows the objection was not a faction’s. The submissions and testimony ran, with near uniformity, against the transfer clauses: child rights organisations and women’s and human rights groups, academics working on the juvenile system, lawyers practising in it, and the children’s law centre at the National Law School all advised the Committee that the mechanism was misconceived, unconstitutional in their submission, or unworkable, while the Ministry’s defence rested on the Bureau’s figures and on the deterrence claim examined below. A committee chaired by a member of the governing party, hearing that balance of evidence, might still have split on party lines; the 264th Report did not treat the provision as a close question. It held that the proposed change lacked a rational basis in the data, that it cut against the Convention and the reformative character of the system, and that the state’s first duty was to the framework it had already enacted. When the report is described in the debate’s later retellings at all, it is usually compressed into the statement that a committee had reservations. The compression is a disservice in both directions: it understates the report for the provision’s critics, who are entitled to the full strength of what a governing party committee concluded on evidence, and it understates it for the provision’s defenders too, since a recommendation that sweeping, from that source, is precisely the kind a government should have to answer with reasons rather than absorb as routine dissent.

The government’s response to the report is the episode’s most instructive document, because it was neither acceptance nor defiance but a division of the two. In April 2015 the Cabinet approved official amendments to the Bill founded on the Committee’s recommendations, and most of the report was in fact accepted across the Bill’s many subjects. On the transfer provision the acceptance was real but peripheral: clause 7, the retrospective clause the Committee had attacked as the Bill’s most damaging feature, was removed, so that no person apprehended as an adult could be tried as an adult for an offence committed as a child; the period for the preliminary assessment was enlarged; and the surrounding safeguards were strengthened. The core of the provision, the assessment and transfer of a child of sixteen or above accused of a heinous offence, was retained without qualification. The pattern is worth naming precisely, because it recurs in Indian legislation and is rarely described accurately: the committee process operated as a genuine filter on the Bill’s drafting and as no filter at all on its policy. Every objection that could be answered by amending a clause was answered; the one objection that went to the provision’s existence was heard, recorded, and overridden. Whether that is a failure of the committee system or its correct functioning is the question the closing sections of this article take up, and it cannot be answered until both sides of the substantive argument have been heard in their strongest form.

What did the government accept from the Standing Committee’s report?

More than the debate usually allows, and less than the report asked. The retrospective clause that would have exposed a person apprehended after twenty one to adult trial for an offence committed as a child was dropped, the assessment period was enlarged and safeguards were strengthened; the transfer provision itself, the report’s central target, was retained entire.

The Lok Sabha stage: acceptance in form, retention in substance

The Bill’s passage through the Lok Sabha in the spring of 2015 is the quiet middle of the episode, and its quietness is evidence of something. The official amendments the Cabinet approved in April were moved when the House took the Bill up for consideration, and they did the work described above: the retrospective clause went, the assessment period was enlarged, safeguards around the mechanism were strengthened, and the Committee’s recommendations were given effect across the Bill’s adoption, institutional and procedural provisions. With those amendments carried, the House passed the Bill on 7 May 2015. The debate in the lower House rehearsed the two cases that the rest of this article sets out, the government resting on the gravity of the triggering case, the deterrence claim and the narrowness of the mechanism as amended, and members opposed to the clause resting on the Committee’s report, which had the unusual authority in that debate of being the work of a committee on which the governing side held the chair and a majority of the membership. What the Lok Sabha stage did not produce was any reconsideration of the clause’s existence: the amendments were the government’s, they presupposed the mechanism, and the question the House answered was how the mechanism should be designed rather than whether it should exist. That ordering, design before existence, is the standard technique by which a determined ministry carries a contested provision through a legislature, because once the debate is about periods, safeguards and the wording of an assessment, the provision has in substance been conceded. The Committee had asked the existence question and answered it no; the House, by taking up the design questions first and only, answered it yes without ever quite putting it. The Bill then went to the Rajya Sabha, where it would wait seven months for the week that decided it, and the wait itself became part of the argument, the government accused of lacking the will to move its own Bill and the opposition accused of lacking the will to let it move, until the release supplied both with a reason to stop waiting.

The week of the release

The Rajya Sabha’s seven month pause ended in a week whose sequence no participant chose and every participant understood. The Bill had been passed by the Lok Sabha in May 2015 and had then waited in the Upper House through the monsoon and winter sessions, listed and not reached, while the young person in the 2012 case served out the final months of the three year term in a special home in Delhi. As the term’s end approached in December 2015, the case returned to the courts in the only form left to it: Subramanian Swamy, whose earlier challenge to the age itself had failed, sought to have the young man’s release deferred, and the Delhi High Court declined in the days before the release to intervene to keep him in custody beyond the term the 2000 Act allowed, the statute containing no power to extend a disposition that had run its course. On Sunday 20 December 2015 he was released, amid protests in the capital by the victim’s parents and others, and the following two days converted the release into a legislative timetable. The victim’s mother met the Minister of State for Parliamentary Affairs, Mukhtar Abbas Naqvi, on the Tuesday morning to press for the Bill’s passage that day; members across parties agreed that the Bill should be taken up at once; and the government, which had been accused through the autumn of allowing the Bill to drift, brought it forward for consideration that afternoon. The debate that followed lasted the better part of five hours and was, by the standards of the episode, a real one: the minister defended the assessment model, the opposition’s leader supported the Bill while reminding the House of the sixteen year line in the Congress government’s own Act of 1986, members of the Nationalist Congress Party and the Dravida Munnetra Kazhagam questioned the hurry and proposed reference to a Select Committee, Brinda Karat of the Communist Party of India (Marxist) placed on record the point that even immediate passage could not touch the case that had produced the week, since the new law would not apply retrospectively, and Sitaram Yechury pressed the Select Committee demand to the point of a walkout by the Left parties when the Chair ruled that no motion for reference was before him. The clauses were then taken up, the amendments moved against the Bill were negatived, and the Bill was passed by voice vote. Assent followed on 31 December 2015, the last day of the year, and the Act commenced on 15 January 2016.

Two features of that week should be preserved against the smoother retellings. The first is that the Bill’s passage changed nothing for the person whose release occasioned it, and everyone in the chamber knew it: the point was made on the floor by the provision’s opponents and was not contradicted by its defenders, whose case was expressly prospective, that the law would govern the next case and not this one. The second is that the timetable inverted the ordinary relationship between scrutiny and passage. The Bill that had received the fullest committee examination of any in its cluster was, at the last, passed in a single sitting convened by an event outside the legislature’s control, with the demand for a second scrutiny refused not on its merits but on a point of procedure. A legislature is entitled to legislate in response to events; that is a large part of its function. What the week of the release shows is the cost structure of doing so, in which the event that supplies the will to legislate simultaneously withdraws the time to legislate carefully, and the institution that had already done the careful work, the Standing Committee, found its work simultaneously vindicated as the debate’s evidence base and set aside as its conclusion.

The case for the provision, at full strength

The strongest form of the case for the transfer provision is not the cry for vengeance that its opponents usually answer; it is a set of propositions about proportionality, individualisation and democratic authority, and it deserves to be met in that form. Begin with proportionality. A criminal law that cannot distinguish, in its consequences, between a theft by a child of thirteen and a planned killing by a person of seventeen years and eleven months has not abolished proportionality; it has abolished the upper half of it. The three year ceiling of the 2000 Act was a single answer to every grade of wrongdoing, and its defenders were obliged to argue that the gravity of the act is simply irrelevant to the state’s response once the actor is below eighteen. Maneka Gandhi, who carried the Bill through both Houses, put the ministry’s position in terms of the system’s own incentive structure: the existing law, she argued, was itself an encouragement, because older children who knew the Act understood that its ceiling protected them whatever they did, and she cited the police view she had been given that a very large share of crime was committed by sixteen year olds who knew the Juvenile Justice Act and counted on its shelter. In the Rajya Sabha debate she framed the mechanism as a discrimination between states of mind: the Board, assisted by psychologists and experts, would decide whether the act had been committed in a childlike frame of mind or in an adult frame of mind, and no child would be sent directly to a jail, a transferred child being held apart from adult prisoners until twenty one and reviewed at that age. Whatever the fate of the empirical claim inside that argument, its structure is not foolish: it says that knowledge of consequences is the thing criminal liability has always turned on, and that a statute which refuses to inquire into it, in the cases where it matters most, protects the calculating and the impulsive indifferently.

The second proposition is about the victim, and it was made most forcefully not by ministers but by the family of the young woman killed in December 2012, whose parents campaigned publicly for the change through the whole passage period and were present in the visitors’ gallery of the Rajya Sabha when the Bill was taken up. Their argument, repeated by the mother in the week of the release, was that had the law been passed earlier the youngest of the accused would not have walked free at the end of three years; and behind the particular claim sat a general one the juvenile system had never had to answer, which is that a process designed entirely around the needs of the child who offended has no place in it for the child or the family who were offended against. The victim of a grave offence committed by a seventeen year old is, in a large proportion of such cases, herself a child or a young woman of the same neighbourhood and the same poverty, and a system that speaks exclusively of the offender’s best interests tells her that her interests are not a statutory value. Parliament, on this view, was not surrendering to a mob; it was correcting an asymmetry that the reformative model had built into itself and had never defended on its merits, and the correction was deliberately narrow, a single grade of offence, a single age band of two years, two successive decisions, and the retention of every protection that prevents the sentence from becoming the adult sentence in its full severity.

A further proposition, less often articulated but doing quiet work throughout the government’s case, concerns public confidence in the juvenile system itself. A welfare jurisdiction survives politically on a fund of public tolerance, and tolerance is spent by its hardest cases. The argument ran that a system which visibly cannot respond proportionately to the worst act a near adult can commit will not be permitted to continue responding gently to the ordinary run of acts either; that the three year ceiling, defended as a principle, would in time bring the whole reformative edifice into the disrepute that a single notorious case can generate; and that a narrow, gated exception was the price of preserving the system for the great majority of children who would never come near it. On this view the provision is not a breach in the wall of the 2000 settlement but a safety valve installed to keep the wall standing, and its narrowness is not an embarrassment to be explained away but the substance of the claim: Parliament took the smallest exception that would answer the confidence problem, hedged it with two decisions and a sentencing floor, and left the settlement otherwise intact. Whether the valve was needed is a judgment about politics rather than law, and the critics’ answer, that the system’s legitimacy problem in the districts was starvation rather than scandal, goes to the implementation ground covered elsewhere in this article; but the argument deserves its place in the strongest form of the case, because it is the one that explains why legislators who accepted much of the critics’ evidence could still vote as they did.

Running through the first two propositions is an equality argument that the provision’s defenders pressed and its critics never quite answered on its own terms. The disparity of December 2012 was not only between offence and consequence; it was between participants in a single act, separated by a few months of age, one of whom faced the gallows and the other a three year ceiling, the line between them being nothing about their conduct, their roles or their understanding, which the evidence at the trial did not suggest differed in kind, but the accident of their birth dates relative to a date neither of them had chosen with the act in mind. A bright line always produces such pairs at its margin, and the existence of a hard case at a boundary is not ordinarily an argument against having the boundary, as the critics rightly said of speed limits and voting ages. The defenders’ refinement was that this boundary was different in kind from those: it did not regulate a licence or a franchise but distributed the state’s most serious response to the most serious acts, and it did so by a criterion, chronological age, that the surrounding law itself treated elsewhere as evidence of capacity rather than as capacity itself. The assessment, on this argument, is not an exception to the principle of equality but an attempt to honour it more exactly, by asking of the marginal pair the question their shared act actually poses: not how many months separate their birthdays, but whether the younger of the two understood what he was doing as the elder did. That the question may be hard to answer reliably is the critics’ ground, and it is strong ground; but it is ground about the instrument, and the equality argument survives it as a statement of what the instrument is for.

The third proposition is the democratic one, and it is the one that survives even if the first two are doubted. The age of criminal responsibility, and the boundary of childhood for penal purposes, are not facts of nature that experts discover and legislatures register; they are choices, made differently by different democracies that share the same developmental science. England and Wales set the age of criminal responsibility at ten and try a child charged with murder in the Crown Court. The United States, which has not ratified the Convention, permits the transfer of children to adult courts in every state by one mechanism or another, and its Supreme Court’s interventions have been directed at the sentence, barring the death penalty for offences committed below eighteen in Roper v. Simmons in 2005, rather than at transfer itself. Germany applies its juvenile law to young adults up to twenty one where their maturity warrants it, which is the mirror image of the Indian mechanism, a discretion to treat an adult as a child rather than a child as an adult, and rests on the same premise: that chronological age is a proxy for maturity, and a legal system may in defined classes substitute a judgment about the person for the proxy. In the Rajya Sabha debate the Leader of the Opposition, Ghulam Nabi Azad, supporting the Bill his own party had doubts about, reminded the House that the Juvenile Justice Act of 1986, enacted under a Congress government, had set the age at sixteen for boys, and that the uniform eighteen was itself a legislative choice of the year 2000, taken in the course of conforming Indian law to the Convention; what one Parliament chose, another Parliament could revisit. And the individualisation argument completes the case. A blanket reduction of the age to sixteen, which is what the public demand after December 2012 principally asked for and what the Verma Committee rejected, would have swept every child of sixteen into the adult system for every offence; the assessment model refuses both blanket rules and asks, of the one child, the only question that could justify the exception: did this child have the capacity to commit the offence and the ability to understand its consequences, and in what circumstances was it committed? A system confident in its own rehabilitative premises, the argument runs, should not fear that question, provided the asking of it is disciplined, expert assisted and repeated by a court; and the safeguards of Section 21, which no transferred child can lose, mark the point at which even a convicted child’s claim on the state’s protection revives.

The case against the provision, at full strength

The strongest form of the case against is likewise not the sentimental one. It is an argument about evidence, about the limits of a particular kind of determination, and about who will in practice stand on the receiving end of a discretion, and each of its strands was placed on the parliamentary record by a named holder before the provision was enacted. The evidence strand belongs first to the two bodies that examined the question with access to the official data. The Verma Committee, which had every political reason to give the public mood what it asked and refused on the age question, reasoned that the developmental science runs the other way: the capacities of judgment, impulse control and foresight that criminal responsibility presupposes are precisely the capacities still forming at sixteen and seventeen, so that the class of cases in which a child most resembles an adult in the act is the class in which he least resembles one in the mind that produced it. The Standing Committee, examining the Ministry’s own figures, found the factual premise of the Bill wanting: the Bureau’s data showed juvenile crime at 1.2 per cent of recorded crime, the share having moved from 1 per cent in 2003 to 1.2 per cent in 2013, and a rising share of older children within a small and roughly stable total is an artefact of the denominator as readily as it is a wave of offending. The Committee also identified the classification trap inside the figures most often quoted in the debate, the rise in recorded sexual offences by children: the Protection of Children from Sexual Offences Act, 2012 had raised the age of consent from sixteen to eighteen, so that sexual activity which had previously been lawful when consensual was reclassified as rape, and a measurable part of the surge in the statistics was the statute book indicting more conduct, not children committing more of it. A legislature that changes a definition and then cites the enlarged count as the reason for a second change is arguing in a circle, and the Committee said so in substance.

The second strand concerns the assessment itself, and it is the strand that has strengthened since enactment. Section 15 asks a Board to determine a child’s mental capacity to commit the offence and ability to understand its consequences. The objection, made by the Centre for Child and the Law’s researchers before the Committee and by the HAQ Centre for Child Rights throughout the debate, its co-director Enakshi Ganguly arguing the point across the Act’s life, is that no instrument exists by which that determination can be made reliably for an individual child: developmental science speaks in population generalisations about adolescents as a class, and the movement from the generalisation to a finding about this child’s mind, made retrospectively about a state of mind at the moment of an act, is not measurement but impression dressed in clinical vocabulary. The objection had, unusually, been conceded in advance by the state’s own side in litigation: in the Subramanian Swamy proceedings the Union’s counsel submitted that the maturity of a person below eighteen cannot be objectively determined on an individual or case to case basis, and the Court’s refusal to read down the age rested in part on the wisdom of a uniform rule over individualised guesswork. Parliament in 2015 nonetheless built the exception on exactly the determination its own government had described as unavailable. Add the composition of the deciding body, a magistrate sitting with two social workers, sitting in districts where the expert assistance Section 15 makes available is often absent in fact, and the assessment risks becoming, in the critics’ phrase, a conclusion in search of a vocabulary: the gravity of the charge supplies the answer, and the capacity language supplies the reasoning.

A third strand, on deterrence, drew on the one jurisdiction with long experience of transfer. The United States has operated waiver and exclusion mechanisms for decades, and its evidence was put before Indian audiences by the provision’s critics throughout the debate, most systematically the finding of the review conducted for the Centers for Disease Control and Prevention’s Task Force on Community Preventive Services, published in 2007, which examined the evaluated studies of transfer laws and concluded that juveniles transferred to the adult system were subsequently rearrested for violent offences at higher rates than comparable juveniles retained in the juvenile system, so that transfer operated in the American record as a generator of the offending it was meant to suppress. The Indian government was not obliged to treat American evidence as decisive for Indian conditions, and its defenders were entitled to answer that the Indian mechanism, with its double gate and its place of safety, differed from the prosecutorial direct file systems that produced much of the American data. But the burden the evidence imposed was real: the only mature body of outcome research on trying children as adults pointed against the deterrent claim, and the passage debate contains no engagement with it at all, the deterrence proposition being carried throughout as an intuition about what a rational sixteen year old would fear, in a debate whose other side consisted precisely of evidence that sixteen year olds do not calculate in that way. An intuition about adolescent rationality, deployed against the developmental evidence about adolescent rationality, is the debate in miniature.

The fourth strand is distributive, and it is the one the debate engaged with least. The children who come before Juvenile Justice Boards are drawn, in the documented record of the system’s own inspections and of the organisations that work within it, overwhelmingly from the poorest households: children of daily wage labour, children out of school, children from the bastis and the railway platforms, children whose families cannot brief counsel, cannot commission a private psychological evaluation, and cannot carry an assessment order into appeal. A discretion structured by impressions of maturity, remorse and circumstance will not be exercised evenly across that population, because the indicia by which maturity is read in a hearing room, fluency, demeanour, dress, the presence of an engaged parent, the ability to narrate one’s own life coherently, are themselves distributed by class. UNICEF, in its submissions on the legislation, made the institutional version of the same point from outside the country: the reform the Indian system needed was the implementation of the framework it had, Boards constituted, members trained, institutions registered and inspected, and a state that had not built the floor of the system was not entitled to conclude that the system’s ceiling was the problem. In the Rajya Sabha, Sitaram Yechury of the Communist Party of India (Marxist) made the process version of the objection and the prophetic one in a single intervention: the debate, he said, was being guided by sentiment, the Bill should go to a Select Committee of the House for the consideration its subject merited, and a House that moved the line to sixteen because of a case at seventeen would be asked, on the next case at fifteen, why the line should not move again. The notices seeking reference to a Select Committee were never formally moved before the Chair, the Deputy Chairman P. J. Kurien ruled that no such motion was before him, the Left parties walked out, and the Bill was passed by voice vote after its clauses were taken up and the amendments moved against it were negatived. The provision’s opponents therefore lost twice in the same week: on the substance, and on the attempt to buy the one thing the committee system exists to supply, which is time.

The science both sides cited

One body of research did more work in this debate than any other, and it armed both sides, which is usually a sign that the research was being read at the level of its conclusions rather than its distinctions. The developmental psychology of adolescence, associated in the debate principally with the work of Laurence Steinberg, distinguishes two maturing systems. The first is basic cognitive capacity: the ability to reason, to understand information, and to grasp the consequences of an act in the abstract, which on Steinberg’s account reaches broadly adult levels by about the age of sixteen. The second is psychosocial maturity: impulse control, resistance to peer influence, orientation towards the future and the tempering of sensation seeking, which continues to develop into the middle twenties and is precisely the set of capacities engaged when an adolescent acts in a charged situation among companions. The Verma Committee relied on this literature, and on the second system, to hold that a person of sixteen or seventeen is not situated as an adult is situated when he offends, and that the law’s uniform line at eighteen tracks the science of the capacities that culpability actually engages. The counter-critique, developed at length in the India Law Journal’s examination of the Committee’s reasoning, answered that the reliance was selective: the same body of work puts the first system, the capacity to understand what one is doing and what follows from it, at adult level by sixteen, and Section 15 of the enacted statute does not ask about psychosocial maturity in the round. It asks about the child’s mental and physical capacity to commit the offence and ability to understand its consequences, which is the first system’s territory, and on the literature’s own terms that inquiry is not the scientific solecism the provision’s critics described. A Board asking whether this child understood what killing means is asking a question that developmental science can inform; a Board asked whether this child’s psychosocial maturity matched an adult’s would be asking one it cannot answer, and the statute, whether by design or by fortune, asked the first question and not the second.

The honest residue is narrower than either side’s rhetoric. The science supports two propositions simultaneously and cannot choose between the policies built on them: that by sixteen the capacity to understand consequences is ordinarily present, and that the capacities governing behaviour in the situations in which grave offences by adolescents typically occur remain immature for years afterwards. Which of those propositions the law should privilege is not a scientific question at all, and the debate’s recurring attempt to make it one, by committees and ministers alike, was a way of borrowing authority for a value judgment. The courts understood this earliest. Salil Bali used the expert understanding to sustain the line the legislature had drawn, and the Subramanian Swamy proceedings recorded the concession that individual maturity cannot be objectively determined case by case; but neither decision held that the science compelled the line, and the distinction mattered in 2015, because it left Parliament free to draw the line elsewhere without defying any judicial finding of fact. What the science did supply, and what the enacted provision has had to live with, is a warning about its own instrument: an assessment regime whose questions track the cognitive system will produce defensible findings, and one that drifts, under the pressure of a grave charge, into judgments about character, remorse and dangerousness will produce the impressionistic findings the critics predicted and the appellate courts have since had to correct. The distance between those two assessments is the distance the Supreme Court travelled in Barun Chandra Thakur when it insisted that the exercise is delicate, evidence based and ordinarily expert assisted.

The factual dispute: what the crime data could and could not show

Underneath the argument about values sat an argument about numbers, and it is the part of the debate most amenable to settlement, because both sides were reading the same publication. The National Crime Records Bureau’s annual Crime in India reports are the only national series on recorded offending by children, and the Ministry’s case for the Bill rested on them. The figures the government placed before Parliament, as the Standing Committee recorded and as PRS Legislative Research’s analysis of the Bill set out, were these: the share of juvenile crime in total recorded crime rose from 1 per cent in 2003 to 1.2 per cent in 2013, and within the population of juveniles accused of crimes, the share of those aged sixteen to eighteen rose from 54 per cent to 66 per cent across the same decade. On those two movements the claim of a surge was built, and Maneka Gandhi’s parliamentary statements added a third and larger claim, attributed by her to the police, that half of all crimes were committed by sixteen year olds who knew the protection the juvenile law gave them. The three claims are of different orders, and the debate went wrong in part by treating them as one. The third claim was never supported by any published series and is irreconcilable with the first, since a group responsible for half of all crime cannot simultaneously account for 1.2 per cent of it; it survives only as a record of what was said in the argument, not as a finding within it. The first claim is true and small: a movement of two tenths of a percentage point across a decade, within a series the Bureau itself compiles from first information reports rather than from convictions, so that it counts accusations and not findings. The second claim is true and ambiguous: a rising older share of a stable total is exactly what one would expect if reporting improved unevenly, if the younger end of the system diverted more children out of it, or if the age composition of the child population itself shifted; it is also what one would expect if older children really were offending more. The data, read with the denominator discipline the Committee applied, establish that recorded offending by children was a small and roughly stable share of recorded crime across the decade in which the demand for the provision arose; they do not establish the proposition for which they were cited, and they do not refute it either, because a series built on reports to police stations cannot carry the weight of a deterrence argument in either direction.

The composition figures cut the same way and were also before the Committee. Of all juveniles apprehended for offences under the Indian Penal Code in 2013, those aged sixteen to eighteen who were apprehended for murder formed 2.17 per cent, and those apprehended for rape 3.5 per cent; and set against the whole of the country’s recorded offending in the same year, the age group accounted for 1.3 per cent of all persons arrested for murder and 3.29 per cent of all persons arrested for rape. A reader can make those figures serve either side, and both sides did: for the Ministry, the point was that the older band dominates the juvenile system’s grave end and that its share of the whole was rising; for the Committee, the point was that a group forming barely one in thirty of those arrested for rape, and one in seventy five of those arrested for murder, was being legislated for as though it were the centre of the country’s violent crime, and that the typical child before a Board was there for theft, burglary or hurt, offences for which the transfer debate was an irrelevance conducted over his head. The figures themselves decide nothing between those readings, because the readings differ over what a small number means: a small number can be a tail risk a legislature is entitled to answer, or a marginal phenomenon a legislature is wrong to reorganise a system around, and statistics cannot say which. What they can do, and what the Committee used them to do, is discipline the adjectives: whatever was happening to recorded offending by older children between 2003 and 2013, the published series does not describe it as large, as fast growing, or as concentrated in the offences the provision targets, and every claim in the debate that implied otherwise was a claim the record would not carry.

Two further cautions belong to any honest use of the numbers, and both were on the record before the vote. The first is the classification effect the Committee identified: the Protection of Children from Sexual Offences Act, 2012 raised the age of consent from sixteen to eighteen, so conduct between adolescents that had been lawful became rape in the statistics, and the same Parliament that was shown a rise in sexual offences by children had itself enlarged the category two years earlier. Any trend drawn across a redefinition is measuring the redefinition until it is proved otherwise. The second is the gap between apprehension and adjudication: the Bureau’s juvenile tables count children apprehended, and a substantial proportion of apprehended children are never found to have committed the offence alleged, so the series measures the system’s intake and not its conclusions. A reader who teaches this debate, or who has to present its evidence to a classroom or a committee, works from these same published tables, and a reader assembling the two rival readings of one table into a single deck can open the presentation directly in the ReportMedic PPT viewer rather than depending on the software in which the file was made. What the episode teaches about data in legislation is not that the Ministry’s figures were fabricated; they were not. It is that a true figure, cited without its base, its source instrument and its reclassifications, functions in a debate as a false one, and that the one body charged with reading the figures slowly did so and was not believed.

Five decision points

The controversy resolves, on analysis, into five distinct decisions, each of which could have gone the other way, and the table below is this article’s map of them. It matters to separate them, because public argument treats the provision as a single act of will, usually attributed to public anger, whereas the record shows a sequence of choices made by identifiable actors at identifiable moments, some of which commanded agreement across the debate and some of which were contested to the end.

Decision point	What was decided	The alternative	Who argued for the alternative	What it has produced
Legislate or implement	A new Act with a transfer mechanism was enacted rather than the 2000 Act being resourced and enforced as it stood	Proper implementation of the existing framework: Boards constituted and trained, institutions registered, timelines enforced	The Verma Committee; the Standing Committee; UNICEF	A mechanism whose working depends on the same Boards and experts whose absence was the old system’s documented failure
Proceed or accept the report	The Bill was enacted with its central provision intact after the examining committee recommended against it	Deletion of the transfer clauses and retention of the uniform age of eighteen for every offence grade	The Standing Committee on Human Resource Development, in its 264th Report	The clearest modern instance of committee scrutiny obtained, considered at Cabinet level, and set aside on the Bill’s central question
The threshold	Sixteen was fixed as the age from which transfer is possible	Retention of eighteen as a uniform line, or a lower threshold of fourteen as some public demand proposed	The Standing Committee and the courts for eighteen; voices in the public campaign for a lower line	A two year band, from sixteen to eighteen, inside which the forum depends on an assessment rather than on the birth date alone
The route	Transfer requires two successive decisions, a Board assessment followed by an independent Children’s Court decision	A single decision by the Board, or a prosecutorial power to file directly in an adult court of the kind used in parts of the United States	Defenders of the enacted route against both simplifications; critics of the Bill against the route itself	A double gate that either body can close, with the assessment disciplined on appeal and by the Supreme Court’s later insistence on expert assistance
The sentencing floor	The bars on death and on life imprisonment without the possibility of release were retained for transferred children	Full adult sentencing, including capital punishment, as the angriest public demand of 2013 proposed	No parliamentary party carried the full adult sentencing case into the enacted text	A ceiling of life imprisonment with the possibility of release, custody in a place of safety until twenty one, and a review at twenty one between release on conditions and jail

Read as a set, the five points show why both the provision’s triumph and its defeat have been overstated. The legislature did not enact the public demand of 2013, which was for the age itself to fall and, in its rawest form, for the youngest accused in the Delhi case to hang; it enacted something considerably narrower, from which the retrospective clause was removed under committee pressure, in which the sentencing floor held, and in which the decision was doubled rather than concentrated. Neither did the legislature merely register a considered judgment on fresh evidence; it acted against the recommendation of the body it had charged with the evidence, on a timetable set by a release date it did not control, in a debate its own members described as driven by sentiment. The five points are therefore the correct unit of judgment. A critic who cannot say which of the five he would reverse has not yet located his objection; a defender who cannot say which of the five he would defend on evidence rather than on authority has not yet located his argument. The comparison of what the two statutes, the Act of 2000 and the Act of 2015, actually do across the whole of their operation belongs to the comparison of the Juvenile Justice Acts of 2000 and 2015; what belongs here is the narrower observation that of the five decisions, the second is the one with no parallel in the statute’s own history, and it is the one to which the series’ wider argument about legislative process returns below.

The third and fourth decisions repay a further paragraph each, because they interact. The choice of sixteen, rather than the fifteen or fourteen that the logic of the triggering case would equally have supported, was defended in the debate as a line with a history: the Juvenile Justice Act of 1986 had drawn the line at sixteen for boys, and the pre-2000 state children Acts had commonly done the same, so that sixteen presented itself to the House not as an invention but as a restoration. The argument is weaker than it sounds, since the 1986 line was one half of a sex differentiated scheme that no one proposed to restore, and the uniform eighteen of 2000 had been chosen deliberately, in the Convention’s wake, to replace it; but it performed the essential parliamentary function of making the number feel inherited rather than improvised. The two body route, by contrast, was a genuine legislative construction with no inheritance behind it, and it is the feature of the provision that has most shaped its life. It converted what could have been a sentencing rule into a jurisdictional inquiry, split the inquiry between a welfare bench and a court, and thereby created the procedural surface on which the Supreme Court would later work: an assessment that can be done well or badly, by experts or without them, is an assessment a higher court can discipline, and the discipline duly arrived. Whether the doubling was designed as a safeguard or accumulated as a compromise, its effect has been to make the provision narrower in operation than in text, which is the opposite of the usual direction of Indian criminal legislation and a large part of why the provision’s predicted consequences, in both directions, have been slow to appear.

The common ground nobody used

The strangest feature of the debate, visible only when the whole record is laid out, is how much of it was agreed. No participant of weight defended the working of the 2000 Act as it stood in the districts. The Verma Committee’s report, while refusing the age reduction, documented the failures of the reformative system the age belonged to: institutions that did not reform, homes that did not educate, a machinery of care whose weakness it treated as the state’s prior obligation. The Standing Committee’s conclusion was that the existing framework should be properly implemented, which presupposes that it was not being implemented. The Ministry’s own Bill was, in four fifths of its length, an implementation statute: registration of institutions behind a penalty, timelines inserted into inquiries, a chapter of offences against children, a statutory adoption authority, monitoring machinery. Those parts of the Bill attracted no organised opposition and little debate, and they passed as the considered product of the same committee process that the transfer clause survived. The debate therefore contained, inside a single Bill, both the consensus remedy and the contested exception, and the parliamentary time, the public attention and the institutional conflict of the entire episode were consumed by the exception. The consensus remedy received the worse fate of the two in the years that followed, not because anyone opposed it but because no event ever again supplied it with a week of its own: the registration duty, the timelines and the committee machinery of the Act depend on the unglamorous appropriations and appointments that the implementation reviews have continued to find wanting, in the same words, under this Act as under its predecessor.

This distribution of attention is worth recording because it corrects the standard moral drawn from the episode, which is that the legislature chose punishment over care. The legislature chose both, in one instrument, on one vote; what it failed to choose, because the failure lies outside any single vote, was the administrative follow through that the care half required and the exception half presupposed. The critics who said implement first were not answered on the point and were not wrong on the point; they were outvoted on a different point, the exception, by a majority that simultaneously enacted their implementation programme and then declined to fund its difference. If the episode has a lesson for the reading of Indian legislation generally, it is this: the contest over a statute’s contested clause is a poor predictor of the statute’s life, because both sides of the contest share an interest in the clause, the defenders to enact it and the critics to oppose it, while the statute’s uncontested machinery has no constituency in the chamber at all and must live on what the administration gives it. The juvenile age debate was fought, on all sides, as though the future of Indian childhood turned on Section 15. The record since suggests it turned at least as much on Sections 27, 41 and 106, the constitution of Committees, the registration of institutions and the district child protection units, which no crowd ever gathered to demand and no committee report ever made famous.

The international obligation question

The argument about the Convention on the Rights of the Child was the debate’s most legal argument, and it is frequently misstated as a slogan by both camps, so its components need separating. India acceded to the Convention on 11 December 1992. Article 1 defines a child as every human being below the age of eighteen years unless majority is attained earlier under the applicable law. Article 3 makes the best interests of the child a primary consideration in all actions concerning children. Article 37 forbids capital punishment and life imprisonment without the possibility of release for offences committed by persons below eighteen, a prohibition Section 21 of the 2015 Act reproduces almost exactly. Article 40, the provision the brief for this debate puts at the centre, requires that every child alleged as or accused of having infringed the penal law be treated in a manner consistent with the promotion of the child’s sense of dignity and worth, which reinforces the child’s respect for the human rights of others and takes into account the child’s age and the desirability of promoting the child’s reintegration; and it requires States to promote the establishment of laws, procedures, authorities and institutions specifically applicable to such children. The Committee on the Rights of the Child, the treaty body that monitors the Convention, addressed the transfer question directly in its General Comment No. 10 of 2007 on children’s rights in juvenile justice, recommending that States not withdraw children of sixteen or seventeen from the jurisdiction of the juvenile justice system, and it went further in General Comment No. 24 of 2019, which replaced the earlier comment, urging States parties to remove provisions under which children are treated as adults and to apply the child justice system to every person below the age of eighteen.

The treaty body’s position, it should be added, was not formed in reaction to the Indian Bill; it long predated it. In its concluding observations on India’s initial report under the Convention, adopted in 2000, the Committee on the Rights of the Child had already expressed concern at the possibility, which then existed under some of the older state children legislation, of trying boys between sixteen and eighteen as adults, and the uniform age of eighteen in the Act of 2000 was in part India’s answer to that concern. The sequence therefore has a symmetry the parliamentary debate rarely acknowledged: the line at eighteen was drawn where it was partly because the treaty body objected to the older patchwork beneath it; the treaty body objected again when the exception was proposed in 2014; and the objection in 2014 was not a new standard invented for the occasion but the same standard India had once legislated to meet. A State is of course free to change its mind about the weight it gives a treaty body’s reading, and States do; but the change is more candidly described as a decision to accept a known and documented divergence than as the discovery that no divergence existed.

On that record, the case that the provision breaches India’s obligations is straightforward to state and was stated, in almost these terms, by the Standing Committee and by UNICEF: the Convention defines the protected class by age alone; its monitoring body has twice addressed the precise device India adopted and advised against it; the Act of 2000 was itself enacted to bring Indian law into conformity with the Convention, so the 2015 exception is a deliberate step away from a conformity Parliament had previously claimed; and the dignity and reintegration language of Article 40 is not honoured by a process whose outcome may be a life sentence and a permanent record. The contrary case is narrower than the provision’s defenders in the street supposed, but it is not empty, and the Ministry maintained it through the passage: the Convention nowhere in terms forbids the trial of a child in an adult forum; its hard prohibitions are those of Article 37, which the Act observes to the letter; Article 40 requires a distinct system for children, and the Indian system remains distinct, since the child enters it at a Board, is transferred only after an individual assessment and a court’s independent decision, is held apart from adults, and is reviewed at twenty one; and several States parties in good standing operate transfer or adult court trial for older children charged with grave offences, England among them, without having been held in breach. Between those two readings the honest position is this. The text of the Convention does not settle the question, and anyone who says it plainly forbids or plainly permits the mechanism is overstating an instrument. The institutional reading of the Convention, by the body authorised to give it, settles the question against the mechanism, and India has offered no reasoned answer to that reading beyond the assertion of compliance. A breach, in a system without a sanction, is a conclusion rather than an event; but the debate established that the conclusion was available, was drawn by the Parliament’s own committee before the vote, and was accepted as a cost. That, rather than the metaphysics of treaty interpretation, is the fact about the international strand that a legislative history should preserve.

There is an irony inside the judicial record that completes this section. When the uniform age of eighteen was challenged in Salil Bali, the Supreme Court upheld it in part by leaning on the Convention: the age line was defensible because it embodied the international standard and the developmental understanding behind it. Two years later Parliament retained the Convention’s own sentencing floor in Section 21 while departing from the Convention’s institutional reading on transfer, and thereby used the same instrument as a shield for the part of the scheme it kept and as a text to be distinguished for the part it changed. Instruments of this kind do not compel; they arm. The episode is a clean demonstration that an international obligation in Indian legislative practice operates less as a boundary than as a resource, available to whichever side of a domestic argument can quote it, and that the side which quoted it first and best in this debate, the committee and the courts of 2013 and 2014, lost to the side that simply proceeded.

The judicial position after enactment

The courts had closed the door to change before 2015; after 2015 they set about narrowing what Parliament had built, and the two principal decisions belong to the debate’s history because each was, in substance, a continuation of the argument by other means. The first is Shilpa Mittal v. State of NCT of Delhi, decided on 9 January 2020. Section 2(33) defines a heinous offence by a minimum punishment of seven years or more. A large family of grave offences prescribes no minimum, or a minimum below seven years, while permitting a maximum well above it, culpable homicide not amounting to murder being the standard instance; such offences fit none of the Act’s three definitions, being too heavy for the serious category as drafted and lacking the minimum the heinous category demands. The Supreme Court held that an offence which does not provide a minimum sentence of seven years cannot be treated as a heinous offence, and that the unclassified fourth category shall be treated as serious offences, in exercise of the Court’s power under Article 142 of the Constitution, from the commencement of the 2015 Act until Parliament takes the call on the matter. The consequence is not marginal. For every offence in the fourth category, no preliminary assessment arises, no transfer can be ordered, and a child of seventeen is dealt with by the Board under the summons case procedure whatever the facts alleged. The provision the House enacted in December 2015 reached every grave charge; the provision that operates reaches only those grave charges whose punishment clause carries the requisite minimum, a subset defined by the drafting habits of the penal code rather than by any judgment about children.

Between the two decisions sits a statutory discipline that belongs to the same story and is usually omitted from accounts of the provision: the appeal. An order of the Board passed after a preliminary assessment is appealable, and the appeal lies not to the Children’s Court, which hears the ordinary run of appeals from Boards and Committees, but to the Court of Sessions, which the Act arms with the assistance of experienced psychologists and medical specialists other than those whose assistance the Board itself obtained. The fresh eyes are built into the design rather than left to the appellate court’s discretion, and the structure repeats the provision’s founding idea at one remove: the capacity question is important enough to be asked twice by different people, and the second asking is not to be conducted by the experts who informed the first. The High Court’s revisional jurisdiction over the orders of Boards, Committees and Children’s Courts stands behind the appeal, so a transfer decision may be examined at three levels before a trial as an adult begins, a density of supervision without parallel elsewhere in the Act and one the debate of 2015, fixed on the clause’s existence, largely failed to price into its predictions.

Beneath the definition’s gap sits a drafting choice the debate never examined, and it is worth isolating because it explains the gap’s existence. Indian penal law had never classified offences by the adjective heinous. The Indian Penal Code grades offences by the punishment it prescribes for each, and the procedural codes classify by cognizability, bailability and the court of trial; heinous was a word from rhetoric and from sentencing judgments, not a term of art with a statutory home. The objection was made at the time, by Swagata Raha of the Centre for Child and the Law in the evidence she gave the Standing Committee: the very concept of heinous crimes had no legal basis in the classification the law otherwise used, and to import it into a statute meant for the reformation of children was to build the system’s most consequential gateway out of a word the system did not define. The enacted Act answered the objection in the only way available, by discarding the adjective’s moral content and defining the category mechanically, by the minimum punishment the offence carries. The mechanical answer is what produced the fourth category problem, since punishment clauses were drafted over a century and a half without any thought that a minimum would one day sort children between forums, and some of the gravest offences in the Code carry no minimum at all. The lesson generalises, and it is one this series meets wherever a legislature legislates by label: a category borrowed from the language of condemnation will have to be converted into a rule before it can be applied, and the conversion, not the condemnation, is where the provision’s real scope is decided. In this case the conversion was begun by the draftsman, continued by the Supreme Court, and left unfinished by Parliament, in that order.

The second decision is Barun Chandra Thakur v. Master Bholu, decided in 2022, in which the Supreme Court examined what a preliminary assessment demands of a Board and held that it is a delicate and demanding task and not a formality to be completed on the papers. The assessment concerns the capacity of a particular child, formed at a particular stage of development; a Board dealing with a heinous allegation against a child above sixteen should ordinarily take the assistance of experienced psychologists or psycho-social workers, as Section 15 permits, rather than resting on judicial impression. The insistence matters because the section makes expert assistance available rather than compulsory, and the early record of the provision included assessments conducted with no psychological input at all. Read together, the two decisions describe the judiciary’s settled posture towards the provision it had no power to prevent: the gate is narrow in law after Shilpa Mittal and is meant to be slow in practice after Barun Chandra Thakur, and between the definition and the discipline, the class of children actually exposed to adult trial is a fraction of the class the December 2015 debate imagined, on either side. Parliament, for its part, has not taken the call Shilpa Mittal reserved to it: the Juvenile Justice (Care and Protection of Children) Amendment Act, 2021 reclassified offences elsewhere in the Act but did not touch the transfer mechanism, the classification thresholds for children in conflict with law, or the Shilpa Mittal problem, which remains where the Court left it, awaiting a legislative call. The argument that began in a committee room therefore ends, for the time being, in an unusual distribution of labour: the legislature supplied the exception, and the judiciary has supplied most of its limits.

Which children can never be transferred for adult trial?

Two classes stand wholly outside the mechanism. A child below sixteen on the date of the offence can never be transferred, and neither can a child accused of an offence whose punishment clause carries no minimum of seven years, the fourth category the Supreme Court directed to be treated as serious rather than heinous.

Scrutiny obtained and rejected: the series’ argument completed

This series has repeatedly criticised legislation enacted without committee scrutiny, and the criminal codes and farm laws clusters supply its standing examples: major penal and agricultural legislation carried at speed, with the committee stage refused or curtailed, so that objections which a committee would have forced onto the record were never answered because they were never formally put. The lesson drawn there was about absence. This episode is the mirror image, and it completes the argument by supplying the contrasting case. Here scrutiny was not refused: the Bill was referred in the ordinary way, the committee took evidence across the affected interests, it reported in time, its report was considered at Cabinet level and most of it was accepted, and the disagreement that remained was the disagreement that mattered. The contrast with the farm laws episode, whose legislative lessons the series treats in farmers’ protest and legislative lessons, is therefore exact. In the one case the legislature would not look; in the other it looked, read, and disagreed. The second failure, if it is a failure, is of a different and arguably more legitimate kind, because considering an objection and rejecting it on a vote is what legislatures are for, and a committee system that could bind the House would be a second legislature and not a scrutiny mechanism at all.

That defence is available, and it should be stated fairly, because it is the strongest answer to this article’s central fact. Committees in the Indian system are advisory; their reports bind no one; governments accept most recommendations in most reports, as this government did across most of this Bill, and reserve the right to carry a provision the committee dislikes where the government’s political judgment differs. On the juvenile provision the government could say, and in substance did say, that the committee’s reading of the data was contestable, that its view of the Convention was a view and not a verdict, and that the elected House, answerable for public safety and for victims in a way a committee is not, was entitled to weigh the same record and conclude otherwise. All of that is constitutionally impeccable. What the record does not show is the weighing. The Lok Sabha passed the Bill in May 2015 with the official amendments; the Rajya Sabha then held it for seven months, through two sessions in which it was repeatedly listed and not taken up, and passed it in a single sitting in the week of the release, in a debate in which the demand for a second scrutiny, by a Select Committee of that House, was made on the floor and defeated by the absence of a formal motion rather than by a vote on its merits. Scrutiny obtained and rejected after deliberation is one thing; scrutiny obtained, rejected, and then overtaken by a timetable nobody chose is another, and the distance between them is the distance between the December sitting as it happened and the December sitting as the defence describes it. The committee system’s product on this Bill was the best evidence based account of the provision that the Indian state produced at any point in the episode. It was produced by the institution designed for the purpose, chaired by a member of the governing party, and it survives as the record against which the provision’s subsequent life is measured. That it did not prevail is the legislature’s right. That it was not answered is the legislature’s choice, and the two should not be confused in the retelling.

The debate the provision began has not ended, and its subsequent career follows the line Yechury drew in the Rajya Sabha with uncomfortable fidelity. The demand the enacted threshold was supposed to satisfy has reappeared in the form he predicted: a private member’s bill has since proposed reducing the age from which transfer is possible from sixteen to fourteen, on reasoning identical in structure to the reasoning of 2015, that grave offences by children below the existing line reveal the line to be wrongly placed. The proposal’s existence is the answer to the argument that sixteen would settle the question; a threshold adopted in response to the case that fell just the wrong side of the old line will always be indicted by the case that falls the wrong side of the new one, because the grievance the threshold answers is not about the number but about the existence of any number that the facts of a shocking case can outrun. This is the ratchet the committee system exists to interrupt, and its operation here is the final piece of evidence in the episode: the recommendation against the provision was, among other things, a prediction that the provision would not conclude the argument, and on that prediction the committee has been proved right by the provision’s own successors in the argument.

What the provision has done in operation

A controversy about a mechanism ends, or should end, in an account of what the mechanism has done, and here the record is thinner than the argument deserves. The one series that exists, the Bureau’s apprehension figures as reported by the Ministry, shows the system’s intake moving both ways after the Act came into force: the Ministry of Women and Child Development, placing the National Crime Records Bureau’s figures before Parliament in December 2017, reported 48,228 juveniles apprehended in 2014, 41,385 in 2015 and 44,171 in 2016, describing the sequence as a mixed trend. Nothing in that sequence can sustain a claim that the transfer provision deterred offending, and nothing in it can sustain the claim that it encouraged it either; apprehensions are not offending, the years straddle the commencement of the Act, and the provision touches only a small fraction of the children the series counts. The deterrence argument, which did some of the heaviest work in the passage debate, is therefore untestable on the data its own proponents cited, and it remains untested on any data collected since, because the data that would test it are not collected at all. Neither the Bureau’s annual series nor the Ministry’s reports to Parliament publish a count of preliminary assessments conducted, of assessments resulting in transfer orders, of transfers confirmed or refused by Children’s Courts, or of children convicted after transfer and the sentences they received. The state therefore does not know, in any public and verifiable form, how often the exception it enacted is used, what outcomes it produces, or whether the deterrent effect predicted for it exists. That is not a gap in the scholarship. It is a gap in the state’s own record keeping, and it sits oddly beside the intensity of the dispute: the provision that cost the system its committee consensus and its judicial consensus is the provision whose working the system does not count.

What an evaluation of the provision would require, were the state minded to conduct one, can be specified exactly, and the specification is itself a finding. It would require, first, a count of preliminary assessments ordered and completed, by State and by offence, so that the mechanism’s reach could be known; second, the division of those assessments between transfers ordered and cases retained, and of transfers between those confirmed and those refused by Children’s Courts, so that the two gates’ relative strictness could be seen; third, the outcomes of transferred trials, convictions, acquittals and sentences, so that the exception’s severity in practice could be compared with the rhetoric of both sides in 2015; fourth, the composition of the assessed population by the markers the class argument turns on, family income, schooling and representation, so that the distributional claim could be tested rather than traded; and fifth, a comparison of recorded offending by sixteen and seventeen year olds before and after commencement, read with the denominator discipline the Standing Committee modelled, so that deterrence could be addressed as a question rather than as a slogan. None of the five requires new law, new institutions or new money of any consequence; all five require only that the bodies already obliged to report on the Act’s working be asked for the numbers the Act’s most contested provision generates. That they have not been asked is the operation record’s most eloquent entry, and it converts the deterrence dispute from an empirical disagreement into a standing reproach: the claim that carried the provision is a claim its carriers have never arranged to verify.

What is documented, though it belongs to the machinery rather than the mechanism, is that the conditions the provision presupposes are the conditions the system’s own reviews have found wanting. The assessment assumes Boards that can obtain expert assistance and children’s courts that apply an independent mind; the implementation record, reviewed by the Supreme Court in Sampurna Behura v. Union of India in 2018 across the country’s Boards and Committees, found the familiar deficits of constitution, staffing, training and inspection, and a parallel finding from the Standing Committee on the later amendment noted that in 2019 only seventeen of thirty five States and Union Territories had all the basic structures and bodies required under the Act in all their districts. A Board that sits with vacancies and without the psychologists Section 15 contemplates does not conduct the assessment the provision’s defenders described; it conducts a shorter version of it, and the higher courts’ discipline on assessments, from Barun Chandra Thakur downwards, has in practice been the state’s only guarantee that the shorter version is noticed. This is the sense in which the first decision point in the table above, to legislate rather than implement, has produced its own verdict. The critics’ strongest prediction was not that the provision would be used widely and badly; it was that the provision would be built on a system that could not supply what the provision needed, and the record of Boards and experts since 2016 has borne the prediction out in the places where it can be checked.

The persistent misreadings

Three beliefs about the episode circulate so widely that a reader of ordinary care will meet all three, and each is wrong in a direction that serves one side of the argument. The first is that the law lowered the age of juvenility to sixteen. It did not, and no amendment since has done so; eighteen remains the definition, the mechanism is a case by case transfer, and the distinction is not technical but substantive, because every protection of the child in conflict with law continues to apply until a transfer order is made and two of them survive it. The second is that transfer is automatic for any grave charge against a child of sixteen or above. It is not, and the two stage structure and the judicial narrowing both stand against the belief: the Board must assess, the Court must agree, and whole classes of grave offence are excluded from the mechanism by the minimum punishment rule and the Court’s reading of it. A fourth, which fed the other three, is that the young person in the 2012 case escaped punishment: he served the full maximum the law allowed, three years in a special home, and was released because the term the statute prescribed had expired, not because any authority had been lenient with him; the grievance was with the statute’s ceiling, which is a complaint about legislation, and it was in the end addressed to legislators, which is where it belonged. The third is that the provision answered the case that prompted it. It did not and could not: criminal law does not apply retrospectively, the Act of 2015 governs offences committed after its commencement, and cases pending at commencement continued under the 2000 Act, so the young person released on 20 December 2015 was beyond the provision’s reach however early the Bill had been passed. The mother’s belief that an earlier passage would have kept him in custody was a belief about timing; the law’s answer is about commencement, and the two do not touch. Together the three misreadings explain why the debate has remained emotive after the text became settled: its subject, in the public mind, has never quite been the provision that was enacted.

What is settled, what is contested on facts, what is contested on values

The closing this subject requires is a sorting of the debate’s claims into three piles, because arguments about legislative process go wrong precisely by mixing them. Settled, on the text and the record, is all of the following. The provision does not lower the age of eighteen and never did. The assessment and the double decision are real procedural requirements and not decorative ones. The committee of Parliament examined the Bill, recommended against the provision, and was overridden, and the government’s own amendments accepted most of the report while retaining the provision’s core. The Supreme Court upheld the uniform age twice before 2015 and narrowed the enacted provision twice after it. The crime figures used in the debate showed a small and roughly stable share of recorded crime and were built on reports rather than convictions, with a reclassification effect in the sexual offence statistics that the committee named. The state does not publish a series on the provision’s use or outcomes. Contested on facts, and therefore capable of resolution by evidence if the evidence were gathered, are the assessment’s reliability, the deterrence question, and the distributional claim: whether Boards can make the capacity determination with the experts they actually have, whether trying children as adults reduces offending by children, and whether the discretion falls disproportionately on poor children. Each of these is an empirical question on which the provision’s defenders and critics make incompatible claims, and on none of them is the public series sufficient to decide. Contested on values, and therefore not resolvable by any study, are the questions underneath all of it: whether proportionate consequences for grave acts are owed to victims even where the offender is a child, whether the dignity of a child forbids adult trial whatever the act, whether a democracy may set aside its own expert committee in the name of public feeling, and whether childhood is a status the law must hold uniform or a capacity it may examine. The provision stands because Parliament had the power to enact it against advice, and it stands in its narrowed form because courts have used the tools the provision gave them to keep it narrow. Everything else in the argument, including this article, is commentary on a decision the record will not reopen.

Frequently Asked Questions

Q: Why was the juvenile age debate reopened in 2015?

Because every other route to change had been closed in sequence. The December 2012 case in Delhi, in which one accused was below eighteen and received the maximum of three years in a special home, produced the demand. The Justice Verma Committee refused it in January 2013, the Supreme Court refused it in July 2013 and again in March 2014, holding the age to be a legislative choice. The general election of 2014 then brought in a government whose Minister of Women and Child Development introduced a Bill containing the transfer provision in August 2014. The debate reopened as legislation because legislation was the only door left.

Q: What did the parliamentary committee say about trying children as adults?

The Department-related Parliamentary Standing Committee on Human Resource Development, chaired by Dr Satyanarayan Jatiya, examined the Bill and in its 264th Report, presented on 25 February 2015, recommended against both reducing the age of juvenility and creating a mechanism for trying children as adults. It held that such a mechanism would violate India’s obligations under the Convention on the Rights of the Child, that the National Crime Records Bureau data showed juvenile crime at 1.2 per cent of the total and did not support drastic change, that sixteen to eighteen years is an extremely sensitive and critical age requiring greater protection, and that the existing framework should instead be properly implemented.

Q: What did the Supreme Court hold about the juvenile age before 2015?

In Salil Bali v. Union of India, decided in July 2013, the Court upheld the uniform age of eighteen in the Act of 2000 as a constitutionally permissible classification resting on intelligible differentia: the age had been fixed on expert understanding that children could still be redeemed and restored to society up to that age, and the line drew support from the Convention on the Rights of the Child. In Subramanian Swamy v. Raju, decided on 28 March 2014, the Court declined to read the definition down or to create an exception for grave offences, treating the age as a considered legislative choice. Pratap Singh, in 2005, had settled that age is reckoned at the date of the offence.

Q: Does trying children as adults breach India’s international obligations?

The Convention on the Rights of the Child, to which India acceded in 1992, defines a child as a person below eighteen and requires treatment consistent with the child’s dignity and reintegration, but does not in terms forbid an adult trial after individual assessment. Its monitoring body has nonetheless advised against the device twice: General Comment No. 10 of 2007 recommended that States not withdraw sixteen and seventeen year olds from juvenile justice, and General Comment No. 24 of 2019 urged States to remove provisions treating children as adults. The Standing Committee and UNICEF treated the provision as a breach; the government maintained compliance, pointing to the retained bars on death and on life imprisonment without release.

Q: What does the data on juvenile crime actually show?

The National Crime Records Bureau figures on which the debate turned show a small and roughly stable share. Juvenile crime formed 1 per cent of recorded crime in 2003 and 1.2 per cent in 2013, while children of sixteen to eighteen rose from 54 per cent to 66 per cent of all juveniles accused across the same decade. In 2013, children of that age apprehended for murder formed 1.3 per cent of all persons arrested for murder, and those apprehended for rape 3.29 per cent of all persons arrested for rape. The figures rest on first information reports rather than convictions, and the sexual offence figures cross the 2012 raising of the age of consent, which reclassified consensual activity as rape.

Q: Which offences count as heinous for adult trial?

An offence is heinous under Section 2(33) of the Act where the minimum punishment prescribed for it is imprisonment for seven years or more, and only a child who had completed sixteen years on the date of such an offence can face a preliminary assessment at all. Murder and rape were the standard instances in the debate, because their punishment clauses carry the requisite minimum. The definition looks to the punishment clause and not to the facts of the case, which is why offences that are grave in fact but carry no minimum, or a minimum below seven years, fall outside it, the position the Supreme Court confirmed in Shilpa Mittal.

Q: What did the Shilpa Mittal judgment decide?

In Shilpa Mittal v. State of NCT of Delhi, decided on 9 January 2020, the Supreme Court held that an offence which does not provide a minimum sentence of seven years cannot be treated as a heinous offence under the Act. A large class of grave offences prescribes a high maximum but no minimum, or a minimum below seven years, and fits none of the Act’s three definitions. The Court directed that this fourth category be treated as serious offences, under Article 142 of the Constitution, from the commencement of the Act of 2015 until Parliament takes the call. For such offences no preliminary assessment arises and no transfer can be ordered.

Q: Has trying children as adults achieved anything?

On the published record, the question cannot be answered, and that is itself the finding. Neither the National Crime Records Bureau nor the Ministry publishes a count of preliminary assessments, transfer orders, confirmations by Children’s Courts, or convictions and sentences after transfer, so the provision’s use and outcomes are not publicly known. The one available series, apprehensions, moved both ways across the commencement period: the Ministry reported 48,228 juveniles apprehended in 2014, 41,385 in 2015 and 44,171 in 2016, calling it a mixed trend. No deterrent effect can be read from those figures in either direction, and none has been demonstrated from any other official source.

Q: Did the 2015 Act lower the age of juvenility to sixteen?

No. Section 2(12) of the Act defines a child as a person who has not completed eighteen years of age, exactly as the Act of 2000 did, and every child enters the system before a Juvenile Justice Board under that definition. What the Act created is a transfer mechanism inside the definition: a child who had completed sixteen years on the date of an offence classified as heinous may, after a preliminary assessment by the Board and an independent decision by the Children’s Court, be tried as an adult. The belief that the age itself fell is the debate’s most durable error, repeated by supporters claiming a tougher law and by critics mourning a lost one.

Q: Did the Justice Verma Committee recommend lowering the juvenile age?

No, it recommended the opposite. The committee under Justice J. S. Verma, constituted after the December 2012 case to review the criminal law on sexual offences, reported on 23 January 2013 and held that the age of juvenility need not be lowered from eighteen. It reasoned from the developmental evidence on adolescent decision making, drawing on the work of the psychologist Laurence Steinberg, and Justice Verma added the practical point that a person of sixteen sentenced to a long term would be released in his middle thirties with little assurance of having reformed. The same report recommended stronger punishments for sexual offences, which Parliament enacted that spring.

Q: What did Maneka Gandhi argue for the transfer provision?

Maneka Gandhi, the Minister of Women and Child Development who introduced the Bill in August 2014 and carried it through both Houses, argued that the existing law encouraged offending because older children knew its three year ceiling protected them whatever they did, citing the police view she had been given that a very large share of crime was committed by sixteen year olds who knew the Juvenile Justice Act. She presented the assessment as a discrimination between an act committed in a childlike frame of mind and one committed in an adult frame of mind, stressed that no child would be sent directly to a jail, and pointed to custody apart from adult prisoners until twenty one, with a review at that age.

Q: What did UNICEF and child rights organisations argue against the Bill?

UNICEF’s submissions on the legislation argued that the need was to implement the existing framework, Boards, Committees and registered institutions, rather than to change the age rule, and that the provision ran against the Convention on the Rights of the Child. The HAQ Centre for Child Rights, whose co-director Enakshi Ganguly argued the point across the Act’s life, contended that the preliminary assessment demands a finding about a child’s capacity that cannot reliably be made, that transfer severs the child from the fresh start principle, and that the children before Boards are overwhelmingly poor, so the discretion carries a class risk. Researchers from the Centre for Child and the Law told the Standing Committee that the category of heinous offence had no basis in the law’s own classifications.

Q: Could the transfer provision have applied to the juvenile in the December 2012 case?

No, and the point was made on the floor of the Rajya Sabha during the passage debate itself. Criminal law cannot be applied retrospectively, and Article 20(1) of the Constitution forbids conviction under an ex post facto law. The Act of 2015 governs offences committed after its commencement on 15 January 2016, and Section 25 keeps cases pending at commencement under the Act of 2000. The young person whose release on 20 December 2015 set the passage timetable was therefore beyond the provision’s reach however early the Bill had passed, a fact his case’s own prominence tended to obscure: the law was always and only about the next case.

Q: Why did the government proceed after the Standing Committee advised against the provision?

Because a standing committee’s report is advisory, and the government divided its response. It accepted most of the 264th Report: the retrospective clause exposing a person apprehended after twenty one was dropped, the assessment period was enlarged, and recommendations across the Bill’s adoption and institutional provisions were given effect. On the central clause it disagreed, treating the Committee’s reading of the crime data as contestable, its view of the Convention as one reading rather than a verdict, and the elected House as entitled to weigh proportionality, deterrence and victims’ interests for itself. The Cabinet approved that course in April 2015, and the Lok Sabha passed the Bill with the transfer provision retained on 7 May 2015.

Q: Why did Parliament choose case by case transfer instead of lowering the age to sixteen?

A general reduction would have swept every child of sixteen into the adult system for every offence, which the Verma Committee had rejected and which the government did not propose. Transfer by assessment let Parliament claim, accurately, that the age of a child was unchanged, that no child would face an adult court on the charge alone, and that the exception would turn on the individual child’s capacity and understanding rather than on a birth date. The design also had a political function: it allowed members who would not vote for a lower age to vote for a gated exception. The price, identified by the critics, is that the hardest question in the system is delegated to Boards district by district.

Q: Why was sixteen chosen as the threshold for transfer?

Sixteen presented itself to the House as an inherited line rather than an invented one. The Juvenile Justice Act of 1986 had set sixteen as the age up to which a boy was a juvenile, a point the Leader of the Opposition, Ghulam Nabi Azad, made in the Rajya Sabha debate, and several of the older state children Acts had drawn similar lines before the Act of 2000 unified the age at eighteen. Supporters added that basic cognitive capacity, the ability to understand an act and its consequences, is on the developmental evidence broadly adult by sixteen. Critics answered that the 1986 line was one half of a sex differentiated scheme nobody proposed to restore, and that an inherited number is not a scientific finding.

Q: Why do critics question whether a Board can assess a child’s maturity reliably?

The objection is that Section 15 requires a finding, about a particular child’s mental capacity and ability to understand consequences at the time of the act, for which no validated instrument exists. Developmental science describes adolescents as a population; it does not certify individuals. The difficulty had been conceded in litigation by the state’s own side: in the Subramanian Swamy proceedings the Union’s counsel submitted that the maturity of a person below eighteen cannot be objectively determined on an individual or case to case basis. Critics add that Boards sit with a magistrate and two social workers, that the psychologists Section 15 makes available are scarce in many districts, and that in practice the gravity of the charge supplies the answer the assessment is meant to find.

Q: What did Barun Chandra Thakur decide about the preliminary assessment?

In Barun Chandra Thakur v. Master Bholu, decided in 2022, the Supreme Court held that a preliminary assessment is a delicate and demanding exercise and not a formality to be completed on the papers. The assessment concerns the capacity of the particular child before the Board, formed at a particular stage of development, and a Board dealing with a heinous allegation against a child above sixteen should ordinarily take the assistance of experienced psychologists or psycho-social workers, as Section 15 permits, rather than proceeding on judicial impression. The judgment matters because the Act makes expert assistance available rather than compulsory, and the early record of the provision included assessments conducted with no psychological input at all.

Q: How did the two Houses treat the Bill differently?

The Lok Sabha considered the Bill in the spring of 2015 after the government moved official amendments founded on the Standing Committee’s report, and passed it on 7 May 2015 in a debate that concentrated on the design of the transfer mechanism rather than its existence. The Rajya Sabha then held the Bill for seven months, through two sessions in which it was listed and not taken up. It was finally debated on 22 December 2015, two days after the release of the young person in the 2012 case, in a sitting of nearly five hours in which a demand to refer the Bill to a Select Committee ended in a walkout by the Left parties, after which the Bill passed by voice vote.

Q: How do other legal systems decide whether a child faces an adult court?

The systems divide by who decides. In England and Wales the age of criminal responsibility is ten, and a child charged with a grave crime such as murder is tried in the Crown Court, an adult court sitting with adaptations, the charge itself settling the forum. In the United States three mechanisms coexist across the states: judicial waiver after a hearing, prosecutorial direct filing, and statutory exclusion of defined offences, while the Supreme Court has limited the sentences available, barring the death penalty for offences committed below eighteen in Roper v. Simmons in 2005. Germany runs the discretion the other way, applying juvenile law to young adults up to twenty one where maturity warrants it. India chose the hearing model, doubled.