The National Investigation Agency is the institution the Indian Constitution appears to forbid and the Indian state decided it could not do without. Police and public order are subjects the Constitution assigns to the states. Investigation of crime is, in the ordinary scheme of the Seventh Schedule, the work of a state police force answerable to a state government. Yet the National Investigation Agency Act, 2008 created a central police agency whose officers exercise police powers throughout India, which can be directed by the Central Government to take an investigation out of the hands of a state police force, and whose entry into a case requires the consent of no state at all. The Agency has operated on that basis since the Act came into force on the last day of 2008. Its constitutionality has been challenged by a state government in an original suit before the Supreme Court, and the challenge goes to the foundations of the federal design rather than to any detail of drafting.

An institution profile in this series asks a single question: what was the body designed to be able to do? The answer here is unusual, because the design choice that matters most in the National Investigation Agency Act is an omission. The statute nowhere requires the Central Government to seek, or the Agency to obtain, the agreement of the state in whose territory a scheduled offence was committed before the Agency investigates it. That absence is not an oversight. It is the deliberate point of contrast with the Central Bureau of Investigation, which works under a statutory consent regime that states can and do withdraw. Consent, as the comparison section of this profile argues, is the whole difference between the two agencies, and it explains almost every federalism dispute about central investigation in India. Everything else in the Act, the Schedule of offences, the fifteen-day direction mechanism, the Special Courts and the appeal structure, is machinery built around that single design decision.

The gap the Act was written to fill

The occasion for the Act was the attacks on Mumbai of 26 to 29 November 2008, in which ten gunmen who had arrived by sea killed 166 people at sites that included a railway station, two hotels, a Jewish centre and a hospital. The attacks exposed many failures, and parliamentary and public discussion after them returned repeatedly to one institutional fact: no agency of the Union had a standing statutory mandate to investigate an act of terrorism that crossed state lines. The investigation of the attacks themselves fell to the Mumbai Police and the Maharashtra Anti-Terrorism Squad, with central agencies assisting rather than leading. Offences of the kind committed in Mumbai, offences against the sovereignty and security of India and offences under statutes enacted to implement international conventions, were scattered across the criminal law, and the power to investigate them belonged, as a matter of constitutional allocation, to whichever state police force had territorial jurisdiction.

The legislative response came in the winter session of Parliament in December 2008, within weeks of the attacks, and it came as a package of two statutes. The first amended the Unlawful Activities (Prevention) Act, 1967, the standing anti-terror statute whose provisions are set out in the complete guide to the UAPA 1967 and whose 2008 amendment and its successors widened the definition of a terrorist act and the machinery of proscription. The second created the investigating institution the amended substantive law had never had: a national agency to investigate and prosecute the offences. The National Investigation Agency Bill was introduced and passed in the same session, received the assent of the President on 31 December 2008, and was brought into force on the same day. It is Act 34 of 2008.

The speed of that passage is a fact about legislative process that should be stated plainly and without decoration. A statute redrawing the boundary between the Union and the states on investigation was moved from introduction to assent in roughly a fortnight, in a session dominated by the aftermath of an atrocity, with none of the extended committee scrutiny that a measure of this constitutional weight would ordinarily receive. Defenders of the timetable, including P. Chidambaram, the Home Minister who piloted the Bill after taking office in the reshuffle that followed the attacks, treated the absence of a federal investigating agency as an emergency gap that the Mumbai attacks had made undeniable. Critics of the timetable, a group that included parliamentarians who voted for the Bill, accepted the need for an agency while recording unease at the manner of its creation. Both things are true, and the haste of December 2008 is part of the explanation for why the constitutional argument about the Act was postponed to litigation rather than settled in debate.

Why was the National Investigation Agency created in 2008?

It was created because the Mumbai attacks of November 2008 demonstrated that no central agency held a standing mandate to investigate terrorism across state lines. Investigation belonged to state police forces by territory, so an attack planned abroad, financed across borders and executed in one city had no single investigator. Parliament filled the gap within weeks.

The Agency the statute builds

Section 3 of the Act empowers the Central Government to constitute a special agency, to be called the National Investigation Agency, for the investigation and prosecution of offences under the Acts specified in the Schedule. The opening words of the section override the Police Act, 1861, the colonial statute under which state police forces are organised, and the override is the first textual signal that the Agency is not a state force on deputation but a creature of the Union. Section 3 then gives officers of the Agency, throughout India, in relation to the investigation of scheduled offences, all the powers, duties, privileges and liabilities that police officers have in connection with the investigation of offences. An officer of or above the rank of Sub-Inspector may exercise, anywhere in India, the powers of the officer in charge of a police station in the area where the officer is present. The jurisdictional unit of ordinary policing, the police station and its territory, is replaced for Agency officers by the whole territory of India.

Section 4 vests the superintendence of the Agency in the Central Government and its administration in a Director General appointed by the Central Government, who exercises such of the powers of a Director General of Police in a state as the Central Government specifies. The chain of command is therefore short and unambiguous: the Agency answers to the Union executive, through a chief who holds office at the Union’s appointment, and no state institution appears anywhere in the structure of command, appointment or superintendence. The manner of the Agency’s constitution and the conditions of service of its members are left to rules under Section 5, which is why the Agency’s internal organisation, its branch offices and its specialised divisions are matters of executive arrangement rather than of the statute. The Act builds a shell of powers and leaves the staffing of the shell to government.

Three features of this skeleton deserve emphasis before the mechanism is examined. The first is that the Agency is defined by its offence list rather than by any general crime mandate; it has no roving commission to investigate crime in India, only the offences Parliament has scheduled. The second is that the Agency does not select its own cases in the ordinary course; the trigger is a direction of the Central Government under Section 6, and the discretion that matters is therefore the executive’s, a point the Chhattisgarh suit presses hard. The third is that the Act is as much about prosecution as investigation. The Agency prosecutes through Public Prosecutors appointed under Section 15 before Special Courts constituted under the Act, so the federal intrusion, if intrusion it is, runs the full length of a criminal case from first information to appeal.

The Agency as an organisation: deputation, command and accountability

The statute’s silence about the Agency’s internal life is itself a design feature worth examining, because it determines what kind of institution the Agency has become. The Act creates the shell and the powers; the people arrive by deputation. Officers of the Agency are drawn from the Indian Police Service, from state police forces, from central armed police forces and from other central services and departments, serving in the Agency for a tenure before returning to their parent cadres. The model has consequences in both directions. It gives the Agency immediate access to experienced investigators and to officers with local knowledge of the states from which they are drawn, and it avoids the decades-long project of building a federal cadre from nothing. It also means the Agency’s institutional memory is carried by officers who will leave, and that the quality of an investigation can turn on the quality of the deputation a state or service chooses to send, a variable no statute can control.

The appointment of the Director General illustrates the same pattern of executive design. The Act contains no selection committee, no statutory tenure and no removal procedure for the Agency’s chief: Section 4 leaves the appointment to the Central Government and the Director General’s powers to executive specification. The contrast with the older agency is instructive once more. The Director of the Central Bureau of Investigation is appointed, under Section 4A of the Delhi Special Police Establishment Act as amended in 2003, on the recommendation of a committee composed of the Prime Minister, the Leader of the Opposition in the Lok Sabha and the Chief Justice of India or a judge of the Supreme Court nominated by the Chief Justice, and holds a tenure of not less than two years secured by Section 4B. Whatever the practical record of that committee, its existence in the statute book shows that Parliament knows how to insulate an investigating chief when it chooses to. In the National Investigation Agency Act it chose not to, and the choice is consistent with the Act’s theory of the institution: the Agency is an instrument of the Union executive for a defined class of offences, and its accountability runs upward to that executive rather than outward to a bipartisan appointment process.

Accountability within the Act takes one express statutory form. Section 18 bars any prosecution, suit or other legal proceeding against a member of the Agency, or a person acting on the member’s behalf, for anything done or purported to be done in exercise of the powers conferred by the Act, except with the previous sanction of the Central Government. Sanction provisions of this kind are a standard feature of Indian security legislation, defended on the ground that officers who act against armed conspiracies must not be answerable in private litigation for operational decisions, and criticised on the ground that the sanctioning authority is the very government that directs the Agency’s work, so the safeguard and the power share an author. The Act offers no alternative forum: there is no statutory inspectorate, no parliamentary reporting requirement written into the Act, and no annual report laid before Parliament by force of the statute. Oversight of the Agency is therefore political and judicial rather than institutional: political, through the Home Ministry’s answerability to Parliament for the directions it issues; judicial, through the trial courts that test its evidence and the constitutional courts that remain open to challenge against its actions in a given case.

It is worth pausing on what that distribution of accountability means for the federalism debate, because the debate is usually conducted as though the Act’s only question were competence. Even a reader who accepts the Union’s competence argument in full is left with a second question the Act answers entirely in the executive’s favour: who watches the watchman, and by what statutory instrument? The framers of the CBI’s modern charter thought the appointment process was the place to build the safeguard. The framers of the NIA Act built the safeguard, such as it is, into the offence list and the direction mechanism, and left the institution itself bare. That bareness is the honest ground on which the Chhattisgarh suit’s arbitrariness argument stands, and it is a ground the Union defends not by pointing to a safeguard but by pointing to the record, a defence examined after the courts that test that record have been described.

The Schedule: a mandate written as a list

The Agency’s jurisdiction is the Schedule, and the Schedule is a list of other statutes. As enacted in 2008, it named the Atomic Energy Act, 1962; the Unlawful Activities (Prevention) Act, 1967; the Anti-Hijacking Act, 1982; the Suppression of Unlawful Acts against Safety of Civil Aviation Act, 1982; the SAARC Convention (Suppression of Terrorism) Act, 1993; the Suppression of Unlawful Acts Against Safety of Maritime Navigation and Fixed Platforms on Continental Shelf Act, 2002; and the Weapons of Mass Destruction and their Delivery Systems (Prohibition of Unlawful Activities) Act, 2005. To those seven statutes it added offences under Chapter VI of the Indian Penal Code, sections 121 to 130, which are the offences against the State, waging war against the Government of India and the offences surrounding them, together with sections 489A to 489E of the Penal Code, the counterfeit currency offences. An offence under any of these provisions is a scheduled offence, and a scheduled offence is the only kind of offence the Agency exists to investigate.

The architecture of the list repays attention, because it explains the Agency’s character better than any ministerial speech. Almost every statute on it gives effect to an international convention or addresses a subject that sits on the Union side of the constitutional line: atomic energy, hijacking and aviation safety, maritime navigation, weapons of mass destruction, terrorism under the SAARC Convention. The Penal Code entries are the offences against the State itself. The preamble of the Act speaks the same language, describing offences affecting the sovereignty, security and integrity of India, the security of the State, friendly relations with foreign States and offences under Acts enacted to implement international treaties, agreements, conventions and resolutions of the United Nations and other international organisations. The Schedule was drawn, in other words, to track the subjects on which the Union’s legislative competence was least contestable, a drafting choice that became the first line of the Union’s defence when the Act was challenged. The Agency sits inside the wider field of Indian anti-terror legislation traced in the lineage of the anti-terror statutes, but it is the only institution in that field whose entire mandate is a cross-reference to other laws.

What is a scheduled offence under the National Investigation Agency Act?

A scheduled offence is an offence under one of the statutes listed in the Schedule to the Act, from the Unlawful Activities (Prevention) Act, 1967 to the offences against the State in Chapter VI of the Indian Penal Code. The Schedule is the boundary of the Agency’s mandate: an offence not on the list is outside its power.

Because the mandate is a list, the mandate can grow by amendment without any change to the Agency’s structure, and that is the route by which the argument about the Agency has shifted since 2008. The original list could be defended, entry by entry, as a catalogue of offences against the Union and its international obligations. The amendment of 2019, examined below, added offences of a different character, and the debate about the Schedule has become a debate about how far the logic of the list can stretch before the institution it defines stops being a counter-terror agency and becomes a general federal police in waiting. That debate cannot be had honestly without the mechanism, because the Schedule alone does not put the Agency into a case. A scheduled offence may be investigated by a state police force in the ordinary way, and most are. The Agency enters only when the Central Government directs it to enter, and the direction is where the constitutional weight of the Act is concentrated.

The direction mechanism: how a case changes hands

Section 6 builds a pipeline with one decision point. When information about a scheduled offence is recorded at a police station under Section 154 of the Code of Criminal Procedure, the officer in charge must forward the report to the State Government forthwith. The State Government must forward it to the Central Government as expeditiously as possible. The Central Government then has fifteen days from receipt of the report to determine two things: whether the offence is a scheduled offence, and whether, having regard to the gravity of the offence and other relevant factors, it is a fit case to be investigated by the Agency. If both questions are answered affirmatively, Section 6(4) obliges the Central Government to direct the Agency to investigate. Section 6(5) adds a power that bypasses the pipeline altogether: if the Central Government is of the opinion that a scheduled offence has been committed and is required to be investigated under the Act, it may direct the Agency to investigate on its own motion, without any report from the state at all.

The consequences of a direction are set out in Section 6(6), and they are absolute in form. The State Government and every state police officer investigating the offence must stop investigating and must forthwith transmit the relevant documents and records to the Agency. There is no handover negotiation, no joint investigation by default, and no statutory role for the state’s view on whether the takeover is wise. The one qualification is Section 6(7), inserted for the removal of doubts: until the Agency actually takes up the investigation, the officer in charge of the police station must continue to investigate, so that evidence is not lost in the interval between the direction and the Agency’s arrival. The Supreme Court gave that qualification practical effect in Naser Bin Abu Bakr Yafai v. State of Maharashtra, decided in 2021, holding that a state agency which had continued to investigate a UAPA case, and had filed a charge sheet, before the Agency took over had acted within the scheme of Section 6 rather than in defiance of the direction.

How long does the Central Government have to decide whether the Agency should take a case?

Fifteen days. Section 6(3) requires the Central Government to determine, within fifteen days of receiving the state report, whether the offence is a scheduled offence and whether its gravity makes it a fit case for the Agency. The clock runs from receipt of the report, not the offence, and the suo motu power in Section 6(5) carries no clock.

The transmission of the case file under Section 6(6) is the point at which the statute touches the physical stuff of investigation: first information reports, case diaries, seizure memos, forensic reports and witness statements, assembled by one force and handed to another, on whose integrity the prosecution will depend. Records that cross institutional boundaries are only as trustworthy as the care taken with their custody, a discipline whose importance extends well beyond policing to any holder of sensitive files; the practices described in accounts of how journalists use VaultBook to protect sources rest on the same principle, that a document’s value in any later proceeding depends on an unbroken and demonstrable chain of keeping. The Act assumes the handover; the quality of the handover is an administrative achievement the statute cannot guarantee.

What the Agency may do with a case, and what it may not

Once directed, the Agency investigates with the powers Section 3 confers, and two further sections widen and qualify the mandate in ways that matter in practice. Section 8 permits the Agency, while investigating a scheduled offence, to investigate any other offence the accused is alleged to have committed if that offence is connected with the scheduled offence, so a terrorism investigation that uncovers connected offences under the ordinary law does not fracture into two investigations in two forces. Section 9 obliges the State Government to extend all assistance and cooperation to the Agency, which converts the displaced state’s role from investigator to auxiliary. Section 7 runs in the opposite direction: the Agency may request a State Government to associate itself with an investigation, and may, with the previous approval of the Central Government, transfer a case back to the State Government for investigation and trial. The door, in other words, opens both ways, but both handles are on the Union side of it.

Section 10 is the saving that the Union relies on when the Act is described as a takeover statute. It provides that nothing in the Act affects the power of the State Government to investigate and prosecute any scheduled offence, except where the Act itself provides otherwise. The exception is the direction: a state investigation of a scheduled offence is lawful and ordinary until the moment a direction under Section 6 issues, at which point Section 6(6) displaces it. The federal settlement the Act enacts is therefore not a partition of cases between two forces but a hierarchy with a trigger: concurrent competence, Union supremacy on direction, and a state duty of assistance once the trigger is pulled. Whether that settlement is within Parliament’s competence is the question the Supreme Court has been asked; that it is the settlement the text enacts is not in doubt.

Reading the Schedule entry by entry

The Schedule rewards a slow reading, because each entry is a small essay on why the Union claims the subject. The Atomic Energy Act, 1962, which stood first in the original list, governs a field that is exclusively the Union’s: atomic energy is Entry 6 of the Union List, the installations are central, the material is the most dangerous the state handles, and an offence against the regime is an offence against the Union in the most literal sense the criminal law knows. The Unlawful Activities (Prevention) Act, 1967 follows, and its position second on the list is the list’s centre of gravity: the UAPA is the standing code of terrorist offences, unlawful associations and terrorist organisations, and the bulk of the Agency’s docket in practice is UAPA offending, alone or in combination with other entries. An agency for scheduled offences that lacked the UAPA would be an agency for the periphery of terrorism; with the UAPA at its heart, the Agency is the UAPA’s dedicated investigator, the institution the substantive statute lacked for the first forty years of its life, during which its offences were investigated by whichever state force the geography of the crime assigned.

The four convention statutes that follow share a single explanation. The Anti-Hijacking Act, the Suppression of Unlawful Acts against Safety of Civil Aviation Act, the SAARC Convention (Suppression of Terrorism) Act and the maritime navigation statute of 2002 each give domestic effect to an international instrument to which India is party: the Hague and Montreal conventions on aviation offences, the regional convention of the South Asian Association for Regional Cooperation, and the Rome convention on unlawful acts against maritime navigation and fixed platforms. Offences under them are, by definition, offences in which a foreign element is present or probable, a hijacked aircraft, a vessel on the high seas, a convention obligation owed to other states. Parliament’s competence to legislate for treaty implementation under Article 253 and the foreign affairs entries is the least contested ground in the entire Union List, and the drafters of the Schedule placed that ground at the list’s centre deliberately, so that the Agency’s mandate would rest, in the majority of its entries, on the Union’s external sovereignty rather than on any claim over internal policing.

The Weapons of Mass Destruction Act, 2005, which closed the original list of statutes, implements the obligations of United Nations Security Council resolution 1540 on the non-proliferation of weapons of mass destruction to non-state actors, and belongs to the same external family. The Penal Code entries that complete the Schedule are of a different but equally deliberate character. Chapter VI contains the offences against the State: waging or attempting to wage war against the Government of India, conspiring to commit such offences, collecting arms with that intention, sedition’s neighbours and companions in the chapter’s scheme. These are the offences in which the victim named by the law is the Union itself. The counterfeit currency sections protect the Union’s currency, issued under Union authority, against forgery organised characteristically across borders and on a scale that implicates the integrity of the monetary system rather than the loss of any individual victim. Entry by entry, in other words, the drafters built a list in which every item answers the competence objection in advance: this offence belongs to the Union because the Union is its victim, its guarantor, its treaty party or its constitutional assignee. The defence of the Schedule, when the Act is challenged, is written into the Schedule.

The 2019 additions disturb that architecture in a way the slow reading makes visible. Human trafficking is an offence whose victim is a person, not the Union; prohibited arms manufacture is an offence against a licensing regime administered substantially by states; cyber-terrorism under Section 66F is an offence against computer systems wherever they sit. Each addition can be defended on the inter-state and international character of the offending, and the trafficking networks and currency rackets of the real world are in fact cross-border enterprises, but the defence has changed its nature: from the identity of the victim to the geography of the crime. Geography is a weaker constitutional argument than identity, because almost every serious offence has some geography beyond one state, and a Schedule justified by geography has no internal stopping point. The profile’s earlier discussion of the Schedule’s growth identified this as the second objection to the Act, distinct from the competence objection to its existence. The slow reading shows why the objection has force: the 2008 list was an argument, and the 2019 list is a precedent, and precedents in a Schedule are enlarged by the same simple majority that made them.

The fifteen days and the fit case: discretion at the centre of the Act

The direction mechanism concentrates the Act’s real power in two phrases, and a profile of the institution has to interrogate them. The first is “the gravity of the offence and other relevant factors” in Section 6(3), the standard by which the Central Government judges whether a scheduled offence is a fit case for the Agency. The Act does not define gravity, does not rank the relevant factors, does not require the determination to be reasoned in writing to the state, and does not provide for the order to be published. The second phrase is the opinion formula of Section 6(5), under which the Central Government’s own opinion that a scheduled offence has been committed and requires investigation under the Act is a sufficient foundation for a direction. Indian administrative law is familiar with such subjective satisfaction clauses and has built, over decades, a limited judicial review around them: a direction may be tested for mala fides, for reliance on irrelevant considerations and for the absence of any material on which the opinion could rest, but the merits of the choice are for the executive. The Act’s critics are therefore right about the width of the discretion, and the Act’s defenders are right that the width is not lawless. Both propositions have to be held together to understand the institution as it actually operates.

The pipeline structure of Section 6 also contains a fact that the rhetoric of takeover tends to obscure: in the ordinary course, the process begins with the state. It is the state police station that records the information, the state police that investigates from the first hour, and the State Government that forwards the report. The Central Government’s fifteen days run against that background, and the material on which it decides is, in the main, material the state itself has supplied. The design presupposes a state that reports promptly and completely, and it gives the state no reciprocal right: a state that wants the Agency in a case cannot compel a direction, and a state that wants the Agency out of a case cannot refuse one. Requests from states for the Agency to take over difficult investigations are a regular feature of the system’s operation in practice, particularly where a conspiracy crosses several states and no single state force can follow it whole, and the direction in such cases is the product of agreement rather than imposition. The statute, characteristically, does not mention any of this. It writes the power as unilateral because the cases it was designed for were the cases in which agreement could not be assumed.

The fifteen-day limit deserves a more sceptical paragraph than it usually receives. Its function in the design is discipline: the state’s investigation must not hang in suspense while the Centre deliberates, and Section 6(7) keeps the state police working in the interval. But the Act prescribes no consequence for a determination made late, no deeming provision under which a delayed case stays with the state, and no bar on a direction issuing long after the event on the suo motu route, which carries no clock at all. The Samjhauta Express investigation, examined in the record section below, was taken over in July 2010 for a bombing committed in February 2007, a perfectly lawful use of the Act that illustrates the elasticity of its timetable. Time limits in the Act thus operate as administrative guidance rather than as jurisdictional conditions, and a reader assessing the Act’s restraints should count them accordingly: the restraints that bind are the Schedule, the requirement of a direction, and the supervision of the courts over the investigation that follows, not the calendar.

There is a final feature of the mechanism that explains much of its subsequent history. The determination under Section 6(3) is made on “information made available by the State Government or received from other sources”, and those other sources include the intelligence agencies of the Union. A direction can therefore rest on material a state has never seen, and the state whose investigation is displaced may be unable to test the basis of its own displacement. That opacity is operationally explicable, since intelligence material cannot be circulated for comment without destroying its value, and constitutionally uncomfortable, since the coordinate government whose constitutional function is being suspended is suspended on evidence it cannot inspect. The Act lives with the discomfort rather than resolving it, and the Chhattisgarh suit can be read, at its deepest level, as an attempt to make the Supreme Court resolve what Parliament in 2008 and again in 2019 declined to.

An investigation in two forces: the seam where cases change hands

The Act’s mechanism is easiest to evaluate as a lived sequence rather than as a list of sections, and the sequence has a seam in the middle of it that determines whether the design works. A scheduled offence is, in the first hours, indistinguishable from any other grave crime: a first information report at a local police station, a scene to be preserved, witnesses to be separated, a post-mortem, seizures. All of that belongs to the state force, and the quality of everything that follows depends on it, because the Agency inherits a case; it does not begin one. The records the state transmits under Section 6(6) are the case in documentary form, and any defect in the first hours, an unsealed seizure, an unrecorded statement, a scene left unguarded, travels with the file into the Special Court and cannot be cured by the eminence of the second investigator. The direction mechanism is sometimes described as though the Agency’s arrival upgrades the investigation. In the respects that decide trials, the opposite is true: the Agency is the prisoner of the investigation it inherits, and its celebrated record is built substantially on case files made by state police officers whose names appear nowhere in the argument about federalism.

The seam has a second side. When the Agency takes over, it does not merely receive paper; it receives a live investigation with accused in custody or at large, witnesses under varying degrees of protection and exposure, and forensic material in laboratories it does not run. The state force that investigated to that point knows the terrain, the informants and the local language of the conspiracy, and the Act’s answer to the loss of that knowledge is Section 9’s duty of assistance and Section 7’s power to request the state’s association with the investigation. In practice the two forces work cases together after a takeover far more often than the takeover metaphor suggests, with state officers attached to the Agency’s teams and the Agency’s officers directing the legal architecture of the case. The statute’s hierarchy is real, but its operation is collaborative, and the collaboration is not a courtesy: it is the only way the design can function, since the Agency maintains no constabulary of its own in the districts where its cases arise.

The seam appears once more at the prosecution stage. A case investigated by a state force is prosecuted by the state’s public prosecutors before the ordinary courts; a case taken over by the Agency is prosecuted by prosecutors appointed under Section 15 before a Special Court, and the change of forum changes the procedural world the accused inhabits: cognizance without committal, day to day trial in principle, witness protection measures under Section 17, and an appeal that goes not through the ordinary hierarchy but directly to a two-judge bench of the High Court under Section 21, with bail itself appealable. Defence lawyers who practise in these courts describe a recognisably different litigation culture, faster in its listing, stricter in its adjournment practice, and dominated by a small bar that specialises in scheduled offence trials. None of that is in the Act’s text, but all of it is the Act’s consequence, and a profile of the institution that confined itself to the text would leave the reader unable to recognise the institution in the world.

There is a corresponding seam in the position of the victim and the witness, whom the Act treats with unusual attention for an investigation statute. Section 17’s protection measures, in camera proceedings, secrecy of identity and address, the bar on publication, exist because the offences on the Schedule are the offences in which witnesses are killed, and the history of Indian terrorism trials before 2008 is littered with prosecutions that collapsed when witnesses were intimidated into silence or hostility. The Samjhauta trial’s hostile witnesses belong to that older pattern persisting inside the newer machinery. The Act cannot compel courage, but it is one of the few Indian statutes to treat witness survival as a problem of institutional design rather than of individual fortitude, and the Special Court’s power to conceal identity is the sharpest single procedural departure the Act makes from ordinary criminal trial. It is also a departure with a cost the accused feels directly, since cross-examination of a witness whose identity is concealed is cross-examination with one hand tied, and the balance Section 17 strikes, secrecy where life is in danger, reasons recorded, contravention punishable, is a balance the courts have to police in every case in which it is invoked.

The Special Courts: a court system for scheduled offences

The Agency investigates, but the Act also builds the courts that try what the Agency investigates, and the design of those courts is where the Act’s idea of a special offence is most fully worked out. Section 11 empowers the Central Government to constitute Special Courts for such area, or for such case or class of cases, as the notification specifies, and questions about a Special Court’s jurisdiction go to the Central Government for a final decision. The presiding judge is appointed by the Central Government on the recommendation of the Chief Justice of the High Court, on an application by the Agency, and the Chief Justice must recommend a name not later than seven days after the application. Only a person who is a Sessions Judge or Additional Sessions Judge immediately before appointment may be appointed, so the specialness of the court is in its jurisdiction and procedure rather than in its personnel: it is staffed from the ordinary higher judiciary, then given a special docket.

Section 13 gives every scheduled offence investigated by the Agency to the Special Court within whose local jurisdiction it was committed, notwithstanding anything in the Code, and provides for transfer of cases between Special Courts by the Supreme Court or the High Court where a fair, impartial or speedy trial cannot otherwise be had, where trial would risk a breach of the peace or endanger the safety of the accused, witnesses, the Public Prosecutor or the judge, or where the interests of justice otherwise require. Section 14 lets a Special Court try connected offences alongside the scheduled offence, and to convict for any other offence disclosed in the course of trial under this Act or any other law. Section 16 gives the Special Court the powers of a Court of Session, permits it to take cognizance without the accused being committed to it, and provides a summary procedure for scheduled offences carrying up to three years’ imprisonment, with a maximum sentence of one year and a fine of up to five lakh rupees in such summary trials. Section 17 protects witnesses: proceedings may be held in camera, the identity and address of a witness whose life is in danger may be kept secret, and contravention of a direction under the section is punishable with imprisonment up to three years and a fine of one thousand rupees.

Two procedural features deserve to be understood precisely because they are often described loosely. Section 16(5) permits a Special Court, for reasons to be recorded, and subject to the provisions of Section 299 of the Code, to proceed with the trial in the absence of the accused or the pleader and to record the evidence of a witness, preserving the accused’s right to recall the witness for cross-examination. It is a narrow provision for evidence preservation, not a general warrant for trial in absentia, and it operates within the discipline of the Code’s own rule on proof in the absence of the accused. Section 19 requires the trial by a Special Court to be held day to day on all working days, to take precedence over any other trial of the accused in any other court, and to be concluded in preference to such other trial, so that the parallel case, if necessary, remains in abeyance. The statute’s theory of speed is therefore not only about the Agency but about the court: the scheduled offence displaces everything else on the judge’s board. Section 20 provides the safety valve, permitting a Special Court which finds the offence before it not triable by it to transfer the case to a court of competent jurisdiction.

The appeal structure is in Section 21, and it is one of the Act’s most distinctive features. An appeal lies from any judgment, sentence or order of a Special Court, not being an interlocutory order, to the High Court, on both facts and law, and every such appeal must be heard by a bench of two judges, which is to dispose of it, as far as possible, within three months of admission. The appeal must be preferred within thirty days, which the High Court may extend for sufficient cause, but no appeal may be entertained after ninety days. Section 21(4) carves out the one interlocutory exception: an appeal lies against an order of the Special Court granting or refusing bail, notwithstanding the ordinary bar on interlocutory appeals. The Supreme Court settled the scope of the bail exception in State of Andhra Pradesh v. Mohd. Hussain in 2013, holding that bail orders are appealable and that such appeals must be heard by a bench of two judges, a ruling that rests on Parliament’s deliberate decision to treat scheduled offences as matters of exceptional gravity in the appeal forum as well as at trial.

Section 22 extends the same court architecture to the states. A State Government may constitute Special Courts for offences under the scheduled enactments, with the Chapter’s provisions applying subject to modifications: references to the Central Government in Sections 11 and 15 read as references to the State Government, the reference to the Agency in Section 13(1) reads as a reference to the state investigating agency, and the reference to the Attorney-General reads as a reference to the Advocate-General. Until such a court is constituted, the Sessions Court of the division exercises the jurisdiction. This is the statutory basis for the practice reported from several states of trying scheduled offences investigated by state police before designated sessions courts, a practice the High Court of Jammu and Kashmir and Ladakh addressed in Mohammad Ayoub Dar v. State of J and K, holding that scheduled offences investigated by a state agency are tried by the Special Court, and that a bail order passed in such a matter is appealable under Section 21.

Where are offences committed outside India tried?

In New Delhi. The 2019 amendment empowers the Central Government to direct the Agency to register and investigate a scheduled offence committed outside India as if it had been committed in India, and it confers jurisdiction for such cases on the Special Court at New Delhi. The forum is fixed by the Act, not by where the offence occurred.

The two central agencies compared

The contrast that gives the profile its namable claim is between the two principal central investigating agencies: the National Investigation Agency and the Central Bureau of Investigation. The table sets them side by side on the seven points the brief identifies. Every row is drawn from the statutes, not from the politics that usually frames the comparison.

Point of comparison	National Investigation Agency	Central Bureau of Investigation
Statutory basis	Constituted under Section 3 of the National Investigation Agency Act, 2008, a special statute for investigation and prosecution	Created by a 1963 executive resolution under the Delhi Special Police Establishment Act, 1946, originally for corruption
Subject matter	Only scheduled offences under the Acts listed in the Schedule, a closed list amendable by Parliament	Offences notified under the DSPE Act, exercised case by case, predominantly corruption and economic offences
State consent	None required. A direction under Section 6(4) or 6(5) displaces state investigation without any consent requirement	General consent of the state required under Section 6 of the DSPE Act, which states may withdraw and several have withdrawn
Who directs investigation	The Central Government alone, on a fit case assessment under Section 6(3), or suo motu under Section 6(5)	The Central Government within the consent a state has given; constitutional courts may also direct a CBI investigation
Court structure	Special Courts under Sections 11 and 22 with a designated judge, day to day trial under Section 19, and protected witnesses under Section 17	Ordinary courts, with special CBI courts for notified offences under the ordinary criminal law
Appeal route	To the High Court on facts and law under Section 21, before a bench of two judges, with bail orders appealable under Section 21(4)	The ordinary appellate hierarchy of the Code of Criminal Procedure
Extraterritorial position	Since 2019, scheduled offences committed outside India against Indian citizens or Indian interests may be investigated, with trial at New Delhi	No equivalent statutory mandate for investigation abroad

The comparison in prose is worth making once more, because the table compresses a history the prose should carry. The Central Bureau of Investigation is the older institution and the more familiar one, and its powers have been litigated and settled for decades. It is born of the Special Police Establishment of 1941, formalised by the 1946 Act, and created in its present form by a Home Ministry resolution of 1 April 1963. Under Section 5 of the 1946 Act the Centre may extend the Establishment’s powers to areas outside Delhi, but Section 6 forbids that extension in a state without that state’s consent. The consent need not be case specific: states grant general consent to the CBI, and general consent is what states have withdrawn, a power exercised by governments of several states including West Bengal, Chhattisgarh, Punjab, Rajasthan, Kerala, Jharkhand, Maharashtra and Meghalaya, some of them subsequently restoring it. The practical effect of withdrawal is that the CBI cannot register a fresh case within that state without the state government’s permission for that case, while continuing to investigate cases already registered under consent given earlier. Courts may still entrust investigations to the CBI regardless of consent, a judge-made exception the Supreme Court has repeatedly affirmed.

The National Investigation Agency has no equivalent regime because the Act creates none. No provision of the NIA Act asks for a state’s consent, general or specific, and no provision permits a state to withhold it. That is why the political argument about consent withdrawal, one of the most common confusions in public discussion of central agencies, applies to exactly one of the two agencies. A state that has withdrawn general consent for the CBI has taken no legal step at all against the NIA, because there is no consent for it to withdraw. The Agency is, in this precise sense, the agency the Constitution’s drafters would have found most surprising: a police force that is constitutionally anomalous and statutorily self-sufficient, needing nobody’s permission and required to give nobody a justification beyond the Central Government’s satisfaction that a scheduled offence has been committed and that the case is a fit one for it.

The judge the Agency asks for: an appointment mechanism under examination

One provision of the Special Court chapter deserves separate notice because it inverts the ordinary relationship between investigator and court. Under Section 11, when a Special Court is to be constituted, the judge is appointed by the Central Government on the recommendation of the Chief Justice of the High Court, and Section 11(4) permits the Agency to make the application to the Chief Justice for the appointment. The Chief Justice must recommend a name within seven days of the application. The investigator, in other words, is the statutory applicant who sets the appointment of its own trial judge in motion, and the constitutional head of the state’s judiciary is put on a seven-day clock to respond. The design is defensible on its own terms: the Agency knows when a case is coming and how large it will be, the pool of eligible judges is confined by the section to serving Sessions Judges and Additional Sessions Judges, and the choice belongs to the Chief Justice, not to the applicant, who merely knocks on the door. No allegation of impropriety in the working of the provision is part of the public record of the Act, and the judges who have presided over the Agency’s major trials have, as the Samjhauta acquittal shows, not behaved as appointees of the prosecution.

The provision nevertheless repays the attention of anyone interested in how the Act distributes institutional dignity. In the ordinary system, the constitution of courts is a function of the state acting through its judicial administration, and an investigating agency has no more standing to initiate the appointment of the judge who will try its cases than it has to select its prosecutor’s opponent. The Act normalises the Agency’s initiative because its courts are not standing institutions with continuing dockets but case-specific creations, constituted for an area or for a case or class of cases as the notification specifies, and somebody must tell the Chief Justice that a court is needed. The 2019 amendment’s shift to designation of existing courts of session was, in part, a retreat from that bespoke model toward the ordinary judiciary: a designated sessions court is an existing institution wearing an additional hat, its judge already appointed through the ordinary channel, and the Agency’s application becomes unnecessary. The trajectory of the provision, from constituted courts on the Agency’s application to designated courts from the ordinary cadre, is the one place where the Act’s design has moved toward the normal rather than away from it, and it moved there for the least dramatic of reasons, the shortage of courts, which suggests that the institution’s practical evolution is being driven less by constitutional theory than by the arithmetic of case load, as the evolution of most institutions is.

The record in court: the Samjhauta case as the machinery in operation

An institution profile that stopped at the statute would miss the point of the profile form, which is to test the design against the thing designed. The Samjhauta Express case is the most instructive single example because it passed through every part of the machinery the Act built. The bombing of the Delhi to Lahore train near Panipat on the night of 18 February 2007 killed 68 people, most of them Pakistani passengers. The case was investigated first by the Haryana police. In July 2010 the Central Government directed the Agency to take it over, an exercise of the direction power in a case three years old, and the Agency filed its charge sheet in 2011 against eight accused, of whom four stood trial before the Special Court at Panchkula, three having been declared proclaimed offenders and the alleged mastermind, Sunil Joshi, having been murdered in December 2007. The trial, in a system whose statute commands day to day hearing and precedence over every other case, concluded on 20 March 2019, when the Special Court acquitted all four accused for want of credible and admissible evidence. More than 250 witnesses had been examined; many had turned hostile; witnesses in Pakistan did not appear despite summonses routed through the Ministry of External Affairs.

Every element of that history maps onto a provision of the Act, which is why it teaches more than a dozen aggregate statistics. The takeover maps onto Section 6, and shows the direction used not for a fresh atrocity but for a stalled investigation the Centre judged, years later, to need a national investigator. The trial maps onto Sections 11 and 19, and shows the distance between the statute’s command of day to day trial and a trial that occupied the better part of a decade, a distance explained by the ordinary realities of a multi-accused prosecution: witnesses scattered across two countries, accused produced from custody, judges transferred. The acquittal maps onto the limit of the entire design: a special agency and a special court can guarantee a forum and a process, but they cannot manufacture evidence, and where witnesses resile and foreign witnesses stay away, the prosecution fails on the ordinary law of proof like any other. The pattern was not confined to Panchkula: the principal accused acquitted there had already been acquitted in the Mecca Masjid and Ajmer Dargah blast cases, prosecutions in which the Agency’s investigations had likewise ended without convictions.

The political argument about the Agency’s record is conducted, predictably, by selecting from this material. The government’s selection is aggregation: the Home Minister’s figures placed before the Rajya Sabha in July 2019, of 272 cases registered, charge sheets in 199, and convictions in 46 of 51 completed prosecutions, describe an institution with a high conversion rate in the cases it finishes. The critics’ selection is the acquittals: cases in which the Agency, having taken over investigations amid public controversy about their direction, failed to prove its charge sheet. Both selections are accurate and neither is complete, because an aggregate rate says nothing about the cases that dominate public argument, and the famous acquittals say nothing about the hundreds of cases that proceed without controversy. What the record does establish, and what matters for the constitutional debate, is narrower and more useful: the Special Courts have not functioned as conviction machinery. They have acquitted in the Agency’s most visible cases where the evidence failed, applied the bail appeal discipline of Section 21, and transferred and managed cases under the Act’s own provisions. A court system that acquits the Agency’s accused in the Agency’s most politically charged prosecution is evidence, of the strongest practical kind, that the judicial half of the 2008 design retains an independence the legislative half is accused of lacking.

The record also bears on the design choice examined at the start of this profile. The Samjhauta takeover was controversial not because the Agency lacked power, nobody suggested the direction was unauthorised, but because the power existed at all in a form that let the Centre redirect a three-year-old state investigation on its own assessment. The objection in that controversy was never really to the investigation’s quality; the acquittal turned on evidence, not on federalism. It was an objection to the allocation of the choice, which is the objection the Chhattisgarh suit later dressed in constitutional language. The history of the Agency in operation therefore confirms the thesis with which the profile began: the disputes the institution generates are disputes about consent and choice, conducted in the vocabulary of misuse or of competence, but rooted in the single legislative decision of December 2008 to build a central investigator that asks nobody’s leave.

One trial law for two investigators: the quiet reach of Section 22

The provision of the Act with the widest practical reach is the one least discussed in the federalism argument, because it operates without any direction, any takeover or any controversy. Section 22 authorises a State Government to constitute Special Courts for the trial of offences under the scheduled enactments, and applies the Act’s Special Court chapter to those courts with the references to the Central Government and the Agency read as references to the State Government and its investigating agency. The consequence is that a scheduled offence investigated from first to last by a state police force, in a case the Centre has never touched, may nevertheless be tried under the procedural code of a central statute: before a designated court, with day to day trial in principle, with witness protection measures, and with appeals governed by Section 21. Until a state designates such a court, the Sessions Court of the division exercises the jurisdiction with the same powers and procedure.

The holding of the High Court of Jammu and Kashmir and Ladakh in Mohammad Ayoub Dar v. State of J and K shows the provision at work on a question of pure procedure. A bail order in a case involving scheduled offences investigated by the state had been treated as an order of an ordinary court; the High Court held that, by force of Section 22, the reference to the Agency in Section 13 must be read as a reference to the state investigating agency, that every scheduled offence investigated even by a state agency is triable only by a Special Court, that the Sessions Court acts as such a court where none has been constituted, and that the bail order was therefore appealable under Section 21 and had to be heard by a bench of two judges. A provision dismissed as machinery had determined the forum, the appeal route and the composition of the bench in an ordinary bail dispute. That is the deeper sense in which the Act of 2008 reordered Indian criminal federalism: not only by creating an agency that can displace state investigation, but by supplying a single trial law for the whole category of scheduled offences, binding on state courts trying state cases, enacted by a Parliament whose competence to legislate the procedure of state trials of state offences rests on the same contested foundation as the Agency itself.

The design has an administrative logic that should be acknowledged before it is criticised. Scheduled offences are grave, their trials are long, their witnesses are vulnerable and their accused are frequently tried for connected offences in several districts at once. A uniform procedural home for such trials, with judges designated for the docket and appeals channelled to a division bench, serves the interests of victims, witnesses and accused alike, and the states have used Section 22 for their own purposes without waiting for the Centre: the designated court trying a state anti-terrorism squad’s case is a state institution exercising a state function under a borrowed procedure. The borrowing is the point at which the competence argument and the administrative argument meet and fail to engage. Administratively, Section 22 is the Act’s most successful provision, used constantly and challenged rarely. Constitutionally, it is the provision that most plainly legislates for the state sphere, since it governs trials in which the Union may have no interest and no involvement. If the Supreme Court sustains the Act, Section 22 will be the quiet reason the judgment matters to every state trial court in the country. If the Court strikes the Act down, Section 22 will be the provision whose disappearance state governments, whatever their politics, will most immediately have to replace.

The caseload in practice: which cases the direction actually reaches

The Act’s design can also be read from the pattern of its use, because the direction power, exercised over a decade and a half, has sorted cases into recognisable classes, and the classes reveal the working meaning of gravity and fitness better than the statute’s silence. The first class is the fresh atrocity with an obvious cross-border or inter-state dimension: the attack whose conspiracy plainly exceeds one state’s territory, where the direction follows within days and the argument for a national investigator makes itself. The second class is the stalled or sensitive investigation taken over late, of which the Samjhauta case is the type specimen: an investigation begun by a state force, overtaken by controversy or by its own difficulties, and transferred to the Agency months or years after the offence, so that the Agency’s task is as much the reconstruction of a cold file as the pursuit of a live one. The third class is generated by the Agency itself: a scheduled offence investigation that uncovers, through Section 8’s connected offence power, a network of offences across several states, financing, recruitment, logistics, each of which becomes an Agency case in its own right, so that the Agency’s docket grows by chains rather than by single directions. The fourth class is the requested handover, in which a state government, facing an investigation whose ramifications it cannot follow beyond its borders or whose local sensitivities it prefers to place at a distance, forwards the case and welcomes the direction that the statute does not require it to welcome.

Against those classes stands the great mass of scheduled offences that never reach the Agency at all. Every offence under the Unlawful Activities (Prevention) Act in every state is a scheduled offence, and such offences are registered, investigated and prosecuted by state police forces and their anti-terrorism squads as the ordinary business of state policing. The Agency’s registered cases, counted in the hundreds across its whole existence at the time of the government’s 2019 statement to Parliament, are a small fraction of the scheduled offences committed in the same period. The direction is therefore a filter of extreme narrowness in practice, whatever its width in law, and the filter’s mesh is set by considerations the Act does not enumerate: the scale of the attack, the presence of a foreign hand, the number of states a conspiracy touches, the failure or compromise of the local investigation, and, critics add, the political salience of the case. A reader who wants to understand the institution should hold both truths at once: the legal power is as wide as its critics say, and the exercised power is as narrow as its defenders say, and the distance between the two is filled by executive choice that no statute governs.

That distance is the practical form of the accountability question. A power exercised rarely and selectively is a power whose selections demand justification, and the Act provides no forum in which the justification must be given: not to the state displaced, not to the state refused, not to Parliament in any statutory report, and not to the public. The justification emerges, when it emerges, in the political process, in parliamentary questions and press conferences after spectacular cases, which is to say it emerges for the cases that least need it and not for the marginal direction or the marginal refusal where the discretion actually bites. The states that framed the federalism objection as a complaint about competence were also, in part, complaining about this: a discretion of that consequence exercised without a duty to explain is, in a federation, a standing irritant even when its every exercise is defensible, because the coordinate government that bears the constitutional responsibility for police learns of its own displacement as a recipient of orders rather than as a partner in a decision.

The defenders’ answer to the practical complaint returns to the character of the offences. Terrorist conspiracies are not ordinary crime with larger casualties; they are operations directed, financed or inspired from outside the state’s territory and frequently outside the country’s, and their investigation is inseparable from intelligence holdings that are themselves centralised. A consent regime for such investigation would not produce partnership; it would produce the 2008 position with better manners, in which the only investigator with the whole picture is the intelligence apparatus that cannot arrest, charge or prosecute anyone. The Agency exists to give the intelligence picture a police arm. That is the functional core of the institution, and it survives every argument about its manners: the question the federation has never answered is not whether such an arm should exist, but on what terms of consultation and explanation it should act inside states whose constitutional province it enters. The Act answers the first question completely and the second not at all, and the docket of the Supreme Court, where the Chhattisgarh suit waits, is where the unanswered half of the design has gone to be argued.

The 2019 amendment: a longer list and a longer reach

The National Investigation Agency (Amendment) Act, 2019, Act 16 of 2019, which received assent on 24 July 2019 and came into force on 2 August 2019, changed the Act in three directions: it lengthened the Schedule, it lengthened the Agency’s territorial reach, and it eased the constitution of Special Courts. The Schedule changes restructured the Penal Code entry and added three new territories of offence. Serial number 1 of the Schedule, which since 2008 had been the Atomic Energy Act, 1962, was renumbered to make room for the Explosive Substances Act, 1908, placed first, with the Atomic Energy Act following as serial 1A. The reference to the Anti-Hijacking Act was updated from the Act of 1982 to its successor, the Anti-Hijacking Act, 2016. Most significantly, entry 8, which had contained Chapter VI of the Penal Code and the counterfeit currency sections, was replaced by a longer entry: Chapter VI at (a); sections 370 and 370A, human trafficking, at (b); the counterfeit currency sections 489A to 489E, retained from the original Schedule, at (c); sub-section (1AA) of section 25 of the Arms Act, 1959, the manufacture or sale of prohibited arms, at (d); and section 66F of the Information Technology Act, 2000, cyber-terrorism, at (e).

A precision belongs here, because accounts of the amendment, including several official summaries, describe all five of the 2019 entries as additions. Four of them are. The counterfeit currency offences are not: sections 489A to 489E stood in the Schedule from the day the Act was passed in 2008, and the amendment carried them from entry 8(b) to entry 8(c) when trafficking was inserted ahead of them. The genuinely new territories of 2019 are human trafficking, prohibited arms, cyber-terrorism and the Explosive Substances Act. The distinction matters to the argument about the Schedule’s growth, because it shows the list has not expanded as dramatically as the summaries suggest, while confirming that its expansion, where it has occurred, has moved into offences with a large ordinary-crime footprint, trafficking being the clearest example.

The territorial change is in Sections 1, 3 and 6. A new clause (d) in Section 1(2) extends the Act to persons who commit a scheduled offence beyond India against Indian citizens or affecting the interest of India. Section 3(2) extends the powers of Agency officers outside India, subject to any international treaty or to the domestic law of the country concerned. A new Section 6(8) empowers the Central Government, where a scheduled offence has been committed at any place outside India to which the Act extends, to direct the Agency to register the case and take up investigation as if the offence had been committed in India, and Section 6(9) gives the Special Court at New Delhi jurisdiction over such cases. The design is careful: the Agency’s foreign reach is declaratory and treaty-bound rather than operational, since an Agency officer in a foreign country acts subject to that country’s law and to agreements the Union has made, and the trial is brought home to Delhi.

The court change is in the language of Sections 2, 11 and 22: the Special Court is no longer described as constituted but as a Court of Session designated as a Special Court, and the Central and State Governments are empowered to designate one or more courts of session as Special Courts, in consultation with the Chief Justice of the High Court. The point of the change was capacity: the dedicated-court model had produced too few courts for the case load, and designation lets an existing sessions court take the scheduled-offence docket without a fresh institution being built for it.

The federalism objection, at full strength

Against all of this stands the objection that the Constitution never authorised it, and the objection deserves to be put at the strength its authors put it. The Seventh Schedule gives “Public order” to the states as Entry 1 of the State List, subject to exceptions for the deployment of Union forces, and “Police” to the states as Entry 2, subject to a provision for the extension of a state’s police beyond its borders in defined circumstances. Entries 1 and 2 together constitute the constitutional assignment of internal security to the states, and the architecture of the Seventh Schedule lists is the framework within which any central police statute must justify itself. The State of Chhattisgarh, in the original suit it filed in the Supreme Court under Article 131 in January 2020, put the objection in its strongest form. Its plaint argued that the Act is beyond the legislative competence of Parliament, because it empowers the Centre to create an agency for investigation when investigation is carried out by the police, a subject in the State List; that no entry in the Union List confers a power of police or investigation of the kind the Act creates; that the Act is therefore a colourable exercise of legislative power and an encroachment on the federal structure; and that once the Act’s machinery is set in motion it completely displaces the state’s power to investigate scheduled offences committed within its own territory, without consent, without coordination and without any rules or guidelines governing the Centre’s choice of which cases to take.

That is a serious argument and it should not be caricatured. The text of the Act does give the Centre what the plaint calls unfettered discretion in one precise respect: Section 6 supplies no criteria beyond gravity and fitness, no requirement of reasons addressed to the state, no consultation, and no review mechanism short of constitutional challenge to a direction itself. The federalism objection is not a complaint about the personality of any government. It is the claim that the Constitution distributes the police power deliberately, that the distribution was a considered choice of the Constituent Assembly against the background of a unitary colonial police tradition, and that a Parliament which cannot amend that distribution by ordinary legislation should not be able to outflank it by scheduling offences and appointing its own investigators. The same objection explains the decision to sue under Article 131, the Supreme Court’s original jurisdiction over disputes between the Union and a state: Chhattisgarh framed the question as one about the existence and extent of a legal right of the state, not about the wisdom of any single investigation.

The Union’s answer, at full strength

The Union’s answer begins by denying the premise that the Act legislates on police or public order at all. On the doctrine of pith and substance, the Union argues, the Act is in substance legislation on the subjects named in its preamble, and those subjects sit in the Union List: the defence of India under Entry 1; the Central Bureau of Intelligence and Investigation under Entry 8, an entry which the Union says shows that the Constitution itself contemplates central investigation; foreign affairs and the implementation of treaties, agreements and conventions under Entries 10 to 14 and Article 253; and the residuary power under Entry 97 for anything that remains. Terrorism directed from abroad, attacks on the sovereignty and integrity of India, offences under convention statutes, are not, on this account, public order matters with a federal flavour; they are defence and external security matters which a state police force encounters only at the point of impact. The Act’s encroachment on the State List, if any, is incidental, and incidental encroachment does not vitiate a law whose pith and substance is a Union subject.

The second strand of the Union’s answer is functional, and it was the strand that carried the winter session of 2008. A conspiracy that is financed in one country, planned in a second, supplied through a third state and executed in a fourth leaves each state police force holding one fragment of evidence, with no force holding the whole. Coordination by conference cannot give a single investigator command of the whole case, and the states’ consent regime of the CBI model demonstrates the failure mode: an investigation that depends on the continuing goodwill of every government whose territory it touches is an investigation that can be switched off by politics. The Union points to the Act’s internal restraints as the answer to the arbitrariness charge: the state does the initial investigation, the report passes through the state government itself, the Central Government must decide within fifteen days, and the Agency’s mandate is confined to offences Parliament has scheduled after debate. Whether those restraints answer the competence question is the very thing in issue, because competence and wisdom are separate questions, and a court deciding the Chhattisgarh suit must decide the first before it reaches the second. The wider framework of Centre and state competence, including the doctrine that governs conflicts between Union and state laws, is set out in the discussion of Article 254 and repugnancy, and the Chhattisgarh suit will be decided within that framework when it is decided. This profile takes no position on the outcome, because the question is pending before the Court and the arguments on both sides are honestly arguable; what can be said is that the Act was drafted, conspicuously and self-consciously, to make the Union’s argument as strong as the materials allowed, by tying its Schedule to Union subjects and its preamble to the language of sovereignty and treaties.

Pith and substance: how the competence question will actually be decided

Because the suit is pending, the honest way to present the legal question is as a method rather than a forecast. A competence challenge of this kind is decided by the doctrine of pith and substance: the court identifies the true nature and character of the impugned legislation, asks which list that character belongs to, and treats encroachment on another list as fatal only if the encroachment is the substance of the law rather than an incidental effect of legislating on a permitted subject. The Chhattisgarh plaint is written to that method. It argues that the true character of the Act is the creation of a police force, because everything the Agency does, recording, investigating, arresting, searching, seizing, prosecuting, is the daily work of police, and that dressing that work in a preamble about sovereignty does not change its character. The Union’s answer is written to the same method from the other direction: the true character of the Act, it says, is the defence of the Union against a defined class of offences, offences which Parliament selects precisely because they are attacks on the Union, on its treaty obligations or on the security of the State, and the investigative machinery is merely the means by which legislation on those subjects is made effective, as any legislation must be.

Two features of the constitutional text will matter to that argument and are worth the reader’s attention. The first is Entry 2A of the Union List, inserted by the Forty-second Amendment, which gives Parliament power over the deployment of armed forces of the Union in aid of the civil power of a state. The entry shows that the Constitution does contemplate Union coercive presence inside a state, but contemplates it as deployment in aid of the state’s power, under a scheme in which the state’s civil authority remains the principal. The Agency is not deployed in aid of a state force; on direction, it replaces the state force. The second feature is Article 355, the Union’s duty to protect every state against external aggression and internal disturbance and to ensure that the government of the state is carried on in accordance with the Constitution. Defenders of the Act have invoked the duty as the constitutional atmosphere in which the statute sits; the plaint’s implicit answer is that a duty to protect is not an entry of legislative competence, and that Article 355 has its own machinery, culminating in Article 356, which the Act does not use. Neither feature decides the case. Together they show why the case is genuinely difficult: the Constitution both reserves police to the states and charges the Union with the security of the whole, and the Act of 2008 is the point at which those two commitments collide in statutory form.

There is also a comparative quietness in the record that both sides will use. The Act has operated since the end of 2008. Directions have been given in cases arising in states governed by every major political formation, and investigations have been handed over by state police forces on many occasions without a recorded instance of a state refusing to transmit records under Section 6(6) or purporting to withhold a cooperation the Act does not ask it to grant. The Union will read that acquiescence as the federal system working: states that need the Agency use it, and the override exists for the exceptional case. Chhattisgarh will read the same record as proof of the plaint’s premise, that a state has no choice in the matter and therefore its compliance proves nothing about consent. The record, like the record of convictions and acquittals, thus supports the structural reading each side brought to it, and the question returns, as it must, to the text of the Seventh Schedule and the character of the Act. That is the question the Supreme Court has been asked, and it remains the Court’s to answer.

The second objection: a Schedule that keeps growing

The competence argument would be unchanged if the Schedule had never been amended, and that is why the second objection must be kept distinct from it. The second objection accepts, for argument, that Parliament may schedule offences and complains instead about what it has scheduled. Human trafficking, prohibited arms manufacture and counterfeit currency are offences with inter-state and international dimensions, but they are also offences of ordinary criminal life, investigated daily by state police forces in enormous numbers. Adding them to the Schedule does not put the Agency into any of those cases by itself; the direction under Section 6 is still required, and the Central Government must still find gravity and fitness. But it changes the class of cases in which a takeover is legally possible, and critics of the 2019 amendment, including members who spoke against it in Parliament, argued that the direction is a political instrument whose use will expand as the list expands, converting a counter-terror agency by stages into a general federal investigative reserve. Defenders of the amendment, led in the Rajya Sabha by Home Minister Amit Shah, answered on the record of the Agency itself: the government placed figures before the House showing that, as on 30 June 2019, of 272 cases registered by the Agency, charge sheets had been filed in 199 and convictions obtained in 46 of the 51 prosecutions completed, and argued that a force with that record should be trusted with a wider list rather than confined to a narrower one. The two arguments do not meet, because one is about institutional design and the other about institutional record, and the design argument is the one that survives changes of government.

The allegation that hangs over both objections, that central agencies are deployed against opposition-governed states, is addressed here only at the level the statute permits it to be addressed. The Act’s triggers are knowable and they are few: a scheduled offence, a report or the Centre’s own information, a fifteen-day determination, a direction. The discretion inside those triggers is real and it is the Centre’s, and the Act gives a state no procedural purchase on it: no notice, no hearing, no reasons addressed to the state, no appeal. A reader who understands that structure can assess any particular allegation of misuse by asking the only questions the statute makes relevant: was the offence scheduled, was the direction given, and did the Agency stay inside the connected-offence limits of Section 8? The statute neither proves the allegation false nor makes it true. It defines the space in which the Centre’s choice operates, and the absence of any state-facing safeguard inside that space is a legislative choice that Parliament made in 2008, confirmed in 2019, and has not revisited.

The objection in politics: the National Counter Terrorism Centre episode

The federalism objection has a political history as well as a forensic one, and the episode that reveals it most clearly is the failure of the National Counter Terrorism Centre. In February 2012 the Home Ministry issued an order constituting the Centre as a counter-terrorism body located within the Intelligence Bureau, with an operations division empowered, by reference to Section 43A of the Unlawful Activities (Prevention) Act, to arrest, search and seize anywhere in India. The proposal was attacked at once by chief ministers across the political spectrum, including Mamata Banerjee of West Bengal, J. Jayalalithaa of Tamil Nadu, Naveen Patnaik of Odisha and Narendra Modi of Gujarat, whose objection, expressed in writing to the Prime Minister, was that the Centre had arrogated operational police powers without consultation and in disregard of the federal structure. A meeting of chief ministers in May 2012 did not reconcile the opponents, and the proposal was kept in abeyance from 2013, unimplemented; a parliamentary committee later recommended that the idea be revived, without effect.

The episode is the control experiment for the argument of this profile. The NCTC would have rested on an executive order and a borrowed power under an existing statute. The National Investigation Agency rests on an Act of Parliament with its own Schedule, its own direction mechanism and its own courts. The first was stopped by the states acting politically; the second has never been stopped by anything, because the states possess no political instrument against it: no consent to withhold, no resolution that binds it, no funding they control. The chief ministers who defeated the NCTC included several whose governments lived, throughout the same years, with Agency investigations in their states under directions they had no power to refuse. The contrast explains why the constitutional challenge, when it came in 2020, came from a state government through Article 131 and in the language of competence rather than through the conference diplomacy that had defeated the NCTC: for a statute, the only forum in which a state can contest the allocation of power is a court, and the only language available there is the Seventh Schedule.

The episode also disciplines the easy claim that the federalism objection is a partisan artefact. The opponents of the NCTC were led by chief ministers of parties in opposition to the government of the day, but the objection they made, that operational policing by the Centre without state consultation offends the constitutional scheme, was the same objection the Chhattisgarh plaint later made against the Act, and it has been made at different times by governments of every complexion, including governments whose parties had supported the Act in Parliament. Positions on central investigation in India track incumbency with embarrassing regularity: parties out of power at the Centre discover the State List, and parties in power at the Centre discover Entry 8. The statute book is indifferent to the rotation. The Act of 2008 was a Congress-led government’s statute, challenged in 2020 by a Congress-led state government, and defended in the 2019 amendment debates by a Bharatiya Janata Party Home Minister whose party’s most prominent state leader had, as a chief minister, led the revolt against the NCTC. A profile that attributed the institution to any party’s design would misdescribe it; the design belongs to the Indian state’s recurring answer to cross-border violence, and the objection belongs to the Constitution’s recurring answer to the Indian state.

Three questions the design leaves open

The first open question is the one before the Supreme Court, and it is genuinely open. Competence litigation of this scale is rare in the field of criminal law, and the Chhattisgarh suit has no close precedent: the Court has considered the distribution of police power in the context of particular investigations and particular directions to the Central Bureau of Investigation, but it has never been asked by a state, in an original suit, whether Parliament may create a standing central police agency at all. The answer will matter beyond the Agency, because the reasoning that sustains or condemns the Act will also govern the space available to every future central investigative body Parliament may design, and because the entries in issue, police on one side and the Central Bureau of Intelligence and Investigation on the other, are the permanent furniture of the Seventh Schedule. A judgment for the Union would confirm that investigation follows the subject of the offence rather than the territory of its commission, a principle with consequences for economic offences, environmental crime and every other field where Parliament legislates across state lines. A judgment for Chhattisgarh would not merely wound one agency; it would require the Agency’s caseload to be returned to state forces or re-founded on state consent, and would stand as the most consequential federalism decision in the field of criminal law since the Constitution came into force.

The second open question is one of discipline rather than validity: what principle, if any, governs the growth of the Schedule? The original list had an intelligible unity, convention offences and offences against the State, and the 2019 additions broke that unity in a deliberate direction, toward offences of inter-state organisation: trafficking networks, currency rackets, arms manufacture, cyber attack. A principle capable of explaining both lists might be that the Agency exists for offences whose investigation characteristically exceeds the capacity of a single state, by geography, by organisation or by the identity of the target. Such a principle would justify the trafficking entry and would equally justify entries Parliament has not made, in organised crime generally, in narcotics, in inter-state fraud. Nothing in the Act states the principle, and the absence of a stated principle means the Schedule’s future shape is a pure question of parliamentary politics, debated afresh at each amendment with no settled test to argue from. The history of the list so far, one enlargement in a decade, suggests restraint in practice; the absence of any legal restraint means the restraint is conventional, and conventions in this field have a short half-life when a spectacular crime rearranges the politics of a session, as the session of December 2008 demonstrated.

The third open question is accountability, and it is the question the institution’s defenders answer least well. The Agency’s chief is appointed without a statutory committee, its officers are protected by a sanction requirement controlled by the government that directs it, its governing Act imposes no reporting duty to Parliament, and the decision that defines its work, the direction, is unreasoned to the state it displaces and unpublished to the public it serves. Each of those features has an operational defence, and the defences are not frivolous: speed, secrecy and the protection of officers are real goods in counter-terrorism. But the accumulation of the features produces an institution that is accountable, in the end, only to the executive that wields it and to the courts that test its finished product in individual cases. The CBI’s history shows where that road leads: an agency whose legal charter is older and weaker became, over decades, the subject of a standing public suspicion that its cases follow the politics of its masters, a suspicion the Supreme Court voiced in celebrated language when it described the Bureau as a caged parrot speaking in its master’s voice. The National Investigation Agency was designed, consciously, to avoid the CBI’s legal weaknesses: a real statute, a defined list, no consent. It was not designed, with anything like the same consciousness, to avoid the CBI’s accountability weakness, and the longer the Agency operates, the more its legitimacy will depend on the part of the design its framers left bare. That, more than the outcome of any single prosecution, is the measure by which the institution of December 2008 will finally be judged.

Why states comply: the politics of the handover

A puzzle remains in the operation of the Act that neither the competence argument nor the record explains: the handover always happens. No state has refused to transmit records under Section 6(6), no state police force has declined to stop investigating on direction, and no government that denounced the Act in opposition has, on taking office in a state, instructed its police to resist it. Part of the explanation is legal habit: Indian police forces are disciplined hierarchies accustomed to court orders and statutory commands, and a direction under the Act arrives with the full form of law. Part is institutional interest: a state force relieved of a difficult investigation is also relieved of its risks, and the failures of a taken-over case belong thereafter to the Agency. And part is the deeper settlement of Indian federalism in the security field, in which the forms of state primacy are preserved, the state investigates first, the state is heard in the pipeline, the state’s courts try the case, while the substance of the final choice rests with the Union. The Act endures because it was designed to be endurable: it takes from the states only the decision, and leaves them the work, the courts and the credit for both. Whether a settlement of that kind is what the Seventh Schedule intended is the question in the Supreme Court. That it is a settlement, stable across governments of every party for a decade and a half, is the fact with which any judgment will have to reckon.

Conclusion: the anomaly, stated precisely

The National Investigation Agency is best understood not as a national police force but as a national exception to the constitutional rule that investigation belongs to the states: an exception defined by a list of offences, triggered by an executive direction, staffed by officers with all-India police powers, prosecuted before special courts, and defended on the ground that the offences on its list are attacks on the Union rather than breaches of a state’s peace. The exception was created in a fortnight of national emergency in December 2008 and has been enlarged once, in 2019, in each of its three dimensions of offences, territory and courts. It has never required and cannot be denied the consent of any state, which is what separates it from the Central Bureau of Investigation and what makes it, in the strictest sense, a constitutional anomaly that works: anomalous because the Seventh Schedule assigns its function to the states, working because the problem it was built for, the cross-border conspiracy that no single state force can investigate whole, has not gone away. This article is offered as legal information about a statute and an institution, not as advice on any case; the constitutional question the Chhattisgarh suit poses remains before the Supreme Court, undecided, and the Agency investigates on.

Frequently Asked Questions

Q: What is the National Investigation Agency and when was it created?

The National Investigation Agency is a central investigating and prosecuting agency constituted under Section 3 of the National Investigation Agency Act, 2008, Act 34 of 2008. Parliament enacted it in December 2008, within weeks of the Mumbai attacks of November 2008, the President assented to it on 31 December 2008, and it was brought into force on the same day. The Agency investigates and prosecutes scheduled offences: offences under the statutes listed in the Schedule to the Act. Its superintendence vests in the Central Government under Section 4, its administration vests in a Director General appointed by the Central Government, and its officers exercise police powers throughout India.

Q: Which offences can the National Investigation Agency investigate?

Only scheduled offences: offences under the statutes listed in the Schedule to the Act. As enacted in 2008, the Schedule covered the Atomic Energy Act, 1962, the Unlawful Activities (Prevention) Act, 1967, the Anti-Hijacking Act, 1982, the civil aviation and SAARC Convention statutes, the maritime navigation statute of 2002, the Weapons of Mass Destruction Act, 2005, the offences against the State in Chapter VI of the Indian Penal Code, sections 121 to 130, and the counterfeit currency offences in sections 489A to 489E of the Penal Code. The 2019 amendment added the Explosive Substances Act, 1908, human trafficking, prohibited arms manufacture or sale and cyber-terrorism. An offence outside the Schedule is outside the Agency’s mandate altogether.

Q: Can the National Investigation Agency take over a case without state consent?

Yes. Nothing in the Act requires the consent of a state government at any point. The state police report a scheduled offence to the State Government, which forwards the report to the Central Government; the Central Government decides within fifteen days whether the case is fit for the Agency and, if it is, directs the Agency to investigate under Section 6(4). It may also direct an investigation on its own motion under Section 6(5). Once a direction issues, Section 6(6) requires the state police to stop investigating and transmit the records to the Agency. This is the central contrast with the CBI, which works under a consent regime.

Q: What did the 2019 amendment to the NIA Act change?

The National Investigation Agency (Amendment) Act, 2019, Act 16 of 2019, assented to on 24 July 2019 and in force from 2 August 2019, changed the Act in three directions. It expanded the Schedule, adding the Explosive Substances Act, 1908, human trafficking under sections 370 and 370A of the Penal Code, manufacture or sale of prohibited arms under the Arms Act, 1959 and cyber-terrorism under section 66F of the Information Technology Act, 2000. It extended the Act to scheduled offences committed outside India against Indian citizens or Indian interests, with investigation subject to treaties and foreign law and trial before the Special Court at New Delhi. And it allowed governments to designate courts of session as Special Courts.

Q: Do NIA cases go to special courts?

Yes. Section 13 provides that every scheduled offence investigated by the Agency is tried only by the Special Court within whose local jurisdiction it was committed. Special Courts are designated under Section 11 by the Central Government, presided over by a Sessions Judge or Additional Sessions Judge appointed on the recommendation of the Chief Justice of the High Court, and they sit with the powers of a Court of Session. Trials are held day to day under Section 19 and take precedence over other cases against the accused. State governments may designate Special Courts under Section 22 for scheduled offences investigated by state agencies, and appeals from either kind of Special Court go to the High Court under Section 21.

Q: Which state challenged the NIA Act on legislative competence grounds?

Chhattisgarh. In January 2020 the State of Chhattisgarh filed an original suit in the Supreme Court under Article 131 of the Constitution, contending that the Act is beyond the legislative competence of Parliament. Its plaint argued that police is Entry 2 of the State List, that investigation is carried out by the state police, that no Union List entry authorises Parliament to create an investigating agency with overriding powers over state police, and that the Act displaces the state’s power to investigate offences within its own territory without consent, coordination or governing rules. The suit raises the federal structure question directly and awaits decision; separate petitions have also challenged the 2019 amendment.

Q: How does the National Investigation Agency differ from the CBI?

The decisive difference is consent. The National Investigation Agency works under a special statute of 2008, investigates only scheduled offences, needs no state consent, takes cases on a direction of the Central Government, and prosecutes before Special Courts with appeals to a two-judge bench of the High Court. The Central Bureau of Investigation was created by executive resolution in 1963 under the Delhi Special Police Establishment Act, 1946, investigates offences notified under that Act, and cannot operate in a state without that state’s consent under Section 6 of the 1946 Act, consent which states grant generally and may withdraw. The CBI also lacks the Agency’s statutory extraterritorial mandate and its special court structure.

Q: Does the National Investigation Agency have jurisdiction beyond India’s borders?

Since the 2019 amendment, yes, within limits the amendment itself sets. Section 1(2) extends the Act to persons who commit a scheduled offence beyond India against Indian citizens or affecting the interest of India. Section 6(8) empowers the Central Government to direct the Agency to register and investigate such an offence as if it had been committed in India, and Section 3(2) extends Agency officers’ powers outside India subject to any international treaty and to the domestic law of the country concerned. The trial of such a case lies before the Special Court at New Delhi under Section 6(9). Before 2019 the Act reached citizens of India abroad, government servants and persons on Indian ships and aircraft, but conferred no general power to investigate offences committed abroad.

Q: Who heads the National Investigation Agency and who controls it?

Section 4 of the Act answers both halves. The superintendence of the Agency vests in the Central Government, and its administration vests in an officer designated as the Director General, appointed by the Central Government. The Director General exercises, in respect of the Agency, such of the powers exercisable by a Director General of Police in respect of a state police force as the Central Government specifies. No state institution has any part in appointing the Agency’s chief, directing its administration or superintending its work, which is one of the structural features the federalism challenge emphasises: the chain of command runs from the Agency to the Union executive and nowhere else.

Q: How does a case reach the National Investigation Agency after a scheduled offence is reported?

Through the chain Section 6 prescribes. The officer in charge of the police station where information about a scheduled offence is recorded under Section 154 of the Code of Criminal Procedure forwards the report to the State Government forthwith. The State Government forwards it to the Central Government as expeditiously as possible. The Central Government then determines, within fifteen days of receipt, whether the offence is a scheduled offence and whether, having regard to its gravity and other relevant factors, it is a fit case for the Agency. If both determinations are affirmative, Section 6(4) requires the Central Government to direct the Agency to investigate, and the investigation changes hands at that point.

Q: Can the Central Government direct the Agency to investigate on its own initiative?

Yes. Section 6(5) provides that, notwithstanding anything else in the section, if the Central Government is of the opinion that a scheduled offence has been committed which is required to be investigated under the Act, it may, on its own motion, direct the Agency to investigate it. This suo motu power bypasses the reporting chain in Sections 6(1) and 6(2): no state report is needed, no fifteen-day determination governs it, and no consent or consultation with the state is prescribed. It is among the provisions the Chhattisgarh plaint singled out, because the discretion it confers is governed by no rules or criteria stated in the Act beyond the Central Government’s own opinion that the offence is scheduled and requires investigation under the Act.

Q: What must state police do when the Agency is directed to take over an investigation?

Two things, in sequence. Under Section 6(6), once a direction is given, the State Government and every state police officer investigating the offence must not proceed with the investigation and must forthwith transmit the relevant documents and records to the Agency. But Section 6(7) qualifies the stop: until the Agency actually takes up the investigation, the officer in charge of the police station must continue to investigate, so that evidence is not lost in the interval. The Supreme Court applied that qualification in Naser Bin Abu Bakr Yafai v. State of Maharashtra in 2021, holding that a state agency which continued investigating, and filed a charge sheet, before the Agency took over had acted within the scheme of the section.

Q: Can the Agency transfer a case back to a state government?

Yes. Section 7 provides that while investigating an offence under the Act, the Agency, having regard to the gravity of the offence and other relevant factors, may request the State Government to associate itself with the investigation, or may, with the previous approval of the Central Government, transfer the case to the State Government for investigation and trial of the offence. The transfer back is therefore possible but it is the Agency’s choice, made with the Centre’s approval, not a right the state can invoke. Read with Section 6, the position is that movement of a case between the Agency and a state is governed at every point by decisions taken on the Union side of the arrangement.

Q: Can state police investigate a scheduled offence themselves?

Yes, unless and until the Agency is directed to take the case. Section 10 saves the powers of the State Government to investigate and prosecute any scheduled offence or other offences under any law in force, except as the Act otherwise provides. The exception is the direction under Section 6: most scheduled offences are in practice investigated by state police forces, and offences investigated by a state agency are tried before Special Courts designated by the State Government under Section 22, or by the Sessions Court of the division until such a court is designated. The Act creates concurrent competence with a Union override, not an exclusive Union jurisdiction over the offences on its list.

Q: Who conducts the prosecution in cases tried under the Act?

Public Prosecutors appointed under Section 15. The Central Government appoints a Public Prosecutor and may appoint Additional Public Prosecutors, and may appoint a Special Public Prosecutor for a case or a class or group of cases; for Special Courts designated by a State Government under Section 22, the references to the Central Government in Section 15 read as references to the State Government. A person is qualified for appointment only after not less than seven years in practice as an advocate, or seven years in a post under the Union or a state requiring special knowledge of law. Every appointee is deemed a Public Prosecutor within the meaning of the Code of Criminal Procedure, and the Code applies accordingly.

Q: Can a Special Court try offences connected with the scheduled offence?

Yes, under two provisions. Section 14(1) provides that when trying any offence, a Special Court may also try any other offence with which the accused may, under the Code of Criminal Procedure, be charged at the same trial, if the offence is connected with the scheduled offence. Section 14(2) goes further: if in the course of the trial it is found that the accused has committed any other offence under this Act or under any other law, the Special Court may convict the accused of that offence and pass any sentence authorised by this Act or the other law. The provisions mirror Section 8 on the investigation side, which lets the Agency investigate connected offences so that a single case is not split between forces or forums.

Q: How are witnesses protected in trials under the Act?

Section 17 builds the protection into the procedure of the Special Court. Proceedings may be held in camera where the Special Court so desires, for reasons recorded in writing. Where the court is satisfied that the life of a witness is in danger, on the application of the witness or the Public Prosecutor or on its own motion, it may take measures to keep the witness’s identity and address secret: holding proceedings at a place it decides, avoiding mention of names and addresses in orders, judgments and publicly accessible records, directing that identity and address not be disclosed, and ordering that proceedings not be published. Contravention of such a direction is punishable with imprisonment of up to three years and a fine of up to one thousand rupees.

Q: Can a trial proceed in the absence of the accused under the Act?

In a narrow sense, yes. Section 16(5) provides that, notwithstanding anything in the Code of Criminal Procedure but subject to Section 299 of the Code, a Special Court may, if it thinks fit and for reasons to be recorded, proceed with the trial in the absence of the accused or the accused’s pleader and record the evidence of any witness, subject to the right of the accused to recall the witness for cross-examination. The provision is directed at preserving evidence that might otherwise be lost, within the Code’s own discipline for recording evidence in an accused’s absence; it is not a general authorisation of trial in absentia, and the cross-examination right it preserves is the safeguard that defines it.

Q: How are appeals from a Special Court decided under the Act?

Under Section 21, an appeal lies from any judgment, sentence or order of a Special Court, other than an interlocutory order, to the High Court, on both facts and law. Every such appeal is heard by a bench of two judges and is to be disposed of, as far as possible, within three months of admission. The appeal must be filed within thirty days, which the High Court may extend for sufficient cause, but no appeal lies after ninety days. Section 21(4) makes one interlocutory exception: an appeal lies against an order granting or refusing bail, and the Supreme Court held in State of Andhra Pradesh v. Mohd. Hussain in 2013 that bail appeals too must be heard by a bench of two judges.

Q: On which Union List entries does the Union rest Parliament’s power to create the Agency?

The Union’s case, as advanced in defence of the Act, is that the statute in pith and substance legislates on Union subjects rather than on police or public order. It relies principally on Entry 1 of the Union List, defence of India; Entry 8, the Central Bureau of Intelligence and Investigation, which it reads as constitutional recognition of central investigation; the foreign affairs and treaty entries, Entries 10 to 14, with Article 253, which empowers Parliament to legislate for implementing international agreements; and the residuary Entry 97 for anything not covered elsewhere. Chhattisgarh’s suit contests exactly this mapping, arguing that an agency exercising police powers over state territory legislates on Entry 2 of the State List whatever its preamble recites.