The Tribunals Reforms Act 2021 did two things of permanent consequence and one thing that could not last. It abolished a set of appellate tribunals and sent their work to the High Courts and the Commercial Courts, and that part of it stands. It wrote a single code of appointment, tenure and service conditions for the tribunals that remained, and that part re-enacted, in a fresh statute, provisions that the Supreme Court had struck down as unconstitutional one month earlier. In November 2025 the Court struck them down again, and in doing so stated, with a clarity the subject had been waiting a decade for, the rule that separates a lawful legislative answer to a judgment from an unlawful one. A legislature may remove the basis on which a judgment rests. It may not simply pass the same provision again.

The Act is therefore two stories sharing a binding. The first is a story about institutional design: which disputes in India are decided by specialised tribunals, which by the ordinary courts, and who controls the people who decide them. The second is a story about the boundary between legislative and judicial power, and it is the clearer case study of that boundary available in Indian law, because the sequence is complete and short. A provision was made by rule in 2017 and struck down in 2019. It was remade by rule in 2020 and struck down in part that year. It was enacted by Ordinance in April 2021 and struck down in July 2021. It was re-enacted by statute in August 2021 and struck down in November 2025. Each step is documented, each judgment reasoned, and the final judgment explains what was wrong with the method, not merely with the content. This article follows that sequence in order, states the Act’s own text before the argument about it, gives the Government’s defence and the petitioners’ attack at full strength, and only then addresses the constitutional question, which is a question about method and is distinct from whether a four year term or a minimum age of fifty is good policy.

Two statutes in one binding: the shape of the Act

The Tribunals Reforms Act 2021 is Act 33 of 2021. It received the assent of the President on 13 August 2021, and by Section 1(2) it is deemed to have come into force on 4 April 2021, the date of the Ordinance it replaced, the Tribunals Reforms (Rationalisation and Conditions of Service) Ordinance 2021. Section 35 repeals that Ordinance and saves everything done under it. The date from which the Act operates, 4 April 2021, is defined in Section 2(c) as the notified date, and it is the hinge on which the Act’s transitional provisions turn.

The long title describes an Act further to amend a list of statutes, and the body bears that out. Chapter I is preliminary. Chapter II, Sections 3 to 7, is the service code: it governs the qualifications, appointment, removal, term of office, reappointment, salary and allowances of the chairpersons and members of the tribunals listed in the First Schedule, and it does so, in terms, notwithstanding anything contained in any judgment, order or decree of any court. Chapters III to XXV then amend a series of parent statutes one by one. Some of those amendments do no more than substitute a reference to Chapter II of this Act for the earlier reference to Section 184 of the Finance Act 2017, so that the service code follows the tribunal into its new statutory home. Others are the abolition chapters, which omit the provisions constituting an appellate body and substitute the High Court, the Commercial Court or a district court in its place. Chapter XXVI is miscellaneous: a power for the Central Government to amend the Schedule, the laying of rules before Parliament, the transitional provisions of Section 33, a power to remove difficulties for three years, and the repeal of the Ordinance.

One chapter of the amending part does something subtly different from the rest, and it is worth noticing because it shows the code reaching a body that is not, in form, a tribunal at all. Chapter XXV amends the Consumer Protection Act 2019 by inserting a sub-section in Section 55 of that Act, providing that the qualifications, appointment, term of office, salaries and allowances, resignation, removal and other terms and conditions of service of the President and members of the National Consumer Disputes Redressal Commission appointed after the commencement of this Act shall be governed by the provisions of this Act. The National Commission is a statutory commission at the apex of the consumer fora, not a tribunal created under Article 323B, and its inclusion in the First Schedule and in this amendment brought the presidency and membership of the consumer system’s highest body within a service code designed for tribunals, including the minimum age, the panel and the four year term. The insertion is a reminder that the code’s reach was defined by its Schedule and its amendments rather than by any constitutional category, and that a body could be drawn into the code, with all the consequences litigated in this line of cases, by the simple device of listing it.

Chapter XXIV deserves separate notice because it closes the Act’s own genealogy. Section 29 omits Sections 183 and 184 and the Eighth Schedule of the Finance Act 2017. The rule-making power over tribunal service conditions that Parliament had enacted in 2017, and around which the litigation of 2019 and 2020 had been fought, was thus repealed by the very statute that re-enacted its substance in a new location. The 2025 judgment fastened on exactly this: what the Ordinance had done through amendments to Section 184 of the Finance Act, the Act did through Sections 3, 5 and 7 of a stand-alone statute.

How many tribunals does the First Schedule cover?

The First Schedule lists sixteen tribunals, appellate tribunals and authorities whose chairpersons and members Chapter II governs, running from the Industrial Tribunal constituted by the Central Government under the Industrial Disputes Act 1947 to the National Green Tribunal, and taking in the Income Tax Appellate Tribunal and the Customs, Excise and Service Tax Appellate Tribunal.

The list continues with the Appellate Tribunal under the Smugglers and Foreign Exchange Manipulators (Forfeiture of Property) Act 1976, the Central and State Administrative Tribunals, the Railway Claims Tribunal, the Securities Appellate Tribunal, the Debts Recovery Tribunal and the Debts Recovery Appellate Tribunal, the Telecom Disputes Settlement and Appellate Tribunal, the National Company Law Appellate Tribunal, the National Consumer Disputes Redressal Commission, the Appellate Tribunal for Electricity, the Armed Forces Tribunal and the National Green Tribunal.

Sixteen is the operative number for the service code, and it explains why the Act mattered far beyond the five bodies it abolished. The tribunals in the First Schedule decide the bulk of India’s tax appeals, company and insolvency appeals, securities appeals, debt recovery, service disputes of government employees, telecom disputes, electricity appeals, environmental cases and armed forces service matters. A rule about how their members are chosen, how long they serve and on what terms they may be removed is a rule about a large fraction of the adjudication the Indian state performs. Section 31 empowers the Central Government to amend the Schedule by notification, laid before Parliament, so the list is not closed; a tribunal can be brought within the code, or taken out of it, by executive notification under a power the Act itself grants.

Two features of the service code should be fixed before the abolitions are described, because they frame everything that follows. The first is the non obstante formula. Sections 3(1), 3(7), 5 and 7(1) each open by overriding not only other law but any judgment, order or decree of any court. That formula is common in Indian statutes and is not by itself objectionable; legislatures use it to make a provision operate on a clean slate. Its significance here is that the slate being cleaned contained constitutional holdings, delivered weeks earlier, on the very provisions being re-enacted. The second feature is that Chapter II is a code about process rather than about any one tribunal. It does not ask what the Income Tax Appellate Tribunal needs or what the Armed Forces Tribunal needs. It prescribes one committee, one minimum age, one term and one retirement structure for all sixteen, and the uniformity was itself part of the argument on both sides: the Government defended it as rationalisation, and the petitioners attacked it as a single point of executive leverage over every tribunal at once.

The abolitions, body by body

The abolitions are effected in two ways. Section 33, read with the Second Schedule, terminates the offices. The named chapters of the Act then rewrite each parent statute so that the work of the abolished body has somewhere to go. The Second Schedule names five bodies. They are the Appellate Tribunal under the Cinematograph Act 1952, known as the Film Certification Appellate Tribunal; the Authority for Advance Rulings under the Income Tax Act 1961; the Airport Appellate Tribunal under the Airports Authority of India Act 1994; the Intellectual Property Appellate Board under the Trade Marks Act 1999; and the Plant Varieties Protection Appellate Tribunal under the Protection of Plant Varieties and Farmers’ Rights Act 2001.

The Intellectual Property Appellate Board was the largest of the five in reach, because although it was constituted under the Trade Marks Act it heard matters arising under five statutes: trade marks, patents, copyright, geographical indications and plant varieties. The Act therefore amends each of those statutes separately. In the Trade Marks Act, the chapters constituting the Board are omitted and the High Court is substituted in the appeal and rectification provisions. In the Patents Act, appeals that lay to the Board lie to the High Court, and the Board’s role in revocation and in the Controller’s proceedings is removed. In the Geographical Indications Act, the same substitution is made. In the Plant Varieties Act, the Tribunal’s provisions are omitted and appeals lie to the High Court. In the Copyright Act, which had its own Appellate Board in form but whose functions had been folded into the Intellectual Property Appellate Board by the Finance Act 2017, the substitution is split: a run of provisions referring to the Appellate Board refer to the Commercial Court, while appeals against the Registrar of Copyrights under the substituted Section 72 lie to the High Court.

A sixth abolition is carried out without a Second Schedule entry. Chapter VII amends the Customs Act 1962 by omitting the provisions constituting the Appellate Authority for Advance Rulings, so that an appeal against an advance ruling lies, under the amended Section 28KA, to the High Court. The Income Tax advance ruling body named in the Second Schedule was being wound into a different arrangement at the same time: the Finance Act 2021 had already created a Board for Advance Rulings in place of the Authority for most purposes, and Section 33(3) of this Act expressly excepts proceedings pending before the Authority for Advance Rulings under the Income Tax Act from the general transfer of pending cases to the courts.

The Film Certification Appellate Tribunal’s work was the most visible of the five to the general public, because it heard appeals against the certification decisions of the Central Board of Film Certification. The Act omits Sections 5D and 5DD of the Cinematograph Act and substitutes the High Court for the Tribunal in Section 5C, so a producer aggrieved by a refusal or by cuts ordered as a condition of certification appeals to the High Court of the relevant jurisdiction. The Airport Appellate Tribunal heard appeals in the eviction scheme of the Airports Authority of India Act, directed at unauthorised occupants of airport premises; the Act omits the constituting sections, substitutes the High Court in the appeal provision of Section 28K, and in Section 28E substitutes the Central Government for the Tribunal in the connected provision. The Tribunal under the Control of National Highways (Land and Traffic) Act 2002 was abolished by the same technique in Chapter XIX, although it is not a Second Schedule body either: the Act inserts a definition of Court as the principal civil court of original jurisdiction in a district, including the High Court in its ordinary original civil jurisdiction, and provides by the substituted Section 14 that appeals from the Highway Administration lie to that Court.

What does Section 33 do to the officeholders of the abolished bodies?

Section 33(1) provides that every person holding office as chairperson, president, presiding officer, vice-chairperson or member of a body named in the Second Schedule immediately before the notified date ceases to hold that office from 4 April 2021, and is entitled to compensation not exceeding three months of pay and allowances.

The compensation is for the premature termination of the term of office or of any contract of service, and officers and employees of the abolished bodies who were on deputation revert to their parent cadre, ministry or department. The remainder of Section 33 moves the institution’s residue to the Centre. Pending appeals, applications and proceedings before the Second Schedule bodies, other than the Income Tax Authority, stand transferred to the court before which they would have been filed had the Act been in force when they began, and that court may proceed from the stage reached, from an earlier stage, or afresh. Unspent balances and all property of the abolished bodies vest in the Central Government. The design is a complete wind-up in a single section: offices ended, staff returned, files moved, assets taken over. Nothing in the section was held unconstitutional in November 2025, and the challenge to Section 33 that was included in the writ petition did not succeed, a point developed later in this article because it is persistently misreported.

Where the work went

Abolition is only half a policy; the other half is the destination. The destinations chosen by the Act fall into three groups, and the differences among them matter to anyone who practised before these bodies.

The largest group goes to the High Courts. Appeals in cinematograph certification, patents, trade marks, geographical indications, plant varieties, airport eviction and customs advance rulings lie to the High Court, and rectification and revocation business that the Intellectual Property Appellate Board conducted in trade marks and patents went with them. The second group goes to the Commercial Courts, and it is confined to the Copyright Act, where the functions of the old Copyright Board concerning licences, assignments and related disputes are exercised by the Commercial Court or the Commercial Division of a High Court, while Registrar appeals go to the High Court. The third group is the district court in the national highways scheme, where the appeal lies to the principal civil court of the district.

The practical consequences of the intellectual property transfer, the largest single block of work moved, are owned by the intellectual property cluster of this series, and the analysis of intellectual property law in India as a complete field takes up what the transfer did to that practice, including the creation and expansion of intellectual property divisions in several High Courts. What belongs here is the structural point. The transfer returned to the High Courts a body of appellate work that had been taken away from them in stages since 1958 in trade marks and since 2003 in the unified Board, and it did so on the reasoning, stated in the objects of the predecessor Bill, that the tribunals had not delivered the speed that justified their existence and that they cost the exchequer a considerable amount for a small caseload. Whether that reasoning was right is a question about each tribunal separately; the Intellectual Property Appellate Board, for example, had been hampered for years by vacancies and by long periods without a technical member for patents, which is a failure of appointment rather than of design. The Act did not distinguish. It abolished the slow and the merely starved alike, and the distinction matters when the abolitions are weighed, because a tribunal that is never given members will always look like a tribunal that does not work.

The bodies abolished were not new, and two of them carried histories that the abolition debate compressed. The Intellectual Property Appellate Board was created by the Trade Marks Act 1999 and began functioning in 2003, seated at Chennai with benches sitting in the principal commercial centres, and it accumulated jurisdiction in stages: appeals and rectification in trade marks from the start, appeals in patents and geographical indications thereafter, and the functions of the Copyright Board and of the plant varieties tribunal folded into it or alongside it by later amendments, including the merger effected by the Finance Act 2017. It decided some of the most consequential intellectual property questions of its period, in pharmaceuticals, in trade mark rectification and in copyright licensing, with benches that paired a judicial member with a technical member drawn from the patent or trade marks administration. It was also, for long stretches, a body that the appointing process failed: it worked for extended periods without a full complement, with vacancies in the office of chairperson and in the technical membership for patents that left benches unable to sit on classes of cases assigned to them. The Copyright Board whose functions it inherited had an older lineage under the Copyright Act 1957, and the film tribunal had for decades been the forum in which the censorship decisions of the certification board were tested, a role with a public profile out of proportion to its caseload. The Authority for Advance Rulings, created under the Income Tax Act to give non-residents and specified applicants a binding ruling on a proposed transaction before it was undertaken, had by contrast a small and specialised docket, and had already been marked for replacement by a Board for Advance Rulings in the Finance Act 2021. The abolitions thus fell on bodies in very different conditions: one being replaced on a parallel track, one small and specialised, one publicly prominent, and one, the largest, whose difficulties were as much a record of unfilled posts as a record of faulty design.

That last point deserves its own emphasis because it recurs throughout this subject. A tribunal’s measured performance is a function of its sanctioned strength actually being filled. Where a board of five sanctioned members sits with two, its disposal rate halves, its pendency doubles, and the case for abolishing it on grounds of delay writes itself. The Act’s objects relied on delay and expense. The litigation record, in the tribunal cases generally and in the July 2021 judgment expressly, relied on the failure to appoint as a cause of that delay. Both cannot be resolved in the abstract; they meet in the proposition that a state which does not appoint to a body cannot then cite the body’s arrears as the reason to abolish it without further enquiry. The Act made no such enquiry in its text. Section 33 simply wound the five bodies up.

For litigants, the transfer changed the forum, the fee, the procedure and the further appeal. A matter that would have been heard by a bench combining a judicial member with a technical member is heard by a single judge or a division bench of a High Court, applying the procedure of that court. The further appeal that lay from the Board to the High Court is gone, since the High Court is the first appellate forum, and its decision is subject to the ordinary routes of intra-court appeal and of special leave. The Act’s defenders treated that as the removal of a layer. Its critics treated it as the loss of a specialist layer and the addition of the same work to dockets that were already, by any measure, the most burdened in the system. Both descriptions are accurate; they differ on which cost counts more.

The tribunals that remained, and why the code mattered to each

The sixteen bodies in the First Schedule are not alike, and a single service code bore on each differently. The Income Tax Appellate Tribunal, constituted in 1941 and the oldest of the tier, decides appeals against the revenue in benches combining a judicial and an accountant member, and its members have historically been drawn from the Bar, the judiciary and the revenue services; for it, the eligibility and age provisions decided who of those three streams could enter and for how long. The Customs, Excise and Service Tax Appellate Tribunal performs the same function for indirect taxes. The Central Administrative Tribunal and the State Administrative Tribunals decide the service disputes of government employees, the class for which Article 323A was designed, and their chairpersons have conventionally been judges or former judges of the High Courts. The Securities Appellate Tribunal hears appeals against the securities regulator, a body whose orders the tribunal exists to check, which makes the executive’s role in appointing its members a live question rather than an abstraction, since the regulator is an arm of the state whose nominee sits in the ministry that administers the tribunal. The Debts Recovery Tribunals and their appellate tribunal decide the claims of banks and financial institutions, overwhelmingly against private borrowers, and their presiding officers are drawn from the district judiciary. The Telecom Disputes Settlement and Appellate Tribunal decides disputes between licensor, licensees and consumers in a sector where the licensor is the Government itself. The National Company Law Appellate Tribunal hears appeals in company and insolvency matters, the National Consumer Disputes Redressal Commission heads the consumer commissions, the Appellate Tribunal for Electricity hears appeals against the electricity regulators, the Armed Forces Tribunal decides service and disciplinary matters of the armed forces, and the National Green Tribunal decides environmental cases under a set of scheduled statutes with a bench combining judicial and expert members.

Two consequences follow from that variety. The first is that the Government’s case for uniformity had a real foundation in the larger bodies and a weaker one at the margins: a common term and a common retirement structure for the tax and company tribunals is an intelligible policy, whereas the same term applied to a part-time appellate authority under a forfeiture statute is uniformity for its own sake. The second is that the independence question is not uniform either. It is sharpest where the state is the counterparty in nearly every case, as in tax, service and debt recovery, and where the sponsoring ministry is also the respondent, as in telecom and securities. It is least sharp where the tribunal’s work is mostly between private parties. The Act’s code did not differentiate, and neither, in fairness, did the judgments that struck it down: the directions of November 2020 were also framed for the tier as a whole. The debate was conducted throughout at the level of the system, and a reader who wants the tribunal-by-tribunal position must assemble it from the parent statutes as the judgments leave them.

The service code in Chapter II

Chapter II is the part of the Act that was struck down, and its text has to be stated exactly, because the judgments turn on its particulars and because much published writing describes a simplified version of it.

Section 3(1) empowers the Central Government, notwithstanding anything contained in any judgment, order or decree of any court or in any law in force, to make rules providing for the qualifications, appointment, salaries and allowances, resignation, removal and other conditions of service of chairpersons and members, after taking into consideration experience, specialisation in the relevant field and the provisions of the Act. A proviso then sets the floor that became one of the centres of the litigation: a person who has not completed the age of fifty years shall not be eligible for appointment as a chairperson or a member. The minimum age is therefore in the Act itself, not in rules made under it. That was a deliberate escalation of the earlier design, in which the age had appeared in subordinate legislation, and it is what made the 2021 litigation a challenge to primary legislation rather than to rules.

Section 3(2) requires appointment by the Central Government on the recommendation of a Search-cum-Selection Committee constituted under Section 3(3). For tribunals other than the State Administrative Tribunals, the committee consists of a chairperson who is the Chief Justice of India or a judge of the Supreme Court nominated by him; two members who are Secretaries to the Government of India nominated by that Government; one member who is, for a chairperson’s appointment, the outgoing chairperson of the tribunal, for a member’s appointment, the sitting chairperson, and where the chairperson seeks reappointment, a retired judge of the Supreme Court or a retired Chief Justice of a High Court nominated by the Chief Justice of India; and the Secretary of the ministry or department under which the tribunal is constituted, as member-secretary. A proviso substitutes a retired Supreme Court judge or retired High Court Chief Justice for that fourth member in the case of the Industrial Tribunal, the Debts Recovery Tribunals and the Debts Recovery Appellate Tribunal, and any other tribunal the Central Government notifies in consultation with the committee’s chairperson, and also wherever the sitting or outgoing chairperson is not himself a retired judge of the Supreme Court or a retired Chief Justice or judge of a High Court. Section 3(4) gives the chairperson a casting vote. Section 3(5) denies the member-secretary a vote. Section 3(6) lets the committee determine its own procedure. For State Administrative Tribunals a parallel committee is prescribed, chaired by the Chief Justice of the High Court of the State, with the Chief Secretary, the Chairman of the State Public Service Commission, the corresponding tribunal member and the Secretary of the General Administration Department as member-secretary.

Count the votes and the constitutional argument becomes concrete. On the committee for a central tribunal, the judicial side consists of the chairperson alone, unless the fourth member happens to be a retired judge. The executive side consists of two nominated Secretaries plus a member-secretary without a vote. The tribunal’s own chairperson is the fourth voting member. The chairperson of the committee has a casting vote, so on a tied vote the judicial voice prevails; but the committee is not a body with a judicial majority, and the earlier judgments had treated judicial primacy in the selection of persons who exercise judicial power as a requirement rather than a preference. That arithmetic, as much as any single provision, is what the litigation was about.

Section 3(7) contains the second centre of the litigation. It provides, again notwithstanding any judgment, that the committee shall recommend a panel of two names for appointment to the post of chairperson or member, and that the Central Government shall take a decision on the recommendations preferably within three months. Section 3(8) saves appointments from invalidity on account of a vacancy or absence in the committee. Section 4 lists the grounds on which the Central Government shall, on the recommendation of the committee, remove a chairperson or member: insolvency, conviction for an offence involving moral turpitude, physical or mental incapacity, acquisition of a financial or other interest likely to affect the functions prejudicially, and abuse of position rendering continuance prejudicial to the public interest. Removal on the last three grounds requires that the person be informed of the charges and heard. The manner of removal is left to rules.

Section 5 fixes the term. The chairperson of a tribunal holds office for a term of four years or until he attains the age of seventy years, whichever is earlier. A member holds office for a term of four years or until he attains the age of sixty-seven years, whichever is earlier. A proviso protects a defined class: a chairperson or member appointed between 26 May 2017, the date from which the Finance Act 2017 scheme operated, and the notified date, whose appointment order specified a longer term or a higher retirement age, keeps that term or age, subject to a maximum term of five years. Section 6 makes chairpersons and members eligible for reappointment in accordance with the Act, with a direction that preference be given to the service rendered, and requires reappointments to be made in the same manner as appointments. Section 7 empowers the Central Government to make rules for salary, and provides that allowances and benefits are payable to the extent admissible to a Central Government officer holding a post carrying the same pay, with a proviso permitting reimbursement of house rent above the ordinary house rent allowance subject to rules, and a protection that neither salary and allowances nor other terms may be varied to a person’s disadvantage after appointment.

What salary and allowances did Section 7 lay down?

Section 7(1) empowers the Central Government to provide by rules for the salary of a chairperson or member, and entitles them to allowances and benefits to the extent admissible to a Central Government officer holding a post carrying the same pay.

A proviso allows reimbursement of house rent higher than the admissible house rent allowance, subject to limits and conditions set by rules, and Section 7(2) bars any variation of salary, allowances or other terms to the holder’s disadvantage after appointment. The protection against disadvantageous variation is one of the few provisions of Chapter II that no judgment had ever questioned, and its presence in a code whose other protections were so much weaker was noticed in the argument: the Act protected a member’s pay packet against reduction while leaving his tenure to a four year term and a renewal he could not count on.

Three features of the code should be noted before the history begins, because each reappears in the judgments. First, the code is skeletal by design: it fixes the age floor, the committee, the panel, the term and the ceiling ages in the Act, and leaves qualifications, salary and the manner of removal to rules made by the Central Government. The litigation was therefore about both levels at once, about what the Act fixed and about how much it left to the executive. Second, the code’s protections run in one direction. Salary cannot be varied to disadvantage, and removal requires a committee recommendation and a hearing on the graver grounds; but tenure is short, reappointment depends on a fresh recommendation through the same process, and the appointment itself is a choice by the Government between two names. Third, the code displaced a patchwork. Before 2017 each parent statute carried its own qualifications, term and retirement age, and the result was the inconsistency the Government relied on: terms of three, four and five years across tribunals, retirement ages of sixty-two, sixty-five, sixty-seven and seventy, and widely differing eligibility rules for the same kind of post. Any account of Chapter II that omits that patchwork understates the case for a code. Any account that omits the litigation history understates the case against this one.

How India came to be governed by tribunals

Tribunals are a constitutional creation, not merely a statutory convenience. The Constitution (Forty-second Amendment) Act 1976 inserted Part XIVA, containing Article 323A, which empowers Parliament to provide for administrative tribunals for service disputes, and Article 323B, which empowers the appropriate legislature to provide for tribunals for taxation, foreign exchange, industrial and labour disputes, land reforms, ceiling on urban property, elections, production and distribution of foodstuffs and certain other matters. Acting on those powers, and on the ordinary legislative entries, Parliament and the State Legislatures built, over four decades, a parallel tier of adjudication: administrative tribunals for government servants, tax tribunals, company law tribunals, debt recovery tribunals, consumer commissions, environmental and electricity appellate bodies, and dozens of others. The reasons given were consistent across the period: the ordinary courts were overburdened, the disputes were technical, and a specialised forum with a member drawn from the field would decide faster and better. The experiment that produced the National Company Law Tribunal and its appellate tribunal, examined in the account of the NCLT and NCLAT and the long litigation over their constitution, is the fullest single illustration of how that reasoning was applied to company and insolvency disputes.

The constitutional settlement governing that tier was reached in stages. In S. P. Sampath Kumar v Union of India in 1987 the Supreme Court upheld the Administrative Tribunals Act 1985, on the footing that a tribunal could substitute for a High Court if it was made an equally effective institution, with members of comparable standing and independence. A decade later, in L. Chandra Kumar v Union of India in 1997, a bench of seven judges revised that footing. The Court held that the clauses of Articles 323A and 323B that excluded the jurisdiction of the High Courts were unconstitutional, that tribunals would function as courts of first instance subject to the writ and supervisory jurisdiction of the High Courts under Articles 226 and 227, and that a tribunal’s decisions could be tested before a division bench of the High Court rather than carried directly to the Supreme Court. Tribunals after 1997 are therefore not substitutes for the constitutional courts. They are a first tier inside a hierarchy that the High Courts supervise, and their members exercise judicial power of a kind that attracts the constitutional requirements attached to judicial office.

L. Chandra Kumar also identified the administrative defect that every later judgment returned to and that the National Tribunals Commission is meant to cure. The tribunals, the Court observed, were administered by the very ministries whose decisions they reviewed, which supplied their staff, their premises and their budgets, and it recommended that tribunals be brought under the supervision of a single nodal ministry, and in time an independent agency, rather than left as appendages of the departments appearing before them. That recommendation was not implemented in the form made. Administration remained dispersed among sponsoring ministries for the following quarter century, and the absence of an administrative home independent of the litigant state is the background condition of everything that happened between 2017 and 2025: a tier of courts in everything but name, maintained as a line item in the budgets of its principal litigant. When the Court in November 2020 directed a separate tribunals wing in the Ministry of Finance as an interim measure, and when it directed a Commission in 2025, it was returning to the point made in 1997 and giving it, at last, a date.

Two later judgments set the terms on which the 2017 to 2025 contest was fought. In Union of India v R. Gandhi, President, Madras Bar Association, decided in 2010, the Court upheld the provisions of the Companies Act providing for the National Company Law Tribunal and the National Company Law Appellate Tribunal, but on conditions: it specified the qualifications their members should hold, held that the selection committee should be constituted with judicial primacy, and laid down that the tenure and conditions of members must be such as to secure independence, since a short tenure and the prospect of favour from the appointing authority are corrosive of it. In Madras Bar Association v Union of India in 2014 the Court struck down the National Tax Tribunal Act 2005 altogether, holding that a tribunal which transferred the High Courts’ jurisdiction over substantial questions of tax law to a body whose members need not be judges, and whose appointments and conditions were executive controlled, impermissibly diluted the judicial function. The through-line of these decisions is a single proposition, stated in different forms: Parliament may create tribunals and may transfer jurisdiction to them, but the body to which judicial power is transferred must be independent of the executive in its appointment, its tenure and its administration, in proportion to the power transferred. Every provision litigated between 2017 and 2025 was tested against that proposition.

The Finance Act 2017 and the Rules of 2017

The modern dispute begins with the Finance Act 2017. Part XIV of that Act did two things. It rationalised the tribunal map, merging eight tribunals into cognate bodies with which they shared a field: among the mergers, the Competition Appellate Tribunal was merged into the National Company Law Appellate Tribunal, the Copyright Board into the Intellectual Property Appellate Board, the Airports Economic Regulatory Authority Appellate Tribunal into the Telecom Disputes Settlement and Appellate Tribunal, and the Employees Provident Funds Appellate Tribunal into the Industrial Tribunal. And it inserted Section 184, which empowered the Central Government to make rules, for the tribunals listed in the Eighth Schedule, on qualifications, appointment, term of office, salaries and allowances, resignation, removal and other conditions of service. Section 184 was the first attempt to govern the service conditions of many tribunals from a single rule-making power held by the executive, and the power was conferred by a Finance Act, a choice that later generated the separate question whether such provisions could travel in a Money Bill at all.

Rules under Section 184 were notified in June 2017 as the Tribunal, Appellate Tribunal and other Authorities (Qualifications, Experience and other Conditions of Service of Members) Rules 2017. They prescribed a term of three years, retirement ages lower than several parent statutes had provided, selection committees in which the executive held the stronger hand, and qualifications that in several tribunals opened judicial posts to persons without judicial or legal training. Petitions followed across the country, and the Supreme Court’s interim orders in Kudrat Sandhu v Union of India in 2018 directed that appointments in the interim be governed by the parent statutes and the position before the Finance Act, a direction to which the Attorney General for India consented and which becomes important again at the end of this story.

The two years between the Rules and the Constitution Bench were governed by interim arrangements, and the arrangements matter because they established the pattern, repeated in 2025, of the Court keeping the appointment system running on the old law while the new law was tested. Petitions against the 2017 Rules were filed in the Supreme Court and in several High Courts. In Kudrat Sandhu v Union of India, a Constitution Bench dealing with the appointment of members to the Central Administrative Tribunal and other bodies issued interim directions in 2018 under which appointments in the interim were to be made in accordance with the parent statutes and the rules as they stood before the Finance Act 2017, and the Attorney General for India stated the Union’s agreement to that course. Selection committees were accordingly reconstituted under the old provisions, and appointments made in that period carried the terms of the parent Acts, including the longer terms and higher retirement ages that the 2017 Rules had cut down. The significance of the episode is easy to miss. It meant that from 2018 onward there were always two regimes in contention, the one on paper and the one under which appointments were actually being made with the Union’s consent, and that the members appointed in the interim acquired terms under the old law that no subsequent scheme could retrospectively shorten without the kind of protection question that eventually reached the Court in the ITAT and CESTAT applications of 2025. When the 2020 Rules then claimed retrospective effect, the collision with that interim settlement was one of the grounds on which the retrospectivity was struck down.

Roger Mathew: the Constitution Bench draws the first line

Rojer Mathew v South Indian Bank Ltd, decided on 13 November 2019 by a Constitution Bench of five judges, was the first full adjudication of the 2017 design. The Court struck down the 2017 Rules in their entirety. The reasoning had two strands. The first was that the rules, in their provisions on search and selection committees, tenure and qualifications, failed to secure the independence that the earlier judgments required: they gave the executive a dominant voice in choosing the members of bodies that decide disputes to which the Government is the most frequent party, and they prescribed tenures too short to give a member security against the appointing authority. The second was excessive delegation. Section 184 delegated the essential features of tribunal design to the executive without legislative guidance on the matters that determine independence, and while the Court upheld Section 184 itself, reading it as conditioned by the constitutional requirements the earlier cases had laid down, it held that the rules made under it did not answer those requirements and could not stand. The Court also referred to a larger bench the question whether the Finance Act 2017 could have been passed as a Money Bill, a question that concerns the route by which Section 184 was enacted rather than its content, and it recorded the need, which Justice Chandrachud’s concurring opinion developed at length, for an independent National Tribunals Commission to take appointments, administration and oversight of tribunals out of the hands of the ministries that appear before them.

The delegation strand of Roger Mathew repays a closer statement, because it is the strand the Act’s design was meant to answer and because it is frequently misstated as a holding that Parliament may not delegate at all. Indian law permits wide delegation; what it forbids is the delegation of essential legislative function without guidance. The essential features of a tribunal scheme, on the analysis the Court had built since 2010, include who may be appointed, by whom they are selected, how long they serve and how they may be removed, because those features determine whether the body can act independently of the executive. Section 184 delegated all of them, in terms that supplied no standard by which the independence of the resulting scheme could be measured, and the rules made under it illustrated the risk by producing committees and tenures that failed the constitutional test. The Court’s response was measured in form: it did not strike down Section 184, but it upheld the section only as conditioned by the law laid down in the earlier judgments, and it struck down the rules as exceeding what a constitutionally read delegation would bear. That construction placed the Government in a position it appears not to have fully reckoned with. Fresh rules under the same section inherited the condition. The only routes to a scheme free of it were primary legislation that enacted the conditions itself, or an amendment of the delegation that wrote the standards in. The 2020 Rules took neither route, being rules again; and the 2021 Act, which did enact the conditions in primary legislation, answered the delegation objection completely while renewing the substantive one, which is why the 2025 judgment contains no delegation holding and decides the case on re-enactment and independence.

Two consequences of Roger Mathew structure the rest of the story. The first is that after November 2019 there existed an authoritative statement, from a Constitution Bench, of what a lawful scheme had to contain, and the Government was not writing on a blank page when it wrote the next scheme. The second is that the defect identified was not only in the content of the rules but in the method of making them: service conditions for the judiciary of the tribunal tier had been set by executive rule under an unguided delegation. A legislature responding to Roger Mathew in the orthodox way had two open routes. It could enact the conditions in primary legislation, curing the delegation objection and meeting the independence requirements the Court had specified. Or it could issue fresh rules within the delegation as the Court had conditioned it. The Government took the second route first, and then, when that failed, took the first route while declining to meet the specified requirements. The judgments of 2020, July 2021 and November 2025 are the record of those choices.

The Money Bill question running underneath

One strand of the litigation was never decided and should be recorded so that it is not mistaken for part of the holdings. The Finance Act 2017 was enacted as a Money Bill, certified as such by the Speaker of the Lok Sabha, which meant that the Rajya Sabha could recommend amendments but could not block it, and that the President’s options were correspondingly narrow. Part XIV of that Act, which contained the tribunal mergers and Section 184, dealt with the constitution and service conditions of adjudicatory bodies, and the challenge in Roger Mathew included the contention that such provisions could not be carried in a Money Bill, whose permissible contents Article 110 confines to taxation, public borrowing, the Consolidated Fund and matters incidental to them. The Constitution Bench in Roger Mathew upheld Section 184 on its merits as conditioned, struck down the Rules, and, doubting the reasoning by which the Aadhaar Act’s Money Bill certification had been upheld, referred the interpretation of Article 110 and the correctness of that reasoning to a larger bench. That reference has not been answered by a bench of the requisite strength, and the Money Bill question therefore hangs over the Finance Act route without having been resolved.

For this Act the question does not arise in the same form, and the reason is worth stating. The Tribunals Reforms Act 2021 is an ordinary statute, passed by both Houses, and its validity does not depend on Article 110. When the Act omitted Sections 183 and 184 of the Finance Act and re-enacted their substance in its own Chapter II, it also, in effect, moved the service code out of the Money Bill’s shadow. The 2025 judgment rests entirely on the re-enactment and independence grounds and not on the Money Bill ground. A reader who encounters the Money Bill argument in commentary on this subject should place it where it belongs: as an unresolved objection to the 2017 route, whose practical importance diminished once the code was re-enacted in an ordinary Act, and which would matter again if a future government returned to the device of carrying institutional design in a Finance Act.

The Rules of 2020 and the judgment of November 2020

Fresh rules were notified in February 2020 as the Tribunal, Appellate Tribunal and other Authorities (Qualifications, Experience and other Conditions of Service of Members) Rules 2020. They lengthened the term from three years to four, and made other adjustments, but they retained short tenure, retirement ages below the levels the Court had indicated for several posts, and selection arrangements that the Madras Bar Association again challenged. The challenge was decided on 27 November 2020, in the decision the later judgments call MBA-III and others call MBA-IV, the numbering varying with whether the 2010 or some other decision is counted first; the date is the safe identifier.

The judgment of 27 November 2020 struck down or read down the 2020 Rules in part and issued a set of directions that together amount to a specification for a lawful scheme. The term of office of a chairperson and of a member was to be five years, with a chairperson holding office until the age of seventy and other members until the age of sixty-seven. Advocates with at least ten years of experience were to be eligible for appointment as judicial members, and members of the Indian Legal Service were to be eligible on the same footing subject to suitability, since a scheme that confined judicial membership to retired judges and serving officials both narrowed the field and deepened the dependence of the tribunal tier on the state. The Search-cum-Selection Committee was to recommend one name for each post, with a wait-listed candidate, not a panel from which the Government would choose. Appointments were to be made within three months of the recommendation. The committee itself was to be constituted with judicial primacy, chaired by the Chief Justice of India or his nominee with a casting vote. House rent allowance was to be substantially enhanced where official accommodation was not provided. The provision giving the 2020 Rules retrospective effect was struck down, subordinate legislation having no power to operate retrospectively without statutory authority. And the Court directed the Union to establish a National Tribunals Commission, with a separate tribunals wing in the Ministry of Finance to serve until the Commission existed, so that the administration of tribunals would pass out of the ministries whose cases the tribunals decide.

The character of these directions matters to the later judgments. They were not obiter observations or expressions of hope. They were issued after argument, in a writ petition, as modifications and directions necessary to bring an impugned scheme within the Constitution, and the Court in 2025 treated them as binding directions under Article 141 whose content the executive and Parliament alike were required to respect unless the law was changed in a constitutionally permissible way. The Attorney General of the day participated in the process: when a modification of the direction on reappointment of advocates was later sought, it was sought by application to the Court, with the amicus curiae not objecting, and the word entitled was replaced by eligible by order. That is how a party who accepts a judgment seeks to vary it. The events of 2021 took a different course.

The Ordinance of April 2021

On 4 April 2021 the President promulgated the Tribunals Reforms (Rationalisation and Conditions of Service) Ordinance 2021. A Bill of the same name had been introduced in the Lok Sabha in February 2021 and had not been taken up. The Ordinance did the two things the Act later did. It abolished the appellate bodies described above and transferred their work, with effect from its own date. And it amended Section 184 of the Finance Act 2017 to write into that section, as provisos and subsections, a service scheme for tribunal members: a minimum age of fifty years for appointment, a term of four years, a Search-cum-Selection Committee that would recommend a panel of two names for each post, and ancillary provisions on allowances.

Measured against the judgment of 27 November 2020, which was then a little over four months old, the Ordinance moved in the opposite direction on each point the judgment had decided. The judgment had set the term at five years; the Ordinance set it at four. The judgment had made advocates of ten years of experience eligible; the Ordinance imposed a minimum age of fifty, which in practice excludes the advocate the judgment had admitted, since an advocate who enrols in the early twenties reaches ten years of practice in the early thirties. The judgment had directed one name per post; the Ordinance directed a panel of two. The judgment had struck at executive dominance in selection; the Ordinance retained a committee design in which the judicial voice was one among several. The Ordinance was challenged at once, by the Madras Bar Association among others, on the ground that it did not merely legislate afresh but legislated against the judgment.

What besides service conditions did the Ordinance carry?

The Ordinance carried the whole abolition programme that the Act later repeated: the dissolution of the Intellectual Property Appellate Board, the Film Certification Appellate Tribunal, the Airport Appellate Tribunal, the Authority for Advance Rulings and the Plant Varieties Protection Appellate Tribunal, with the consequential amendments to their parent statutes.

The pending work of those bodies was transferred to the High Courts and, for copyright, to the Commercial Courts. Those parts of the Ordinance were not struck down in July 2021 and were re-enacted without material change in the Act, which is why the abolitions were never in suspense during the litigation over tenure and age.

July 2021: the first strike down of the provisions themselves

The challenge to the Ordinance was decided on 14 July 2021 by a bench of three judges, Justices L. Nageswara Rao, Hemant Gupta and S. Ravindra Bhat. Justices Rao and Bhat formed the majority; Justice Gupta dissented. The majority judgment, delivered by Justice Rao, declared unconstitutional the first, second and third provisos to Section 184 insofar as they fixed fifty years as the minimum age for appointment and dealt with house rent allowance contrary to the 2020 judgment, declared Section 184(7), which required the committee to recommend two names for each post, unconstitutional, and held Section 184(11), prescribing a tenure of four years, contrary to the principles of separation of powers, independence of the judiciary, the rule of law and Article 14 of the Constitution. The majority also made clear that appointments made after 4 April 2021 would be governed by the Ordinance as modified by the judgment, that appointments made to the Customs, Excise and Service Tax Appellate Tribunal under interim orders of the Court would be governed by the statute and rules as they stood before the Finance Act 2017, and that even interim orders of the Court could not be overruled by a legislative act.

The majority’s reasoning on the minimum age illustrates the method of the whole judgment. Fixing fifty as the minimum age, Justice Rao held, would act as a deterrent to competent advocates, would exclude the class the 2020 judgment had deliberately admitted, and would leave a member appointed at fifty who was not reappointed after a single four year term unable, as a practical matter, to rebuild a practice. The reasoning on tenure was structural: a member who holds office for four years and whose continuance depends on reappointment through a process the executive influences will decide, or will be perceived to decide, with the appointing authority in view, and the authority most often before a tribunal is the Government. Justice Bhat’s concurring opinion agreed in reasoning and result, and addressed directly the argument that the Court was trespassing on legislative ground, holding that the prescription of conditions destructive of independence is not a policy choice the separation of powers protects.

Justice Gupta’s dissent is an essential part of the record and is stated here at its own strength, since it represents the strongest form of the argument the Union later ran again. In his view, a law cannot be struck down merely because it is not in accordance with the precedents of the Court; directions issued in earlier judgments on matters of policy, such as the length of a term or a minimum age, do not bind Parliament when it legislates; and the challenge to most of the Ordinance’s provisions was not legally sustainable. The disagreement between the majority and the dissent is thus not about tribunals alone. It is about the juridical status of directions: whether a direction that a scheme must contain a five year term, issued as part of a judgment that found a four year term unconstitutional for stated reasons, is a constitutional holding that binds the legislature, or a remedial preference that the legislature may discard. The majority treated it as the former, because the direction was the operative content of the finding of unconstitutionality: the four year term fell because of what it did to independence, and a legislature that re-enacts it re-enacts the defect. That analysis, adopted in July 2021 against an Ordinance, was applied in November 2025 against an Act of Parliament.

Two further features of the July 2021 judgment complete its picture. The first is what it upheld. The majority did not disturb the abolition provisions of the Ordinance, did not question the power of Parliament to prescribe qualifications and conditions for tribunal posts as a subject of legislation, and upheld the proviso to Section 184(11) that protected the terms of persons appointed in the window after the Finance Act scheme began, directing that appointments made to the Customs, Excise and Service Tax Appellate Tribunal under the Court’s interim orders be governed by the statute and rules as they stood before 26 May 2017. The judgment is therefore a partial strike down of an appointments scheme, not a general veto on tribunal legislation, and the Government’s later description of the litigation as the Court obstructing reform does not survive a reading of its operative part. The second feature is the treatment of house rent allowance. The Ordinance had carried provisos overriding the 2020 judgment’s direction substantially enhancing house rent allowance where accommodation was not provided; the majority struck those down as well, and the Union thereafter amended Rule 15 of the 2020 Rules by a notification of 30 June 2021 to conform to the direction, a detail the judgment itself recorded. The episode is small and instructive: on the one matter where the Government complied by amending its rules, the judgment noticed the compliance and the matter ended. On the three matters where it did not, the matter continued for four more years and ended in a strike down of an Act of Parliament.

Justice Bhat’s concurrence added a strand that the 2025 bench later drew on. Concurring fully with Justice Rao, he treated independence of the judiciary and separation of powers as themes whose relevance is not confined to the constitutional courts, and he could not persuade himself to agree with the dissent that a minimum age or a condition of service is a matter of legislative wisdom into which the Court should not enter when earlier judgments had decided the point. That exchange, between a concurrence that treats the directions as constitutional law and a dissent that treats them as remedial advice, is the whole of the 2025 case in embryo. The bench that decided the Act’s fate adopted the concurrence’s premise.

The judgment closed with an observation that reads, in retrospect, as the hinge of the whole decade. Recording that the 2020 judgment had been delivered after a detailed dialogue with the Attorney General, the majority noted the existence of a large number of vacancies of members and chairpersons and the inordinate delay in filling them, which had resulted, in its words, in the emasculation of the tribunals, and said that it was high time a serious effort was made to fill vacancies without delay, since the main reason for tribunalisation, speedy justice, was not being achieved while tribunals wilted under the weight of their dockets. The Court was, in the same judgment, striking down the Government’s scheme and urging the Government to appoint. Both halves of that sentence matter to the vacancy question taken up below.

The Act of August 2021

Parliament’s response came within weeks. The Tribunals Reforms Bill 2021 was introduced in the Lok Sabha on 2 August 2021, passed by the Lok Sabha on 3 August 2021 and by the Rajya Sabha on 9 August 2021, and received the President’s assent on 13 August 2021. The Act repealed the Ordinance and re-enacted its contents as primary legislation, with the service provisions moved out of Section 184 of the Finance Act and into Chapter II of the new Act, and with the Finance Act provisions themselves omitted.

On the three points decided in July 2021, the Act repeated the Ordinance. The minimum age of fifty stands in the proviso to Section 3(1). The panel of two names stands in Section 3(7). The term of four years stands in Section 5. Each provision is prefaced by the non obstante formula overriding any judgment, order or decree. The Act added the proviso to Section 5 protecting the terms of persons appointed between 26 May 2017 and the notified date, and it carried the abolition chapters and Section 33 in substantially the Ordinance’s terms. The passage of the Bill was rapid and was contested on precisely the ground the Court later accepted. The Bill was introduced by the Minister of Law and Justice, Kiren Rijiju, and in both Houses the objection pressed by opposition members was that the Bill re-enacted, three weeks after the event, provisions the Supreme Court had struck down on 14 July 2021, and that enacting them in the teeth of the judgment invited a further strike down and demeaned the Court’s authority. The Government’s answer in the debate was the answer it later gave in court: that the legislature was within its rights to prescribe the conditions of tribunal service, that the judgment on the Ordinance did not exhaust Parliament’s power over the subject, and that the system needed a uniform code rather than continued government by direction. Each House passed the Bill within days of its introduction, in a session dominated by other controversies, and the assent followed on 13 August. The speed is part of the record the 2025 judgment sits on. A legislature that takes three weeks to re-enact a struck down provision has not, in the ordinary sense, reconsidered it; and while speed is not a constitutional defect, the judgment’s language about a conscious defiance of the law declared earlier is intelligible only against that chronology.

The Madras Bar Association challenged Sections 3(1), 3(7), 4, 5, 6, 7(1) and 33 in Writ Petition (Civil) No. 1018 of 2021, and that petition was decided on 19 November 2025.

The sequence to this point is the article’s findable artifact, and it repays being seen whole, because the individual steps are each reported and the whole is rarely set out in one place.

Step	Instrument or decision	What it provided or held	Position immediately afterwards
2017 provisions	Finance Act 2017, Part XIV and Section 184	Eight tribunals merged into cognate bodies; the Central Government empowered to make rules on qualifications, appointment, term, salaries and removal for the scheduled tribunals	Service conditions of the tribunal tier passed to executive rule-making
2017 Rules	Tribunal Rules of 2017 under Section 184	Three year terms, lower retirement ages, executive-weighted selection committees	Appointments under the rules challenged across the country
2019 judgment	Rojer Mathew v South Indian Bank, 13 November 2019	The 2017 Rules struck down in full; Section 184 upheld as conditioned by earlier law; National Tribunals Commission urged	The field returned to the parent statutes and interim arrangements
2020 Rules	Tribunal Rules of 2020	Four year terms and other adjustments, with retrospective effect claimed	Challenged by the Madras Bar Association
2020 judgment	Madras Bar Association, 27 November 2020	Rules struck down or read down in part; five year terms, retirement at seventy and sixty-seven, advocates of ten years eligible, one name per post, National Tribunals Commission directed	A judicial specification for a lawful scheme stood on record
2021 Ordinance	Ordinance of 4 April 2021	Five appellate bodies abolished; Section 184 amended to impose a minimum age of fifty, a four year term and a panel of two names	Abolitions took effect; service provisions challenged at once
July 2021 judgment	Madras Bar Association, 14 July 2021	The minimum age, the panel of two names and the four year term struck down, two to one; abolitions untouched	Appointments to be made under the Ordinance as modified by the judgment
August 2021 Act	Tribunals Reforms Act 2021, Act 33 of 2021	The Ordinance re-enacted as a statute; the same age, panel and term placed in Sections 3 and 5 with effect from 4 April 2021	The struck down provisions returned in statutory form; challenged in Writ Petition (Civil) 1018 of 2021
November 2025 judgment	Madras Bar Association v Union of India, 19 November 2025	Sections 3 to 7 struck down; the 2020 and 2021 directions restored; National Tribunals Commission to be established within four months; earlier selections protected	Tribunal appointments governed by the earlier directions until Parliament enacts a compliant law

The Government’s case, at full strength

The Union’s defence of the Act, advanced by the Attorney General for India, R. Venkataramani, deserves to be stated as it was put, because it is not a frivolous case and because the 2025 judgment rejected it as a reprise of arguments already considered, which is a different thing from its being unarguable.

The first limb is institutional. Tribunals are created by the executive’s sponsoring ministries, funded from the Consolidated Fund through those ministries, housed and staffed by them, and accountable through them to Parliament for their budgets and their performance. A member of a tribunal is not a judge of a constitutional court; the office is a statutory post of a defined kind, and the conditions of statutory posts are, in the ordinary run of Indian public law, a matter for the legislature and the executive. To constitutionalise a five year term, a retirement age and a selection method is to transfer the design of a large part of the state’s administrative apparatus from the organs answerable for it to the courts.

The second limb concerns the four year term. A term of four years with eligibility for reappointment, the Union argued, enables performance review: a member who has not disposed of cases, who has been unable to sit regularly, or whose conduct of a bench has generated difficulty can be let go at the end of a term without the confrontation of a removal proceeding, and a member who has performed well can be continued, with Section 6 directing that preference be given to the service rendered. A long, non-renewable term, by contrast, converts tribunal membership into a settled entitlement insulated from any assessment, and the prospect of renewal is common in public appointments without being thought to destroy independence; members of many regulatory bodies serve renewable terms. The third limb concerns the minimum age. Tribunals decide complex technical disputes in taxation, securities, insolvency and environmental science, frequently against senior counsel and on records of great volume. A minimum age of fifty was defended as a proxy for the substantial experience such work requires, and as a uniformity measure, since the parent statutes had prescribed no consistent eligibility floor. The fourth limb concerns the panel of two names. A selection committee that recommends a single name leaves the appointing authority no choice at all, which is not the ordinary position in public appointments and which concentrates the effective power of appointment in the committee; a panel of two preserves a real, if narrow, executive choice while keeping the field confined to persons the committee has found suitable. The fifth limb is the constitutional one, and it is the limb on which the case was lost. The Attorney General argued that Parliament has discretion to legislate on a subject within its competence, that principles such as judicial independence and separation of powers cannot be invoked to invalidate a statute unless the statute violates an express constitutional provision, and that the decisions of the Court in the tribunal cases, insofar as they prescribed the contents of a scheme, did not bind Parliament to those contents. On that footing the Act was not an override at all, but the exercise of a legislative power that the Court’s remedial directions had never displaced.

A sixth limb drew on the constitutional text. Articles 323A and 323B commit the creation of tribunals to the legislature and expressly empower it to prescribe matters ancillary to them, and the parent statutes of the tribunal tier have always contained the qualifications, terms and retirement ages of their members as ordinary legislative choices, varying from tribunal to tribunal without constitutional objection. On this argument, the Court’s case law had moved, by increments, from reviewing tribunal design for constitutionality to specifying it, and the specification had never been anchored in a provision that fixes five years, or forbids a minimum age, or prescribes the composition of a selection committee. What the judgments call requirements are, on this view, standards of good design dressed as constitutional law, and the appropriate judicial response to a badly designed scheme is the response given to bad policy in other fields: it is for the legislature to answer to the electorate for it. The 2025 bench met this argument directly, in the passage holding that the norms are distilled from Articles 323A and 323B read with separation of powers, independence of the judiciary and Article 14, and the meeting of the two positions is the jurisprudential centre of the case. A reader who wants the strongest statement of the losing side should read it here, in its textual form, rather than in the weaker forms in which it is sometimes summarised.

Supporting considerations were pressed alongside. The State is the largest litigant before the tribunals, but it is also the litigant before the High Courts and the Supreme Court, whose judges it appoints through a process in which it participates, without that participation being thought to disqualify the forum. Vacancies, on the Union’s account, were a problem of supply and of litigation, not of design: every scheme since 2017 had been under challenge, selection processes had been stayed or unsettled by interim orders, and the executive could not appoint on a stable basis while the governing rules were in flux. And the abolitions, which were not struck down, showed the direction of the policy: fewer, better-resourced bodies, with the work of weak appellate tribunals returned to the courts.

The petitioners’ case, at full strength

The petitioners, led by senior advocates Arvind P. Datar and C. S. Vaidyanathan for the Madras Bar Association, with a number of other senior advocates appearing for applicants including affected members, answered each limb on the same ground: that the question is not whether the executive may design tribunals, but whether a member who decides cases against the Government can be kept dependent on the Government for the continuance of his office.

On tenure, their case was that a four year term with reappointment in the gift of the executive is not a performance review mechanism but a dependence mechanism. A member in the third year of a four year term, hearing appeals by the revenue, by a regulator or by a ministry, knows that his continuance, and with it his salary, his residence and his post-retirement prospects, will be decided shortly by the party before him. The harm does not require any member actually to bend; a system that gives the most frequent litigant a recurring, lawful opportunity to end the adjudicator’s tenure has already compromised the appearance and the reality of independence. Security of tenure is not a perquisite of office. It is the mechanism by which the Constitution makes impartiality possible, and it is why the Court had fixed five years as a floor and why even that was tied to retirement ages that gave a member a reasonable span.

On the minimum age, the case was that fifty is not a proxy for experience but a filter with a direction. It excludes the advocate in the forties with twenty years of practice, the class the 2020 judgment had deliberately brought in to widen the field beyond retired judges and serving officers. Its practical effect is to confine appointment to persons at or near the end of other careers, for whom a tribunal post is a terminal appointment, and to exclude the one class whose members might serve, return to practice, and carry the tribunal’s standards back into the profession. Combined with the four year term, it produces the deterrent the July 2021 majority identified: an advocate appointed at fifty who is not reappointed at fifty-four cannot rebuild a practice, and able advocates will decline to apply, leaving the field to those with least to lose.

On the panel of two names, the case was that the choice it reserves to the executive is the substance of the appointment power. A committee that recommends one name performs the selection; a committee that recommends two performs a shortlisting, and the appointing authority performs the selection, with no obligation to record why the first name was passed over. Where the committee has a judicial chairperson but no judicial majority, the panel device compounds the position: the executive members of the committee can ensure that both names on the panel are acceptable to the executive, and the Government then chooses between them. On the committee’s composition more generally, the petitioners relied on the arithmetic set out earlier in this article and on the principle, running from the 2010 judgment through Roger Mathew, that bodies exercising judicial power must be selected by a process in which the judicial voice predominates.

Underlying all three limbs was the argument about method, which is the argument that won. Whatever the merits of four years, of fifty, or of two names, those provisions had been considered and struck down, with reasons, in July 2021, and the Act re-enacted them without altering the features the Court had held fatal. A litigant who loses may appeal, seek review, or invite a larger bench to reconsider. The State as legislator has a fourth route unavailable to any other litigant: it may change the law. But the petitioners’ case was that the fourth route is a route to a different law, not to the same law with a new citation, and that to treat it as the latter is to make every constitutional judgment provisional on the legislature’s agreement with it. The broader appointments question, of how far the executive may go in designing the bodies that check it, is the same fault line that runs through the litigation over judicial appointments to the constitutional courts themselves, traced in the account of the NJAC Act and the collegium judgment; the tribunal cases are that argument conducted at a lower altitude and over a much larger number of offices.

The constitutional question, separated from the merits

Strip the tribunal content away and the question in Writ Petition (Civil) 1018 of 2021 is one that arises wherever a constitutional court has struck down a statute: what may the legislature do next? Indian law’s answer is old, stable and twofold, and the 2025 judgment is an application of it rather than an invention.

The first half is permissive. A judgment ordinarily decides a case on the law as it stands, and the legislature, which makes the law, may change the law on which the judgment rests. A court construes a word in a taxing statute; the legislature may amend the statute to define the word otherwise, retrospectively if it chooses, and the judgment then has nothing to operate on. A court finds a provision missing that the Constitution requires; the legislature may supply it. A court strikes down a levy because the statute lacked a validating feature; the legislature may enact a validating Act that cures the defect and reimposes the levy. The classic statements are in Shri Prithvi Cotton Mills Ltd v Broach Borough Municipality in 1969, which upheld a validating Act that removed the defect the Court had identified, and in Indian Aluminium Co v State of Kerala in 1996, which set out the principles governing validating legislation: the legislature may render a judgment ineffective by altering its basis, retrospectively, provided it acts within its competence and does not transgress constitutional limits. The State of Tamil Nadu v State of Kerala decision of 2014, on the Mullaperiyar legislation, applied the same analysis to uphold a statute that altered the legal position on which an earlier decree rested. This half of the doctrine is not a concession extracted from legislatures. It is the ordinary working of a system in which courts interpret and legislatures enact, and it is used constantly and uncontroversially.

The second half is prohibitory. The legislature may not, by a mere declaration that a judgment is wrong or shall be deemed never to have been given, set the judgment aside while leaving the law unchanged. That is not legislation but adjudication, and adjudication is not a legislative function. The line was drawn in the Cauvery reference of 1992, where an Ordinance that purported to nullify an interim order of a tribunal constituted under a parliamentary law was held to be an impermissible legislative judgment on a dispute, and it is reflected in a line of cases, including Madan Mohan Pathak v Union of India and Municipal Corporation of Delhi v Gurnam Kaur, holding that a legislature cannot overrule a decision of a court by simply declaring it ineffective. The test that emerges from the two halves is about what the new law does. If it changes the underlying legal position so that the judgment, however correct on the old law, no longer governs, it is legislation. If it leaves the legal position exactly where it was and differs from the struck down provision only in its label, its location or its overriding formula, it is an assertion that the court was wrong, and it fails.

Applying that test to the Tribunals Reforms Act was not difficult, and the difficulty the case presented was institutional rather than analytical. The defects identified in 2020 and July 2021 were specific: a term too short to secure independence, an age floor that excluded the class the Court had admitted and deterred the class it sought, and a selection device that transferred the effective choice to the executive. A legislature answering those judgments permissibly had real choices. It could have enacted a longer term and defended it as sufficient. It could have replaced the age floor with an experience requirement measured in years of practice, which would have met the stated policy of seniority without the exclusionary effect. It could have designed a committee with a judicial majority, or an independent commission, and given it the appointment. It could have sought a reference to a larger bench, or a review, on the footing that the earlier decisions were wrong in principle, which is the route the dissent of Justice Gupta in July 2021 would have supported. What it could not do, on the doctrine, was enact the same term, the same age and the same panel, in a new Act, behind a clause overriding any judgment. That is the sense in which the 2025 judgment is about method rather than about tenure. The Court did not hold that four years is an unconstitutional number in the abstract for all time. It held that a provision struck down for a stated constitutional defect cannot be re-enacted with the defect intact, because the second enactment does not engage with the judgment at all; it merely contradicts it.

Two clarifications keep the doctrine from being overstated, and the 2025 judgment made both. The first is that the Court cannot dictate the contents of a statute. The bench expressly disavowed any power to prescribe the text of a law or to mandate a policy choice, and drew the distinction as one between directing legislation, which is forbidden, and reviewing legislation, which is indispensable. The directions in the tribunal cases sit on the reviewing side of that line as the Court drew it: they are the specification of the constitutional minimum that a reviewing court found the impugned scheme to fall below, stated so that a compliant scheme can be written. The second clarification is about the source of the standards applied. The bench held that the norms on tenure, age limits, selection and independence are not abstract judicial preferences but constitutional requirements distilled from Articles 323A and 323B read with separation of powers, independence of the judiciary and Article 14, and that a statute designing the justice system may be tested against structural principles of that kind. Whether that distillation is right is the live jurisprudential question in the case, and it is the point on which the Union’s argument, and Justice Gupta’s earlier dissent, were directed. It should be recorded as a question on which the Court has spoken with finality in this litigation, not as a proposition beyond argument in the abstract.

The doctrine applied to the three provisions, one at a time

The general test becomes clearer when it is run against each of the three provisions separately, because each fails it in a slightly different way and each had a compliant alternative of the same policy family.

Take the term of office. The defect identified in November 2020 and again in July 2021 was not that four is a small number but that a four year term, combined with reappointment at the executive’s pleasure and retirement ceilings that in several tribunals left a member well short of even that, made tenure insecure in a way the Court held unconstitutional for a body exercising judicial power. A legislature persuaded that five years is too long had an altered basis available to it: a single non-renewable term of four years, for example, with a bar on subsequent government employment, would have changed the dependence structure that the judgments condemned, because the vice lay in the renewal and not only in the number. It could also have differentiated by tribunal, giving the heavy appellate bodies a longer term and the smaller authorities a shorter one, and defended the differentiation on the work each body does. Section 5 did neither. It re-enacted four years, renewable, on the same ceilings, which is repetition.

Take the minimum age. The defect identified was exclusionary and deterrent in effect: it shut out the advocate of ten years of practice whom the Court had admitted, and it confined the judicial membership of tribunals in practice to retired judges and retired officials. A legislature persuaded that tribunal members need seniority had an altered basis available: a requirement of a stated number of years of practice, of judicial service or of work in the relevant field, measured by experience rather than by birthdays, would have achieved seniority without excluding the class the judgment had brought in, and would have been defensible as a qualification rather than an age bar. The proviso to Section 3(1) did not do that. It repeated fifty.

Take the panel of two names. The defect identified was the transfer of the effective choice from the committee to the executive, aggravated by a committee without a judicial majority. A legislature persuaded that the appointing authority needs a choice had altered bases available: a committee with a judicial majority whose recommendation of a single name is binding unless returned once for stated reasons, or a panel accompanied by a duty to record reasons for departing from the committee’s order of preference, or a commission independent of the ministries altogether. Section 3(7) adopted none of them. It repeated the bare panel.

In each case the pattern is the same, and it is the pattern the series comparison below generalises: the policy objective survived the judgments and could have been pursued; the particular instrument chosen to pursue it had been held unconstitutional for a stated reason; and the Act chose the instrument again. That is the line between a legislative answer and a legislative repetition, and it is a line about drafting choices rather than about the distribution of power in the abstract.

The four years between the Act and the second strike down

The Act was challenged within weeks of its enactment, and the petition then waited four years for a decision, a period during which the tribunal system was administered under a statute whose central provisions two benches had already, in substance, condemned in their Ordinance form. Appointments were made under Chapter II. Selections begun before the Act were completed under it, which produced the class of members whose protection the eventual judgment had to arrange. Benches sat with members whose tenure the Act had fixed at four years and whose reappointment depended on the process the petition attacked. The delay had causes that the parties disputed: the Union sought adjournments at stages of the final hearing, and the bench in November 2025 recorded its displeasure at repeated adjournment requests by the Attorney General in a case challenging the Act, a matter of public reporting of the hearings. The petitioners, for their part, pressed the point that a statute under challenge on this ground should not be allowed to govern appointments by default while the challenge aged.

The interregnum is itself part of the subject’s lesson, because it shows what a strike down cannot do. A judgment of November 2025 can restore the directions of 2020 and 2021 as the governing law for the future, and can protect defined classes of appointees, but it cannot give back the four years in which selections were made, deferred or declined by candidates who would not accept a four year term under a statute of doubtful validity. The strongest form of the petitioners’ argument about vacancies belongs here rather than in the merits: whatever the correct design is, a system in which the governing design is under a constitutional cloud for four years will lose candidates, postpone selections and accumulate arrears, and that cost is incurred by litigants who were not parties to the argument. It is also the practical reason the Court’s earlier insistence on time-bound appointments, three months from recommendation, was not a procedural nicety. Delay in this field is not neutral between the branches. It operates in favour of whichever design is provisionally in force, and in this period that was the executive’s.

The judgment of 19 November 2025

Madras Bar Association v Union of India and Another, Writ Petition (Civil) No. 1018 of 2021, was decided on 19 November 2025 by a bench of the Chief Justice of India, B. R. Gavai, and Justice K. Vinod Chandran, and is reported with the neutral citation 2025 INSC 1330. The Chief Justice authored the principal judgment; Justice Chandran concurred in a separate opinion. The Court struck down Sections 3, 4, 5, 6 and 7 of the Tribunals Reforms Act 2021.

The holdings can be stated in the Court’s own sequence. The Indian constitutional framework, the Court held, subscribes neither to parliamentary sovereignty nor to unqualified judicial supremacy; its architecture is rooted in constitutional supremacy. Parliament must operate within the bounds of the Constitution, respecting separation of powers, fundamental rights and structural values such as judicial independence that form part of the basic framework. Once the Court has struck down a provision or issued binding directions after identifying a constitutional defect, Parliament cannot simply override or contradict that decision by re-enacting the same measure in a different form. The provisions of the impugned Act could not be sustained: they violate the constitutional principles of separation of powers and judicial independence embedded in the text, structure and spirit of the Constitution; they directly contradict binding judicial pronouncements that had repeatedly clarified the standards governing appointment, tenure and functioning of tribunal members; and instead of curing the defects identified, the Act merely reproduces, in slightly altered form, the very provisions earlier struck down. That, the Court held, amounts to a legislative override in the strictest sense, an attempt to nullify binding judicial directions without addressing the underlying constitutional infirmities, and because the Act fails to remove the defects identified in the prior judgments and instead re-enacts them under a new label, it falls afoul of the doctrine of constitutional supremacy. The Court rejected the Attorney General’s submission that Parliament has discretion to ignore the decisions of the Court, and found the Union’s defence to be, in substance, a reprise of arguments already considered and rejected in the earlier cases.

Three features of the remedy matter as much as the strike down. First, the Court restored the earlier position as the operative law: until Parliament enacts a law consistent with the prior judgments, the guidelines and directions in the judgments of 27 November 2020 and 14 July 2021 govern appointments, tenure and conditions of service in the tribunal tier. Those directions include the five year term, the retirement ages of seventy for chairpersons and sixty-seven for members, the eligibility of advocates with at least ten years of practice, the recommendation of one name for each post, appointment within three months of recommendation, and a selection committee with judicial primacy. Second, the Court directed the Union of India to establish a National Tribunals Commission within four months of the judgment, and specified the principles the Commission must embody: independence from executive control, professional expertise, transparent appointment and administrative processes, and robust oversight mechanisms that reinforce public confidence. Third, the Court protected defined classes of appointments already made, in terms examined in the next section of this article. This is legal information about a public judgment, not advice on any individual appointment or case.

What did Justice Chandran add in his concurring opinion?

Justice K. Vinod Chandran concurred with the reasoning and the directions, and added the sentence by which the judgment will be remembered outside the reports: the Act, he wrote, is a replica of the struck down Ordinance, old wine in a new bottle, where the wine whets not the judicial palate but the bottle merely dazzles.

The concurrence matters beyond the epigram, because it locates the vice in the replication itself rather than in any single provision, which is the doctrinal core of the majority’s holding. A reader who takes only the concurrence away has the case in miniature: the container changed, from an amendment of the Finance Act to a stand-alone statute, and the contents did not.

The structural point about the Commission deserves emphasis, because it is the only part of the judgment that builds rather than forbids. A National Tribunals Commission had been recommended by the Law Commission of India in its 272nd Report of 2017 on the assessment of the statutory frameworks of tribunals, urged by Justice Chandrachud’s concurrence in Roger Mathew, directed as an interim objective in November 2020, and left uncreated through the whole of the period in which four successive schemes were litigated. The 2025 direction differs from its predecessors in carrying a period: four months from 19 November 2025. What such a commission would have to contain, on the principles the Court stated, is an appointments function removed from the sponsoring ministries, a cadre and budget of its own, authority over infrastructure, staffing and the rationalisation of benches, and an oversight function over the administration of the tribunals that does not extend to their adjudication. The direction is a direction to establish the body; the design of a compliant statute remains Parliament’s, and proceedings concerning the implementation of the judgment are the route by which its performance is tested.

The members already appointed: protection and its limits

Every strike down of an appointments scheme creates a class of people who took office under it, and the treatment of that class was litigated as hard as the scheme itself. The 2025 judgment’s treatment rests on a record that begins in 2018. In Kudrat Sandhu, the Court’s interim directions, consented to by the Attorney General, provided that appointments made before the new framework would be governed by the parent Acts and the rules under them. The 2020 judgment struck down the retrospective operation of the 2020 Rules. In the proceedings of 2020 and July 2021 the Attorney General expressly submitted that appointments of chairpersons and members made prior to the enactment of the impugned framework would stand protected. The 2025 bench recorded that submission, at paragraph 143 of the judgment, and recorded also that the appointment orders under challenge did not honour it.

The concrete grievance came from the Income Tax Appellate Tribunal and the Customs, Excise and Service Tax Appellate Tribunal. Members selected in 2018, in processes completed before the Act, received their appointment orders only in 2021, by orders dated 11 September 2021 and 1 October 2021 in the Income Tax Appellate Tribunal’s case, and were subjected on the strength of those dates to the Act’s truncated tenure and altered conditions. The Court held those orders to be inconsistent with the Attorney General’s earlier statements and directed that the service conditions of the members so appointed be governed by the old Act and the old Rules, meaning the parent statute and the position before the impugned framework. It clarified that the age of superannuation for the 2018 selections in the Income Tax Appellate Tribunal and the Customs, Excise and Service Tax Appellate Tribunal would be sixty-two years for members and sixty-five years for chairpersons, as those parent schemes provided. It then generalised the protection: all appointments of chairpersons and members whose selection or recommendation by the Search-cum-Selection Committee was completed before the commencement of the Tribunals Reforms Act, but whose formal appointment notifications issued after the Act came into force, are protected, and continue to be governed by the parent statutes and their conditions of service rather than by the Act’s truncated tenure and altered conditions. The brief’s framing of the point, that in the earlier proceedings an assurance protecting appointments already made had been given and in the present proceedings the appointments made were found inconsistent with it, is the accurate summary: the protection in 2025 was not a new indulgence but the enforcement of a position the Union had itself taken in court.

The limits of the protection are as important as its terms. It is addressed to selections completed before the Act and notified after it. It does not confer the Act’s benefits on anyone, and it does not validate appointments made under the struck down provisions on their own terms. And it leaves the general body of serving members in the position the restored directions create: their tenure and conditions are governed by the parent statutes as read with the 2020 and 2021 directions until Parliament legislates afresh. Stability of tenure and protection of vested rights, the Court said, are essential components of judicial independence, and its earlier directions on the subject could not be lightly departed from. That sentence does double duty in the judgment. It explains why the members were protected, and it explains why the Act’s treatment of term and age was not a neutral administrative choice.

What the judgment did not disturb

A recurring error in published accounts of the November 2025 judgment is to describe the Tribunals Reforms Act as struck down, without qualification, and to infer that the abolished tribunals revive or that a practitioner with an intellectual property appeal should return to a board that no longer exists. The judgment does not do that. The challenge in the writ petition was directed at Sections 3(1), 3(7), 4, 5, 6, 7(1) and 33. The provisions struck down were Sections 3 to 7, the service code. The abolition chapters, Chapters III to XXV insofar as they dissolve bodies and transfer their work, were not struck down, and the challenge to Section 33, the transitional provision that terminated the offices of the Second Schedule bodies, did not succeed. The Intellectual Property Appellate Board, the Film Certification Appellate Tribunal, the Airport Appellate Tribunal, the Authority for Advance Rulings, the Plant Varieties Protection Appellate Tribunal and the Customs appellate authority for advance rulings remain abolished, and their work remains where the Act put it, in the High Courts, the Commercial Courts and the district court in the highways scheme.

One drafting consequence of a partial strike down should be noted, because practitioners will meet it in the parent statutes. The parent statutes of the tier contain provisions, inserted or substituted over the years, that refer the service conditions of their tribunals’ members to Chapter II of this Act. Those references remain in the statute book, and the chapter to which they point has lost its operative sections. The working position is supplied in the way the judgment directs: the parent statutes, as they stood apart from the impugned scheme, read with the restored directions of November 2020 and July 2021, govern appointments and conditions until Parliament enacts a compliant law. Consolidations and commercial databases that present the Tribunals Reforms Act as a current statement of tribunal service law are, in that respect, out of date in a way that matters, and the error is common enough in published material to be worth naming: the Act’s service provisions are frequently described as the governing position by sources that do not record the judgment of 19 November 2025.

The distinction is not a technicality; it is the whole shape of the outcome. Parliament’s power to decide which forums exist was never in issue. What was in issue was Parliament’s power to set the conditions of the forums that remain by re-enacting conditions already held unconstitutional. A legislature that wants fewer tribunals may abolish them, subject to the ordinary constitutional limits on abolishing a forum without providing an effective substitute. A legislature that wants the tribunals it has kept to be staffed on terms the courts have struck down must either cure the defect or persuade a court that the earlier holdings were wrong. The 2025 judgment leaves the first power intact and denies the second shortcut. Reporting that collapses the two misleads practitioners in both directions: it tells the intellectual property litigant that his forum may return, which it will not on this judgment, and it tells the serving tribunal member that the Act stands, which in the respects that govern his tenure it does not.

What a compliant scheme would have to contain

The judgments, read together, amount to a specification from which a drafter can work, and setting it out is the most useful thing an account of this Act can do for a practitioner advising on the next statute. A scheme for the tribunal tier that answers the case law of 2010 to 2025 would contain the following elements, each traceable to a holding rather than to a preference.

Appointments would be made on the recommendation of a body in which the judicial voice predominates, chaired by the Chief Justice of India or his nominee with a casting vote, and that body would recommend one name for each post, with a wait-listed name, the Government being required to appoint within three months or to return the recommendation once with reasons recorded. Eligibility would be defined by qualification and experience, including a route for advocates with at least ten years of practice to be appointed as judicial members, and for members of the Indian Legal Service on the same footing, and would not be defined by a minimum age. The term would be five years, with retirement at seventy for a chairperson and sixty-seven for a member, and reappointment, where allowed, would follow the same selection process with preference given to service rendered. Salary and allowances would be charged or otherwise protected against variation to disadvantage, and house rent would be met at a level the 2020 judgment required where accommodation is not provided. Removal would require an inquiry and a recommendation from a body independent of the sponsoring ministry. Administration, budget, infrastructure and the posting of members would vest in a National Tribunals Commission independent of the ministries that litigate before the tribunals, with a tribunals wing as an interim arrangement. And the whole would be enacted in primary legislation, or in rules made under a delegation that itself states these requirements, so that the Roger Mathew objection of unguided delegation does not revive.

Two cautions belong with that specification. The first is that it is a floor derived from holdings on particular schemes, not a code the Court has enacted; a Parliament that wishes to design differently may do so, provided it can show that its design secures the independence the Constitution requires, and the burden of that showing is heavier after 2025 than it was before 2017. The second is that the specification is silent on the questions that will decide whether the tier works: the number and location of benches, the staffing of registries, the case management powers of chairpersons, the quality of the technical members available in fields such as accountancy, valuation and environmental science, and the enforcement of the tribunal’s own orders. Those are administrative and legislative questions on which the case law has little to say and on which the Commission, if built, will be judged. This article is legal information about a statute and the judgments on it, and not advice on any individual appointment or case.

Four legislative answers: the pattern across the series

This series has documented Parliament answering the courts in several fields, and the comparison is the fastest way to see what decided this case. Read together, the episodes yield a rule rather than a tendency, and the tribunals episode is the one that completes the set. Where Parliament displaced a judicial construction of a statute by changing the statutory text on which the construction rested, as it did when the competition amendment of 2023 redefined the turnover on which penalties are calculated after the Court had read the word down to relevant turnover, the answer stood and was not struck down. Where Parliament supplied the provision a court had found missing and thereby removed the foundation of a judicial direction, as it did in the Scheduled Castes and Scheduled Tribes amendment of 2018, which answered directions that had conditioned arrest and prosecution on preliminary enquiry and prior approval by writing a contrary scheme into the Act, the answer stood, the amendment was upheld, and the Court recalled the directions it had issued. Where a constitutional question was put to the Court and decided on the Constitution’s own terms, as in the Article 370 judgment, there was no legislative route around the holding at all, and none was attempted. And where Parliament re-enacted, without curing the defect, provisions that had been struck down on constitutional grounds, as in the Tribunals Reforms Act, the answer failed. The variable that separates the first two episodes from the fourth is not the subject matter, the size of the majority or the temper of the times. It is whether the legislature altered the legal basis or merely repeated the provision.

Stated as a working rule for anyone drafting or challenging legislation, the pattern is this. A judgment that rests on a statute can be answered by amending the statute, and the amendment will be judged on its own constitutional merits. A judgment that rests on the Constitution can be answered only by satisfying the Constitution, which usually means curing the defect identified, or by amending the Constitution itself within the limits of the basic structure, or by persuading the Court in review or before a larger bench that it was wrong. What cannot be done is to treat a constitutional holding as though it were a statutory construction, and to answer it with a non obstante clause. The Tribunals Reforms Act is the case study because its authors had every alternative open to them, chose the one the doctrine forecloses, and did so in the face of a judgment delivered five weeks before the Bill was introduced.

Vacancies: the cost of a decade of unsettled design

The practical consequence most often attributed to this litigation is the vacancy position in the tribunals, and it should be stated with attribution and with its causes separated. The Supreme Court itself made the point in July 2021, recording that a large number of vacancies of members and chairpersons and the inordinate delay in filling them had resulted in the emasculation of the tribunals. In September 2021 a bench headed by the Chief Justice of India, N. V. Ramana, took the Union to task over unfilled posts across the tribunal system and over the enactment of the Act while the litigation was pending, and the Union thereafter cleared a round of appointments to the National Company Law Tribunal and the Income Tax Appellate Tribunal. In December 2021, finding the Debts Recovery Tribunals and their appellate tribunal without presiding officers in several places, the Court requested the High Courts to entertain under Article 226 the applications that would otherwise have gone to those tribunals until the appointments were made, which is as close as a court can come to recording that a statutory forum has ceased to function. In proceedings in December 2022 it was placed before the Court that of forty-one names recommended for the Income Tax Appellate Tribunal in 2019, twenty-two had been approved and appointed, a figure that measures the selection pipeline rather than the litigation. Press and parliamentary reporting across the period repeatedly recorded substantial fractions of sanctioned strength unfilled in the company law and tax tribunals, the two with the heaviest commercial dockets.

Three causes run together in those figures and should not be merged. The first is the litigation itself, which unsettled the governing scheme four times in six years and repeatedly changed the terms on which appointments were to be made, so that selections begun under one scheme were completed, or stalled, under another. The second is executive delay within whatever scheme was in force: recommendations not acted on within the directed period, names neither accepted nor returned, and selection processes not initiated in advance of foreseeable retirements, which is a failure that no judgment caused and several judgments condemned. The third is supply: for several technical posts the eligible field is thin, and for judicial posts the terms on offer, a short term at a distance from a member’s home bench with no certainty of continuance, deter exactly the candidates the scheme most needs. A vacancy figure quoted without that separation is an argument disguised as a statistic. What can be said on the record is narrower and stronger: the Court identified vacancies as a constitutional problem in its own right in 2021, the executive’s own pipeline was shown in 2022 to have converted fewer than half of a completed set of recommendations into appointments, and a design dispute about tenure was conducted for a decade over tribunals that were, for much of it, short of the members needed to sit.

Should tribunals exist at all

Underneath the service conditions dispute sits an older question that recurs whenever a tribunal fails, and the stance of this article requires that it be argued both ways rather than assumed.

The case for tribunals is a case about volume and expertise. The Indian state generates disputes of a volume that the ordinary courts have never been resourced to absorb: tax assessments by the lakh, service disputes of government employees, debt recovery for the banking system, company and insolvency proceedings that require decisions in weeks if the underlying business is to survive. A specialised forum can be designed for that volume, with members drawn from the field, procedures simpler than the civil procedure code, and benches sitting where the disputes arise. The 272nd Report of the Law Commission of India, which assessed the statutory frameworks of tribunals in 2017, proceeded on that footing: it did not propose abolition of the tier, but uniformity of qualifications, a nodal agency, and appointments and conditions that would make the tier work. On this view the failures in the record are failures of staffing and design discipline, and the remedy is the Commission the Court has directed, not a retreat to the High Courts.

The case against is a case about what has actually been built. Tribunals in India are administered by the ministries whose decisions they review, staffed through processes those ministries influence, housed in premises those ministries provide, and dependent on them for budgets, staff and even the notification of their own members’ appointments. Their members are frequently retired judges and retired officials for whom the post is a final appointment, their tenures have been short and repeatedly rewritten, and their decisions are in any event subject to the writ jurisdiction of the High Courts, so that the tier adds a stage to litigation without adding finality. On this view, the decade from 2017 to 2025 is not an accident but a disclosure: a forum whose independence must be litigated for in every generation is a forum whose design invites the litigation, and the honest alternatives are to give the tier genuine constitutional protection or to return its work to the courts, as the Act did for the five bodies it abolished. The abolitions themselves are evidence that can be read either way. They show a government willing to close tribunals it considered useless, which supports the view that the tier is a policy instrument to be pruned. They also show work returning to High Courts whose dockets were the original reason for taking it away, which supports the view that abolition without judicial capacity merely relocates delay.

The verdict this article takes is confined to the ground the judgments have settled, and it is a verdict about method. On the constitutional question, the 2025 judgment is right, and it is right for the reason the doctrine gives: a system in which a struck down provision can be re-enacted behind an overriding clause is a system in which no constitutional judgment on legislation is final, and the separation of powers objection runs against the legislature in that case, not against the court. On the institutional question, the record supports a harsher judgment on all concerned than either side’s advocacy admits. The executive wrote four successive schemes that failed on grounds stated in advance. The legislature enacted the last of them five weeks after the provisions were struck down, in a session in which the Bill passed each House within days. The courts, for their part, governed the design of the tribunal tier by directions for a decade, an arrangement that no one defends as an ideal way to run a branch of the justice system and that persisted because the political organs would not build the institution, the National Tribunals Commission, that every judgment since 2019 identified as the way out of judicial administration of the subject. The Commission stands directed, with a period attached. Whether it is built, and built to the principles the Court specified, is the test by which the next decade of this subject will be judged, and it is a test of the executive and of Parliament rather than of the courts.

The afterlife of the abolished bodies’ case law

Abolition dissolves an institution; it does not dissolve its decisions. The orders of the Intellectual Property Appellate Board in trade mark rectification, patent revocation and copyright licensing remained operative after 4 April 2021, the proceedings behind the undecided ones moved to the courts under Section 33(3), and the Board’s reasoned decisions continue to be cited before the High Courts that inherited its jurisdiction, which treat them as the decisions of the predecessor appellate forum whose place they have taken. The same is true, on a smaller scale, of the film tribunal’s certification jurisprudence and of the advance rulings given by the Authority for Advance Rulings on transactions already ruled upon. For a researcher the consequence is a split archive: the law stated by the tribunals up to April 2021 in their reports and databases, and the law stated thereafter by the High Courts and Commercial Courts in theirs, with the transfer date as the seam. The seam matters in practice, because a proposition about patent revocation procedure or about the registrability of a mark may rest on Board authority that a search confined to court databases will miss, and because the Board’s technical member reasoning on questions of fact in the sciences has no exact equivalent in the courts that replaced it, one of the costs of the transfer that its defenders accepted and its critics predicted. Section 33’s transfer of records and property to the Central Government completed the institutional winding up; the intellectual winding up is the slower process by which a body of precedent is absorbed into the case law of the courts that succeeded the body that made it.

The paper trail of a decade of litigation

A dispute conducted over nine instruments generates a documentary record that practitioners and researchers have to work with directly: the Gazette text of the Act and of the Ordinance it repealed, the two sets of Rules, the parent statutes as amended, cause lists and orders of a line of judgments running from 2019 to 2025, and rosters showing which members of a tribunal were appointed under which scheme and on what terms. Much of that record circulates as scanned Gazette copies, word documents and spreadsheets rather than as clean text, and the working problem is often simply opening and reading a file in the form in which a registry, a colleague or a ministry has shared it. For that purpose a practitioner may use an office file viewer for opening spreadsheets, documents and presentations in the browser, which is the kind of tool this series points its readers to for the documentary side of the work. The substantive point is the one the litigation itself makes: in this field the primary materials are unusually accessible and unusually decisive, because the judgments quote the provisions they strike down and the statutes print the provisions that replaced them, so the whole controversy can be reconstructed from the record without dependence on commentary. Readers preparing for examinations, in particular, will find that the sequence makes most sense read in the order in which it happened, judgment against instrument, rather than summarised.

Frequently Asked Questions

Q: What did the Tribunals Reforms Act restructure?

The Act restructured the tribunal tier in two directions at once. It abolished five appellate bodies named in its Second Schedule, together with the Customs appellate authority for advance rulings, and transferred their work to the High Courts, the Commercial Courts and a district court. And it imposed a single code of appointment, tenure and service conditions, in Chapter II, on the sixteen tribunals listed in its First Schedule, replacing the varied provisions of their parent statutes. The abolitions stand. The service code was struck down in November 2025, so the restructuring that endured is the map of forums, not the code of conditions.

Q: Which tribunals were abolished and where did their work go?

The Second Schedule names the Film Certification Appellate Tribunal, the Authority for Advance Rulings under the Income Tax Act, the Airport Appellate Tribunal, the Intellectual Property Appellate Board and the Plant Varieties Protection Appellate Tribunal, and the Act also abolished the Appellate Authority for Advance Rulings under the Customs Act and the tribunal under the national highways statute. Appeals in film certification, patents, trade marks, geographical indications, plant varieties, airport eviction and customs advance rulings went to the High Courts. Copyright Board functions went to the Commercial Courts, with Registrar appeals to the High Court. Highways appeals went to the principal civil court of the district.

Q: What conditions of service did the Act prescribe for tribunal members?

Chapter II prescribed a minimum age of fifty years for appointment as chairperson or member, appointment by the Central Government on the recommendation of a Search-cum-Selection Committee, a panel of two names to be recommended for each post, a term of four years with retirement at seventy for a chairperson and sixty-seven for a member, eligibility for reappointment with preference for service rendered, removal on stated grounds on the committee’s recommendation, and salary and allowances set by rules at the level of a Central Government officer holding a post carrying the same pay, protected against variation to disadvantage after appointment.

Q: Why was the tribunals Act described as a legislative override?

Because its central provisions repeated, in a new statute, provisions the Supreme Court had struck down weeks earlier. The minimum age of fifty, the four year term and the panel of two names had been declared unconstitutional in the judgment of 14 July 2021 on the Ordinance, and the Act, passed in August 2021 and effective from 4 April 2021, enacted the same age, term and panel in Sections 3 and 5, each prefaced by a clause overriding any judgment. The Court in 2025 called this a legislative override in the strictest sense: an attempt to nullify binding directions without curing the defects identified.

Q: What did the Supreme Court decide about the tribunals Act?

In Madras Bar Association v Union of India, decided on 19 November 2025 by the Chief Justice of India, B. R. Gavai, and Justice K. Vinod Chandran, the Court struck down Sections 3 to 7 of the Act, holding that they violated separation of powers and judicial independence and fell afoul of the doctrine of constitutional supremacy by re-enacting provisions already struck down. It restored the directions of November 2020 and July 2021 as the governing position, directed the Union to establish a National Tribunals Commission within four months, protected defined classes of appointments already made, and left the abolition provisions untouched.

Q: Why did the Court direct a National Tribunals Commission?

Because every judgment in the line had identified the same structural defect: tribunals are administered by the ministries whose decisions they review, which control their budgets, staff, premises and appointments machinery. An independent commission, on the principles the Court stated, would take appointments, administration and oversight out of the sponsoring ministries, supply professional expertise and transparent processes, and give the tier an institutional home of its own. The idea had been recommended by the Law Commission in 2017, urged in Roger Mathew in 2019 and directed as an objective in November 2020. The 2025 judgment attached a period of four months.

Q: What happens to tribunal appointments already made?

The 2025 judgment protected the class whose position was in dispute. Appointments of chairpersons and members whose selection or recommendation by the Search-cum-Selection Committee was completed before the Tribunals Reforms Act commenced, but whose appointment notifications issued afterwards, continue to be governed by the parent statutes and their conditions of service, not by the Act’s truncated tenure. The Court applied this to the Income Tax Appellate Tribunal members appointed by orders of 11 September 2021 and 1 October 2021, and fixed superannuation for the 2018 selections in that tribunal and the Customs, Excise and Service Tax Appellate Tribunal at sixty-two for members and sixty-five for chairpersons.

Q: What does the sequence show about tribunal independence?

The sequence shows that independence in the tribunal tier is a function of design details rather than of declarations. A short term with reappointment controlled by the executive, an age floor that confines appointment to the end of other careers, and a selection device that reserves the effective choice to the Government were each held, in combination, to make members dependent on the litigant most often before them. It also shows that the dependence was structural rather than personal: no finding was made against any member. The remedy the judgments built was equally structural, covering tenure, eligibility, selection and, at the end, an independent commission for administration.

Q: What is the difference between removing the basis of a judgment and re-enacting a provision that was struck down?

Removing the basis means changing the law on which the judgment rests, so that the judgment, correct on the old law, no longer governs the new position: amending a definition a court construed, supplying a provision a court found missing, or curing the defect a court identified in a validating Act. Re-enacting means passing the same provision again, with the defect intact, under a new label or behind a clause overriding judgments. The first is legislation and is permissible, as cases from Shri Prithvi Cotton Mills in 1969 to Indian Aluminium in 1996 establish. The second is an assertion that the court was wrong, which is adjudication, not legislation, and it fails.

Q: Which sections of the Tribunals Reforms Act were struck down in the judgment of November 2025?

Sections 3, 4, 5, 6 and 7, the whole of Chapter II. Those sections contained the rule-making power over qualifications and conditions with the minimum age of fifty in Section 3(1), the Search-cum-Selection Committee and the panel of two names in Section 3, the removal provision in Section 4, the four year term and the retirement ages in Section 5, reappointment in Section 6, and the salary and allowances power in Section 7. The rest of the Act, including the abolition chapters and the transitional provisions of Section 33, was not struck down, and the challenge to Section 33 did not succeed.

Q: What tenure and retirement ages apply to tribunal members after the 2025 judgment?

The restored position is that of the directions issued on 27 November 2020 and 14 July 2021, which govern until Parliament enacts a compliant law. The term of office of a chairperson and of a member is five years. A chairperson holds office until the age of seventy and a member until the age of sixty-seven. Advocates with at least ten years of practice are eligible for appointment as judicial members. Appointments are to be made within three months of the selection committee’s recommendation, and reappointment follows the same process with preference given to service rendered. Individual parent statutes and the protection given to earlier appointees qualify the general position in particular cases.

Q: Why did the Court object to the selection committee recommending a panel of two names?

Because the panel transfers the effective choice from the committee to the executive. A committee that recommends one name performs the selection, and the Government’s role is to appoint or, exceptionally, to seek reconsideration. A committee that recommends two names performs a shortlisting, and the Government chooses between the two without any duty to record why the other was passed over. The objection was sharpened by the committee’s composition, in which the judicial voice was the chairperson alone against two nominated Secretaries and the tribunal’s own chairperson, the chairperson of the committee having only a casting vote. The direction in the earlier judgments was one name for each post, with a wait-listed candidate.

Q: How did the Finance Act 2017 change the law governing tribunal appointments?

Part XIV of the Finance Act 2017 did two things. It merged eight tribunals into cognate bodies, among them the Competition Appellate Tribunal into the National Company Law Appellate Tribunal and the Copyright Board into the Intellectual Property Appellate Board. And by Section 184 it empowered the Central Government to make rules, for the tribunals in the Eighth Schedule, on qualifications, appointment, term of office, salaries and allowances, resignation, removal and other conditions of service. That section moved the service conditions of the tribunal tier from the individual parent statutes, where Parliament had set them tribunal by tribunal, into a single executive rule-making power, and it was the foundation of the Rules of 2017 and 2020 and of the litigation that followed.

Q: What did the Constitution Bench decide in Roger Mathew about the Tribunal Rules of 2017?

Rojer Mathew v South Indian Bank Ltd, decided on 13 November 2019, struck down the Tribunal Rules of 2017 in their entirety. The Constitution Bench held that the rules failed to secure the independence the earlier judgments required, giving the executive a dominant voice in selecting members of bodies that decide disputes to which the Government is the most frequent party, and prescribing tenures too short for security. It upheld Section 184 of the Finance Act 2017 as a delegation, but only as conditioned by the constitutional requirements in the earlier case law, and it referred the Money Bill question concerning the Finance Act to a larger bench. It also recorded the need for a National Tribunals Commission.

Q: What did Section 33 provide for the staff, funds and property of the abolished bodies?

Section 33 wound each abolished body up completely. Officers and other employees who were serving on deputation stood reverted to their parent cadre, ministry or department from 4 April 2021. The balance of all money received by or advanced to the body and not spent by that date stood transferred to the Central Government. All property of whatever kind owned by or vested in the body stood transferred to and vested in the Central Government. Read with the termination of the officeholders’ tenure on compensation not exceeding three months of pay and allowances, and the transfer of pending proceedings to the courts, the section left nothing of the institution behind: no staff, no funds, no premises and no docket.

Q: Did the Supreme Court strike down the abolition of the tribunals in the 2025 judgment?

No. The writ petition challenged Sections 3(1), 3(7), 4, 5, 6, 7(1) and 33, and the Court struck down Sections 3 to 7, the service code. The abolition chapters were not struck down and the challenge to Section 33, which terminated the offices of the abolished bodies and transferred their work, did not succeed. The Intellectual Property Appellate Board and the other abolished bodies therefore remain abolished, and their work remains with the High Courts, the Commercial Courts and the district court in the highways scheme. Accounts describing the Act as struck down without qualification misstate the result, which was a strike down of the conditions imposed on the tribunals that remain.

Q: Who sits on the Search-cum-Selection Committee under the Act?

For a tribunal other than a State Administrative Tribunal, Section 3(3) constitutes a committee chaired by the Chief Justice of India or a judge of the Supreme Court nominated by him, who has a casting vote. Its other voting members are two Secretaries to the Government of India nominated by that Government, and one member who is the outgoing chairperson for a chairperson’s appointment, the sitting chairperson for a member’s appointment, or, where reappointment is in issue, a retired judge of the Supreme Court or a retired Chief Justice of a High Court nominated by the Chief Justice of India. The Secretary of the administering ministry is member-secretary without a vote. A parallel committee, chaired by the Chief Justice of the High Court, is prescribed for State Administrative Tribunals.

Q: Why did Justice Hemant Gupta dissent in the July 2021 Ordinance judgment?

Justice Gupta held that a law cannot be struck down merely because it is not in accordance with the precedents of the Court. In his view, directions issued in earlier judgments on matters such as the length of a term or a minimum age are remedial preferences on questions of policy, and they do not bind Parliament when it legislates within its competence; on that footing most of the challenge to the Ordinance was not legally sustainable. The majority, Justices Rao and Bhat, held the contrary: the directions were the operative content of findings that the earlier schemes were unconstitutional for stated reasons, so a provision repeating the defect repeated the unconstitutionality. The 2025 bench adopted the majority’s premise against an Act of Parliament.

Q: How could the Central Government remove a tribunal chairperson or member under the Act?

Section 4 required removal to be made by the Central Government on the recommendation of the Search-cum-Selection Committee, in the manner provided by rules, on five grounds: the person has been adjudged insolvent; has been convicted of an offence involving moral turpitude; has become physically or mentally incapable of acting; has acquired a financial or other interest likely to affect his functions prejudicially; or has so abused his position as to render continuance in office prejudicial to the public interest. For removal on the grounds of incapacity, conflicting interest or abuse of position, the person had first to be informed of the charges and given an opportunity of being heard. Section 4 was among the sections struck down in November 2025.

Q: What did the Court hold about the appointments of members selected in 2018 but appointed in 2021?

It held that they were protected by the position the Union had itself taken in the earlier cases. Members of the Income Tax Appellate Tribunal selected in 2018 received appointment orders dated 11 September 2021 and 1 October 2021, after the Act commenced, and were subjected to its shorter tenure. The Court held those orders inconsistent with the Attorney General’s submissions in the 2020 and 2021 proceedings that appointments made before the impugned framework would stand protected, and directed that the members’ service conditions be governed by the old Act and the old Rules. It generalised the protection to every appointment whose selection was completed before the Act but notified after it.