The Collegium Cannot Be Reformed Without Being Replaced

There is a pattern in Indian constitutional argument that you can set your watch by. A political government proposes an answer to the judicial appointments question. The proposed answer is graceless, badly drafted, and shaped by short-term political incentives. The Court strikes it down. The lawyers cheer. Editorial pages call it a defence of the Constitution. Two years pass. The collegium continues to do what the collegium has always done - select judges in private rooms, communicate decisions in single-line orders, recall recommendations without explanation - and the same lawyers begin to complain, in different forums, about exactly that. The cycle has been running, more or less unchanged, since 1981.⁴ I want to argue that the cycle is the problem. Both sides of the standard debate are right about their diagnoses and wrong about their prescriptions. NJAC was a clumsy answer to a real failure. The 2015 reflex that struck it down was a clumsier defence of an indefensible system. The collegium cannot be reformed in place because there is nothing inside it to reform. The doctrine that built it - that judicial independence requires a judicial veto over judicial appointments - has to go first.

The settled view, stated as honestly as I can state it

Let me put the strongest version of the defence on the table before I argue against it. The argument I am about to attack is held in good faith by people who know more constitutional law than I do, and it has more force than its critics usually concede.

The defence runs roughly as follows. The framers built Articles 124 and 217 against a particular memory: the Government of India Act 1935, under which the executive had final say over judicial appointments and judges held office during pleasure. The framers replaced pleasure with fixed tenure during good behaviour, and they wrote into both Articles a requirement that the President consult the judiciary before appointment. The word 'consultation' was placed there, as Justice Khehar reconstructed in the NJAC judgment, precisely to constrain executive choice. The First Judges Case in 1981 hollowed that constraint out. The Second Judges Case in 1993 restored it by reading 'consultation' as 'concurrence' in substance, with the Chief Justice's view binding the executive.¹ The Third Judges Case in 1998 refined that structure into the collegium as we know it.² When Parliament in 2014 tried to dismantle it through the Ninety-ninth Amendment and the National Judicial Appointments Commission Act, the Court struck the package down because giving two unelected 'eminent persons' an effective veto over a unanimous judicial recommendation - and giving the Union Law Minister a vote at the same table - amounted to handing the executive a hold on the judiciary that the Constitution does not contemplate.³

The defence has more institutional force than this short version suggests. Indira Gandhi's government superseded three judges of the Supreme Court in 1973 and a fourth in 1977; the names are Shelat, Hegde, Grover, and HR Khanna, and the lesson the system drew from those four supersessions is that a Parliament with a majority and an Executive without restraint can reach into the bench whenever it wants to.¹² Justice (Retd) BN Srikrishna's defence of the collegium, in the Oxford handbook chapter on judicial independence, returns to those four names more than once.⁹ The collegium, on this view, is not an institutional design choice. It is a scar.

It is also, on this view, a doctrinally narrow scar. The defenders are careful to insist - and they are not wrong - that the collegium does not exclude the executive from the appointments process. Memoranda of procedure issued by the Ministry of Law and Justice after the Three Judges Cases describe a participatory mechanism in which the executive can return a recommendation once, with reasons, and a reiterated recommendation by the collegium binds. The system, then, is not a judicial monopoly in the formal sense. The executive participates, has a voice, can object, can delay. What it cannot do is impose a candidate the judiciary refuses. What it cannot do is veto a candidate the judiciary insists on, beyond a single round of reasoned objection. The defenders treat that limited executive role as proof that the system is participatory. The critics treat it as the proof that it is not. Both readings of the same set of facts are tenable. The choice between them is a constitutional choice about where the burden of persuasion lies. The 2015 majority resolved that choice by treating any retreat from the collegium as an attack on basic structure. That was the closure I want to argue against.

I take all of that seriously. Indira Gandhi's supersessions did happen. The 99th Amendment did contain bad clauses. Article 124C - the part that let Parliament regulate the appointments procedure by ordinary legislation - was a constitutional aggression that the Court was right to refuse.⁸ The two 'eminent persons' chosen by a Prime Minister, a Chief Justice, and a Leader of the Opposition - with no statutory criteria, no public reasoning, no removal mechanism - were a wide-open backdoor into the bench, and the Court was right to refuse them too. If the only question on the table in 2015 was whether the NJAC as drafted should survive, the answer was correctly given.

The trouble is that the Court did not stop at that answer. It went further, and the further it went, the more it gave away. The four concurring majority opinions did not merely strike down the impugned amendment. They wrote, into Articles 124 and 217, a constitutional requirement of judicial primacy that had survived as an interpretive choice for twenty-two years and was now elevated to a permanent feature of the constitutional order. That elevation foreclosed the design space in which any future reform could operate. The 2015 judgment is the moment the collegium stopped being a contested practice and became a doctrine. Doctrines, in Indian constitutional law, do not unwind. They accrete.

What the 2015 judgment actually held, and why it cannot bear the weight

Strip away the elegiac prose and the four concurring opinions in the NJAC judgment hold three distinct propositions. First, judicial primacy in appointments is part of the basic structure of the Constitution. Second, the basic structure is violated when the judiciary loses a controlling say over who joins it. Third, anything that gives the political branches an effective veto - including a single Law Minister at the table, including two unelected outsiders - is constitutionally fatal.

None of these propositions follows from the constitutional text. Article 124(2) and Article 217(1) require consultation with the Chief Justice. They do not require concurrence. They certainly do not require an appointments monopoly. Justice Chelameswar's dissent stated this with a clarity that the literature has worked hard to ignore: the basic feature is the absence of an executive monopoly over the bench, not the presence of a judicial one.⁶ Both propositions can be true at the same time. Both protect independence. Only one of them is a power grab dressed as a doctrine.

The interpretive move that produced the collegium was small in scale and enormous in consequence. In Sankalchand, Chandrachud J said 'consult' meant a conference, not concurrence.⁵ By the Second Judges Case sixteen years later, 'consult' had become an obligation on the executive to accept the Chief Justice's view in substance. By the Third Judges Case it had become a binding decision of a four-judge plurality. By 2015 it had become a constitutional basic feature, immune to amendment. Each step was reasoned. Each step took its authority from the step before. None of them was textually compelled, and the cumulative product - a court that picks itself, removes its members from political accountability, refuses to publish the reasons for its choices, and treats every proposal to alter that arrangement as an attack on judicial independence - is what we live with now.

Where the settled view actually fails

Three failures need to be named, because each of them is invisible from inside the doctrine.

The first is conceptual. Judicial independence and judicial veto are not the same thing. A judge is independent when she can decide a case without fear of executive sanction, legislative reprisal, or financial punishment - the muniments the Constitution actually provides, fixed tenure, undiminishable salary, immunity for judicial acts, restricted grounds of removal, the bar on legislative discussion of conduct under Article 121. A judge does not become more independent by being chosen by other judges. She becomes a different kind of dependent: dependent on the seniority logic of the body that selected her, dependent on the Chief Justice who sits at the apex of that body, dependent on a system that has every incentive to reproduce its own preferences and very few to disrupt them. The collegium does not insulate the judge from power. It transfers the power that selects her from one branch to another. Independence in the constitutional sense - the sense Articles 121 to 125 and 217 to 221 actually protect - does not require that transfer.

The second failure is institutional. The collegium has now had three decades to demonstrate what an appointments process run by senior judges, with no published criteria, no record of deliberations, no statutory composition, and no external check, actually produces. The record is on the wall. Nick Robinson's chapter on judicial architecture in the same handbook lays out, with the politeness of academic prose, what the inside of the system looks like: appointments by recommendation that no one outside the room can audit, transfers that arrive without explanation, recalls that the system absorbs in silence, a senior-puisne hierarchy in which seniority does almost all the work that merit is supposed to do.¹⁰ The complaints about incestuousness and sycophancy that prompted the Ninety-ninth Amendment did not come from a hostile government. They came from inside the legal profession, from former judges, and from inside the Court itself. The defenders of the collegium answer those complaints by pointing again at 1973 and 1977. The answer does not address the complaint.

The third failure is the one the doctrine cannot see at all. The system the 2015 majority defended is not the system the framers designed. The framers designed an appointments process in which the executive proposed and the judiciary checked, with the Chief Justice's view treated with the seriousness of an institutional voice but not with the finality of a binding vote. That is what the text of Articles 124 and 217 says when you read it straight, without the gravitational pull of forty years of intervening cases. It is also what BR Ambedkar said the clauses meant in the Constituent Assembly debates - Khehar J's own reconstruction in 2015 conceded as much, before he reached the opposite conclusion. The Court in 2015 did not defend a constitutional design. It defended an interpretive overlay that it had itself built between 1993 and 1998. Calling that overlay basic structure was the move that closed the door.

The defence of the defence, taken at its strongest

Here is the response the steel-manned defender of the collegium will make to all of this, and it is not a weak response.

Indian politics is what Indian politics is. We do not have a culture of executive self-restraint. We do not have a second chamber that meaningfully checks majority government. We have a political class that has, on multiple occasions in living memory, demonstrated a willingness to use whatever institutional levers the Constitution gives it to reshape institutions to its convenience. The 39th Amendment, retrospectively validating a prime ministerial election; the Emergency, in which the Supreme Court itself crumpled; the 42nd Amendment, which tried to legislate a basic-structure-proof Parliament; the 99th Amendment itself, whose stated aim of 'transparency, accountability and objectivity'⁷ sat awkwardly next to a design that gave the Prime Minister two nominated members and no statutory criteria for them. In an environment of that kind, the argument runs, doctrine must compensate for what the political culture refuses to supply. The collegium is what compensation looks like when the alternative is capture.

I find this argument morally serious and constitutionally wrong. It is morally serious because the danger it points to is real. It is constitutionally wrong because it treats every reform proposal as a Trojan horse and so guarantees that no reform proposal can ever succeed. If the only acceptable appointments process is one in which the judiciary has the final word, then 'reform' means tinkering with which judges have the final word, when they have it, and how they communicate it. That is not reform. That is interior decoration. The doctrine has eaten the question.

And there is a second cost to the argument that the defenders do not register. A doctrine that justifies itself by pointing at the worst behaviour of the political branches has, over time, an inflationary effect on its own authority. If the only thing standing between the Constitution and the Emergency is a judicial veto over appointments, then a judicial veto over administrative tribunals begins to look indispensable too, and a judicial veto over electoral bonds, and a judicial veto over the conduct of investigations, and a judicial veto over the procedure of Parliament itself. L Chandra Kumar already gestured in that direction.¹¹ The collegium is not an isolated overreach. It is the appointments-stage instance of a larger pattern in which the Court treats every politically risky question as a basic-structure question and every basic-structure question as one only the Court is competent to answer. The pattern is internally coherent. It is also institutionally exhausting, and it is not what the Constitution sets up.

What replacing the frame would actually look like

Replacing the analytic frame does not mean restoring SP Gupta. SP Gupta was wrong in the same way that the Second Judges Case was wrong, in the opposite direction: it read the appointments clauses as giving the executive a free hand, when the text clearly contemplates a structured dialogue. The frame I am proposing is neither.

It looks like this. Judicial independence, properly understood, is what Articles 121 to 125 and 217 to 221 actually protect: tenure, salary, immunity, removal procedure, restricted grounds of discussion, restricted post-retirement avenues. Those guarantees should be strengthened, not relaxed. An independent appointments commission - whose composition is fixed by the Constitution itself, not by ordinary legislation in the way Article 124C contemplated - would draw from the judiciary, from the executive, and from a smaller, statutorily defined non-political element. The commission's decisions should be reasoned in writing, communicated to the Chief Justice and to the public in a manner that does not violate candidate privacy, and reviewable on narrow administrative-law grounds. The commission's composition should not give any one branch a numerical majority - which is what the Ninety-ninth Amendment in fact tried to do, and what the Court legitimately criticised, before going further and rejecting any non-judicial role at all.

None of this is novel. It is roughly the design that exists, with national variations, in the United Kingdom after the Constitutional Reform Act 2005, in South Africa under the Judicial Service Commission established by the 1996 Constitution, and in several Canadian provinces. None of these systems is perfect. None has produced a captured judiciary. All of them are constitutionally specified, statutorily framed, and publicly reasoned, which the Indian collegium is not.

A commission of this kind would have failed the 2015 challenge for the simple reason that it would not have looked like the NJAC. It would not have given the Law Minister a vote. It would not have allowed two undefined 'eminent persons' to veto a judicial unanimity. It would not have invited Parliament to set its procedure by ordinary law. It would have given the judiciary a structural role - probably a plurality, possibly a majority on the appointments stage itself, with non-judicial members participating in selection but not in veto - within a process whose criteria and reasoning were specified in advance and visible afterwards. This is not a compromise between the collegium and NJAC. It is a different question, asked in a different frame.

The vocabulary problem, and why it matters

There is a smaller point I want to make before the closing one, because it is the point that the doctrinal literature most consistently misses. The defenders of the collegium and its critics, when they argue, use the same words to mean different things. Both sides say 'judicial independence'. The defenders mean: a constitutional condition in which the judiciary is structurally insulated from the political branches, including at the moment of appointment. The critics mean: a constitutional condition in which the judge, once appointed, can decide cases without fear of the political branches. These two meanings are not the same. The first requires an appointments monopoly or near-monopoly. The second does not require anything of the kind. The Constitution, when it speaks of independence, is doing the second kind of work, not the first. The textual muniments - Articles 121, 124(4), 124(5), 125, 217 to 221 - all relate to the conditions of service and the conduct of office. They do not speak to the question of who gets to enter the office. That question is settled by Articles 124(2) and 217(1), and they require consultation, not concurrence.

The slide from one meaning to the other is what made the collegium possible. Once 'independence' came to mean 'structural insulation including at the appointments stage', the textual silence on appointments-stage primacy became, in interpretive practice, a textual command for it. This is the move I want every reader of this essay to be able to see. It is the small move that closed the constitutional design space. Reform of the collegium begins with insisting that the two meanings be kept apart. A judge who is selected by a properly designed commission - one in which the judiciary has a structural role but not a veto - is no less independent, in the sense the Constitution actually protects, than a judge selected by the collegium. The salary is the same. The tenure is the same. The immunity is the same. The grounds of removal are the same. The bar on legislative discussion is the same. What changes is the room in which the selection happens, and whether the public is allowed to know what happened in it.

Treating that change as a violation of independence is a category error. It is the error the 2015 majority committed, and the error every defender of the collegium has committed since. Naming it does not by itself produce a reform. But it removes the doctrinal padlock from the door behind which reform has to happen.

Why we will not get there

The honest closing of an essay like this one is not a programme. The honest closing is to say what the obstacle is. The obstacle is not Parliament. Parliament has shown no appetite for a properly designed commission - the Ninety-ninth Amendment was not a thwarted attempt at one, it was a different thing, designed to give a particular political coalition particular kinds of control. The obstacle is not the bar. The bar is divided, and parts of it have been pushing for a serious commission for fifteen years. The obstacle is the doctrinal frame the Court itself has built. So long as judicial independence and judicial appointments monopoly are treated as the same constitutional thing, every proposed commission will look, to the Court that has to clear it, like an attack on independence. The Court will strike it down. The cycle will resume.

The Court that wrote the 2015 NJAC judgment did not preserve the Constitution. It preserved its own interpretive history. The interpretive history is defensible step by step and indefensible as a whole. The next time a government attempts an appointments reform - and one will be attempted, because the political incentives to do so do not go away - the Court will be asked the same question again. The question has a different answer. The answer is that the system the Constitution actually wrote into Articles 124 and 217 is not the system the Court built between 1993 and 1998. The collegium was a temporary correction to a particular abuse. It has been institutionalised into a permanent doctrine and treated as the only design loyal to the framers. Neither is true. Saying so is the first thing reform requires. The rest is design work that other constitutional democracies have already done. We are simply choosing, judgment after judgment, not to do it.

One last word, because polemics that ignore their own implication are dishonest. I am not arguing for a system that lets the government of the day place its loyalists on the bench. I have no taste for that outcome and the Constitution has no warrant for it. I am arguing that the choice between a judicial monopoly and an executive monopoly is a false choice the Court itself has manufactured. The framers did not write either monopoly into the text. They wrote a structured dialogue, in which the judiciary's voice was protected by being placed at the front of the consultation requirement. That structure can be reconstructed as a properly designed appointments commission with public criteria, written reasons, and a fixed composition. The reconstruction is harder than defending the collegium and easier than designing the NJAC. The fact that we have, for a decade now, refused to attempt it tells you something about the doctrinal frame inside which we are operating. The collegium cannot be reformed without being replaced. Replacement does not mean abolition. It means returning to the constitutional design that the Court, in 2015, chose to call something else.

Get the week's biggest decisions, distilled.

A weekly digest for lawyers, students and the legally curious. Every Friday morning. Free.