The Master of the Roster is the Most Powerful Office in Indian Law

The most powerful office in Indian law is not held by the Prime Minister, the Attorney General, or the seniormost advocate at the Bar. It is held by the Chief Justice of India, and the power in question is not the power to write judgments. The power in question is the power to decide which judges will write which judgments. This power is called, by its holders and its admirers, the master-of-the-roster doctrine. The doctrine has no constitutional anchor. It has no statutory frame. It is justified by reference to administrative convenience and judicial collegiality, and its limits, such as they are, are limits the office itself has declared and the office itself can revise. A power that decides outcomes is not administration. It is jurisdiction without a vesting clause. The settled view treats this as unproblematic. I want to say why the settled view is wrong, what the doctrine actually does, and what real limits would look like.

The settled view, taken at its strongest

Let me state the settled view before I argue against it. The defence of the master-of-the-roster doctrine has three parts and each is held in good faith.

The first is institutional. The Supreme Court of India hears, on the most recent figures, almost seventy thousand admission matters a year. It disposes of roughly seven and a half thousand regular hearings.¹ It sits in benches of two, three, occasionally five, very occasionally seven or nine. Someone has to decide which judges sit together on which bench, and which bench hears which case. The Court could not function if every assignment were litigated. The Chief Justice, as the seniormost administrative officer of the Court, is the only person institutionally placed to make those calls. To call this 'administrative' is not a euphemism; it describes what the function literally is. A court of thirty-four judges cannot operate without a master of the roster, and no parliamentary statute or constitutional amendment has ever assigned the function to anyone else. The settled view says: this is how every appellate court works in the world. The Court of Appeal in England has a Senior President of Tribunals; the United States Supreme Court has a Chief Justice who assigns opinions; the Federal Court of Australia has a Chief Justice who allocates judges. The Indian Supreme Court has the Chief Justice of India. The function is unremarkable.

The second is doctrinal. The Court has held, repeatedly and through benches of three judges and five, that the master-of-the-roster power is a settled feature of the institution. In Asok Pande v Supreme Court of India, a bench led by Dipak Misra CJ confirmed that the Chief Justice alone has the authority to constitute benches and assign cases, and that no individual judge can take that work onto himself.⁶ In Shanti Bhushan v Supreme Court of India, a Constitution Bench reaffirmed the doctrine months after the unprecedented press conference of 12 January 2018 in which four sitting judges of the Court raised concerns about case allocations.⁷ The argument is not new and the Court has had its chance to reconsider. It has chosen not to. Stare decisis, at the level of doctrine, is settled.

The third is comparative. Every system has to choose between two failure modes. Either the assignment power sits with one office, in which case that office can be criticised for arbitrariness. Or the assignment power is distributed, in which case the system can be criticised for incoherence. Every comparator jurisdiction has chosen the first failure mode. The objection that the Chief Justice has too much power is, on this view, an objection that any system of central case allocation has too much power, and the only alternative on offer is one no one runs.

Take these three propositions together and the case for the doctrine looks formidable. The Court is overworked, the function is necessary, the rule has been reaffirmed, and the comparator courts run the same way. I want to argue that each leg of this defence is half right and the conclusion still wrong, because the defence does not engage with what the master-of-the-roster doctrine actually does in the Indian system, as opposed to what it claims to do.

What the doctrine actually does

The argument against the doctrine begins not with what the Chief Justice does on any given Monday morning when he draws up the roster, but with what the Indian Supreme Court is. It is, as Nick Robinson has insisted in the only sustained empirical study of the question, a 'polyvocal court'¹ — a court that does not, in practice, speak with a single voice. Pathak CJ in Union of India v Raghubir Singh recommended that division benches should as a general rule comprise at least three judges 'for the purpose of imparting certainty and endowing due authority'; the recommendation has not been followed.⁴ Article 141 of the Constitution provides that 'the law declared by the Supreme Court shall be binding on all courts within the territory of India',¹² but the law declared by a two-judge bench on Tuesday is regularly inconsistent with the law declared by another two-judge bench on Thursday. The Indian Supreme Court is, on any given week, thirteen or fourteen co-equal smaller courts producing parallel doctrine in parallel benches, with the formal claim that all of it is the law of the land. Upendra Baxi has called this an 'assembly of empanelled judges'; Rajeev Dhavan, more pointedly, has noted that the Court's judges lack 'precedent consciousness'.¹

Now place the master-of-the-roster doctrine inside this picture. If the Court spoke with one voice — if the Court sat en banc, or if every important matter went to a constitution bench of five or more — then which bench heard which case would be a question of scheduling. The substantive law would be the same whichever bench was assigned. In a polyvocal court, this is not the case. Different benches decide cases differently. Different benches have different tolerances for the State, for the Bar, for the petitioner, for the policy. The choice of bench is the choice of doctrine. The choice of doctrine is the choice of outcome. And the choice of bench, in our system, rests with one person.

The empirical record on this point is striking. Robinson's analysis of constitutional benches since Independence finds that the Chief Justice of India has been six and a half times less likely to be in dissent than other judges on constitution benches.² That number is not, on a moment's reflection, surprising. The Chief Justice constitutes the constitution bench. The Chief Justice chooses who sits on it. The Chief Justice puts the case up. The Chief Justice presides. A litigation system that produces a six-and-a-half-times dissent gap between the office that composes the bench and the office's colleagues on the same bench is not a system in which the master-of-the-roster doctrine is doing administrative work. It is a system in which the doctrine is doing substantive work, and the substantive work is the achievement of pre-determined outcomes.

This is not a hypothetical concern. The list of cases in which the constitution of the bench was, with hindsight, the case itself, is long enough to constitute its own literature. The defence of Asok Pande and Shanti Bhushan says: in all of these cases, the assigned bench reached the decision it was always going to reach, and the choice of bench was incidental. The empirical record says: in many of these cases, a different bench would have reached a different decision, and everyone involved — the petitioner, the State, the senior advocate, the bench, the Chief Justice — knew it. The choice of bench was not incidental. It was the litigation strategy.

The doctrine has no constitutional anchor

The second problem with the master-of-the-roster doctrine, beyond what it does, is what it is. It is not in the Constitution. It is not in the Supreme Court Rules in any form that constrains the power. It is in Asok Pande and Shanti Bhushan, both of which are decisions of the Court itself, both of which were authored by benches the Chief Justice of the day had constituted to hear them. The doctrine is, in this strict sense, self-declared.

This matters because the Indian Supreme Court has, in every other domain, been alive to the basic-structure objection to self-declared power. The Three Judges Cases¹⁰ — and the 2015 judgment striking down the Ninety-ninth Amendment and the National Judicial Appointments Commission — were grounded in the proposition that the appointments process must not be exclusively executive, because such exclusivity would breach judicial independence. The reasoning of those judgments cannot be lifted out of context and read narrowly. It rests on the more general proposition that a power which affects the composition of the Court must be checked by something other than the discretion of the office that wields it. Once that proposition is accepted for appointments, it is hard to see why it does not apply to bench allocation. The composition of a Court matters at two moments: once when the judges are selected for the Court, and again when the judges are selected for the bench. The first selection is checked by the collegium and, in principle, by judicial review. The second selection is checked by nothing.

The standard reply is that bench allocation is internal court management and judicial review of it would be self-defeating — the Court cannot review its own administrative decisions about who sits on which bench. This is true as a procedural matter. It is not a reason to accept the doctrine. It is a reason to seek structural alternatives that do not require ex-post judicial review at all. The fact that you cannot litigate the roster does not mean the roster should be unconstrained. It means the constraint, if any, has to be built in.

The deeper point is that the doctrine sits in a place no other power in Indian constitutional law sits. Article 124(2) tells us how Supreme Court judges are appointed. Article 145(3) tells us how many judges sit on a constitution bench. Article 141 tells us the legal force of a Supreme Court judgment. Article 142 tells us the Court can do 'complete justice'. None of these provisions says anything about which judges of the appointed set sit on which bench to deliver which judgment. The Constitution is silent on the question. Into that silence the Court has read, by judgment of its own, a plenary discretion in the Chief Justice. The reading is not absurd — someone has to do the work — but it is also not constitutionally constrained, and it is not, in any meaningful sense, drawn from constitutional text.

The Dipak Misra moment and what it told us

If the doctrine is a settled administrative convenience, the four sitting judges of the Supreme Court who held a press conference on 12 January 2018 to complain about its operation should not have done so. They did. Justices Chelameswar, Gogoi, Lokur and Joseph released a letter they had earlier written to the Chief Justice of India alleging 'less than desirable' assignments of important matters.⁸ One of the matters specifically alluded to was the bench allocation in the petition concerning the death of Judge BH Loya. Other matters had been mentioned in earlier correspondence. The letter was not a legal pleading. It was an institutional warning, made by judges who could not, by virtue of their office, file a writ petition against their own Chief Justice.

What the press conference revealed was something the doctrine's defenders do not engage with: that the four most senior judges of the Court, the very judges who would sit on the collegium with the Chief Justice for purposes of appointments, did not regard bench allocation as a merely administrative function. They regarded it as a function whose exercise affected substantive outcomes, and on their account, the exercise in those months had been irregular enough to warrant going public. They were not arguing that the Chief Justice did not have the power. They were arguing that the power, exercised without constraint, was producing patterns they could not endorse.

The Court's response in Asok Pande and Shanti Bhushan was to reassert the doctrine and to reject the implied invitation to constrain it. The bench that did the reasserting was, of course, constituted by the same Chief Justice the four judges had publicly criticised. The argument is circular and the doctrine is structurally indefensible. A power exercised by an office cannot be reviewed by a bench composed by the same office; that is not a procedural defect but a definitional impossibility. The four judges who held the press conference knew this, and that is why they held the press conference. There was no legal remedy available within the system to address what they were describing. There still isn't.

The defenders of the doctrine treat the January 2018 event as an outlier, an institutional crisis brought about by the temperaments of particular judges and a particular Chief Justice. I read it differently. I read it as a structural feature of the doctrine becoming visible because the office in question had begun to use the power in ways that other judges of the same Court were unable to tolerate in silence. The doctrine had always permitted what those judges complained of. What changed was that the holder of the office did the thing the doctrine permitted at a volume the institution could no longer absorb. The structural defect was always there. The press conference was a tell.

The counter-argument addressed honestly

The strongest reply to all of this concedes that bench allocation is consequential and that the Chief Justice's power is large, but argues that the alternative is worse. Let me state it in its most serious form, because a polemic that ducks its target's best case is shouting.

The argument runs as follows. Any rule-based alternative to chief-justice discretion runs into a familiar problem. The rule could be random — assign cases to benches by lot. But random assignment cannot accommodate subject-matter expertise (the tax judge for tax cases, the criminal-side judge for criminal appeals), cannot accommodate workload balancing (some benches are overburdened, others under-burdened), and cannot accommodate urgency (the vacation bench for emergency matters does not exist by lot). The rule could be subject-matter-coded, with cases routed to standing benches by classification — environmental cases to bench A, taxation to bench B, criminal appeals to bench C. But subject-matter coding produces silos, predictable bench reasoning, and what the American literature calls 'panel-shopping' by counsel who time their filings to particular rosters. The rule could be that all important constitutional questions go to a constitution bench of five or more — that is, the framers' Article 145(3) test, properly applied.⁹ But this would, as Robinson notes, require either many more judges or many fewer matters, and the Court's docket cannot bear either change in the medium term.¹

I take this seriously. I take it seriously enough to say that I do not propose to abolish the master-of-the-roster doctrine. The doctrine is necessary. The function is real. The question is not whether the Chief Justice should assign cases. The question is whether the assignment power should be exercised entirely at the Chief Justice's discretion, or whether it should be exercised within a structured framework whose departures from default require recorded reasons.

The serious reply also concedes too much to the institutional convenience claim. Yes, the Court is overworked. Yes, the function must rest somewhere. None of this answers the empirical question: why does the Chief Justice end up six and a half times less likely to be in dissent on constitution benches than other judges? The convenience explanation does not predict that number. The agenda explanation does. A doctrine whose practical operation produces the agenda effect is not, by virtue of being convenient, also benign.

What real limits would look like

The case for the doctrine is at its strongest in the abstract and at its weakest at the point where the abstract meets practice. The case against the doctrine is at its strongest where it asks for specific, modest, structural limits. Let me set out four.

First, default rules with recorded departures. The Court could adopt, by its own rules, a presumptive bench-allocation table — for example, that all matters under Article 32 in which a fundamental rights question is principally raised go to the bench of the senior-most three judges available; that all challenges to constitutional amendments go to a constitution bench of five composed by seniority; that all special leave matters within a defined value or subject classification go to the relevant subject roster. Departures from the default would be permitted, but would require a recorded order of the Chief Justice setting out reasons, served on the parties, and placed on the Court's website. The Chief Justice's discretion would remain. What would change is that the exercise of discretion would now be visible. Visibility is not judicial review, but it is a discipline.

Second, the constitution bench should not be the Chief Justice's bench. The Court has held in Central Board of Dawoodi Bohra Community that any larger-bench decision binds smaller benches.³ It follows that the constitution bench is the structurally most important bench in the system. The Chief Justice should not, as a matter of practice, sit on his own constitution benches as the presiding judge whenever he so chooses. A simple rule — the Chief Justice does not sit on constitution benches save in cases involving the basic structure or the structure of the judiciary itself — would defuse the agenda-setting risk without disturbing the doctrine.

Third, the reference power needs structure. At present, two judges who disagree with a smaller-bench precedent may refer the matter to a larger bench; the larger bench is then constituted by the Chief Justice. The 2002 decision in Pradip Chandra Parija confirmed that the reference goes through the Chief Justice.⁵ The structural change required is straightforward: once a reference is made, the larger bench must be composed by seniority, not by discretion. The Chief Justice can preside or not. The composition is not the Chief Justice's call.

Fourth, the in-house procedure must permit a senior-judges committee, sitting independently of the Chief Justice, to call up specific bench allocations for review where four or more of the seniormost puisne judges request it. This is what the January 2018 press conference was, in substance, asking for: an institutional channel through which the office of the Chief Justice could be constrained by his immediate colleagues without the dispute spilling onto the lawns of the Court. There is no constitutional barrier to such a channel being created by Supreme Court Rules, which the Court itself may amend under Article 145.

None of these proposals abolishes the master-of-the-roster doctrine. None of them transfers the function elsewhere. What they do is convert an unconstrained discretion into a structured discretion. The structured discretion is administratively defensible in a way the unstructured one is not.

What changes if this is right

If the master-of-the-roster doctrine is structurally indefensible as currently practised, three things follow.

The first is that constitutional litigation in India is, at the most fundamental level, not a contest between rival readings of the law. It is a contest between rival readings after a separate, prior contest between rival proposals about who reads. That prior contest happens in a forum with no rules of evidence, no published reasons, no judicial review, and no appeal. To say this is not to say the Court is corrupt; it is to say that the structural place where the most important decisions get made is the place where the system has placed the fewest constraints. A constitutional democracy that organises its highest court this way is making a particular bet about the trustworthiness of one office. The bet may be vindicated by character. It will not be vindicated by structure.

The second is that the polyvocal-court problem is downstream of the roster problem, not upstream of it. The standard diagnosis of polyvocality, drawing on Baxi and Dhavan, treats it as a problem of bench size and precedent-consciousness — that two-judge benches produce uncertain doctrine, and that judges do not always read what other judges have decided. This is true so far as it goes. But the polyvocal Court is only a problem because the bench you get matters. If every constitutional case went to a constitution bench composed by seniority, the polyvocality complaint would shrink. The doctrinal incoherence that Article 145(3) was designed to prevent has come back not because the constitution-bench rule failed but because the master-of-the-roster doctrine has made the question of which bench hears the case the antecedent question. Fixing one without the other will not work.

The third is that the case for reforming the Court — fewer admission matters, more constitution benches, clearer guidelines under Article 136¹¹ — depends on first reforming the office that decides which bench gets which case. The Law Commission has recommended, several times, that the Court adopt structural reforms in its docket. Those recommendations cannot be implemented by the Court collectively if their implementation depends on the discretion of one office that benefits from the status quo. The Chief Justice has, on the current rules, no institutional incentive to reform the master-of-the-roster doctrine. The reform must come from the constitutional text or from a collective rule of the Court that binds future Chief Justices.

The case for the doctrine, in the end, is the case for trusting the office. The case against it is the case for structuring the office. The latter is not anti-judicial. It is the same case the Court itself made, against the executive, in the Three Judges Cases. The Court argued there that the appointments power could not be left to the discretion of one branch because discretion without structure is rule without law. The argument was correct. It applies to bench allocation with the same force. The master of the roster is the most powerful office in Indian law because it is the only office in Indian law whose central power is exercised without rules. The remedy is not to take the power away. The remedy is to give it rules.

Get the week's biggest decisions, distilled.

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