The continuing mandamus made the Court an administrator

The continuing mandamus is an unusual writ. A conventional mandamus tells a public authority what to do and ends. A continuing mandamus tells a public authority what to do, then keeps the file open and asks the authority to come back next month and report on progress. The technique was christened in Vineet Narain v Union of India in 1997-98¹ and was, at the time, a creative answer to a discrete institutional failure: the Central Bureau of Investigation was not investigating the Jain hawala case, and the Supreme Court could not get the investigation moving by issuing a single order. By holding the matter on its board and calling the Director of the CBI back month after month, the Court did what an ordinary mandamus could not do. It worked.

Twenty-five years later, the device has spread across the constitutional courts' docket. The forest matter in TN Godavarman² has been alive since 1996, generating thousands of orders on subjects ranging from sawmill licensing to the size of permitted timber lots. Bonded-labour cases run for decades. Prison-reform petitions return to the same bench every three months for a fresh affidavit. The Court's Covid-period writ on the oxygen-allocation crisis⁴ functioned, in effect, as a continuing mandamus over the allocation of a scarce medical commodity in real time. None of this was contemplated by the framers of Article 32. None of it has a doctrinal source other than the Court's own judgments. And almost none of it has been examined honestly against the question that ought to govern any judicial device: in which cases does this work, and in which cases does it fail?

The answer, on the evidence, is that the continuing mandamus works in a narrow class of cases and fails in a wider one. Where the underlying problem is a discrete statutory authority failing a measurable duty, the device produces results that an ordinary mandamus could not. Where the underlying problem is polycentric — many actors, many trade-offs, no single right answer, budget implications that the Court cannot price — the device strains the Court's competence and converts it into an administrator without an administrator's tools. The Court has not drawn the distinction, and the failure to draw it is itself a doctrinal cost. This essay traces the device from its origin, sets out where it has worked and where it has not, and argues that the conditions for its proper use can be stated.

The origin: a workable improvisation

Vineet Narain was a hawala case. A petitioner had brought to the Court evidence that senior politicians and bureaucrats had received payments from Jain brothers' hawala diaries; the CBI had refused to investigate. The traditional remedies available were narrow. The Court could issue a mandamus directing the CBI to investigate, but it could not, by a single order, force an investigation through to its conclusion against a reluctant agency. It could issue a writ of certiorari, but there was no decision to quash. The petition lay in a familiar zone: an executive agency was failing a statutory duty, and the executive above it had no incentive to intervene.

Justice JS Verma's bench, in 1997-98, did three things at once. It directed the CBI to investigate. It kept the matter on its board. And, by stating that it was retaining seisin until the investigation reached an honest conclusion, it converted the writ into a supervisory order. The Court called this a 'continuing mandamus' and tied it to its plenary power under Article 32. The decision identifies the device by name and stakes the jurisdictional claim. Almost everything that followed in this field rests on the brief and confident statement in Vineet Narain that the Court had this power.

The case is the strongest single example of the device's value. Its conditions are worth listing. The respondent was a specific statutory agency, not an open-ended class. The duty was discrete: investigate, file a chargesheet, prosecute. The benchmarks were measurable: the case-diary was either being maintained or it was not; the investigation was either proceeding or it was stalled. The remedy did not require the Court to balance competing public interests; it required the Court to refuse to allow one public interest — the impunity of senior officials — to suppress another. And the political ecology was a particular one: a Government of India that had inherited the case from a predecessor regime and was happy, on the whole, to let the courts force a result it did not wish to be seen forcing itself.

These conditions matter because they are not present in most of the cases that followed.

The expansion: from agency to ecosystem

The earliest expansion of the device was into environmental cases, and it began before Vineet Narain. The Dehradun limestone-quarrying case, Rural Litigation and Entitlement Kendra v State of Uttar Pradesh³, was a continuing mandamus in everything but name from the early 1980s onwards. Justice PN Bhagwati's bench visited the Doon Valley, set up a monitoring committee, ordered the closure of some mines and the rehabilitation of others, and held seisin of the matter for years. Shyam Divan, surveying the field for the Oxford Handbook, treats it as the forerunner. The Court was already, by the late 1980s, doing the work the continuing mandamus would later be asked to describe.

What Vineet Narain did was to give the technique a name and a doctrinal anchor in Article 32. What TN Godavarman did, between 1996 and the present, was to push it beyond anything the device's name suggests. Godavarman began as a petition against the felling of timber in a small forest tract in the Nilgiris. It has, since December 1996, expanded into a roving mandate over India's forests. The Court has, in Godavarman, defined what counts as a 'forest' for the purposes of the Forest (Conservation) Act 1980. It has prohibited the movement of timber across state lines without its permission. It has constituted a Central Empowered Committee — a body unknown to the statute — to advise it on whether forest land should be diverted to non-forest use. It has, through that Committee, become, in practice, the licensing authority for forest land in India.

None of this is, strictly, what Vineet Narain contemplated. The CBI is a single agency; the forests of India are an ecosystem of state forest departments, mining companies, panchayats, tribal communities, central ministries, and statutory authorities, with no single Director the Court can summon and instruct. The 'duty' the Court has been enforcing in Godavarman is not a statutory duty in the conventional sense — it is the Court's own evolving view of what the Forest Conservation Act, read against the right to a healthful environment in Subhash Kumar, requires the executive to do. The Central Empowered Committee — whose constitution Divan describes as an institution that has 'gained near-institutional status over a decade of functioning' — has no statutory mandate other than the Court's own order continuing it after its original five-year term ended.

The result is what one might call Godavarman jurisprudence: a body of law on Indian forests that exists only in the orders of the Supreme Court, that binds the Union and the states, that creates a parallel administrative regime, and that has no exit. The case has not been disposed of in nearly thirty years. The Court that took it up in 1996 is not the Court that hears it today; the device has outlived the personalities who created it.

Where the device works

Before turning to the harder cases, it is worth being concrete about where the continuing mandamus does what its defenders say it does. Four conditions, drawn from the cases, mark out the zone in which the device produces results.

First, the respondent is a discrete authority with a statutory mandate. The Central Vigilance Commissioner in Centre for Public Interest Litigation v Union of India⁸, the CBI in Vineet Narain, the Election Commission in cases on the disclosure of candidates' antecedents, the High-Powered Committee on the appointment of integrity-institution heads, the petroleum ministry in Common Cause v Union of India¹⁰ on the discretionary allotment of petrol pumps. The Court can identify a single counterparty and hold it accountable to a single text.

Second, the duty is binary or close to it. Investigate or do not investigate. Disclose or do not disclose. Recover or do not recover. The Court is not asked to choose among policy options; it is asked to determine whether one option — non-performance of a duty owed in law — is permissible.

Third, the benchmarks are observable. An affidavit will tell the Court whether something has happened. A status report will record the progress. The Court does not need to develop expertise in the underlying subject matter to know whether the order is being complied with.

Fourth, the remedy is exit-ready. The Court can say at the outset what would close the matter — the chargesheet has been filed, the appointment has been made, the antecedents have been disclosed. The continuing mandamus is, in this class of cases, a device with a planned end. It is not a permanent office.

The CVC appointment matter is the clean illustration. The Court in 2011 was asked whether the appointment of the Central Vigilance Commissioner had been made with due regard to the eligibility criteria and the institutional integrity that the office demanded. It held that the High-Powered Committee had failed to consider relevant material, set aside the appointment, and laid down the standards for future appointments. The order ended. The institution moved on. The device had done what the device is for.

Where the device strains

The harder cases share none of those four features. The Covid oxygen-allocation litigation in the second wave was a case in which the Court — in fairness, with no other institution willing to take the call — undertook to supervise the distribution of medical oxygen across states in real time. The respondent was not one agency. The duty was not a single statutory mandate but the entire executive function of pandemic management. The benchmarks were not observable in the way that an investigation is observable; they required scientific judgments on hospital demand, supply-chain logistics, and budgetary allocation. And the remedy had no obvious exit: when does a pandemic end, and at what point does the Court stop allocating oxygen?

The Court did, in fact, do useful work in those weeks. It pushed states to share data they would not otherwise have shared. It forced the Union government to disclose its allocation formula. It compelled a level of transparency that the bureaucratic process had not produced. None of this is denied. But the question is not whether the Court did some good in that crisis; it is whether the device the Court used was the right one for that work, and what the cost of using it was.

The cost was structural. Lon Fuller's classic distinction between bipolar and polycentric problems⁹ remains, fifty years later, the cleanest framework for thinking about this. A bipolar problem is one with two parties and a discrete legal question between them. A polycentric problem is one in which many parties are connected, in which a decision affecting one of them reshapes the position of the others, and in which the consequences of a particular outcome cannot be reasoned to in advance from a legal premise. Fuller's claim — modest, defensible — was that adjudicative process is well-suited to bipolar problems and poorly suited to polycentric ones. Oxygen allocation in a pandemic is polycentric in Fuller's exact sense. Distribution to one state means non-distribution to another. Reservation of supply for one hospital means non-reservation for another. The Court can rule on procedure, on transparency, on the duty to share data; it cannot, without becoming the allocator, decide who gets the oxygen.

The same observation applies to Godavarman. A decision to permit a particular diversion of forest land has consequences for the state's revenue, for the displaced communities, for the carbon balance, for the downstream water table, and for the eight or nine departments and agencies that will be affected by the project that follows the diversion. The Central Empowered Committee, set up to advise the Court, is now the polycentric decision-maker the Court itself cannot be. The Court has not stopped being involved; it has simply outsourced the substance of its supervision to a body it created. This is the device's strain made visible: the Court keeps the mandamus running, the underlying decisions migrate to a committee outside the constitutional architecture, and the question of who is in charge stops being answerable.

The institutional case against the line being drawn

The strongest defence of the present practice is institutional. India does not have a functional ombudsman. Its environmental tribunals are limited in jurisdiction. Its administrative regulators — RBI, SEBI, CCI, TRAI — operate in sectoral silos and do not address general-administrative failures. Its parliamentary committees do not have the authority or the staff to monitor compliance over years. In this institutional landscape, the argument runs, the continuing mandamus is the only mechanism by which a constitutional body can hold the executive to account beyond the first order. To insist that the Court draw a line — to confine itself to the bipolar cases — would, on this view, be to insist that a working remedy be withdrawn because it has been deployed in some cases where it should not have been.

This is the steel-manned version of the position and it deserves to be taken seriously. TV Somanathan's chapter on the regulatory state in the Oxford Handbook makes the related observation that Indian regulatory bodies are weaker than their counterparts in mature jurisdictions, that the legal aspect of regulation has 'come to acquire a very large influence over the regulatory process, even overshadowing the technical aspect', and that this has happened in part because the institutions that ought to handle the technical and political aspects have not done so. The continuing mandamus, on Somanathan's reading, is a symptom of an institutional vacuum that the Court has been pulled into filling. The Court did not seek the role; it was the only institution available.

There is force here. But the institutional defence proves something narrower than its proponents claim. It proves that there is real work to be done, and that no other institution is doing it. It does not prove that the continuing mandamus is the right device for that work. The cost of substituting judicial supervision for ordinary administration is not zero. The administrative state, deprived of the responsibility for hard calls, gets worse at making them. The Court, asked to make hard calls outside its competence, gets worse at the calls it is competent to make — the bipolar legal questions on which doctrinal coherence depends. The same Court that hears Godavarman in the morning hears bail applications in the afternoon. The judges who supervise the Central Empowered Committee are also the judges who decide constitutional questions on which the law turns. The supervisory work is not free; it is paid for in attention diverted from the work the Court alone can do.

Jalan and Rai, in their chapter on review of administrative action, point to a parallel cost. Article 14 review has, over the years, drifted from a doctrine of equality to a general supervisory power over executive reasonableness — and the Court has, as they show, struggled to articulate a 'judicially manageable test or standard' for that supervision. The continuing mandamus is the procedural counterpart of that drift. The substantive law has become open-ended; the procedural device has become open-ended; and the cumulative effect is a Court whose reach is now defined by what it chooses to engage with rather than by what the Constitution authorises it to engage with.

What the line would look like

A principled use of the continuing mandamus would limit it to cases that satisfy the four conditions identified above: a discrete respondent, a binary or near-binary duty, observable benchmarks, an exit. It would refuse the device in cases where the underlying problem is polycentric in Fuller's sense — where the Court cannot reason from a legal premise to a single right outcome, and where supervising the outcome would require the Court to choose between trade-offs that the Constitution has assigned to political and administrative organs. In those cases, the Court can do what an ordinary writ permits: declare the right, identify the duty, set the procedural standards. It cannot, without converting itself into something else, supervise the implementation indefinitely.

This is not a programme of judicial retreat from rights enforcement. The Court would retain — must retain — the power to enforce fundamental rights against egregious failures. The line is between enforcement and administration. Vineet Narain enforces. The CVC appointment matter enforces. The disqualification-of-candidate-with-criminal-antecedents litigation enforces. Godavarman, by contrast, administers. The Covid oxygen orders administered. The river-interlinking project that emerged from a continuing mandamus⁶ — flagged by Divan as a 'disastrous' judgment — administered, badly. The Aravali Golf Club bench⁷ warned, briefly and to little effect, that the Court was drifting into that latter zone. The warning has not been heeded.

One implication of the line is that some cases the Court now hears as continuing mandamus would have to be handed off to other institutions or — if no other institution can take them — left to the legislative and political process. The forest matter, on this view, should have been resolved through the Forest Conservation Act and its successor regimes; the failure of those regimes is a political failure, not a judicial deficit that the Court is obliged to remedy. The Covid oxygen problem should have been resolved through the Disaster Management Act 2005 and the constitutional machinery for centre-state coordination; the failure of that machinery, in a public-health emergency, is a political failure that the Court cannot mend through orders. There is real loss in this position. Citizens whose rights are being violated by administrative failure will, on the line proposed, sometimes be told that the Court cannot help them. The discomfort is not denied. But the cost of refusing to draw the line — a Court that increasingly cannot tell, even to itself, where its competence ends — is the larger one.

The unfinished doctrinal task

What is striking is how little doctrinal writing the continuing mandamus has produced. Vineet Narain creates the device; Godavarman, Manohar Lal Sharma⁵, and the prison-reform line use it. No constitution bench has examined the device and laid down the conditions on which it may be deployed. The Court that uses the mandamus most often is also the Court that has not been asked, by any litigant or by any government, to articulate when it would refuse to use it. Manohar Lal Sharma, when the Union tried to argue that Section 6-A of the Delhi Special Police Establishment Act limited the Court's continuing mandamus, was dismissed in two paragraphs: a statutory provision could not bind 'the plenary power of this Court of issuing orders in the nature of a continuing mandamus under Article 32'. The plenary-power formulation is the formulation of an unreviewable power. The Court has not, since 1998, given that power a doctrinal shape.

It is not impossible to give it one. Raeesa Vakil's chapter on jurisdiction in the Oxford Handbook describes the broader pattern: the Court has, across many of its powers, refused to lay down guidelines for the exercise of its discretion, preferring to invoke 'phrases such as grave injustice or shocking the conscience of the court'. As HM Seervai's observation, quoted by Vakil, has it: 'to coin a phrase is not to indicate a standard'. The continuing mandamus sits within that broader pattern. The Court has the phrase. It has not articulated the standard.

The standard, if articulated, would have three elements. The first is a threshold question of whether the case is the kind of case in which a single order will not deliver the right — that is, whether ordinary mandamus would fail. The second is a competence question of whether the supervision the Court is undertaking can be discharged through legal reasoning or whether it requires the Court to make polycentric decisions for which it has no tools. The third is an exit question of whether the Court can specify, at the time of issuing the supervisory order, what would close the file. A case that fails any of the three is not a case for continuing mandamus.

The doctrinal exercise is not academic. The device was created in 1998 by a small bench in a single case. It has, since then, reshaped the Court's relationship to the executive. The Court that signed off on Vineet Narain was a court of last resort against impunity; the Court that runs Godavarman and oxygen allocation is an administrator with case files going back a quarter of a century. Both selves of the Court exist in the same institution, and both speak with the same authority. The doctrinal task is to tell them apart.

The continuing mandamus, properly confined, is one of the more useful institutional innovations of the post-Emergency Court. It does, in cases such as Vineet Narain and the CVC appointment matter, what the orthodox writs cannot. The case for keeping it is straightforward; the device meets a real need. The case against using it as it is now used is also straightforward; the device, deployed beyond its competence, makes the Court an administrator, deprives the executive of accountability for hard calls, and produces orders the Court itself cannot enforce. The line between the two uses is not impossible to draw. The Court has, in twenty-five years, not drawn it. The failure to draw the line is not an absence of doctrine; it is a doctrine, by silence, that the device has no limits. That is the doctrine this Court should be willing to disown.

Get the week's biggest decisions, distilled.

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