National Judicial Reforms Commission: A proposal we ignore at our peril [Part I]

Speaking at the first Supreme Court Bar Association National Conference in Bengaluru, Justice BV Nagarathna of the Supreme Court suggested the establishment of a National Judicial Reforms Commission. She set this in a broader context: the government’s paradoxical role as a driver - perhaps the primary driver - of rising case pendency while expressing concern and anguish about it, the lack of institutional discipline in government agencies (settling disputes invites criticism or even disciplinary action, while litigation frivolity is seen as diligence) and a crumbling judicial infrastructure.

While these reasons are not new, they have seldom been set out with such directness. But the idea of a dedicated national-level Reforms Commission, one that sits outside established structures like the Law Commission, is more than something of a breakthrough. For an informal group (DAKSH) that has been working towards a coherent framework for judicial process reforms since mid-2024, Justice Nagarathna’s recommendation is a much-needed (and long-desired) shot in the arm.

In mid-2024 Justice Patel, then recently retired from the Bombay High Court and Senior Advocate Harish Narasappa (co-founder of DAKSH) met in Mumbai. They rapidly agreed that while organisations like DAKSH (and there are many) had done work with separate judicial nodes, nobody seemed to have attempted a nation-wide systemic overhaul across the whole judiciary. For DAKSH itself, working on data analysis and modelling was rapidly becoming a limitation. Continuing that first conversation later that year with a larger team in Bengaluru, where DAKSH is based, the initial group agreed that rather than spend time looking at different apparent ‘solutions’, it made sense to first agree on an overarching objective. Was this to reduce pendency? Expand digitisation? Something else altogether? Over the next few days, each member of the group set out their list of prioritised objectives. We did not exchange them.

The result was something of a surprise. Many of us anticipated that ‘reducing pendency’ might be the lead objective. It wasn’t. Instead, almost counter-intuitively, the staggering pendency was viewed as a problem to be addressed while moving to the final goal, which was to reduce the amount of time required for the disposal of any case, carefully calibrated for different case-types. Not all cases are the same; some, by their nature, take more time than others. What had never been attempted was to establish a framework of a typical time-to-completion of any particular case-class.

We also eliminated as pointless any discussion about judicial appointments, increasing the number of judges, improving the judge-to-population ratio and so forth. Not only was this a huge political issue, but on a cold reasoning, we believed that you could pack the judiciary to sanctioned capacity, but as long as you had antiquated, cumbersome and opaque processes, there was unlikely to be any meaningful change.

This led to our central realisation: that the judicial processes baked into the entire judicial system from case retrieval to classification and on to decision-making were so wildly inconsistent, incoherent and mutually incompatible that efficiency could not possibly be achieved without wholesale top-to-bottom process re-engineering. This meant, we believed, taking every single nugget of civil and criminal procedure and rationalising it for consistency across the entire judiciary.

The most fundamental and sweeping reform, we came to believe, was to ensure consistency and predictability in the backend handling of cases and the manner in which they were dealt with. Accepting that the actual decision will vary, the processes underlying that decision-making need to be, we believed, known, transparent, easy to navigate, predictable and entirely consistent throughout the country.

What horrified us was the sheer extent of inconsistencies we found. From the simplest and the most immediately visible to the farthest recesses of internal case classification systems, we found no two systems in sync. Across the judiciary, a basic and elementary requirement is to be able to retrieve case details from a publicly accessible website. We accept that individual High Courts may have their preferences about visual elements, but surely a search-and-retrieval system aimed at the litigant can be expected to be the same whether one is searching the Supreme Court databases or those of a High Court? Not so; every High Court has its own search syntax. The case types listed in the retrieval system are different. Even the security requirements - the infamous and fabulously annoying CAPTCHA - on a simple search form varied widely, with at least one website demanding a complex, multi-character alphanumeric input.

Now this is not to suggest that everything on a court website should be exactly the same. Of course, there will be and should be variations that are specific to that court (recruitment, history, resources and so on). But core functionality that affects every litigant everywhere must surely be common. Nothing in experience or logic suggests that a litigant in one court is never before another court or has no need to retrieve orders from other jurisdictions. As a beginning, this underscores our thinking that for core and essential functioning, the judiciary must be treated as one unified system, not a federated coalition of wildly individualistic sub-systems.

There persists one central myth to judicial ‘data’. Somehow, the thinking has taken root that this data is ‘proprietary’. It is not. There is no copyright to a judgment or an order. Every single order is public domain the moment it is signed and released. There is a difference too often occluded between what is ‘confidential’ and what is ‘proprietary’. The judiciary owns no data. It may keep some of it confidential and away from public disclosure for a while, but it cannot claim ownership of any judicial data. And most certainly not over orders and judgments. Automatically, this demands that access to judicial data must be consistently open.

It may be argued that the CAPTCHA exists to prevent random threats or automated bots from overloading the system. This may be true, but today’s technology has moved far beyond the input-random-characters control of the antiquated CAPTCHA we see on almost every single court website. The functionality is also seriously crippled; one should not have to repeatedly enter a CAPTCHA for every single query during a single session.

The problem of inconsistency and incoherences get infinitely worse as one digs deeper. Cases are filed across the length and breadth of the country, in courts at various echelons of the established tiered structure. But all cases are not the same. Nor is the civil versus criminal division sufficient. A more careful case classification is essential so that these can be assigned both in dedicated registry-level departments for checking and then placed before courts to which they are assigned. The problem is that the identical case is classified differently in different courts. It is one case type in one High Court and a completely different case type with a wholly different case type code in another court. Classification is at the root and it has a ripple effect: it affects the purpose for which a case is listed, when it is listed, before whom it is listed and how the assignment of cases between courts is decided. Even this is entirely inconsistent.

There are two powerful forces at play when a case gets listed. The first of these is the famous, or infamous, ‘Master of the Roster’ doctrine - the Chief Justice decides which bench hears what type or class of case. Self-evidently, this is also a type of case classification. But the problem in many High Courts is that this rostering classification is almost wholly unrelated to the case classification in the database. Every case thus has two classifications - one on filing and entered in the database; and another, frequently unconnected one, that appears in the roster.

This has led to the emergence of an entire sub-genre of legal learning: the interpretation of the roster. Planning a roster then becomes a nightmare and for no reason we are able to tell, the most obvious solution has remained unexplored: using a simple drag-and-drop modular system where modules linked to registry/filing-level case types are simply dropped into a bench’s roster. To make this even more efficient, artificial intelligence tools can be used to suggest which case types are best ‘clubbed’ together. This should be deployed country-wide for consistency and predictability.

Add to this another layer of complication: the daily uncertainty of getting an actual listing. Received wisdom would have it that while the Chief Justice (or the Principal District Judge or Principal Judge) is the ‘Master of the Roster’, each judge has complete dominion over her or his own listing: which of the cases assigned to that bench should be listed when, in what order of priority and for what ‘purpose’. Too often have judges been heard to say that it is they who will decide what gets listed, on what date and why and that no computer will decide this for them. There is absolutely no merit in this. It is a formula for catastrophic chaos, for it often means -especially when lists are released late evening - that lawyers and parties do not know until the last minute whether their cases are listed at all and, if they are, where they are positioned.

To make matters worse, this is also untied to any established protocol and totally divorced from the classifications in the databases. The result is wholesale unpredictability, inconsistency, a complete lack of uniformity and, inevitably, delays. Nobody wants to tell a judge how a case should be decided. But what to list, in what order and on what day is surely something that lends itself to an easily achievable algorithmic solution. It also takes out that other menace - of manual overrides - sometimes nefariously by a clerk of the court.

The algorithm does need some level of sophistication. This is because of the nature of the system. Some High Courts have what is called ‘original’ jurisdiction; they are courts of first instance for certain case classes. This means that such courts have ‘intra-court appeals’. High Courts that do not have original jurisdiction do not have intra-court appeals. Consequently, there is no one-size-fits-all algorithm that will work across the country, but an intelligently constructed program can certainly eliminate the daily vagaries of the listing system. Overall, some ground rules must be established to do away with ad hoc practices and these can be established country-wide.

High Courts have their own ‘rules’. Sometimes, these conflict with the centrally enacted procedural statutes and settled law is that where there is a conflict, the rules must prevail. Some attempts have been made in the past to harmonise the rules and the codes. There was at least one attempt about 20 years ago to evolve a common code, but it came to nothing. This is one of the most urgent reforms needed.

A few examples tell us why. The Civil Procedure Code, for instance, tells us when an ex parte decree (a decree in the absence of the defendant) may be passed. In the Bombay High Court, the court rule is totally different on the original side. It says that the plaintiff must file a motion (application) for a judgment for ‘want of written statement’. And that motion must be served, then numbered, then listed, then taken up. Result: monumental delay. Consequence of not following the rule? The decree is liable to be set aside.

There is a deliciously ironic story to illustrate this. An extremely eminent judge of the Bombay High Court once passed an ex parte decree. The appeal was admitted. The judge in question progressed to the Supreme Court. There, he delivered a judgment to the effect that the court rules would prevail over the code. The direct result was that his own ex parte decree was set aside: it had not followed the rules.

Or consider this: in one High Court, a party may file a single interim application for several different interim reliefs. In another High Court, the identical case requires the same party to file multiple applications, one per relief sought. What is the rationale?

In some High Courts, an opponent need not file a substantive application to urge that an ad interim order - especially one with a defined life-span - should be vacated or cancelled. For the very same case, just filed in another court, the opponent must file a substantive application for recall, vacating or cancellation. And that will receive orders for filing multiple affidavits. Again, no rationale.

Some laws, especially our criminal laws, allow for local state-level amendments. Yet, there is enormous inconsistency even here. A state-level amendment that permits compounding of an offence in one state is not accepted in the immediately neighbouring state. Why? What is the rationale? What is the benefit?

And then there is that ultimate judicial discretionary power: adjournments. Before we start pillorying courts and lawyers and judges for ‘tareek pe tareek’ (and limited hours and so-called ‘vacations’), there are a few things about the system that must be accepted. First, that some adjournments are actually necessary. Not all cases are ‘ripe’ for disposal on the first day. Most have to be seen through a gestational period before they can be decided. Second, there are days when an adjournment is inevitable: there is just not enough time in the day to attend to everything listed. Neither of these is an insoluble problem. The difficulty is in the random applications for and grant of adjournments for no stateable reason. This urgently needs reform, and wider discussion. Should judicial discretion be eliminated completely? What guardrails need to be provided? Should not our procedural laws prescribe, as the Commercial Courts Act does, strict timelines for all types of cases and simultaneously provide for equally strict, non-discretionary consequences of default?

If it is argued that our Code of Civil Procedure already has timelines and provisions for default, one needs only to look at the empirical data available on the National Judicial Data Grid to understand that the timelines are breached without consequence: written statements are routinely allowed to be filed years after their due date. How many ‘free’ or ‘without cause’ adjournment applications should each party have? What happens if there are multiple defendants or respondents? What if the judge believes that an adjournment is indeed necessary for a further filing?

The point is not that there are ready-made solutions; there are not. Rather, the point is that the problem needs to be tackled head-on and systematised, for transparency, predictability and certainty.

Justice Gautam Patel is a former judge of the Bombay High Court.

Harish Narasappa is a Senior Advocate practicing at the Karnataka High Court and co-founder of DAKSH.

BS Surya Prakash is Fellow and Programme Director at DAKSH.

Leah Verghese is Research Manager at DAKSH.