Five Crore Cases and the Arithmetic of Delay

Pritiraj BrahmaFounding Partner, Pritiraj & Partners · 19 March 2026

India’s courts carry 5.05 crore pending cases across 20,580 judges. But the number that identifies the remedy is different: of 4.4 crore cases in the district courts, 1.8 crore are recorded at appearance or service. Two in five have not passed the first step — and the first step is not the court’s work.

ProvenancePublished 19 March 2026; revised 28 August 2026 to rest the argument on the stage distribution and appointment arithmetic rather than on the pendency stock, and to add the adjournment objection. Pendency, judge strength and stage figures are from the Supreme Court’s Centre for Research and Planning report on the state of the judiciary. High Court vacancy and proposal figures are as placed before Parliament for 6 March 2026; appointment counts are for calendar 2025. The Law Commission’s 50-judges-per-million recommendation is several decades old and is cited as a historical benchmark, not a derived current requirement. Pendency figures move daily; the National Judicial Data Grid is a live series.

A backlog is not a scandal. It is arithmetic: cases arriving faster than judges can dispose of them, sustained for decades. India carries 5.05 crore pending cases across 20,580 judges — but the number that identifies the remedy is different. Of 4.4 crore cases in the district courts, 1.8 crore are recorded at appearance or service. Two in five have not passed the first step, and the first step is not the court’s work.

India’s courts are carrying more than five crore pending cases across all tiers. The Supreme Court’s own Centre for Research and Planning, in its report on the state of the judiciary, put the position plainly: a pendency of 5.05 crore cases handled by 20,580 judges across the Supreme Court, the High Courts and the district judiciary.

The temptation is to read this as a story about lazy courts or endless adjournments. The evidence points somewhere less satisfying and more tractable — and it points there through the stage data rather than the headline total, because a stock of pending cases is a poor verdict on performance and this essay does not rest its case on one.

Six findings anchor this analysis:

1. Two of every five district-court cases have not passed the first step. Of the 4.4 crore pending in the district and taluka courts, the stage recorded for 1.8 crore is appearance or service — the defendant has not yet been brought before the court. These are not complex matters awaiting a difficult judgment. A judiciary can only dispose of what is properly before it.
2. Sanctioned strength is not filled, and the gap is structural rather than political. The government told Parliament that 312 High Court judge posts were vacant as of 6 March 2026, with 132 appointment proposals at various stages between the government and the Collegium. This is not a hiring freeze; it is a process that does not clear at the rate posts fall vacant.
3. Appointments are being made at scale, and vacancies persist anyway. 157 judges were appointed to the High Courts during 2025 — a substantial year — and the vacancy figure above is what followed it. That is the signature of a system where inflow and attrition together match the appointment rate, leaving the stock untouched.
4. The judge-population comparison is real but weaker than usually stated. India has around 14.2 judges per million people against a Law Commission recommendation of 50 per million. India plainly has too few judges. But that recommendation is several decades old and was derived from assumptions about caseload and court structure that no longer obtain; required judicial capacity is a function of disputes brought for adjudication, not of population, and comparable outcomes are achieved at very different ratios across jurisdictions. The shortfall is a finding; the specific multiple of four is a number that should be re-derived rather than cited.
5. The state is the largest litigant, and much of its litigation is unchosen. Government departments and public undertakings account for a substantial share of cases and appeals, many pursued not because they are winnable but because no official is authorised to decide not to appeal. This is the one category of Indian litigation that could be reduced by an administrative order rather than a procedural amendment.
6. Digitisation has improved visibility, not throughput. The National Judicial Data Grid publishes pendency and disposal at national, state, district and individual court level, updated daily. India can see its backlog in real time and has not thereby reduced it. Transparency is a precondition for reform, not a substitute for it — and the Grid holds the clearance-rate data that would settle whether the system is gaining or losing ground, which is not what gets quoted.

The first-step problem

That reframes the problem usefully. A large share of Indian judicial delay is not judicial at all — it is administrative and executive: summons that are not served, process servers who are unavailable, addresses that are stale, and police bandwidth diverted elsewhere. Investment in service of process is unglamorous, cheap relative to court construction, and sits largely with state governments rather than with the courts.

Why the backlog compounds

Three mechanisms make delay self-reinforcing, and each deserves naming because each has a different remedy.

Delay is an asset to one party. In a very large share of civil litigation, one side benefits from the passage of time — possession retained, payment deferred, an order stayed. Where delay is valuable and cheap, it will be purchased. Adjournment discipline is not procedural fussiness; it is the removal of a subsidy.

The state litigates because nobody is allowed to stop. A settlement policy that protects officers who withdraw weak cases would remove more litigation than any procedural amendment, because the present incentive is entirely one-directional: an officer who appeals and loses has followed procedure, while an officer who declines to appeal has made a decision that can be questioned for the rest of a career.

Old cases lose their advocates. Cases pending for decades — and India has tens of thousands in the High Courts alone that are more than thirty years old — outlive the interest of the parties, sometimes the parties themselves. They cannot be closed without an application, and no one applies. Legacy stock therefore requires a deliberate purge mechanism rather than ordinary disposal.

The capacity equation, and why appointments alone do not solve it

Retirements are known years in advance. A process that begins when a post falls vacant will, by construction, run permanently behind — and the 312 open posts are the accumulated arithmetic of that lag rather than evidence of anybody’s bad faith. This is the cheapest fix available in the whole essay: appointing against a rolling forecast rather than against vacancies requires no change to the Collegium system, no legislation, and no additional posts.

Why this is an economic question

It is easy to treat judicial delay as a matter of individual injustice, which it certainly is. It is also a first-order economic constraint, and this is the argument that tends to move governments.

Contract enforcement is the substrate of investment. A lender who cannot foreclose within a predictable period prices that risk into every loan. A firm that cannot enforce a supply contract vertically integrates instead of contracting, which is precisely the inefficiency that keeps Indian firms small. A land title that takes fifteen years to establish is a land title that cannot be collateralised. Every rupee of the Union’s capital programme runs through contracts that ultimately rest on the expectation of enforcement — and that expectation, in India, is weaker than the balance sheet assumes.

The counter-case, honestly stated

Three objections deserve a hearing.

First, pendency is a stock, and a stock is a poor verdict on performance. A court disposing of more cases than it did last year will still report rising pendency if filings rise faster — and filings rise with population, with commercial activity, and with improved access to justice. On one reading, a growing docket is what a legal system looks like when more citizens can actually reach it. This essay accepts the point, which is why its findings rest on the stage distribution and the appointment arithmetic rather than on the 5.05 crore, and why publishing clearance rates appears among the recommendations. It is entirely possible that the Grid’s clearance data shows a system roughly holding its position.

Second, the first-step diagnosis assigns one cause to a category that aggregates several. Cases recorded at appearance or service include matters newly filed and awaiting first listing, matters where a defendant is evading service deliberately, matters stayed by a superior court, and matters the plaintiff has stopped pursuing. Some substantial share is certainly failed process-serving, and reading the entire 1.8 crore as an executive failure overstates what a stage classification can tell you. Professionalising service remains the highest-return intervention available; its yield is less certain than the number implies.

Third, an adjournment is often the correct order. Counsel falls ill, a witness cannot travel, a document is genuinely awaited — and a court that refuses adjournments mechanically produces bad decisions faster. Adjournment discipline has to mean recording reasons and publishing the pattern, not imposing a quota, because the version of this reform that a target-driven system would actually implement is the harmful one.

What we would do

Five interventions, in order of cost-effectiveness.

1. Fix service of process before anything else. If 1.8 crore cases are recorded as stalled at appearance, a professionalised, adequately staffed process-serving system — with electronic service where addresses permit and verified registries where they do not — would move more cases than any number of new benches. It is a state government function, and it is not being done.
2. Appoint against a rolling forecast, not against vacancies. Beginning the appointment process twelve months before a post falls vacant would eliminate a large share of the 312 currently open, without changing the Collegium system at all. It is the single cheapest structural fix in Indian judicial administration.
3. Create an authorised route for the state to withdraw weak cases. A standing committee empowered to close litigation, with statutory protection for the officers who recommend closure, addresses the largest single litigant in the country and requires no court’s consent.
4. Run a legacy purge with a statutory basis. Cases beyond a defined age, where no party has appeared for a defined period, should be closable on notice. The stock will not clear itself, and ordinary disposal was never designed to reach it.
5. Publish court-level clearance rates, not just pendency. The Grid already holds the data. Clearance rate — disposals against institutions — is the number that tells you whether a court is gaining or losing ground, and until it is the headline figure the national conversation will keep being conducted in a measure that cannot answer the question.

India has spent two decades making its judicial backlog visible, and the instrumentation is now genuinely excellent. The arithmetic beneath it is unchanged: too few judges appointed too late, too many cases the state was never required to file, and a first step that in two of five cases has not happened. None of that is fixed in a courtroom.