Plea Bargaining Doesn't Belong in Indian Criminal Procedure

Plea bargaining has been on the statute book in India for twenty years. It does not work. It is barely used. It produces neither the disposal volume the Law Commission's 154th Report promised in 1996 nor the relief from undertrial pendency the Malimath Committee promised in 2003. It produces, year on year, a footnote in the criminal-procedure curriculum and a vanishingly small number of dispositions in the courts.⁸ The standard explanation is that the reform is a good idea badly implemented — that prosecutors are reluctant, that the defence bar is uninformed, that the judiciary is suspicious, that the procedure under Chapter XXI-A of the CrPC is too cumbersome. The standard explanation is wrong. The reform did not fail in implementation. It was never going to do what its drafters said it would, because the procedure that was lifted from the American criminal-justice system cannot do that work in the Indian one, and Chapter XXI-A was drafted by people who either did not understand this or did and went ahead anyway.

The correct verdict on Indian plea bargaining is not that it needs reform. It is that it does not belong here. The 2005 amendment retained in BNSS Chapter 21 was a transplant from one system to another, conducted without the surrounding institutional architecture that makes the original work, and produced — predictably — a vestigial procedure that runs on a track of its own and intersects the actual criminal-justice system almost nowhere.¹ Twenty years is long enough to call this what it is.

The settled view, and why it is wrong

The settled view, in the law schools and in the reform reports, is that plea bargaining is a sensible procedural innovation that India has not yet learned to use. It is sometimes credited with potential to reduce undertrial pendency, sometimes with potential to fast-track minor offences, sometimes with potential to give victims agency they do not have in the ordinary trial. All of these claims rest on the same underlying picture: that plea bargaining is a tool, the tool exists, and the question is why we are not using it.

This picture has its strongest version, and I should state it before I disagree with it. The strongest case for plea bargaining in India runs as follows. Indian criminal trials take five to ten years. Undertrials languish in jail for periods that exceed the maximum sentence for the offence they are charged with. The system is structurally incapable of giving every accused a full trial within constitutional time. Some mechanism for negotiated disposition is necessary. The American system, for all its faults, processes ninety-five per cent of its criminal cases through plea bargaining and lives with the resulting compromises because the alternative — full trial for every accused — is unaffordable. India faces the same affordability constraint with an even tighter resource base. The 154th Report and the Malimath Committee saw the problem honestly and recommended a tool that other jurisdictions had used to address it.¹¹ The drafting in Chapter XXI-A is conservative — applicable only to offences punishable with less than seven years, excluding socio-economic offences, excluding offences against women and children below fourteen, excluding juveniles — precisely because the drafters were sensitive to the dangers of the American model.² The Chapter is a careful, calibrated introduction of a tool that the system needs. The right response is to use it more, train prosecutors better, and let the volumes grow.

This is the strongest version. It is also wrong. The Chapter XXI-A drafters did not understand what plea bargaining does in the American system, and the careful calibration of the Indian provision has produced not a conservative tool but a useless one. The defenders of the transplant point to the small volume of plea-bargained dispositions in India as a sign of teething trouble. Twenty years of teething trouble is not teething trouble. It is the design.

What plea bargaining actually is in the American system

To see why the Indian transplant cannot do what its drafters said it would, it helps to be clear about what plea bargaining is in the system it was borrowed from. Plea bargaining in the United States is not a procedural shortcut. It is the constitutional default for criminal adjudication. Brady v United States (1970) made guilty-plea bargains constitutionally permissible.⁴ Santobello v New York (1971) made them judicially enforceable. Bordenkircher v Hayes (1978) authorised prosecutors to threaten more serious charges to induce a plea. These cases collectively produced a system in which the trial is the exception and the plea is the rule. Ninety-five per cent of federal convictions and a comparable fraction of state convictions are obtained by plea, not by trial.

This works — to the extent it works — because it sits inside a particular institutional architecture. American prosecutors have charging discretion. They decide what offence to charge, whether to charge enhancements, whether to seek the death penalty in a homicide, whether to invoke the federal three-strikes rule. They have sentencing-recommendation power that the trial judge typically follows. They have the threat of a jury, which means the marginal cost of a trial to the defendant is the variance of jury outcomes, which is high. They operate under Brady v Maryland disclosure rules that give the defence a partial preview of the prosecution's case before the plea. They operate inside determinate-sentencing or guideline-sentencing regimes that let prosecutor and defence price the trade-off in months and years that both can compute.

Most importantly, on the defence side, the system is staffed by public defenders carrying caseloads of three hundred to a thousand felony cases per attorney per year. The defender does not have the time to take every case to trial. The defender's job, in the typical case, is to extract the best plea the prosecutor will offer, to advise the client whether to take it, and to move on. Plea bargaining in America is not a sometimes-used alternative to trial. It is the criminal-justice system. The trial is the residual category.

Each of these institutional features matters for whether plea bargaining produces something the system can call justice. None of them exists in India. Indian prosecutors do not have charging discretion in the American sense — the charge is framed by the Magistrate or the Sessions Judge after considering the police report, and the prosecutor opens the case on the charge so framed.⁶ Indian prosecutors do not have sentencing-recommendation power that the trial judge follows; the trial judge sentences after evidence is taken and after hearing the accused under section 235 of the CrPC. There is no jury threat to bargain against. There is no Brady-style pre-plea disclosure regime. There is no determinate-sentencing or guideline-sentencing framework — the Indian Penal Code (now BNS) provides a maximum and sometimes a minimum, and the trial judge picks a point inside that range using considerations that are not standardised across courts.

And there is, critically, no public-defender system to carry the criminal docket. Legal-aid representation under Article 39A is patchy, undercapacity, and confined to a small fraction of cases. Most Indian criminal defendants are represented either by private counsel who have an interest in stretching the matter or by counsel they cannot afford. The structural pressure that drives ninety-five per cent of American cases into plea bargains — defenders without time and prosecutors with charging discretion — is simply absent.

What Chapter XXI-A does instead

Chapter XXI-A, accordingly, does not import plea bargaining. It imports a procedure that bears the name. The procedure works as follows. The accused files an application before the trial court under section 265-B, accompanied by an affidavit swearing that the application is voluntary and that he has not previously been convicted of the same offence. The court issues notice to the prosecutor, the complainant if any, the victim and the police investigator. The court examines the accused in camera to confirm voluntariness. If satisfied, it invites the prosecutor, the complainant, the victim and the accused to a meeting to work out a 'mutually satisfactory disposition' under section 265-C, with the court superintending throughout to ensure that the process is voluntary.⁹ The disposition is reduced to a report under section 265-D. The court then sentences under section 265-E, applying the statutory discount: one-half of the minimum punishment where a minimum is prescribed, otherwise one-fourth of the maximum.⁵ Judgment is delivered in open court under section 265-F and is non-appealable under section 265-G except by way of Article 136 or Articles 226 and 227.¹⁰

Read against the American architecture, every step in this procedure is doing the opposite of what plea bargaining does in the United States. The voluntariness inquiry is judicially supervised, not adversarial. The prosecutor at the bargaining table is not a party negotiator but, under the settled law of Habeeb v State of Hyderabad and State of Bihar v Ram Naresh, an officer of the court whose duty is to represent the administration of justice — not to extract concessions.⁶ The victim and the police investigator must be present, which is the right policy choice for a system that takes victims seriously but is doctrinally incompatible with an adversarial bargain. The sentencing discount is statutorily fixed, which means there is nothing for the prosecutor to offer; he cannot tell the accused 'plead and I will recommend three years instead of five' because the recommendation is irrelevant — the judge applies the statutory fraction.

What this leaves is a procedure that looks like a bargain on paper and is, in operation, a judicially supervised guilty-plea hearing dressed up with extra notice requirements and a fixed-rate sentence reduction. It is more cumbersome than the existing guilty-plea route under section 229 of the CrPC for sessions trials.⁷ It offers the accused a sentence discount that he could often have obtained anyway by pleading at the framing of the charge. And it costs the prosecutor and the victim a meeting that the rest of their docket does not give them time for.

This is why the Chapter is under-used. Not because prosecutors are uninformed. Because the procedure was drafted on top of an inquisitorial trial system and gives nobody a serious reason to use it. The accused has a parallel route to a plea-based conviction under section 229 that does not require an application, an affidavit, an in-camera examination and a meeting. The prosecutor has no incentive to negotiate, because there is nothing to negotiate — the discount is statutory and the charge cannot be revised downward. The victim has no incentive to participate, because the compensation under section 265-E(a) is rarely larger than what a civil action would yield and the criminal stigma against the accused is in any case attenuated by the discount.

What the offence-list reveals

The exclusion list in section 265-A is where the design pretence collapses most clearly. Chapter XXI-A does not apply to offences punishable with death, imprisonment for life or imprisonment exceeding seven years. It does not apply to offences notified as affecting the socio-economic condition of the country. It does not apply to offences against women. It does not apply to offences against children below fourteen.² It does not apply to juveniles, which section 265-L makes explicit.

What is left? Strip out homicide. Strip out rape. Strip out POCSO offences. Strip out section 498-A and dowry death. Strip out economic offences notified under the socio-economic-condition clause. Strip out narcotics offences above the small-quantity threshold. Strip out juveniles. What remains is a residue of mid-tier offences — simple hurt under section 319 IPC (now section 115 BNS), criminal breach of trust below a notified threshold, certain motor-vehicle offences, certain offences under the Indian Penal Code that have neither a death nor a life imprisonment ceiling nor a socio-economic-condition notification. The residue does not include the offences that fill Indian jails. It does not include the offences responsible for the undertrial pendency that the Malimath Committee identified.

So the Chapter was sold as a solution to undertrial pendency and was drafted to exclude all the cases responsible for the pendency. Either the drafters did not notice the contradiction or they did and went ahead. Either reading is bad. The first means the reform was incompetent. The second means it was performative — a reform passed to be seen to have passed, with the substantive scope of application kept tight enough that the procedure could not change anything.

The pre-Chapter XXI-A jurisprudence is also worth recalling here, because it casts the 2005 enactment in a particular light. The Gujarat High Court in State v Lakshaman (1998) had set aside a sentence on the ground that the Magistrate's caution to the accused before plea was 'illicit plea bargaining'.³ The Supreme Court in Harbhajan Singh v State of UP (2002) had 'deprecated plea bargaining in cases involving serious offences'. Balram Kumawat v Union of India (2003) had held that 'plea bargaining shall not be supported for economic crime'. The courts, in other words, had a settled position by 2005 that plea bargaining was constitutionally suspect. The 2005 Amendment did not address that suspicion. It legislated around it by excluding the categories where the Court had said plea bargaining was unacceptable. The Chapter is what is left after you remove the cases that would actually have raised the question.

The defence that the procedure protects the vulnerable

The strongest defence of the offence-list — and of the careful, supervised character of the bargaining process — is that the design protects vulnerable accused and vulnerable victims from the pathologies of American plea bargaining. American plea bargains coerce innocent defendants, especially poor defendants of colour, into pleading guilty to offences they did not commit, by stacking the threatened charges so high that the rational choice is to take the plea even where the accused has a defence. American plea bargains exclude victims from the bargain, often producing dispositions that the victim experiences as a betrayal. American plea bargains let prosecutors mass-produce convictions at the cost of due process. India, the defence runs, has looked at these pathologies and built a procedure that avoids them — by capping the offences, by requiring judicial supervision, by requiring victim participation, by fixing the discount statutorily so that the bargain cannot be coercive.

The defence is honest and it is partly correct. A non-supervised plea-bargaining regime in India would replicate the worst features of the American system in a setting that has fewer institutional protections than the American one — no constitutional right to counsel comparable to Gideon, no exclusionary rule of comparable bite, no jury to deter prosecutorial overreach, no Brady disclosure regime. If we were going to import plea bargaining at all, the supervised version is plainly better than the unsupervised one.

But the defence concedes the polemic's premise. If the supervised version is so attenuated that it cannot do the work it was advertised to do — if it cannot, on the offence-list, address the undertrial-pendency problem; if it cannot, on the structural-incentive analysis, attract prosecutorial participation; if it cannot, given the statutory discount, give the parties anything to negotiate — then the right conclusion is not that we have a careful supervised plea-bargaining regime. The right conclusion is that we have a procedure that bears the name 'plea bargaining' for political reasons and does not function as plea bargaining for any practical purpose. We have, in other words, what the American legal economist Robert Ellickson would call a Potemkin reform: visible from a distance, hollow on inspection. The careful design did not produce a tool. It produced a piece of statutory furniture that everybody can point to and nobody uses.

What the system actually needs

The pendency problem that the 154th Report and the Malimath Committee identified is real and Chapter XXI-A is not the solution to it. The solution to under-trial pendency lies in three places, and plea bargaining is not one of them.

First, in the offence-grading work that BNS could have done and did not — re-graduating the offence-list so that offences punishable with less than three years are tried summarily by default, with the option to elect a regular trial only on the accused's application. The summary trial is the under-used reform; the summons-case procedure in Chapter XX of the CrPC already gives the structural pathway. Reducing the volume of regular trials by re-classifying the offence-list is a substantive reform that does not require a transplant.

Second, in disposal of long-pending undertrials by application of section 436-A of the CrPC (now section 479 BNSS) — release of undertrials who have served half the maximum sentence — and by genuine application of the constitutional time-bar that the Supreme Court has repeatedly articulated since Hussainara Khatoon. The vehicle for clearing the pendency exists. Its under-use is the problem. Adding a parallel plea-bargaining route does not address the under-use of the existing route.

Third, in resourcing legal aid. The structural feature of American plea bargaining that does the actual work — the public-defender system — is the one feature India would have to build first to make any plea-bargaining regime functional, and is also the one feature India has not built. Legal-aid representation under Article 39A and the Legal Services Authorities Act has been chronically under-funded. Without it, no negotiated-disposition regime can produce just outcomes, because the accused will not be advised on what to accept and what to refuse. Without it, plea bargaining will either be unused (the current outcome) or used coercively (the American outcome). There is no good third option.

The verdict

Chapter XXI-A of the CrPC, retained as Chapter 21 of the BNSS, is a procedural reform that imported the name of a tool without the institutional architecture that gives the tool its function. It cannot reach the offences responsible for the criminal-justice system's pendency. It cannot offer the parties anything to negotiate. It cannot, in the absence of a public-defender system, be used safely at scale. It has not been used at scale, twenty years on, because the design did not permit it to be — and because, in the inquisitorial system India runs, the design could not have permitted it. The persistence of the Chapter in the BNSS in 2023 is not retention of a working reform; it is retention of a piece of statutory drapery that the legislature has not yet been embarrassed enough to remove.

If we want a negotiated-disposition route in Indian criminal procedure, we should design one for an inquisitorial system: a Magistrate-led process at the framing stage, with statutory sentence discounts that vary by stage of admission (highest discount for the earliest plea, lowest for a plea at sentencing), without the parallel-track structure that Chapter XXI-A imposes. That reform is a real reform. The 2005 Chapter is not. It is what India did instead of reform.

Get the week's biggest decisions, distilled.

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