The Supreme Court's Prajwala judgment treats trafficking as what it is: a business that runs on other people's poverty. It then converts rehabilitation from a favour the State may grant into a right the State must fund. The harder question opens where the judgment closes: who pays, and whether the money will actually move.

Read the early pages of the Prajwala judgment, and you are reading a balance sheet written on a woman's body. The price a trafficker pays to procure her becomes a debt. That debt is not treated as settled until she has earned back somewhere between fifty and a hundred times the principal, and until then, her income belongs to the brothel keeper, the pimp, the local enforcers who live off her. When a rescue raid arrives, she has often been schooled to give a false name, a false age, a false village, partly because she has been told the government home will be worse than the brothel and emptier of money, and partly because she has been taught that the police belong to the same machinery that holds her.

The Court sets this down as a mechanism. It describes how a market clears.

That is the world the bench of Justices J.B. Pardiwala and R. Mahadevan was looking at when it decided Prajwala v. Union of India on 29 May 2026, reported as 2026 INSC 609. The litigation is old. It began with a writ petition in 2004, survived on a miscellaneous application filed in 2022, and accumulated two decades of orders that repeatedly promised a comprehensive plan for victims and never produced one. By the Court's own account, the Union had admitted as far back as 2004 that there were at least thirty lakh victims in the country. Against that admission, the National Crime Records Bureau recorded about 2,250 trafficking cases in 2022, with roughly 1,500 registered under the Immoral Traffic (Prevention) Act, 1956, that year and about 2,200 in 2023. The crime is, statistically, almost invisible to the State that admits its scale.

To a lawyer, Prajwala is a Section 17 case about who gets produced before a magistrate and what the magistrate may do with them. To a development economist, it reads as something else, a judgment about the economics of exploitation and the economics of repair. This essay is about reading it that way and why that reading matters for what happens next.

The argument the Court builds, stripped to its frame, runs in three moves. Trafficking for commercial sexual exploitation is a business, and like every business, it has a supply chain that begins in deprivation. The Constitution's answer to that business goes beyond punishment and rescue to the duty to rebuild the material conditions that deprivation destroyed, which the Court treats as a positive obligation under Articles 21 and 23. And the State, on the Court's findings, has not performed that duty, because rescue is the cheap and visible half of the work, while rehabilitation is the expensive half that nobody funds. The judgment is a good one. It is also, by its own logic, only the first entry in a ledger that the executive now has to settle.

The crime as a market

Siddharth Kara, whose study of sex trafficking the Court cites, made his name by doing something most human-rights writing avoids. He costed the business out. Trafficking persists because it is, for the trafficker, a high-margin and low-risk enterprise: the acquisition cost of a victim is recovered many times over, the marginal cost of each transaction is low, and the probability of punishment is negligible. The conviction rate in Indian trafficking cases was about sixteen per cent in 2021, with several States recording none at all. A rational, amoral operator looking at those numbers would invest. Many do.

The Court understands the debt-bondage device in exactly these terms. The principal sum is a form of financing. It converts a one-time acquisition into a recurring claim on the victim's future earnings, and it does the ideological work of making her feel that she owes her own captivity. This is why rescue without anything behind it changes so little. The asset, in the trafficker's accounting, is replaceable, and the supply of replacements is the part of the system the law has never touched.

The supply has sources, and the judgment names them without flinching: poverty, the absence of schooling, chronic unemployment, gender and caste discrimination, and the dearth of economic opportunity in the regions from which the victims are taken.

Traffickers recruit with the promise of work, of marriage, of rescue from a bad home. They do it at a distance from the victim's origin, the Court notes, precisely to sever her from the people and the institutions that might help her. The recruitment has migrated online, into matrimonial sites, job portals, gaming platforms and social media, which lowers the traffickers' cost further and widens the pool of the exposed. Svati Shah, whose ethnography of Mumbai the Court also draws on, has shown how thin and shifting the line is between distress migration for work and entry into sexual labour, which is the same point seen from the supply side.

Two features of the Indian state structure deepen the problem, and an economist should say so plainly. Policing and public order are State subjects under the Seventh Schedule, which means the front lines of identification, rescue, and prosecution rest with State governments, whose capacities and priorities vary widely. And the recorded data is only a fraction of the real figure, making the crime hard to budget for, hard to monitor, and easy to ignore. A government that cannot count a problem will not fund its solution.

I have argued a version of this before, in the context of the gap between a health-insurance promise and the public system that was supposed to deliver care behind it. The pattern repeats here. The State keeps building the visible front of a service while the load-bearing structure behind it stays empty.

Article 23 and the coercion of poverty

The most interesting doctrinal move in Prajwala is one the headlines missed, and it is the one a law-and-economics reader should sit with. It concerns the meaning of "force".

Article 23 of the Constitution prohibits traffic in human beings and all forms of forced labour. The Court traces its reading of that article back to the Asiad Workers case of the early 1980s, People's Union for Democratic Rights, where an earlier bench refused to read "force" narrowly. Force, on the Court's reading, reaches past physical compulsion and past the legal compulsion of a threatened penalty. It includes what an earlier bench called the compulsion of economic circumstances. A person driven by hunger and destitution to accept work below the statutory minimum wage is not exercising a free choice between alternatives. He is accepting the only option that grinding poverty has left him, and labour extracted under that condition is forced labour within the meaning of Article 23.

This is, in the vocabulary of economics, a statement about constrained choice. Choice is rarely all or nothing. It is the size of the option set a person faces, and severe deprivation shrinks that set until what passes for consent is closer to capitulation. The Constitution, on this reading, does not pretend that people experiencing poverty are unfree. It recognises that their material circumstances condition their freedom, and it accordingly places a duty on the State. B.R. Ambedkar, who chaired the committee that drafted these provisions and whom I have written about as a development economist rather than only a jurist, understood Article 23 as economic law as much as moral law. Prajwala keeps faith with that understanding.

The companion authority is Bandhua Mukti Morcha, the bonded-labour case, and the Court quotes its hardest sentence: a released bonded labourer who is not rehabilitated will, the earlier bench wrote, prefer slavery to hunger. Freedom without the means to live it is a door opened onto the same trap. The released labourer slides back into bondage. No one forces him a second time; nothing has changed in the conditions that delivered him there in the first place. The Court applies that logic directly to the trafficked woman. Rescue her, return her, and if the poverty and the absence of a livelihood are exactly as they were, she is likely to be trafficked again. Re-trafficking is then the predicted equilibrium of a system that intervenes once and withdraws. The system reproduces the outcome it was meant to prevent.

Rehabilitation as a positive right

Having established that deprivation is both the cause of trafficking and a constitutional harm in itself, the Court takes the step that gives the judgment its teeth. It holds that victims of trafficking have a right to rehabilitation, and it locates that right in Articles 21 and 23 read together.

The reasoning on dignity is careful and worth following because it is doing economic work under a constitutional name. Drawing on Francis Coralie Mullin and the dignity-rights scholarship of writers like Erin Daly, the Court separates dignity into three components. There is inherent dignity, the worth a person holds simply by being human, which the State cannot grant or remove and which trafficking violates by treating a person as an object to be priced. There is dignity as material well-being, the proposition that a life of dignity requires the means to live it, food, shelter, health, and the capacity to earn. And there is dignity as recognition, freedom from the humiliation and social exclusion that mark the trafficked and the stigmatised.

The second of those components is, in substance, Amartya Sen's capability argument arriving in a Supreme Court judgment. Dignity is not an abstraction that floats above a person's circumstances. It depends on what she is actually able to do and be, and that, in turn, depends on resources, opportunities, and the removal of the constraints imposed by deprivation. The right to rehabilitation, on this account, is the right to have the material basis of one's agency restored.

The obligation that follows is positive. Here, the Court borrows, without naming it as such, the machinery of socio-economic rights review that runs through the International Covenant on Economic, Social and Cultural Rights and the South African jurisprudence that built on it. It declines to specify the exact contents of the right, accepting that resource allocation is for the legislature and the executive. Instead, it sets itself a narrower test: whether the State is taking reasonable measures toward the progressive realisation of the minimum core of the right, and whether those measures are real on the ground or only on paper. That formula, reasonableness plus a minimum core plus progressive realisation, is the standard apparatus for adjudicating a positive economic right without the Court becoming the budget-maker. It is the right tool, and the Court deliberately reaches for it.

The consequence is that rehabilitation ceases to be discretionary welfare and becomes a justiciable claim. The State does not get to decide whether to rehabilitate. It gets to decide, within limits, how. And a court can now ask whether what it has decided is reasonable, and whether it has been delivered.

The reasonable-measures test and the empty homes

On the reasonableness test, the Court's verdict is unsparing, and the evidence it relies on is administrative rather than rhetorical.

There is no comprehensive Victim Protection Plan, the absence that brought the petition in 2004 and that the Union once accepted needed remedying before it changed position and argued the existing framework was adequate. The Court finds that the framework has barely moved since 2015. The older Ujjawala and Swadhar homes were folded into the Shakti Sadan scheme under Mission Shakti, and little else changed. When the bench examined what those homes provide, the picture was of provision on paper and emptiness in fact. Several States reported no mental-health care, no de-addiction support, and no education or vocational training for residents. Some States reported either no Shakti Sadan homes at all or only one. No State could report how many halfway homes, the institutions meant to bridge the move from shelter to independent life, it had set up. The implementation, the Court concludes, is poor and non-uniform, and the measures that exist on paper have not been translated into reality.

This is where the development economist parts company with the assumption that money is the binding constraint. The Nirbhaya Fund, created after the 2012 Delhi gang rape to finance exactly this kind of women's-safety and victim-support infrastructure, was built as a non-lapsable corpus, so unspent money does not vanish at the end of the year. By the mid-2020s, the corpus had grown past six thousand crore rupees, and most of it sat unspent. By the government's own accounting, projects worth about Rs 9,549 crore had been appraised, of which roughly Rs 6,213 crore was allocated, and about Rs 4,242 crore was released to the ministries and States meant to spend it. A parliamentary committee found that only about a third of the fund had been spent, and that twenty-two of thirty-six States and Union Territories had drawn down less than two-thirds of their share, several of them almost nothing.

The Accountability Initiative, the public-finance tracking group where Avani Kapur led much of this expenditure work, has shown the pattern in finer grain. The One Stop Centre scheme, one of the Nirbhaya-funded women's safety programmes, saw its allocation climb from about Rs 105 crore to Rs 385 crore over three years, while the share it spent fell. The binding constraint here is administrative, not fiscal: the capacity and the will to turn the envelope into functioning homes, trained counsellors, vocational programmes and reintegration support, in a domain where the front-line responsibility sits with States of uneven administrative strength and where women's safety has rarely been a political priority that survives the news cycle.

I have made the same structural argument about the Union's accounts, where money was parked under opaque heads and votes left unspent allowed a government to claim a commitment it never delivered. The Shakti Sadan homes that exist on a dashboard but offer no de-addiction support are the welfare-state version of that accounting: the promise recorded, the service never rendered. Prajwala matters because it refuses to accept the recorded promise as performance. It treats the empty home as a constitutional failure, which is what it is.

Agency under constraint

The part of the judgment that drew the headlines is the part on consent, and the Court handles it with more honesty than the public debate usually allows.

Indian law on prostitution has long conflated two different things. The Immoral Traffic (Prevention) Act defines prostitution itself as exploitation, and its offence-creating provisions, for the most part, treat the involvement of any third party as criminal regardless of force, coercion or deception. The effect is that everyone swept up in a brothel raid, the trafficked, the once-trafficked who now work voluntarily, and those who entered sex work by their own choice, get processed through the same Section 17 machinery as though they were identical. They are not.

The Court works through the long feminist disagreement that sits behind this, and it does not pretend to resolve it. On one side are the abolitionists, in the tradition of Andrea Dworkin and Dorchen Leidholdt, for whom prostitution is itself a form of violence and consent to it is largely a fiction, so that all those in it are victims to be rescued. On the other hand, there are those who, following Martha Nussbaum's argument that taking money for bodily services is not categorically different from other wage labour, treat sex work as work and insist on distinguishing it from trafficking. Prabha Kotiswaran, whose study of sex work and the law in India the Court relies on heavily, has spent a career showing what gets lost when the law collapses the two, and the harm that anti-trafficking enforcement can do to the very women it claims to save.

The Court's resolution is the one the economics supports. Following Giorgia Serughetti, it holds that vulnerability and agency coexist. Poverty and the absence of alternatives narrow the range of a woman's choices, sometimes drastically, but they do not abolish her capacity to choose. Treat her only as a victim, and you talk over what she wants. See nothing but an agent, and you lose sight of the conditions that constrained her choice, and of the help she may still need. The same insight that drives Article 23's reading of economic coercion drives this: choice is real, and choice is conditioned, both at once, and policy has to hold both.

From this, the Court derives two operative principles—first, a threshold inquiry. Before a person produced before a magistrate is put through the full apparatus of inquiry and possible detention, the magistrate must first ask whether she is a voluntary adult sex worker who does not wish to be there. If she is, the machinery should not run on her at all. This is the principle of non-interference, drawn from the Law Commission's 64th Report and from the Court's own directions in Budhadev Karmaskar in 2022, which had already told the police not to harass adult consensual sex workers during raids. Second, where a person is a trafficking victim, her consent becomes the governing factor in what the magistrate decides about detention and reintegration, displaceable only where her safety is genuinely at risk or where her stated wishes are the product of coercion, threat or tutoring, and then only with written reasons.

For children, the Court draws the line absolutely. Consent is irrelevant. Any child in commercial sexual exploitation is a victim, full stop, and the bench flags as an anomaly that the Bharatiya Nyaya Sanhita still requires proof of the "means" element even for children, a deviation from the Palermo Protocol that it asks Parliament to correct.

There is a careful formulation late in the judgment that deserves to outlast the news cycle: that one can protect the rights of sex workers without thereby creating a right to sex work. It is a lawyerly distinction, and it does real work. It lets the State extend dignity, safety and access to law to adult sex workers, who at present are isolated by stigma and failed by officials who share that stigma, without the State having first to settle the unwinnable question of whether sex work should exist. The rights are owed to the person, not to the trade.

The incentive logic behind the recommendations

Underneath the constitutional language, several of the Court's recommendations are purely incentive-based analyses, and they are the parts most likely to change outcomes if anyone acts on them.

Take the criminalisation of victims. Sections 7, 8 and 20 of the ITPA penalise soliciting and being a prostitute in certain places, and in practice, the police reach for these provisions against the rescued women rather than the harder-to-prove provisions against brothel keepers, financiers and recruiters. The Court records that this has been known and lamented in official advisories since at least 2009, and that a 2006 Bill to delete two of these provisions lapsed. The advisories failed for a reason an economist would predict: they were suggestions, not mandates, so the discretion that lets an officer book the victim instead of the trafficker remained intact, and the path of least resistance won. The Court's fix is to make the protection a matter of law rather than advice, by writing into the statute that these provisions do not apply to someone who is or may be a trafficking victim. Change the rule, not the exhortation.

There is a sharper observation buried in the section on shelters. The ITPA's model is institutional and time-bound: a magistrate can order a victim detained in a protective home for one to three years, which the Court says begins to resemble a carceral stay. The bench points out the perverse incentive this creates. If the only rehabilitation on offer is something that feels like custody, fewer women will come forward to identify as victims and accept it. Improve the quality and range of rehabilitation, including alternatives to compulsory detention, so that more victims choose to be rehabilitated and more succeed in doing so. The take-up of a public service depends on whether people want it. The Court is reasoning about participation constraints, which is the right way to approach the issue.

And then there is the money trail. Among the closing recommendations is the point that trafficking persists because it is profitable. Hence, an effective response has to attack the profit through asset attachment and prosecution under the Prevention of Money Laundering Act, and not stop at the criminal charge for the act itself. This is the oldest result in the economics of crime, the one Gary Becker formalised in 1968: deterrence is a function of the expected cost of offending, and if you cannot raise the probability of conviction, you can still raise the cost by stripping the gains. A raid that frees a few women while leaving the financiers and their balance sheets untouched does not alter the enterprise's economics. Following the money does.

On one structural demand the petitioners made, a new Organised Crime Investigation Agency, the Court declined to issue a direction, holding that the functions already exist across the National Investigation Agency, the CBI's anti-trafficking unit and the State bureaus, and that creating a wholly new body would stretch its powers under Articles 32 and 142 too far. It is a defensible restraint and consistent with a court that elsewhere insists allocation is for the executive. It also leaves the coordination problem the petitioners identified, agencies working in silos, formally unsolved and handed back to the government's discretion.

What a development economist would still ask

The judgment is strong, and the praise it has drawn is earned. It names the cause of trafficking correctly, it converts rehabilitation from charity into a right with a reasonableness standard attached, it refuses the lazy binary of victim and agent, and it gets the incentive structure right on criminalisation, take-up and the money trail. A bench could have done far less and still produced a quotable judgment. This one did the work.

The questions that remain are the questions the Court was careful to leave to the executive, and they are fiscal-federal questions before they are anything else. A right to rehabilitation with a minimum core is a spending commitment, and the spending falls in a domain where the Union holds the corpus and the schemes, while the States hold the police, the homes and the front line. The history of the Nirbhaya Fund tells us that money in a non-lapsable corpus does not rehabilitate anyone on its own. It has to be allocated to schemes, released to States, converted into staffed homes, and funded vocational programmes, and then monitored for outcomes rather than the count of buildings notified. None of that is automatic, and the record so far is of allocation rising while utilisation falls. The minimum core the Court has now made justiciable will be realised, if it is realised, through the unglamorous business of expenditure that moves and for which someone is held accountable.

That points to where the real test lies. The Court has asked for compliance with its additional directions within three months and has listed the matter to return in September 2026. The honest measure of this judgment will not be the elegance of the Victim Protection Plan that has been written. It will be whether, by the time the matter comes back, the Shakti Sadan homes that reported no counselling have counsellors, whether the States that reported no halfway homes have built any, and whether the Nirbhaya corpus has been drawn down for the purpose for which it was created rather than parked. A plan that the executive can record as compliance while changing nothing on the ground would be the welfare equivalent of the fiscal theatre I have written about elsewhere. This dashboard reports a service the citizen never receives.

There is a second, harder gap the judgment cannot close on its own, and it sits on the supply side. The Court is right that material deprivation is what makes a person trafficable, and right that rehabilitation must rebuild a livelihood if it is to mean anything. But rebuilding a livelihood for the rescued is the same problem, at the level of the individual, as the absence of decent work that the Indian economy has been generating in aggregate, the drift of workers back into low-productivity, low-security activity that should worry anyone watching the labour data. A rehabilitation that trains a woman for work that does not exist returns her to the same constrained choice that delivered her to the trafficker. The right to rehabilitation, taken seriously, eventually runs into the question of whether the economy offers people with low incomes any alternative worth choosing. That is beyond what a Section 17 case can decide. It is not beyond what the judgment's own logic implies.

For now, the achievement should be stated without hedging. Prajwala has told the State that it may not treat the rescued as objects to be processed, that it must fund their repair as a matter of right, and that the women who choose sex work are owed dignity and law rather than raids. It has put the bill on the table. The paying is the part the Indian State has always found hardest, and a judgment, however good, cannot spend the money for it.

Varna Sri Raman is an independent development economist.

Further reading

The judgment and its reporting: The full text of Prajwala v. Union of India, 2026 INSC 609; LiveLaw's reports on the primacy of the victim's consent and on the Victim Protection Plan and the call for legislative reform.

The economics of trafficking and sex work Siddharth Kara, Sex Trafficking: Inside the Business of Modern Slavery (Columbia University Press, 2008); Prabha Kotiswaran, Dangerous Sex, Invisible Labor: Sex Work and the Law in India (Princeton University Press, 2011); Svati P. Shah, Street Corner Secrets: Sex, Work, and Migration in the City of Mumbai (Duke University Press, 2014); Martha C. Nussbaum, "Whether from Reason or Prejudice: Taking Money for Bodily Services", 27 Journal of Legal Studies 693 (1998); Giorgia Serughetti, "Rethinking Force and Consent, Victimisation and Agency", 3 Femeris 79 (2018).

Dignity, capability and the constitutional economics of exploitation Amartya Sen, Development as Freedom (Oxford University Press, 1999); Erin Daly, Dignity Rights: Courts, Constitutions, and the Worth of the Human Person (University of Pennsylvania Press, 2012); Gary Becker, "Crime and Punishment: An Economic Approach", 76 Journal of Political Economy 169 (1968); the bonded-labour and forced-labour line of cases, People's Union for Democratic Rights v. Union of India and Bandhua Mukti Morcha v. Union of India.

The fiscal record; The Accountability Initiative's analysis of the unspent Nirbhaya Fund; Business Standard on nearly half the fund left unspent; the Mission Shakti scheme guidelines; the United States State Department's 2022 Trafficking in Persons report on India, on shelter conditions and return to families.