The collapse of high-profile criminal prosecutions in India frequently follows a pattern: witnesses who gave statements to police turn hostile in court, deny their earlier statements, or become unavailable. The standard explanation is that witnesses have been intimidated or induced. The standard response is prosecutorial frustration and judicial resignation. What is rarely examined is why, after decades of acknowledging the problem, India's legal system still has no effective mechanism to protect the people on whom every criminal prosecution ultimately depends.
The Witness Protection Scheme, 2018 was approved by the Supreme Court in Mahender Chawla v. Union of India and directed to be implemented by all states and Union territories. The Scheme provides for three categories of threat (Category A, B, and C, in descending order of severity) and specifies corresponding protective measures, identity concealment, accommodation changes, relocation, and close protection. On paper, it is a serious framework. In practice, it is institutionally fragile: underfunded, understaffed, and almost entirely unused in the states where witness intimidation is most prevalent.
The structural problem begins with the Scheme's legal status. It was approved by the Supreme Court as a direction under Article 142, not enacted by Parliament as a statute. This means it has no independent enforcement mechanism, no dedicated funding stream, and no penalty for non-compliance by state governments. States that choose not to implement it face no consequence beyond a potential contempt application, which is a slow and uncertain remedy when a witness faces an immediate threat. A statutory framework, with mandatory state obligations, ring-fenced funding, and dedicated witness protection cells, would be categorically different. It does not exist.
The consequences fall disproportionately on witnesses from marginalised communities. A witness from a dominant caste or with significant economic resources can, to some extent, protect herself through private means, relocation, legal representation, social networks. A witness who is a daily-wage worker, a woman in a rural area, a person from a Scheduled Caste community challenging a dominant-caste accused, these witnesses have no such recourse. They are fully dependent on a state protection system that is effectively absent. The result is a systematic bias in which perpetrators with social power can reliably neutralise the testimony of witnesses without social power.
The Bharatiya Sakshya Adhiniyam, 2023, the replacement for the Indian Evidence Act, does not materially address witness protection. It retains the existing provisions on hostile witnesses and the procedure for contradicting a witness's prior statements. Section 23 of the BSA, dealing with admissions, and the provisions on examination and cross-examination, preserve the trial architecture that makes witness neutralisation possible. The new statute's failure to address the protection gap is a significant missed opportunity. The Bharatiya Nagarik Suraksha Sanhita similarly makes no substantial changes to the framework within which witnesses operate.
What would an effective witness protection system require? At minimum: a dedicated statutory authority with state-level branches; funding specified in the statute and not subject to annual appropriation discretion; a threat assessment process that can be triggered by the witness, the prosecutor, or the court; Category A protections that include genuine identity changes and relocation support; and criminal penalties for witness tampering that are prosecuted rather than left as theoretical provisions. Several high-income countries have implemented versions of this architecture. The United States' Witness Security Programme, operated by the US Marshals Service, has been running since 1971 with documented success in protecting witnesses in organised crime prosecutions.
The political economy of witness protection reform is unfavourable. The witnesses most in need of protection are often witnesses against powerful accused, politicians, organised crime figures, dominant community leaders. Those accused have incentives and resources to oppose effective protection systems. The witnesses themselves are diffuse, unorganised, and unable to lobby. Prosecutorial agencies have interests aligned with reform, but they are bureaucratic actors dependent on political sanction. Reform requires political will that is difficult to generate precisely because the problem is most acute in cases involving politically powerful defendants.
The Witness Protection Scheme of 2018 was a meaningful first step. Seven years on, it remains a first step. India's criminal justice system cannot function effectively when the testimony of witnesses, the central mechanism of adjudication, is systematically vulnerable to neutralisation by defendants with resources or social power. A statutory witness protection framework, adequately funded and independently administered, is not a peripheral reform. It is a prerequisite for the basic credibility of the criminal justice system.