By Sunil Garodia
First publised on 2026-02-28 10:29:13
On February 27, 2026, a special court at Rouse Avenue discharged all 23 accused in the CBI's Delhi Excise Policy case, including former Chief Minister Arvind Kejriwal, former Deputy Chief Minister Manish Sisodia, and BRS leader K. Kavitha. The immediate political reactions were predictable. The deeper legal significance is more serious.
This was not an acquittal after trial. It was a discharge at the stage of charge framing. That distinction matters. At this stage, the court does not ask whether the accused are guilty. It asks a narrower question: is there enough material to justify putting them on trial at all?
The court's answer was no.
To understand how consequential that is, one must understand how low the threshold is at the stage of framing charges. The prosecution does not have to prove guilt beyond reasonable doubt. It does not even have to prove that conviction is likely. It only has to show that the material on record, if taken at face value, discloses a prima facie case. Courts routinely allow cases to proceed even on relatively thin evidence, reserving deeper scrutiny for trial.
When a court refuses to frame charges, it is effectively saying that even this minimal threshold has not been met.
The case began as a policy reform. The Delhi Excise Policy for 2021-22, introduced by the Aam Aadmi Party government, sought to exit state-run liquor retail and allow private participation. The opposition, led primarily by the Bharatiya Janata Party and supported by the Indian National Congress, alleged irregularities in licence allocation and claimed that the policy facilitated cartelisation.
In July 2022, following a report from the Delhi Chief Secretary estimating financial losses, the Central Bureau of Investigation registered a case. The Enforcement Directorate followed with a money laundering investigation. Arrests came in stages. Sisodia was jailed for roughly 530 days. Kejriwal spent about 156 days in custody before receiving bail from the Supreme Court of India in September 2024.
The public narrative that accompanied these arrests was clear: this was a major corruption scandal involving high-level conspiracy. The chargesheet ran into thousands of pages. The theory advanced was that policy decisions were manipulated to benefit select private players in exchange for illicit gain.
Yet when the case reached the court for charge framing, the prosecution could not persuade the judge that the material on record even justified a trial.
The discharge order reportedly found no overarching conspiracy and no material demonstrating criminal intent embedded in the policy. It noted internal inconsistencies in the chargesheet and observed that some averments were misleading. It also found that portions of the prosecution's narrative were not supported by witness statements. With respect to Sisodia, the court held that there was no evidence linking him to criminal conduct and no recovery to substantiate allegations of illegal proceeds.
Most importantly, the court appears to have rejected the attempt to build the case primarily on the statement of an approver without adequate independent corroboration. Indian criminal jurisprudence has consistently treated approver testimony with caution. It is admissible, but rarely sufficient on its own. If corroboration is missing, the structure built around it collapses.
What makes this episode troubling is not merely that a prosecution failed. Prosecutions fail after trial all the time. What makes it troubling is that this case collapsed before trial, after arrests, extended incarceration, and years of investigation.
In theory, discharge is a procedural safeguard. It prevents individuals from being dragged through a full criminal trial when the foundational material is insufficient. In practice, however, discharge does little to compensate for the consequences already suffered. Pre-trial incarceration in India can stretch for months or years. In this case, one accused spent over seventeen months in custody. A discharge order cannot return that time.
The CBI has challenged the order before the Delhi High Court, and the High Court may yet revive the prosecution. That possibility must be acknowledged. But even if the discharge is overturned, the fact remains that a trial court found the evidentiary foundation too weak to proceed.
The institutional question is unavoidable. The CBI is India's premier investigative agency. It operates under the administrative control of the central government and frequently handles politically sensitive matters. When such an agency arrests senior opposition leaders and later fails to establish even a prima facie case, the damage is not confined to the individuals involved. It affects public confidence in the neutrality and rigour of the investigative process.
There are only two plausible explanations for this outcome. Either the investigation did not generate legally sustainable evidence, or the prosecution misjudged the strength of its own case. Both scenarios raise concerns about standards applied in high-profile cases.
Criminal law carries immense coercive power. Arrest, custody, reputational harm, political destabilisation, and years of litigation are not minor inconveniences. They are consequences of the state's most serious authority. That authority demands discipline in evidence collection and restraint in prosecution. The lower threshold at the stage of charge framing is designed to give the prosecution latitude. When even that latitude is insufficient, something has gone wrong.
For Kejriwal and Sisodia, the discharge offers significant, though provisional, relief. For the CBI, it is a moment of institutional reckoning. The High Court may alter the trajectory of this case, but it cannot erase the broader lesson. In criminal law, narrative is irrelevant unless it is anchored in proof. If the proof is weak, the case fails. And when the case involves elected officials and prolonged custody, the failure resonates far beyond the courtroom.










