P.V. Narasimha Rao v. State (1998): the JMM bribery case and parliamentary immunity

The paradox of P.V. Narasimha Rao v. State (CBI/SPE) — the case the public came to know as the "JMM bribery case" — is that a charge of buying votes to save a government was met not with a straightforward criminal trial but with a constitutional shield. A five-judge Constitution Bench of the Supreme Court divided 3:2 over whether legislators who allegedly took money to vote a certain way could be tried at all. The majority concluded that some of them could not. That holding — counterintuitive on its face, and stated with a fine and consequential distinction — governed the law for a quarter of a century before it was unanimously overruled.

The facts in brief

In July 1993 a no-confidence motion was moved against the minority government headed by Prime Minister P.V. Narasimha Rao. The motion was defeated, and the government survived. It was later alleged — the CBI laid charges in 1996 — that the survival had been bought: that bribes were paid to Members of Parliament belonging to the Jharkhand Mukti Morcha (JMM) and to the Janata Dal (Ajit Singh faction) so that they would vote against the motion and keep the government in office.

The prosecution named two sets of accused. Rao and others figured as the alleged bribe-givers; the JMM and Janata Dal (Ajit Singh) MPs as the alleged bribe-takers. A factual detail that proved decisive: most of the alleged bribe-takers voted against the no-confidence motion, but Ajit Singh, who was also alleged to have been part of the bargain, abstained — he did not cast a vote at all.

The questions

Two questions of principle ran through the case. The first concerned the reach of Article 105(2) of the Constitution, which provides that no Member of Parliament shall be "liable to any proceedings in any court in respect of anything said or any vote given" by him in Parliament — and the parallel guarantee for State legislators under Article 194(2). How far does that immunity stretch? Does it cover only proceedings that attack the speech or vote itself, or does it extend to any proceeding that bears a connection — a nexus — to the vote?

The second question turned on the meaning of the phrase "in respect of." The Attorney General urged a narrow reading; the accused MPs urged a broad one. On the narrow reading, a bribery prosecution is about the antecedent act of taking money, not about the vote, and so falls outside the immunity. On the broad reading, "in respect of" embraces anything arising out of or connected with the vote — and a prosecution premised on how a member voted would be foreclosed.

Beneath both lay the sharpest issue of all, and the one on which the Bench would divide: does a member who takes a bribe and then votes stand differently from a member who takes the bribe but does not vote?

What the Court held

The Bench split 3:2.

The majority — S.P. Bharucha, J., writing for himself and S. Rajendra Babu, J., with G.N. Ray, J. concurring — read Article 105(2) broadly. The phrase "in respect of," the majority held, must receive a wide meaning: the immunity is not confined to proceedings whose cause of action is the speech or vote, but reaches "all civil and criminal proceedings that bear a nexus to their speech or vote." The purpose of the protection, the majority reasoned, is to let members participate fearlessly in debate and voting, and that purpose would be defeated if a member could be hauled into court over the motivation behind a vote.

Applied to the facts, this produced the holding the case is remembered for. The alleged bribe-takers who had actually voted against the no-confidence motion had, on the majority's view, cast votes that were protected; a prosecution alleging they had been paid to vest those votes a particular way was a proceeding "in respect of" those votes and was therefore barred. They could not be tried.

But the same logic cut the other way for a member who never voted:

Because Ajit Singh abstained, there was no protected vote to which a prosecution could attach. He fell outside the shield — and, the majority added, the bribe-givers who were alleged to have conspired with him could likewise be prosecuted in relation to him, the immunity attaching only to those bribe-takers who had actually voted. The result was a precise and unusual line: vote, and be immune; take the money but do not vote, and be exposed.

The minority — S.C. Agrawal, J., writing the dissent, joined by A.S. Anand, J. — would have permitted prosecution in every case. For the dissent, the offence of bribery is complete the moment the illegal gratification is accepted; it is antecedent to, and independent of, any vote. Agrawal, J. put it directly: "The offence is complete with the acceptance of the money or on the agreement to accept the money being concluded and is not dependent on the performance of the illegal promise by the receiver." On that footing the liability of a bribe-taking member "arises independently of the making of the speech or giving of vote" — and a prosecution for it is therefore not a proceeding "in respect of" a vote at all. The dissent declined to stretch "in respect of" into "arising out of," and so found no immunity whether or not the member ultimately voted.

Analysis

The fault line between the opinions is, at bottom, a disagreement about what the prosecution is about. The majority saw a bribery case against a member who voted as inescapably an inquiry into the motive for that vote: to prove the charge, one would have to show the vote was bought, and that is exactly the kind of scrutiny Article 105(2) was meant to foreclose. Drawing on United States authority — the majority found the dissent of Brennan, J. in United States v. Brewster more persuasive than that case's majority — it treated the act of taking a bribe to vote as bound up with the vote itself, not "casually or incidentally" related to it.

The dissent located the wrong elsewhere. Bribery, in its analysis, crystallises on acceptance; the later vote is irrelevant to whether the offence was committed, because "the receiver of the money will be treated to have committed the offence even when he defaults in the illegal bargain." If the crime is complete before any vote is cast, then prosecuting it does not require, and is not "in respect of," the vote. On this reasoning the abstaining member and the voting member stand in exactly the same position — both committed bribery on acceptance — and there is no principled basis to immunise one and not the other.

That last point exposed the discomfort in the majority's line. It is, on its face, perverse that the law should reward the member who completed the corrupt bargain by voting while exposing the one who took the money and reneged. The majority did not deny the gravity of the conduct — it spoke of bribe-takers who "bartered a most solemn trust" and "enabled a Government to survive" — but held that the constitutional text, broadly read, plainly afforded them protection, and that "our sense of indignation should not lead us to construe the Constitution narrowly." The dissent's reply was that no such indignation-driven narrowing was needed: read correctly, the immunity simply did not cover bribery at all.

Why it matters

For twenty-six years, P.V. Narasimha Rao was the governing statement of the law on legislative bribery immunity in India, and a textbook illustration of how a broad reading of a privilege clause can collide with the integrity of the legislative process itself. Its central distinction — that immunity followed the vote, so that taking a bribe and voting attracted protection while taking a bribe and abstaining did not — was always its most criticised feature, and the one the Agrawal–Anand dissent had warned against.

In September 2023 the question was referred to a larger bench, and on 4 March 2024 a seven-judge Constitution Bench in Sita Soren v. Union of India (2024 SCC OnLine SC 229) unanimously overruled the Narasimha Rao majority. The later Court held that there is no immunity under Articles 105(2) or 194(2) for legislators against prosecution for bribery, because the offence is complete on acceptance of the bribe — squarely adopting the position of the 1998 dissent. Sita Soren is covered separately on Valkya.

For practitioners, Narasimha Rao now matters chiefly as the decision that Sita Soren displaced — but it remains essential reading. It frames the precise question the seven-judge Bench had to answer, it preserves the most fully argued statement of the broad-immunity view, and its dissent reads today as the law. Where a question turns on the reach of Articles 105(2) or 194(2), the Narasimha Rao majority is no longer good law; but understanding why it was wrong begins with understanding exactly what it held.

Related on Valkya

• Sita Soren v. Union of India
• Kihoto Hollohan v. Zachillhu
• Subhash Desai v. Governor of Maharashtra

Sources

• Bar & Bench, "Bribery in Parliament: What did the 1998 PV Narasimha Rao judgment say?" — barandbench.com
• Supreme Court Observer, "Sita Soren v Union of India: Judgement Summary" — scobserver.in
• SCC Online Blog, "No Immunity for MPs and MLAs from prosecution for bribery: Supreme Court overrules PV Narasimha Rao" — scconline.com
• Supreme Court Observer, P.V. Narasimha Rao v. State (CBI/SPE) judgment (hosted PDF) — scobserver.in