The Anti-Defection Law Killed the Whip and Made the Speaker King

The Tenth Schedule was meant to do one thing. It was meant to stop the Aaya-Ram-Gaya-Ram floor-crossing that had toppled State governments through the 1960s and 1970s and made Indian legislative politics a sour joke. It was meant to discipline the rare and corrupting defection — the MLA who switched sides for a ministerial berth, the cluster of legislators who walked out together for a deal. The 52nd Amendment of 1985 was framed in that spirit and was sold to Parliament in those terms. Four decades on, the law does something larger and worse. It binds every individual legislator on every party direction, not only on confidence motions. It vests the question of disqualification in a Speaker who is a partisan officer and whose timing is not under anyone's control. And it has, through the case law that has accumulated around it, produced a constitutional ecology in which the high commands of political parties are the strongest single force in Indian legislative life. The law has lost its original purpose. It has become its opposite.

I want to argue four things in what follows. The first is that the Tenth Schedule's text was drafted too broadly, and that the breadth — the disqualification on any 'direction' from the party, not only on motions of confidence — is the structural source of most of what has gone wrong. The second is that Kihoto Hollohan, the 1992 decision that upheld the Schedule, accepted the breadth on a reasoning that has not survived contact with how the law has actually been used. The third is that the line of cases from Rajendra Singh Rana through Keisham Meghachandra to Subhash Desai is the Court repeatedly trying, and repeatedly failing, to put discipline on the Speaker's adjudicatory monopoly. The fourth is that the right repair is the one the Court has, on each occasion, declined to make: cutting the Schedule back to the motions on which government survival actually depends and moving the adjudicatory function out of the chair. The full essay defends each of these claims in turn.

What the Schedule was meant to do, and what it actually says

The 52nd Amendment of 1985 inserted clause (2) into Articles 102 and 191 and added the Tenth Schedule to the Constitution.¹ The Statement of Objects and Reasons named 'the evil of political defections' as the mischief and identified its principal symptoms — the formation and fall of State governments on the strength of unprincipled cross-overs, the corruption that lubricated the cross-overs, and the public contempt the practice had earned. The drafters' starting picture was a defined kind of conduct: a legislator who, after election on one party's symbol, joined a different party in exchange for office or money, or who walked out of his own party to bring down a government in a confidence vote. The picture was real. The Haryana of Gaya Lal in 1967, the Bihar and Uttar Pradesh of the 1970s, the post-Janata-collapse politics of the early 1980s — these are not invented. The defections were destabilising and they were corrupting.

What the Schedule actually says is something else. Paragraph 2(1)(a) disqualifies a member who has 'voluntarily given up his membership' of the party. So far the text tracks the picture. Paragraph 2(1)(b) disqualifies a member who 'votes or abstains from voting in such House contrary to any direction issued by the political party to which he belongs', subject only to the fifteen-day grace period in which the party may condone the act.³ The breadth of this clause is the heart of the problem. The text draws no distinction between a direction issued on the confidence motion that decides whether a government stands or falls, a direction issued on a money bill, a direction issued on an ordinary policy bill, and a direction issued on a private member's resolution. Every direction, on every vote, on every subject — the disqualification is the same. The legislator's conscience is not protected on subjects that have nothing to do with government survival. It is not protected on subjects on which the party itself was internally divided and on which the whip was the leadership's instrument against its own dissidents. It is not protected at all.

This was not an oversight. The 1985 drafters could have written paragraph 2(1)(b) to cover only confidence and money-bill votes; the Private Member's Bill of 2010 proposed exactly that wording.¹³ The 1985 wording was a choice, and the consequence of the choice is the law we have. In practice, as MR Madhavan's recent survey records, political parties issue whips on most issues, not only those that affect government stability.⁴ When Parliament debated foreign direct investment in retail in December 2012 — a policy question on which both the ruling alliance and the principal opposition had internal divisions — news reports of the time captured both whips arguing not with their own MPs but with the leaders of smaller parties. The negotiation was with party leaders, not with members. The members were already locked in by the whip. Madhavan's observation that this 'effectively converts each Member of Parliament to a mere number that the party leader can count on' is not a rhetorical flourish. It is a description of how the Schedule has reshaped the everyday work of Parliament. The legislator who reads the bill and disagrees with the party's position is not free to vote her conscience. She is free to lose her seat.

Kihoto Hollohan and the reasoning the case did not stress-test

The constitutional challenge to the Tenth Schedule reached the Supreme Court within four years of its enactment, and it was decided in 1992 in Kihoto Hollohan v Zachillhu.² The five-judge bench split three-to-two. The majority opinion, delivered by Venkatachaliah J, upheld the constitutionality of the Schedule with one structural exception. Paragraph 7, which had purported to bar any court from entertaining a matter connected with a disqualification under the Schedule, was struck down — but on a procedural ground rather than a substantive one. The Court held that paragraph 7 effected a 'change' in Articles 136, 226, and 227, and therefore required ratification by half the State legislatures under the proviso to Article 368(2). It had not been so ratified. Paragraph 7 was severable from the rest of the Schedule, and so the rest stood. The dissent of Verma and Sharma JJ would have struck down the entire Schedule on the substantive ground that it violated parliamentary democracy as a basic feature of the Constitution.

The majority's reasoning on the breadth of paragraph 2(1)(b) is worth quoting at length because it has not aged well. Venkatachaliah J wrote that 'a political party functions on the strength of shared beliefs' and that 'any freedom of its members to vote as they please independently of the political party's declared policies will not only embarrass its public image and popularity but also undermine public confidence in it which, in the ultimate analysis, is its source of sustenance — nay, indeed, its very survival.'⁵ This is a recognisable theory of representative government — the responsible-party theory, associated with Edmund Burke's letter to the Sheriffs of Bristol and with the British Labour Party's twentieth-century practice. But it is a theory, and it was being deployed to defend a constitutional disqualification, and the deployment skipped the work of distinguishing the cases in which the theory is strong from the cases in which it is weak.

The responsible-party theory is strongest on the question of government survival. A party that wins an election on a manifesto, forms a government, and faces a confidence motion has a strong claim that its members owe it their vote, because the alternative is a government formed by people who did not win the election and were not authorised by the manifesto. The theory is weakest on questions of conscience — when the bill before the House is one on which the party itself did not take a manifesto position, or one on which the party's manifesto position contradicted its current government's position, or one on which the legislator's constituents have a view that differs from the party's. The Tenth Schedule treats all of these as the same. Kihoto Hollohan's reasoning treats them as the same. That treatment is not defensible on the theory the case itself relies on. A theory that requires party discipline on the confidence vote does not require party discipline on the criminal-procedure bill, the personal-law bill, the constitutional amendment bill that was not in the manifesto, the foreign-investment bill that the party itself was divided on. The reasoning in Kihoto Hollohan applied the strongest version of the theory to the entire field. The field, as Madhavan documents, has been the entire legislative session.⁴

The second piece of Kihoto Hollohan that has aged badly is the holding on the Speaker. The majority held that the Speaker, while exercising the adjudicatory function under paragraph 6(1) of the Schedule, acts as a tribunal — and that the Speaker's decision, while 'final' in the sense of paragraph 6(1), is amenable to judicial review for violation of constitutional mandates, mala fides, non-compliance with the rules of natural justice, and perversity. The Court refused to anticipate the Speaker's likely partiality. Verma J's dissent did anticipate it, on the simple ground that the Speaker is the nominee of the ruling party and that the disqualification of opposition members or of dissidents within the ruling party will systematically benefit the ruling party. This was treated by the majority as speculation. Three decades of practice have made it not speculation. The Speaker is the ruling party's nominee, the Speaker's decision is reviewable in principle but not in time, and the time at which the Speaker decides — early or late, before the floor test or after, in the same term or in the next — is the variable that, more than any other, determines the political consequences of the Schedule. Kihoto Hollohan's assumption that judicial review would discipline the office has held only in the most egregious cases. The everyday administration of the Schedule belongs to the chair. A parallel doctrinal expansion of the Speaker's discretion came in Ravi S Naik v Union of India, where the Court read paragraph 2(1)(a)'s 'voluntarily given up' formulation to be inferable from a member's conduct — including from newspaper reports and photographs — without a formal resignation.⁶ The evidentiary discretion the chair acquired through Ravi S Naik, combined with the timing discretion the chair already had through Kihoto Hollohan's silence on timelines, produced the office I am describing.

Rajendra Singh Rana: the Court tries to put a clock on the chair

The first sustained judicial attempt to discipline Speaker timing came in Rajendra Singh Rana v Swami Prasad Maurya in 2007.⁷ The facts came out of the Uttar Pradesh Assembly's 2003 crisis, in which a faction of the Bahujan Samaj Party had voted against its party in a confidence motion. Disqualification petitions were filed against the dissenting members. The Speaker did not decide them. Months passed; then a year. The petitioners eventually moved the High Court, which directed the Speaker to decide. The Speaker again did not decide. The matter reached the Supreme Court. Sabharwal CJ's bench held that the Speaker's failure to decide for an extended period was, in itself, a 'failure to exercise jurisdiction' and therefore amenable to judicial review on the basis that Kihoto Hollohan had left open. The Court ordered the petitions to be disposed of within two months.

Rajendra Singh Rana was important not for its remedy — the remedy was specific to the petition before it — but for the principle. It established that Speaker delay was itself a constitutional wrong, that the chair did not have a free hand to sit on disqualification petitions until the political question they raised had been overtaken by events. The principle, however, depended on the Court's willingness to enforce it in subsequent cases. The willingness was uneven. Through the 2010s, in case after case from Goa, Karnataka, Madhya Pradesh, Manipur, and Maharashtra, the pattern repeated. Disqualification petitions were filed; the Speaker delayed; the disqualification, if it eventually came, came after the political crisis had been resolved by other means. The two-month timeline of Rajendra Singh Rana had no enforcement architecture. The Court could direct in a particular case but the next case began the cycle again.

The deeper problem was that the Speaker's delay was not, in most cases, plainly mala fide. The Speaker was busy. The Speaker had given hearings, reserved orders, recorded the volume of materials placed before the chair. The order was not refused; it was deferred. The constitutional vocabulary of Kihoto Hollohan — mala fides, non-compliance with natural justice — did not have a category for the indefinite deferral that did the political work without ever being plainly improper. Rajendra Singh Rana's 'failure to exercise jurisdiction' was the closest the Court got to that category, and it required a long pattern of inaction before a court would intervene. By the time the long pattern of inaction had accumulated, the term of the legislature had often ended.

Keisham Meghachandra: the clock with no enforcement

In 2020 the three-judge bench in Keisham Meghachandra Singh v Speaker, Manipur Legislative Assembly tried again.⁹ The facts were similar: disqualification petitions filed; Speaker delaying; the political moment passing. The Court's response was to articulate a presumptive timeline. Disqualification petitions should ordinarily be decided 'within a reasonable period of three months', the Court said, unless 'exceptional circumstances' were shown for a longer period. The Court also did something more interesting. It urged Parliament to consider amending the Constitution to vest the disqualification decision in a separate, independent tribunal rather than in the Speaker. This was the closest the Supreme Court has come to saying that the Speaker's adjudicatory role under the Schedule is the architectural flaw.

The presumptive three-month timeline has not worked. The 'exceptional circumstances' qualifier swallowed the rule almost immediately. A Speaker who wishes to delay can point to the volume of pleadings, the request for cross-examination, the holiday calendar of the assembly, the pending writ petition before the High Court, the unavailability of counsel, the need for additional written submissions. None of these is plainly an abuse. Each is a step in a process. The cumulative effect is the same as before: the decision comes after the political moment has passed. The Court's request to Parliament for a constitutional amendment has gone unanswered. The architectural flaw is still there.

The reason the architectural flaw matters is structural and worth being clear about. The Speaker is elected by the House and, by long convention, comes from the ruling party. The Speaker is not required to resign from the party on election. The Speaker does, by the practice of every legislature in the Republic, retain the political loyalties that secured the election. The Tenth Schedule then gives that Speaker the power to decide whether members of the opposition, or dissidents within the ruling party, lose their seats for voting contrary to a party direction. The asymmetry is structural. It is not the fault of any particular Speaker. It is the consequence of placing an adjudicatory function in an office that is, by design, partisan. Sidharth Chauhan's recent survey of legislative privileges makes the same point in a different context — that legislative committees and Speaker-led adjudications tend to become 'self-interested inquisitorial bodies that take disproportionate action against political opponents'.¹² The pattern is not specific to the Tenth Schedule. The Tenth Schedule is where the pattern is most consequential.

The merger exception and how it reshapes the calculation

The 1985 Schedule originally contained two exceptions to the disqualification rule. Paragraph 3 protected one-third of a party's legislators who 'split' from the original party. Paragraph 4 protected two-thirds of a party's legislators who 'merged' with another party. The 91st Amendment of 2003 removed paragraph 3.⁸ Paragraph 4 — the two-thirds merger exception — remains.

The retention of paragraph 4 has been more consequential than is generally noticed. The Schedule, in its current form, disqualifies the individual legislator who votes against the whip on a routine bill but does not disqualify the legislative group that organises a two-thirds defection from the party as a 'merger'. The threshold is high enough that it requires real political work to clear it, but not so high that it cannot be cleared. The Maharashtra crisis of June 2022, in which approximately two-thirds of the Shiv Sena's legislators broke from the Uddhav Thackeray faction and aligned with the Eknath Shinde camp, was precisely the kind of action paragraph 4 was supposed to cover — if the two-thirds threshold was crossed and if the action could be characterised as a merger with a different party rather than a split within the original party. Whether it could be so characterised was the constitutional question that ended up before the Supreme Court in Subhash Desai v Principal Secretary, Governor of Maharashtra.¹¹

The structural point is this. The Schedule disciplines the individual legislator more sharply than it disciplines the organised faction. The legislator who reads the bill and votes her conscience loses her seat. The faction that crosses the two-thirds threshold and engineers a merger does not. This inverts the framers' picture of what the Schedule was supposed to do. The framers were worried about organised cross-overs orchestrated by political fixers. The Schedule, as it operates, leaves a route open for organised cross-overs (the paragraph 4 merger) while shutting the door on the individual conscience vote (the paragraph 2(1)(b) whip). The 91st Amendment closed the smaller of the two routes (the one-third split) but left the larger one open. The legislative-discipline architecture is built around the wrong target.

Subhash Desai: the doctrine confirmed, the asymmetry preserved

The five-judge bench decision in Subhash Desai v Principal Secretary, Governor of Maharashtra in May 2023 was the Court's most recent opportunity to address what the Schedule has become.¹¹ The facts I have just summarised. The constitutional questions were several, and not all of them concerned the Tenth Schedule. The most important Tenth Schedule question was who decides which of two rival factions is the 'real' Shiv Sena, for the purpose of identifying the 'political party' under paragraph 2(1)(b) and the 'two-thirds' threshold under paragraph 4. The Court held that the question was for the Speaker to decide, not the Election Commission, and not the Court itself in the first instance. The disqualification petitions were remitted to the Speaker, who was directed to decide them 'within a reasonable period'. No hard timeline was set.

The decision is doctrinally consistent with Kihoto Hollohan. It is also a missed opportunity. The Court was sitting on a record that documented exactly the architectural problems I have described: a Speaker who had been elected during the political crisis the petitions arose from, a Speaker whose impartiality on the rival-faction question was contested by both sides, a Speaker who had every political incentive to favour the faction whose support kept the government in office. The Court's answer — that the Speaker should decide, and that the Speaker should decide within a reasonable period — was the same answer that Kihoto Hollohan had given in 1992 and that Keisham Meghachandra had repeated in 2020. The earlier line of Speaker-power cases — Mayawati v Markandeya Chand and Speaker, Haryana Vidhan Sabha v Kuldeep Bishnoi — had read paragraph 6(1) of the Schedule as conferring 'finality' subject only to the narrow exceptions Kihoto Hollohan had carved out; Subhash Desai declined to widen those exceptions.¹⁰ It was, by 2023, an answer the Court itself had three decades of evidence against. The evidence did not move the architecture. The Speaker continued to be the adjudicator, the timeline continued to be 'reasonable', and the parties continued to negotiate the political crisis in the time the architecture gave them.

The lesson of Subhash Desai is that the Court is unwilling, or unable, to repair the Schedule by reading it down. The Court has now had three principal moments to do so — 1992, 2020, 2023 — and at each moment has chosen to work within the architecture rather than against it. The reason for this is intelligible. The Schedule's architecture is what Parliament enacted. The Court that read it down to its proper scope would be exercising a power closer to amendment than to interpretation. The Court has been, since Kihoto Hollohan's majority, unwilling to take that step. The consequence is that the repair, if it comes, must come from Parliament.

The counter-argument: discipline holds the Republic together

The strongest case for the Schedule as it stands is the case that defenders of the law have made, in good faith, since 1985, and the case deserves to be put in its strongest form. Coalition politics in India is structurally unstable. The years between 1989 and 2014 were the years of coalition government at the Centre and through much of the State system. Governments rose and fell on the votes of small parties and individual legislators. The Tenth Schedule, on this account, is what made coalition government possible at all. Without it, every confidence motion would have been an auction. Every policy bill that touched the interest of a faction would have been an opportunity for blackmail. The 1985 Parliament, this account says, looked at the politics of the 1970s and saw what unprincipled defection had done to the State governments of Haryana, Bihar, and Uttar Pradesh, and decided that the cost of an over-broad discipline measure was lower than the cost of legislators continuously available to the highest bidder. The discipline matters because the alternative is worse.

I take this seriously. The post-Aaya-Ram-Gaya-Ram politics was a real political failure of the 1960s and 1970s. The 52nd Amendment had a real mischief to remedy. And it is not true, in any straightforward sense, that the Schedule has not worked at all. Floor-crossing for personal gain in a confidence vote, of the kind that brought down State governments through the 1970s, is much rarer now than it was. The headline number — how many State governments have fallen on the basis of unprincipled cross-overs since 1985 — is lower than it was in the comparable period before 1985. The Schedule has done some of the work it was meant to do.

But the strongest version of the case has to acknowledge two things. The first is that the headline number is misleading, because the cross-overs have not stopped; they have organised. The two-thirds merger exception in paragraph 4 has produced a politics in which the deal-making moves from the individual to the faction, and the faction-level cross-over is harder to police because it has constitutional cover. Maharashtra in 2022 is the most visible recent example; the Madhya Pradesh switch of 2020 and the Karnataka episode of 2019 are others. The Schedule, in its current form, channels the politics it was meant to suppress rather than ending it. The second is that the price the Schedule extracts for whatever discipline it produces is borne in a place the 1985 drafters did not look. It is borne by the everyday business of the legislature. The deliberative space that a parliament is supposed to be — where members read bills, hear arguments, change minds — has narrowed. The legislator who is bound to vote the party line on every motion is not deliberating; she is recording a decision someone else made. The legislature is more disciplined and less a legislature. The defenders of the Schedule have to accept this trade. The trade is not a free good.

Even granting the trade, the question is whether the trade can be made on better terms. The Schedule as drafted is not the only possible discipline measure. A Schedule that disqualified only on confidence motions, money bills, and adjournment motions would preserve the government-stability rationale while restoring conscience on every other vote. The 2010 Private Member's Bill proposed exactly this and Madhavan's recent survey endorses the proposal as the right balance.¹³ The case for the Schedule as it stands cannot rest on a comparison with no Schedule at all. It has to rest on a comparison with a Schedule properly tailored to the mischief. On that comparison, the current Schedule loses.

The Speaker, the timeline, and the institutional repair

The second piece of the repair is the Speaker. The case for moving the adjudicatory function out of the chair is not a case against any particular Speaker. It is a case against placing an adjudicatory function in an office that the Constitution has designed to be partisan. The Speaker is elected by the House, comes from the ruling party, is not required to resign from the party, and serves at the political pleasure of the majority that elected her. None of these features is wrong as a feature of the chair's legislative function. All of them are wrong as features of an adjudicatory function. The Schedule combined the two functions in 1985 and has paid the price for the combination ever since.

What would the alternative look like? Keisham Meghachandra's proposal — an independent tribunal — is one form. A panel of retired judges of the Supreme Court and the High Courts, with the panel's chair selected by the Chief Justice of India and the Leader of the Opposition, would have the institutional credibility the Speaker's office cannot manufacture. A constitutional amendment to insert a fresh paragraph 6 in the Tenth Schedule, vesting the disqualification decision in such a tribunal and binding the tribunal to a hard timeline of, say, sixty days from the filing of the petition, is the architectural fix. The Court asked for the fix in 2020. Parliament has not provided it. The reason is that no government — whichever party is in power — wants to lose the Speaker's discretionary control over the timing of disqualification. The political incentive to retain the current arrangement is bipartisan. So is the reason the arrangement should be changed.

Two smaller repairs are also available without a constitutional amendment. The first is a judicial repair: the Court could read the 'reasonable period' standard hard, treating delay beyond ninety days as itself constituting a jurisdictional default, with the consequence that the petition is deemed dismissed (which protects the dissident) or that the petition is decided by a default rule (which protects the petitioner). The Court has not been willing to take this step, but the doctrinal opening from Rajendra Singh Rana and Keisham Meghachandra is there. The second is a procedural repair through the Rules of Procedure: each House could amend its rules to require the Speaker, on receipt of a Tenth Schedule petition, to publish a calendar of hearings within seven days and to deliver an order within sixty days, with the calendar reviewable in writ jurisdiction. The Speaker would still be the adjudicator, but the timing discretion would be reduced. These are second-best repairs. They are still repairs.

What changes if the argument is right

If the case I have made is correct, three things follow.

The first is that the standard description of the Tenth Schedule as a discipline measure for floor-crossing has to be replaced by a more accurate description: it is a whip-enforcement instrument that gives party high commands a constitutional sanction over the everyday conduct of their legislators, and it is administered through an office whose impartiality is not designed into the office. The 1985 sale is no longer the truth of the law in 2026. The textbook entries that describe the Schedule's purpose as 'curbing political defections' are describing the Schedule's marketing material, not its operation.

The second is that the locus of repair is Parliament, not the Court. The Court has done what the doctrine of Kihoto Hollohan permitted it to do and is not willing to do more. Rajendra Singh Rana, Keisham Meghachandra, and Subhash Desai are a sequence of attempts to police timing within an architecture the Court will not redesign. The architecture has to be redesigned by amendment, and the amendment has to do two things: cut paragraph 2(1)(b) back to confidence motions, money bills, and adjournment motions; and move paragraph 6(1) out of the Speaker's office.

The third is that the political coalition for the repair has to come from outside the high commands. The high commands have no incentive to give up the whip on every motion. The Speaker's office has no incentive to surrender its adjudicatory monopoly. The repair, if it comes, will come from backbenchers who have lived inside the Schedule for forty years and who understand what it has done to their working lives. The 2010 Private Member's Bill came from that constituency. The Bill lapsed. A Bill that does not lapse is the constitutional politics this question now requires.

The Tenth Schedule is not a small provision. It governs the daily conduct of about five thousand elected legislators across the Union and the States, and through them, the deliberative work of the country's parliamentary system. Forty years is long enough to know what a law has become. The Tenth Schedule has become a law for the high commands. It should be a law for the legislatures. The architecture for the change is not difficult. The political will is. The case for finding that will is the case this essay has made.

Get the week's biggest decisions, distilled.

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