Listen to this article:

The Justice Kurian Joseph Committee on Union-State Relations, constituted by the Government of Tamil Nadu in April 2025 – and of which the author is a member – submitted Part I of its Report in February 2026. This article draws upon, and expands, Chapter 2 (“Amendment of the Constitution”) and Chapter 3 (“Territorial Integrity of States”) of that report.

In The Indian Constitution: Cornerstone of a Nation (1966), Granville Austin described India’s amendment procedure as “one of the most ably conceived aspects of the Constitution.” Yet even he acknowledged early criticism – that it was “too flexible” and the Constitution’s sanctity had already begun to erode due to the seventeen amendments enacted by that time.

One can only imagine what Austin might have written had he witnessed what followed: 106 constitutional amendments in 76 years, of which at least 15 moved from introduction to Presidential assent in ten days or less. The lesson is stark: a Constitution that is too easy to amend can, in the wrong hands, be too easy to bend.

The 39th Constitutional Amendment (1975), which placed election disputes concerning the President, Vice-President, Prime Minister and Speaker beyond judicial review and retrospectively validated formper prime minister Indira Gandhi’s electoral victory after it had been set aside by the Allahabad High Court, exemplified this danger.

The 42nd Constitutional Amendment (1976) enacted at the height of the Emergency – went even further. In one fell swoop, it altered the Preamble, amended 40 Articles, inserted 14 new Articles and two new Parts, revised the Seventh Schedule and, most ominously, inserted clauses (4) and (5) into Article 368 (the amending provision), declaring that no amendment “shall be called in question in any court on any ground” and that there shall be “no limitation whatever” on Parliament’s constituent power.

It is true that the 44th Constitutional Amendment (1978), together with a vigilant Supreme Court, restored aspects of the constitutional equilibrium disturbed by the excesses of the Emergency. But a Republic cannot rely indefinitely on judicial correction to remedy structural infirmities in its constitutional design. The amendment process itself must embody adequate safeguards.

Recent amendment proposals bring this danger into sharper focus. The 129th Constitutional Amendment Bill (2024) proposes “One Nation, One Election”. The 130th Amendment Bill (2025) contemplates the removal of the Prime Minister, Chief Minister or a Minister upon prolonged detention. The recently defeated 131st Amendment Bill (April 2026) sought to lift the 50-year freeze on inter-state Lok Sabha representation while rushing through sweeping delimitation changes within three days, including expanding the House to 850 seats.

These proposals reveal a central truth: only a pliant constitution permits measures of such import to be advanced with such haste and so little deliberation. In mature federations, whose amendment procedures are designed with far more exacting safeguards, such aberrations would be inconceivable.

The amending provision

The amending provision is the most vital part of any constitution – its very lifeline. If one were to trace the moral biography of the Indian constitution, nothing would illuminate it more clearly than the evolution of Article 368, its amending clause. The story is familiar, but merits brief recall.

On September 17–18, 1949, the Constituent Assembly debated a three-tier scheme of amendment, of which the first two tiers found expression in Article 368.

First, the bulk of constitutional provisions were to be amendable by a special majority in Parliament – requiring both a majority of the total membership of each House and a two-thirds majority of members present and voting.

Second, a narrow set of federally sensitive provisions was to be subject, in addition, to ratification by not less than one-half of the State Legislatures.

Third, provisions of a minor, transitional or administrative character were kept outside Article 368 and could be altered by parliament through ordinary legislation – by a simple majority of members present and voting in each House – under specific enabling provisions.

This amendment scheme rendered India, in Dr B.R. Ambedkar’s evocative phrase, “a flexible federation” – perhaps, in retrospect, more flexible than prudence would justify. The framers’ preference for a flexible amendment procedure was understandable in 1950. India was then a newly independent and fragile Union – recovering from Partition, integrating hundreds of princely states, and anxious to avoid constitutional paralysis. Flexibility appeared indispensable for institution-building and national consolidation. But what was once a necessity has, over time, become a vulnerability.

The constitutional limits of the amending power were tested early. In Shankari Prasad v. Union of India (1951), the Supreme Court upheld the 1st Constitutional Amendment (1951) on the reasoning that the word “law” in Article 13(2) referred only to ordinary legislation and not to a constitutional amendment under Article 368. Accordingly, the amending power was held not to be subject to Article 13(2)’s prohibition against laws that take away or abridge fundamental rights.

This interpretation was reaffirmed by a 3:2 majority in Sajjan Singh v. State of Rajasthan (1965), which upheld the 17th Constitutional Amendment (1964) and endorsed parliament’s authority to amend any part of the constitution, including fundamental rights. However, the dissents of Justices J.R. Mudholkar and M. Hidayatullah questioned whether an unlimited amending power could extend to altering the very identity of the Constitution itself.

A decisive reversal came in I.C. Golak Nath v. State of Punjab (1967). By a slender 6–5 majority, the Supreme Court held that constitutional amendments were “law” within the meaning of Article 13(2) and therefore could not abridge fundamental rights. The judgment placed fundamental rights beyond the reach of the amending power, setting the stage for a constitutional confrontation.

Parliament responded through the 24th Constitutional Amendment (1971). It amended the title of Article 368 from a mere “procedure for amendment” to “Power of Parliament to amend the Constitution and procedure therefor”. It inserted a new clause (1) affirming parliament’s constituent power to amend “any provision” of the constitution – and redesignated the original Article 368 as clause (2). Presidential assent was made obligatory. It added clause (3) to Article 368, expressly excluding the application of Article 13 to constitutional amendments, and introduced Article 13(4) to reinforce the same position. Through this, parliament sought to restore its supremacy in matters of constitutional change.

The issue reached its culmination in Kesavananda Bharati v. State of Kerala (1973). By a slender 7–6 majority, the Supreme Court held that parliament’s amending power, though wide, is not unlimited: it cannot alter or destroy the “Basic Structure” of the constitution. This judicially evolved doctrine recognises that the Constitution derives its identity not merely from individual provisions, but from deeper organising principles – constitutional supremacy, democracy, republicanism, federalism, secularism, separation of powers, judicial review and judicial independence.

These principles pervade and interlink the entire constitutional scheme. To abrogate them would not be to amend the constitution, but to destroy its essential identity.

Subsequent events vindicated this limitation. The 42nd Constitutional Amendment (1976) sought to place constitutional amendments beyond judicial scrutiny and to confer unbounded constituent power on Parliament. In Minerva Mills v. Union of India (1980), the Supreme Court struck down these provisions, reaffirming that the amending power is subject to inherent constitutional limits.

This trajectory reveals a deeper truth: the Basic Structure doctrine emerged as a judicial response to an overly pliant amendment process and the absence of robust procedural safeguards.

Rethinking Article 368

The recommendations that follow seek to articulate a modernised amendment framework that restores federal balance, deepens democratic legitimacy, and ensures meaningful institutional restraint.

1. Mandate prior public consultation for constitutional amendments

The Preamble proclaims that “We, the People of India … give to ourselves this Constitution”. Yet, paradoxically, the people play no formal role – direct or consultative – in amending the very charter they are said to have enacted. This omission stands in sharp contrast to the far more participatory precedent set by the Constituent Assembly.

The Drafting Committee under B.R. Ambedkar released its first draft of the constitution on February 21, 1948 for wide public scrutiny, inviting comments from citizens, provincial governments, central ministries, the judiciary, and members of the assembly. The revised draft, substantially shaped by this extensive consultation, was introduced over eight months later, on November 4, 1948.

No comparable process exists today. Constitutional amendments are often rushed through parliament, and even the Union government’s pre-legislative consultation policy, applied sporadically to ordinary bills, is almost never invoked for constitutional change.

Institutionalising a mandatory, time-bound process of public consultation for all constitutional amendments would reduce opacity, check legislative haste, and significantly enhance democratic legitimacy. It would tap into the collective intelligence of the citizenry – the “wisdom of the crowds” – which no closed parliamentary process, however well-intentioned, can fully replicate. More importantly, it would revive the participatory ethos that guided the framing of the constitution itself.

Insert a new clause (2A) in Article 368 stipulating that no Bill to amend the Constitution shall be introduced unless the text of the proposed amendment has been published for public consultation for a period of not less than three months.

2. Mandate minimum deliberation periods for constitutional amendments

A constitution is not an ordinary legal instrument; it is a founding charter that gives shape to the nation’s political order. Its amendment, therefore, should be undertaken with caution, restraint, and deep deliberation, lest its fundamental identity be unsettled. Regrettably, the counsel of H.V. Kamath in the Constituent Assembly that no constitutional amendment should be finally passed until at least six months had elapsed from its introduction to permit informed public debate, went unheeded. The record – 106 amendments in 76 years – speaks for itself, often marked by episodes of disquieting legislative haste, even in matters of profound constitutional consequence.

The 67th Constitutional Amendment (1990), which extended President’s Rule in Punjab by one year, was introduced in the Lok Sabha on October 4, 1990, passed by both Houses, and received presidential assent on the very same day.

The 103rd Constitutional Amendment (2019), which introduced 10 per cent reservation for the Economically Weaker Sections (EWS) outside the SC, ST, and OBC categories – over and above the long-standing 50 per cent ceiling on total reservations fixed by the Supreme Court – was introduced in the Lok Sabha on January 8, 2019, passed there the same day, approved by the Rajya Sabha on January 9, and received Presidential assent on January 12.

The controversial 39th Constitutional Amendment (1975), referred to earlier, represents the high-water mark of legislative velocity in India. Introduced in the Lok Sabha on August 7, 1975, it was passed by the Lok Sabha the same day, approved by the Rajya Sabha on August 8, ratified by one-half of all the states on August 9, and received Presidential assent on August 10, all in just three days.

An early episode underscores the need for procedural restraint. The 3rd Constitutional Amendment Bill (1954) received Presidential assent before the Mysore Legislature could even consider it, on the ground that the requisite number of states had ratified it. Though constitutionally valid, propriety required that assent be deferred until all states had a fair opportunity to deliberate.

These instances are illustrative, not exhaustive. Whatever the merits of the proposals, such procedural velocity would be questionable even for ordinary legislation; in the case of constitutional amendments, it is plainly inappropriate. As Alexander Hamilton observed presciently in The Federalist No. 70 (1788):

“In the legislature, promptitude of decision is oftener an evil than a benefit. The differences of opinion, and the jarring of parties in that department of the government, though they may sometimes obstruct salutary plans, yet often promote deliberation and circumspection, and serve to check excesses in the majority.”

“In the legislature, promptitude of decision is oftener an evil than a benefit. The differences of opinion, and the jarring of parties in that department of the government, though they may sometimes obstruct salutary plans, yet often promote deliberation and circumspection, and serve to check excesses in the majority.”

Comparative federations incorporate deliberate temporal safeguards to moderate the pace of constitutional change and guard against impulsive or ill-considered reform.

Australia (Section 128): A proposed amendment must be put to a referendum not earlier than two months and not later than six months after passage by Parliament.

Canada (Section 39): No proclamation may be issued for one year after Parliament adopts a resolution unless all required Provinces act sooner.

Belgium (Articles 195 & 196): The declaration that certain articles are open for amendment automatically dissolves Parliament, and fresh elections must be held. The new Parliament may amend only those articles previously declared open.

Spain (Article 168): There is a similar process for major amendments: dissolution of Parliament and fresh elections; re-approval by a two-thirds majority of the new Parliament; and a mandatory referendum.

India would benefit from comparable guardrails of time, transparency, and reflection.

It may be contended that the introduction of the Tenth Schedule through the 52nd Constitutional Amendment (1985), and its tightening by the 91st Amendment (2003), has reduced meaningful legislative debate to a formality. By compelling adherence to party whips on pain of disqualification, the anti-defection regime has significantly narrowed the space for independent deliberation. This distortion is further compounded by the growing dominance of the Executive, with most enacted laws originating as Government Bills.

Insert a new clause (2B) in Article 368 providing that no Constitutional Amendment Bill shall be taken up for final voting in any House or Legislature until at least three months have elapsed from its introduction in that body. Further, where state ratification is required, the President shall not grant assent until at least one year has elapsed from the date of its first introduction, unless all legislatures concerned have completed their consideration and communicated their decisions earlier.

Amend Paragraph 2 of the Tenth Schedule to confine the binding force of party whips strictly to ‘core stability’ votes – confidence motions, no-confidence motions and money bills. This would restore space for principled intra-party dissent on policy, legislation, and constitutional amendments, enabling legislators to act as reasoned representatives of their constituents while preserving the stability of the government.

3. Codify the basic structure of the constitution as unamendable

Modern constitutional design recognises that certain foundational principles must lie beyond the reach of transient majorities. Article 79(3) of Germany’s Basic Law of 1949, the celebrated Eternity Clause, exemplifies this approach. Forged in response to the failures of the Weimar Republic and the rise of Nazi totalitarianism, it renders unamendable – even by overwhelming majorities – the core guarantees of human dignity, fundamental rights, democracy, republicanism, the rule of law, and separation of powers. It also entrenches federalism, and guarantees the Länder (states) a permanent role in federal legislation through the Bundesrat.

Similarly, Brazil’s Cláusulas Pétreas (Article 60(4) of the 1988 Constitution) prohibit amendments that would dismantle federalism, undermine direct and periodic elections, disturb the separation of powers, or curtail fundamental rights – thereby safeguarding the constitutional order against regression.

By contrast, India relies on the Basic Structure doctrine – a judicial innovation that, while vital, leaves scope for contestation and interpretive uncertainty. In Indira Nehru Gandhi v. Raj Narain (1975), the Supreme Court affirmed that “basic features” must be identified case by case, declining to offer an exhaustive catalogue. The doctrine has thus evolved incrementally, with different benches recognising different features over time.

Of the 76 Constitutional Amendment Acts enacted since Kesavananda Bharati (1973), the Supreme Court has invoked the Basic Structure doctrine to invalidate only eight – seven of them in part. The sole amendment struck down in its entirety was the 99th Constitutional Amendment (2015), which sought to replace the collegium system with the National Judicial Appointments Commission. Almost all cases involved attempts to dilute judicial review or judicial independence; otherwise, the court has been reluctant to invalidate constitutional amendments.

India should now codify its core unamendable principles, drawing from models such as Germany and Brazil. A textual formulation would enhance clarity for Parliament, predictability for the states, and reduce the judiciary’s burden of case-by-case delineation.  Above all, it would fortify the constitution’s identity against transient political majorities, illiberal impulses and the risks of judicial inconsistency.

Insert a new clause (2C) in Article 368 stipulating that there shall be no amendment abridging, taking away, or destroying any of the following basic features of the constitution – namely, the supremacy of the constitution, the unity and integrity of the nation, democracy, republicanism, federalism, secularism, the separation of powers, the rule of law, the independence of the judiciary, judicial review and free and fair elections (the list being illustrative).

4. Narrow the list of provisions amendable by simple majority

In the Constituent Assembly, R.K. Sidhwa cautioned that no serious constitution permits alteration by a mere simple parliamentary majority. Yet, several provisions – originally designed for the transitional needs of a nascent Republic – remain amendable by Parliament through a simple majority of members present and voting. Granville Austin identified at least 22 such provisions, noting that “several more can be added.”

These fall into two broad categories: first, Articles that explicitly state, “No such law … shall be deemed to be an amendment of this Constitution for the purposes of Article 368,” and second, provisions that use formulations such as “Parliament may by law” or “until Parliament by law otherwise provides,” where Parliament, while ostensibly legislating, effectively exercises constituent power.

Crucially, many of these provisions directly impact state identity, powers, and the federal balance. These include Article 169 (Legislative Councils), Article 293 (State Borrowing), Article 312 (All-India Services), Article 343 (Official language of the Union), Article 348 (Language of the higher judiciary and legislation), etc.

The most serious threat to federal equilibrium lies in Article 3, which empowers Parliament to create, alter, or rename states – even to reduce them to Union Territories – through ordinary legislation. While such flexibility was justified in the aftermath of Partition and linguistic reorganisation, that rationale has long since expired. The Jammu and Kashmir Reorganisation Act, 2019 illustrates the extent of this power: a state was bifurcated and simultaneously downgraded into two Union Territories while under President’s Rule. If states can be altered or downgraded by a simple Parliamentary majority, federalism – recognised as part of the Constitution’s Basic Structure – risks becoming a hollow abstraction.

This vulnerability is compounded by the low quorum requirement under Article 100(3)merely one-tenth of the total membership of each House. Theoretically, a Bill could pass the Lok Sabha with just 55 members present and 28 votes in favour; and it could pass the Rajya Sabha with as few as 25 members present and 13 votes in favour. Such a threshold is unacceptably low even for ordinary legislation; for measures affecting the federal compact, it is wholly inadequate.

Only a narrow class of provisions should remain amendable by simple parliamentary majority:

Article 2: Admission of new territory into the Union.

Article 11: Regulation of citizenship rights.

Articles 239, 239A, 239AA, and 241: Governance of Union Territories and Delhi.

Second Schedule: Salaries and Allowances of constitutional office-holders.

All other constitutional provisions should be amendable only in accordance with Article 368.

5. Restore Federal Balance by Expanding State Ratification Requirements

The proviso to Article 368(2) presently requires ratification by “not less than one-half of the states” only for a narrow and incomplete set of federal provisions. While amendments to Union-state legislative relations require state ratification, Parliament anomalously retains unilateral authority to amend comparable domains such as Union-state administrative and financial relations. This asymmetry is difficult to defend. The wide range of provisions affecting the federal balance that remain amendable without state consent sits uneasily with the Constitution’s federal character.

By contrast, mature federations embed multi-level consent – through state ratification, referendums, or both – at the heart of constitutional change.

The United States (Article V): Every constitutional amendment requires ratification by three-fourths of the states.

Canada (Part V): Section 38 prescribes the General Amendment Procedure for most provisions; it requires approval by Parliament and at least seven of the ten provinces representing 50 per cent of the population (the “7/50 Rule”). Section 41 requires ratification by all Provinces for ‘core’ constitutional features.

Australia (Section 128): Every constitutional amendment must be approved in a referendum by both a national majority and a majority of states (at least four of six states).

Switzerland (Articles 140 and 142): Amendments to the Federal Constitution must be approved by a double-majority of voters and Cantons.

India’s linguistic States are today politically mature and territorially stable. Their near-exclusion from most constitutional amendments is no longer defensible. The Rajamannar Committee had recommended extending state ratification to all amendments. A calibrated approach would be to require state ratification for amendments affecting Union–state relations, democratic institutions, and fundamental rights – thus covering most substantive provisions, while excluding a narrow class amendable by ordinary law or through state-specific procedures. This would align India with established federal practice and ensure that constitutional change rests on genuine multi-level consent.

Rather than vastly expanding the existing proviso to Article 368(2), a more coherent reform would be to invert the present structure: recast Article 368(2) to make state ratification the general rule, subject to limited exceptions clearly set out in a new proviso.

Raise special majority threshold and introduce enhanced double-majority rule for state ratification

Under Article 368(2), India’s special majority requirement in Parliament for constitutional amendments consists of (i) a majority of the total membership of each House, and (ii) a majority of not less than two-thirds of the members present and voting.

While this appears rigorous, it is, in practice, susceptible to executive dominance. The ruling establishment, by definition, commands a majority in the Lok Sabha and often enjoys a substantial, if not decisive, presence in the Rajya Sabha. The “members present and voting” criterion further weakens the threshold: strategic absenteeism, orchestrated walkouts, or large-scale suspensions can artificially reduce the effective majority required. The sweeping changes of the 42nd Constitutional Amendment (1976), enacted during the Emergency when Opposition members were detained, remains a stark illustration of this vulnerability.

It is a telling anomaly that impeaching the President under Article 61(4) requires not less than a two-thirds majority of the total membership of each House – making it, in effect, more difficult to remove the President than to amend the Constitution itself.

By contrast, other major federations employ significantly more demanding parliamentary thresholds:

Germany (Article 79): Every constitutional amendment requires the support of two-thirds of the members of the Bundestag and two-thirds of the votes of the Bundesrat.

Brazil (Article 60): All amendments require a three-fifths majority of the total membership of each House of the National Congress, to be achieved in two successive rounds of voting.

Belgium (Article 195): Parliament must first specify amendable provisions by an absolute majority, with a quorum of a majority of members. This triggers automatic dissolution and fresh elections. The newly elected Parliament may amend only the specified provisions, requiring at least two-thirds of members present in each House and a two-thirds majority of votes cast.

Spain (Articles 167–168): Under Article 167, ordinary constitutional amendments require a three-fifths majority in both Houses, or alternatively, by a two-thirds majority in the Congress of Deputies and by an absolute majority in the Senate. A referendum is triggered if one-tenth of members of either House so demand. For “total revision” or fundamental changes under Article 168, a two-thirds majority in both Houses is required, followed by dissolution, fresh elections, re-approval by the new Parliament, and mandatory ratification by a national referendum – making Spain’s procedure among the most stringent.

India should therefore strengthen the amendment procedure by tying the parliamentary special majority to the total membership of each House and raising the threshold to two-thirds, ensuring broader national consensus.

The double majority solution

The current framework for state ratification is equally problematic. The proviso to Article 368(2), which requires ratification by not less than one-half of the states for certain amendments, allows a group of small states representing a fraction of the population to determine outcomes for the entire Union. With 28 states, the smallest 14 could, in theory, ratify an amendment while representing barely over one-tenth of the population.

This democratic deficit is compounded by the low quorum requirement in state assemblies under Article 189(3)merely ten members or one-tenth of total strength whichever is greater. In the 14 smallest states, this permits quorums as low as ten to thirteen members. In theory, constitutional amendments affecting the federal structure could be ratified by as few as six or seven MLAs in each of these states.

At the same time, reliance solely on population creates the opposite distortion. Based on the 2011 Census, five of India’s largest states account for nearly half the population, and eight for almost two-thirds. Even after the bifurcation of Andhra Pradesh in 2014, this imbalance persists. If population alone governs ratification, a handful of large states could effectively control constitutional change, reducing smaller states to near irrelevance.

As discussed earlier, mature federations such as Canada (Section 38), Australia (Section 128), and Switzerland (Articles 140 & 142) avoid both distortions by adopting a “double-majority rule”: constitutional change must secure (i) the support of a majority of constituent units and (ii) the approval of a majority of the national population. These safeguards ensure that neither a coalition of small units nor a handful of populous units can unilaterally impose constitutional change.

For contemporary India, an appropriate rule would be ratification by not less than two-thirds of the states (19 states) representing not less than two-thirds of the national population. This process upholds democratic legitimacy and preserves federal balance.

A related issue is the internal threshold for state ratification of a constitutional amendment. The current rule – a simple majority of members present and voting – is inadequate. Conversely, requiring a two-thirds majority of the total membership of the House for two-thirds of the states is unduly stringent. A balanced approach is to require ratification by a majority of the total membership of each House of the state legislature.

Amend Article 368 to:

Require every Constitutional Amendment Bill to secure the support of not less than two-thirds of the total membership of each House of Parliament; and

Introduce a double-majority rule for state ratification – approval by not less than two-thirds of the states representing not less than two-thirds of India’s population.

Ratification in each state should require a majority of the total membership of the Legislative Assembly and, where applicable, the Legislative Council.

7. Enable State-Initiated Amendments

The first draft of the Constitution (February 1948), permitted state legislatures to initiate amendments regarding the method of choosing the Governor and the number of Houses in the state legislature. However, this was omitted from the final text, resulting in a striking constitutional asymmetry: although India is a “Union of States,” the states possess no formal power to initiate constitutional change or compel Parliament to consider federal reforms.

In mature federations, constitutional change is a shared endeavour, not a unilateral prerogative of the Union.

The United States (Article V): Two-thirds of state legislatures can require Congress to summon a Constitutional Convention for proposing amendments.

Canada (Section 46): Provincial legislature can initiate an amendment proposal, which the federal Parliament then considers under the appropriate amending formula.

Brazil (Article 60): Proposals for constitutional amendments can be initiated by more than half of the State Legislative Assemblies.

Switzerland (Articles 138 and 139): These provisions allow at least 100,000 citizens to initiate Constitutional amendments via the “Popular Initiative.”

Germany (Articles 76 and 79): The Länder participate in the federal legislative and constituent process through the Bundesrat, composed of delegates of state governments who vote as a bloc under binding instructions from their respective state Cabinets. Through Article 76, the Bundesrat may initiate constitutional amendments, and under Article 79(2), its approval by a two-thirds majority is mandatory. This confers on the Länder both a right of initiative and a decisive veto over constitutional amendments.

Insert a new clause (2D) in Article 368 providing that where a proposal to amend the Constitution is adopted by not less than two-thirds of the state Legislatures, and those states together represent not less than two-thirds of India’s population, Parliament shall be constitutionally obliged to take up the proposal for consideration.

8. Create a state-specific amendment procedure

A core principle of mature federalism is that constitutional amendments affecting particular subnational units should require the consent of those units alone, not a generalised nationwide process. Those who bear the consequences must have the decisive voice.

Comparative federal practices illustrate this principle:

Canada (Section 43): Amendments applicable to specific provinces require approval only by Parliament and the legislatures of the affected provinces.

Germany (Article 29): Territorial reorganisation of the Länder requires federal approval and, in many cases, referendums in the affected regions.

Switzerland (Article 53): Any change to the existence, status or territory of a Canton requires the consent of the affected electorates and the Cantons, along with federal approval – safeguarding them against unilateral federal action.

India too requires such a tailored mechanism.

Insert a new clause (2E) in Article 368 to establish a state-specific ratification procedure, under which only the Legislature of the affected state – or the Legislatures of the affected states, as the case may be – must concur in amendments that relate exclusively to their constitutional rights, territorial arrangements, or special protections. This procedure should apply, in particular, to amendments concerning:

Formation of new states and alteration of areas, boundaries, or names (Article 3);

Special provisions relating to certain states (Articles 371 to 371J); and

The First and Sixth Schedules.

Given the gravity of such decisions, ratification within each affected state should require a two-thirds majority of the total membership of its Legislative Assembly and, where applicable, its Legislative Council.

9. Introduce referendums as exceptional democratic safeguards

Across constitutional democracies, referendums are used in varying degrees to legitimise constitutional change. In some countries, they are mandatory for all amendments: Australia requires a double-majority of voters and states, Switzerland mandates approval by both voters and Cantons, and Ireland requires a popular referendum after parliamentary passage. Other countries, such as Spain, reserve referendums for “total” or core constitutional revisions, making them optional for ordinary amendments.

In the Constituent Assembly, Brajeshwar Prasad advocated referendums as a democratic expression of popular sovereignty, a check on parliamentary absolutism, a “conservative” safeguard against hasty change, and as a means to resolve Union-state deadlocks. B. R. Ambedkar, however, opposed the idea, citing India’s scale, logistical constraints, and a preference for representative democracy.

In the wake of the Emergency, the 44th Constitution Amendment Bill (1978) sought to amend Article 368 to mandate referendums for changes affecting core features – secularism, democracy, fundamental rights, free elections, and judicial independence – with at least 51 per cent electorate participation. The clause was adopted by the Lok Sabha but rejected by the Rajya Sabha, and was not enacted.

Referendums should not become a routine feature of constitutional amendment in India. Yet, in exceptional situations where representative institutions cannot remedy a democratic deficit, a carefully designed and narrowly confined referendum mechanism is justified. Two contexts are particularly compelling.

First, while the territorial integrity of existing states should be safeguarded by requiring legislative concurrence before new states are carved out, a referendum may provide a democratic pathway for sub-state minorities to seek separate statehood where the parent state withholds consent.

Second, Union Territories – especially those without legislatures – suffer from a serious democratic deficit. Their continued governance by the Union must remain an exception, not a permanent condition. Periodic referendums in Union Territories (except the National Capital Territory of Delhi) can provide a structured route either to full statehood or to integration with neighbouring states.

Insert a narrowly tailored referendum mechanism in the Constitution, providing that:

Where creation of a new state lacks the parent state legislature’s consent, the President may direct a referendum in the proposed territory. The proposal shall be deemed approved if not less than three-fourths of registered electors participate, and not less than two-thirds of votes cast support statehood.

Mandate decennial referendums in all Union Territories (other than the National Capital Territory of Delhi), offering two options: (a) full statehood; or (b) merger with a neighbouring state, subject to that state’s concurrence.

A proposal shall be carried only if not less than three-fourths of registered electors participate and not less than two-thirds of votes cast support the chosen option.

10. Introduce sunset clauses and periodic review

Sunset clauses ensure that experimental, transitional, or exceptional constitutional provisions – especially those conferring centralising or emergency powers – do not harden into permanent distortions. By mandating periodic review, they enable redesign in light of experience and prevent temporary arrangements from becoming entrenched structural imbalances.

The Constitution contains only two sunset clauses. Article 334 originally imposed a ten-year limit on SC/ST reservations and Anglo-Indian representation, subject to periodic extension; Anglo-Indian representation ceased after 2020, while SC/ST reservations continue until 2030. Article 334A, inserted by the 106th Constitution Amendment (2023), introduces a 15-year sunset for women’s reservation in legislatures.

Seventy-six years of constitutional practice indicate that several provisions warrant periodic reconsideration through sunset clauses. These include state reorganisation (Articles 3–4), the role and discretion of Governors (Articles 155–158, 163–164, 200–201), the GST Council (Article 279A), President’s Rule (Article 356), emergency-related fiscal controls (Articles 354 and 360), transitional provisions (Articles 372–392), and the Tenth Schedule on anti-defection.

Insert a new clause (2F) in Article 368 requiring that specified constitutional provisions – whether newly enacted or existing – be subject to mandatory review every 10 or 15 years, and cease to have effect unless re-enacted within that period.

India’s 76-year constitutional journey yields two abiding lessons. First, an over-flexible amendment procedure, when combined with concentrated political power, becomes a structural vulnerability; reliance on judicial valour alone to correct such excesses is neither sound in theory nor secure in practice. Second, conventions and pious appeals to restraint cannot compensate for gaps or ambiguities in the constitutional text. Reform must therefore begin with Article 368, for the coherence and integrity of the constitutional order ultimately rest upon the amending provision.

Comparative experience across mature federations offers clear guidance. The amending power must never rest with a single institution but should require layered consent – Parliament, state legislatures, and, where appropriate, the people. states should have a meaningful role in amendments affecting the federal balance, including the right to initiate proposals. Amendment thresholds should be strengthened through a double-majority rule of states and population, preferably with higher thresholds (two-thirds majority). The Constitution’s core identity should be explicitly protected. Additional safeguards such as transparency, public consultation, minimum deliberation periods, state-specific amendment procedures, and sunset clauses are also essential.

India must modernise Article 368 along these lines. Otherwise, past aberrations such as the 42nd Amendment and the recent 131st Amendment Bill will recur, steadily eroding fundamental rights, the federal balance, and democratic legitimacy.

K. Ashok Vardhan Shetty is a retired IAS officer of Tamil Nadu cadre, former Vice-Chancellor of the Indian Maritime University, Chennai and presently a member of the High-Level Committee on Union-State Relations constituted by the Government of Tamil Nadu.

© The Wire