Federalism — India's asymmetric bet

A federation answers a small set of structural questions before it answers anything else. Who are the constituent units? Are they equal to one another? What does the centre owe each of them, and what may it ask of each in return? The American Constitution answers these questions in one register. Australia answers them in much the same register. India answers them in a different register, and has done so from the founding moment. The Indian Union is a Union of States that has never treated all its members alike. The unequal treatment is not a series of compromises grafted onto a symmetric design — it is the design.

This essay surveys the Indian bet on asymmetric federalism. It looks at what the framers chose, what the asymmetric features have produced over seventy-five years, and what the symmetric alternative — the road taken by the United States and, in modified form, by Australia — would have meant. The argument is not that India should have copied either model. The argument is that the choice between the two ways of building a federation is not, as it is sometimes presented, a choice between principle and pragmatism. It is a choice between two different theories of what holds a large country together.

Two ways of building a federation

The United States built itself out of thirteen colonies that were already separate units of government, each with its own legislature, militia, and tax system. The Constitution of 1787 federated those units. The basic compact was equality: every State, regardless of population, sent two senators; every State was admitted on what came to be called an 'equal footing' with the original thirteen; every State enjoyed the same residual powers reserved by the Tenth Amendment. The asymmetries that crept in over the two centuries that followed — the Insular Cases, the District of Columbia, the territorial status of Puerto Rico — were treated as anomalies pending resolution rather than as features of the system. The structural template remained one of formally equal States.

Australia, federating in 1901 from six colonies, kept the template. Each State, large or small, sends twelve senators to the Commonwealth Parliament. The Northern Territory and the Australian Capital Territory have a constitutionally inferior position, but the six original States are formally equal. Section 99 of the Australian Constitution forbids the Commonwealth from giving 'preference to one State or any part thereof over another State'. The thought behind that clause is the thought behind the symmetric template generally: federation is a contract among equals, and the centre owes each unit the same treatment.

India did not federate equals. It built itself in 1950 out of three categories of preceding unit — the Part A States that had been British provinces, the Part B States that had been Princely States with their own pre-existing constitutional arrangements, and the Part C centrally administered territories that had no comparable history of self-governance.¹ The framers did not pretend that these three groups stood in the same relationship to the Union. They wrote different terms for different units into the original Constitution. Part XXI — 'Temporary, Transitional and Special Provisions' — was where most of the asymmetry was placed, and the words 'temporary' and 'transitional' have remained in the title even as some of those provisions have proved among the most permanent in the document.

What replaced Parts A, B and C in 1956 was a new dual structure of States and Union Territories, but the asymmetric impulse survived. Within the category of States, some were created with rights and protections that others did not have. The asymmetry was negotiated and renegotiated over the decades through constitutional amendments — the Article 371 family grew from a single clause for Maharashtra and Gujarat to a long list covering Nagaland, Assam, Manipur, Andhra Pradesh, Sikkim, Mizoram, Arunachal Pradesh, Goa, and Karnataka. Each was added in response to a specific pressure: a reorganisation that left a region feeling shortchanged, an insurgency that needed a political settlement, an accession that came with conditions. The result is a Constitution in which the question 'what does the Union owe a State?' cannot be answered without first asking 'which State?'

The framers' choice, in their own words

Ambedkar's much-cited line from the Constituent Assembly is that the framers had chosen the term 'Union' in preference to 'federation' precisely because the Indian Union was 'not the result of an agreement by the States to join in a Federation, and the Federation not being the result of an agreement, no State has the right to secede from it'. The line is usually quoted to make a point about the strength of the centre. It also makes, by implication, a point about asymmetry. A federation built by agreement among pre-existing equals must, as a matter of contractual logic, treat them equally. A Union built by central act, into which units are admitted on terms set by the centre, has no such constraint.

That is the structural reason — not the only reason, but a foundational one — why the Indian Constitution permits Article 2 admissions of new States 'on such terms and conditions as it [Parliament] thinks fit'. A constitutional bench in RC Poudyal v Union of India held in 1993 that those terms and conditions need not be the same as those that govern other States, provided they are compatible with the basic features of the Constitution.⁹ That holding ratified, after the fact, what the framers had built into the document: the absence of a constitutional symmetry requirement.

The same point is visible in the second chamber. The Rajya Sabha, often described as a 'Council of States' in homage to the federal model, departs from the federal template in two material ways. First, the States have unequal representation, allocated by population, not by sovereign equality as in the American Senate or the Australian Senate. Uttar Pradesh sends thirty-one members; Sikkim sends one. Second, since the Representation of the People (Amendment) Act 2003, a Rajya Sabha member need not even be a resident of the State they represent, a change the Court upheld in Kuldip Nayar v Union of India.⁴ Sabharwal CJ's reasoning was that residence becomes a constitutional requirement only if the Constitution expressly imposes one — which India's, unlike the American or Canadian, does not.

If the second chamber were the only departure, one could read it as an oddity. It is not. The Constitution allows Parliament under Article 3 to alter State boundaries, redraw the map, carve a State out of an existing State, or abolish a State, on the views — but not the consent — of the affected legislature. State of West Bengal v Union of India in 1963 confirmed that even on the assumption of State sovereignty, the Union's legislative power over property in a State is unrestricted.³ No federation built on the symmetric template would tolerate the Article 3 power. The American Constitution requires the consent of the affected State legislature for the formation of a new State out of its territory; the Indian Constitution requires only that the affected State be heard.

Article 370 and the Sixth Schedule — the original asymmetric architecture

The first asymmetric provision in the Constitution was Article 370, drafted into Part XXI to govern the special status of Jammu and Kashmir. Its origin is not constitutional principle but the specific facts of the State's accession in October 1947 — a Princely State with a Muslim majority and Hindu ruler, acceding to India under conditions of armed conflict, on the express understanding that only defence, foreign affairs, and communications would pass to the Union. The State retained the right to draft its own Constitution through its own Constituent Assembly. Article 370, as N Gopalaswami Ayyangar told the Constituent Assembly of India, was 'an interim arrangement' until the Kashmir Assembly met.⁵

The history of Article 370 between 1954 and the early years of this century is the history of a provision being read down through Presidential Orders and Supreme Court rulings until the autonomy it was meant to preserve had been substantially extended. Ninety-four of ninety-seven entries in the Union List, and 260 of the 395 articles of the Constitution, had been extended to Jammu and Kashmir by the mid-1990s. The Sadar-i-Riyasat — an elected Head of State — was replaced by a Governor appointed by the Centre. The Court, in Sampat Prakash v State of J&K (1970) and Mohd Maqbool Damnoo v State of J&K (1972), accepted the use of the President's Article 370 powers long after the Kashmir Constituent Assembly had been disbanded. AG Noorani's argument that all post-1956 amendments to the Article were constitutionally infirm was never formally answered by the Court. Article 370 was finally read down in 2019 by Presidential Order C.O. 272, a step whose constitutionality was upheld by the Supreme Court in 2023. What the constitutional position now is — and how much of the State's autonomy remains — is a different essay. The point here is that Article 370, even at its most attenuated, was an asymmetric arrangement that no symmetric federation could have written.

Less contested but more deeply embedded is the Sixth Schedule, which governs the autonomous district councils in tribal areas of Assam, Meghalaya, Tripura and Mizoram. The Schedule gives those councils legislative power over land use, forest management, social customs, marriage, inheritance, and the administration of customary justice. It draws, as Ambedkar acknowledged in the Assembly, on the American model for the governance of Native American tribes. The provision is asymmetric in a different way from Article 370: it does not give the State a special status against the Union; it gives a region within a State a special status against both. It carves out a third tier where the ordinary Article 246 distribution of powers does not apply.

The body of case law on the Sixth Schedule has been mixed. Samatha v State of Andhra Pradesh (1997), interpreting the Fifth Schedule, struck down mining leases to non-tribals in Scheduled Areas; subsequent benches in Balco Employees Union v Union of India (2002) drew the line back, holding that the Madhya Pradesh and Chhattisgarh land regimes were governed differently from the Andhra Pradesh regime that Samatha had addressed.⁷ The most consequential recent ruling on the Sixth Schedule, Pu Myllai Hlychho v State of Mizoram (2005), held that the Schedule is not a 'constitution within a constitution' — that the asymmetric architecture, however far it goes, sits inside a single legal order, not a plural one.⁸ The distinction matters. Asymmetry permitted by a single legal order is amendable, justiciable, and ultimately reversible by the institutions of the Union. Legal pluralism is none of those things.

The Article 371 family — asymmetry built case by case

Where Article 370 was an exceptional negotiated settlement and the Sixth Schedule a developmental design choice, the Article 371 family is something stranger: an accumulation of separate constitutional bargains, struck at different moments with different States for different reasons, all collected under a single article number. Article 371 itself, as originally drafted, made provision for development boards and equitable allocation in the new States of Maharashtra and Gujarat after the bifurcation of Bombay. Article 371A came in 1962 with Nagaland's statehood, providing that no Act of Parliament shall apply to Nagaland in respect of religious or social practices, customary law, civil and criminal justice involving customary law, or land and its resources, unless the State Assembly so decides. Article 371G did the same for Mizoram in 1986, with the omission of 'land and its resources' — a small textual difference that has produced a substantial dispute over Nagaland's claim to authority over petroleum and natural gas reserves.

Article 371D, inserted in 1973 after the agitation for separate Telangana statehood, gives the President power to require Andhra Pradesh to provide regional reservations in public employment and education for residents of different parts of the State. It is the most litigated of the asymmetric provisions — at least a dozen Supreme Court judgments turn on it. P Sambamurthy v State of Andhra Pradesh (1987) struck down clause 5 of the Article on basic-structure grounds, finding that allowing the State Government to override the Administrative Tribunal's awards on civil service appointments violated the rule of law. V Jagannadha Rao v State of Andhra Pradesh (2001) held that special rules framed under Article 309 could not displace the non-obstante protection of Article 371D. The pattern is one of the Court strengthening the asymmetric protections within Article 371D, even as those protections proved insufficient to head off the eventual Telangana statehood demand.

Article 371F, governing Sikkim's accession in 1975, goes further still. Its non-obstante clause has been read by the Court to permit derogation from the fundamental rights themselves. State of Sikkim v Surendra Prasad Sharma (1984) confirmed that pre-accession laws continue in force in Sikkim even where they conflict with Part III. RC Poudyal v Union of India (1994) upheld the reservation of twelve seats for Sikkimese of Bhutia-Lepcha origin and one seat for the Sangha out of a State Assembly of thirty-two — a clear departure from one-person-one-vote that would be unimaginable in a State elsewhere in the Union.⁹ The Court's reasoning was that Article 2 gives wide latitude on terms of admission, provided the new State's system of government is not 'fundamentally different' from what the Constitution envisages.

None of this is incidental. The Article 371 family covers ten States. It includes some of the largest in population (Andhra Pradesh, Karnataka) and some of the smallest (Sikkim, Arunachal Pradesh). It covers reorganisation grievances, ethnic-conflict settlements, accession terms, and developmental gradient corrections. There is no single principle that explains it, because it was never built on a single principle. It was built one bargain at a time, ratified in constitutional text, and left to the Court to police.

Fiscal federalism — where the asymmetry is quieter but real

The asymmetry is not only territorial. The fiscal architecture also reads differently from a symmetric federation, though the asymmetries here are less openly named. Three institutions carry the load.

The Finance Commission, appointed every five years under Article 280, recommends the distribution of central tax revenues between the Union and the States and among the States. Its formula has, since the Tenth Commission's reforms ratified by the Eightieth Amendment in 2000, used population, area, income distance, and demographic and forest indicators to weight horizontal allocation.¹¹ The formula is, by design, asymmetric in effect: poorer States receive more per capita than richer States. This is consistent with most fiscal federations, but the Indian model layers it with a separate 'Special Category' status — historically extended to the eight northeastern States plus Jammu and Kashmir, Himachal Pradesh, and Uttarakhand — under which Plan transfers came in a 90:10 grant-loan ratio rather than the 30:70 applicable to the major States. The Special Category mechanism has been substantially diluted since the Fourteenth Finance Commission, but its imprint on the federal architecture is real, and the political demand for the status (most recently from Andhra Pradesh after the Telangana bifurcation, and earlier from Bihar) shows that the asymmetry is understood by the units themselves to matter.

The GST Council, created by the One Hundred and First Amendment in 2016 and given constitutional form in Article 279A, is asymmetric in a different and counter-intuitive sense.¹⁰ Its voting structure gives the Centre a one-third weight and the States collectively a two-thirds weight; any decision requires three-fourths of the weighted votes cast. The arithmetic gives the Centre a veto without giving any combination of States a veto without it. In a symmetric federation that took its second chamber seriously, this kind of intergovernmental body would have been unnecessary; tax legislation would have been worked out in the upper house. India needs the GST Council precisely because the Rajya Sabha is not the federal chamber the comparative literature assumes a Council of States would be.

The Inter-State Council, contemplated in Article 263 and finally constituted in 1990 on a Sarkaria Commission recommendation, is the third leg.¹² Its meeting record is intermittent, its outputs largely advisory, and its political weight depends on the prime minister of the day. But its very existence, as a forum where Centre and States bargain outside the legislative chamber, is itself an admission that the formal federal architecture in India does not contain enough deliberative space for the kind of disputes a federal system regularly throws up. A symmetric federation either resolves those disputes through its second chamber or leaves them to the courts. India does both, and adds an Inter-State Council on top.

The friction points — Bommai, the Governor, and Article 356

The asymmetric design produces friction in places a symmetric federation would not. Three are worth naming.

The first is Article 356 — the President's Rule provision, which permits the dismissal of an elected State government on the report of a Governor that the State cannot be carried on in accordance with the Constitution. S.R. Bommai v Union of India (1994) is the leading authority and is usually celebrated as the case that brought Article 356 under judicial review.² A close reading of the nine-judge bench's six opinions tells a more complicated story. Federalism was discussed in five of the six opinions, but it was not the doctrinal pivot of any of them. The judicial review of the President's proclamation was rooted by the bench in the principle of democracy and in the separation of powers, not in federalism as such. Bommai achieved a great deal — the dismissals it considered were largely upheld, but the Court set out a framework that has made later misuse of Article 356 politically and legally costlier. What it did not do was to ground that framework in the federal architecture. The federal feature of the Constitution that Article 356 most directly threatens — the elected State government as a separate locus of authority — was named in the case but did not do the doctrinal work.

The second is the Governor. The constitutional design places a centrally appointed officer at the head of every State, holding office during the pleasure of the President, with discretionary powers in matters such as the appointment of a Chief Minister where no party has a majority, the reservation of bills for Presidential assent, and the framing of recommendations under Article 356. The Sarkaria and Punchhi Commissions both recommended procedural constraints on the Governor's discretion — fixed tenure, consultation with the Chief Minister on appointment, removal only for cause — but no constitutional amendment has implemented them. The Governor remains, in practice, a federal pressure point: the office through which the Centre's preferences enter State politics. This is not a feature a symmetric federation would have written, and it sits awkwardly even within the asymmetric one.

The third is the money-bill route, which has, in the post-2014 period, allowed the Union to pass legislation on subjects of substantive policy through the Lok Sabha alone, bypassing the Rajya Sabha — itself, as we have noted, a chamber whose federal credentials are already weak. The Aadhaar Act and several Finance Acts have been challenged on this ground. The asymmetric Rajya Sabha, designed without sovereign equality of States, becomes an even thinner federal check when its scrutiny can be circumvented procedurally. The federal scheme survives the manoeuvre because it was never built on the assumption that the Rajya Sabha would be the principal federal chamber. A federation that had built the federal check into its second chamber — as the United States and Australia have — would have nowhere to retreat to.

The counter-argument — what asymmetry buys

The strongest defence of India's asymmetric bet is also the most empirical. It is that the bet has produced cohesion. The federation has held — through linguistic reorganisation, through three wars, through the Emergency, through three decades of regional-party government in major States, through insurgencies in the Northeast and in Punjab and in Kashmir, through the dismantling of the dominant-party system in 1989 and the coalition era that followed. No major unit of the Union has seceded. The accommodative capacity of the asymmetric design is, on this account, the reason.

The case is not merely India's. Stepan, Linz and Yadav put it in comparative terms.⁶ Multinational federations that try to operate on the symmetric model have, with some regularity, fractured along the lines they refused to recognise. The Soviet Union, Yugoslavia, Czechoslovakia — all symmetric in their formal architecture, all unable to absorb the centrifugal pressures their formal symmetry suppressed. Spain's accommodation of the Catalan and Basque demands has been asymmetric because symmetric federalism could not contain those demands. Belgium has gone further still. Canada's struggle with Quebec is, on one reading, the struggle of a federation built on the symmetric template trying to accommodate an asymmetric reality without quite admitting that it is doing so.

Against this background, India's choice looks less idiosyncratic. The Sixth Schedule did not create the cultural distinctness of the Khasi or the Garo or the Mizo; it recognised it, and gave it a constitutional home. Article 371A did not invent the Naga claim to differentiated sovereignty; it absorbed enough of it into the constitutional order to convert an insurgency into a State within the Union. Article 371F did not establish Sikkim's pre-1975 separateness; it carried that separateness forward in a form the Supreme Court was prepared to enforce. In each case, the asymmetric device did the work of recognition that a symmetric federation either could not do or would not do. The cost of doing it was heterogeneity. The benefit was that the federation held.

This is, I think, the right place to put the strongest version of the case. The asymmetric design is not a series of grudging concessions. It is a positive answer to the question of how a country with India's diversity holds together at all. A symmetric design would have been more elegant and less litigated. It would also have had to find some other way of carrying the kinds of recognition that Article 370 and the Sixth Schedule and the Article 371 family carry. There is no reason to believe such a way exists.

Where the bet rests

Three observations follow, none of them the kind of summary that ties a debate off.

First, the asymmetric architecture has been, with some local exceptions, sustained rather than dismantled. Of the major asymmetric provisions in the original Constitution, only Article 370 has been substantially read down, and even that was a process of seventy years rather than a single moment. The Sixth Schedule has been preserved and in places extended through the Bodo Accord and other settlements. Article 371D has been strengthened by the Court even as Telangana itself was reorganised out of Andhra Pradesh. The Sarkaria and Punchhi Commissions both endorsed the broad framework.¹² The asymmetric bet is being honoured, not regretted.

Second, the asymmetry has migrated from the territorial provisions to the fiscal ones. The Special Category status, the GST Council's voting structure, the Finance Commission's horizontal weighting formula — these are the live sites of intergovernmental bargaining in a way that Article 371 itself, with the partial exception of the Nagaland petroleum dispute, no longer is. A federation that began its asymmetric career with provisions about land, language, custom, and accession terms is now arguing about transfer pools and weighted votes. The terrain has shifted; the structural grammar has not.

Third, the unfinished settlements are visible. Jammu and Kashmir's constitutional position after 2019 remains contested. The Naga claim to sovereignty over land and its resources, including hydrocarbons, has not been resolved. Demands for separate statehood from Bodoland, Gorkhaland, Vidarbha, and the Saurashtra region recur. New asymmetric devices — for Delhi after the 69th Amendment and the GNCTD (Amendment) Act 2023, for Ladakh as a Union Territory without a legislature — are being added to the pile. The asymmetric bet is open-ended in a way the symmetric bet would not be.

The Indian framers wagered that a federation can hold together by treating its members differently. The wager was not principled in the way the American wager on equal States was principled; it was instead empirical, made under the pressure of partition and accession and reorganisation, and revised one constitutional amendment at a time as new pressures appeared. Seventy-five years on, the bet has not failed. It has not been won either, in the sense of producing a settled constitutional architecture the country can stop arguing about. What it has produced is a federation in which the question of who counts as a full member, on what terms, and against which guarantees, is part of the ongoing political settlement rather than a question the Constitution closed in 1950. That is not what the United States chose. It is not what Australia chose. It is what India chose, and it is what India has had to keep choosing.

Get the week's biggest decisions, distilled.

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