Indian Constitutional Law New Challenges Notes | LL.M. Semester 1 | Mumbai University | munotes

The same logic runs the other way. An article that is taken away is not deleted from the numbering. Article 31, which was the right to property, was repealed in 1978, and the official text still prints "31." followed by the heading in square brackets and the words "Rep. by". Nothing else moves. So the gaps in the sequence are themselves information.

Four printing conventions that will catch you out

These are in the official text, they are not errors, and a student reading the bare Act needs all four.

1. An amended provision is printed inside its footnote marker. The right to education appears as:

2[21A. Right to education. The State shall provide free and compulsory education to all children of the age of six to fourteen years in such manner as the State may, by law, determine.]

In the printed book a long dash separates the marginal heading "Right to education" from the words of the article; this book uses a full stop there instead, and does so throughout. The little "2" and the square brackets are not part of the article. They mean "footnote 2 at the bottom of this page tells you which amendment put this here", and the bracket shows how much of the text the amendment inserted. Read the footnote: for article 21A it records the Constitution (Eighty-sixth Amendment) Act, 2002. Those footnotes are the most reliable amendment history you will find anywhere.

2. Some articles are printed with a hyphen. Goa's special provision is printed 371-I, not 371I. So is 243-O in the arrangement of articles at the front of the book. The reason is typographical: a capital I next to a digit reads as a 1, and a capital O reads as a 0, so the Legislative Department separates them. If you search a downloaded copy for "371I" you will find nothing at all, and you may conclude the article does not exist. It does.

3. For the same reason, some letters are skipped. Part IXB, on co-operative societies, runs 243ZH and then 243ZJ. There is no article 243ZI. That is a deliberate gap in the Constitution's own numbering, not a missing page.

4. There are two texts of every article in the official edition. It is published in diglot form, Hindi and English on facing pages. Article 348 makes the authoritative text of the Constitution the English one, and article 394A provides for the authoritative Hindi translation. For examination purposes you work from the English text.

The Preamble, and what it is for

The Preamble is the opening declaration: sovereign, socialist, secular, democratic republic, and the four objectives of justice, liberty, equality and fraternity. The words "socialist" and "secular", and the word "integrity", were inserted by the Forty-second Amendment in 1976.

Two things about it are examinable. It is not a source of power, so no authority can be exercised on the strength of the Preamble alone. But it is part of the Constitution and it is used to resolve an ambiguity, which is why the objectives it names keep appearing in judgments about the basic structure. Both propositions are developed in [The Basic Structure].

What this book calls things

A provision is any article or clause. An amendment is an Act passed under article 368, cited by its number and year, as the Constitution (Forty-second Amendment) Act, 1976. A Bill is not an amendment: it becomes one when it is passed by the required majorities and receives assent, and until then it changes nothing. That distinction matters more than it sounds, because two Bills to amend the Constitution were before Parliament while this book was being written and neither has become law.

The state of the Constitution as this book is written

One hundred and six amendments have been enacted. The current official text is headed "As on 1st May, 2026" and its own preface records that it incorporates all amendments up to the Constitution (One Hundred and Sixth Amendment) Act, 2023.

Two later Bills are often written about as though they were amendments, and they are not. The Constitution (One Hundred and Thirtieth Amendment) Bill, 2025 was introduced in the Lok Sabha on 20 August 2025 and is pending. The Constitution (One Hundred and Thirty-first Amendment) Bill, 2026 was introduced on 16 April 2026 and was defeated in the Lok Sabha the following day, 17 April 2026, having failed to obtain the special majority article 368 requires.

So the correct answer to "how many times has the Constitution been amended" is one hundred and six, and a student who writes one hundred and thirty has confused a Bill with an Act. The consequences of that particular Bill's defeat are the subject of [The 106th Amendment, and Why It Has Not Started].

Where the law in this book comes from

Every article quoted in this book is quoted from the official Legislative Department text described above. Every case is one whose judgment has been read: the citation, the facts and what was actually held. Where a proposition is contested, this book says so rather than choosing a side and presenting it as settled, because at LL.M. level the contest is usually the question.

A worked example

A student is given this line from a downloaded copy of the Constitution and asked what it tells them:

1[371-I. Special provision with respect to the State of Goa. Notwithstanding anything in this Constitution, the Legislative Assembly of the State of Goa shall consist of not less than thirty members.]

What is the "1["? The marker of footnote 1 on that page, which will say which amendment inserted the article. It is not part of the text, and the closing bracket shows how far the insertion runs.

Why the hyphen in 371-I? Because a capital I beside digits reads as a 1. The article is article 371I, and searching the file for "371I" without the hyphen returns nothing.

What does the letter tell you? That the article was added after article 371 and before 371J, by amendment. Goa became a State in 1987, which dates it.

Where does it sit? Part XXI, the temporary, transitional and special provisions, which is also where article 370 sat. That placement is an argument in itself, and it was used in the article 370 litigation.

What is the "notwithstanding" doing? Overriding anything in the Constitution that would otherwise fix a different minimum size for a Legislative Assembly.

Quick revision

• The Constitution is a Preamble, 395 articles in twenty-five Parts, and twelve Schedules.
• A letter after an article number means the provision was inserted later; the footnote says by which amendment.
• A gap in the numbering means a provision was repealed and the number was not reused.
• Articles 371-I and 243-O are printed with hyphens, and there is no article 243ZI.
• One hundred and six amendments are in force. The 130th and 131st are Bills, and the 131st was defeated on 17 April 2026.

Test yourself

1. What do the square brackets and the small numeral round an article in the official text mean, and what is the one thing a student should always do when they see them?

2. Why is there no article 243ZI, and what does that tell you about the Constitution's numbering?

3. How many amendments to the Constitution are in force, and why is "one hundred and thirty" a wrong answer in 2026?

4. A provision is numbered 51A. Without knowing anything else about it, what three things can you say?

An independent judiciary. Somebody has to decide disputes about the line, and it cannot be either of the parties to them. In a federation the court that does this is exercising a function no other kind of State needs.

Dual polity, and direct authority over the citizen. Both governments act directly on the same citizen, each in its own field. This is what distinguishes a federation from a confederation, in which the centre acts on the units and only the units act on the individual.

Federation, confederation and a unitary State

Three arrangements, distinguished by where authority originates.

In a unitary State all authority is in the centre. Regional bodies may exist and may be given wide powers, but they hold those powers as a delegation, and the centre may take them back by ordinary legislation. The United Kingdom before devolution is the standard example, and it remains formally unitary even after it.

In a confederation the units are the source of authority and the centre exercises only what they have agreed to give it. The centre generally acts on the units rather than on individuals, and a unit may in principle leave. The United States under the Articles of Confederation is the standard example, and it did not work.

In a federation the constitution is the source of both governments' authority, and neither is the delegate of the other. That is the crucial idea: in a federation the States do not hold their powers by permission of the Centre, and the Centre does not hold its powers by permission of the States. Both hold them by grant of the constitution.

How federations come into being, and why it shows

There are two routes, and the route leaves a permanent mark.

By coming together. Independent States agree to unite, surrendering some of their powers to a new centre they create. The United States, Switzerland and Australia are of this kind. Because the units existed first, they usually keep the residue: whatever the constitution did not give to the centre remains with them. They also usually get an equal or weighted voice in a second chamber, and a strong say in amendment.

By holding together. A single State, already unitary or nearly so, devolves power on regions in order to hold a diverse country together. India, Canada, Spain and Belgium belong here in varying degrees. Because the whole existed first, the centre tends to keep the residue, the units are creatures of the constitution rather than parties to it, and the arrangement is generally asymmetric, with different regions getting different terms.

This distinction explains most of what a student will otherwise have to memorise about India. Almost every feature that is said to make India less than fully federal follows from its having been created by holding together rather than by coming together.

Cooperative federalism

The classical model imagines two governments operating in watertight compartments. No modern federation works that way, because the problems do not divide neatly: public health, transport, education, environment and disaster relief all need both levels.

Cooperative federalism describes the arrangements by which the two levels work together instead of apart: joint bodies, shared finance, conditional grants, consultation, and legislation by one level that the other administers. It is not a departure from federalism but a description of how federations behave once government becomes large.

Its shadow is competitive federalism, in which regions compete with each other for investment, and the centre encourages that competition by ranking and rewarding them. The two are usually discussed together, and India's official language has used both. Both return in [Cooperative and Competitive Federalism].

What federalism is not

Three confusions worth clearing up now, because they cost marks.

Federalism is not decentralisation. A unitary State can decentralise heavily, and a federation can be highly centralised. The question is not how much power the regions exercise but whether they hold it by right or by permission.

Federalism is not the same as having a second chamber. Most federations have one, and it usually represents the units, but a second chamber can exist without federalism and a federation can exist without a strong one.

Federalism is not a guarantee of liberty. It divides power, which tends to check it, and that is a real argument for it. But a federation can protect an oppressive regional government against a liberal centre just as easily as the reverse, and the history of federalism includes both.

A worked example

State X has a written constitution which divides subjects between a central government and four regional governments. The central legislature may, by a two-thirds vote of its own members, transfer any subject from a regional list to the central list. Regional governments administer their own subjects and raise their own taxes, and a supreme court decides disputes about the division.

Is State X a federation?

What points towards yes. A written constitution; a distribution of powers; dual administration; separate revenue; and an independent court to police the line. Four of the classical tests are met.

What decides it is the amendment power. The central legislature may move a subject out of the regional list by its own vote alone. The regions have no say. That means the division of powers is held at the centre's pleasure, and the regional governments exercise their powers by permission rather than by right.

Conclusion. State X is not a federation on the classical tests, because the constitution is not rigid in the respect that matters: the article dividing powers is amendable by one of the two parties to it acting alone. It is a decentralised unitary State with federal machinery.

Compare India. Article 368 requires, for a change to the distribution of legislative powers, not only a special majority in Parliament but ratification by the legislatures of half the States. That is the safeguard State X lacks, and it is why the Indian answer is different.

Quick revision

• A federation divides power by a constitution neither level can alter alone; the units are not delegates of the centre, nor the centre of the units.
• The classical tests are a written and supreme and rigid constitution, a distribution of powers, an independent judiciary, and a dual polity acting directly on the citizen.
• A confederation puts authority in the units; a unitary State puts it in the centre; a federation puts it in the constitution.
• Coming-together federations leave the residue with the units; holding-together federations leave it with the centre. India is the second kind.
• The real question is never federal or not, but where the line is, who drew it, and who may move it.

Test yourself

1. State Wheare's federal principle and list the classical tests of a federation.

2. Distinguish a federation from a confederation and from a unitary State by reference to where authority originates.

3. What is the difference between a coming-together and a holding-together federation, and what does the difference predict about the residuary power?

4. "Federalism is decentralisation." Discuss.

A dual polity. A Government of India and a government for each State, each with its own legislature, executive and civil service, each acting directly on the same citizen.

Distribution of legislative powers. Article 246 with the Seventh Schedule allocates subjects across three Lists. This is the core of the arrangement and is taken up in [The Legislative Lists].

Distribution of executive power. Articles 73 and 162 make the executive power of each level coextensive with its legislative power, so the division of subjects carries the administration with it.

Distribution of financial resources. Separate heads of taxation, a divisible pool, and a Finance Commission to recommend the sharing, dealt with in [Fiscal Federalism].

An independent judiciary with the power to decide the line. Article 131 gives the Supreme Court exclusive original jurisdiction over a dispute between the Union and a State, and the courts decide questions of legislative competence.

A bicameral Parliament with a chamber representing the States. The Council of States under article 80.

The features no true federation has

A single citizenship. There is no citizenship of a State. Compare the United States, where a person is a citizen of the Union and of a State.

An integrated judiciary. One hierarchy of courts administers both Union and State law, with the Supreme Court at its head. There is no separate system of State courts.

All-India services. Officers recruited centrally serve the States and hold key posts in them, while remaining members of a service the Centre controls.

Parliament may redraw the map. Under article 3 Parliament may form a new State, alter the area, boundaries or name of an existing one, and it needs only to refer the Bill to the State legislature for its views, which do not bind. No true federation permits the centre to abolish a unit. This is [Making and Remaking States].

A single integrated audit and election machinery. One Comptroller and Auditor General, one Election Commission.

The Governor. Appointed by the President, holding office during pleasure, and empowered to reserve a State Bill for the President's consideration. See [The Governor].

Emergency provisions. During a Proclamation under article 352 Parliament may legislate on any State subject and the Union may direct a State on any matter; under article 356 the State government may be dismissed altogether; under article 360 the Centre may direct the States on financial matters. See [Emergency and the Federal Balance].

Parliament may legislate on a State subject even in normal times, on a resolution of the Council of States under article 249, with the consent of two or more States under article 252, or to implement a treaty under article 253.

A strong Union List, and the residue with the Centre under article 248, the reverse of the American and Australian arrangement.

The judicial position, and how it moved

Two decisions bracket the argument, and an answer that gives only one of them is out of date.

Facts. State of West Bengal v. Union of India, AIR 1963 SC 1241, (1964) 1 SCR 371, decided 21 December 1962. Under the Coal Bearing Areas (Acquisition and Development) Act 1957 the Union proposed to acquire coal bearing land owned by the State of West Bengal. The State sued, arguing that the Act did not reach land owned by a State, and that if it did, Parliament lacked competence, because the States are sovereign within their sphere and their property cannot be taken by a Union law.

Held. By a majority, with Subba Rao J. dissenting, the Act applied to State land and was within Parliament's competence under Entry 42 of the Concurrent List read with Entries 52 and 54 of the Union List. On federalism the Court said in terms that the Constitution of India is not truly federal in character. The basis of distribution is that only powers concerned with the regulation of local problems are vested in the States, and the residue, especially what maintains the economic, industrial and commercial unity of the country, is left to the Union. Full sovereignty does not vest in the States. Parliament, competent under article 3 to destroy a State altogether, cannot be held incompetent to acquire a State's property.

Why it matters. It is the leading early authority for the proposition that India is not a true federation, and it puts the argument at its strongest: a unit that can be abolished cannot be sovereign.

Facts. S.R. Bommai v. Union of India, AIR 1994 SC 1918, (1994) 3 SCC 1, decided 11 March 1994 by nine judges. A group of appeals about Proclamations under article 356 dismissing the governments of Karnataka, Nagaland, Meghalaya, Madhya Pradesh, Rajasthan and Himachal Pradesh.

Held. Among twelve summarised conclusions, the Court held that the Constitution of India has created a federation but with a bias in favour of the Centre, and that within the sphere allotted to them the States are supreme. It also held that a Proclamation under article 356 is not immune from judicial review, that the satisfaction must rest on relevant material, and that the court may require the Union to produce that material. Two of the six Proclamations were held unconstitutional.

Why it matters. It converts federalism from a description into a constraint. If a Proclamation dissolving a State government can be struck down, the States' sphere is protected by a court and not merely by the Centre's forbearance. This is also the case that makes federalism part of the basic structure, so that it limits even the amending power, as [The Basic Structure] explains.

The vocabulary, and which word to use

Several descriptions are in circulation and an examiner will recognise all of them.

Quasi-federal is Wheare's, and it is the most commonly quoted: a unitary State with subsidiary federal features rather than a federal State with subsidiary unitary features. It is a fair description of the text of 1950 and a poor description of the practice since 1994.

Federal with a strong centre is the description the Supreme Court itself has settled on, and it is the safest phrase to use.

Co-operative federalism describes the working relationship rather than the formal structure.

A bargaining federalism is the political scientist's phrase for what the arrangement has become, in which the real division is settled by negotiation over money and schemes rather than by litigation over Lists.

A good answer does not simply pick one. It says that the formal position is a Union of States with a pronounced tilt to the Centre, that the tilt was deliberate and is explained by partition and by the holding-together origin, and that the judicial and political practice has since moved the working position closer to a genuine federation without changing a word of article 1.

"Trust and faith between the Centre and the States"

MU's own question adds this phrase, and it is not decoration: it points at the parts of the arrangement that are not legal rules at all.

Several of the Constitution's federal provisions only work if both sides act in good faith. Article 3 requires the views of the State legislature to be sought but not followed, so the safeguard is a habit rather than a rule. Article 263 permits an Inter-State Council but does not require the Centre to use it. Article 356 turns on a satisfaction, and before Bommai the honesty of that satisfaction was the only protection there was. The Governor's discretion, the reservation of Bills for the President, and the sharing of tax revenue all depend on restraint by the stronger party.

The doctrine that gives this a legal name is federal comity: the proposition that the Union and the States must each exercise their powers with regard to the legitimate interests of the other, and neither should use a power it undoubtedly has in a way calculated to defeat the other's functioning. It is developed in [Administrative Relations and Federal Comity].

So an answer to MU's question should say that the relationship of trust and faith is not merely a sentiment. It is the substitute for the guarantees a coming-together federation would have written down, and where the trust has broken down, the courts have had to supply a rule instead: Bommai for article 356, and the recent litigation about a Governor's delay in assenting to Bills, which is [The Governor].

A worked example

Answer this stem, which MU has set in six of the eleven printed papers: "Explain fully, the principle of Federalism embodied in the constitution of India and relationship of trust and faith between Centre and States."

Open with the text, not with a conclusion. Article 1 says "Union of States", and Dr Ambedkar gave two reasons: the Union is not the result of an agreement, so no State may secede; and the States are not creatures of the Centre. Both halves matter, and quoting them settles the frame.

Give the federal features as a list, each with its article: written, supreme and rigid Constitution with article 368's ratification proviso; dual polity; article 246 and the Seventh Schedule; articles 73 and 162; the Finance Commission; article 131; the Council of States.

Give the unitary features as a list: single citizenship; integrated judiciary; all-India services; article 3 with article 4; the Governor; articles 352, 356 and 360; articles 249, 250, 252 and 253; article 248 and Entry 97.

Then the two cases, in order. West Bengal v. Union of India (1962): not truly federal, and a Parliament that may destroy a State cannot be incompetent to acquire its property. Bommai (1994): a federation with a bias in favour of the Centre, the States supreme within their sphere, and the article 356 power reviewable.

Answer the second half of the stem, which most candidates ignore. "Trust and faith" is federal comity. Say that article 3's consultation, article 263's Council and article 356's satisfaction all depend on restraint the text does not compel, and that where the restraint failed the courts supplied a rule: Bommai for article 356, and the assent litigation for article 200.

Conclude. A federation with a strong centre, whose federal character is now a basic feature, so the tilt is a design choice and not a licence.

Quick revision

• Article 1 says "Union of States" on purpose: no unit created the Union and none may leave it, but the States are not creatures of the Centre either.
• The federal features are the written, supreme, rigid Constitution, the dual polity, the three Lists, the separate executive and financial spheres, and an independent judiciary.
• The unitary features are single citizenship, an integrated judiciary, all-India services, article 3, the Governor, the emergency provisions, and the residue with the Centre.
• West Bengal (1962) said not truly federal. Bommai (1994) said a federation with a bias in favour of the Centre, the States supreme within their sphere. Give both.
• Federalism is now a basic feature, so the tilt towards the Centre is a design choice, not a licence.

Test yourself

1. Why does article 1 say "Union of States" rather than "federation"? Give the Constituent Assembly's two reasons.

2. List five features of the Constitution that no true federation has, with the article for each.

3. Contrast the view of federalism in State of West Bengal v. Union of India with that in S.R. Bommai, and explain what changed between them.

4. What is federal comity, and which provisions of the Constitution depend on it?

Two Union territories are special. Delhi has, since the Sixty-ninth Amendment, a legislative assembly and a council of ministers under article 239AA, with three State List entries expressly withheld from it: public order, police and land. The long litigation about who controls the services in Delhi turns on the wording of that article. Jammu and Kashmir became a Union territory with a legislature in 2019, and Ladakh a Union territory without one, which is dealt with in [Article 370].

A Union territory can become a State and a State can be converted into a Union territory, and both have happened. That is a consequence of article 3, and it is the sharpest illustration of how much weaker the Indian units are than the units of a coming-together federation.

The First Schedule, and how it is amended

The First Schedule lists the States and their territories, and then the Union territories. It is the operative list: what makes a place a State is that it appears there.

The Schedule is changed constantly. Every reorganisation, every new State, and every boundary change alters it. Article 4 is what makes that easy. A law made under article 2 or article 3 may contain provisions amending the First Schedule and the Fourth Schedule, which is the allocation of seats in the Council of States, and any supplemental, incidental and consequential provisions Parliament thinks necessary. Then comes the sentence that matters:

No such law as aforesaid shall be deemed to be an amendment of this Constitution for the purposes of article 368.

So a Bill that creates a new State, redraws a boundary and reallocates Rajya Sabha seats passes by an ordinary majority, like any other Act, and does not need the special majority or the ratification by half the State legislatures that article 368 would otherwise require for a change to the federal structure.

That is a remarkable provision, and it is the single strongest argument that India is not a true federation. In the United States, article IV section 3 forbids the formation of a State from the territory of another without the consent of the legislature concerned. Under article 3 the Indian States are consulted and nothing more.

Acquisition and cession of territory

Article 1(3)(c) covers acquisition. Territory acquired by India becomes part of the territory of India, and Parliament may by law admit it or establish it as a State under article 2, or absorb it into an existing State or Union territory under article 3. India acquired territory in this way more than once after 1950, including the former French and Portuguese possessions.

Cession, giving territory away, is the harder question, because article 3 speaks of diminishing the area of a State but says nothing about transferring Indian territory to a foreign country.

The settled position is that ceding Indian territory to another country requires an amendment of the Constitution under article 368, not merely a law under article 3. The reason is that the territory of India is defined by article 1 and the First Schedule, and reducing it alters the Constitution itself rather than merely rearranging the units within it. That position was reached when the question arose over an agreement with Pakistan about the Berubari Union, and it produced the Constitution (Ninth Amendment) Act, 1960, which the official text records as the instrument by which the transfer was carried through. The same route was followed for the exchange of enclaves with Bangladesh by the Constitution (One Hundredth Amendment) Act, 2015, whose details the official text sets out in Appendix I.

Distinguish settling a boundary that was always uncertain from ceding territory that was undoubtedly India's. The first can be done by executive action, because it determines what the border already was; the second cannot. That distinction is what most examination questions on this point are really testing.

Why Part I matters for the rest of the module

Three consequences carry forward.

The units are not indestructible. Parliament may abolish a State by merging it, may divide it, may rename it, and may reduce it to a Union territory. Every later argument about how federal India is runs into this fact, and it is why the Court in State of West Bengal v Union of India could say that a Parliament competent to destroy a State cannot be incompetent to acquire its property.

The map is a matter of ordinary legislation. Because of article 4, changing the composition of the federation does not engage the amendment procedure that protects the rest of the federal structure.

The territory can grow. The Constitution anticipates acquisition and provides for it without amendment, which is a unitary rather than a federal feature.

A worked example

Parliament passes an Act which (a) transfers a small border district from State P to State Q, (b) renames State Q, (c) reallocates two Rajya Sabha seats between them, and (d) cedes a strip of territory to a neighbouring country under a boundary agreement. The Bill was introduced on the President's recommendation and referred to both State legislatures, which objected. It was passed by a simple majority.

(a), (b) and (c) are valid. Transferring territory between States, altering a name, and amending the First and Fourth Schedules are all within article 3 read with article 4. The State legislatures' objections do not bind, and article 4 says such a law is not an amendment for the purposes of article 368, so a simple majority suffices.

(d) is not. Ceding Indian territory to a foreign country reduces the territory of India as defined by article 1 and the First Schedule. That alters the Constitution itself and requires an amendment under article 368, as the Ninth Amendment 1960 and the Hundredth Amendment 2015 both show.

The distinction to draw. Had the strip been land whose ownership was genuinely uncertain, settling the boundary would have determined what the border always was, and could have been done by executive action. Ceding land that was undoubtedly India's is different.

What would fix it. Sever the cession from the rest, and carry it through by a separate amendment under article 368.

Quick revision

• Article 1: India is a Union of States; the territory is the States, the Union territories, and anything acquired.
• A State is a unit of the federation; a Union territory is administered by the President under article 239, with Delhi and Jammu and Kashmir as special cases.
• Article 2 admits or establishes new States; article 3 reshapes existing ones.
• Article 4 lets both be done by ordinary majority and says such a law is not an amendment for the purposes of article 368.
• Acquiring territory needs no amendment. Ceding territory does, and the Ninth Amendment 1960 and the Hundredth Amendment 2015 are the examples.

Test yourself

1. Set out the three limbs of the definition of the territory of India in article 1(3).

2. What does article 4 provide, and why is it the strongest single argument that India is not a true federation?

3. Distinguish a State from a Union territory, and name the two Union territories that are special cases.

4. Does India need a constitutional amendment to acquire territory? To cede it? Explain the difference.

The reference to the State legislature is the only provision in article 3 that looks like a federal safeguard, and three features make it very weak.

First, the State legislature expresses views; it does not consent. The proviso says the Bill shall be referred for the legislature "to express its views thereon within such period as may be specified". Nothing says the views must be accepted or even considered in any particular way.

Second, the period is fixed by the President, and if the legislature says nothing within it, Parliament may proceed.

Third, and least obviously, Parliament may change the proposal after the reference without referring it again.

Facts. Babulal Parate v. State of Bombay, AIR 1960 SC 51, (1960) 1 SCR 605, decided 28 August 1959. A Bill to reorganise the State of Bombay was referred to the State legislature under the proviso to article 3. The proposal referred was for three units. Parliament then adopted a materially different scheme, the bifurcation into Maharashtra and Gujarat. The petitioner said Parliament could not enact a scheme different from the one on which the State had been consulted without a fresh Bill and a fresh reference.

Held. No fresh reference is required. What the President must refer is the proposal contained in the Bill, and on a true construction the proviso does not contemplate that a subsequent modification by Parliament requires a fresh Bill or a fresh reference. The word "State" in article 3 refers back to article 1 and the First Schedule, and there is no warrant for importing doctrines of democratic theory prevalent in other countries, nor for giving the word an extended meaning. The requirements of article IV section 3 of the American Constitution are materially different.

Why it matters. It settles how thin the safeguard is. A State may be consulted on one scheme and reorganised under another, and the courts will not intervene.

Why the safeguard was made so weak

This looks like a drafting failure until you remember what the Constituent Assembly was doing.

In 1950 the map was provisional. The Constitution inherited the Provinces of British India, several hundred princely States which had acceded in 1947 and 1948, and territories still under foreign control. The First Schedule then divided the units into Part A, Part B, Part C and Part D States, which were governed differently. Nobody expected that arrangement to last, and the Assembly deliberately left Parliament the power to rationalise it without going through the amendment procedure each time.

The consequence is that a power created for a transitional problem is permanent. It has been used repeatedly since the transition ended.

How it has been used

A short chronology is enough for the short note, and the dates are worth having.

1953, Andhra was created from the Telugu-speaking areas of Madras, the first State formed on a linguistic basis, and it prompted the appointment of the States Reorganisation Commission.

1956, the States Reorganisation Act redrew the map wholesale on linguistic lines and abolished the Part A, Part B, Part C and Part D classification. The Constitution (Seventh Amendment) Act, 1956 accompanied it.

1960, Bombay was bifurcated into Maharashtra and Gujarat, which is the reorganisation Babulal Parate was about.

1963 to 1987, the north-eastern States were created in stages: Nagaland, Meghalaya, Manipur, Tripura, Mizoram, Arunachal Pradesh, and Goa in 1987, most of them carrying a special provision in article 371.

1966, Punjab was reorganised and Haryana created; Chandigarh became a Union territory serving as the capital of both.

2000, three States were created in a single year: Chhattisgarh from Madhya Pradesh, Uttaranchal, later Uttarakhand, from Uttar Pradesh, and Jharkhand from Bihar.

2014, Telangana was created from Andhra Pradesh, the most recent new State.

2019, the State of Jammu and Kashmir was reorganised into two Union territories, Jammu and Kashmir with a legislature and Ladakh without one. This is the only occasion on which a State has been reduced to Union territories, and it is dealt with in [Article 370].

What this does to the federal argument

The point to make in an essay is not that Parliament has used the power often, but what it means that the power exists at all.

Facts. State of West Bengal v. Union of India, AIR 1963 SC 1241, (1964) 1 SCR 371, decided 21 December 1962. The Union sought to acquire coal bearing land owned by the State of West Bengal under a central Act. The State argued that it was sovereign within its own sphere and its property could not be taken by Union legislation.

Held. The Act was valid. The Constitution is not truly federal in character; full sovereignty does not vest in the States; and, in the sentence that matters here, Parliament, which is competent under article 3 to destroy a State, cannot be held on a theory of absolute State sovereignty to be incompetent to acquire by legislation the property owned by that State.

Why it matters. It uses article 3 as the premise for a much wider conclusion about the nature of the Indian union. If the greater power exists, the lesser cannot be denied.

The counter-argument is worth knowing too. The existence of a power says nothing about the constraints on exercising it. Reorganisation in practice has followed sustained popular demand, usually after a commission or a long agitation, and no State has ever been abolished against a settled political consensus. On that view article 3 is a power that political reality restrains even though law does not. Whether a restraint that the Constitution does not impose is a federal guarantee at all is exactly the argument MU's phrase "trust and faith" is pointing at.

A worked example

A Bill under article 3 proposing to divide State R into two States is referred by the President to the legislature of State R, which is given sixty days to express its views. The legislature passes a resolution opposing the division outright. Parliament then amends the Bill so that instead of two States there are three, and passes it without any further reference.

Is the Act valid?

On the resolution opposing it. The proviso requires that the Bill be referred to the State legislature "to express its views thereon". It does not require consent, and Parliament is free to proceed against the view expressed.

On the modification. This is Babulal Parate exactly. What must be referred is the proposal contained in the Bill, and the proviso does not contemplate that a subsequent modification by Parliament requires a fresh Bill or a fresh reference. The three-State scheme is valid without a second reference.

On the majority. Article 4 provides that a law under article 3 is not an amendment for the purposes of article 368, so an ordinary majority suffices and the First and Fourth Schedules may be amended by the same Act.

What if the President had never referred it at all? That is different: the reference is a condition on the introduction of the Bill, not a formality about its content. Article 255 cures a missing recommendation or previous sanction where assent was given, and whether it reaches this proviso is a question worth flagging rather than asserting.

Quick revision

• Article 2 admits or establishes new States from outside the Union, on such terms as Parliament thinks fit; article 3 rearranges the units already inside it.
• Article 3 needs the President's recommendation and a reference of the Bill to the State legislature for its views. The views do not bind.
• Babulal Parate: Parliament may modify the proposal after the reference without referring it again.
• Article 4 makes such a law an ordinary law, expressly not an amendment under article 368.
• The chronology worth quoting: Andhra 1953, the States Reorganisation Act 1956, Bombay 1960, the north-east from 1963, Punjab 1966, three States in 2000, Telangana 2014, and Jammu and Kashmir reduced to Union territories in 2019.
• The federal significance: a unit that can be abolished by ordinary majority is not sovereign, and West Bengal v Union of India draws exactly that conclusion.

Test yourself

1. Distinguish article 2 from article 3, and give the five limbs of article 3.

2. What are the two conditions in article 3, and how far does the second protect a State? Answer with Babulal Parate.

3. Why does article 4 exist, and what would change if a law under article 3 had to satisfy article 368?

4. Give the chronology of State reorganisation from 1953 to 2019, and say what the 2019 reorganisation did that no earlier one had done.

Clause (3) gives the State legislature exclusive power over List II, and it is expressed to be "subject to clauses (1) and (2)".

Clause (4) gives Parliament power to legislate for any part of India not included in a State, with respect to any matter, including a matter in the State List.

Read the non obstante and subject-to phrases together and the hierarchy falls out: List I prevails over List III and List II; List III prevails over List II. The State List is the only one that is subordinate to both the others, so a State's exclusive field is exclusive only where nothing in the other two Lists covers the ground.

That is the structural reason Indian federalism tilts to the Centre, and it is worth saying in an essay in exactly this form: the tilt is not in the length of the Union List, it is in the ordering of the clauses.

The three Lists

List I, the Union List, has ninety-seven entries and covers what has to be uniform for the country: defence, foreign affairs, citizenship, railways, currency, banking, insurance, corporation tax, income tax other than agricultural income, inter-State trade, and the residuary entry 97.

List II, the State List, has sixty-six entries and covers what can vary locally: public order, police, local government, public health and sanitation, agriculture, land, liquor, betting and gambling, and taxes on land, buildings and agricultural income.

List III, the Concurrent List, has forty-seven entries and covers what both levels have a real interest in: criminal law and criminal procedure, marriage and divorce, transfer of property other than agricultural land, contracts, bankruptcy, trusts, forests, education, electricity, factories, labour welfare, and economic and social planning.

Three entries moved from the State List to the Concurrent List by the Forty-second Amendment in 1976: education, forests, and weights and measures, together with the protection of wild animals and birds and the administration of justice. That single amendment did more to shift the balance than most of the emergency provisions, and it is worth naming in a federalism answer.

Interpreting an entry

Courts have settled a few rules of construction, and an examiner expects them.

An entry is given the widest possible meaning. Each general word is held to extend to all ancillary or subsidiary matters which can fairly be said to be comprehended in it. A power to legislate on a subject carries the power to legislate on everything reasonably incidental to it.

The entries are fields of legislation, not powers. They describe subject matter; the power comes from article 246.

Taxation entries are separate from general entries. A general entry on a subject does not carry a power to tax that subject: the taxing entries are set out separately in each List, and a tax must be traced to a taxing entry.

An apparent conflict is reconciled if possible. The court first tries to give effect to both entries by reading one as limited by the other; only if reconciliation is impossible does the hierarchy of clauses decide.

Facts. State of Bombay v. F.N. Balsara, AIR 1951 SC 318, 1951 SCR 682, decided 25 May 1951. The Bombay Prohibition Act 1949 prohibited the possession and sale of liquor, including imported liquor. Intoxicating liquor was a Provincial entry; import and export across customs frontiers was a Dominion entry.

Held. There was no conflict. "Possession and sale" in the Provincial entry must be read without qualification, and "import" standing alone does not include the sale or possession of the article imported. Even assuming that prohibiting purchase, possession, transport and sale would affect import, the Act was in pith and substance a law on intoxicating liquors, and an incidental encroachment on the Dominion field did not affect its validity. The American doctrine of the "original package" has no application in India, given how precisely the entries are worded.

Why it matters. It shows the reconciliation rule at work before any doctrine is needed: two entries that look as though they collide are read so that they do not.

Residuary power

Article 248 gives Parliament exclusive power to make any law with respect to any matter not enumerated in the Concurrent or State Lists, including the power to impose a tax not mentioned in either. Entry 97 of List I says the same thing from inside the List.

This reverses the arrangement in the United States and Australia, where the residue stays with the units, and it follows from India's being a holding-together federation.

Facts. Union of India v. H.S. Dhillon, AIR 1972 SC 1061, (1972) 2 SCR 33, decided 21 October 1971. Section 24 of the Finance Act 1969 amended the Wealth Tax Act to include agricultural land in the assets on which net wealth is computed. The High Court held this beyond Parliament's competence, because a tax touching agricultural land belonged to the States under Entry 49 of List II, "taxes on lands and buildings".

Held. The amendment is valid. Articles 246 and 248 with Entry 97 mean that any matter, including a tax, not allotted exclusively to the States under List II or concurrently under List III falls within List I. It follows, and this is the part to remember, that when a central Act is challenged the only enquiry is whether it legislates on a matter in List II, because that is the only field from which Parliament is barred. If the central Act does not enter that field, there is no point asking which entry of List I or List III it fits.

Why it matters. It is the leading authority on residuary power and it inverts the usual enquiry. For a State law you ask "which entry authorises this"; for a central law you ask only "does this trespass on List II".

Article 246A, the entry that is not in a List

The Constitution (One Hundred and First Amendment) Act, 2016 introduced the goods and services tax, and it did not do so by adding an entry to a List. It inserted article 246A, a free-standing provision which begins "Notwithstanding anything contained in articles 246 and 254" and gives Parliament and every State legislature power to make laws with respect to goods and services tax.

Two features are examinable. It is a concurrent power that sits outside the Concurrent List, with Parliament given exclusive power where the supply is inter-State. And its non obstante clause displaces article 254, so the ordinary repugnancy rule does not apply to it in the same way.

This is dealt with fully in [The Goods and Services Tax and Its Council], because the interesting question is not the drafting but the Council that article 279A creates.

Where this leads

The Lists overlap. They were bound to: no list of subjects drawn in the abstract survives contact with a real statute, because a real statute touches several subjects at once. The Constitution therefore needs techniques for deciding which List a law belongs to and what happens when two valid laws collide. Those are the next three chapters:

• Which List does this law really belong to? [Pith and Substance].
• Is the legislature pretending to legislate on its own subject while really legislating on another's? [Colourable Legislation and Incidental Encroachment].
• Both laws are valid, and they conflict. Which prevails? [Repugnancy and the Occupied Field].

A worked example

A State legislature enacts a law taxing the entry of goods into a local area for consumption, and requiring every carrier to obtain a permit before crossing the State border. A carrier challenges both.

Step one, competence for the tax. Find a taxing entry in List II. If one covers entry of goods into a local area, the State is competent, and a general entry alone would not do, because a taxing power must be traced to a taxing entry.

Step two, competence for the permit. Trade and commerce within the State is a State entry; inter-State trade and commerce is a Union entry. Ask what the permit requirement is in pith and substance about. If it regulates carriers operating within the State, it is within List II; if it operates on the movement of goods across the border as such, it is not.

Step three, is there a central law? If Parliament has legislated on the same Concurrent matter, article 254 arises, which is a separate question from competence.

Step four, Part XIII. Even a competent State law must satisfy articles 301 to 304. The tax must not discriminate against goods from other States under article 304(a); the permit requirement is a non-fiscal restriction on the freedom under article 301 and needs article 304(b), which requires the President's previous sanction.

The order matters. Competence first, then repugnancy, then the freedom of trade. A candidate who begins with article 301 will miss the point that an incompetent law never reaches it.

Quick revision

• Article 245 is reach; article 246 is subject matter.
• A Union law is not invalid for extra-territorial operation; a State law needs a territorial nexus.
• List I, ninety-seven entries; List II, sixty-six; List III, forty-seven. List I prevails over both, List III over List II.
• The tilt is in the ordering of the clauses of article 246, not in the length of the Lists.
• Entries get the widest meaning; taxing power must be traced to a taxing entry; conflicts are reconciled where possible.
• Residuary power is Parliament's under article 248 and Entry 97, and Dhillon says the only question about a central Act is whether it invades List II.
• Article 246A gives GST a concurrent power outside the Lists, and displaces article 254.

Test yourself

1. Explain how the non obstante and subject-to clauses of article 246 create the hierarchy among the three Lists.

2. What does article 245(2) provide, and what is the corresponding position of a State legislature?

3. State the rules for construing an entry in a List, and explain why a taxing power must be traced to a taxing entry.

4. What is the test in Union of India v. H.S. Dhillon for a central Act, and why does it reverse the usual enquiry?

the extent of the invasion by the Provinces into subjects enumerated in the Federal List has to be considered. No doubt it is an important matter, not, as their Lordships think, because the validity of an Act can be determined by discriminating between degrees of invasion, but for the purpose of determining what is the pith and substance of the impugned Act.

The question, the Board said, is not whether the Act has trespassed more or less, but whether the trespass, whatever it be, is such as to show that the pith and substance of the Act is not money-lending but promissory notes or banking. Once that is determined the Act falls on one side of the line or the other. The priority of the Federal List does not prevent a Provincial legislature from dealing with a matter that incidentally affects a Federal entry.

Why it matters. It is the classical statement of the doctrine for India, and it corrects the error almost every weak answer makes. The size of the encroachment is evidence, not a test. A law may trespass a great deal and still be valid if the trespass does not change what the law is about; a law may trespass slightly and be invalid if the slight trespass shows the law was really about the other subject all along.

The doctrine at work

Facts. State of Bombay v. F.N. Balsara, AIR 1951 SC 318, 1951 SCR 682, decided 25 May 1951. The Bombay Prohibition Act 1949 forbade the possession and sale of liquor, including imported liquor, and was said to encroach on the Union entry for import and export.

Held. No conflict arose on a proper reading of the two entries, and in any event the Act was in pith and substance a law on intoxicating liquors within the Provincial entry. An incidental encroachment on the Union field did not affect its validity.

Why it matters. It is the standard Indian illustration, and it shows the order of the enquiry: reconcile the entries first, and reach for pith and substance only if reconciliation fails.

Facts. M. Karunanidhi v. Union of India, AIR 1979 SC 898, (1979) 3 SCC 431, decided 20 February 1979. A State Act on the criminal misconduct of public men was said to be repugnant to central anti-corruption legislation and to trespass on the Union field.

Held. Among the propositions the Court set out was this one: where a law passed by a State legislature, while being substantially within the scope of the entries in the State List, entrenches upon any of the entries in the Union List, the constitutionality of the law may be upheld by invoking the doctrine of pith and substance if, on an analysis of its provisions, it appears that by and large the law falls within the four corners of the State List, the entrenchment being purely incidental or inconsequential.

Why it matters. It is the modern restatement, in a form drafted to be quoted, and it links pith and substance to repugnancy, which is the subject of a later chapter.

How a court actually applies it

Four steps, in this order.

One, read the impugned law as a whole. Not the section complained of in isolation: the object, the scope, and the effect of the enactment taken together. The preamble and the scheme are legitimate aids.

Two, try to reconcile the competing entries. Give each its widest meaning, and see whether they can both stand without collision. Balsara did exactly this and never needed the doctrine.

Three, if they cannot be reconciled, identify the true nature and character of the law. What is the law substantially about? What is the mischief it addresses, and to which List does that mischief belong?

Four, having allotted the law to a List, treat everything else it touches as incidental. An incidental encroachment is not a ground of invalidity. If the law belongs to the legislature that made it, it is valid in its entirety, including in its incidental operation on the other field.

What the doctrine does not do

It does not cure a want of power. If the law in pith and substance belongs to the other List, the doctrine does not save it, however small the trespass appears.

It does not decide a conflict between two valid laws. Pith and substance answers "was this legislature competent". When both legislatures were competent and their laws collide, the question is repugnancy under article 254, which is a different enquiry with a different test. Keep them apart in an answer: competence first, repugnancy only if both laws are competent.

It does not turn on the motive of the legislature. Whether the legislature meant well is irrelevant. What its law is really about is a question of the law's content, not of anyone's intention, and the point where motive does come in, or rather where a pretence does, is the subject of the next chapter.

A worked example

A State legislature enacts the (imaginary) State Agricultural Credit Act. It caps the interest a lender may charge a farmer, requires every such loan to be registered with a State officer, and provides that a cheque given as security for such a loan shall be unenforceable beyond the capped amount. Money-lending and agriculture are State subjects; cheques are negotiable instruments and belong to the Union.

Reconcile the entries first. Ask whether the money-lending entry and the negotiable instruments entry can both be given effect. If the Act can be read as regulating the lending relationship and not the instrument's negotiability, there may be no conflict at all, which is the Balsara move.

If they cannot be reconciled, find the pith and substance. Read the Act as a whole. Its object is the protection of agricultural borrowers; its scope is loans to farmers; its effect is to limit what a lender recovers. That is money-lending.

Apply Prafulla Kumar to the cheque provision. Do not ask whether the trespass on negotiable instruments is large or small. Ask whether the trespass shows that the Act is really about cheques rather than about money-lending. It does not: the provision exists only because loans are secured by cheques, and it operates on the lender's recovery, not on negotiability generally.

Conclusion. The Act is valid in its entirety. The effect on cheques is an incidental encroachment, and a power over a subject carries the power over what is reasonably ancillary to it.

What would change the answer? A provision regulating the transfer of cheques between third parties who are strangers to any loan. That would operate on negotiability itself and would suggest the Act's true nature lies in the Union field.

Quick revision

• Pith and substance asks what a law is really about, and allots it to the List its true nature belongs to.
• It exists because lists of abstract subjects always overlap with real statutes.
• Prafulla Kumar: the extent of the invasion is evidence of what the law is about, not a test of validity. Not "has it trespassed more or less" but "does the trespass show the law is about the other subject".
• Balsara: reconcile the entries first; use the doctrine only if they will not reconcile.
• Karunanidhi: a State law substantially within List II that entrenches incidentally on List I is saved.
• An incidental encroachment never invalidates. A law whose true nature lies in the other List is never saved.

Test yourself

1. State the doctrine of pith and substance and explain why the overlapping of the Lists makes it necessary.

2. Quote and explain the Privy Council's statement in Prafulla Kumar about the extent of the invasion. Why is it the passage most often misused?

3. In what order does a court apply reconciliation of entries and the doctrine of pith and substance? Answer with Balsara.

4. Distinguish the question the doctrine answers from the question article 254 answers.

If the answer is "the List belonging to the legislature that made it", the law is valid, and whatever else it touches is an incidental encroachment. That is pith and substance.

If the answer is "the List belonging to the other legislature", the law is void, and the fact that it was drafted to look like something else does not save it. That is colourable legislation.

So the two doctrines share the whole of their reasoning and differ only in the outcome. An answer that says so is stronger than one which treats them as separate tests to be applied one after the other.

The classic illustration

The case cited in every account of this doctrine is K.C. Gajapati Narayan Deo v. State of Orissa, decided by the Supreme Court in 1953. Orissa amended its agricultural income-tax law shortly before acquiring estates under its estates abolition legislation, in a way that raised the assessed income of large landholders and so reduced the compensation payable to them, compensation being calculated by reference to that income. The landholders argued that the amendment was not really a tax law at all but a device to cut compensation, and therefore colourable.

The Court upheld the amendment, holding that the State was competent to legislate on agricultural income-tax and that the motive with which it exercised a power it undoubtedly had could not affect the validity of the law. In the course of doing so it gave the classical exposition of the doctrine, including the proposition that the idea conveyed by the expression colourable legislation is that although apparently a legislature in passing a statute purports to act within the limits of its powers, yet in substance and in reality it transgresses those powers, the transgression being veiled by what appears on proper examination to be a mere pretence or disguise.

The report of this judgment could not be opened from the sources this book is built from, so no citation is printed for it here. The account above is given as the standard textbook account and should be checked against the report before it is relied on in a written answer. The doctrine itself does not depend on it: the reasoning is fully stated in Prafulla Kumar, below, and in the modern restatements.

The other half: incidental encroachment

The positive side of the same coin, and the one that decides far more cases.

An incidental or ancillary encroachment is an effect a law has on a field outside its own, which follows naturally from doing what the law is properly about. Because a power to legislate on a subject carries with it the power to legislate on matters reasonably ancillary to it, such an encroachment does not invalidate.

Facts. Prafulla Kumar Mukherjee v. Bank of Commerce Ltd, Khulna, (1947) 49 Bom LR 568, also reported at [1947] FCR 28, Privy Council, 11 February 1947. The Bengal Money-Lenders Act 1940 limited what a money-lender could recover, including on a promissory note. Money-lending was Provincial; promissory notes and banking were Federal.

Held. The Act was valid. Where subjects overlap the court asks what in pith and substance the enactment is, and in what List its true nature is to be found. The extent of the invasion into the Federal field is important as evidence of what the law is about, and not because validity turns on discriminating between degrees of invasion. The question is not whether the Act has trespassed more or less, but whether the trespass shows that the pith and substance of the Act is promissory notes or banking rather than money-lending. The priority of the Federal List does not prevent a Provincial legislature from dealing with a matter that incidentally affects a Federal entry.

Why it matters. It supplies the test for incidental encroachment as well as for pith and substance, and it makes clear that the two are the same test producing different answers.

How to tell the two apart in a problem

Four questions, in order, will resolve almost any problem on this material.

Is the encroachment necessary to the operation of the law? If the law cannot do its own work without touching the other field, the encroachment is ancillary and the law survives. A money-lending law that could not reach promissory notes would not be a money-lending law at all.

Does the law provide for the other subject in its own right? If the statute contains machinery that only makes sense as regulation of the other subject, that is evidence its true nature lies there.

Would the law still make sense with the encroaching part removed? If yes, the encroachment is probably incidental. If the whole scheme collapses, the encroachment may be the point of the law.

Is the form of the law doing work that its substance does not? If the statute is labelled and structured as one thing but every operative provision does another, that is the disguise the doctrine is aimed at.

Where the doctrine applies, and where it does not

It applies to any limit on legislative power, not only to the Lists. A legislature is limited by subject matter, and it is also limited by the fundamental rights and by any express constitutional prohibition. Any of those limits may be evaded by drafting, and the doctrine reaches all of them.

It does not apply where the legislature has plenary power over the subject. If the subject belongs to the legislature, there is no limit to evade, and the enquiry stops. This is why Gajapati failed on its facts: Orissa was competent to tax agricultural income, so its reason for doing so was beside the point.

It does not apply to a delegated instrument in the same form. Subordinate legislation is tested for whether it is within the enabling Act, which is the ordinary doctrine of ultra vires and is dealt with in [Ultra Vires].

A worked example

A State is competent to tax agricultural income. It is also acquiring land under a land reform statute which fixes compensation as a multiple of the landholder's assessed agricultural income. Shortly before the acquisitions, it amends its agricultural income-tax law in a way that sharply increases the assessed income of large holders, so that the compensation payable to them falls.

The landholders' argument. The amendment is not really a tax measure at all. Its purpose and effect are to cut compensation, which the State could not have done directly without breaching the compensation provisions. It is therefore colourable.

The first question is competence, not motive. Is the State competent to tax agricultural income? Yes, it is a State entry. That is the beginning and, on the standard reasoning, the end.

Why motive does not save the challenge. The doctrine has nothing to do with the bona fides of the legislature. Where the legislature possesses the power, the reason it exercised the power is irrelevant, and a court will not enquire into it. This is the reasoning by which the Supreme Court upheld the Orissa amendment in Gajapati.

When would the answer differ? If the amendment had no operation as a tax at all: if it applied only to the persons whose land was being acquired, or only for the period of the acquisitions, or if its machinery could not function as a tax. Then the form would be doing work the substance did not, and the pith and substance would be compensation, not taxation.

The lesson. Colourable legislation is about what the law IS, not about why it was passed.

Quick revision

• Quando aliquid prohibetur ex directo, prohibetur et per obliquum: what cannot be done directly cannot be done indirectly.
• The doctrine asks about competence, never about motive or bona fides.
• It is the negative face of pith and substance: same enquiry, opposite outcome.
• An incidental or ancillary encroachment never invalidates, because a power over a subject carries power over what is reasonably ancillary to it.
• Prafulla Kumar supplies the test: not more or less trespass, but whether the trespass shows the law is really about the other subject.
• Gajapati is the standard illustration and is named here without a citation because its report could not be read.

Test yourself

1. State the maxim on which the doctrine rests and explain what "colourable" describes.

2. Why is the doctrine of colourable legislation the negative face of pith and substance rather than a separate test?

3. "The doctrine has nothing to do with the motive or bona fides of the legislature." Explain, with the Orissa example.

4. Give four questions that will separate an incidental encroachment from a colourable exercise of power.

Held. The State Act was not repugnant. It was complementary and ran pari passu with the central Acts, providing an additional safeguard rather than an inconsistent rule, and although the ingredients of the offence were substantially the same, the punishment under the State Act was more severe, not less. In reaching that conclusion the Court set out the conditions that must be satisfied before any repugnancy can arise:

(i) That there is a clear and direct inconsistency between the Central Act and the State Act; (ii) that such an inconsistency is absolutely irreconcilable; (iii) that the inconsistency between the provisions of the two Acts is of such a nature as to bring the two Acts into direct collision with each other and a situation is reached where it is impossible to obey the one without disobeying the other.

The Court also summarised the circumstances in which repugnancy operates: where central and State laws in the Concurrent field are fully inconsistent and absolutely irreconcilable, the central Act prevails and the State Act becomes void; where the State Act has been passed in accordance with article 254(2), it prevails in that State to the extent of the repugnancy; and where a State law substantially within the State List entrenches on the Union List, it may be saved by pith and substance if the entrenchment is incidental.

Why it matters. It is the standard three-condition test, drafted in a form made to be quoted, and it makes the threshold high. Mere difference is not repugnancy; nor is greater severity; nor is duplication. There must be a direct collision such that obeying one means disobeying the other.

The two ways a conflict can arise

The test above catches the first way. The case law recognises a second.

Direct conflict. The two laws command incompatible things. One permits what the other forbids, or one requires what the other prohibits. Obedience to both is impossible. This is what Karunanidhi's three conditions describe.

The occupied field. Parliament has legislated on the subject so completely that the central law is intended to be a complete, exhaustive and exhaustive code on the matter, leaving no room for the State to add to it. Here there may be no direct contradiction at all: the State law may be perfectly consistent in its terms, and still be void because Parliament has covered the ground and evinced an intention that its scheme alone should apply.

The second is harder to establish and depends on the intention to be gathered from the central Act as a whole: its scheme, its comprehensiveness, and whether it contains its own provision about the operation of other laws. A saving clause preserving State laws is strong evidence against an intention to occupy the field.

Article 254(2), and the President's assent

The exception is the most practically important part of article 254, and its effects are frequently misstated.

What it requires. The State law must be with respect to a matter in the Concurrent List; it must be repugnant to an earlier law of Parliament or to an existing law; it must be reserved for the consideration of the President; and it must receive his assent.

What it achieves. The State law then prevails in that State. It does not repeal the central law, and it has no effect anywhere else. So the same central provision may be displaced in one State and fully operative in the next, which is one of the genuinely federal features of the Indian scheme.

What it does not achieve. By the proviso, Parliament may at any time enact a law with respect to the same matter, including a law adding to, amending, varying or repealing the law so made by the State legislature. The State's victory is therefore permanent only for as long as Parliament allows it. That is the point at which article 254(2) stops looking like a federal guarantee.

Two limits are examinable. The assent protects the State law only against the earlier central law that was placed before the President; it does not immunise it against a later central law. And the assent operates only where the President actually applied his mind to the repugnancy, so a State that reserves a Bill without disclosing the conflict cannot rely on the assent to cure it.

Article 251, and the temporary displacement

Article 251 deals with a different situation and belongs beside article 254 in an answer.

Where Parliament has legislated on a State subject under article 249, on a resolution of the Council of States, or under article 250 during a Proclamation of Emergency, a State legislature is not thereby disabled from legislating on the same matter. But if the two laws are repugnant, the law of Parliament prevails, and the State law is inoperative to the extent of the repugnancy only so long as the central law remains in force. When the central law lapses, the State law revives.

That word "inoperative" is doing work. Contrast article 254(1), under which the State law is void to the extent of the repugnancy. Under article 251 it is merely suspended.

Eclipse, and whether a void State law can revive

The revival question also arises under article 254(1), and the answer comes from the eclipse line of cases.

Facts. Deep Chand v. State of Uttar Pradesh, decided 15 January 1959. The Uttar Pradesh Transport Service (Development) Act 1955 authorised a State road transport scheme. Parliament then passed the Motor Vehicles (Amendment) Act 1956 on the same subject. The question was whether the State Act became void under article 254(1) and what happened to a scheme already framed under it.

Held. The State Act did not become wholly void; it continued as a valid subsisting law supporting the scheme already framed, and even if the central Act had repealed it under article 254(2), section 6 of the General Clauses Act would have saved the scheme. On the wider point, Das C.J. and Sinha J. held that there is no reason why the doctrine of eclipse should not apply to a post-Constitution law that infringes a fundamental right conferred on citizens alone: such a law is shadowed as against citizens but remains effective against non-citizens, and when the shadow is removed by amendment it applies to citizens without re-enactment.

Why it matters. It is the bridge between article 254 and the doctrine of eclipse, and it establishes that "void to the extent of the repugnancy" is a narrower thing than it sounds: the rest of the State law stands, and what has already been done under it may survive.

Read the citation for this case with care. The report's own equivalent-citations line was not captured when the judgment was fetched, so the AIR and SCR references recorded in the case bank are marked unread and must be confirmed against a report before they are copied into an answer. The holding above was read in full from the judgment.

A worked example

Parliament enacts a central Act on a Concurrent subject in 2015. In 2018 State S enacts a law on the same subject which sets a higher penalty and adds a procedural safeguard the central Act does not contain. The State Bill was reserved for the President and received his assent. In 2022 Parliament amends the central Act, raising its own penalty above the State's.

Step one, competence. Both legislatures are competent: the subject is in List III.

Step two, is there repugnancy at all? Apply the three conditions from Karunanidhi: a clear and direct inconsistency, absolutely irreconcilable, such that it is impossible to obey one without disobeying the other. A higher penalty and an extra safeguard are not necessarily irreconcilable with the central Act; a law that is complementary and runs pari passu is not repugnant. This is exactly what the Court held about the Tamil Nadu Act.

Step three, if there is repugnancy, what does the assent do? Article 254(2): a Concurrent State law repugnant to an EARLIER central law, reserved for and assented to by the President, prevails in that State. So between 2018 and 2022 the State law governs in State S, and the central Act is displaced there to the extent of the repugnancy.

Step four, the 2022 amendment. The proviso to article 254(2) preserves Parliament's power at any time to enact a law with respect to the same matter, including one adding to, amending, varying or repealing the State law. The 2022 amendment is a later central law and the assent does not protect the State law against it. From 2022 the central provision prevails.

The trap. The assent is not a permanent immunity. It protects only against the earlier law placed before the President.

Quick revision

• Order of enquiry: competence first, then whether the field is Concurrent, then repugnancy.
• Article 254(1): central law prevails, whether earlier or later, and the State law is void to the extent of the repugnancy.
• Karunanidhi's three conditions: clear and direct inconsistency, absolutely irreconcilable, direct collision so that obeying one means disobeying the other.
• Two routes to repugnancy: direct conflict, and the occupied field where Parliament intended its law to be a complete code.
• Article 254(2): a Concurrent State law repugnant to an earlier central law, reserved for and assented to by the President, prevails in that State. The proviso lets Parliament undo it at any time.
• Article 251: a State law inconsistent with a law made under article 249 or 250 is inoperative, not void, and revives when the central law goes.

Test yourself

1. In what order must competence and repugnancy be considered, and why does the order matter?

2. State the three conditions that must be satisfied before repugnancy can arise, from M. Karunanidhi.

3. Distinguish direct conflict from the occupied field, and say what evidence establishes the second.

4. Explain the effect of the President's assent under article 254(2), and its two limits.

5. Contrast the word "void" in article 254(1) with "inoperative" in article 251.

The duplication is deliberate belt and braces. Entry 97 makes the residuary power part of the Union List, so that everything said about List I in article 246 applies to it. Article 248 makes it a free-standing grant, so that it does not depend on the construction of an entry. Either route reaches the same place.

Article 247 is a different provision that sits nearby and is often confused with these. It empowers Parliament to provide for the establishment of additional courts for the better administration of laws made by Parliament or of existing laws with respect to a matter in the Union List. It has nothing to do with residuary legislative power.

The case that fixed the method

Facts. Union of India v. H.S. Dhillon, AIR 1972 SC 1061, (1972) 2 SCR 33, decided 21 October 1971. Section 24 of the Finance Act 1969 amended the Wealth Tax Act 1957 so that agricultural land was included among the assets by reference to which net wealth is computed. The High Court held the amendment beyond Parliament's competence, reasoning that a tax which took account of agricultural land could be imposed only by a State under Entry 49 of List II, "taxes on lands and buildings".

Held. The amendment is valid. Article 248 with Entry 97 of List I means that any matter, including a tax, which has not been allotted exclusively to the State legislatures under List II or concurrently under List III falls within List I. The Court then drew the consequence that has become the working rule:

If this is the true scope of residuary powers of Parliament, then when dealing with a Central Act the only enquiry is whether it is legislation in respect of any matter in List II, for, this is the only field regarding which there is a prohibition against Parliament. If a Central Act does not enter or invade these prohibited fields there is no point in trying to decide as to under which entry or entries of List I or List III a Central Act would rightly fit.

Why it matters. It inverts the enquiry for a central law. For a State law the question is "which entry of List II or List III authorises this", and if no entry does, the law fails. For a central law the question is only "does this invade List II", and if it does not, the law stands, whether or not any particular entry of List I can be identified. That is a large practical advantage for Parliament and it is worth saying so in a federalism essay.

The limits on the residuary power

The power is wide but it is not unlimited, and an answer that says only "the residue goes to the Centre" is incomplete.

It cannot be used to reach a subject that is in List II. That is the whole of the Dhillon test read the other way. Entry 97 is "any other matter not enumerated in List II or List III", so a matter that IS enumerated there is outside it by definition.

It cannot be used to enlarge an entry that already exists. If a subject falls within an entry of List II, Parliament cannot reach it by calling it residuary merely because the State entry is narrower than Parliament would like.

It is subject to the fundamental rights. Legislative competence and constitutional validity are different questions, and a law within the residuary power is still void if it violates Part III.

It is subject to the basic structure when exercised to amend. That is [The Basic Structure], and it does not arise from article 248 itself.

Why the residue matters more than it looks

Two consequences run through modern constitutional argument, and both are worth a paragraph in an essay.

New subjects belong to the Centre by default. Everything that did not exist in 1950 arrives in Parliament's hands unless it can be fitted into a State entry. Over seventy-five years that is a substantial and continuing transfer of power towards the Union, achieved without a single amendment.

The taxing residue is where the money is. Article 248 expressly includes the power to impose a tax not mentioned in the Lists, and taxes on new forms of wealth and new kinds of transaction have repeatedly been traced to it. Service tax, before the goods and services tax absorbed it, was levied under the residuary power for years before an entry was inserted for it. The revenue consequences of that are dealt with in [Fiscal Federalism].

A worked example

Parliament enacts a law imposing a levy on the transmission of data by satellite, and regulating who may operate such a transmission. A State challenges it, saying that no entry of the Union List mentions satellites or data, that broadcasting is not the same thing, and that in the absence of an entry the subject must belong to the States.

The State's argument inverts the correct test. For a central Act the question is not "which entry of List I authorises this". Dhillon holds that the only enquiry is whether the Act legislates on a matter in List II, because List II is the only field from which Parliament is barred.

So ask: is data transmission by satellite in List II? If no entry of the State List covers it, the enquiry ends and the Act is valid, whether or not any entry of List I or III can be identified, because article 248 with Entry 97 puts the residue with Parliament.

On the levy. Article 248 expressly includes the power to impose a tax not mentioned in the Lists, so the absence of a taxing entry is not a defect. But note the corollary: if the levy could be brought within a State taxing entry, the residuary power is unavailable, because Entry 97 is "any other matter not enumerated in List II or List III".

The limit worth stating. The residuary power cannot be used to enlarge a narrow State entry or to reach a subject List II already covers. And a law within the residuary power is still void if it violates Part III: competence and validity are different questions.

Quick revision

• Article 248 and Entry 97 of List I put the residue with Parliament, including the power to tax what the Lists do not mention.
• India follows the Canadian model, not the American or Australian one, because it is a holding-together federation.
• Article 247 is about additional courts and is not a residuary power at all.
• Dhillon: for a central Act the only question is whether it invades List II; if not, no entry of List I or III need be identified.
• The power cannot reach a subject already in List II, cannot enlarge an existing entry, and is subject to Part III.
• Everything invented since 1950 falls to the Centre by default, which is a quiet and continuing centralisation.

Test yourself

1. Where does the residuary power lie in India, and how does that differ from the United States and Australia? Explain the difference by reference to how each federation was formed.

2. Why does the Constitution state the residuary power twice, in article 248 and in Entry 97?

3. State the test in Union of India v. H.S. Dhillon and explain why it inverts the enquiry for a central Act.

4. What is article 247, and why is it not a residuary power?

5. "Everything invented since 1950 belongs to the Centre by default." Discuss.

This is the widest of the five and needs no resolution, no consent and no specification of subjects. It converts the Constitution into a unitary one for the duration, which is the standard description and is accurate. A law so made ceases to have effect six months after the Proclamation ceases, except as to things done before.

Again, article 251 applies: the State's own power continues, and a repugnant State law is inoperative rather than void.

The wider consequences of an emergency for the federal balance are in [Emergency and the Federal Balance].

Article 252: legislation by consent

How it works. If the legislatures of two or more States pass resolutions that a State List matter should be regulated by Parliament, Parliament may pass an Act for those States. Any other State may afterwards adopt the Act by a resolution of its own legislature, and the Act then applies there too.

What is distinctive about it. This is the one route that is genuinely consensual, and it is the one that produces a law which the States cannot then unmake: once Parliament has legislated under article 252, the Act may be amended or repealed only by Parliament, not by the State legislatures that asked for it.

That is the trap in the provision and it is examinable. A State surrenders the subject permanently, or at least until Parliament chooses to give it back.

What it is used for. Subjects on which uniformity is desirable but which the Constitution left with the States: estate duty, prize competitions, urban land ceiling, transplantation of human organs, and water pollution control have all been legislated in this way.

Article 253: the treaty power

The widest of the five, and the one that matters most today.

The text. Notwithstanding anything in the two preceding articles, Parliament has power to make any law for the whole or any part of the territory of India for implementing any treaty, agreement or convention with any other country or countries, or any decision made at any international conference, association or other body.

What that means. The non obstante clause displaces the distribution of powers entirely. If the law implements an international obligation, it may deal with a State List subject, and no resolution, consent, emergency or time limit is required. The law is permanent.

Why it matters more each year. India is party to a very large number of treaties and conventions, and many of them touch subjects that the Seventh Schedule gives to the States: environment, forests, wildlife, labour, health, human rights, agriculture and intellectual property among them. Environmental legislation in India is largely traceable to this power, enacted to give effect to the Stockholm Declaration of 1972 and the Rio Declaration of 1992, and the preambles of those Acts say so.

The obvious objection. If the Union executive negotiates the treaty and the Union Parliament implements it, the States can be legislated out of their own List by a process in which they take no part at all. The counter-argument is that the alternative, a country that cannot honour an international obligation because a State declines to legislate, would make India an unreliable party to every agreement it signs.

Note the limit that does exist. Article 253 authorises a law for implementing the treaty. A law that goes beyond what the international obligation requires cannot be justified under it, and must find competence elsewhere. In practice courts have read the implementing requirement generously, but the limit is there and should be stated.

Article 356: the legislature displaced

Where a Proclamation under article 356 is in force, the President may declare that the powers of the State legislature shall be exercisable by or under the authority of Parliament. Parliament may then legislate for the State on any subject, and may confer the power on the President or on another authority.

This is different in kind from the other four. Under articles 249, 250, 252 and 253 the State legislature continues to exist and to hold its own power. Under article 356 it has been dissolved or suspended and Parliament is acting in its place. See [President's Rule].

Reading the five together

Put in order of how much the State is consulted:

Article 252, the State asks. Article 249, the chamber representing the States agrees, by a special majority. Article 250, an emergency is in force and the States are not asked. Article 356, the State government is gone and the States are not asked. Article 253, the Union has made a treaty and the States are not asked, with no time limit and no expiry.

An answer that arranges them this way, rather than listing them in numerical order, is making a federalism argument rather than reciting provisions. The argument is that the Constitution's protections for the State List are graduated, that four of the five have a time limit or a consent requirement, and that the fifth, which has neither, is the one whose use has grown most.

A worked example

India ratifies an international convention requiring parties to regulate the disposal of a class of industrial waste. Waste disposal, sanitation and public health are State subjects. Parliament enacts a central Act implementing the convention, which occupies the field completely and displaces existing State legislation. Three States object that Parliament has legislated on a State subject without their consent and without any emergency.

Which route is being used? Article 253: Parliament may make any law for the whole or any part of India for implementing any treaty, agreement or convention with any other country, or any decision made at any international conference or other body.

Does it need consent? No. Article 253 opens "Notwithstanding anything in the two preceding articles", which displaces the distribution of powers. Unlike article 249 it needs no resolution of the Council of States, unlike article 250 it needs no emergency, and unlike article 252 it needs no request from the States.

Does it expire? No. A law under article 249 lapses six months after the resolution, and one under article 250 six months after the Proclamation. A law under article 253 is permanent.

Is there any limit? Yes, and it is the one to state. Article 253 authorises a law FOR IMPLEMENTING the treaty. A provision going beyond what the obligation requires cannot be justified under it and must find competence elsewhere. In practice the implementing requirement has been read generously, but the limit exists.

The federalism point. The Union executive negotiates the treaty and the Union Parliament implements it, so the States can be legislated out of their own List by a process in which they take no part. The answer given for that is that a country which cannot honour an obligation because a State declines to legislate would be an unreliable party to every agreement it signs.

Quick revision

• Five routes: articles 249, 250, 252, 253 and 356.
• Article 249 needs a two-thirds resolution of the Council of States, lasts a year, and is renewable without limit.
• Article 250 operates during a Proclamation of Emergency and covers any State subject.
• Article 252 needs two or more States to ask, and the resulting Act can then be amended or repealed only by Parliament.
• Article 253 implements treaties, displaces the distribution of powers entirely, needs no consent and never expires.
• Under articles 249 to 252 the State's own power survives and article 251 makes a conflicting State law inoperative, not void.

Test yourself

1. List the five routes by which Parliament may legislate on a State subject, with the article for each.

2. What majority does article 249 require, in which House, and for how long does the resolution last?

3. Why is a law made under article 252 unusual, and what can a State that asked for it no longer do?

4. Arrange the five routes in order of how far the States are consulted, and say what that ordering shows about the Constitution's protections for the State List.

5. What is the limit on article 253?

The doctrine of territorial nexus

A State legislature has no equivalent of article 245(2), and it does not need one for the ordinary case: almost every State law operates inside the State. The doctrine deals with the case where a State law has some operation outside.

The rule. A State law is not invalid merely because it affects a person or thing outside the State, provided there is a sufficient territorial nexus between the State and the subject matter of the law. Two conditions have to be met:

One, the connection must be real and not illusory. Some genuine link between the State and the object of the legislation, not a contrived one manufactured to found jurisdiction.

Two, the liability sought to be imposed must be pertinent to that connection. The tax, duty or obligation must relate to what connects the subject matter to the State, and not be a general liability hung on a slender local peg.

Where it typically arises. Sales tax on a transaction with elements inside and outside the State; a tax on a company incorporated elsewhere but doing business in the State; betting and gambling legislation where the bet is placed in one State and accepted in another; and legislation on entertainment broadcast from outside.

How to use it in a problem. Identify the element of the transaction that is inside the State. Ask whether that element is a real part of the transaction or an incident of it. Then ask whether what the State is imposing is related to that element or to the transaction as a whole. A State that taxes the whole of an inter-State transaction because one small part of it happened locally will fail the second condition even if it passes the first.

Delegated legislation and the reach of a law

Two related points that examiners like.

A law may be made to operate prospectively or retrospectively. Neither Parliament nor a State legislature is confined to the future, subject to article 20(1) in the case of criminal liability, which forbids conviction for an act that was not an offence when committed. That is [Article 20: The Protections of an Accused].

A law may be brought into force in different parts of the territory at different times, and may leave the date of commencement to the executive, because fixing a date is not an essential legislative function. The limits on what may be left to the executive are the doctrine of excessive delegation, in [Delegated Legislation].

Article 255: a saving for procedure

Several provisions of the Constitution require a recommendation or a previous sanction before a Bill is introduced or a law is passed: the President's recommendation under article 3, or under article 117 for a money Bill, or the Governor's under article 207.

Article 255 provides that no Act, and no provision in an Act, shall be invalid by reason only that some recommendation or previous sanction required by the Constitution was not given, if assent to that Act was given by the Governor, or the President, or by the President after having been reserved for his consideration.

What this means. These requirements are matters of procedure, not of competence. A defect in obtaining them is cured by the assent that follows. It is a sensible provision: a law should not be void years later because a formality was missed at the outset, when the authority whose recommendation was required has since assented to the whole thing.

What it does not cure. A want of legislative competence. If the legislature could not have made the law at all, no assent makes it valid. Article 255 saves procedure and nothing else, and the distinction between a procedural requirement and a substantive limitation is the whole of the article.

A worked example

State T imposes a tax on every lottery ticket sold anywhere in India by a lottery organised, promoted or conducted by State T's own lottery department, and also on tickets of other States' lotteries sold within State T. A ticket buyer in another State challenges the first limb.

Which limit is in question? Not subject matter but territory. Betting and gambling is a State entry, so competence by subject is not the difficulty. Article 245(1) confines a State legislature to making laws for the whole or any part of the State.

Apply the territorial nexus test, both limbs. First, is the connection between State T and the subject matter real and not illusory? For a ticket sold outside the State, in a lottery organised and conducted by State T, with the draw held in State T and the prize paid from State T, there is a genuine connection. Second, is the liability pertinent to that connection? The tax is on the lottery activity that State T itself conducts, so it is.

The second limb is where such laws usually fail. A State that taxes the whole of a transaction because one incidental element occurred locally satisfies the first limb and fails the second. Identify the element that is inside the State, then ask whether what is being imposed relates to that element or to the transaction as a whole.

Contrast Parliament. Article 245(2) would remove the objection entirely: a law of Parliament is not invalid on the ground that it has extra-territorial operation. A State has no such protection and must rely on nexus.

Quick revision

• Article 245(1) is territorial reach; article 246 is subject matter. Both must be satisfied and they are independent.
• Article 245(2): a law of Parliament is not invalid for extra-territorial operation. That removes one ground of challenge and no others.
• A State law reaching outside the State needs a territorial nexus: a connection that is real and not illusory, and a liability pertinent to that connection.
• Article 255 makes a missing recommendation or previous sanction a curable procedural defect, provided the necessary assent was given.
• Article 255 never cures a want of competence.

Test yourself

1. Distinguish the limit in article 245 from the limit in article 246, and explain why both must be satisfied.

2. What does article 245(2) protect a law of Parliament against, and what does it not protect it against?

3. State the two conditions of the doctrine of territorial nexus, and explain which of the two most often defeats a State law.

4. What does article 255 cure, and what does it never cure?

Notice that this is a power to direct the manner of compliance with a law of Parliament. It is not a general power to direct a State on anything.

Article 257: the State must not impede the Union

Article 257(1) states the converse duty in general terms: the executive power of a State shall be so exercised as not to impede or prejudice the exercise of the executive power of the Union, and the Union may give such directions as appear necessary for that purpose.

Clauses (2) and (3) then give two specific powers, which are narrower and more concrete.

Communications of national or military importance. The Union may direct a State as to the construction and maintenance of such means of communication, even though they may be a State subject. Roads are in the State List.

Protection of the railways. The Union may direct a State on measures to be taken for protecting railways within the State.

Article 257(4) provides that where a State incurs extra costs in carrying out such a direction, the Union must pay the State the amount of those costs, as agreed or as determined by an arbitrator appointed by the Chief Justice of India. That is a genuinely federal provision, and it is worth naming.

Articles 258 and 258A: functions moving both ways

Article 258(1) lets the President, with the consent of the State Government, entrust to that Government or its officers functions relating to any matter to which the executive power of the Union extends. This is how central schemes are run through State machinery.

Article 258(2) goes further and does not require consent. A law made by Parliament which applies in a State may confer powers and impose duties upon the State or its officers, even in relation to a matter outside the State's own executive power. The State must comply, and article 258(3) requires the Union to pay any extra administrative cost.

Article 258A, inserted by the Seventh Amendment, allows the traffic in the other direction: the Governor of a State may, with the consent of the Government of India, entrust to that Government or its officers functions relating to any matter to which the executive power of the State extends. It was added because the original Constitution allowed entrustment only downwards, which turned out to be a gap.

Articles 259 and 260 complete the picture. Article 260 permits the Government of India, by agreement with the government of a territory not part of India, to undertake executive, legislative or judicial functions in relation to it.

Article 261: full faith and credit

Full faith and credit shall be given throughout the territory of India to public acts, records and judicial proceedings of the Union and of every State. Parliament may by law provide how such acts and records are to be proved, and final judgments or orders delivered by civil courts in any part of India are capable of execution anywhere in India according to law.

The provision is modelled on article IV of the United States Constitution. In India it matters less than it does there, because the judicial system is integrated and the substantive law is largely uniform, but it is the article that makes a decree of one State's court executable in another.

Article 365: the sanction behind the directions

The direction powers in articles 256 and 257 would be advisory without a sanction, and article 365 supplies one:

Where any State has failed to comply with, or to give effect to, any directions given in the exercise of the executive power of the Union under any of the provisions of this Constitution, it shall be lawful for the President to hold that a situation has arisen in which the Government of the State cannot be carried on in accordance with the provisions of this Constitution.

That sentence connects Chapter II to article 356. Non-compliance with a direction is not itself a ground for dismissal; it is a ground on which the President may hold that the article 356 condition is satisfied. The word is permissive, and after S.R. Bommai the holding is reviewable like any other satisfaction under article 356. See [President's Rule].

Federal comity

This is the doctrine MU's phrase about trust and faith points at, and it deserves a name in an answer.

The idea. In a federation the two levels are not adversaries and not strangers. Each holds powers that could, if pressed to their limit, frustrate the other's functioning. The Constitution does not spell out most of the restraints that stop that happening, so the courts have supplied a principle: the Union and the States should each exercise their powers with due regard to the legitimate interests of the other, and neither should use a power it undoubtedly possesses in a manner calculated to obstruct the other's functioning.

Where it shows up. In the construction of a direction under article 256 or 257, in the exercise of the Governor's discretion, in the reservation of Bills for the President's consideration, in the timing of an article 356 Proclamation, and in the way the Union deploys central agencies within a State.

What kind of rule it is. For most of its life federal comity has been a principle of interpretation and of political practice rather than a rule producing a remedy. It tells a court how to read a power; it has rarely told a court to strike something down. That is changing at the margins, and the recent litigation about a Governor's indefinite delay in dealing with Bills is the clearest example, dealt with in [The Governor].

How to use it in the essay. MU's question asks about a "relationship of trust and faith". The honest answer is that the Constitution left a great deal of the federal relationship to good faith because a holding-together federation drafted in 1949 could not anticipate every way the Centre and a State might fall out, and that where the good faith has broken down, the courts have had to convert an expectation into a rule. Bommai did that for article 356. The assent litigation is doing it for article 200. The chapters on both are the evidence for the proposition.

A worked example

State U declines to enforce a central Act within its territory, instructing its police not to register offences under it. The Union issues a direction under article 256 requiring the State to ensure compliance. The State ignores the direction. The Union proposes to act under article 356.

Is the direction lawful? Article 256 obliges every State to exercise its executive power so as to ensure compliance with laws made by Parliament, and empowers the Union to give such directions as appear necessary for that purpose. A direction to enforce a central Act is squarely within it. Note that article 256 is about compliance with a law of Parliament; it is not a general power to direct a State on anything.

What does non-compliance do? Article 365 provides that where a State has failed to comply with a direction, it shall be lawful for the President TO HOLD that a situation has arisen in which the government of the State cannot be carried on in accordance with the Constitution. The word is permissive. Non-compliance does not automatically justify a Proclamation; it makes the holding available.

Is the holding reviewable? Yes. After Bommai the satisfaction under article 356 is a conditioned power resting on relevant material, and the Proclamation may be struck down if mala fide or based on wholly irrelevant or extraneous grounds. The Union must produce the material.

What should the Union do first? Sarkaria recommended a warning and the exhaustion of alternatives, and Bommai says those recommendations merit serious consideration. Article 356 is a last resort, and a Union that reaches for it without first pressing article 256 and article 257 has weakened its own case.

The comity point. A State that refuses to enforce a valid central law, and a Union that dismisses a government at the first opportunity, are each using a power they have in a way that defeats the other's functioning. That is what federal comity forbids.

Quick revision

• The States administer central law, so the Union needs a power to direct them: article 256 for compliance with Union law, article 257 for not impeding the Union.
• Article 257 adds two specific powers, over communications of national or military importance and the protection of railways, with the extra cost payable by the Union under 257(4).
• Articles 258 and 258A move functions down and up; 258(2) can impose duties on a State by a law of Parliament without its consent.
• Article 261 is full faith and credit, making a civil decree executable throughout India.
• Article 365 is the sanction: failure to comply with a direction lets the President hold that the article 356 condition is met, and that holding is reviewable after Bommai.
• Federal comity is the name for the restraint the text does not impose, and it is what MU's phrase "trust and faith" is describing.

Test yourself

1. Why does the Constitution give the Union a power to direct the States on administration, when it does not divide administration the way it divides legislation?

2. Distinguish the direction power in article 256 from that in article 257, and name the two specific powers in article 257 with the cost provision that accompanies them.

3. What do articles 258 and 258A each permit, and which of them does not require consent?

4. Explain article 365 and its relationship with article 356. What is the significance of the word "may"?

5. What is federal comity, and how does it relate to MU's phrase "trust and faith"?

Section 5. The Tribunal investigates and forwards a report and a decision. Under the 2002 amendment it is to do so within three years, extendable by two.

Section 6. The decision, when published in the Official Gazette, is final and binding on the parties, and has the same force as an order or decree of the Supreme Court. That equivalence was added by the 2002 amendment and matters: before it, enforcement was a genuine difficulty.

Section 11. The bar. Notwithstanding anything in any other law, neither the Supreme Court nor any other court shall have or exercise jurisdiction in respect of any water dispute which may be referred to a Tribunal under the Act.

Why they take decades

MU's question asks for the reasons for delay, so this section is the answer and should be given as a list.

One, the reference itself is discretionary in practice. A Tribunal is constituted when the Central Government is of opinion that negotiation has failed. Before 2002 there was no time limit, and requests waited years. The political cost to the Union of constituting a Tribunal, when both States are governed by parties whose support it needs, is real.

Two, a new Tribunal is constituted for every dispute. Each starts from nothing: new members, new registry, new procedure, new experts, and no accumulated jurisprudence. Nothing is reusable.

Three, the science is slow and contested. These disputes turn on hydrological data over long periods, on rainfall and yield estimates, on crop patterns and on what a "dependable flow" means. Each side commissions its own studies and the Tribunal must resolve them.

Four, there is no deadline that bites. The 2002 amendment introduced three years extendable by two, but extensions have been repeatedly granted, and the Act provides no consequence for exceeding them.

Five, the further reference under section 5(3). A party may, within three months of the report, seek explanation or guidance from the Tribunal, which then reconsiders. In practice this has functioned as a second round of litigation.

Six, the section 11 bar has not stopped litigation, it has redirected it. The bar excludes jurisdiction over the water dispute. It does not exclude a challenge to the constitution of the Tribunal, or to the validity of the Act, or a petition under article 32 alleging violation of a fundamental right, or a proceeding to enforce or interpret an award. Parties have used all of these, so the dispute reaches the Supreme Court by another door, and the exclusion adds a preliminary jurisdictional argument to every case rather than removing litigation.

Seven, the politics is unresolvable by adjudication. Water is existential for farmers in the basin and therefore decisive at elections in the basin States. A State government that accepts an adverse award pays for it, so no government accepts one willingly, and compliance becomes a further dispute.

Eight, enforcement machinery is weak. Even a binding award needs an authority to operate the reservoirs and release the water in accordance with it, and creating that authority has itself been litigated.

The 2019 amendment, and what it tries to fix

The Inter-State River Water Disputes (Amendment) Act 2019 replaced the pattern of a Tribunal per dispute with a single standing Inter-State River Water Disputes Tribunal with multiple benches, and inserted a Disputes Resolution Committee stage before adjudication, intended to settle the dispute by negotiation within a fixed period. The stated aim is to cure the first two causes of delay: the standing Tribunal removes the setting-up time and preserves institutional memory, and the Committee formalises the negotiation the Act had always assumed.

Whether it works is a fair question for an answer to raise rather than pretend to settle. The causes of delay that the amendment does not touch, the politics, the science, the enforcement and the litigation around the section 11 bar, are the larger ones.

The disputes worth naming

A short list, with what each illustrates.

The Cauvery dispute, between Karnataka, Tamil Nadu, Kerala and Puducherry. Referred to a Tribunal in 1990, final award 2007, and the Supreme Court delivered a judgment modifying the allocation in 2018 and directing the constitution of a scheme to implement it. It illustrates every cause of delay in the list above, and also the point that the section 11 bar did not keep the Supreme Court out.

The Mullaperiyar dispute, between Kerala and Tamil Nadu, about a dam built in 1895 and the safe level of its water. It illustrates the point that these are not only allocation disputes: safety, maintenance and old agreements produce the same deadlock.

The Ravi and Beas dispute, between Punjab, Haryana and Rajasthan, in which a State legislature enacted a law purporting to terminate its water-sharing agreements, and a Presidential Reference under article 143 followed. It illustrates the interaction between this subject and [Advisory Jurisdiction].

The Krishna and Godavari disputes, which had to be revisited after the creation of Telangana in 2014, illustrating that reorganisation under article 3 creates new water disputes as a by-product.

The judgments in these matters are in the case bank's queue and have not been read, so no citation is given for any of them here and none should be invented. What is stated above about each dispute is the procedural history, not a holding.

The constitutional argument to make

An answer that only lists the causes of delay is describing a problem. The federalism point is this.

Article 262 is the only place in the Constitution where an entire class of dispute between States may be removed from the Supreme Court. Article 131 gives the Court exclusive original jurisdiction over disputes between the Union and States and between States, and article 262(2) carves this one class out of it.

The justification is that these disputes are technical, polycentric and continuing, and that a court deciding a lis between two parties on the evidence before it is a poor instrument for allocating a shared resource over decades. That justification is respectable.

The cost is that the class of dispute most likely to set State against State, and most likely to need an authoritative and final answer, is the one class for which the Constitution's own final authority is unavailable. Whether that trade was worth making is the question, and the honest answer is that the machinery article 262 substituted has not delivered the speed that was its whole justification.

A worked example

States V and W share a river. V, the upper riparian, builds a barrage that reduces the flow to W. W asks the Central Government to constitute a Tribunal; nothing happens for two years. W then files a suit in the Supreme Court under article 131, and separately a writ petition under article 32 alleging that its residents' right to water under article 21 is being violated.

The article 131 suit. Section 11 of the Inter-State River Water Disputes Act 1956 provides that neither the Supreme Court nor any other court shall exercise jurisdiction in respect of a water dispute which may be referred to a Tribunal under the Act. Parliament enacted that under article 262(2). So far as the suit asks the Court to allocate the water, it is barred.

What is not barred. The bar is on the water dispute. It does not prevent a challenge to the constitution of a Tribunal, or to the validity of the Act, or a proceeding about the interpretation or enforcement of an award, or a claim that the Central Government has failed to perform a statutory duty. W's real complaint, that the Union has not constituted a Tribunal within the time the Act allows, is of that kind, and a direction to perform the statutory duty is available.

The article 32 petition. A State is not a "person" for article 32, so W cannot bring it in its own right. Its residents could, but the relief they seek would in substance be an allocation of water, and the bar cannot be circumvented by changing the petitioner.

What the 2019 amendment changes. A single standing Tribunal removes the delay in constituting one, and a Disputes Resolution Committee must attempt settlement first. Neither cures the science, the politics or the enforcement.

The point to make. The section 11 bar has not removed litigation; it has redirected it, and added a preliminary jurisdictional argument to every case.

Quick revision

• Entry 17 List II gives water to the States, subject to Entry 56 List I on inter-State rivers, which needs a Parliamentary declaration.
• Article 262(1) lets Parliament provide for adjudication; article 262(2) lets Parliament oust the courts.
• The Inter-State River Water Disputes Act 1956: section 4 constitutes a Tribunal, section 6 makes the award binding with the force of a Supreme Court decree, section 11 bars the courts.
• The 2002 amendment added time limits and the decree equivalence; the 2019 amendment created a single standing Tribunal and a Disputes Resolution Committee.
• The causes of delay: discretionary reference, a new Tribunal each time, contested science, unenforced deadlines, the section 5(3) further reference, litigation around the section 11 bar, unresolvable politics, and weak enforcement.
• The federalism point: this is the one class of inter-State dispute the Constitution allows to be taken away from the Supreme Court, and the substitute has been slower than the thing it replaced.

Test yourself

1. Set out the three provisions that create the problem: Entry 17 of List II, Entry 56 of List I and article 262.

2. What does section 11 of the Inter-State River Water Disputes Act 1956 bar, and what does it not bar?

3. Give six reasons why these disputes take decades, and say which of them the 2019 amendment addresses.

4. What did the 2002 amendment change about the status of a Tribunal's award?

5. "Article 262 removes from the Supreme Court the one class of inter-State dispute most likely to need a final answer." Discuss.

When the party system fragmented, that informal machinery stopped working, and the demand for a constitutional forum revived. It was the Sarkaria Commission that pressed for it, and the Council was constituted by a Presidential order of 28 May 1990.

The Council as constituted

Composition. The Prime Minister as chairman; the Chief Ministers of all States; the Chief Ministers of Union territories having a legislative assembly and the Administrators of those that do not; and six Union Ministers of Cabinet rank nominated by the Prime Minister. A Standing Committee, chaired by the Union Home Minister, does the continuous work, and a secretariat supports it.

Function. The Presidential order assigned it the second and third of the article 263 duties, the deliberative and the co-ordinating, and not the first. So the Council as constituted does not inquire into and advise upon disputes between States. That omission is worth naming: the one function that would have given it real work was left out.

Practice. It is required to meet at least three times a year and has met far less often than that, with long gaps between meetings. Its recommendations are not binding and there is no mechanism to follow them up.

The Zonal Councils, which are not constitutional

Students confuse these with the Inter-State Council and they should not.

The Zonal Councils were created by the States Reorganisation Act 1956, an ordinary statute, not by the Constitution. There are five, with a sixth body, the North Eastern Council, created by a separate Act in 1971. Each groups the States of a region under the chairmanship of the Union Home Minister, with the Chief Ministers as members, and provides a regional forum for common problems.

The distinction to draw: the Inter-State Council is constitutional and national; the Zonal Councils are statutory and regional. Both are advisory.

The Commissions on Centre and State relations

Three inquiries have examined the federal relationship, and an answer on federalism is stronger for naming them and knowing what each recommended.

The Rajamannar Committee, appointed by the Government of Tamil Nadu in 1969, was a State initiative and reported in 1971. It recommended the immediate establishment of an Inter-State Council, the deletion of articles 356, 357 and 365, the transfer of the residuary power to the States, and the abolition of the Planning Commission in favour of a statutory body. The Union did not act on it. It is worth citing precisely because it is the States' own view of the arrangement.

The Sarkaria Commission, appointed by the Union in 1983 under a retired Supreme Court judge, reported in 1988. It examined the whole relationship and made recommendations across two hundred and forty-seven items. It did not recommend abolishing article 356; it recommended that it be used sparingly, as a measure of last resort, after all alternatives had failed, and that a warning be given to the State first. It recommended that the Governor be an eminent person from outside the State, not belonging to the ruling party, appointed after consultation with the Chief Minister. And it recommended the establishment of the Inter-State Council, which is what produced the 1990 order. Its recommendations on article 356 were substantially adopted by the Supreme Court in S.R. Bommai and are quoted in that judgment.

The Punchhi Commission, appointed in 2007 under a former Chief Justice, reported in 2010. Among its recommendations: that article 356 be used only in the affected area rather than the whole State where possible, a "localised emergency"; that the Governor's tenure of five years be protected against removal at pleasure except by a procedure with reasons; that the Governor's discretion in appointing a Chief Minister be guided by a stated order of preference; that a State's consent be required before central agencies investigate offences within it; and that the Inter-State Council be given a stronger role and meet regularly.

A note for the essay. The pattern across all three reports is the same. The problems identified are the Governor, article 356, the residuary power and the absence of a working consultative forum. Each has been examined and reported on repeatedly over fifty years, and the structural recommendations have largely not been implemented. What has changed the position is judicial decision, not constitutional amendment. That contrast is the point to make.

A worked example

A State government complains that centrally sponsored schemes are designed without consulting the States, and asks what constitutional machinery exists to raise the grievance.

The forum the Constitution provides. Article 263 permits the President to establish an Inter-State Council charged with investigating and discussing subjects in which the Union and one or more States have a common interest, and with making recommendations for the better co-ordination of policy and action. The design of a scheme operating in a State's own legislative field is squarely such a subject.

What the Council can and cannot do. It can discuss and recommend. It cannot decide, and its recommendations do not bind. It also meets rarely, and the Presidential order of 1990 did not assign it the first of the article 263 duties, advising on disputes between States.

What else is available. The Zonal Councils under the States Reorganisation Act 1956, which are statutory and regional, not constitutional. NITI Aayog, which is an executive body created by Cabinet resolution and allocates no funds. And, for a money grievance, the Finance Commission under article 280, though its terms of reference are set by the Union.

Why the grievance is hard to press. The schemes are funded under article 282, which permits either government to make a grant for any public purpose notwithstanding that the purpose is not one on which it may legislate. A State that wants the money accepts the design. There is no constitutional mechanism that compels consultation.

What has been recommended. Rajamannar, Sarkaria and Punchhi all pressed for a stronger consultative forum, and Punchhi specifically for a Council that meets regularly with a real role. None of the structural recommendations has been implemented.

Quick revision

• Article 263 lets the President establish an Inter-State Council with three possible duties: advising on inter-State disputes, discussing common interests, and recommending co-ordination.
• The Council exists at the Union's discretion and its output is advisory.
• Constituted only in 1990, on the Sarkaria recommendation. The dispute-advising function was not assigned to it.
• Composition: Prime Minister as chairman, all Chief Ministers, six Union Ministers, with a Standing Committee under the Home Minister.
• Zonal Councils are statutory and regional, under the States Reorganisation Act 1956; not to be confused with the constitutional Council.
• Rajamannar 1971, Sarkaria 1988, Punchhi 2010. Sarkaria's article 356 recommendations were adopted judicially in Bommai; most structural recommendations were never implemented.

Test yourself

1. What three duties may be assigned to a Council under article 263, and which of them was not assigned in 1990?

2. Why was the Inter-State Council not constituted until forty years after the Constitution came into force?

3. Distinguish the Inter-State Council from the Zonal Councils.

4. What did the Sarkaria Commission recommend about article 356 and about the Governor, and which of its recommendations were adopted, and by whom?

5. Name the three Commissions on Centre and State relations with their years, and state what the pattern across all three shows.

Article 265, and the rule of law in taxation

"No tax shall be levied or collected except by authority of law."

Three points make this more than a formality. Law means a statute, not an executive order, so a tax imposed by a government resolution is void. The whole of the levy must be authorised: the taxable event, the person liable, the rate and the measure must all be found in the law, and a tax cannot be saved by an authority's assurance that it is reasonable. And "levied or collected" covers both imposition and recovery, so money collected without authority is recoverable.

Article 265 is the taxation limb of the rule of law and belongs with [The Rule of Law] as well as here.

The three funds

The Consolidated Fund of India, and of each State, under article 266(1). All revenues received, all loans raised and all money received in repayment of loans go into it. No money may be appropriated out of it except in accordance with law and for the purposes and in the manner provided in the Constitution. This is the fund that the annual budget and the appropriation Act operate on.

The Public Account under article 266(2). Money received by or on behalf of the government that is not revenue: provident funds, small savings, deposits. The government holds it as a banker, and payments out of it do not need an appropriation by the legislature.

The Contingency Fund under article 267. Placed at the disposal of the President, or the Governor, to meet unforeseen expenditure pending authorisation by the legislature. It is an imprest and must be recouped.

Where the tax money goes

The pre-2016 scheme distinguished several categories, and the goods and services tax has since absorbed a great deal of it. The categories still in the text, and still examinable, are these.

Article 268. Duties levied by the Union but collected and appropriated by the States: stamp duties on the instruments listed in the Union List. The Union sets the rate; the State collects and keeps it.

Article 269. Taxes on the sale or purchase of goods and on the consignment of goods in the course of inter-State trade, levied and collected by the Union but assigned to the States. The proceeds do not form part of the Consolidated Fund of India.

Article 269A. The goods and services tax on inter-State supply, levied and collected by the Union and apportioned between the Union and the States. See [The Goods and Services Tax and Its Council].

Article 270. The general rule. All taxes and duties in the Union List, except those in articles 268, 269 and 269A, surcharges under article 271 and cesses for a specific purpose, are levied and collected by the Union and distributed between the Union and the States in the manner the Finance Commission recommends. This is the divisible pool, and it is the main channel.

Article 271. Parliament may at any time increase any of the duties or taxes in articles 269 and 270 by a surcharge for the purposes of the Union, and the whole proceeds of such a surcharge form part of the Consolidated Fund of India.

That last sentence is the most contested provision in the whole chapter. A surcharge, and equally a cess levied for a specific purpose, is outside the divisible pool, so none of it is shared with the States. A Union that raises revenue through surcharges and cesses rather than through the base rate of a shared tax increases its own resources without increasing anybody's share. That is the single most frequently made criticism of Indian fiscal federalism in the last decade, and an answer should make it in those terms.

Grants

Article 275. Statutory grants in aid of the revenues of States which Parliament determines to be in need of assistance, charged on the Consolidated Fund of India. Different sums may be fixed for different States. There are also specific grants for schemes of development for the Scheduled Tribes and Scheduled Areas.

Article 282. Discretionary grants. The Union or a State may make any grant for any public purpose, notwithstanding that the purpose is not one with respect to which the legislature concerned may make laws. This is a striking provision: it lets the Union spend on State subjects.

Article 282 is the constitutional basis of the centrally sponsored schemes, which is how a very large part of the Union's transfer to the States is actually made, and it is why the Union has effective influence over policy in fields the Seventh Schedule gives to the States. A State that wants the money accepts the scheme's design. Whether an article intended for occasional grants should have become a principal channel of transfer is a fair question to raise.

The Finance Commission

Article 280 requires the President to constitute a Finance Commission within two years of the commencement of the Constitution and every fifth year thereafter, or earlier if he considers it necessary.

Composition. A chairman and four other members, with qualifications determined by Parliament.

Duties. To make recommendations to the President on the distribution between the Union and the States of the net proceeds of taxes which are to be divided between them, and the allocation of the respective shares between the States; on the principles which should govern grants in aid of the revenues of the States out of the Consolidated Fund of India; on measures needed to augment the Consolidated Fund of a State to supplement the resources of the panchayats and municipalities on the basis of the recommendations of the State Finance Commission; and on any other matter referred to it by the President in the interests of sound finance.

Status of its recommendations. They are advisory. Article 281 requires the President to cause every recommendation, together with an explanatory memorandum as to the action taken on it, to be laid before each House of Parliament. The convention has been that the core recommendations on tax devolution are accepted, and the strength of the Commission rests on that convention rather than on any legal obligation.

Two features worth naming. The terms of reference are set by the Union, which shapes what the Commission can recommend. And the Commission's recommendations operate on the divisible pool, so revenue raised outside that pool by surcharge or cess is beyond its reach whatever it recommends.

The Planning Commission and NITI Aayog

Alongside the constitutional machinery there was, from 1950 to 2014, a parallel and entirely extra-constitutional channel. The Planning Commission was created by a Cabinet resolution, not by the Constitution or by statute, and it allocated plan assistance to the States. For much of that period the plan channel was comparable in size to the Finance Commission channel, and it operated on criteria the Union set.

It was replaced in 2015 by NITI Aayog, also created by a Cabinet resolution, which does not allocate funds at all. Its function is advisory and co-ordinating. The consequence is that the Finance Commission is now the principal formula-based channel, which strengthens the constitutional machinery, while the discretionary channel under article 282 remains.

The federalism argument

Bring the chapter together in one paragraph, which is what an examiner is looking for.

Fiscal federalism is where the formal division of powers meets the practical question of who can actually govern. A State with exclusive legislative competence over public health cannot deliver public health without money, and if the money comes from the Union on the Union's terms, the exclusivity in article 246 is worth less than it reads. The Constitution recognised the problem and answered it with an independent Finance Commission whose formula the Union has by convention accepted. The pressure on that answer comes from three directions: revenue raised outside the divisible pool by surcharge and cess, the growth of conditional transfers under article 282, and the fact that the Commission's terms of reference are written by one of the two parties whose shares it is dividing. All three are live and all three are proper material for an answer on whether India is a federation.

A worked example

The Union raises the rate of a cess levied for a specified purpose on a tax in the Union List, and reduces the base rate of the same tax by an equivalent amount. Total collections are unchanged. A State complains that its revenue has fallen sharply.

Why the State's revenue falls. Article 270 provides that taxes in the Union List are levied and collected by the Union and DISTRIBUTED between the Union and the States in the manner the Finance Commission recommends. That is the divisible pool. But article 270 expressly excludes surcharges under article 271 and cesses levied for a specific purpose. A rupee moved from the base rate to a cess leaves the divisible pool entirely.

Is it unconstitutional? No. Article 271 permits Parliament at any time to increase any of the duties or taxes referred to in articles 269 and 270 by a surcharge for the purposes of the Union, and the whole proceeds form part of the Consolidated Fund of India. A cess for a specified purpose is likewise outside the pool. The Union is exercising powers it plainly has.

What the State can and cannot do. It cannot challenge the levy for want of competence. It may argue that a cess collected for a specified purpose must actually be spent on that purpose, which is a real limit, and that a levy called a cess but functioning as general revenue is a colourable exercise. Whether that succeeds depends on the facts.

What the Finance Commission can do. Nothing directly: it operates on the divisible pool, so revenue raised outside the pool is beyond its reach whatever it recommends. It may comment, and Commissions have.

The point for an essay. This is the sharpest current criticism of Indian fiscal federalism, and it needs no allegation of bad faith: the structure permits it.

Quick revision

• Vertical imbalance is deliberate: the States spend, the Union collects, and Part XII moves the money.
• Article 265: no tax except by authority of law, and law means a statute.
• Consolidated Fund, Public Account, Contingency Fund: articles 266 and 267.
• Article 270 is the divisible pool; article 271 surcharges and specific-purpose cesses sit outside it and are not shared.
• Article 275 is statutory grants; article 282 is discretionary grants and is the basis of centrally sponsored schemes on State subjects.
• Article 280: a Finance Commission every fifth year, advisory, with terms of reference set by the Union, its recommendations laid before Parliament under article 281.
• The Planning Commission was never constitutional; NITI Aayog replaced it in 2015 and allocates nothing.

Test yourself

1. Explain the vertical imbalance and why it is deliberate.

2. What is the divisible pool, and what falls outside it? Why does that matter?

3. Distinguish articles 275 and 282, and explain why article 282 has become a principal channel of transfer.

4. How often must a Finance Commission be constituted, what must it recommend, and what is the status of its recommendations?

5. Was the Planning Commission a constitutional body? What replaced it, and what does the replacement do?

It displaces article 254. The repugnancy rule that would ordinarily resolve a conflict between a central and a State law on a shared subject does not apply either. So if a central GST law and a State GST law conflict, article 254 supplies no answer, and, as the Supreme Court observed, article 246A envisages no repugnancy provision at all.

Both levels hold the same power over the same tax. Not "the Union taxes this and the State taxes that", but both taxing the same supply, the Union by central GST and the State by State GST, simultaneously.

Except for inter-State supply, where clause (2) gives Parliament exclusive power. That is integrated GST, and article 269A governs what happens to the money.

Article 269A: the money on inter-State supply

Integrated GST is levied and collected by the Government of India, and apportioned between the Union and the States in the manner Parliament provides by law on the recommendations of the Council. The amount apportioned to a State does not form part of the Consolidated Fund of India.

Two consequences worth noting. Import of goods or services is treated as supply in the course of inter-State trade, so IGST applies to imports. And the apportionment is not a share of a divisible pool under article 270 but a distinct mechanism, so the Finance Commission's formula does not govern it.

Article 279A: the Council

Constitution. Within sixty days of the commencement of the amendment the President was to constitute the Council. Its members are the Union Finance Minister as chairperson, the Union Minister of State for revenue or finance, and the minister in charge of finance or taxation, or any other minister nominated, from each State. The members from the States choose one of themselves as vice-chairperson.

Functions. To make recommendations on the taxes, cesses and surcharges to be subsumed; the goods and services to be subjected to or exempted from the tax; model GST laws, principles of levy, apportionment of IGST and the principles governing place of supply; the threshold turnover below which goods and services may be exempted; the rates including floor rates with bands; special rates for a specified period to raise resources during a natural calamity; special provision for certain States; and any other matter the Council decides.

Quorum and voting. This is the part to learn precisely, because it is the constitutional expression of the bargain.

• The quorum is one half of the total number of members.
• Every decision must be taken by a majority of not less than three-fourths of the weighted votes of the members present and voting.
• The vote of the Central Government has a weightage of one-third of the total votes cast.
• The votes of all the State Governments taken together have a weightage of two-thirds of the total votes cast.

Work out what that means. The Union alone cannot carry a decision, because one-third is less than three-fourths. The States alone cannot carry a decision, because two-thirds is less than three-fourths. Neither side can act without some of the other, and the Union holds a blocking third. That is a genuine federal bargain written into the Constitution, and it is unique in the Indian text.

Article 279A(11) requires the Council to establish a mechanism to adjudicate disputes between the Union and States, or among States, arising out of its recommendations.

Are the recommendations binding?

This was the central constitutional question and it has been answered.

Facts. Union of India v. Mohit Minerals Pvt Ltd, decided 19 May 2022. The issue was whether integrated GST could be levied on the ocean freight component of an import on cost, insurance and freight terms, on a reverse charge basis. The Union argued in support of the levy partly on the footing that the Council had recommended it and that its recommendations bind.

Held. The recommendations of the GST Council are not binding on the Union and the States. The Court gave three reasons. The Constitution Amendment Act as introduced had contained an article 279B providing for a dispute-settlement authority, and that provision was deleted while article 279A(11) was included, which indicates that Parliament intended the recommendations to have only persuasive value. Article 279A does not begin with a non obstante clause, and article 246A is not expressed to be subject to it; both legislatures possess simultaneous power, and article 246A envisages no repugnancy provision, so treating the recommendations as binding edicts would disrupt fiscal federalism. And it is not necessary that one federal unit must always hold the higher share of power for the units to be able to decide.

The Court added the qualification that matters in practice: the Government is bound by the recommendations when exercising its rule-making power under the CGST and IGST Acts, but that does not make every recommendation binding on the legislature's power to enact primary legislation.

It also gave the sentence most worth quoting in a federalism answer:

Indian federalism is a dialogue between cooperative and uncooperative federalism where the federal units are at liberty to use different means of persuasion ranging from collaboration to contestation.

Why it matters. It settles the status of the Council and it reframes the vocabulary. Cooperative federalism is not a duty to agree; it is a structure that makes agreement necessary and leaves disagreement lawful.

The federalism argument, both ways

For the arrangement. It is the first genuine instance of shared sovereignty in Indian fiscal arrangements. The States gave up their most important tax and got in exchange a seat at a table where the Union cannot act alone. The weighted voting is a real protection, not a formality, and the Council has in fact functioned by consensus for most of its life.

Against it. The States surrendered a power they held exclusively and received in return a vote in a body whose output is not binding, in which the Union holds a blocking third that no combination of States can overcome. Their remaining independent revenue is narrow: land, buildings, alcohol, petroleum products for the time being, electricity and vehicles. The compensation that cushioned the transition was time-limited. And because article 246A displaces article 254, there is no constitutional rule to resolve a real conflict between a central and a State GST law, only politics.

The honest conclusion, and the one to write, is that GST made India a single market at the cost of making the States fiscally dependent on a forum rather than on a power, and that the Supreme Court, by holding the recommendations non-binding, gave the States back the one thing the design had taken from them: the ability to say no.

A worked example

The GST Council recommends by the required majority that a particular service be taxed at a specified rate. State Y disagrees, and its legislature enacts a State GST law taxing that service at a lower rate. The Union says the recommendation binds.

Does article 246A permit the State law? Yes. Article 246A confers on Parliament and on the legislature of every State power to make laws with respect to goods and services tax, and the supply here is intra-State, so the State legislature has the power.

Does the recommendation bind? No. Mohit Minerals holds that the recommendations of the GST Council are not binding on the Union and the States. The reasons are that the proposed article 279B was deleted while 279A(11) was included, indicating persuasive value only; that article 279A has no non obstante clause and article 246A is not expressed to be subject to it; and that treating the recommendations as binding edicts would disrupt fiscal federalism.

Is the State law repugnant to the central GST law? Article 246A opens "Notwithstanding anything contained in articles 246 and 254", so the ordinary repugnancy rule is displaced, and article 246A envisages no repugnancy provision at all. There is no constitutional rule that resolves the conflict.

So what actually resolves it? Politics and the Council, not law. That is what the Court meant by describing Indian federalism as a dialogue between cooperative and uncooperative federalism, with the units at liberty to use means of persuasion ranging from collaboration to contestation.

One qualification. The Government IS bound by the recommendations when exercising its rule-making power under the CGST and IGST Acts. The freedom is the legislature's, not the executive's.

Quick revision

• Article 246A grants GST power directly, notwithstanding articles 246 and 254; it is not an entry in any List.
• Both Parliament and every State legislature hold the power simultaneously; Parliament alone for inter-State supply.
• Article 269A: IGST is levied and collected by the Union and apportioned; imports count as inter-State supply.
• Article 279A: the Council, chaired by the Union Finance Minister, quorum one half, decisions by three-fourths of weighted votes, Union one-third and all States together two-thirds.
• Neither side can carry a decision alone, and the Union holds a blocking third.
• Mohit Minerals: the recommendations are persuasive, not binding, though the Government is bound by them in its rule-making. Indian federalism is a dialogue between cooperative and uncooperative federalism.

Test yourself

1. Why could neither the Union nor the States tax the whole of a commercial transaction before 2016?

2. What do the opening words of article 246A displace, and what follows for repugnancy between a central and a State GST law?

3. Set out the quorum and voting rule in article 279A, and show by arithmetic why neither side can carry a decision alone.

4. Are the recommendations of the GST Council binding? Give the three reasons in Mohit Minerals, and the one qualification.

5. "The States surrendered a power and received a vote." Discuss.

Article 303(2) is the one exception: Parliament may do so if it declares by law that it is necessary for the purpose of dealing with a situation arising from scarcity of goods in any part of India. Note that no such power is given to a State.

Article 304(a) permits a State to tax goods imported from other States, but only so as not to discriminate: the tax must be one to which similar goods manufactured or produced in that State are subject. The test is equality of treatment, not the absence of a tax.

Article 304(b) permits a State to impose reasonable restrictions on the freedom in the public interest, and attaches a procedural condition: no Bill or amendment for that purpose may be introduced or moved in the State legislature without the previous sanction of the President.

Article 305 saves existing laws and laws providing for State monopolies. Article 306 is spent. Article 307 allows Parliament to appoint an authority to carry out the purposes of articles 301 to 304, and Parliament has never done so.

The compensatory tax doctrine, and its end

For fifty years the central question was whether a tax is a restriction on the freedom at all, and the answer was a doctrine the Constitution does not mention.

The problem. If every tax that made goods more expensive to move were a restriction under article 301, then almost every State tax on transport or entry would need the President's previous sanction under article 304(b), and States would be unable to fund the roads and bridges the trade actually uses.

The doctrine. The courts developed the idea of a compensatory tax: a levy which is really a charge for the use of trading facilities provided by the State, such as roads, is not a restriction on trade but a payment for what makes the trade possible, and therefore falls outside article 301 altogether. A compensatory tax needed no presidential sanction.

Its development. The doctrine emerged in Atiabari Tea Co Ltd v. State of Assam (1961), where a tax on goods carried by inland waterways was struck down as a direct and immediate restriction on movement, and was refined the following year in Automobile Transport (Rajasthan) Ltd v. State of Rajasthan (1962), which added the compensatory tax exception. For half a century the argument in every case was whether a particular levy was compensatory, and the tests for that grew increasingly elaborate.

Its abandonment. In Jindal Stainless Ltd v. State of Haryana (2016) a nine-judge Bench held that the compensatory tax doctrine has no basis in the Constitution and abandoned it. The reasoning, in outline, was that Part XIII contains no such exception, that the doctrine had been imported from an Australian and American context that the Indian text does not share, and that it had become unworkable. The Court held that a tax simpliciter is not a restriction on the freedom under article 301, and that what article 304(a) requires is non-discrimination: a State may tax goods from other States provided it taxes its own like goods in the same way.

The judgments in all three cases are in the case bank's queue and have not been read, so no citation is printed for any of them and none should be invented. What is described above is the doctrinal history, and a student writing on it should confirm the citations against a report.

The position now

Stated as a working test, which is what a problem question needs.

Is the impugned measure a tax? If it is, it is not by that fact alone a restriction under article 301. The question becomes whether it discriminates.

Does it discriminate against goods from other States? Compare the treatment of the imported goods with that of like goods produced within the State. If the imported goods bear a heavier burden, article 304(a) is violated and the levy is bad, and no presidential sanction can save it, because 304(a) is not subject to that proviso.

Is the measure a non-fiscal restriction? A licensing requirement, a quota, a checkpost delay, a prohibition on movement. Then article 301 is engaged directly, and the State needs article 304(b): a reasonable restriction in the public interest, with the previous sanction of the President, or subsequent assent, which article 255 permits to cure the want of previous sanction.

Is it Parliament that has acted? Then article 302 supplies the power, subject to the anti-discrimination rule in article 303(1) and its scarcity exception in 303(2).

What Part XIII means for federalism

Two observations belong in an essay.

It is a limit on the States that has no equivalent limit on the Union. Parliament may restrict the freedom in the public interest under article 302 with no procedural condition; a State needs the President's previous sanction under article 304(b). Article 303(1) does bind both, but only article 303(2) provides an escape, and only for Parliament. The asymmetry is deliberate and it is another instance of the general tilt.

Its practical importance has fallen sharply since 2016. The goods and services tax subsumed entry tax, octroi, central sales tax and most of the levies that generated Part XIII litigation, and abolished the check posts that were the visible form of the problem. What Part XIII was trying to achieve by prohibition, article 246A has largely achieved by unification. That is worth saying, because it is an example of a constitutional problem solved by redesign rather than by adjudication.

A worked example

State Z levies an entry tax on goods brought into the State for consumption, at a rate of two per cent. Goods manufactured within the State are subject to a local levy at one per cent. The Bill was passed without the President's previous sanction. Separately, State Z requires every vehicle carrying goods into the State to obtain a permit, for which it charges a fee said to cover the cost of maintaining the roads used.

The entry tax. Since Jindal Stainless a tax simpliciter is not a restriction under article 301, so the question is not whether the tax burdens trade but whether it DISCRIMINATES. Article 304(a) permits a State to tax goods imported from other States only so as not to discriminate, that is, only a tax to which similar goods produced in the State are subject. Two per cent against one per cent is discriminatory on its face, and the tax fails. Note that article 304(a) is not subject to the previous-sanction proviso, so no assent could cure it.

The permit requirement. This is not a tax but a non-fiscal restriction on movement, so article 301 is engaged directly. The State needs article 304(b): a reasonable restriction in the public interest, with the previous sanction of the President. That sanction was not obtained.

Can the want of sanction be cured? Article 255 provides that an Act is not invalid for want of a recommendation or previous sanction if the necessary assent was in fact given. So if the Bill was reserved and received the President's assent, the defect is cured; if it was simply passed and assented to by the Governor, it is not.

What the compensatory tax doctrine would have done. Before 2016 the State would have argued that the permit fee was compensatory, a charge for the trading facility of the roads, and therefore outside article 301 altogether. Jindal Stainless abandoned that doctrine, so the argument is no longer available.

Quick revision

• Article 301: trade, commerce and intercourse throughout the territory of India shall be free, subject to the rest of Part XIII.
• Article 302: Parliament may restrict in the public interest. Article 303(1): no preference or discrimination between States by either legislature, with a scarcity exception in 303(2) available only to Parliament.
• Article 304(a): a State may tax imported goods only so as not to discriminate against them. Article 304(b): reasonable restrictions in the public interest, with the previous sanction of the President.
• The compensatory tax doctrine, from Atiabari and Automobile Transport, held that a charge for trading facilities fell outside article 301. Jindal Stainless (2016), nine judges, abandoned it.
• The current test is non-discrimination: a tax simpliciter is not a restriction, and article 304(a) asks whether like goods produced in the State bear the same burden.
• GST has taken away most of the litigation by removing the taxes that caused it.

Test yourself

1. How does article 301 differ from section 92 of the Australian Constitution in its wording, and what follows?

2. Distinguish the powers of Parliament under article 302 from those of a State under article 304, and identify the asymmetry.

3. What was the compensatory tax doctrine, in which cases did it develop, and on what reasoning was it abandoned?

4. State the current working test for a fiscal and for a non-fiscal measure under Part XIII.

5. Why has the practical importance of Part XIII fallen since 2016?

Article 163 and the extent of discretion

Facts. Nabam Rebia v. Deputy Speaker, Arunachal Pradesh Legislative Assembly, decided 13 July 2016 by five judges. During a political crisis the Governor advanced the session of the Legislative Assembly by his own order, and fixed its agenda to include a resolution for the removal of the Speaker, acting without the advice of the Council of Ministers and against the advice of the Chief Minister.

Held. Under article 163(1) the Governor is bound by the advice of his Council of Ministers. There are only three exceptions, and they are introduced by the words "except in so far as": where he may act in his discretion as conferred by the Constitution; where he may act in his discretion as conferred under the Constitution; and where he may act in his individual judgment in instances specified by the Constitution. A further narrow category has evolved by constitutional necessity in extraordinary situations, such as those arising under article 356, or in choosing a person to lead the Assembly as Chief Minister. Those limits do not prevent the Assembly framing rules under article 208 or the Governor framing rules of business under article 166.

Why it matters. It is the modern statement that discretion is the exception and advice the rule, and that the exceptions are textual and closed rather than a general reserve of power. Learn the three limbs in the Court's own words.

The functions in which discretion is admitted

Gathered from the text and the case law, and worth listing in an answer:

• Reserving a Bill for the consideration of the President under article 200, on which see below, because the position has just changed.
• Recommending a Proclamation under article 356.
• Appointing a Chief Minister where no party has a clear majority, or where the incumbent has lost it.
• Dismissing a Council of Ministers that has lost the confidence of the House and will not resign.
• Dissolving the Assembly where the Council of Ministers has lost its majority.
• Seeking information from the Chief Minister under article 167.
• The special responsibilities conferred by article 371 on the Governors of certain States, dealt with in [Asymmetric Federalism: Articles 371 to 371-J].

Article 200: what the Governor may do with a Bill

The article gives three courses and one procedure.

Assent. The Bill becomes an Act.

Withhold assent. The Bill does not become an Act.

Reserve the Bill for the consideration of the President. Article 201 then applies: the President may assent or withhold assent, and may direct the Governor to return the Bill to the House, which must reconsider it within six months, after which it is presented to the President again, who is not obliged to assent.

The first proviso. Where the Governor withholds assent to a Bill that is not a Money Bill, he may, as soon as possible after presentation, return it with a message requesting the House to reconsider it or any specified provisions, and to consider the desirability of introducing recommended amendments. If the House passes it again, with or without amendment, and it is presented again, the Governor shall not withhold assent therefrom.

The second proviso. The Governor shall not assent to, but shall reserve for the President, a Bill which in his opinion would, if it became law, so derogate from the powers of the High Court as to endanger the position that Court is designed to fill. This is the one case of compulsory reservation.

The gap in the article is obvious and it is what the litigation is about: no time limit is stated anywhere. A Governor who simply does nothing is not, on the face of the text, doing anything the article forbids.

The two decisions, and which one governs

This is the part that has changed, and an answer must have both and in the right order.

Facts. State of Tamil Nadu v. Governor of Tamil Nadu, 2025 INSC 481, decided 8 April 2025. Ten Bills passed by the Tamil Nadu Assembly were pending with the Governor for long periods. After the Court's decision in the Punjab Governor's case, he reserved all ten for the President, including Bills already returned once and repassed.

Held. Article 200 gives three options; the first proviso is attached to withholding assent and is not a fourth course; neither a pocket veto nor an absolute veto exists, because "shall declare" leaves no room for inaction and "as soon as possible" imports urgency. As a general rule the Governor may not reserve a Bill for the President in the second round after it has been returned and repassed, unless the repassed Bill is materially different. The reservation of the ten Bills was illegal and was set aside. The Court held that it could prescribe timelines and laid down maxima of one month and three months for the various situations, and it declared the ten Bills deemed to have been assented to on the date they were presented after reconsideration.

Why it matters. It was the furthest a court had gone in controlling the office, and it is the decision that provoked the Reference.

Facts. In re Assent, Withholding or Reservation of Bills by the Governor and the President, advisory opinion under article 143 pronounced on 20 November 2025 by five judges, Gavai C.J. with Surya Kant, Vikram Nath, Narasimha and Chandurkar JJ. The President referred fourteen questions about articles 200 and 201 and about whether a court may fix time limits.

Held. Summarised by the Court itself at paragraph 165. The Governor has the three options and the first proviso restricts rather than adds to them, the third option being unavailable for a Money Bill. The Governor enjoys discretion in choosing among the three and is not bound by the aid and advice of the Council of Ministers when exercising the article 200 function. The discharge of that function is not justiciable and there can be no merits review; but where inaction is prolonged, unexplained and indefinite, the Court may issue a limited mandamus to act within a reasonable time, without commenting on the merits. Article 361 bars proceedings against the Governor personally, but the constitutional office remains subject to the Court's jurisdiction. It is not appropriate for the Court to prescribe timelines under article 200, and the same applies to the President under article 201, whose assent is likewise not justiciable. Neither decision is justiciable before the Bill becomes law. And article 142 does not permit the concept of deemed assent.

Why it matters. This is the current law. It leaves standing the proposition that the Governor cannot sit on a Bill for ever, and it removes the two remedies the April judgment had supplied. The result is a duty with a much weaker sanction: a court may order the Governor to decide, and may not tell him what to decide, may not fix how long he has, and may not deem the Bill assented to.

How to write this in an answer. State the position under article 200; state the April 2025 decision; then state that the Presidential Reference of 20 November 2025 held that timelines may not be judicially prescribed and that there is no deemed assent, and that discretion under article 200 is not controlled by ministerial advice. Do not present the April 2025 timelines as the law. An answer that gives only the earlier case is out of date by a year.

What is left of the control

Three things survive and should be stated as the residue.

Inaction is not a lawful option. Article 200 says "shall declare", and the Court has not resiled from that.

A limited mandamus is available where inaction is prolonged, unexplained and indefinite. It compels a decision and says nothing about which decision.

The office, though not the person, is subject to the Court's jurisdiction. Article 361's immunity is personal.

What has gone is the timetable and the self-executing remedy. Whether that leaves enough is the argument, and both sides of it are respectable. Against the Reference: a duty to act with no deadline and no consequence is the pocket veto returning by another name. For it: fixing a period the Constitution does not fix is amendment by judgment, and the Court said so.

The reform proposals

The recurring recommendations, all of which are still recommendations:

Sarkaria (1988). The Governor should be an eminent person from outside the State, not belonging to the party in power at the Centre, appointed after consultation with the Chief Minister, and should not be removed before the end of the term except for compelling reasons.

Punchhi (2010). The doctrine of pleasure should be qualified so that removal requires a procedure and reasons; the five-year tenure should be protected; and the Governor's discretion in appointing a Chief Minister should follow a stated order of preference.

Neither has been implemented. This is the point at which the chapter connects to MU's phrase about trust and faith: the office was designed on the assumption that the Union would appoint people who would behave as constitutional heads, and the Constitution supplied almost no rule for the case where that assumption fails.

A worked example

The Legislative Assembly of State A passes a Bill and presents it to the Governor in January. The Governor does nothing for fourteen months. In March of the following year, the State moves the Supreme Court seeking a direction that the Governor decide, a declaration that the Bill is deemed assented to, and a timetable for the future.

Is inaction lawful? No. Article 200 says the Governor "shall declare" one of three things, and the Court has held that this leaves no room for inaction and that neither a pocket veto nor an absolute veto exists.

Is the Bill deemed assented to? No, and this is where the law changed. The April 2025 decision did deem ten Bills assented. The Presidential Reference of 20 November 2025 held that article 142 does not permit the concept of deemed assent. So that relief is unavailable.

Will the Court fix a timetable? No. The Reference held that in the absence of constitutionally prescribed time limits it is not appropriate for the Court to prescribe timelines under article 200, and that the same applies to the President under article 201.

What relief IS available? Where the inaction is prolonged, unexplained and indefinite, the Court may issue a LIMITED MANDAMUS requiring the Governor to discharge his function under article 200 within a reasonable time, without any observation on the merits of the decision he then makes. Fourteen months, unexplained, is capable of meeting that description.

Can the Court review which of the three options he chooses? No. The Reference held the function is not justiciable and there can be no merits review, and that the Governor enjoys discretion in choosing among the three and is not bound by ministerial advice when exercising it.

Answer. A direction to decide, and nothing more.

Quick revision

• Appointed by the Union, removable at pleasure, bound by the advice of a State Council of Ministers. That is the structural problem.
• Nabam Rebia: advice is the rule, discretion the exception, and there are exactly three textual exceptions plus a narrow category of constitutional necessity.
• Article 200: assent, withhold, or reserve; the first proviso requires return "as soon as possible" and forbids withholding after repassage; the second proviso compels reservation where a Bill would endanger the High Court.
• April 2025: no pocket or absolute veto, timelines of one and three months, ten Bills deemed assented.
• 20 November 2025, and this governs: discretion under article 200 is the Governor's and is not controlled by ministerial advice; the function is not justiciable on the merits; no judicially prescribed timelines; no deemed assent; but a limited mandamus lies against prolonged, unexplained and indefinite inaction.
• Sarkaria and Punchhi both proposed protecting the office from the Centre. Neither was implemented.

Test yourself

1. Identify the four structural facts about the office of Governor that create the constitutional problem.

2. State the three exceptions to the rule in article 163(1), in the words used in Nabam Rebia.

3. Set out the three courses open under article 200 and explain what the first and second provisos each do.

4. What did the April 2025 decision hold, and which parts of it no longer stand after 20 November 2025?

5. What survives of judicial control over a Governor who does nothing, and is it enough? Argue both sides.

Article 250. Parliament may make laws for the whole or any part of India with respect to any matter in the State List. See [When Parliament May Legislate on a State Subject].

Article 354. The President may by order direct that the provisions of articles 268 to 279, the financial distribution provisions, shall have effect subject to such exceptions or modifications as he thinks fit. So the money can be redirected too.

Article 83(2) and article 172(1) provisos. The life of the House of the People and of a State Assembly may be extended by Parliament for one year at a time while the Proclamation is in force.

Put together, for the duration the Union may direct the State executive on anything, legislate on any State subject, and alter the financial settlement. The standard sentence is that the Constitution becomes unitary in form for the duration of the emergency, and it is accurate.

Article 355: the duty, and the argument about it

Article 355 is short and it has grown in importance.

It shall be the duty of the Union to protect every State against external aggression and internal disturbance and to ensure that the government of every State is carried on in accordance with the provisions of this Constitution.

Two readings compete, and both are examinable.

The narrow reading. Article 355 is a preamble to article 356. It states the duty that article 356 exists to discharge, and confers no independent power. On this view the Union's only remedy for a failure in a State is a Proclamation.

The wider reading. Article 355 imposes a positive duty, and a duty carries with it the means of performing it. On this view the Union may take measures short of dismissal, deploying central forces or issuing directions, to protect a State against internal disturbance, without proclaiming under article 356.

The wider reading has attractions, because it offers something less drastic than dismissing an elected government, and the Punchhi Commission recommended a "localised emergency" confined to the affected area, which builds on it. It also has an obvious danger: an article that authorises central intervention without any of article 356's safeguards, without a Proclamation, without parliamentary approval and without the review that Bommai established. That danger is the reason to be careful with it, and an answer should say so.

Articles 358 and 359, and the Forty-fourth Amendment

The rights position during an emergency is where the Constitution was most changed by experience.

Article 358 as originally enacted. While a Proclamation of Emergency was in operation, article 19 was suspended: nothing in it restricted the State's power to make any law or take any executive action, and the State could act in ways article 19 would otherwise forbid.

Article 359 as originally enacted. The President could by order suspend the right to move any court for the enforcement of such of the rights in Part III as were specified, for the period of the Proclamation. This suspends the remedy rather than the right, but with the remedy gone the distinction is thin.

What happened. During the emergency proclaimed in 1975, orders under article 359 suspended the right to move any court for the enforcement of articles 14, 21 and 22, and the Supreme Court held by majority that in consequence no person had any locus to move a High Court under article 226 for habeas corpus, even to challenge a detention alleged to be unlawful or mala fide. The dissent of Khanna J. is the most celebrated dissent in Indian constitutional history. The decision itself, and the constitutional position it produced, is dealt with in [Preventive Detention] and in [Article 32 and the Writs].

The Forty-fourth Amendment, 1978. The changes it made are the answer to any question about safeguards, and they should be listed.

• "Internal disturbance" as a ground for a Proclamation under article 352 was replaced by "armed rebellion", so that a domestic political crisis can no longer found a national emergency.
• A Proclamation may be issued only on the written recommendation of the Cabinet, not on the Prime Minister's advice alone.
• Parliamentary approval must be by a special majority and within one month, instead of a simple majority within two.
• The Proclamation continues for six months at a time and needs renewed approval.
• One-tenth of the members of the House of the People may requisition a special sitting to disapprove the Proclamation, and a simple majority then revokes it.
• Article 358 was confined to a Proclamation on the ground of war or external aggression, so article 19 is not suspended in an armed rebellion emergency; and it now applies only to laws which recite that they are in relation to the emergency.
• Articles 20 and 21 were excluded from article 359 altogether. No order may suspend the right to move a court for the enforcement of the protections of an accused or of the right to life and personal liberty. This is the single most important change the amendment made.

Article 360: the financial emergency

Where the President is satisfied that a situation has arisen whereby the financial stability or credit of India or of any part of it is threatened, he may make a Proclamation. During its operation the Union may direct a State to observe canons of financial propriety, may direct the reduction of the salaries of persons serving a State including judges of the High Courts, and may require all Money Bills of a State to be reserved for the President.

It has never been proclaimed. It belongs in an answer as the third emergency and as an illustration of how far the text allows the Union to reach into a State.

Emergency, federalism and the basic structure

Bring it together with the case.

Facts. S.R. Bommai v. Union of India, AIR 1994 SC 1918, (1994) 3 SCC 1, nine judges, decided 11 March 1994. Proclamations under article 356 dismissing six State governments.

Held. Among the Court's twelve summarised conclusions: the Constitution has created a federation but with a bias in favour of the Centre, and within the sphere allotted to them the States are supreme; the power under article 356 is a conditioned power resting on relevant material and is not immune from judicial review; and secularism is a basic feature.

Why it matters here. It establishes that the emergency provisions are not a licence. They are powers with conditions, and the conditions are enforceable. That is what distinguishes the position now from the position in 1975, when the same text was read as putting the Union's satisfaction beyond challenge.

The federalism argument to make is this. The emergency provisions were included because the drafters, writing in the year after partition, believed a country of this size and diversity could not survive without a power to centralise in a crisis. That judgment was defensible. What the drafters did not provide was any mechanism to decide whether the crisis was real. The 1975 emergency demonstrated the consequence, and the answer came from two directions: the Forty-fourth Amendment tightened the text, and Bommai supplied the review. Neither removed the powers. India remains a federation whose federal character can be suspended, with the difference that suspending it is now a decision somebody can be made to justify.

A worked example

A Proclamation of Emergency is issued on the ground of armed rebellion in one State. Parliament then enacts a law regulating agricultural marketing throughout India, a State subject. The Union also directs the government of a different State, where there is no disturbance at all, on the manner in which it is to exercise its executive power. A person detained under a law made during the emergency challenges the detention as violating articles 19 and 21.

The law on a State subject. Article 250 permits Parliament to legislate on any matter in the State List while a Proclamation is in operation, and it applies to the whole or any part of India, not only the disturbed State. The law is valid, and it ceases to have effect six months after the Proclamation ends. The State's own power survives, and under article 251 a repugnant State law is inoperative rather than void, reviving when the central law goes.

The direction to the undisturbed State. Article 353(a) provides that while the Proclamation is in operation the executive power of the Union extends to giving directions to a State as to the manner in which its executive power is to be exercised, and it is not confined to the State where the disturbance is. The direction is valid.

Article 19. After the Forty-fourth Amendment, article 358 applies only to a Proclamation on the ground of WAR or EXTERNAL AGGRESSION. This Proclamation is on the ground of armed rebellion, so article 19 is not suspended and the challenge on that ground is available.

Article 21. Article 359 permits the President to suspend the right to move a court for the enforcement of specified rights, but the Forty-fourth Amendment excluded articles 20 and 21 from it altogether. So no order can take away the right to move a court under article 21, and the detention may be challenged.

The lesson. Almost every element of this answer is the Forty-fourth Amendment responding to 1975.

Quick revision

• Three emergencies: article 352 national, article 356 State, article 360 financial. The third has never been used.
• During an article 352 emergency: article 353 directions to the State executive and wider legislative power, article 250 legislation on any State subject, article 354 alteration of the financial provisions, and extension of the life of legislatures.
• Article 355 is a duty; whether it is also an independent power is contested, and the wider reading has no safeguards attached.
• Article 358 suspends article 19; article 359 suspends the remedy for specified rights.
• The Forty-fourth Amendment: armed rebellion replaces internal disturbance, written Cabinet recommendation, special majority within a month, six-month renewals, requisition by one-tenth of the House, article 358 limited to war and external aggression, and articles 20 and 21 removed from article 359 entirely.
• Bommai: a federation with a bias in favour of the Centre, the States supreme in their own sphere, and the emergency power reviewable.

Test yourself

1. Name the three emergencies with their articles, and say which has never been used.

2. List five things that change in the federal structure while a Proclamation under article 352 is in operation.

3. Give the two competing readings of article 355 and the danger in the wider one.

4. Set out seven changes the Forty-fourth Amendment made to the emergency provisions, and identify the most important.

5. Distinguish article 358 from article 359 in what each suspends.

Article 365 connects Chapter II of Part XI to this article. Where a State has failed to comply with or give effect to a direction given by the Union under any provision of the Constitution, it shall be lawful for the President to hold that a situation has arisen in which the government of the State cannot be carried on in accordance with the Constitution. The word is permissive, and after Bommai the holding is reviewable like any other satisfaction.

The history, which is the reason the law changed

Two figures tell the story and both belong in an answer.

Between 1950 and 1994 article 356 was invoked over a hundred times. Very few of those were cases of the constitutional machinery actually failing. The recurrent patterns were these: dismissal of a State government of a different party after a change of government at the Centre; dismissal on the ground that the ruling party in the State had lost the general election to the House of the People; dismissal of a ministry which had not been tested on the floor of the House; and dismissal on a Governor's report asserting a loss of majority which the Chief Minister denied.

The Sarkaria Commission, reporting in 1988, examined the use of the article and recommended that it be used sparingly, as a measure of last resort, after all alternatives had failed and after a warning to the State. It also listed situations that would not justify its use, and those are quoted at length in Bommai. Among them: maladministration in a State where a duly constituted ministry enjoys majority support; a recommendation of President's rule without exploring the possibility of an alternative government or fresh elections; a refusal to let a ministry demonstrate its majority through a floor test; and the massive defeat of the ruling State party in a general election to the House of the People.

The case that changed the law

Facts. S.R. Bommai v. Union of India, AIR 1994 SC 1918, (1994) 3 SCC 1, decided 11 March 1994 by nine judges. A group of appeals and transferred cases about Proclamations dismissing the governments of Karnataka in April 1989, Nagaland, Meghalaya in October 1991, and Madhya Pradesh, Rajasthan and Himachal Pradesh in January 1993 following the demolition of the disputed structure at Ayodhya.

Held. The Court summarised twelve conclusions. Those that matter for this chapter are:

The satisfaction under article 356 is subjective, but the power is a conditioned power and not an absolute one: the existence of material is a pre-condition, and the satisfaction must be formed on relevant material. The Sarkaria recommendations merit serious consideration.

The Proclamation is not immune from judicial review. The Supreme Court or a High Court may strike it down if it is found to be mala fide or based on wholly irrelevant or extraneous grounds. The deletion of clause (5) by the Forty-fourth Amendment removed the cloud on reviewability. When called upon, the Union must produce the material on which it acted, and cannot refuse if it seeks to defend the action. The court will not go into the correctness or the adequacy of the material; its enquiry is limited to whether the material was relevant.

Article 74(2) does not shield the material. It bars an enquiry into what advice was tendered, but it does not bar the court from calling on the Union to disclose the material on which the President formed his satisfaction, because the material on which advice was based does not become part of the advice.

The Assembly may not be dissolved before parliamentary approval. Though the power to dissolve is implicit in clause (1), it may be exercised only after both Houses approve the Proclamation under clause (3). Until then the President may only suspend the Assembly. Dissolution is not a matter of course.

If Parliament does not approve, the dismissed government revives and a suspended Assembly is reactivated, and acts done in the two months are not retrospectively invalidated though they may be reviewed or repealed.

If the court strikes down the Proclamation it may restore the dismissed government and revive the Assembly even if dissolved, and may declare that what was done in the interim remains valid.

And the two conclusions of wider importance: the Constitution has created a federation but with a bias in favour of the Centre, and within the sphere allotted to them the States are supreme; and secularism is a basic feature, so a State Government which pursues unsecular policies acts contrary to the constitutional mandate and renders itself amenable to action under article 356.

On the facts, the Karnataka and Meghalaya Proclamations were unconstitutional; those for Madhya Pradesh, Rajasthan and Himachal Pradesh were not.

Why it matters. It converted article 356 from a political power into a legal one. Three things in particular changed the practice: the requirement of relevant material, the disclosure obligation, and the rule that the Assembly may not be dissolved before Parliament approves, which preserves the possibility of restoring the government. The use of article 356 fell sharply after 1994, and that is the practical measure of the decision.

The floor test

One point deserves separate statement because it is where Bommai bites hardest.

The Court held that the proper forum for testing whether a ministry commands a majority is the floor of the House, not the Governor's subjective assessment and not a headcount conducted at Raj Bhavan. Where the question is whether the government still has the confidence of the Assembly, the Governor should require it to prove that on the floor.

The consequence has been a settled modern practice: where a government's majority is in doubt, a court will order a floor test, and will fix the time and sometimes the manner of it. That is the most frequently applied part of the decision.

What remains open

An answer should not present the matter as closed.

The satisfaction is still subjective and the review is narrow. The court asks whether the material was relevant, not whether it was sufficient or correct. A Union that assembles relevant material is difficult to challenge.

Restoration is a remedy that is rarely available in practice. By the time a case is decided, fresh elections have usually been held. In Bommai itself the Court said it would have struck down the Karnataka and Meghalaya Proclamations and restored the governments but for the intervening elections.

Article 355 offers a route around the safeguards, on the wider reading discussed in [Emergency and the Federal Balance].

The Governor's report remains the usual trigger, and the office that produces it is the one described in [The Governor], appointed and removable by the party the report may benefit.

A worked example

The ruling party in State B loses every seat in the State at a general election to the House of the People. The Governor reports that the ministry has lost the confidence of the people, though it has not been defeated in the Assembly and has not been asked to prove its majority. A Proclamation under article 356 is issued and, within a fortnight and before Parliament has met, the Assembly is dissolved.

Is the ground good? No. Bommai quotes the Sarkaria list of situations that do not justify article 356, and this is one of them in terms: the massive defeat of the ruling State party in a general election to the House of the People is not a failure of the constitutional machinery. So is the refusal to let a ministry demonstrate its majority through a floor test.

Is the satisfaction reviewable? Yes. The power is conditioned, not absolute; the satisfaction must rest on relevant material; and the Proclamation may be struck down if mala fide or based on wholly irrelevant or extraneous grounds. The court will not weigh the material's adequacy, only its relevance, and a ground the Sarkaria list expressly excludes is not relevant.

Must the Union disclose the material? Yes. Article 74(2) bars an enquiry into what advice was tendered, but not the production of the material on which the satisfaction was formed, because that material does not become part of the advice. If the Union wants to defend the Proclamation it cannot refuse.

Was the dissolution lawful? No. Bommai holds that although the power to dissolve is implicit in clause (1), it may be exercised only AFTER both Houses approve the Proclamation under clause (3). Until then the Assembly may only be suspended.

What relief follows? If the Proclamation is struck down the court may restore the dismissed government and revive the Assembly, and may declare that acts done meanwhile remain valid. The practical difficulty is timing: in Bommai itself fresh elections had intervened.

Quick revision

• Article 356: satisfaction that the government of a State cannot be carried on in accordance with the Constitution, on the Governor's report or otherwise. The High Court's powers are excluded from the Proclamation.
• Approval by both Houses within two months, six months at a time thereafter, three years maximum.
• Article 365: failure to comply with a Union direction may found the satisfaction.
• Sarkaria 1988: last resort, after a warning, and a list of situations that do not justify it.
• Bommai 1994: conditioned power, relevant material, reviewable for mala fides or wholly extraneous grounds, material must be disclosed, article 74(2) is no shield, no dissolution before parliamentary approval, restoration possible, floor test is the proper forum, federation with a bias to the Centre, secularism a basic feature.
• What remains: subjective satisfaction, narrow review, and restoration that usually comes too late.

Test yourself

1. Why does everything about article 356 follow from the fact that its trigger is a satisfaction rather than a fact?

2. What does article 356(3) require, and what happens if Parliament does not approve?

3. Give four situations from the Sarkaria list that do not justify a Proclamation.

4. State six of the conclusions in S.R. Bommai that bear on article 356.

5. Why is the floor test the proper forum, and what has that produced in modern practice?

6. What remains unsatisfactory about the position even after Bommai?

Article 370(3). The President may by notification declare that article 370 shall cease to be operative, subject to the proviso that the recommendation of the Constituent Assembly of the State shall be necessary before he issues such a notification.

Over sixty years dozens of Constitutional Orders were issued under 370(1)(d), the most important being the Constitution (Application to Jammu and Kashmir) Order 1954. By that route most of the Indian Constitution came to apply to the State, with modifications, and the special position eroded steadily long before 2019. That is a point worth making: article 370 was described as permanent by its defenders and temporary by its critics, and in practice it had been used as an instrument of gradual integration by both.

The problem the proviso created. The Constituent Assembly of the State dissolved in 1957. If its recommendation was a condition of any notification under 370(3), and the body that could make it no longer existed, then article 370 could never be brought to an end at all. That was the argument for its permanence, and it was the central question in 2023.

What was done in August 2019

Three instruments, in order, and an answer must keep them apart.

Constitutional Order 272, 5 August 2019, issued under article 370(1)(d). It applied all the provisions of the Constitution of India to Jammu and Kashmir. Its paragraph 2 also amended article 367, the interpretation clause, by adding a provision that the reference to the Constituent Assembly of the State in the proviso to article 370(3) should be read as a reference to the Legislative Assembly of the State.

Constitutional Order 273, 6 August 2019, issued under article 370(3), declaring that all clauses of article 370 should cease to be operative, subject to a saving.

The Jammu and Kashmir Reorganisation Act 2019, dividing the State into the Union territory of Jammu and Kashmir, with a legislature, and the Union territory of Ladakh, without one.

All of this was done while the State was under President's rule and its Legislative Assembly stood dissolved, so the "concurrence of the Government of the State" was given by the Governor, and the "views" of the State legislature under article 3 were expressed by Parliament acting in its place.

The decision

Facts. In re Article 370 of the Constitution, decided 11 December 2023 by five judges, Chandrachud C.J. with Kaul, Khanna, Gavai and Surya Kant JJ. Challenges to CO 272, CO 273 and the Reorganisation Act.

Held. The abrogation was upheld unanimously, but not every step of the route was.

On sovereignty. The State retained no element of internal sovereignty after the Instrument of Accession and the Proclamation of 25 November 1949 by which the State adopted the Constitution of India. Article 370 was a feature of asymmetric federalism and not of sovereignty. Kaul J. differed on this, holding that the State did retain an element of internal sovereignty which article 370 recognised, though he agreed the abrogation was valid.

On temporariness. Article 370 was intended to be a temporary provision, as its historical context, its text and its placement in Part XXI show.

On the dissolved Constituent Assembly. The power under article 370(3) did not cease when the State's Constituent Assembly dissolved. What ceased was only the transitional power of that Assembly to make a recommendation. The President's power survived, and could be exercised afterwards. CO 273 is therefore valid, and the President could issue it unilaterally.

On CO 272, and this is the part that failed. Article 370 cannot be amended by an exercise of power under article 370(1)(d). If article 370 was to cease to operate or be modified, recourse had to be had to article 370(3). Paragraph 2 of CO 272, which amended article 370 through article 367, is ultra vires article 370(1)(d), because it modified article 370 in effect without following the procedure for modifying it: an interpretation clause cannot be used to bypass the procedure laid down for amendment. The remainder of CO 272, applying all the provisions of the Constitution to the State, is valid, and is not mala fide merely because everything was applied at once.

On President's rule. The exercise of power by the President after an article 356 Proclamation is subject to judicial review; the exercise must have a reasonable nexus with the object of the Proclamation; the challenger must make out a prima facie case of mala fides or extraneous exercise, and the onus then shifts to the Union to justify it. Parliament's power under article 356(1)(b) extends to the non-legislative as well as the legislative powers of the State legislature.

On the reorganisation. The views of the State legislature under the first proviso to article 3 are recommendatory. The Court upheld the carving out of Ladakh as a Union territory under article 3(a) read with Explanation I. It did not decide whether article 3 permits an entire State to be converted into Union territories, because the Solicitor General stated that statehood would be restored. And it directed that the Election Commission hold elections to the Legislative Assembly by 30 September 2024, and that restoration of statehood take place at the earliest and as soon as possible.

On the State's Constitution. Following the application of the whole Constitution of India by CO 273, the Constitution of Jammu and Kashmir is inoperative and declared redundant.

Why it matters. Three propositions of general constitutional law come out of it, and they are usable well beyond Kashmir. What cannot be done directly cannot be done indirectly, and an interpretation clause is not a route round an amendment procedure. Presidential action during President's rule is reviewable, with a shifting onus. And a State legislature's views under article 3 are recommendatory, which confirms Babulal Parate on a much larger scale.

The criticism, which an LL.M. answer should engage with

Two lines are usually pressed and both deserve a sentence.

The self-consent problem. Every consent, concurrence and view that the process required was given by the Union or its appointee, because the State was under President's rule and had no elected government. The Governor concurred on behalf of the State Government; Parliament expressed the views of the State legislature. The Court's answer was that these were the constitutional consequences of a Proclamation whose validity had not been challenged in time, and that the powers exercised were reviewable. The criticism is that a safeguard which the beneficiary of the decision may satisfy on the other party's behalf is not a safeguard.

The unanswered question. The Court expressly declined to decide whether a State may be converted into Union territories under article 3, relying on the assurance that statehood would be restored. That leaves the most far-reaching question in the case open, and it is a question that affects every State, not only this one.

The direction on elections and statehood is stated above as the Court gave it. A student writing in a later year should check what has since happened before asserting the current position, because that is a fact about the world and not a proposition of law.

A worked example

A student is asked: "The Supreme Court upheld the abrogation of article 370, so the Government won on every point." Is that accurate?

No, and the inaccuracy matters. The outcome was upheld; one of the two instruments was not.

What was upheld. CO 273, the notification under article 370(3) declaring that article 370 ceased to operate, is valid. The power under 370(3) survived the dissolution of the State's Constituent Assembly in 1957, because what ceased was only that Assembly's transitional power to recommend, not the President's power. The President could therefore act unilaterally.

What was struck down. Paragraph 2 of CO 272, which amended article 367 so that the reference to the Constituent Assembly in the proviso to article 370(3) would be read as a reference to the Legislative Assembly. That is ultra vires article 370(1)(d): article 370 cannot be amended under 370(1)(d), an interpretation clause cannot be used to bypass an amendment procedure, and what cannot be done directly cannot be done indirectly.

Why the outcome survived anyway. Because CO 273 did not need CO 272's paragraph 2. Once the Court held that the President's power under 370(3) was never conditioned on a body that no longer existed, the workaround was unnecessary.

What the Court did not decide. Whether article 3 permits an entire State to be converted into Union territories. It relied on the Solicitor General's statement that statehood would be restored, upheld only the carving out of Ladakh under article 3(a) with Explanation I, and directed elections to the Assembly by 30 September 2024.

The wider propositions. Presidential action after an article 356 Proclamation is reviewable with a shifting onus, and the views of a State legislature under the first proviso to article 3 are recommendatory.

Quick revision

• Article 370 sat in Part XXI among the temporary and transitional provisions, and limited Parliament's power over the State to the acceded subjects and whatever the State concurred in.
• Article 370(1)(d) applied provisions of the Constitution by Presidential order; article 370(3) could end the article, on the recommendation of the State's Constituent Assembly, which dissolved in 1957.
• August 2019: CO 272 applied the whole Constitution and amended article 367 to read Constituent Assembly as Legislative Assembly; CO 273 ended article 370; the Reorganisation Act made two Union territories.
• 2023: abrogation upheld. No internal sovereignty; article 370 temporary and a feature of asymmetric federalism; the 370(3) power survived the Constituent Assembly's dissolution; CO 273 valid.
• Paragraph 2 of CO 272 was ultra vires: an interpretation clause cannot bypass an amendment procedure.
• Presidential action under article 356 is reviewable with a shifting onus; article 3 views are recommendatory; Ladakh's carving out upheld; whether a whole State may become Union territories was left undecided.

Test yourself

1. What did article 370 actually do, and what were the two mechanisms in clauses (1)(d) and (3)?

2. What was the argument that article 370 had become permanent, and how did the Court answer it?

3. Distinguish CO 272 from CO 273, and say what the Court held about each.

4. Why was paragraph 2 of CO 272 ultra vires? State the principle in general terms, without reference to Kashmir.

5. What question did the Court expressly leave undecided, and why does it matter to every State?

Notice what article 371A does. It gives a State legislature a veto over the application of a law of Parliament in defined fields. That is not a protection any other State has, and it goes further than article 370 ever did on those subjects.

Article 371B, Assam. The President may provide for a committee of the Legislative Assembly consisting of members elected from the tribal areas.

Article 371C, Manipur. The President may provide for a committee of the Assembly consisting of members elected from the hill areas, and may give the Governor special responsibility for its proper functioning.

Article 371D, Andhra Pradesh and Telangana. Equitable opportunities in public employment and education for people of different parts of the State. The President may require the State to organise civil posts into local cadres and to specify local areas, and may provide for an Administrative Tribunal outside the ordinary judicial hierarchy to decide service disputes, whose jurisdiction excludes that of all courts except the Supreme Court. Article 371E provides for the establishment of a Central University in Andhra Pradesh.

Article 371F, Sikkim. The most elaborate, because Sikkim joined in 1975 by the Thirty-sixth Amendment. It fixes a minimum Assembly size, provides for seats to protect the rights and interests of sections of the population, gives the Governor special responsibility for peace and equitable arrangements between sections, continues laws in force before Sikkim became a State, and provides for the transformation of the Sikkim High Court.

Article 371G, Mizoram. In the same form as Nagaland: no Act of Parliament on religious or social practices of the Mizos, Mizo customary law and procedure, administration of civil and criminal justice involving Mizo customary law, or ownership and transfer of land, applies unless the Assembly so resolves.

Article 371H, Arunachal Pradesh. Special responsibility for the Governor with respect to law and order, exercised in his individual judgment after consulting the Council of Ministers.

Article 371-I, Goa. The shortest in the series: the Legislative Assembly of Goa shall consist of not less than thirty members.

Article 371J, Karnataka. Inserted in 2012 for the Hyderabad-Karnataka region: a separate development board, equitable allocation of funds, and reservation in education and State employment for persons of that region.

What the pattern shows

Read as a group rather than as a list, four things emerge, and these are the marks in an essay.

The provisions answer different problems. Articles 371, 371D and 371J address intra-State regional inequality. Articles 371A, 371B, 371C and 371G address the position of tribal and customary communities. Articles 371F and 371H address the terms on which a territory joined or was upgraded. They are not one policy.

Most were added by amendment, not by the original Constitution. Only article 371 was there in 1950. The rest were inserted between 1962 and 2012 as particular problems arose. Asymmetry in India is not a founding design but an accumulated practice.

The strongest of them survive, and the one that was abrogated was different in kind. Article 371A gives Nagaland a legislative veto that no other State has, and it is untouched. Article 370 was ended. The difference is that article 370 was placed among the temporary provisions, was drafted with its own termination mechanism in clause (3), and was tied to a Constituent Assembly that had ceased to exist. The 371 series has no equivalent clause, and can be altered only by amendment under article 368.

Asymmetry is compatible with equality. The standard objection is that treating States differently offends the equality of units. The answer is the one article 14 gives in a different context: treating unequals alike is as unequal as treating equals differently, and a State whose social organisation rests on customary law is not in the same position as one whose does not.

Asymmetric federalism after 2023

The decision on article 370 supplies the vocabulary.

Facts. In re Article 370 of the Constitution, decided 11 December 2023, five judges. Challenges to the abrogation of article 370 and the reorganisation of the State.

Held. Among the conclusions, the State of Jammu and Kashmir retained no element of internal sovereignty after the Instrument of Accession and the Proclamation of 1949, and article 370 was a feature of asymmetric federalism and not of sovereignty. The article was temporary; the power under article 370(3) survived the dissolution of the State's Constituent Assembly; CO 273 was valid; but paragraph 2 of CO 272, which amended article 370 by way of article 367, was ultra vires.

Why it matters here. The Court drew the distinction this chapter depends on. A special provision is a differentiated term of membership of the Union, and it does not make the State a partner rather than a unit. On that reasoning the 371 series is secure as a matter of constitutional description: it does not claim sovereignty, so nothing said about sovereignty threatens it. What follows for its permanence is a different question, since a provision inserted by amendment can in principle be removed by amendment, subject to the basic structure.

A worked example

Parliament enacts a law regulating the transfer of agricultural land, and provides that it shall apply throughout India. A resident of Nagaland says it does not apply there. A resident of Goa says the same about his State.

Nagaland. Article 371A provides that no Act of Parliament in respect of ownership and transfer of land and its resources shall apply to Nagaland UNLESS the Legislative Assembly of Nagaland by a resolution so decides. Ownership and transfer of land is one of the four subjects named, alongside religious or social practices of the Nagas, Naga customary law and procedure, and the administration of civil and criminal justice involving decisions according to Naga customary law. So unless the Assembly has resolved, the Act does not apply. The resident is right.

Goa. Article 371-I provides only that the Legislative Assembly of Goa shall consist of not less than thirty members. It confers no immunity from central legislation of any kind. The resident is wrong.

The lesson. "Special provision" is not a single thing. The article 371 series answers three different problems: intra-State regional imbalance (articles 371, 371D, 371J), the position of tribal and customary communities (371A, 371B, 371C, 371G), and the terms on which a territory joined or was upgraded (371F, 371H). Only the second group contains a legislative veto, and only two articles in it.

And note where Mizoram sits. Article 371G is in the same form as 371A, so the same answer would be given for a resident of Mizoram.

The wider point. Article 371A gives a State legislature a veto over the application of a law of Parliament, which goes further on those subjects than article 370 ever did, and it is untouched.

Quick revision

• Asymmetric federalism means units on different constitutional terms, and India is asymmetric by special provision, by category of unit, and by scheduled area.
• Article 371 Maharashtra and Gujarat, 371A Nagaland, 371B Assam, 371C Manipur, 371D and 371E Andhra Pradesh, 371F Sikkim, 371G Mizoram, 371H Arunachal Pradesh, 371-I Goa, 371J Karnataka.
• Articles 371A and 371G are the strongest: an Act of Parliament on customary law, social and religious practice, justice according to custom, and land does not apply unless the State Assembly resolves that it shall.
• The provisions answer three different problems: regional imbalance within a State, customary and tribal communities, and terms of joining.
• All but article 371 were added by amendment, so asymmetry here is accumulated practice, not founding design.
• In re Article 370: article 370 was a feature of asymmetric federalism and not of sovereignty. That description now governs the whole series.

Test yourself

1. What does asymmetric federalism mean, and in what three distinct ways is India asymmetric?

2. Match each of articles 371, 371A, 371D, 371F, 371G, 371-I and 371J to its State and to the problem it answers.

3. What is the legislative veto in articles 371A and 371G, and over which four subjects does it operate?

4. Why is article 370 gone while the 371 series remains? Give three differences.

5. "Treating States differently offends the equality of units." Answer the objection.

A State suing the Union under article 131 may not add a private party as a defendant, and a suit which is really between private parties does not become an article 131 suit because a government is joined to it.

The legal right requirement

The words "if and in so far as the dispute involves any question, whether of law or fact, on which the existence or extent of a legal right depends" are the real gate.

What satisfies it. A claim that a law of Parliament is beyond its legislative competence and so invades the State's field; a claim to property or to money as of right; a claim that a constitutional provision confers an entitlement the other government has denied.

What does not. A grievance about policy. A complaint that the Union has been unfair in allocating resources, without pointing to a right to a particular allocation. A political dispute dressed as a legal one.

The contested question. Whether a State may use article 131 to challenge the constitutional validity of a central law that does not directly affect the State's own legal rights, but affects its residents. The arguments both ways are worth knowing: for, that a State has a legal interest in the constitutionality of what it will be required to administer within its territory; against, that a challenge to a law's validity on the ground that it violates fundamental rights belongs to the persons whose rights are affected, under articles 32 and 226, and that allowing a State to bring it converts article 131 into a general jurisdiction to review legislation. The question has been argued in the Supreme Court and an answer should present it as open rather than settled.

The case that shaped it

Facts. State of West Bengal v. Union of India, AIR 1963 SC 1241, (1964) 1 SCR 371, decided 21 December 1962. The State sued the Union under article 131, complaining that the Coal Bearing Areas (Acquisition and Development) Act 1957 could not apply to land owned by the State, and that if it did, Parliament lacked competence to enact it, because the States are sovereign in their own sphere.

Held. The Act applied to State land and was within Parliament's competence under Entry 42 of the Concurrent List read with Entries 52 and 54 of the Union List. The Constitution is not truly federal in character; full sovereignty does not vest in the States; and Parliament, competent under article 3 to destroy a State, cannot be held incompetent to acquire its property. Article 294 vests property in a State but contains no prohibition on its transfer, and what may be transferred may be acquired.

Why it matters here. It is the leading example of article 131 doing exactly what it was designed for: a State and the Union litigating the extent of legislative power, in the only forum that could hear them, with the answer binding both. Note that the State lost. The jurisdiction is a guarantee of a hearing, not of an outcome.

The two exclusions

The proviso to article 131. Disputes arising out of a treaty, agreement, covenant, engagement, sanad or other similar instrument entered into before the commencement of the Constitution and still in operation, or which itself provides that the jurisdiction shall not extend to such a dispute. This was aimed at the covenants of accession and the settlements with the princely States, and its practical importance has faded.

Article 262(2) and inter-State water. Parliament may by law provide that neither the Supreme Court nor any other court shall exercise jurisdiction over an inter-State water dispute, and section 11 of the Inter-State River Water Disputes Act 1956 does exactly that. This is the large exclusion, and it removes from article 131 the very class of dispute most likely to set State against State. See [Inter-State Water Disputes].

Article 131 and the other routes

Keep these apart, because a problem question will mix them.

Article 131 is a suit between governments about a legal right. Article 32 is a petition by a person whose fundamental right has been infringed, and a State is not a person for that purpose. Article 136 is special leave against a decision of a court or tribunal, which an article 131 suit is not, because it starts in the Supreme Court. Article 143 is a Presidential reference for an advisory opinion, which is not a dispute between parties at all, and which the Union alone can set in motion; see [Advisory Jurisdiction].

A State that wants to challenge a central law therefore has a choice of route only in a narrow sense: article 131 if it can show its own legal right is in question, article 32 not at all, and a writ petition in a High Court only in the limited circumstances in which a government may be a petitioner.

A worked example

State C wishes to challenge a central Act on two grounds: that Parliament lacked legislative competence to enact it, because it falls within List II; and that it violates the fundamental rights of the State's residents. State C proposes to sue the Union in the Supreme Court under article 131, joining as a co-plaintiff a company incorporated in the State which is affected by the Act.

The company cannot be a party. Article 131 lies only between the Government of India and one or more States, or between States. A private person may neither invoke it nor be joined, and a suit that is really between private parties does not become an article 131 suit because a government is added to it. The company must be dropped.

The competence ground is squarely within article 131. A claim that a central Act invades List II asserts that the State's own legislative field has been trespassed on, which involves a question of law on which the existence or extent of a legal right depends. This is the classic use of the jurisdiction, and it is what State of West Bengal v. Union of India was.

The fundamental rights ground is doubtful. A State is not a "person" for article 32 and cannot assert its residents' fundamental rights as its own. Whether a State may nevertheless use article 131 to challenge the validity of a central law that affects its residents rather than its own legal rights is contested, and an answer should present it as open: for, that the State has a legal interest in the constitutionality of what it must administer; against, that this would convert article 131 into a general jurisdiction to review legislation belonging to those whose rights are affected.

One further check. Is this a water dispute? If it were, section 11 of the 1956 Act would bar the suit under article 262(2).

Quick revision

• Article 131: original and exclusive jurisdiction over disputes between the Union and States, or between States.
• Only governments may be parties; a private person cannot be joined, and a public undertaking is not the government.
• The dispute must involve a question on which the existence or extent of a legal right depends. Policy grievances do not qualify.
• Whether a State may use it to challenge the validity of a central law that affects its residents rather than its own rights is contested and should be presented as open.
• West Bengal v. Union of India is the model case, and the State lost: the jurisdiction guarantees a forum, not a result.
• Excluded: pre-Constitution covenants under the proviso, and inter-State water disputes under article 262(2) with section 11 of the 1956 Act.

Test yourself

1. Who may and who may not be a party to a suit under article 131?

2. What must a dispute involve before article 131 is available, and give an example of a grievance that does not qualify.

3. State the two exclusions from article 131, and say which is the practically important one.

4. Distinguish articles 131, 32, 136 and 143 from one another.

5. May a State use article 131 to challenge the validity of a central law on the ground that it violates the fundamental rights of the State's residents? Argue both ways.

Notice how much of the real machinery is executive rather than constitutional. That is the standing criticism of Indian cooperative federalism: the forum the Constitution provided was left unused for forty years, and the forums that did the work were created by Cabinet resolution and could be abolished the same way, as the Planning Commission was.

The GST Council as the working model

The Council under article 279A is the clearest instance of cooperative federalism written into the Constitution rather than improvised, and it is worth using as the worked example.

Both levels hold the same taxing power under article 246A; neither can carry a decision in the Council alone, because the Union's third and the States' collective two-thirds are each less than the three-fourths a decision requires; and the machinery is constitutional, not executive, so neither side can dissolve it. That is cooperation made structurally necessary rather than merely encouraged.

Facts. Union of India v. Mohit Minerals Pvt Ltd, decided 19 May 2022. Among the questions was whether the recommendations of the GST Council bind the Union and the States.

Held. They do not. The deletion of the proposed article 279B and the inclusion of article 279A(1) indicate that Parliament intended the recommendations to have persuasive value only, read with the objective of fostering cooperative federalism. Article 279A has no non obstante clause and article 246A is not subject to it; both legislatures have simultaneous power and article 246A envisages no repugnancy provision, so treating recommendations as binding edicts would disrupt fiscal federalism. It is not imperative that one federal unit must always possess a higher share in the power for the units to make decisions. The Government is nevertheless bound by the recommendations when exercising its rule-making power under the CGST and IGST Acts.

And the sentence to quote:

Indian federalism is a dialogue between cooperative and uncooperative federalism where the federal units are at liberty to use different means of persuasion ranging from collaboration to contestation.

Why it matters. It is the most important modern statement of what cooperative federalism means in India, and it corrects the usual misuse of the phrase. Cooperative federalism is not a duty to agree, and a State that refuses is not acting unconstitutionally. The Court gave contestation the same standing as collaboration.

Competitive federalism

The second phrase, and the one that has grown in official use.

The idea. States compete for private investment, for skilled people and for a reputation for good administration. The Union encourages this by publishing rankings and indices: ease of doing business, health, education, water management, and the aggregate development indices NITI Aayog produces.

The case for it. Comparison creates pressure to improve where central direction cannot reach, because the subjects are the States' own. A State that improves its administration attracts investment and is rewarded politically. It respects the division of powers, since nobody is being told what to do.

The case against it. Competition between unequal units entrenches inequality: a State that begins with better infrastructure attracts more investment and so improves faster. Competition for investment invites a race to the bottom in taxation, labour standards and environmental regulation. And a ranking published by the Union is a form of central influence over State subjects that is not accountable in the way legislation is.

How the two fit together. They are not opposites. Cooperative federalism concerns the relationship between the Union and the States, vertically. Competitive federalism concerns the relationship among the States, horizontally. A federation can be, and India is, both at once.

The new challenges, which is where the topic is going

The paper is called "New Challenges", and the following are the live ones for this topic. An answer that names them is answering the question that was set rather than the question that was set thirty years ago.

Delimitation and the southern States. Article 82 requires readjustment of the allocation of seats in the House of the People after each census, and the allocation has been frozen since 1976, most recently until the first census after 2026. When it is unfrozen, States whose population grew slowly, largely in the south, will lose relative weight to States whose population grew faster. Those States argue that they are being penalised for succeeding at a national policy. This is the most serious federal question India now faces, and it interacts with the 106th Amendment, whose commencement under article 334A waits on the same delimitation. See [The 106th Amendment, and Why It Has Not Started].

The narrowing of independent State revenue. After GST the States' own tax base is small, and the growth of surcharges and cesses outside the divisible pool reduces the share of what remains.

Central agencies operating within States. The Punchhi Commission recommended that a State's consent be required before central investigating agencies act within it, and the withdrawal of general consent by several States has made this a live constitutional dispute rather than a recommendation.

The Governor. Dealt with in [The Governor], and now the subject of two decisions in eight months.

Union territories with legislatures. The long litigation over the control of services in Delhi under article 239AA has produced a body of law about what a Union territory with a legislature actually is, and the question is not confined to Delhi.

A worked example

A commentator writes: "Cooperative federalism means the States must fall in line with national policy; a State that refuses is acting against the constitutional spirit." Assess that statement.

The description of the machinery is roughly right. Cooperative federalism does describe joint institutions and shared work: the Inter-State Council under article 263, the Finance Commission under article 280, the GST Council under article 279A, legislation by consent under article 252, entrustment under articles 258 and 258A, and, outside the Constitution, NITI Aayog and the centrally sponsored schemes funded under article 282.

The normative claim is wrong, and there is authority for saying so. In Mohit Minerals the Supreme Court held that the recommendations of the GST Council do not bind, precisely because treating them as binding edicts would disrupt fiscal federalism, and said that Indian federalism is a dialogue between cooperative and uncooperative federalism in which the federal units are at liberty to use different means of persuasion ranging from collaboration to contestation.

So refusal is not unconstitutional. Contestation is given the same standing as collaboration. A State that declines a scheme, withholds consent under article 252, or votes against a recommendation in the Council is exercising a constitutional freedom, not breaching a duty.

Where the commentator has a point. Article 279A does make agreement structurally necessary: neither the Union's one-third nor the States' collective two-thirds reaches the three-fourths a decision needs. Cooperation is compelled by arithmetic, not by duty.

The distinction to draw. Cooperative federalism is vertical, between the Union and the States. Competitive federalism is horizontal, among the States. A federation can be both, and India is.

Quick revision

• Cooperative federalism is a description of practice, not a rule in the text: joint institutions, shared finance, consultation.
• Its constitutional instruments are articles 263, 280, 279A, 252 and 258; much of the real machinery has been executive and therefore abolishable.
• The GST Council is cooperation made structurally necessary: neither side can decide alone.
• Mohit Minerals: recommendations are persuasive, not binding, and Indian federalism is a dialogue between cooperative and uncooperative federalism, with contestation as legitimate as collaboration.
• Competitive federalism is horizontal, between States; the objection is that it entrenches inequality and invites a race to the bottom.
• The live challenges: delimitation, the shrinking State tax base, central agencies within States, the Governor, and Union territories with legislatures.

Test yourself

1. Why is dual federalism impossible in a modern State? Give three reasons.

2. List the constitutional instruments of cooperative federalism with their articles, and say why so much of the real machinery has been executive instead.

3. Why is the GST Council the clearest structural example, and what does the voting arithmetic compel?

4. Quote and explain what Mohit Minerals says Indian federalism is.

5. What is competitive federalism, and what are the two standard objections to it?

6. Name four live federal challenges and say why delimitation is the most serious.

Four categories are named. The Government and Parliament of India; the Government and the Legislature of each State; all local authorities; and all other authorities. The first two are the executive and legislature of both levels; the third is a defined class; the fourth is the open one.

Two territorial formulas. "Within the territory of India" or "under the control of the Government of India". The second reaches bodies outside India that the Government of India controls, so a body operating abroad under Indian government control is not outside Part III merely because of where it is.

Article 36 carries the same definition into Part IV, so a body that is "State" for the fundamental rights is also "State" for the directive principles.

Why the question keeps changing

The definition has not been amended since 1950. What has changed is the State.

In 1950 the State was the government and its departments. Public functions were performed by ministries, and the list in article 12 caught almost everything that mattered.

From the 1950s to the 1980s the State expanded through corporations. Functions were hived off into statutory bodies, government companies and registered societies, each a separate legal person, none of them a "Government" or a "Legislature". The question became whether those bodies were caught, and the courts said yes, because otherwise the State could escape Part III by choosing a corporate form. That is the widening described in [The Instrumentality Test].

From the 1990s the movement reversed. The State began to withdraw: disinvestment, privatisation, contracting out, public private partnership, and regulation in place of ownership. Now the question is whether the fundamental rights follow the function out of government hands, and if not, what protects the citizen who deals with the private body that replaced the department. That is MU's own topic label, and it is [Liberalisation, Privatisation and the Case for Widening Article 12].

The single sentence that captures the whole topic is this: article 12 asks who holds public power, and the answer to that question has changed twice since 1950 while the words have not changed at all.

The first widening

Facts. Rajasthan State Electricity Board, Jaipur v. Mohan Lal, AIR 1967 SC 1857, (1967) 3 SCR 377, decided 3 April 1967. Government employees were placed at the disposal of the State Electricity Board, a body constituted under the Electricity (Supply) Act 1948. When the first respondent was refused consideration for promotion while others identically placed were promoted, he moved the High Court under articles 226 and 227 alleging breach of articles 14 and 16. The Board answered that it was not "State" under article 12, so no direction could issue to it.

Held. By the majority, the Board is an "other authority" and therefore "State". The expression is wide enough to include every authority created by a statute on which powers are conferred to carry out governmental or quasi-governmental functions. It does not matter that some of those powers are for commercial activity, because under articles 19(1)(g) and 298 the State itself may carry on trade. And the principle of ejusdem generis must not be applied to "other authorities", because the bodies named before it, the executive governments, the legislatures and local authorities, share no distinct genus running through them.

Why it matters. It is the case that opened the field. Before it, the argument that "other authorities" must be read as limited to bodies of the same kind as those named would have confined article 12 to organs of government proper. The rejection of ejusdem generis is what made everything that follows possible.

The shape of the rest of the topic

The chapters that follow work through the categories in order:

• The three named categories, in [The Named Authorities].
• The early struggle over "other authorities", in ["Other Authorities": The Early Cases].
• The tests that emerged, in [The Instrumentality Test].
• The narrowing, in [The Retreat, and What Survived It].
• Two special problems: [Is the Judiciary "State"?] and [Private Bodies and Public Functions].
• MU's own question, in [Liberalisation, Privatisation and the Case for Widening Article 12].
• And the companion article, in [Article 13: Laws Inconsistent with Fundamental Rights].

A worked example

A student writes in an examination: "The petitioner's fundamental rights under article 14 were violated by the private company, so she may move the Supreme Court under article 32." Identify what has gone wrong and set out the correct chain of reasoning.

The error. Article 14 binds the State. Article 32 lies against the State. A private company is not the State unless it satisfies article 12, and the answer assumes what has to be proved.

The correct chain, in order.

One. Is the body "State" within article 12? It is not the Government or a legislature, and it is not a local authority, so the question is whether it is an "other authority".

Two. Apply the test. Is the body, on the cumulative facts, financially, functionally and administratively dominated by or under the control of the Government, that control being particular to it and pervasive rather than merely regulatory?

Three. If yes: articles 14 and 21 bind it, article 32 lies, and its rules are "law" for article 13(2).

Four. If no: article 32 does not lie. Ask instead whether the body discharges a public function or owes a public duty, in which case article 226 is available in the High Court, being wider in reaching "any person or authority" and extending "for any other purpose".

Five. If neither: the remedy is the ordinary law, a suit or a statutory remedy.

The point to hold on to. Article 12 is not a definition to be recited. It is the gate, and everything in Part III is on the other side of it.

Quick revision

• Article 12 defines "State" for Part III and, through article 36, for Part IV.
• It matters because the rights bind only the State, article 32 lies only against the State, and article 13 tests only what the State makes.
• "Includes" makes the definition inclusive, and "unless the context otherwise requires" makes it flexible. Both are the textual basis of judicial widening.
• Four categories: the Union executive and legislature, the State executive and legislature, local authorities, and other authorities. Two territorial formulas: within India, or under the control of the Government of India.
• Rajasthan State Electricity Board: "other authorities" covers every statutory authority with governmental or quasi-governmental powers, commercial activity does not matter, and ejusdem generis does not apply.
• The words have not changed since 1950; what the State is has changed twice.

Test yourself

1. State three consequences that follow from a body being "State" within article 12.

2. What is the significance of the words "includes" and "unless the context otherwise requires"?

3. Give the two territorial formulas in article 12 and explain what the second adds.

4. Why has the article 12 question changed twice since 1950, when the words have not changed at all?

5. What did Rajasthan State Electricity Board decide about ejusdem generis, and why was that decision the foundation of everything that followed?

Distinguish also the legislature acting in its legislative capacity from the House exercising its privileges. The privileges of a House under articles 105 and 194, and its power to punish for contempt, raise a separate question about how far the fundamental rights control them, dealt with in [Legislative Privilege].

Local authorities

The Constitution does not define "local authority", and the definition applied is the one in section 3(31) of the General Clauses Act 1897: a municipal committee, district board, body of port commissioners or other authority legally entitled to, or entrusted by the Government with, the control or management of a municipal or local fund.

From that definition and the case law, a body is a local authority if it has these characteristics.

A separate legal existence, so that it is a corporate body and not a department.

A defined area within which it functions.

A function of governmental character in that area, of the kind a municipality or a district board performs.

The power to raise its own funds, generally by levying taxes, rates, charges or fees.

A substantial degree of autonomy, so that it is not merely an agent carrying out instructions, though government control by way of supervision does not prevent it being a local authority.

Elected or partly elected composition is typical, though it is the function and the fund that matter more.

So a municipal corporation, a municipality, a nagar panchayat, a zila parishad, a panchayat samiti, a gram panchayat, an improvement trust, a port trust and a cantonment board are local authorities. The panchayats and municipalities now have constitutional status of their own under Parts IX and IXA, dealt with in [Grass Root Democracy: The Panchayats] and [Urban Local Government: The Municipalities], but their status under article 12 does not depend on that.

Why the fourth category exists

The three named categories share a feature: each is a recognisable organ of government, created as such, exercising governmental power in an obvious way.

The fourth category exists because the State does not confine itself to such organs. It creates statutory corporations, forms companies, registers societies, and endows trusts, and each of them is a legal person distinct from the government that made it. If article 12 stopped at the third category, the State could place any activity beyond Part III by the simple device of incorporating it.

That is the problem the courts have spent sixty years on, and it begins in the next chapter.

A worked example

A municipal corporation refuses a trade licence to an applicant on a ground she says is discriminatory. A parallel refusal is made by a State-owned company operating a market. Both are challenged under article 14.

The corporation. A municipal corporation is a "local authority" within article 12, taking the meaning from section 3(31) of the General Clauses Act 1897. Test it against the marks: separate legal existence, a defined area, functions of a governmental character in that area, power to raise its own funds by rates and fees, and a real measure of autonomy. A municipal corporation satisfies all five. Article 14 binds it directly and both articles 32 and 226 are available.

The company. A State-owned company is not a local authority: it has no defined area of local government and no power to levy rates. It falls, if anywhere, in the fourth category, "other authorities", and must be tested on the instrumentality question in the following chapters. Ownership of the shares by the State is a significant fact but not by itself decisive.

Why the distinction matters here. For the corporation the question of article 12 does not arise at all: it is answered by the third category, and the argument goes straight to the merits of the discrimination. For the company the case may be lost on the threshold.

A further point on the corporation. Since the Seventy-fourth Amendment, municipalities have constitutional status under Part IXA. That does not change their position under article 12, which never depended on it, but it is worth noticing that the same body is now reached by two different routes.

Quick revision

• The Government and Parliament of India, and the Government and Legislature of each State, are named. Executive action is caught as well as legislation.
• Article 33 permits Parliament to restrict the rights for the armed forces.
• Whether a constitutional amendment is "law" under article 13(2) is a question about article 368, not about article 12.
• "Local authority" takes its meaning from section 3(31) of the General Clauses Act 1897: control or management of a municipal or local fund.
• The marks of a local authority: separate legal existence, a defined area, governmental function, power to raise its own funds, and a real measure of autonomy.
• The fourth category exists because the State acts through separate legal persons, and a definition that stopped at the third would let it escape Part III by incorporation.

Test yourself

1. Name the four categories in article 12 and say which of them has generated the litigation.

2. Is executive action, as distinct from legislation, caught by article 12? What does article 33 permit?

3. Where does the meaning of "local authority" come from, and what are the marks of one?

4. Why does the Constitution need a fourth, residual category at all?

5. Is a constitutional amendment "law" for article 13(2)? Explain why that is not an article 12 question.

Held. By the majority, the Board is an "other authority" and therefore "State", to which directions may be given under articles 226 and 227. The expression is wide enough to include every authority created by a statute on which powers are conferred to carry out governmental or quasi-governmental functions, functioning within the territory of India or under the control of the Government of India. It is not material that some of the powers conferred may be for carrying on commercial activities, because under articles 19(1)(g) and 298 the State itself is empowered to carry on trade or business. And ejusdem generis must not be applied, because for that rule to operate there must be a distinct genus or category running through the bodies previously named, and there is none: the executive governments of the Union and the States, the legislatures of the Union and the States, and local authorities have no common genus.

Why it matters. It rejects two of the three readings in a single judgment. Ejusdem generis goes, and with it the argument that article 12 is confined to organs of government; and the sovereign function test goes, because commercial activity is expressly held immaterial. What is left is the statutory power reading, and every later development builds on it.

The next step: what a statutory corporation's rules are

Facts. Sukhdev Singh v. Bhagatram Sardar Singh Raghuvanshi, AIR 1975 SC 1331, (1975) 1 SCC 421, decided 21 February 1975. Employees of the Oil and Natural Gas Commission, the Life Insurance Corporation and the Industrial Finance Corporation were removed from service in breach of the regulations framed under the Acts constituting those bodies. Two questions arose: whether a removal contrary to the regulations entitled the employee to a declaration that he continued in service, or only to damages; and whether an employee of a statutory corporation may claim articles 14 and 16 against it, which is to ask whether such a corporation is an authority within article 12.

Held. The regulations framed by the three bodies have the force of law. They are subordinate legislation made under a specific statutory power, not merely terms of a contract of employment, and the statutory bodies have no free hand to depart from them. It follows that the employees hold a statutory status and are entitled to a declaration of continuance in service where their removal contravened the regulations. And all three statutory bodies are "authorities" within article 12, though the Court stated expressly that their employees are not servants of the Union or of a State.

The concurrence that mattered more than the holding. Mathew J. agreed in the result but reasoned differently. He treated the question as whether the corporation is an instrumentality or agency of the State, looking to the public function it performs and to the reality of State control, rather than only to the statute that created it. That reasoning is the seed of everything in the next chapter.

Why it matters. Two propositions in one case: that the internal rules of a statutory corporation are law, so that breach of them is a legal wrong rather than a breach of contract; and that such corporations are State. And Mathew J's concurrence reframed the question from "was this body created by statute" to "is this body really the government acting under another name", which is a question that can be asked of a company or a society as easily as of a corporation.

What the early cases had not yet answered

Both cases concerned bodies created by a statute. That left the harder case untouched.

Suppose the State does not create a statutory corporation. Suppose it registers a society under the Societies Registration Act, or incorporates a company under the Companies Act, and puts the function there. The body is then created by an ordinary law of general application, in the same way as any private club or company, and there is no special statute conferring powers on it at all.

Is such a body "State"? On the Rajasthan Electricity Board formula, arguably not, because it is not "created by a statute on which powers are conferred". On Mathew J's approach, the answer depends on whether it is an instrumentality of the State, and its legal form is beside the point.

That question is the subject of the next chapter, and the answer given to it is the reason article 12 reaches as far as it does.

A worked example

A State sets up a Housing Board by statute, giving it power to acquire land compulsorily, to allot houses, to fix rents and to evict defaulters, and providing that it shall be run on commercial lines and shall aim to break even. An allottee complains that the allotment rules discriminate, and the Board answers that it is a commercial body and not "State".

Take the three early readings in turn.

Ejusdem generis? The Board is not a government, a legislature or a local authority, so on that reading it would be outside. But Rajasthan State Electricity Board rejects the rule here, because the bodies named in article 12 share no distinct genus, and without a genus the rule cannot operate.

Sovereign function? The Board allots houses and charges rent, which is not obviously a sovereign function. But that reading was also rejected: it is not material that some of the powers conferred are for carrying on commercial activities, because under articles 19(1)(g) and 298 the State itself may carry on trade.

Statutory power? The Board is created by a statute and has been given powers, including compulsory acquisition and eviction, to carry out governmental or quasi-governmental functions. On the surviving reading it is an "other authority" and therefore "State".

Answer. The Board is State, and article 14 binds its allotment rules.

A second point the Board may not have expected. On Sukhdev Singh, regulations framed by such a body under its statute have the force of law. So the allotment rules are not merely administrative instructions: an allottee dealt with in breach of them has a legal right, not just a grievance.

Quick revision

• Three readings competed: ejusdem generis, sovereign function, and statutory power. The third won.
• Rajasthan State Electricity Board: every authority created by statute with governmental or quasi-governmental powers is caught; commercial activity is immaterial; ejusdem generis does not apply because the named bodies share no genus.
• Sukhdev Singh: the regulations of a statutory corporation have the force of law, its employees have statutory status, and such corporations are State.
• Mathew J's concurrence in Sukhdev Singh reframed the enquiry as whether the body is an instrumentality or agency of the State.
• Both cases concerned bodies created by statute. The question they left open was the government company and the registered society.

Test yourself

1. Set out the three competing readings of "other authorities" and say which prevailed.

2. Why does ejusdem generis have no application to article 12? Give the reason, not just the conclusion.

3. Does it matter that a body carries on commercial activity? Which two articles answer that?

4. What two propositions come out of Sukhdev Singh, and which of them was in a concurring judgment?

5. What question did the early cases leave open, and why could their formula not answer it?

On article 12, Bhagwati J. asked when a corporation becomes an instrumentality or agency of government, and answered that there is no cut and dried formula which would provide the correct division of corporations into those which are instrumentalities of government and those which are not. Relevant considerations include: the holding of the entire share capital by government, which goes a long way; extensive and unusual financial assistance, where the government's purpose in giving it coincides with the purpose for which the body uses it and that purpose is of a public character; a monopoly status conferred or protected by the State; deep and pervasive State control; functions of public importance closely related to governmental functions; and the transfer to the body of a department of government.

The judgment also produced the sentence that has outgrown its subject: every action of the executive government must be informed with reason and should be free from arbitrariness, and that is the very essence of the rule of law and its bare minimal requirement.

Why it matters. It is the origin of the test, and it is also the origin of the proposition that arbitrariness in State action is itself unconstitutional, which is developed in [Arbitrariness].

The tests restated, and extended past statutory bodies

Facts. Ajay Hasia v. Khalid Mujib Sehravardi, AIR 1981 SC 487, (1981) 1 SCC 722, decided 13 November 1980. Admission to the Regional Engineering College, Srinagar, run by a society registered under the Jammu and Kashmir Registration of Societies Act 1898. Candidates challenged a viva voce lasting two or three minutes a candidate, consisting of formal questions about parentage and residence with no relevance to the subject, and carrying a third of the total qualifying marks. The society denied that it was "State" at all.

Held. The society is "State" within article 12 and is bound by article 14. Its composition is dominated by nominees of the Central and State Governments; the money to run the college comes entirely from those Governments; its rules require their prior approval; its accounts go to them; it must comply with their directions; it cannot dispose of immovable property without their approval; and the Governments may appoint and remove members. That control is so deep and pervasive that the society is merely a projection of the Governments: the voice is that of the State and the Central Governments.

The Court held that the tests are the same whether the body is a statutory corporation or a society registered under an ordinary Act, because the form in which the State chooses to act cannot decide whether a fundamental right applies. On the merits, allotting a third of the marks to an interview conducted in that way was excessive and arbitrary.

Why it matters. It is the case usually cited for the settled formulation of the tests, and it is the case that carried them past bodies created by statute. After Ajay Hasia the enquiry is genuinely about substance, and the legal instrument of creation is only one fact among several.

Using the test in a problem

Six questions, and the answer to none of them is decisive on its own.

Who owns it? Is the entire share capital, or the whole of the corpus, held by government?

Who funds it? Is the financial assistance extensive and unusual, and does it meet substantially the whole of the expenditure? Occasional grants and ordinary subsidies available to everyone in the sector do not count.

Does it hold a monopoly, and if so, who gave it? A monopoly conferred or protected by the State is significant; a monopoly the body achieved for itself is not.

Is the control deep and pervasive, or merely regulatory? This is the question that decides most modern cases, and it is developed in [The Retreat, and What Survived It].

What does it do? Are its functions of public importance and closely related to governmental functions?

Where did it come from? Was a department of government transferred into it?

Then stand back and ask the real question, of which those six are only evidence: is the voice that of the government?

What the test achieved, and what it left unstable

What it achieved. It made article 12 follow public power rather than legal form, which is what the drafters' inclusive definition invited. Through the 1980s it brought a very wide range of bodies within Part III, and gave employees and contractors of public sector undertakings the protection of articles 14 and 16.

What it left unstable. A list of six indicia, none decisive, with no rule about how to weigh them, is not a test so much as a licence to reach either answer. Different Benches applied it differently, and a body could be held to be State in one case and not in another on facts that were hard to distinguish. That instability is what produced the correction described in the next chapter.

A worked example

A society is registered under the Societies Registration Act to run a college. Its governing body has eleven members, seven nominated by the State Government. Ninety-five per cent of its recurring expenditure is met by State grants. Its rules require the Government's prior approval, its accounts are submitted to the Government, it must comply with directions issued by the Government, and it cannot sell its buildings without approval. It was formed by transferring a Government-run college into it. A candidate refused admission alleges that the admission procedure is arbitrary.

Is the society "State"?

Ownership and composition. Seven of eleven members are Government nominees, so the composition is dominated by the Government.

Funding. Ninety-five per cent of recurring expenditure is extensive and unusual financial assistance, not an ordinary grant available across a sector.

Control. Prior approval of rules, submission of accounts, compliance with directions and a bar on disposing of property are the marks of deep and pervasive control, and they are particular to this body rather than applying to every society in the State.

Function. Education, closely related to a governmental function.

Origin. A Government college was transferred into it.

Legal form. A registered society, not a statutory corporation. On Ajay Hasia that makes no difference, because the tests are the same and the form in which the State chooses to act cannot decide whether a fundamental right applies.

Answer. The society is State, the voice being that of the Government, and article 14 binds the admission procedure. On the merits, an interview carrying a third of the marks and conducted in two or three minutes on questions about parentage and residence was held excessive and arbitrary in Ajay Hasia itself.

Quick revision

• The test asks whether a body is an instrumentality or agency of the State, because the State's choice of legal form cannot decide whether Part III applies.
• R.D. Shetty: no cut and dried formula. Indicia are entire share capital, extensive and unusual financial assistance, State conferred or protected monopoly, deep and pervasive control, public functions closely related to governmental functions, and transfer of a government department.
• R.D. Shetty is also the source of the rule that every action of the executive must be informed with reason and free from arbitrariness.
• Ajay Hasia: the same tests apply to a registered society, and control so deep and pervasive that the body is a projection of government makes it State. The voice is that of the government.
• Six indicia, none decisive, and no rule for weighing them. That is what made the test unstable and invited the retreat.

Test yourself

1. Why can the State's choice of legal form not decide whether Part III applies?

2. List the six indicia from R.D. Shetty, and say what R.D. Shetty said about whether they amount to a formula.

3. What does R.D. Shetty say about arbitrariness and the rule of law, and where else in this book does that proposition reappear?

4. What did Ajay Hasia add to R.D. Shetty?

5. What made the test unstable, and what did that instability lead to?

Why it matters. Four changes in one paragraph. The indicia become evidence, not tests, so satisfying one is not enough. The question becomes cumulative: the facts are weighed together. The domination must be financial, functional and administrative, which is conjunctive in form and has been treated as requiring an overall picture rather than each separately. And regulatory control is expressly excluded, which is the new element and the one that decides cases.

The regulatory control exception

This is the part to understand rather than memorise, because it is where the modern cases turn.

Pervasive control is control over how the body conducts itself: who runs it, what it may spend, what rules it makes, whether it may dispose of its property, what directions it must obey. It is the control an owner exercises.

Regulatory control is control of the kind the State exercises over an entire field of activity, applying to everyone in that field alike. Licensing, inspection, safety standards, reporting requirements and approval of tariffs are regulatory. A bank is heavily regulated and is not thereby an instrumentality of the Reserve Bank.

The distinction is one of kind and of particularity, not of degree. The Court said the control must be "particular to the body in question". A body subject to intense regulation shared with all its competitors is regulated; a body subject to direction that applies to it alone is controlled.

The test applied, with the opposite result

Facts. Zee Telefilms Ltd v. Union of India, decided 2 February 2005 by five judges. Zee challenged the termination of its telecast rights by the Board of Control for Cricket in India, and the preliminary question was whether the Board is "State" so that article 32 lies against it.

Held. The Board is not "State". Applying Pradeep Kumar Biswas, the established facts were that the Board is not created by a statute; no part of its share capital is held by government; practically no government financial assistance is given towards its expenditure; it enjoys a monopoly in cricket but that status is not State conferred or State protected; there is no deep and pervasive State control, and such control as exists is only regulatory in nature, of the kind applicable to other similar bodies and not exercised under any special statute; not all its functions are public functions or closely related to governmental functions; and it was not created by transferring a government owned corporation.

But the Court added the qualification that matters. The Board does discharge duties, such as selecting the Indian team and controlling the activities of players, which are akin to public duties or State functions. Although a remedy under article 32 is not available against it, that does not mean the violator of a right goes free: an aggrieved party may seek a remedy in the ordinary course of law, or by a writ petition under article 226, which is much wider than article 32.

Why it matters. It shows what the narrowed test does in practice, and it supplies the answer to the objection that the narrowing leaves people unprotected. The protection moves from article 32 to article 226, from the fundamental rights to the public function jurisdiction, and that is [Private Bodies and Public Functions].

What survived

Three things from the earlier line are untouched, and an answer should say so rather than presenting the retreat as a reversal.

Legal form is still irrelevant. Ajay Hasia's central proposition, that the tests apply equally to a statutory corporation, a government company and a registered society, was not doubted. CSIR itself is a registered society and was held to be State.

The indicia are still the evidence. The six considerations from R.D. Shetty remain the facts a court looks for. What changed is that they are weighed rather than counted.

Deep and pervasive control still makes a body State. The formula in Ajay Hasia, that a body so controlled is a projection of government and the voice is that of the government, remains good.

What went is the possibility of establishing that a body is State by pointing to one indicium alone, and the treatment of ordinary sectoral regulation as if it were control.

The criticism

Two lines, and an LL.M. answer should engage with both.

For the retreat. A definition that reaches every body receiving public money or subject to regulation would make Part III apply to most of organised life, which is not what article 12 says and not what the drafters intended. The Court restored a boundary that had become unfindable.

Against it. The narrowing arrived exactly when the State was withdrawing from direct provision, so the test contracted at the moment when the functions were moving out of government hands. The combination is that neither the departing State nor the arriving private body is caught: the function is performed, the power is exercised, and Part III reaches neither. That is MU's own topic, and it is [Liberalisation, Privatisation and the Case for Widening Article 12].

A worked example

A national sporting federation selects the country's team, frames the rules of the sport, disciplines players, and is the only body through which anyone can play the sport at national level. It is a registered society. It receives no government funding, no share capital is held by government, and the Ministry's involvement is limited to clearances for foreign tours of the kind every sporting body needs. A player suspended without a hearing petitions the Supreme Court under article 32.

Apply Pradeep Kumar Biswas, not Ajay Hasia's list. The question is whether, on the CUMULATIVE facts, the federation is financially, functionally and administratively dominated by or under the control of the Government, that control being particular to it and pervasive.

Financially: no. Administratively: no; the Ministry's clearances are the kind required of every similar body. Functionally: it performs functions of public importance, but that alone is one indicium and, since Pradeep Kumar Biswas, no single indicium is enough.

The monopoly point. The federation does enjoy a monopoly, but the question under R.D. Shetty was always whether the monopoly is State conferred or State protected. A monopoly a body achieved for itself does not count.

Is the control regulatory or pervasive? Clearances applying to the whole field, not exercised under any special statute directed at this body, are regulatory. Pradeep Kumar Biswas holds expressly that merely regulatory control does not make a body a State.

Answer on article 32. The petition fails at the threshold. This is Zee Telefilms.

But the player is not without a remedy. Selecting a national team and disciplining players are akin to public duties, and article 226 reaches any person or authority and extends for any other purpose. The High Court is the forum.

Quick revision

• Pradeep Kumar Biswas, seven judges: the Ajay Hasia tests are not rigid; the question is whether the cumulative facts show the body is financially, functionally and administratively dominated by or under the control of government; that control must be particular and pervasive; merely regulatory control does not suffice.
• Regulatory control is control over a whole field applying to everyone alike; pervasive control is particular to the body and is the control of an owner.
• Zee Telefilms: the BCCI is not State, because its monopoly is not State conferred and its control is only regulatory. But its functions are akin to public duties, and article 226 is available even though article 32 is not.
• What survived: legal form is irrelevant, the indicia remain the evidence, and deep and pervasive control still makes a body State.
• The criticism: the test narrowed at the moment the State began withdrawing, so the two movements together leave a gap.

Test yourself

1. Why was a Bench of seven constituted, and what two problems had the Ajay Hasia formulation produced?

2. Quote the test in Pradeep Kumar Biswas and identify the four changes it made.

3. Distinguish pervasive control from regulatory control, and say why the distinction is one of kind and particularity rather than degree.

4. On what facts did Zee Telefilms hold the BCCI not to be State, and what did the Court add about the remedy?

5. What survived the retreat? Name three things.

6. State the criticism that the retreat came at the wrong moment.

Held. By the majority, the order was within the inherent power of the High Court. A court has inherent jurisdiction to hold a trial in camera where the ends of justice clearly and necessarily require it, and the power to prohibit excessive publication of part of the proceedings is a facet of that power; section 151 of the Code of Civil Procedure saves the inherent power to make such orders as are necessary for the ends of justice.

On the constitutional question, the principle that has been carried forward was put by Sarkar J.: if a judicial tribunal makes an order which it has jurisdiction to make, by applying a law which is valid in all respects, the order cannot offend a fundamental right. An order is within jurisdiction if the tribunal had jurisdiction to decide the matters litigated before it and the law it applied was valid, and a tribunal having that jurisdiction does not act without jurisdiction if it makes an error in the application of the law. The order was a judicial order within the jurisdiction of the judge who made it, even though it restrained persons who were not parties. He added that the law empowering a court to prohibit publication of its proceedings is in any event protected by article 19(2), relating as it does to contempt of court.

Hidayatullah J. dissented, holding that a court which had heard a witness in open court could not suppress publication of that evidence on the witness's request that his business would suffer, and that section 151 could not be used to turn proceedings which should be open into closed ones.

Why it matters. It is the authority for the proposition that a judicial order of a competent court is not challengeable under article 32 as violating a fundamental right. Note carefully what the reasoning is: not that the judiciary is outside article 12, but that an order within jurisdiction applying valid law cannot by definition contravene Part III, and that an error of law within jurisdiction is not a constitutional violation. The remedy for an erroneous order is appeal or review, not article 32.

Where the line actually runs

Three refinements are worth having.

An order without jurisdiction is a different case. Mirajkar's reasoning depends on the order being within jurisdiction. Where a court acts wholly without jurisdiction, or where the law it applies is itself invalid, the reasoning does not apply, and the order may be attacked.

A judicial order can be corrected by the Supreme Court on its own account. The Court has recognised a jurisdiction to reconsider its own final judgment where a gross miscarriage of justice or a violation of natural justice has occurred, exercised by a petition heard by a Bench of senior judges. That is a self-correcting mechanism inside the judicial hierarchy, not an application of article 12, and the decisions establishing it are in the case bank's queue and are named here without citations.

A court that violates natural justice is amenable to correction. Where a court makes an order affecting a person without hearing him, the objection is real, but it is taken by appeal, review or the Supreme Court's own corrective jurisdiction, and not by treating the court as a violator of Part III.

Article 12 and the judiciary in other jurisdictions

Worth one sentence in an answer, because the comparison sharpens the Indian position.

In the United States the constitutional guarantees are treated as binding on the courts as organs of the State, and a judicial order enforcing a private discriminatory agreement has been held to be state action. India has not gone that way. The reason is structural: the Indian Constitution provides an integrated hierarchy with appeals to the Supreme Court and article 136 special leave over every court and tribunal in the country, so the machinery for correcting judicial error already exists and does not need to be built out of article 12.

A worked example

A High Court, on its administrative side, frames a rule for recruitment to the subordinate judiciary which a candidate says discriminates on a prohibited ground. In a separate matter, a district judge passes an order in a civil suit which a stranger to the suit says violates his right under article 19(1)(a). Both petitioners move the Supreme Court under article 32.

The recruitment rule. Here the High Court is acting administratively, under article 235 in relation to control over the subordinate courts, and its rule-making power under article 227(2). In that capacity it is treated as State: the rule is subordinate legislation, it is "law" for article 13, and articles 14 and 16 apply to it as they would to any public employer. The petition is maintainable.

The judicial order. Different in kind. On Mirajkar, if a judicial tribunal makes an order which it has jurisdiction to make, by applying a law valid in all respects, the order cannot offend a fundamental right; and a tribunal does not act without jurisdiction merely because it errs in applying the law. The order in Mirajkar bound journalists who were not parties, and that did not change the analysis.

So what must the stranger show? Not that the order was wrong. He must show either that the court acted wholly without jurisdiction, or that the law it applied was itself invalid. Neither is an ordinary complaint about an erroneous order.

What is his remedy? Appeal, revision or review within the hierarchy, and in an extreme case the Supreme Court's own corrective jurisdiction over its final judgments.

The reasoning to notice. It is not that the judiciary is outside article 12. It is that an order within jurisdiction applying valid law cannot by definition contravene Part III.

Quick revision

• Article 12 does not mention the judiciary, and because the definition is inclusive the omission does not settle the question.
• On its rule-making and administrative sides, including employment under articles 229 and 235, the judiciary is treated as State.
• Mirajkar: an order made by a court within its jurisdiction, applying valid law, cannot offend a fundamental right, and an error within jurisdiction is not a constitutional violation.
• The reasoning is about the nature of a judicial order, not about the judiciary being outside article 12.
• The remedy for an erroneous judicial order is appeal, review or the Supreme Court's corrective jurisdiction, not article 32.
• India has not adopted the American state-action approach to judicial orders, because the integrated hierarchy and article 136 already supply the correction.

Test yourself

1. Why would it be unacceptable to hold that the judiciary is State for all purposes? And why unacceptable to hold that it is never State?

2. Distinguish the three capacities in which the judiciary acts, and give the article that governs the second.

3. State the principle in Mirajkar in Sarkar J's words, and say what Hidayatullah J. held.

4. In what circumstances does Mirajkar's reasoning not apply to a judicial order?

5. Why has India not followed the American state-action approach to judicial orders?

What makes a function public

There is no exhaustive definition, and an answer should say so and then give the indicia.

The function is one the State would otherwise perform. Education, health, the supply of essential services, the allocation of a public resource.

The body has a monopoly or near monopoly over it, so that a person who wants the service has nowhere else to go.

The body's decisions affect the public or a section of it, not only the parties to a contract.

The body has been entrusted with the function by statute or by the State, or has been permitted to occupy the field in place of the State.

There is a public element in the duty, meaning an obligation owed to the public rather than only under a private agreement.

The consistent limitation is that a purely contractual obligation is not a public duty. A private employee dismissed in breach of contract by a private employer has a suit, not a writ, however large the employer. The public element must be in the duty sought to be enforced, not merely in the identity of the body.

The authority

Facts. Zee Telefilms Ltd v. Union of India, decided 2 February 2005 by five judges. Zee challenged the termination of its telecast rights by the Board of Control for Cricket in India. The preliminary question was whether the Board is "State" so that article 32 lies.

Held. The Board is not "State", because it is not created by statute, no share capital is held by government, it receives practically no government funding, its monopoly is not State conferred or State protected, and the control over it is only regulatory. But the Board does discharge duties, such as selecting the Indian cricket team and controlling the activities of players, which are akin to public duties or State functions; and although the remedy under article 32 is not available, that does not mean the violator of a right goes free. An aggrieved party may seek a remedy in the ordinary course of law or by way of a writ petition under article 226, which is much wider than article 32. The Court cited Andi Mukta Sadguru Trust v. V.R. Rudani for the proposition that article 226 confers wide powers on the High Courts to issue writs, a striking departure from English law.

Why it matters. It is the clearest modern statement that the two jurisdictions are not coextensive, and that the narrowing of article 12 does not leave the citizen without a public law remedy. It also identifies the test on the article 226 side as the nature of the duty, not the status of the body.

Andi Mukta is the leading authority on the public duty test and it is in the case bank's queue; its report has not been read in the session that wrote this chapter, so it is named without a citation and nothing is attributed to it beyond the proposition Zee Telefilms itself quotes.

How to answer a problem

Four steps, in order.

One, is the body State under article 12? Apply Pradeep Kumar Biswas: cumulative facts, financial, functional and administrative domination, particular and pervasive, and not merely regulatory. If yes, article 32 and article 226 are both available and Part III binds directly.

Two, if not, is the duty a public duty? Look at the function, the monopoly, the effect on the public, the entrustment, and the public element. If yes, article 226 is available.

Three, is the obligation purely contractual? If the complaint is only that a contract has been broken, the writ jurisdiction will usually be declined however public the body, and the remedy is a suit.

Four, what right is being enforced? Under article 226 it need not be a fundamental right: "any other purpose" covers a statutory right and a legal right generally.

What this leaves unresolved

Two things, and an LL.M. answer gains by naming them.

The remedy is procedural, not substantive. Article 226 gives access to a court. It does not make Part III bind the private body. So a court may quash an arbitrary decision as a breach of a public duty, and may draw on the standards of article 14 in deciding what arbitrariness means, without holding that the body was bound by article 14 at all. Whether that distinction survives scrutiny is a fair question.

It works only where a High Court will take the case. Article 226 is discretionary, and the discretion is exercised against a petitioner with an alternative remedy. A person dealing with a large private provider of an essential service may be told to sue, which is precisely the outcome the doctrine was developed to avoid.

A worked example

A private university, established under a State Act but privately funded and privately managed, expels a student without a hearing. Its own statutes require a hearing. The student petitions the High Court under article 226 and, in the alternative, the Supreme Court under article 32.

Article 32 first. The university must be State under article 12. Being established under a State Act is not enough on its own: the question is cumulative domination that is particular and pervasive, and privately funded and privately managed points away from it. Assume it fails. Article 32 is unavailable.

Article 226. The words are "any person or authority", and the jurisdiction extends "for any other purpose", so it reaches beyond the State and beyond fundamental rights. Ask whether there is a public function or a public duty.

Is there? Conferring degrees is a function the State would otherwise perform and which the university performs under statutory authority. Its decisions affect a section of the public. Expulsion determines a student's academic standing in a way no private contract does. There is a public element.

The limitation to test. A purely contractual obligation is not a public duty. If the complaint were about fees or a hostel booking, the writ would likely be refused. But the duty here arises under the university's own statutes made under the Act, not under a private agreement, and depriving a student of a hearing those statutes require is a breach of a public duty.

Answer. Article 32 fails; article 226 lies. Note what that gives her: a forum and a remedy, not a holding that the university is bound by article 14. The court may still measure the decision by standards of fairness drawn from administrative law.

Quick revision

• Article 226 reaches "any person or authority" and extends "for any other purpose", so it is wider than article 32 in both respects.
• A body outside article 12 may still be amenable to article 226 if it discharges a public function or owes a public duty.
• The indicia: a function the State would otherwise perform, monopoly, effect on the public, entrustment, and a public element in the duty.
• A purely contractual obligation is not a public duty, however large the body.
• Zee Telefilms: the BCCI is not State, but its functions are akin to public duties and article 226 is available, being much wider than article 32.
• The limitation: article 226 supplies a forum, not the substantive application of Part III to a private body.

Test yourself

1. Compare articles 32 and 226 on four points: against whom, for what, where, and with what discretion.

2. What are the indicia that a function is public?

3. Why is a purely contractual obligation not a public duty, however large the body?

4. What did Zee Telefilms say about the remedy available against a body that is not State?

5. "Article 226 supplies a forum, not the substantive application of Part III." Explain, and say why that matters.

Reservation. Reservation under articles 15(4), 15(5), 16(4) and 16(4A) binds the State. It does not bind a private employer, and where the State transfers an undertaking, the obligation goes with it. This is exactly what MU asks in its other recurring question, and the two topics meet here: see [Privatisation and Affirmative Action].

Fairness in contracting. The rule in R.D. Shetty, that every action of the executive must be informed by reason and free from arbitrariness, governs how a public body awards a contract. A private body allocating the same opportunity is under no such duty.

Reasons, hearing and non-arbitrariness generally. The whole apparatus of administrative law that attaches to State action attaches to the body, not to the function.

The direct remedy. Article 32 is itself a fundamental right and cannot be refused for the existence of an alternative remedy. Article 226 is discretionary.

The arguments for widening

The function argument. Part III protects people against the exercise of public power. What matters is that the power is public, not who holds it. If the State may divest itself of constitutional obligations by divesting itself of the enterprise, the obligations are worth only as much as the State's willingness to keep the enterprise.

The monopoly argument. The justification for treating the State differently from a private person is that a person cannot walk away from it. Where a private body holds a monopoly over an essential service, the citizen cannot walk away from it either, and the reason for the special treatment applies unchanged.

The textual argument. Article 12 says "includes", and the definition operates "unless the context otherwise requires". Both phrases invite an interpretation responsive to changed conditions, and the Court used them to widen the definition once already, in the era when the State was expanding. Nothing in the text confines the technique to one direction.

The consistency argument. Ajay Hasia held that the form in which the State chooses to act cannot decide whether a fundamental right applies. Privatisation is a choice of form on the largest possible scale.

The arguments against

The boundary argument. If public function is enough, the definition has no stopping point. Private schools, hospitals, banks, insurers, transport operators and employers all perform functions the State performs somewhere, and treating them all as State would make Part III the general law of the land, which the Constitution did not do. Part III was written against the State because the State is uniquely dangerous, and diluting the distinction weakens rather than strengthens it.

The remedy argument. The gap is smaller than it looks, because article 226 already reaches a public function or a public duty, as Zee Telefilms held. What is lost is the article 32 route, not the substantive protection.

The legislative argument. Where a private body needs to be bound by an equality obligation, that is properly done by legislation, which can be tailored, debated and confined to the sector. Consumer protection, anti-discrimination provisions in specific fields, sectoral regulators with fairness obligations and licence conditions are the appropriate instruments, and Parliament may enact them under article 35 and its general powers.

The federalism and separation argument. Deciding that a whole class of private activity is bound by Part III is a policy judgment with large economic consequences, and a court reaching it by reinterpreting a definition is doing what a legislature should do.

The middle positions

Most serious writing does not take either extreme, and an answer that presents the middle ground is doing better than one that picks a side.

Function-specific application. A body is State for the purposes of the particular function it performs publicly, and not for everything it does. Article 12's own words, "unless the context otherwise requires", would support this, and it mirrors the treatment of the judiciary in [Is the Judiciary "State"?].

Duty-specific application through article 226. Leave article 12 alone and develop the public duty jurisdiction, which is what the Court has in fact done. Its weakness is that it gives a forum rather than a right.

Obligations that travel with the enterprise. Require, by the terms of the disinvestment or the concession, that the transferee assume specified obligations. This is a contractual and regulatory solution to a constitutional problem, and it is the one governments have generally adopted.

Horizontal application of particular rights. Some rights in Part III already bind private persons: article 15(2) on access to shops, public restaurants, hotels, places of public entertainment, wells, tanks, bathing ghats, roads and places of public resort; article 17 on untouchability; article 23 on traffic in human beings and forced labour; and article 24 on the employment of children. The Constitution therefore already knows how to bind a private person where it wants to, and the argument from that is two-edged: it shows the technique exists, and it shows the drafters used it selectively.

Writing the answer

MU's question is "explain the need for redefining the State in view of the policy of economic liberalisation", or "in the backdrop of liberalisation justify the need to widen the nature and scope of Article 12". Both invite a case for widening, and a good answer makes that case without pretending the other side does not exist.

A structure that works: what article 12 is and why it matters; how the definition was widened when the State expanded, with Rajasthan Electricity Board, R.D. Shetty and Ajay Hasia; how it was narrowed in Pradeep Kumar Biswas and Zee Telefilms; what liberalisation did over the same period; what is actually lost, with the concrete examples above; the arguments each way; the middle positions; and a conclusion. The conclusion that the material supports is that the gap is real, that article 226 has absorbed part of it, and that the part it has not absorbed, the substantive application of equality to bodies exercising monopoly public power, is the part that still needs an answer.

A worked example

A State sells its entire shareholding in a State-owned bank to a private group. Before the sale, the bank reserved posts under articles 16(4) and 16(4A), followed a published and non-arbitrary recruitment procedure, and was answerable under article 32. After the sale it does none of these. A candidate refused a reserved post asks what the Constitution now gives her.

Article 12. Apply Pradeep Kumar Biswas. No shareholding, no funding, no particular and pervasive control; banking regulation applies to the whole sector and is regulatory. The bank is no longer State. So articles 14 and 16 do not bind it and article 32 does not lie.

Reservation. Articles 16(4) and 16(4A) are enabling provisions addressed to the State. They do not bind a private employer, and no obligation to reserve survives the transfer unless the terms of the sale imposed one.

Article 226. Banking is a public function in a loose sense, but recruitment by a private employer is ordinarily a matter of contract, and the writ jurisdiction is generally declined for it. So the article 226 route, which answers the Zee Telefilms problem, does not obviously answer this one.

What is actually left. Labour legislation, any anti-discrimination provision that binds private employers in the field, the terms of the disinvestment if they imposed obligations, and nothing else of constitutional force.

The point MU's question is asking. The function has not changed and the power over the candidate has not changed. What changed is who holds it, and on the present law that decides whether the Constitution applies. Whether that should be so is the argument, and a good answer gives the case for widening, the boundary and remedy objections against it, and at least one of the middle positions.

Quick revision

• Two movements: the State withdrew from provision after 1991, and the definition of State narrowed after 2002. The problem is the combination.
• What is lost: articles 14 and 16 in employment, reservation, the duty of non-arbitrariness in contracting, and the article 32 remedy.
• For widening: public power is what Part III is about; monopoly removes the citizen's exit; "includes" and "unless the context otherwise requires" invite it; and Ajay Hasia said form cannot decide.
• Against: no stopping point; article 226 already covers public duty; legislation is the proper instrument; and it is a policy judgment.
• Middle positions: function-specific application, the article 226 public duty route, obligations attached to the transfer, and the horizontal rights the Constitution already has in articles 15(2), 17, 23 and 24.

Test yourself

1. Describe the two movements, in the economy and in the doctrine, and explain why the combination is the problem.

2. Name four things that are lost when a public sector undertaking is privatised.

3. Give three arguments for widening the definition and three against.

4. Set out four middle positions between leaving article 12 alone and abandoning it.

5. Which provisions of Part III already bind private persons, and what does their existence prove, for and against?

6. Answer MU's stem: "In the backdrop of Liberalisation justify the need to widen the nature and scope of Article 12 of the Constitution.

Held. By a majority, article 13(1) does not make an existing inconsistent law void ab initio. It renders such a law ineffectual and void only with respect to the exercise of fundamental rights, and only on and after the commencement of the Constitution. It has no retrospective effect. So an act done before the Constitution, in contravention of a law that was valid at the time, may still be prosecuted and punished after the Constitution comes into force. Fazl Ali and Mukherjea JJ. dissented.

Why it matters. It settles the temporal reach of article 13(1), and it establishes the proposition on which eclipse depends: "void" in article 13(1) means unenforceable going forward, not obliterated. A pre-Constitution law remains on the statute book and remains effective for everything it did before 26 January 1950, and for anything the fundamental rights do not touch.

Two corollaries follow and are examinable. A pre-Constitution law inconsistent with a right conferred only on citizens remains fully effective against non-citizens, because as to them there is no inconsistency. And article 13(1) operates only on the inconsistency: the rest of the law stands, which is where severability comes in.

What counts as "law"

Article 13(3)(a) is deliberately wide, and the width is the point: a prohibition that applied only to statutes could be evaded by using any other instrument.

Included expressly: an Ordinance, an order, a bye-law, a rule, a regulation, a notification, and custom or usage having the force of law.

Included by decision: the regulations of a statutory corporation.

Facts. Sukhdev Singh v. Bhagatram Sardar Singh Raghuvanshi, AIR 1975 SC 1331, (1975) 1 SCC 421, decided 21 February 1975. Employees of the Oil and Natural Gas Commission, the Life Insurance Corporation and the Industrial Finance Corporation were removed from service contrary to the regulations framed under the Acts constituting those bodies, and claimed a declaration that they continued in service rather than damages.

Held. The regulations have the force of law. They are subordinate legislation made under a specific statutory power, binding on the bodies themselves, which have no free hand to depart from them, and are not merely terms of a contract of employment. It follows that the employees hold a statutory status. The three bodies are also authorities within article 12.

Why it matters here. Their regulations are therefore law made by the State, so article 13(2) applies to them and a regulation that abridges a fundamental right is void. It also shows how far article 13(3)(a) reaches: the instrument need not be a statute or even a rule made by government.

Custom and usage. Expressly included where it has the force of law. Whether personal law, meaning the religious law of a community as distinct from a custom, is "law" for article 13 is a distinct and unresolved question, and it belongs with [Women, Religion and Personal Law] rather than here. An answer should note that it is open rather than assert an answer.

Administrative instructions that do not have the force of law are generally not "law" for article 13, though State action taken under them is still State action and is tested against Part III directly.

Article 13(4) and constitutional amendments

Article 13(2) forbids the State to make any law taking away or abridging Part III. Parliament is the State. Is a constitutional amendment a "law"?

The question produced the longest argument in Indian constitutional history: the early view that an amendment is not law for article 13(2); the reversal holding that it is, so that Parliament could not abridge a fundamental right; the Twenty-fourth Amendment, which inserted article 13(4) providing that nothing in article 13 shall apply to any amendment made under article 368, and correspondingly amended article 368; and the eventual answer that Parliament may amend any provision but may not damage the basic structure.

That sequence is [The Amending Power] and [The Basic Structure], and it should be cross-referenced rather than compressed here. What belongs in this chapter is the narrow point: since the Twenty-fourth Amendment, article 13 does not apply to a constitutional amendment, and the limit on the amending power is not article 13 but the basic structure doctrine.

Article 372 and the continuance of existing laws

Article 13(1) operates on "laws in force" and article 372 is what keeps them in force to operate on.

Article 372(1) provides that notwithstanding the repeal of the enactments referred to in article 395, all the law in force in the territory of India immediately before the commencement of the Constitution shall continue in force until altered, repealed or amended by a competent legislature or other competent authority. Article 372(2) empowered the President, for a limited period, to make adaptations and modifications to bring such laws into accord with the Constitution.

So the scheme is: article 372 continues the old law; article 13(1) tests it against Part III and renders it void to the extent of any inconsistency; and article 13(2) governs everything enacted afterwards.

The order of enquiry

For any challenge, four questions in this order.

One, is the actor "State" within article 12? If not, Part III does not bind it directly, and the enquiry moves to article 226 and the public duty route.

Two, is the impugned instrument "law" within article 13(3)(a)? A statute, an Ordinance, a rule, a regulation, a bye-law, a notification, or a custom with the force of law. If it is not law, the action is still tested as State action, but article 13 does not operate on the instrument.

Three, was it made before or after the Constitution? Before, and article 13(1) applies prospectively on Keshavan Madhava Menon. After, and article 13(2) applies as a prohibition on power.

Four, how much is void? Only "to the extent of the inconsistency" or "to the extent of the contravention". That is severability, in the next chapter.

A worked example

A pre-Constitution Act punishes the publication of specified political material. In 1949 X publishes such material and is prosecuted; the trial is still pending on 26 January 1950. In 1951 the same Act is applied to Y, a non-citizen, and in 1952 to Z, a citizen. Assume the Act is inconsistent with article 19(1)(a), which is a right of citizens only.

X. Article 13(1) is prospective. On Keshavan Madhava Menon it does not make an existing inconsistent law void ab initio; it renders it ineffectual only with respect to the exercise of fundamental rights and only on and after the commencement of the Constitution. X's act was done in 1949 under a law then valid, and the prosecution may proceed.

Z. The Act is inconsistent with a right Z holds, so as against Z it is void to the extent of the inconsistency from 26 January 1950. The prosecution fails.

Y. Article 19 is conferred on citizens only. As against a non-citizen there is no inconsistency at all, so the Act remains fully effective and the prosecution succeeds. This is the corollary of Keshavan Madhava Menon that most answers miss.

Suppose the Constitution is later amended so that the restriction becomes permissible. The Act is not dead and needs no re-enactment: on Bhikaji Narain it was overshadowed, not obliterated, and when the shadow lifts it operates again.

And if the same Act had been passed in 1955? Then article 13(2) applies, the legislature had no power to make it, and the analysis is different. Whether eclipse can help is the Deep Chand question, and the answer given is that it can where the right is conferred on citizens alone.

Quick revision

• Article 13(1): pre-Constitution law is void to the extent of inconsistency with Part III. Article 13(2): the State shall not make any law taking away or abridging Part III, and such a law is void to the extent of the contravention.
• Keshavan Madhava Menon: article 13(1) is prospective, not retrospective. An offence committed before the Constitution under a then valid law may still be prosecuted.
• "Law" in article 13(3)(a) includes an Ordinance, order, bye-law, rule, regulation, notification and custom or usage having the force of law.
• Sukhdev Singh: the regulations of a statutory corporation have the force of law, so they are within article 13.
• Whether personal law is "law" for article 13 is unresolved and should be presented as open.
• Article 13(4), inserted by the Twenty-fourth Amendment, takes constitutional amendments outside article 13. The limit on amendment is the basic structure, not article 13.
• Article 372 continues pre-Constitution law; article 13(1) then tests it.

Test yourself

1. What is the difference in character between article 13(1) and article 13(2), and what doctrinal consequence follows?

2. State the holding in Keshavan Madhava Menon and its two corollaries.

3. What does "law" include under article 13(3)(a)? Name two things included by decision rather than by the words.

4. What did article 13(4) do, and what is now the limit on the amending power?

5. How do articles 372 and 13(1) work together?

6. Set out the four questions, in order, for any challenge under article 13.

Facts. Deep Chand v. State of Uttar Pradesh, decided 15 January 1959. The Uttar Pradesh Transport Service (Development) Act 1955 authorised a State road transport scheme; Parliament then legislated on the same subject. The question was the effect on the State Act and on a scheme already framed under it.

Held. The State Act was not wholly void and continued to support the scheme already framed; even if the central Act had repealed it, section 6 of the General Clauses Act would have saved the scheme. On eclipse, Das C.J. and Sinha J. held that there is no reason why the doctrine as explained in Bhikaji Narain should not also apply to a post-Constitution law that infringes a fundamental right conferred on citizens alone: such a law is shadowed as against citizens but remains effective against non-citizens, and when the shadow is removed by amendment it applies to citizens without re-enactment.

Why it matters. It supplies the qualified answer that is now generally given: eclipse applies to a post-Constitution law where the right infringed is one conferred only on citizens, because as against non-citizens there was never any inconsistency and the law was always valid for them. Where the right is conferred on all persons, the law is void from inception and there is nothing to eclipse.

The citation recorded in the case bank for Deep Chand is marked unread: the report's own equivalent-citations line was not captured when the judgment was fetched. The holding above was read in full from the judgment, but the AIR and SCR references should be confirmed before they are written into an answer.

Severability

The problem it answers. A statute is usually a package of many provisions. If one of them offends Part III, is the whole Act void, or only that provision?

Article 13 answers in principle: void to the extent of the inconsistency or contravention. Severability is the working out of that phrase.

Facts. R.M.D. Chamarbaugwalla v. Union of India, AIR 1957 SC 628, 1957 SCR 930, decided 9 April 1957. Promoters of prize competitions challenged sections 4 and 5 of the Prize Competitions Act 1955 and rules 11 and 12. They argued that "prize competition" as defined covered not only gambling but competitions where success depended substantially on skill, that restricting the latter breached article 19(1)(g), and that the enactment was a single inseverable whole which must therefore fail entirely.

Held. Restrictions on gambling competitions were unassailable, gambling not being within article 19(1)(g) at all. On construction the Act reached only competitions of a gambling nature. But even assuming it went wider, the provisions were severable in their application and remained valid as to gambling. The Court also laid down that in construing an enactment a court ascertains the legislature's intention not merely from the literal words but from the history of the legislation, its purpose and the mischief it seeks to suppress.

Why it matters. It is the leading Indian authority on severability, and it establishes both forms of the doctrine: severability of provisions, where the bad section is cut out, and severability of application, where the same provision is upheld as to part of its field and struck down as to the rest.

The working principles, drawn from the case law and worth reproducing as a list:

• The intention of the legislature is the determining factor: would it have enacted the valid part alone?
• If the valid and invalid parts are so inextricably mixed that they cannot be separated, the whole falls.
• If they are distinct and separate, only the invalid part falls, even if they are contained in the same section.
• If what remains after severance is so thin and truncated as to be substantially different from what the legislature intended, the whole falls.
• If the valid and invalid parts form part of a single scheme intended to operate as a whole, the whole falls.
• Severability is determined on the Act as a whole, not on the particular provision alone, and the preamble, history and object may be looked at.

Waiver

The question. A fundamental right is a right. May the person entitled to it agree not to insist on it, so that the State may then do what the right would otherwise forbid?

Facts. Basheshar Nath v. Commissioner of Income-tax, Delhi, AIR 1959 SC 149, 1959 Supp (1) SCR 528, decided 19 November 1958. The appellant's case had been referred to the Income-tax Investigation Commission under section 5(1) of the Taxation of Income (Investigation Commission) Act 1947, a provision already declared void as discriminatory under article 14. He nevertheless applied for and obtained a settlement under section 8A, agreeing to pay a large sum. The revenue argued that even if the section was void he had waived his right by settling.

Held. The settlement was covered by the earlier decision and violated article 14. On waiver, Bhagwati and Subba Rao JJ. held that there can be no waiver of the right in article 14 or of any other fundamental right in Part III. The Constitution draws no distinction between rights enacted for an individual's benefit and rights enacted in the public interest or on grounds of public policy, and there is no justification for importing American notions to whittle down the transcendental character of those rights. Article 13(2) is a constitutional mandate addressed to the State, and no citizen can by waiver relieve the State of the obligation it lays on it.

S.K. Das J. took a narrower view: whether waiver is possible must depend on the nature of the right and on whether it is primarily for the benefit of the individual or of the general public, and where the Constitution vested the right in the individual, primarily to benefit him, and it does not impinge on the rights of others, waiver could be possible if not forbidden by law or contrary to public policy.

Why it matters. It is the authority for the rule against waiver, and it must be cited accurately. The broad proposition is the view of two judges, with a reasoned narrower opinion alongside, and the modern statement of the rule is Bhagwati J's. The reasoning to remember is the structural one: article 13(2) is a prohibition on the State, and a prohibition on one party cannot be lifted by the consent of the other.

The related propositions. A fundamental right cannot be waived, but this does not mean that a person may never do anything that has the practical effect of losing the benefit of one: a person may decline to exercise a freedom, may enter a contract that restricts what he does, and may fail to raise a point in time. What he cannot do is confer on the State a power the Constitution denies it. Nor can there be estoppel against the Constitution, so a person who has taken a benefit under a void law is not thereby precluded from challenging it.

A worked example

A statute has twelve sections. Section 5 imposes a restriction that violates article 19(1)(g) as against citizens. Sections 1 to 4 and 6 to 12 create an entire licensing scheme which can operate without section 5, though less effectively. A licensee who accepted a licence under the Act, and took its benefits for three years, now challenges section 5. The State argues that he has waived the right, and alternatively that if section 5 goes the whole Act must go.

Severability. Article 13 makes a law void only "to the extent of" the inconsistency. On Chamarbaugwalla the question is the legislature's intention: would it have enacted the valid part alone? Here the remaining eleven sections form a workable scheme, they are not inextricably mixed with section 5, and what survives is not so truncated as to be substantially different from what was intended. Section 5 alone falls. Note also the second form of the doctrine: severability of application, by which the same provision may be upheld as to part of its field and struck down as to the rest.

Waiver. On Basheshar Nath there can be no waiver of a fundamental right. Bhagwati and Subba Rao JJ. held that the Constitution draws no distinction between rights enacted for an individual's benefit and rights enacted in the public interest, and that article 13(2) is a mandate to the State which no citizen can relieve it of. Taking the benefit of the Act does not preclude the challenge, because there is no estoppel against the Constitution. Cite the case accurately: S.K. Das J. took the narrower view that waiver might be possible for a right primarily for the individual's benefit.

And if the Constitution is later amended so that the restriction becomes permissible? Section 5 is a post-Constitution provision, so ask the Deep Chand question. Since article 19 is a citizens' right, the provision was always valid against non-citizens, and on that reasoning it can be eclipsed rather than still-born, and would revive.

Quick revision

• All three doctrines come out of the words "to the extent of".
• Eclipse: Bhikaji Narain. A pre-Constitution inconsistent law is shadowed, not obliterated, and revives without re-enactment when the inconsistency is removed.
• Eclipse and post-Constitution laws: Deep Chand supports its application where the right infringed is conferred on citizens only, because the law was always valid against non-citizens.
• Severability: Chamarbaugwalla. Both severability of provisions and severability of application. The test is the legislature's intention, and whether what survives is substantially what was intended.
• Waiver: Basheshar Nath. No waiver of a fundamental right, because article 13(2) is a mandate to the State which no citizen can relieve it of. Cite it as the view of Bhagwati and Subba Rao JJ., with S.K. Das J. taking a narrower view.
• No estoppel against the Constitution.

Test yourself

1. Which words in article 13 do all three doctrines come out of?

2. State the doctrine of eclipse and the facts of Bhikaji Narain that illustrate it.

3. Does eclipse apply to a post-Constitution law? Give the reasoning both ways and the qualified answer from Deep Chand.

4. Set out six working principles of severability, and distinguish severability of provisions from severability of application.

5. State the rule against waiver, name the judges who stated it and the judge who differed, and give the structural reason for the rule.

6. What is meant by saying there is no estoppel against the Constitution?

Held. The plea of mala fides failed on the facts. On principle, article 16 is an instance of the application of the concept of equality in article 14: article 14 is the genus, article 16 a species, and the basic principle informing both is equality and the inhibition against discrimination. Equality is a dynamic concept with many aspects and dimensions which cannot be cribbed, cabined and confined within traditional and doctrinaire limits. From a positivistic point of view, equality is antithetic to arbitrariness, and where an act is arbitrary it is by that fact unequal and violates article 14.

Why it matters. It supplies the organising idea of the whole code, and it is the passage from which the arbitrariness doctrine grows. Cite it early in any equality answer.

Who each article protects

A table students get wrong under pressure.

Article 14: any person. Citizens, non-citizens, and legal persons such as companies. This is why a foreign company may complain of discriminatory treatment.

Article 15: citizens only, and the prohibition binds the State. Clause (2), which forbids restriction on access to shops, public restaurants, hotels, places of public entertainment, wells, tanks, bathing ghats, roads and places of public resort, is one of the few provisions of Part III that binds private persons as well.

Article 16: citizens only, and confined to employment or appointment to an office under the State.

Article 17: binds everyone, including private persons. It is the clearest instance of horizontal application in Part III.

Article 18: binds the State in clauses (1) and (2) and, in clauses (3) and (4), persons holding office of profit or trust under the State.

The two kinds of clause

Within articles 15 and 16 the clauses do two different jobs, and confusing them produces a muddled answer.

Prohibitory clauses forbid: article 15(1) and (2), article 16(1) and (2). They are commands.

Enabling clauses permit: article 15(3), (4), (5) and (6); article 16(3), (4), (4A), (4B) and (6). They confer no right on anybody and impose no duty. A citizen cannot compel the State to make a reservation, because an enabling clause enables and does not oblige. That proposition is examined in [Reservation in Public Employment], and it is the single most useful thing to know about the reservation articles.

Where the module goes

The rest of topic 3 works through the code:

• Article 14's two limbs, in [Equality Before the Law and Equal Protection].
• Its first test, in [Reasonable Classification], and its second, in [Arbitrariness].
• The prohibited grounds, in [Articles 15 and 16: The Prohibited Grounds].
• Then the reservation chapters, from [Reservation in Education] to [Economically Weaker Sections].
• The lists and the Commissions, in [The Lists, and the Commissions That Watch Them].
• The doctrine, in [Protective Discrimination], and MU's own question in [Privatisation and Affirmative Action].
• The saving provisions, in [Saving of Laws: Articles 31A, 31B and 31C].
• And the two abolitions, in [Untouchability and Titles].

A worked example

A State law provides that in any prosecution under a particular Act, the accused shall be presumed guilty unless he proves otherwise, and applies only to persons carrying on a specified trade. A trader challenges it under articles 14 and 19. He is not a citizen.

Which articles are available to him?

Article 19 is not. Its rights are conferred on citizens only, and he is not one.

Article 14 is. It protects every PERSON. This is the practical reason article 14 is the floor: where the specific rights are closed to a claimant, the general guarantee is not.

On what basis is the article 14 challenge framed? Two, and both should be pleaded. First, classification: is the singling out of one trade founded on an intelligible differentia, and does that differentia have a rational nexus with the object of the Act? Second, arbitrariness: even a law that survives the classification test may be struck down if it is manifestly arbitrary, which is the Royappa line.

Would article 15 help? No, on two grounds. It protects citizens only, and trade is not one of the prohibited grounds in article 15(1), which are religion, race, caste, sex and place of birth, and those only.

The lesson. Identify the claimant before choosing the article. Half the equality problems set in examinations turn on the difference between "person" and "citizen".

Quick revision

• Five articles, one scheme: article 14 the general guarantee, articles 15 and 16 specific applications, articles 17 and 18 abolitions.
• Article 14 protects every person; articles 15 and 16 protect citizens; article 17 binds private persons too, as does article 15(2).
• Royappa: article 14 is the genus, article 16 a species; equality is a dynamic concept; equality and arbitrariness are sworn enemies.
• Articles 15 and 16 are not exceptions to article 14 but instances of it, and the enabling clauses are instances too.
• Prohibitory clauses command; enabling clauses permit, confer no right and impose no duty.
• Where a specific right is closed to a claimant, article 14 remains.

Test yourself

1. Set out the five articles of the equality code and say what job each does in the scheme.

2. Which of them protect a non-citizen, and which bind a private person?

3. What did Royappa decide about the relationship between articles 14 and 16?

4. Distinguish a prohibitory clause from an enabling clause, and state the practical consequence of the difference.

5. Why is it wrong to describe article 16(4) as an exception to article 16(1)?

The proposition that treating unequals alike is itself unequal

This is the single most useful sentence in the whole topic and it should be stated in every equality answer.

If equality meant identical treatment, then a law giving a blind candidate extra time in an examination would be unconstitutional, and so would a law taxing higher incomes at higher rates. Nobody thinks either result is right. The reason is that formal equality, treating everyone the same, produces substantive inequality where people are not the same to begin with.

So the guarantee is understood as substantive equality: like should be treated alike, and unlike differently in proportion to their unlikeness. The whole of the reservation jurisprudence rests on this, and so does the answer to the standard objection that reservation is itself discrimination. See [Protective Discrimination].

What article 14 forbids

Three things, and they are worth separating because a problem may involve any of them.

A law that classifies unreasonably. The classification test, in [Reasonable Classification].

A law or executive act that is arbitrary. The arbitrariness test, in [Arbitrariness].

Unequal administration of an equal law. A law fair on its face may be applied with an evil eye and an unequal hand. The law is then valid but its application is not, and the remedy is against the application. This is a distinct ground and it is often forgotten: a challenge may be to the statute, or to the executive action under it, or to both.

Article 14 binds every kind of State action

Not only legislation. Article 14 controls executive action, administrative decisions, policy, the award of contracts, and the framing of rules. That proposition comes from the same line of cases as the instrumentality test:

Facts. E.P. Royappa v. State of Tamil Nadu, AIR 1974 SC 555, (1974) 4 SCC 3, decided 23 November 1973. A Chief Secretary complained that his transfers to other posts were mala fide and offended articles 14 and 16.

Held. The plea failed on the facts, but the Court held that articles 14 and 16 strike at arbitrariness in State action and ensure fairness and equality of treatment. State action must be based on relevant principles applicable alike to all similarly situated, and must not be guided by extraneous or irrelevant considerations, because that would be a denial of equality. Where the operative reason for State action is not legitimate and relevant but extraneous, it amounts to a mala fide exercise of power and is hit by articles 14 and 16.

Why it matters. It confirms that an administrative decision, not merely a statute, is measured against article 14, and it supplies the standard: relevant principles, applied alike, free of extraneous considerations.

The exceptions the Constitution itself makes

An honest answer names these, because they are the limits of the guarantee.

Article 361. The President and the Governors are not answerable to any court for the exercise of their powers, and no criminal proceeding may be instituted against them during their term.

Articles 105 and 194. Members of Parliament and of the State legislatures are not liable in any court for anything said or any vote given in the House.

Article 31C, so far as it survives, saves a law giving effect to the directive principles in article 39(b) and (c) from challenge under articles 14 and 19. See [Saving of Laws: Articles 31A, 31B and 31C].

Diplomatic and sovereign immunities recognised by international law and given effect by statute.

Article 359, under which the right to move a court for the enforcement of article 14 may be suspended during an emergency, though not articles 20 and 21 after the Forty-fourth Amendment.

A worked example

A municipal law provides that any building erected without permission may be demolished after notice, and gives the Commissioner a discretion to decide which of two procedures to follow: a summary procedure with a short notice, or an elaborate one with a hearing and an appeal. The Act states no principle for choosing between them. Two owners in identical positions are dealt with under different procedures.

Which limb of article 14 is engaged?

The classification challenge. Ask first whether the Act itself classifies. It does not: it does not say which buildings get which procedure. So this is not a case of an unreasonable classification in the statute.

The Anwar Ali Sarkar objection. A statute that confers an uncontrolled discretion to pick who is treated differently, with no principle stated to guide the choice, is bad for that reason alone, even though the differential treatment could have been justified had the Act itself drawn the line. That is the point of [Reasonable Classification].

The unequal-administration objection. Even if the Act laid down a principle, applying it to one owner and not to another identically placed is administration with an evil eye and an unequal hand, and the executive action falls though the Act stands.

The arbitrariness objection. On Royappa, State action must rest on relevant principles applied alike to all similarly situated. A choice of procedure made on no principle at all is arbitrary, and is unequal for that reason.

How to write it. Take the statute and the action separately. A student who runs them together usually loses the second point.

Quick revision

• Article 14 has two limbs: equality before the law, English and negative, no special privilege and equal subjection to the ordinary law; and equal protection of the laws, American and positive, like treated alike.
• The positive limb is what permits classification and therefore makes reservation possible.
• Treating unequals alike is as unequal as treating equals differently. Substantive, not formal, equality.
• Three things are forbidden: unreasonable classification, arbitrariness, and unequal administration of an equal law.
• Article 14 binds executive and administrative action, not only legislation: Royappa.
• Constitutional exceptions: articles 361, 105 and 194, 31C so far as it survives, diplomatic immunity, and suspension under article 359 except for articles 20 and 21.

Test yourself

1. Give the source and the sense of each of the two expressions in article 14, and explain why the Constitution needed both.

2. "If equality meant identical treatment, a progressive income tax would be unconstitutional." Explain what that shows.

3. Name the three distinct things article 14 forbids, and say why the third is often missed.

4. Does article 14 control executive action? Give the authority and the standard it lays down.

5. List five exceptions to equality that the Constitution itself makes.

So the enquiry has three steps, not two: what is the object; is the object legitimate; and is the differentia rationally connected to it.

The rules of application

Courts have settled a set of working propositions, and reproducing them is what a good answer does after stating the twin test.

There is a presumption of constitutionality, and the burden of showing a violation of article 14 lies on the person alleging it.

The presumption may rest on matters of common knowledge, and a court will assume that the legislature correctly appreciated the needs of its own people.

A legislature is free to recognise degrees of harm and may confine its restriction to those cases where the need is deemed clearest. It need not attack every aspect of a problem at once.

A law may be constitutional even though it applies to a single individual, if on account of some special circumstance that individual may be treated as a class by himself.

Mathematical nicety is not required. A classification may be rough at the edges, and some inequality in practice does not invalidate.

The classification may be made by the legislature or by an authority under a statute, provided the statute lays down the policy or principle for making it.

What the doctrine forbids

Facts. State of West Bengal v. Anwar Ali Sarkar, AIR 1952 SC 75, (1952) 1 SCR 284, decided 11 January 1952. The West Bengal Special Courts Act 1950 was intituled an Act to provide for the speedier trial of certain offences. Section 5(1) empowered the State Government to direct that a Special Court try "such offences or classes of offences or cases or classes of cases" as it directed. The special procedure dispensed with committal and with a jury, restricted adjournments and dispensed with de novo trial on transfer. The respondent and forty-nine others, charged over an armed raid on a factory, were convicted by a Special Court to which their case had been referred by notification, and moved the High Court saying section 5(1) denied them the equal protection of the laws.

Held. The special procedure was not on its face calculated to prejudice a fair trial, and the need for a speedier trial could form the basis of a reasonable classification, so section 5(1) was unobjectionable so far as it allowed the reference of offences or classes of offences or classes of cases. What was fatal was that it also allowed the State Government to refer "any case", including an individual case, and so vested an absolute and arbitrary power to select, with no principle laid down in the Act to guide the selection. To that extent the section was discriminatory and void.

Why it matters. It is the foundational Indian authority on classification, and its real teaching is about who draws the line. A legislature may classify. What it may not do is hand an unguided discretion to an official to decide, case by case, who will be treated differently. The Act failed not because differential treatment was unjustifiable but because the Act itself did not justify it.

Classification and the reservation question

MU's 2026 paper asked how the judiciary has interpreted reasonable classification in the context of affirmative action, so the connection has to be made explicitly.

Reservation is a classification: it marks off a group and gives it different treatment. On the equality code as set out in [The Equality Code], articles 15(4) and 16(4) are not exceptions to the guarantee but instances of the classification the guarantee already permits. That is why the Court has been able to say that reservation does not derogate from equality but effectuates it.

The twin test then supplies the discipline. The differentia must be intelligible: social and educational backwardness, or inadequate representation, must be identifiable and must be shown, not assumed. And it must have a nexus with the object: the object of reservation is the adequate representation of a class that is not adequately represented, so a reservation for a group that is in fact adequately represented has no nexus. That reasoning is why the creamy layer must be excluded and why data must be collected, dealt with in [The Creamy Layer] and [Reservation in Promotion].

An illustration of the test in operation

Facts. Ajay Hasia v. Khalid Mujib Sehravardi, AIR 1981 SC 487, (1981) 1 SCC 722, decided 13 November 1980. Admission to an engineering college was decided partly by a viva voce which lasted two or three minutes a candidate, consisted of formal questions about parentage and residence with no relevance to the subject, and carried one third of the total qualifying marks.

Held. The society running the college was "State" within article 12 and was bound by article 14. Allotting a third of the marks to an interview conducted in that manner was excessive and arbitrary.

Why it matters. It shows the second half of the article 14 enquiry doing what the classification test cannot. The rule applied equally to every candidate, so there was no classification to attack. What was wrong was that the method bore no rational relation to the object, selecting candidates on merit. That is the bridge to the next chapter.

A worked example

A State law provides that offences under a food adulteration statute committed in a municipal area shall be tried summarily, while the same offences committed elsewhere follow the ordinary procedure. A trader convicted summarily challenges the provision under article 14.

Step one, is there a classification? Yes: by geography, municipal area against the rest.

Step two, is the differentia intelligible? Yes. Municipal areas are defined by law and a person can tell which side of the line he is on. Geography is a recognised basis of classification, and territorial classification has repeatedly been upheld.

Step three, what is the object of the Act? Suppose the preamble says it is to prevent the adulteration of food and to secure prompt punishment where the risk to public health is greatest.

Step four, is there a nexus? This is where the case is won or lost. If the State can show that adulteration is concentrated in dense urban markets and that delay there causes wider harm, the nexus holds. If the object is simply to punish adulteration wherever it occurs, geography has no connection with it and the classification fails.

Step five, who drew the line? Here the statute itself drew it, so Anwar Ali Sarkar is not offended. Had the Act instead allowed a Collector to decide case by case which trader would be tried summarily, with no principle stated, it would have failed on that ground alone however good the geographical justification.

Step six, is the object itself legitimate? If the real purpose were to burden traders of a particular community concentrated in those areas, the classification would serve the object perfectly and still be void.

Quick revision

• Article 14 forbids class legislation, not reasonable classification.
• The twin test: an intelligible differentia, and a rational nexus between that differentia and the object of the Act. Both are required.
• The object must itself be legitimate, or a discriminatory purpose would validate its own means.
• Presumption of constitutionality; burden on the challenger; degrees of harm may be recognised; a single individual may be a class; mathematical nicety is not required.
• Anwar Ali Sarkar: speedier trial can be a valid basis, but an unguided power to pick an individual case is void. The vice is uncontrolled discretion, not differential treatment.
• Reservation is a classification, so the same test disciplines it: backwardness or inadequate representation must be shown, and must connect to the object.

Test yourself

1. State the twin test and explain which of the two conditions does the real work.

2. Why must the object of the Act itself be legitimate? What hole does that close?

3. Give five working propositions courts apply when testing a classification.

4. In Anwar Ali Sarkar, what was held valid and what was held void? What exactly was the vice?

5. Apply the twin test to reservation, and say what it requires the State to prove.

Articles 14 and 16 strike at arbitrariness in State action and require that it rest on relevant principles applicable alike to all similarly situated, free of extraneous or irrelevant considerations.

Why it matters. It is the origin of the doctrine. Note what it does structurally: it makes arbitrariness itself a violation, without requiring a comparator. That is the whole innovation.

Extended to administrative action

Facts. Ramana Dayaram Shetty v. International Airport Authority of India, AIR 1979 SC 1628, (1979) 3 SCC 489, decided 4 May 1979. The Authority published a tender notice confining eligibility to registered second class hoteliers with five years' experience, and then accepted a tender from a party who satisfied neither condition.

Held. Accepting a tender from a party who did not meet the Authority's own published standard was discriminatory, because it excluded others similarly situated, and was arbitrary and without reason. The Court laid down the wider principle:

Every action of the executive Government must be informed with reason and should be free from arbitrariness. That is the very essence of the rule of law and its bare minimal requirement.

It added that it makes no difference whether the exercise of power affects a right or denies a privilege: the licences, contracts and benefits the modern State dispenses are a new kind of wealth and are not gratuities the State may grant or revoke at pleasure.

Why it matters. It carries arbitrariness from service law into the whole field of government dealing: contracts, tenders, licences, allotments and concessions. It is also the source of the proposition most often quoted in administrative law answers.

Applied to strike down a rule

Facts. Air India v. Nergesh Meerza, AIR 1981 SC 1829, (1981) 4 SCC 335, decided 28 August 1981. Air hostesses retired at thirty-five, or on marriage within four years of joining, or on first pregnancy, whichever occurred earlier, with the Managing Director holding a discretion to extend service.

Held. The provisions were a clear case of official arbitrariness. The words "or on first pregnancy whichever occurs earlier" are unconstitutional, void and violative of article 14 and stand deleted. The Managing Director's uncontrolled option to extend was also struck down, because it allowed him to discriminate between one air hostess and another; the effect is that extensions must be granted as a matter of course where she is medically fit. The offending words being severable, the whole regulation was not struck down.

Why it matters. It shows both doctrines working together: the discretion clause fails on the Anwar Ali Sarkar ground of unguided power, and the pregnancy clause fails as arbitrary in itself. Read it accurately: the Court did NOT strike down the different retirement age, and its treatment of air hostesses and male cabin crew as separate classes has been criticised.

Manifest arbitrariness, and legislation

The harder question is whether a statute, as distinct from executive action, can be struck down simply for being arbitrary. For many years the position was contested: one line of authority held that arbitrariness is a doctrine for administrative action and that a law can be tested only by classification.

The modern position uses the phrase manifest arbitrariness, and it means something narrower than mere unreasonableness: legislation which is capricious, irrational or without an adequate determining principle, or which is excessive and disproportionate.

Facts. Shayara Bano v. Union of India, decided 22 August 2017 by five judges. The petitioner's husband pronounced talaq three times at once. She sought a declaration that talaq-e-biddat is void and that the practice is unconstitutional as violating articles 14, 15 and 21; the respondents said it was personal law protected by article 25 and recognised by section 2 of the Muslim Personal Law (Shariat) Application Act 1937.

Held. The order of the Court records that "by a majority of 3:2 the practice of talaq-e-biddat, triple talaq, is set aside". The three in the majority did not agree on why. Kurian Joseph J. held the practice was not part of Islamic law at all. Nariman and Lalit JJ. held it was manifestly arbitrary and therefore void under article 14. Khehar C.J. and Nazeer J. dissented, holding it was protected personal law and that Parliament should legislate.

Why it matters. It is the leading modern application of manifest arbitrariness, and it is also a warning about citation: the result is clear, the ratio is not, and a student should say which opinion he is relying on.

The criticism

An LL.M. answer should engage with the objection rather than recite the doctrine approvingly.

It has no textual anchor. Article 14 speaks of equality, not of reasonableness. Reading arbitrariness into it converts a guarantee of equal treatment into a general power to review the quality of legislation.

It has no standard. "Arbitrary" is not a test in the way the twin test is a test. It gives a court a conclusion rather than a method, and different judges reach different answers on the same facts.

It trespasses on the legislature. Striking down a statute because a court thinks it lacks an adequate determining principle comes close to reviewing policy, which is what the separation of powers reserves to Parliament. See [Judicial Activism, Restraint and Overreach].

The answer to the criticism is that the alternative is worse: a doctrine confined to classification cannot reach a law that harms everybody equally and for no reason, and the qualifier "manifest" is what keeps the standard from swallowing ordinary policy disagreement.

A worked example

A university frames a rule that every applicant for a teaching post must submit a certificate of physical fitness issued by one named private hospital, at the applicant's cost. The rule applies to all applicants without distinction. An applicant challenges it under article 14.

Can the classification test reach it? No. The rule applies to everybody identically, so there is no group treated differently and no differentia to test. A student who stops here concludes wrongly that article 14 is not engaged.

The arbitrariness route. On Royappa, State action must rest on relevant principles and be free of extraneous considerations. Requiring fitness may be relevant to a teaching post; requiring it from one named private hospital is not, unless some reason is shown. On R.D. Shetty, every action of the executive must be informed with reason.

What the university must show. Not that the rule is wise, but that there is a reason for it: that the named hospital alone has some capacity the others lack, or that a uniform standard could not otherwise be secured. If no reason is offered, the rule is arbitrary.

And if the rule were in a statute? Then the higher threshold applies. The question would be whether it is manifestly arbitrary: capricious, irrational, or without an adequate determining principle. A merely unwise legislative choice survives.

The lesson. Where nobody is treated differently, plead arbitrariness. Where a group is singled out, plead both.

Quick revision

• The classification test needs a classification. Arbitrariness needs none, which is why the second doctrine exists.
• Royappa: equality and arbitrariness are sworn enemies; an arbitrary act is by that fact unequal and violates article 14.
• R.D. Shetty: every action of the executive must be informed with reason and free from arbitrariness, the bare minimal requirement of the rule of law; and it applies to privileges as well as rights.
• Nergesh Meerza: retirement on first pregnancy is official arbitrariness and void; an unguided power to extend is void on the Anwar Ali Sarkar ground.
• For legislation the standard is MANIFEST arbitrariness: capricious, irrational, without an adequate determining principle, or excessive.
• Shayara Bano set aside triple talaq 3:2, but only two of the three used manifest arbitrariness. State which opinion you rely on.
• The criticism: no textual anchor, no standard, and a risk of reviewing policy.

Test yourself

1. What is the structural limit of the classification test, and how does the arbitrariness doctrine answer it?

2. Quote the Royappa passage and explain what it changed.

3. What did R.D. Shetty add, and why does its reference to privileges matter?

4. What exactly was struck down in Nergesh Meerza, and what was not?

5. Distinguish arbitrariness from manifest arbitrariness, and say why the qualifier is needed.

6. Set out three criticisms of the doctrine and the best answer to them.

The differences between the two articles

Set them side by side, because problems turn on the differences.

Field. Article 15 is general; article 16 is confined to employment or appointment to an office under the State.

Grounds. Five against seven, as above.

Who is bound. Article 15(1) binds the State. Article 15(2) binds private persons too, and is one of the few horizontal provisions in Part III: a private hotel or restaurant may not refuse a person on a prohibited ground. Article 16 binds the State as employer.

The enabling clauses.

• Article 15(3): special provision for women and children.
• Article 15(4): special provision for the advancement of any socially and educationally backward classes of citizens or for the Scheduled Castes and Scheduled Tribes.
• Article 15(5): special provision for their advancement so far as it relates to admission to educational institutions, including private ones, aided or unaided, other than minority institutions under article 30(1).
• Article 15(6): special provision for the economically weaker sections other than those covered by (4) and (5), including reservation of up to ten per cent in admissions.
• Article 16(3): residence requirements, by a law of Parliament.
• Article 16(4): reservation in appointments for any backward class of citizens not adequately represented in the services under the State.
• Article 16(4A): reservation in promotion, with consequential seniority, for the Scheduled Castes and Scheduled Tribes.
• Article 16(4B): the carry forward of unfilled reserved vacancies, which are not to be counted against the ceiling of the year in which they are filled.
• Article 16(5): a law providing that the incumbent of an office connected with a religious institution be of a particular religion or denomination.
• Article 16(6): reservation for the economically weaker sections, up to ten per cent.

Which were original and which came later is worth knowing, because the sequence is the history of the topic. Only 15(1) to (3) and 16(1) to (5) are original. Article 15(4) came with the First Amendment 1951; 16(4A) with the Seventy-seventh 1995 and its seniority limb with the Eighty-fifth 2001; 16(4B) with the Eighty-first 2000; 15(5) with the Ninety-third 2005; and 15(6) and 16(6) with the One Hundred and Third 2019.

Each of those amendments was a response to a judgment. That pattern, court decides, Parliament amends, is the shape of the whole reservation story and should be said out loud in an essay.

Enabling, not obliging

The single most useful proposition about the reservation clauses.

Article 16(4) says the State "shall not be prevented from" making provision. Article 15(4) says "nothing in this article... shall prevent the State from making any special provision". Neither confers a right on anybody. It follows that:

• No citizen has a fundamental right to a reservation, and a writ will not issue to compel the State to make one.
• A State that chooses not to reserve is not acting unconstitutionally.
• But once the State does reserve, the exercise is subject to article 14, so the reservation itself must be reasonable and non-arbitrary, and this is where the ceiling, the creamy layer and the requirement of data come in.

The relationship between the clauses

Facts. Indra Sawhney v. Union of India, AIR 1993 SC 477, decided 16 November 1992 by nine judges. Challenges to the office memoranda implementing the Mandal Commission's recommendation of twenty-seven per cent reservation for the socially and educationally backward classes, and a further ten per cent for other economically backward sections.

Held. Among the fourteen summarised answers: clause (4) of article 16 is not an exception to clause (1) but an instance and an illustration of the classification inherent in clause (1). Article 16(4) is exhaustive of the subject of reservation in favour of a backward class, but reservations may also be provided under clause (1), adjusted so as not to exceed the level of representation prescribed for backward classes. A provision under article 16(4) need not be made by the legislature; the executive may make it, and it is enforceable the moment it is issued.

Why it matters. It settles the architecture. Because 16(4) is an instance of 16(1) and not a derogation from it, reservation is not a departure from equality requiring narrow construction; and because the same equality principle governs both, the limits the Court has read in are limits the equality guarantee itself supplies.

A worked example

A State advertises posts in its Health Department and provides that only persons who have resided in the State for at least ten years are eligible, and that within the posts so filled, twenty per cent are reserved for women. A candidate resident elsewhere, and a male candidate, both challenge it.

The residence condition. Residence is a prohibited ground under article 16(2), so the requirement is on its face bad. The saving is article 16(3), but read it precisely: it permits Parliament, not a State legislature, to make a law prescribing a residence requirement for a class of employment under a State. A condition imposed by the State's own executive order or its own legislation is not saved. So the resident-elsewhere candidate succeeds unless a Parliamentary law authorises it.

The reservation for women. Sex is a prohibited ground under article 16(2), so this too needs an enabling provision. Article 16(4) speaks of a backward class, not of women. Article 15(3) permits special provision for women, but article 15 is the general article and article 16 governs public employment. The orthodox answer is that a reservation for women in public employment is sustained by reading articles 15(3) and 16(1) together, on the footing that 16(4) is not exhaustive of every permissible measure and that reservations may be provided under clause (1), which is what Indra Sawhney says.

What the male candidate must show. Not that he was treated differently, which is admitted, but that the measure is not a permissible special provision: that women are not disadvantaged in that service, or that the extent is excessive.

The order of reasoning. Prohibited ground first, enabling provision second, and the limits on the enabling provision third.

Quick revision

• Article 15(1): five grounds, religion, race, caste, sex, place of birth. Article 16(2): those five plus descent and residence.
• "On grounds only of" does not mean the ground must be the sole reason; that reading would empty the articles.
• Article 15(2) binds private persons; article 15 is general, article 16 is confined to public employment.
• The enabling clauses: 15(3) women and children, 15(4) backward classes, 15(5) admissions including private institutions, 15(6) EWS; 16(3) residence by Parliament, 16(4) reservation, 16(4A) promotion, 16(4B) carry forward, 16(5) religious offices, 16(6) EWS.
• Only 15(1) to (3) and 16(1) to (5) are original. Every other clause is an amendment answering a judgment.
• The clauses ENABLE, they do not oblige. Nobody has a fundamental right to a reservation.
• Indra Sawhney: article 16(4) is an instance of 16(1), not an exception to it, and may be made by the executive.

Test yourself

1. List the prohibited grounds in each article and explain why article 16(2) has two the other lacks.

2. What does "on grounds only of" mean, and what would the narrow reading do to the articles?

3. Which provisions of articles 15 and 16 were in the Constitution as enacted, and which came by amendment? What does the pattern show?

4. "Nobody has a fundamental right to a reservation." Explain, and say what does control a reservation once it is made.

5. Why does it matter that article 16(4) is an instance of article 16(1) rather than an exception to it?

"Or for the Scheduled Castes and the Scheduled Tribes". They are named separately and do not have to be shown to be socially and educationally backward: their inclusion in the Presidential Lists under articles 341 and 342 does that work.

It is an enabling clause. No citizen can compel a State to make a provision under it.

Article 15(5) and the private institution

The Ninety-third Amendment answered a different judgment. After the eleven-judge decision in T.M.A. Pai Foundation and the decisions that followed it, the position was that the State could not impose its reservation policy on a private unaided educational institution, because that would infringe the institution's own right to occupation under article 19(1)(g) and the autonomy those decisions recognised. That line is treated in [The Eleven-Judge Answer] and [After the Eleven-Judge Answer].

Article 15(5) reverses that result for reservation, in terms: special provision may be made "in so far as such special provisions relate to their admission to educational institutions including private educational institutions, whether aided or unaided by the State", other than the minority educational institutions referred to in clause (1) of article 30.

Three consequences follow.

Private unaided non-minority institutions may be required to reserve. That was the whole object.

Minority institutions are expressly excluded. So a minority institution, aided or unaided, is outside article 15(5). This is the point at which topic 3 meets topic 7, and it is developed in [Minority Institutions and the Right to Education Act].

The exclusion of minority institutions has itself been attacked as producing an inequality between two kinds of private institution. The answer given is that article 30(1) is itself a constitutional right, so excluding its holders from a burden is not an arbitrary favour but respect for another provision of the same Constitution.

Article 15(6) and the economically weaker sections

The One Hundred and Third Amendment inserted articles 15(6) and 16(6) in 2019, permitting special provision, including reservation of up to ten per cent, for economically weaker sections of citizens other than the classes mentioned in clauses (4) and (5). The Explanation leaves "economically weaker sections" to be notified by the State on the basis of family income and other indicators of economic disadvantage.

Two features are examinable and both are novel.

Economic criteria alone. Every earlier basis of reservation required social backwardness. Indra Sawhney had held in terms that a backward class cannot be identified only and exclusively by economic criteria, and had struck down a ten per cent reservation for economically backward sections on that ground.

Exclusion of the backward classes. The benefit is confined to those NOT covered by clauses (4) and (5), so a poor member of a Scheduled Caste is excluded from it.

Facts. Janhit Abhiyan v. Union of India, decided 7 November 2022 by five judges. Challenge to the 103rd Amendment on three grounds: that reservation on economic criteria alone breaches the basic structure; that excluding the backward classes violates the Equality Code; and that ten per cent in addition to existing reservations breaks the fifty per cent ceiling.

Held. By 3:2 the amendment was upheld. Reservation structured singularly on economic criteria does not violate any essential feature and does not damage the basic structure. Excluding the classes covered by articles 15(4), 15(5) and 16(4) balances non-discrimination against compensatory discrimination and does not violate the Equality Code. And ten per cent in addition does not damage the basic structure on account of the ceiling, because that ceiling is itself not inflexible and in any case applies only to the reservations envisaged by articles 15(4), 15(5) and 16(4). The extension to private unaided institutions was likewise upheld.

Why it matters. It changes what the fifty per cent rule means, which is why it belongs in [The Ceiling on Reservation] as well as here. And it establishes that economic disadvantage is now a constitutionally sufficient basis for a special provision, which every earlier decision had denied.

A worked example

A State notifies that in admissions to its medical colleges, twenty-seven per cent of seats are reserved for socially and educationally backward classes, fifteen per cent for the Scheduled Castes, seven and a half per cent for the Scheduled Tribes, and ten per cent for economically weaker sections, and applies the same percentages to private unaided colleges, including one run by a linguistic minority. A general-category candidate and the minority college both challenge it.

The general candidate on the total. Sixty per cent taken together exceeds fifty. On Indra Sawhney the ceiling applies to reservations under article 16(4); on Janhit Abhiyan the ceiling is not inflexible and in any event does not govern the EWS reservation, which rests on articles 15(6) and 16(6). So the correct analysis separates the two: the fifty per cent rule bites on the 49.5 per cent under clauses (4) and (5), which is within it, and the further ten per cent stands outside.

The general candidate on the EWS quota. After Janhit Abhiyan the objections that it rests on economic criteria alone and that it excludes the backward classes both fail.

The private unaided non-minority college. Article 15(5) expressly reaches private institutions whether aided or unaided, so it must comply.

The minority college. Article 15(5) expressly excludes the minority educational institutions referred to in article 30(1). It is outside the clause and cannot be compelled. Whether it is a minority institution is a separate question of fact, and that is [Who Is a Minority?].

Quick revision

• Article 15 as enacted had clauses (1) to (3) only. It could not authorise reservation in education.
• Champakam struck down the Communal Government Order; the First Amendment 1951 inserted article 15(4) in response. The Constitution's own footnote records it.
• Article 15(4): "special provision for the advancement" is wider than reservation; "socially AND educationally backward" is conjunctive; SC and ST are named separately.
• Article 15(5), Ninety-third Amendment 2005, reaches private institutions aided or unaided, and expressly excludes minority institutions under article 30(1).
• Article 15(6), 103rd Amendment 2019, permits up to ten per cent for economically weaker sections other than those in clauses (4) and (5).
• Janhit Abhiyan upheld it 3:2: economic criteria alone are permissible, the exclusion is permissible, and the fifty per cent ceiling is not inflexible and applies only to reservations under 15(4), 15(5) and 16(4).

Test yourself

1. Why did the First Amendment happen, and what does it tell you about article 15 as enacted?

2. Distinguish "socially and educationally backward classes" in article 15(4) from "backward class of citizens" in article 16(4).

3. What judgment did the Ninety-third Amendment answer, and which institutions did it deliberately leave out?

4. State the three grounds of challenge in Janhit Abhiyan and the answer given to each.

5. What did Janhit Abhiyan change about the meaning of the fifty per cent ceiling?

A caste can be, and quite often is, a social class in India. If it is backward socially it is a backward class for article 16(4). Among non-Hindus there are occupational groups, sects and denominations which for historical reasons are socially backward and are likewise backward classes.

No procedure of identification is prescribed by the Constitution or by law, and it is neither possible nor advisable for a court to lay one down. It must be left to the authority appointed to identify, which may adopt any method so long as its survey covers the entire populace. One may begin with castes, or with occupational groups, applying the criteria evolved for determining backwardness.

A class need not be similarly situated to the Scheduled Castes and Scheduled Tribes to be designated backward.

A backward class cannot be identified only and exclusively by economic criteria, though identification on the basis of occupation-cum-income without reference to caste is permissible. On that ground the ten per cent for economically backward sections was struck down. That last proposition has since been overtaken for a different purpose by article 15(6) and 16(6) and by Janhit Abhiyan, dealt with in [Economically Weaker Sections].

There is no constitutional bar to classifying backward classes into backward and more backward.

Why it matters. It is the controlling authority on who may be reserved for. Its treatment of caste as a permissible starting point, coupled with the insistence that caste is used as an indicator of social backwardness rather than as an end in itself, is the compromise Indian reservation law runs on.

Article 335 and efficiency

Article 335 is regularly misquoted as a limit on reservation. Read it.

It provides that the claims of members of the Scheduled Castes and Scheduled Tribes shall be taken into consideration, consistently with the maintenance of efficiency of administration, in the making of appointments. A proviso added by the Eighty-second Amendment permits relaxation in qualifying marks or standards of evaluation for reservation in matters of promotion.

Facts. State of Punjab v. Davinder Singh, decided 1 August 2024 by seven judges. Whether a State may sub-classify within the Scheduled Castes, giving preference to the most backward among them, when article 341 provides a single Presidential list.

Held. Sub-classification is permissible, and E.V. Chinnaiah, which had held otherwise, is overruled. Article 341 creates a deeming fiction whose consequence is that the listed castes receive the benefits the Constitution provides; it does not create an integrated homogeneous class. Sub-classification does not violate article 341(2) because no caste is included in or excluded from the List; it would offend only if preference or exclusive benefit were given to certain castes over all the reserved seats. The State may sub-classify under articles 15(4) and 16(4) if there is a rational principle for differentiation with a nexus with the purpose, and must collect data on inadequacy of representation and establish that it flows from backwardness. On article 335, the Court held that it is not a limitation on the exercise of power under articles 16(1) and 16(4) but a restatement of the necessity of considering the claims of the Scheduled Castes and Scheduled Tribes. Gavai J., concurring, held that the creamy layer principle should apply to the Scheduled Castes and Scheduled Tribes too.

Why it matters. Two things, and both are recent enough that most textbooks do not have them. Sub-classification within the Scheduled Castes is now lawful, so a State may direct benefits to the most disadvantaged within the list. And article 335 is authoritatively characterised: it is not a brake on reservation but an instruction to consider the claims.

The mechanics an answer should know

Vertical and horizontal reservation. A vertical reservation is one made under article 16(4) for a backward class, and the categories are exclusive of one another. A horizontal reservation, for example for women or for persons with disabilities, cuts across all the vertical categories: the seats are taken from within each vertical category rather than in addition to them. The distinction decides how a candidate who belongs to both a vertical and a horizontal category is counted.

The roster and the post-based system. Reservation is worked through a roster of points rather than by reserving individual named posts, and the ceiling is applied to the roster.

Reservation is applied to the year's vacancies, not to the total cadre strength. That is one of Indra Sawhney's answers and it matters arithmetically.

A reserved-category candidate selected on merit is counted against the general seats, not against the reserved quota. Otherwise the reservation would operate as a ceiling on the class rather than a floor.

A worked example

A State notifies that in a particular service, twenty per cent of posts are reserved for a caste it has newly identified as backward. The material before it is a report showing that the caste is poor. A candidate challenges the notification.

Is the identification good? Poverty alone is not enough. Indra Sawhney holds that a backward class cannot be identified only and exclusively by economic criteria; what article 16(4) looks to is social backwardness, of which economic condition is evidence but not a substitute. Identification on the basis of occupation-cum-income without reference to caste would have been permissible; a report showing income alone is not.

Is inadequate representation shown? The clause requires the State's opinion that the class is not adequately represented in the services. Backwardness and under-representation are two conditions, not one, and material is needed on the second. A court will ask whether there was material, not whether the State's assessment was right.

Could the State have sub-classified instead? If the caste is already within the Scheduled Castes list, Davinder Singh now permits sub-classification in its favour, provided there is a rational principle of differentiation with a nexus to the purpose, and provided the preference does not swallow all the reserved seats.

And article 335? The candidate cannot use it as a trump card. On Davinder Singh it is not a limitation on the power but a restatement of the need to consider the claims of the Scheduled Castes and Scheduled Tribes, consistently with efficiency.

Quick revision

• Article 16(4) has four conditions: a backward class of citizens, in the opinion of the State, not adequately represented, in the services under the State.
• The expression in article 16(4) is wider than that in article 15(4), with the accent on social backwardness.
• Indra Sawhney: a caste can be a class; no method of identification is prescribed; a class need not resemble the SC and ST; economic criteria alone will not do; backward and more backward may be distinguished.
• Davinder Singh (2024), seven judges: sub-classification within the Scheduled Castes IS permissible, Chinnaiah overruled; article 341 creates no homogeneous class; a rational principle with a nexus is required, and data on inadequacy of representation.
• Article 335 is not a limitation on articles 16(1) and 16(4) but a restatement of the need to consider the claims, consistent with efficiency.
• Vertical reservation is exclusive; horizontal reservation cuts across. A reserved candidate selected on merit occupies a general seat.

Test yourself

1. Set out the four conditions in article 16(4) and say which of them is about the class and which about the service.

2. What did Indra Sawhney decide about the use of caste in identifying a backward class?

3. Why is economic criteria alone insufficient under article 16(4), and how has that been altered for a different purpose since 2019?

4. What did State of Punjab v. Davinder Singh decide, what did it overrule, and what conditions did it attach?

5. How was article 335 characterised in Davinder Singh, and why does that matter?

6. Distinguish vertical from horizontal reservation, and say how a reserved-category candidate selected on merit is counted.

Held, in the answers relevant here:

"The reservations contemplated in Clause (4) of Article 16 should not exceed 50%." But the Court immediately qualified it: while fifty per cent shall be the rule, it is necessary not to put out of consideration certain extraordinary situations inherent in the great diversity of this country and the people. In far-flung and remote areas the population may, on account of being out of the mainstream of national life, need to be treated differently, and some relaxation of the strict rule may become imperative. In doing so extreme caution is to be exercised and a special case made out.

"The rule of 50% should be applied to each year." It cannot be related to the total strength of the class, category, service or cadre.

Devadasan was overruled to the extent inconsistent with the judgment.

Why it matters. Two things are usually got wrong. The rule was stated with an exception built in from the start, so a State that makes out a special case for an extraordinary situation is not defying Indra Sawhney but using it. And the rule operates on the year's vacancies, which is an arithmetical point with real consequences for carry-forward.

Article 16(4B) and the carry-forward problem

The year rule creates a practical difficulty. If reserved vacancies in a year go unfilled for want of candidates, and are carried forward to the next year, the next year's reservation may exceed fifty per cent even though each year's original quota did not.

Parliament answered by the Eighty-first Amendment, inserting article 16(4B): unfilled reserved vacancies of a year, carried forward to any succeeding year, are to be considered as a separate class and are not to be considered together with the vacancies of the year in which they are being filled for determining the ceiling of fifty per cent.

So the ceiling survives in form while the arithmetic that would have enforced it is set aside by the Constitution itself. An answer that notices this is making a real point about how the ceiling has been eroded by amendment rather than by decision.

What Janhit Abhiyan changed

Facts. Janhit Abhiyan v. Union of India, decided 7 November 2022 by five judges. Challenge to the 103rd Amendment, which inserted articles 15(6) and 16(6) permitting up to ten per cent reservation for economically weaker sections in addition to existing reservations. One ground was that this breaks the fifty per cent ceiling, which the petitioners said had become part of the basic structure, relying on the repeated affirmation of the rule in more than fifty decisions over sixty years.

Held. By 3:2 the amendment was upheld. On the ceiling:

Reservation for economically weaker sections of citizens up to ten per cent. in addition to the existing reservations does not result in violation of any essential feature of the Constitution of India and does not cause any damage to the basic structure of the Constitution of India on account of breach of the ceiling limit of fifty per cent. because, that ceiling limit itself is not inflexible and in any case, applies only to the reservations envisaged by Articles 15(4), 15(5) and 16(4) of the Constitution of India.

Why it matters. Two distinct propositions, and both should be stated. The ceiling is not inflexible, which confirms and widens the qualification Indra Sawhney had itself included. And it applies only to reservations under articles 15(4), 15(5) and 16(4), so a reservation resting on a different enabling clause is outside it altogether. That second proposition is the one that decides cases: it means the arithmetic must be done separately for the backward-class reservations and for the EWS reservation.

The corollary, which the petitioners pressed and lost, is that the fifty per cent rule is not part of the basic structure. That is now the position, and a book that says otherwise is out of date.

Where the ceiling stands today

Stated as a working rule for a problem.

Reservations under articles 15(4), 15(5) and 16(4) should not exceed fifty per cent, applied to the year's vacancies, subject to relaxation in an extraordinary situation for which a special case is made out with extreme caution.

Carried-forward vacancies under article 16(4B) are a separate class and are not counted for the ceiling in the year they are filled.

Reservation under articles 15(6) and 16(6) for economically weaker sections is outside the ceiling.

The ceiling is not part of the basic structure, so Parliament may provide for reservation beyond it by amendment.

A worked example

A State provides, for direct recruitment in a year with one hundred vacancies: twenty-seven per cent for other backward classes, sixteen per cent for Scheduled Castes, eight per cent for Scheduled Tribes, ten per cent for economically weaker sections, and a further six per cent carried forward from the previous year's unfilled Scheduled Tribe vacancies. Total reserved posts in the year: sixty-seven.

Do the arithmetic the way the law does it, not the way it looks.

The article 16(4) group is 27 plus 16 plus 8, which is fifty-one per cent. That exceeds the ceiling by one point, and the State must justify it as an extraordinary situation with a special case made out, or reduce it.

The carried-forward six per cent is excluded by article 16(4B): unfilled reserved vacancies carried forward are a separate class and are not counted together with the year's vacancies for the ceiling.

The EWS ten per cent is outside the ceiling altogether on Janhit Abhiyan, because the ceiling applies only to reservations under articles 15(4), 15(5) and 16(4).

So the figure to test is fifty-one, not sixty-seven. A candidate who argues that sixty-seven per cent is plainly unconstitutional has not done the analysis, and will be answered by the two exclusions.

The remaining argument. Even fifty-one requires justification, and a State relying on an extraordinary situation must show it: remoteness from the mainstream of national life, or comparable circumstances, established by material and not asserted.

Quick revision

• No provision of the Constitution fixes a percentage. The rule is judicial and structural.
• The derivation: article 16(4) is an instance of article 16(1), and an application of a guarantee cannot destroy it.
• Indra Sawhney: fifty per cent is the rule, subject to relaxation in extraordinary situations with extreme caution and a special case made out, applied to each year's vacancies and not to cadre strength.
• Article 16(4B): carried-forward unfilled reserved vacancies are a separate class and are outside the ceiling for the year they are filled.
• Janhit Abhiyan: the ceiling is NOT inflexible, and applies ONLY to reservations under articles 15(4), 15(5) and 16(4).
• The fifty per cent rule is not part of the basic structure.

Test yourself

1. Where does the fifty per cent rule come from, given that no article states it? Set out the derivation.

2. What qualification did Indra Sawhney itself attach to the rule, and to what does the rule apply?

3. What does article 16(4B) do, and why does it matter for the ceiling?

4. State the two propositions about the ceiling in Janhit Abhiyan and explain the practical effect of the second.

5. Is the fifty per cent rule part of the basic structure? What follows from the answer?

Constitutional posts. The children of the holder of a high constitutional office are excluded.

Service category. The children of officers of a certain rank in the all-India, State and public sector services, generally Group A or its equivalent, and of officers in the armed forces above a specified rank.

Professional class and those engaged in trade or industry. Persons of a professional or business standing at or above a stated level.

Property. Holders of agricultural land above a stated proportion of the ceiling area, and of plantations or urban property above a stated size.

Income and wealth. Family income above a notified limit, revised from time to time.

Two refinements matter. The test looks to the status of the parents, not of the candidate, because the object is to catch a family that has escaped the disadvantage. And the income limit is applied to income other than salary and agricultural income in the standard formulation, which is a frequent source of dispute.

The Scheduled Castes and Scheduled Tribes question

This is where the argument has run, and an answer must handle it in sequence.

The starting position. Indra Sawhney's discussion of the creamy layer was in the context of the other backward classes. The Scheduled Castes and Scheduled Tribes are identified not by an assessment of backwardness but by inclusion in the Presidential Lists under articles 341 and 342, which Parliament alone may alter.

The argument against applying it to them. If the List is conclusive, then excluding some members of a listed caste amounts to altering the List otherwise than by Parliament, contrary to articles 341(2) and 342(2). And Indra Sawhney itself treats the Scheduled Castes and Scheduled Tribes as the most backward among backward classes.

The argument for applying it. The purpose of reservation is the same whoever the beneficiary, and a person who has escaped disadvantage is not within the mischief. Excluding him from a benefit is not removing his caste from the List: he remains a member of a Scheduled Caste for every other purpose.

Where the law got to. The requirement was extended to promotion for the Scheduled Castes and Scheduled Tribes in M. Nagaraj, and the Union then asked for that to be reconsidered.

Facts. Jarnail Singh v. Lachhmi Narain Gupta, decided 26 September 2018 by five judges. The Attorney General asked that Nagaraj be referred to seven judges on two points: that its requirement of quantifiable data showing the backwardness of the Scheduled Castes and Scheduled Tribes contradicts Indra Sawhney, and that Nagaraj had misread Indra Sawhney in applying the creamy layer to them, since once they are in the Presidential List the List may be altered only by Parliament under articles 341 and 342.

Held. Nagaraj need not be referred to a Bench of seven. But its conclusion that the State must collect quantifiable data showing the backwardness of the Scheduled Castes and Scheduled Tribes, being contrary to the nine-judge Bench in Indra Sawhney, is held to be invalid to that extent. The Court declined to disturb Nagaraj on the creamy layer.

Why it matters. It separates two things the Attorney General had run together. Backwardness need not be proved for the Scheduled Castes and Scheduled Tribes, because the List settles it. The creamy layer nevertheless applies, because excluding an advanced individual is not the same as questioning the class.

Facts. State of Punjab v. Davinder Singh, decided 1 August 2024 by seven judges, on whether a State may sub-classify within the Scheduled Castes.

Held. Sub-classification is permissible and E.V. Chinnaiah is overruled; article 341 creates a deeming fiction and not an integrated homogeneous class; sub-classification does not violate article 341(2) because no caste is added to or removed from the List. Gavai J., concurring, held that the creamy layer principle should apply to the Scheduled Castes and Scheduled Tribes as well.

Why it matters. The reasoning of the majority on article 341 removes the main objection to applying the creamy layer to the Scheduled Castes: if sub-classifying within the List does not alter the List, then neither does excluding an advanced family. Note the status carefully. The application of the creamy layer to the Scheduled Castes and Scheduled Tribes rests on Nagaraj as left undisturbed by Jarnail Singh, and on a concurring opinion in Davinder Singh. It is not a holding of the seven-judge majority, and an answer should say so rather than overstate it.

The criticism

Worth engaging with at LL.M. level.

For the doctrine. It is the only mechanism that makes reservation self-limiting. Without it a small advanced section within each class captures the benefit permanently, and the constitutional purpose is never achieved.

Against it, as applied to the Scheduled Castes. The disadvantage the Scheduled Castes suffer is not only economic; it is the social disability of untouchability, which income does not remove. A wealthy member of a Scheduled Caste may still face exclusion of a kind a wealthy member of an other backward class does not. On that view an income-based exclusion misdescribes the injury the reservation addresses.

The middle position. Apply the doctrine but calibrate the criteria to the nature of the disadvantage, so that the tests for the Scheduled Castes are not simply the tests for the other backward classes with the same figures.

A worked example

A candidate belongs to a caste on the Presidential List of Scheduled Castes. His father is a serving officer of Group A rank in an all-India service, and the family's income is well above the notified limit. He applies for a post reserved for the Scheduled Castes, and separately for promotion under a reservation in his own service. The State's rules provide for a creamy layer exclusion in both.

Is the exclusion permissible in direct recruitment? On the reasoning above, yes: excluding him is not altering the Presidential List, since he remains a member of the Scheduled Caste for every other purpose, and Davinder Singh's holding that article 341 creates no homogeneous class removes the objection. Note the qualification: this rests on Nagaraj as undisturbed, and on a concurrence, not on a seven-judge holding.

Is the exclusion permissible in promotion? Nagaraj applied the creamy layer to promotion for the Scheduled Castes and Scheduled Tribes, and Jarnail Singh expressly declined to disturb it. So yes.

Must the State also prove his caste is backward? No. Jarnail Singh held that requirement invalid: the Presidential List settles backwardness, and requiring quantifiable data on it contradicts Indra Sawhney.

What must the State prove for the promotion reservation? Inadequacy of representation in the cadre, and it must not compromise efficiency under article 335, which on Davinder Singh is a restatement rather than a limitation. That is [Reservation in Promotion].

Which test applies to him? The parent's service rank, not his own means. The doctrine looks to the family that has escaped the disadvantage.

Quick revision

• The creamy layer is the advanced section of a backward class whose members no longer answer the description on which the class was identified.
• It follows from the clause: no backwardness, no differentia, no nexus.
• Indra Sawhney: it CAN be, and MUST be, excluded. Not a discretion.
• The tests look to constitutional posts, service rank, professional and business standing, property and income, and they look at the PARENTS.
• For the Scheduled Castes and Scheduled Tribes the argument was that the Presidential List is conclusive. Jarnail Singh held backwardness need not be proved but declined to disturb the creamy layer; Davinder Singh's reasoning on article 341 removes the main objection, and Gavai J. concurring would apply it.
• Its application to the SC and ST rests on Nagaraj undisturbed and a concurrence, not on a seven-judge holding.

Test yourself

1. Explain why the creamy layer exclusion follows from the words of article 16(4) rather than from judicial policy.

2. What exactly did Indra Sawhney say about it, and why does the choice of verb matter?

3. Name five heads on which the creamy layer is identified, and say whose position is looked at.

4. Set out the argument that the doctrine cannot apply to the Scheduled Castes, and the answer to it.

5. What did Jarnail Singh hold invalid, and what did it leave standing?

6. What is the precise status of the creamy layer for the Scheduled Castes after Davinder Singh?

Move three, the conditions. The amendments were challenged, and in M. Nagaraj v. Union of India the Supreme Court upheld them but read conditions into their exercise. As the argument in Jarnail Singh records, Nagaraj required the State, before making a promotional reservation, to collect quantifiable data showing the backwardness of the class, the inadequacy of its representation, and to consider the maintenance of efficiency of administration under article 335. It also applied the creamy layer exclusion to the Scheduled Castes and Scheduled Tribes in this context.

The report of Nagaraj could not be opened from the sources this book is built from. It is named without a citation, and what is stated about it above is taken from the account of it given in Jarnail Singh, which was read. Confirm it against the report before relying on it in writing.

Move four, the correction.

Facts. Jarnail Singh v. Lachhmi Narain Gupta, decided 26 September 2018 by five judges. The Attorney General asked that Nagaraj be reconsidered by seven judges on two points: that requiring quantifiable data on the backwardness of the Scheduled Castes and Scheduled Tribes contradicts the nine-judge Bench in Indra Sawhney, which treats them as the most backward once they are in the Presidential Lists under articles 341 and 342; and that Nagaraj had misread Indra Sawhney in applying the creamy layer to them.

Held. Nagaraj need not be referred to a Bench of seven. But its conclusion that the State has to collect quantifiable data showing the backwardness of the Scheduled Castes and Scheduled Tribes, being contrary to Indra Sawhney, is held to be invalid to this extent. The Court declined to disturb Nagaraj on the creamy layer, and declined the Attorney General's invitation to lay down proportionality to population as the test of adequacy, noting the difference in language between article 330, which mandates proportionality for reserved seats in the House of the People, and article 16(4A), which does not.

Why it matters. It leaves a precise, and examinable, list of what a State must now do.

What a State must show today

Not backwardness. Jarnail Singh removed that requirement for the Scheduled Castes and Scheduled Tribes. Their inclusion in the Presidential Lists settles it.

Inadequacy of representation. Still required, and it must rest on data about the cadre in question, not on a global figure for the service.

Efficiency of administration. Article 335 must be considered. On State of Punjab v. Davinder Singh, decided 1 August 2024 by seven judges, article 335 is not a limitation on the exercise of power under articles 16(1) and 16(4) but a restatement of the necessity of considering the claims of the Scheduled Castes and Scheduled Tribes. So it is a factor to be weighed, not a veto.

The creamy layer. Applies, on Nagaraj as left undisturbed by Jarnail Singh, and supported by Gavai J.'s concurrence in Davinder Singh. See [The Creamy Layer] for the precise status.

Cadre-wise, not service-wide. Because adequacy of representation is judged in relation to the posts to which promotion is being made.

The arguments

For promotional reservation. Reservation at entry alone produces a service in which the disadvantaged class is present at the bottom and absent at the top. Representation that stops at the lowest grade is not representation in any sense that matters, and the higher the post the greater its influence on policy.

Against. Promotion is by definition a comparison of those already in the service, all of whom have crossed the entry barrier. The justification for assistance at entry, that the competition was unequal, is weaker once the candidates have been in the same service for years. Consequential seniority compounds the effect, since a promoted candidate carries the advantage forward at every later stage.

The compromise the law has reached is the list above: permitted, confined to the Scheduled Castes and Scheduled Tribes, and conditional on data about representation and consideration of efficiency.

A worked example

A State proposes to reserve twenty-five per cent of promotion posts in the grade of Superintending Engineer for the Scheduled Castes and Scheduled Tribes, with consequential seniority. It relies on a departmental note stating that these communities are backward and that reservation is required as a matter of social justice. It has figures for the department as a whole but not for that grade.

Is the source of power good? Yes. Article 16(4A) permits reservation in promotion with consequential seniority for the Scheduled Castes and Scheduled Tribes.

Is the material good? No, on two counts. Relying on backwardness is unnecessary and beside the point after Jarnail Singh; and departmental figures are not figures for the grade. Adequacy of representation must be judged in the cadre to which promotion is made, because that is the field article 16(4A) operates on.

What else is missing? Consideration of article 335, which on Davinder Singh is a restatement of the need to weigh the claims consistently with efficiency, not a bar, but which must actually be weighed.

And the creamy layer? The scheme must provide for the exclusion, on Nagaraj as undisturbed by Jarnail Singh.

Does the ceiling apply? Yes, to the article 16(4) group, applied to the year's vacancies, with carried-forward vacancies excluded by article 16(4B).

Answer. The power exists; the exercise fails for want of cadre-specific data on representation and for want of consideration of efficiency, and it must provide for the creamy layer.

Quick revision

• Indra Sawhney: article 16(4) does NOT permit reservation in promotion, prospectively, with five years for existing schemes.
• Parliament's answer: article 16(4A) 1995 (promotion, SC and ST only), 16(4B) 2000 (carry forward), the words "with consequential seniority" 2001, and the proviso to article 335 in 2000.
• Nagaraj upheld the amendments but required quantifiable data on backwardness and inadequacy, consideration of efficiency, and the creamy layer.
• Jarnail Singh 2018: no reference to seven judges, but the requirement of data showing BACKWARDNESS of the SC and ST is invalid, being contrary to Indra Sawhney. Creamy layer left undisturbed. Proportionality to population rejected as the test.
• What a State must now show: inadequacy of representation, cadre-wise, on data; consideration of article 335; and provision for the creamy layer.
• Davinder Singh: article 335 is a restatement, not a limitation.

Test yourself

1. Set out the four moves in this topic in order, with the year of each.

2. What exactly did Indra Sawhney hold about promotion, and what did it do about existing schemes?

3. Which classes may benefit from a reservation in promotion, and which may not?

4. What did Nagaraj require, and which of its requirements did Jarnail Singh hold invalid, and why?

5. List what a State must establish today before reserving in promotion.

6. Give the argument for and against consequential seniority.

Two, the exclusion of the backward classes. The benefit is confined to those NOT covered by articles 15(4), 15(5) and 16(4). So a poor member of a Scheduled Caste is excluded from a benefit available to an equally poor person of the general category. The petitioners said an exclusion defined by caste, operating on a criterion of poverty, is discrimination on a prohibited ground.

Three, the ceiling. Ten per cent in addition breaks the fifty per cent rule, which the petitioners argued had become part of the basic structure through more than fifty decisions over sixty years.

The decision

Facts. Janhit Abhiyan v. Union of India, decided 7 November 2022 by five judges. Challenge to the 103rd Amendment on the three grounds above, and on its extension to admission in private unaided institutions.

Held. By 3:2 the amendment was upheld.

On economic criteria. Reservation is an instrument of affirmative action to ensure an all-inclusive march towards an egalitarian society, and it is an instrument not only for the inclusion of socially and educationally backward classes but for the inclusion of any class or section so disadvantaged as to answer the description of a weaker section. Reservation structured singularly on economic criteria does not violate any essential feature and does not damage the basic structure.

On the exclusion. Excluding the classes covered by articles 15(4), 15(5) and 16(4) from the benefit is in the nature of balancing the requirements of non-discrimination and compensatory discrimination, and does not violate the Equality Code or damage the basic structure.

On the ceiling. Ten per cent in addition does not damage the basic structure on account of a breach of the fifty per cent limit, because that ceiling limit itself is not inflexible and in any case applies only to the reservations envisaged by articles 15(4), 15(5) and 16(4).

On private unaided institutions. The challenge fails for the same reasons, read with the decision upholding article 15(5) in relation to such institutions.

Why it matters. Three separate propositions of general importance come out of it: economic disadvantage alone can found a special provision; the fifty per cent rule is not inflexible and is confined to the backward-class provisions; and the fifty per cent rule is therefore not part of the basic structure.

The dissent, which an LL.M. answer should give

Two of the five would have struck the amendment down, and the reasoning is worth reproducing because it is the strongest form of the objection.

The dissent's central point was not that poverty is an unworthy ground. It was that the exclusion is the vice: a measure for the economically weak which excludes the poorest precisely because they belong to castes that suffer the deepest disadvantage inverts the purpose of the equality code. On that view the amendment practises the discrimination it claims to remedy, and it is the exclusion, rather than the economic criterion or the ten per cent, that damages the basic structure.

An answer that gives only the majority is incomplete, and MU's questions on this topic ask for critical examination.

What follows in practice

The two reservations are counted separately. The fifty per cent ceiling is tested against the article 15(4), 15(5) and 16(4) group; the EWS ten per cent stands outside it. See the worked example in [The Ceiling on Reservation].

Identification is by notification. The Explanation leaves economically weaker sections to be notified by the State from time to time on the basis of family income and other indicators of economic disadvantage, so the criteria are executive and variable rather than constitutional.

The creamy layer has no application, because the class is defined by economic condition in the first place: the exclusion of the advanced is built into the criterion instead of being an exception to it.

Nobody is obliged to make it. Like every other clause in articles 15 and 16, clause (6) enables and does not oblige.

A worked example

A State notifies that economically weaker sections are those whose family income is below a stated figure and who do not hold agricultural land or urban property above stated limits. A candidate belonging to a Scheduled Caste, whose family income is below the figure, applies under the EWS quota and is refused. He challenges the exclusion under articles 14 and 15(1).

Is he within the class? No. Article 16(6) confines the benefit to economically weaker sections "other than the classes mentioned in clause (4)", and article 15(6) likewise excludes those covered by clauses (4) and (5). His caste puts him outside by definition.

Is the exclusion discrimination on a prohibited ground? That is exactly the second ground of challenge in Janhit Abhiyan, and the majority answered it: the exclusion balances non-discrimination against compensatory discrimination and does not violate the Equality Code. He is excluded not because his caste is disfavoured but because he is already within a scheme of compensatory provision.

What is his real complaint, and does the answer meet it? His complaint is that he receives nothing from the EWS quota while remaining subject to the fifty per cent competition within his own quota, so a poorer person of the general category may be better placed than he is. The dissent thought that unanswerable. The majority's answer is that the two schemes rest on different justifications and are not to be compared seat for seat.

How should he have framed it? Not as an attack on the amendment, which has been upheld, but on the notified criteria if they are arbitrary, or on the operation of the scheme if the ten per cent is in practice being taken from the seats otherwise available to the reserved categories rather than from the general pool. The second is a factual question and it is where such challenges now go.

Quick revision

• The 103rd Amendment 2019 inserted articles 15(6) and 16(6): up to ten per cent for economically weaker sections OTHER than those covered by 15(4), 15(5) and 16(4), in addition to existing reservations.
• It displaced three settled propositions: economic criteria alone were insufficient, reservation was for the socially disadvantaged, and fifty per cent was the ceiling.
• Indra Sawhney had struck down an identical ten per cent in 1992 for economically backward sections.
• Janhit Abhiyan upheld it 3:2. Economic criteria alone are permissible; the exclusion balances non-discrimination against compensatory discrimination; the ceiling is not inflexible and applies only to the backward-class provisions.
• The dissent's point was the exclusion, not the criterion: a measure for the poor which excludes the poorest by caste inverts the equality code.
• The creamy layer has no application; identification is by executive notification; the clause enables and does not oblige.

Test yourself

1. Name the three settled propositions the 103rd Amendment displaced, and give the authority for each.

2. State the three grounds of challenge in Janhit Abhiyan and the majority's answer to each.

3. What did the case decide about the fifty per cent ceiling, and what follows for the basic structure argument?

4. Set out the dissenting objection in its strongest form.

5. Why does the creamy layer doctrine have no application to the EWS reservation?

The Constitution (One Hundred and Fifth Amendment) Act, 2021 reversed that consequence. It amended article 342A and article 366(26C) to make clear that a State or Union territory may prepare and maintain its own list of socially and educationally backward classes for its own purposes, which may differ from the Central List.

The pattern to notice, and it is the pattern of the whole reservation story: an amendment intended to strengthen the machinery produced a federal consequence nobody wanted, and a second amendment was needed to undo it within three years.

Whether the list is a single homogeneous class

Facts. State of Punjab v. Davinder Singh, decided 1 August 2024 by seven judges. Whether a State may sub-classify within the Scheduled Castes so as to give preference to the most backward among them, when article 341 provides a single Presidential list. An earlier five-judge decision, E.V. Chinnaiah, had held that the Scheduled Castes form a single homogeneous class which may not be sub-divided, on the footing that sub-division would amount to varying the list otherwise than by Parliament.

Held. Sub-classification is constitutionally permissible, and the holding in Chinnaiah is overruled.

On article 341 as a deeming fiction. The phrase "deemed" means the notified castes shall be "regarded as" the Scheduled Castes. Even if the fiction creates a constitutional identity, the only logical consequence is that the listed castes receive the benefits the Constitution provides. The operation of the provision does not create an integrated homogeneous class.

On article 341(2). Sub-classification does not violate it, because the castes are not per se included in or excluded from the List. It would violate the provision only where preference or exclusive benefit is given to certain castes or groups over all the seats reserved for the class.

On the evidence. Historical and empirical evidence demonstrates that the Scheduled Castes are a socially heterogeneous class.

On the conditions. The State may sub-classify under articles 15(4) and 16(4) if there is a rational principle for differentiation and that principle has a nexus with the purpose of sub-classification. The State must establish that the inadequacy of representation of a caste or group is because of its backwardness, and must collect data on inadequacy of representation in the services of the State, that being used as an indicator of backwardness.

On article 335. It is not a limitation on the exercise of power under articles 16(1) and 16(4) but a restatement of the necessity of considering the claims of the Scheduled Castes and Scheduled Tribes.

Why it matters. It is the most important change in reservation law in a decade and it postdates most textbooks. A State may now direct the benefit within the list towards those who have received least of it, subject to proof.

The Commissions

Three constitutional bodies, and the differences are examinable.

Article 338, the National Commission for Scheduled Castes. Constitutional since the Sixty-fifth Amendment. A Chairperson, Vice-Chairperson and three other members. Its duties are to investigate and monitor all matters relating to the safeguards provided for the Scheduled Castes, to inquire into specific complaints, to advise on planning, to report to the President annually, and to make recommendations. In investigating and inquiring it has the powers of a civil court trying a suit. The Union and every State Government must consult the Commission on all major policy matters affecting the Scheduled Castes.

Article 338A, the National Commission for Scheduled Tribes. Created by the Eighty-ninth Amendment 2003, which split the earlier joint Commission in two. Identical structure and functions for the Scheduled Tribes.

Article 338B, the National Commission for Backward Classes. Given constitutional status by the 102nd Amendment 2018. Identical structure and functions for the socially and educationally backward classes.

Article 339 empowers the President to appoint a Commission on the administration of Scheduled Areas and the welfare of the Scheduled Tribes, and gives the Union executive power to give directions to a State as to the drawing up and execution of schemes for the welfare of the Scheduled Tribes.

Article 340 is the Backward Classes Commission provision, and it is the source of the two national Commissions whose reports shaped the topic: the first, which reported in 1955, and the second, whose report of 1980 was implemented in 1990 and produced Indra Sawhney.

How identification is actually done

Facts. Indra Sawhney v. Union of India, AIR 1993 SC 477, decided 16 November 1992 by nine judges.

Held, on identification: no procedure or method is prescribed by the Constitution or by law, and it is neither possible nor advisable for a court to lay one down; it must be left to the authority appointed to identify, which may adopt any method so long as its survey covers the entire populace. Identification may be done with reference to castes among and along with other occupational groups, classes and sections. One may begin with castes, apply the criteria evolved for determining backwardness, and see whether a caste satisfies them; the same process may be applied to occupational groups. A caste can be and often is a social class; among non-Hindus, occupational groups, sects and denominations may equally be backward social collectives. A class need not be similarly situated to the Scheduled Castes and Scheduled Tribes. And the Court directed that a permanent mechanism in the nature of a Commission be created to examine requests for inclusion and complaints of over-inclusion or non-inclusion, whose advice should ordinarily bind the Government, which must record reasons if it does not accept it.

Why it matters. The direction in the last sentence is what produced the statutory National Commission for Backward Classes, later given constitutional status as article 338B. So the Commission machinery is itself a judicial creation given constitutional form.

A worked example

A State enacts a law providing that within the fifteen per cent reserved for the Scheduled Castes in State services, half shall go to four named castes identified by a State commission as having received almost none of the benefit over fifty years. A member of a Scheduled Caste outside the four challenges it.

Is sub-classification permissible at all? Yes, since Davinder Singh. Chinnaiah is overruled and article 341 creates no integrated homogeneous class.

Does it vary the Presidential List? No. No caste is included in or excluded from the List; every one of them remains a Scheduled Caste for all purposes. That is the express reasoning on article 341(2).

Where is the line? Davinder Singh drew it: sub-classification offends article 341(2) only where preference or exclusive benefit is given to certain castes over all the seats reserved for the class. Half of fifteen per cent leaves the other half open to the whole class, so the line is not crossed. A law giving the four castes the entire fifteen per cent would be a different case.

What must the State prove? A rational principle for the differentiation, a nexus between that principle and the purpose, data on inadequacy of representation in the services, and that the inadequacy is because of backwardness. A commission report recording fifty years of near-total exclusion is exactly the kind of material contemplated.

Any other requirement? Article 335 must be considered, but it is a restatement rather than a limitation.

Quick revision

• Articles 341 and 342: Presidential notification, variable only by a law of Parliament, and no later notification may vary it.
• Article 342A, inserted 2018, created a central list for the socially and educationally backward classes; the 105th Amendment 2021 restored the power of a State to maintain its own list.
• Davinder Singh (2024), seven judges: sub-classification within the Scheduled Castes is permissible; Chinnaiah overruled; article 341 is a deeming fiction creating no homogeneous class; the vice is only preference over ALL the reserved seats; a rational principle with a nexus and data are required; article 335 is a restatement, not a limitation.
• Commissions: article 338 for the Scheduled Castes, 338A for the Scheduled Tribes, 338B for the Backward Classes. Civil-court powers, annual reports, and mandatory consultation on major policy.
• Article 339 on Scheduled Areas; article 340 on a Backward Classes Commission, the source of the two national Commissions.
• Indra Sawhney: no method of identification is prescribed; caste may be the starting point; and a permanent Commission was directed, which is why article 338B exists.

Test yourself

1. Distinguish the machinery for the Scheduled Castes from the machinery that existed for the backward classes before 2018.

2. What did the 102nd Amendment do, what unintended consequence followed, and how was it undone?

3. State the reasoning in Davinder Singh on article 341 as a deeming fiction, and identify where the Court drew the line.

4. What conditions must a State satisfy before sub-classifying?

5. Name the three constitutional Commissions with their articles, and state two powers they hold.

6. What did Indra Sawhney direct about a permanent mechanism, and what came of it?

The Indian constitutional vocabulary is "special provision", which is wider than reservation and covers all of these. An answer that notices that article 15(4) permits "special provision for the advancement" rather than "reservation" is making a real point: reservation is one instrument among several, and it is the most contested one because it is the only one that operates by exclusion.

The forms it takes

Reservation of seats or posts, under articles 15(4), 15(5), 15(6), 16(4), 16(4A) and 16(6).

Relaxation of standards: lower qualifying marks, higher age limits, extra attempts. The proviso to article 335 expressly permits relaxation in qualifying marks or standards of evaluation for reservation in promotion.

Financial assistance: scholarships, fee concessions, hostels, maintenance grants, traceable to article 46.

Preferential treatment short of reservation: preference between equally qualified candidates, which Indra Sawhney said would not be impermissible in promotion even where reservation was not.

Reservation of seats in legislatures, under articles 330 to 334, which is a different thing again because it is not about opportunity but about representation.

Protective legislation, such as laws against untouchability under article 17 and laws for the protection of tribal land in Scheduled Areas under the Fifth Schedule.

The limits, and why they exist

Every limit the courts have imposed follows from the justification. That is the argument to make, because it shows the limits are not hostility to the doctrine but its logic.

The class must actually be disadvantaged. Otherwise there is no differentia, and on the twin test the classification has no nexus with its object.

The advanced within it must be excluded. [The Creamy Layer].

There must be inadequate representation. The object of article 16(4) is representation, so a class already represented adequately falls outside the clause.

The measure must not destroy the guarantee it applies. [The Ceiling on Reservation]. An application of article 16(1) cannot abolish article 16(1).

Efficiency must be considered. Article 335, which on Davinder Singh is a restatement of the need to consider the claims rather than a limitation.

It must not be arbitrary. On E.P. Royappa v. State of Tamil Nadu, AIR 1974 SC 555, (1974) 4 SCC 3, decided 23 November 1973, equality and arbitrariness are sworn enemies, and an arbitrary act is unequal and violates article 14. A special provision made without material, or extended to a group for reasons unconnected with disadvantage, is arbitrary however benevolent its label.

The criticisms, and the answers

An LL.M. answer must engage rather than defend.

It entrenches the categories it means to dissolve. By distributing benefits on the basis of caste, the State keeps caste politically salient and gives every group an interest in being classified as backward. The answer: the categories were not created by the Constitution and will not disappear by being ignored, and the creamy layer and the requirement of data are designed to make the classification temporary in application even if the category persists.

It sacrifices merit. The answer: merit measured by a competitive examination reflects the opportunity a candidate has had as much as the ability he possesses, so a standard that ignores unequal preparation is not neutral. The Court has also said that reservation cannot be called anti-meritarian, while accepting that there are services and posts to which it may not be advisable to apply the rule.

It benefits the already advantaged within the class. The answer: that is precisely what the creamy layer exclusion is for, and Davinder Singh now permits sub-classification so that benefits can be directed to those within the list who have received least.

It has no end. This is the strongest objection, and the honest answer is that the Constitution provided an end only for reservation of seats in legislatures, in article 334, which has been extended repeatedly, and provided none for reservation in employment or education. Whether a remedy with no terminal date remains a remedy is a fair question and an answer should say so rather than resolve it.

A worked example

A State provides that in recruitment to the police, candidates from a listed community shall be given five additional marks, and that the physical standards required of them shall be lower. A general-category candidate challenges both.

Is a preference short of reservation permissible? Yes. "Special provision for the advancement" in article 15(4) and the power in article 16(4) are not confined to reservation, and Indra Sawhney said concessions and relaxations short of reservation are permissible even where reservation is not.

Do the same limits apply? Yes, because the justification is the same. The community must be disadvantaged on material; the measure must be connected to the disadvantage; and it must not be arbitrary.

Are the two measures on the same footing? Not necessarily, and this is the point the question is testing. The five additional marks address unequal preparation for a written test, which is a disadvantage the community plausibly suffers, so the nexus is straightforward. The lower physical standards address a physical requirement of the post itself. Unless the State can show that the physical disadvantage is itself a product of the deprivation the provision is meant to remedy, for example through nutrition, the nexus is missing, and the measure looks like a benefit unconnected to the reason for conferring it.

How to frame the answer. The power exists; the first measure survives the twin test; the second requires justification that may not be available; and neither is saved merely because the object was benevolent, because on Royappa an act without a relevant reason is arbitrary and therefore unequal.

Quick revision

• Protective discrimination is treating a disadvantaged group differently so that equality becomes substantive rather than formal.
• Textual basis: the enabling clauses of articles 15 and 16, with article 46 as the directive principle behind them.
• The justification in four steps: equal protection requires like treated alike; unlikes need not be; some groups began behind for reasons the Constitution condemns; correcting for that serves the guarantee.
• Indra Sawhney: article 16(4) is an instance of 16(1), not an exception.
• "Special provision" is wider than reservation and includes relaxation, finance and preference.
• The limits all follow from the justification: real disadvantage, creamy layer, inadequate representation, no destruction of the guarantee, efficiency, and non-arbitrariness.
• The strongest criticism is that the remedy has no terminal date except in article 334.

Test yourself

1. Set out the justification for protective discrimination as a four-step argument.

2. Distinguish protective, compensatory and affirmative action, and say what the Indian constitutional term is.

3. Name five forms the doctrine takes, with the provision authorising each.

4. Show how each of the judicial limits on reservation follows from the justification rather than contradicting it.

5. State the four standard criticisms and the best answer to each.

Existing employees' protections. Service conditions that had statutory or regulation force under Sukhdev Singh become contractual on transfer.

What is NOT lost. Statutory protections that bind private employers: labour legislation, the law on equal remuneration, provisions protecting persons with disabilities where they apply to private establishments, and the general law of contract and tort. The point is that these are ordinary law, alterable by ordinary majority, and not constitutional guarantees.

Why the doctrine moved the wrong way at the wrong time

This is the analytical heart of the topic and it should be stated as a paradox.

Facts. Pradeep Kumar Biswas v. Indian Institute of Chemical Biology, decided 16 April 2002 by seven judges. Whether the Council of Scientific and Industrial Research, a registered society, is "State".

Held. It is. But the test was restated: the Ajay Hasia indicia are not a rigid set of principles such that satisfying one makes a body State; the question is whether on the cumulative facts the body is financially, functionally and administratively dominated by or under the control of the Government; such control must be particular to the body and pervasive; and merely regulatory control, whether under statute or otherwise, does not suffice.

Facts. Zee Telefilms Ltd v. Union of India, decided 2 February 2005 by five judges. Whether the Board of Control for Cricket in India is "State".

Held. It is not. It is not created by statute, no share capital is held by government, it receives practically no government funding, its monopoly is not State conferred or State protected, and the control over it is only regulatory, of the kind applicable to similar bodies. But the Board does discharge duties akin to public duties, and although article 32 is unavailable, an aggrieved party may seek a remedy under article 226, which is much wider than article 32.

The paradox. Between 1991 and 2005 the State was withdrawing from direct provision while the definition of "State" was contracting. The two movements are independent, and neither was designed with the other in mind, but together they open a gap: the function leaves government hands just as the test stops following it.

What survives, and how far it helps

Article 226 and the public function route. The most important survival, and Zee Telefilms is its authority. A privatised utility, a private university conferring degrees under statute, a licensee holding a monopoly over an essential service, may be answerable in a writ petition for breach of a public duty.

Its limits, which must be stated. It supplies a forum, not the substantive application of Part III. A court may quash an arbitrary decision as a breach of public duty, and may draw on article 14 standards in deciding what arbitrariness means, without holding the body bound by article 14. Article 226 is discretionary, and is generally declined where the obligation is purely contractual, which is what a recruitment dispute with a private employer usually is. So the route that answers the Zee Telefilms problem does not obviously answer the reservation problem.

Conditions attached to the transfer. A government may require the transferee to assume specified obligations as a term of the disinvestment or the concession. This is the mechanism actually used, and its weakness is that it is contractual and negotiable rather than constitutional.

Legislation. Parliament may impose obligations on private employers under article 35 and its general legislative powers, and there is a standing proposal for reservation in the private sector. Nothing of the kind has been enacted, and an answer should say so rather than imply otherwise.

The arguments, both ways

For extending the obligation. The justification for reservation is that a class was excluded from opportunity by a social order, not that the government happened to be the employer. The organised private sector now holds most of the desirable employment, so a guarantee confined to public employment reaches a shrinking fraction of the opportunity that exists. And on Indra Sawhney's own reasoning, article 16(4) is an instance of the equality principle, which is not by its nature confined to State employers.

Against. Part III was written against the State because the State is uniquely dangerous and uniquely inescapable, and diluting that focus weakens the guarantees rather than extending them. Imposing selection obligations on private employers is a policy decision with economic consequences that a legislature is equipped to weigh and a court is not. And there is no stopping point: if the private sector must reserve because the function is public, the same reasoning reaches every employer of any size.

The middle positions. Obligations travelling with the enterprise on transfer; a statutory scheme confined to sectors that depend on State licences, concessions or land; incentives rather than compulsion; and function-specific application of article 12 where a private body exercises monopoly public power.

A worked example

A State disinvests its entire shareholding in a State Road Transport Corporation, which becomes a private company holding an exclusive licence to operate on notified routes. Before the transfer, recruitment was governed by regulations reserving posts under article 16(4). After it, the company advertises posts with no reservation. A candidate from a Scheduled Caste petitions under article 32, and in the alternative under article 226.

Article 32. Apply Pradeep Kumar Biswas on the cumulative facts. No shareholding, no funding, and the licensing regime is regulatory rather than a control particular to this body. The company is not "State", so article 32 does not lie and articles 14 and 16 do not bind it. This is Zee Telefilms in substance.

Does the reservation survive the transfer? Not of its own force. The regulations had statutory force while the Corporation was a statutory body, on Sukhdev Singh; on transfer the employment becomes contractual. The obligation lapses with the precondition.

Article 226. Better, but not decisive. Operating an exclusive licence over notified public routes is a strong candidate for a public function, so the company may be amenable in principle. The difficulty is that the duty complained of is a recruitment obligation, and the writ jurisdiction is generally declined where the matter is contractual employment with a private employer. The petitioner's best framing is not "you must reserve" but "you exercise a monopoly conferred by the State and your selection process must therefore be fair and non-arbitrary".

What would have secured the position? A condition in the disinvestment agreement, or a licence condition, requiring the transferee to maintain the reservation. That is the practical answer, and it is a contractual answer to a constitutional problem.

Answer. The candidate loses on article 32, has an arguable case on article 226 framed as fairness rather than reservation, and has no constitutional right to the quota.

Quick revision

• Reservation is an exercise of an enabling power over employment UNDER THE STATE. Remove the State and both the power and the field disappear.
• What is lost: the reservation, equality of opportunity in recruitment, the duty of non-arbitrariness, the article 32 remedy, and the statutory force of service regulations.
• The paradox: the State withdrew from provision after 1991 while article 12 narrowed after 2002. Pradeep Kumar Biswas and Zee Telefilms are the two decisions to cite.
• Article 226 survives and reaches a public function or public duty, but it gives a forum rather than the substantive application of Part III, and is usually declined for contractual employment.
• Practical answers: conditions attached to the transfer, sectoral legislation, incentives, and function-specific application of article 12.
• MU's stem is "privatisation affects programme of affirmative action, discuss". Answer it as a legal question about the precondition of the power, not as a policy lament.

Test yourself

1. Why does the reservation obligation lapse on privatisation without any government deciding to withdraw it?

2. List five things a candidate loses when a public employer is privatised, and one thing that is not lost.

3. State the two movements that together create the gap, and give the leading case for each.

4. How far does article 226 answer the problem, and what are its two limits here?

5. Give three arguments for extending the obligation to private employers and three against.

6. Name four middle positions and say which of them is actually used in practice.

That is the structural objection. Article 31B does not describe a class of law at all. It protects whatever Parliament chooses to list, which means Parliament can, by a constitutional amendment listing a statute, place any statute beyond the reach of any fundamental right. A device created to solve a specific problem became a general escape from Part III.

Article 31C, and the two attempts to widen it

As enacted in 1971, article 31C had two limbs. The first protected a law giving effect to article 39(b) and (c) from articles 14 and 19. The second provided that no law containing a declaration that it is for giving effect to such policy shall be called in question in any court on the ground that it does not give effect to such policy.

The second limb was struck down in Kesavananda Bharati, on the ground that it made the legislature the judge of whether its own law fell within the protection, and so excluded judicial review. The first limb survived. Kesavananda is in the case bank's queue and is named here without a citation.

The Forty-second Amendment 1976 widened the first limb, by section 4, so that it protected a law giving effect to all or any of the directive principles in Part IV, not merely article 39(b) and (c).

Facts. Minerva Mills Ltd v. Union of India, AIR 1980 SC 1789, (1981) 1 SCR 206, decided 31 July 1980 by five judges. Challenge to sections 4 and 55 of the Forty-second Amendment. Section 55 had inserted clauses (4) and (5) in article 368: clause (4) that no amendment shall be called in question in any court, clause (5) that there shall be no limitation whatever on the constituent power of Parliament.

Held. Clause (5) is unconstitutional: it demolishes the pillars on which the Preamble rests by conferring constituent power without any limitation whatever, and the power to destroy is not a power to amend. A limited amending power is itself a basic feature, and the donee of a limited power cannot by exercising it convert that power into an unlimited one. Clause (4) is equally void, being interlinked with clause (5): the Constitution rests on a balance among the three wings, it is the duty of judges to pronounce on the validity of laws, and rights without remedies are as writ in water. On section 4, the majority struck down the widening of article 31C, holding that giving the directive principles blanket primacy over articles 14 and 19 destroys the balance between Parts III and IV, which is itself a basic feature. Bhagwati J. dissented on this point, holding that fundamental rights and directive principles are both part of the broad spectrum of human rights and that the primacy did not damage the basic structure.

Why it matters. It restored article 31C to its 1971 width, and it made the balance between Parts III and IV a basic feature. It is also the case that made the limited nature of the amending power itself unamendable.

The end of the sanctuary

Facts. I.R. Coelho v. State of Tamil Nadu, decided 11 January 2007 by nine judges. Whether a law placed in the Ninth Schedule, and so protected by article 31B, may nevertheless be attacked as damaging the basic structure.

Held, in six numbered propositions.

A law that abrogates or abridges Part III rights may or may not violate the basic structure; if it does, whether by amending Part III or by insertion in the Ninth Schedule, it must be invalidated.

The validity of each new amendment is judged on its own merits, by the actual effect and impact of the law on Part III rights. That is the impact test.

All amendments adding laws to the Ninth Schedule on or after 24 April 1973, the date of Kesavananda Bharati, must be tested against the basic or essential features as reflected in article 21 read with articles 14 and 19 and the principles underlying them. An Act in the Ninth Schedule is therefore open to attack if the right taken away pertains to the basic structure.

The protection is not blanket: the nature and extent of the infraction is examined by the rights test and the essence of the right test.

Where the validity of a Ninth Schedule law has already been upheld it may not be challenged again on these principles; but a law held violative of Part III and afterwards incorporated in the Schedule may be challenged.

Actions taken and transactions finalised are not open to challenge.

Why it matters. It ends the Ninth Schedule as a sanctuary. The date, 24 April 1973, is the single most quotable fact in this chapter: it is the day Kesavananda was decided, and it marks the point from which Parliament knew the amending power was limited.

The 24 April 1973 cut-off was drawn earlier, in Waman Rao v. Union of India, decided 9 May 1980. That case is in the case bank's queue and is named without a citation.

A worked example

In 2020 Parliament amends the Constitution to insert a State statute into the Ninth Schedule. The statute abolishes a class of tenancy without any hearing, and bars any suit or proceeding in relation to it. A tenant challenges the amendment and the statute.

Does article 31B protect the statute from Part III? On its terms, yes: article 31B says none of the Acts specified in the Ninth Schedule shall be deemed void on the ground of inconsistency with any of the rights in Part III.

Is that the end of it? No. On Coelho, an amendment adding a law to the Schedule on or after 24 April 1973 is tested against the basic structure as reflected in article 21 read with articles 14 and 19. This amendment is from 2020, so it is squarely within the rule.

What test is applied? Not whether the statute is inconsistent with a fundamental right, which article 31B forgives, but the impact test: what is the actual effect of the law on the rights in Part III, and does the right taken away pertain to the basic structure. The rights test and the essence of the right test are applied, taking a synoptic view of Part III.

Apply it. Abolishing a right of property is not by itself basic structure, since property ceased to be a fundamental right in 1978. But excluding any hearing and barring any proceeding engages article 21 read with article 14, and the exclusion of judicial remedy engages the reasoning in Minerva Mills that rights without remedies are as writ in water. The challenge is therefore arguable on the second limb and weak on the first.

Would article 31A help the State instead? Only if the statute falls within one of its five classes, and only against articles 14 and 19, not against article 21. And a State law needs the President's assent under the proviso.

Would article 31C? Only if the law gives effect to article 39(b) or (c), and only against articles 14 and 19. After Minerva Mills it cannot be invoked for the directive principles generally.

Quick revision

• Article 31A: immunity by SUBJECT MATTER, five classes, from articles 14 and 19 only, with a proviso requiring the President's assent for a State law.
• Article 31B: immunity for NAMED STATUTES in the Ninth Schedule, from the whole of Part III, notwithstanding any judgment.
• Article 31C: immunity by PURPOSE, for a law giving effect to article 39(b) and (c), from articles 14 and 19.
• The Ninth Schedule was drafted for thirteen land laws and now holds hundreds, many unrelated to land. It protects whatever Parliament lists.
• Kesavananda struck down article 31C's second limb, which had excluded judicial review of the declaration.
• Minerva Mills struck down clauses (4) and (5) of article 368 and the Forty-second Amendment's widening of article 31C. Power to destroy is not a power to amend; rights without remedies are as writ in water; Bhagwati J. dissented on article 31C.
• I.R. Coelho: laws added to the Ninth Schedule on or after 24 APRIL 1973 are testable against the basic structure as reflected in article 21 with articles 14 and 19, by the impact test.

Test yourself

1. What historical problem produced all three articles, and why did the Constitution's original text not solve it?

2. Distinguish articles 31A, 31B and 31C by what each immunises and from which rights.

3. What is the structural objection to article 31B, as distinct from an objection to any particular listed law?

4. Trace article 31C through 1971, Kesavananda, 1976 and Minerva Mills.

5. State the six holdings in I.R. Coelho, and explain the significance of 24 April 1973.

6. Does a law in the Ninth Schedule still get any protection at all after Coelho? Explain precisely.

The inverted commas round the word in the article's own text are a signal to that effect: the Constitution is naming a specific historical institution rather than a general concept.

The consequence for an answer. Article 17 has a narrow subject matter and, within it, absolute force. That combination is unusual and worth stating: the article does not do a little about a lot, it does everything about one thing.

The legislation under it

The Protection of Civil Rights Act 1955, originally the Untouchability (Offences) Act, is the law article 17 contemplates. It makes punishable the enforcement of disabilities in relation to entry to places of public worship, access to shops and public places, the use of water sources, and the practice of occupations, and it provides that an offence under it is cognisable.

The Scheduled Castes and the Scheduled Tribes (Prevention of Atrocities) Act 1989 goes further, creating a separate class of offences of atrocity, with special courts and provisions on bail and investigation. It is not enacted under article 17 alone but is part of the same constitutional project.

Article 15(2) works alongside article 17. It forbids subjection to any disability, liability, restriction or condition with regard to access to shops, public restaurants, hotels and places of public entertainment, and the use of wells, tanks, bathing ghats, roads and places of public resort. Like article 17 it binds private persons, so the two together are the horizontal core of the equality code.

Article 25(2)(b) completes the picture: nothing in article 25 prevents the State making a law providing for the throwing open of Hindu religious institutions of a public character to all classes and sections of Hindus.

Article 18: what it abolishes and what it does not

Clause (1) forbids the State to confer a title, with two exceptions: military distinctions and academic distinctions. So a rank, a decoration for gallantry and a university degree are all outside the prohibition.

Clause (2) forbids a citizen to accept a title from a foreign State, without exception.

Clauses (3) and (4) deal with persons holding an office of profit or trust under the State, and require the President's consent for accepting a foreign title, present, emolument or office.

The purpose is not modesty. Hereditary titles were an instrument of the social order the Constitution was dismantling, and a title conferred by the State creates a class marked out by the State itself, which is inconsistent with the equality the code guarantees. The article abolishes the machinery, not merely the practice.

The national awards question

The recurring examination point, and it should be answered precisely.

The national awards, Bharat Ratna, Padma Vibhushan, Padma Bhushan and Padma Shri, were instituted in 1954, discontinued in 1977, revived in 1980, and challenged as titles forbidden by article 18(1).

The position that has been reached is that they are not titles within article 18(1), on the reasoning that they are decorations or awards in recognition of exceptional work, not titles of nobility conferring a status, and that the Constitution does not forbid the State to recognise merit. The decisive practical point is that they are not to be used as suffixes or prefixes to the recipient's name. A person who appends the award to his name is using it as a title, and that is what the article forbids.

So the distinction is between recognising achievement, which is permitted, and conferring a status that attaches to the person and marks him out, which is not.

A worked example

A village temple trust, a private body, refuses entry to persons of a particular caste, and the village panchayat passes a resolution supporting the refusal. Separately, a State Government announces a scheme of hereditary honorific titles for descendants of former rulers, to be used before their names, and confers one on a person who also accepts a decoration from a foreign government while serving as a State officer.

The temple refusal. Article 17 binds private persons, so the trust cannot answer that it is not "State". Enforcing a disability arising out of untouchability is directly forbidden and is an offence in accordance with law, and the Protection of Civil Rights Act 1955 makes the enforcement of disabilities as to entry to places of public worship punishable. A defence of religious freedom under article 25 fails, because article 17 is absolute and because article 25(2)(b) expressly permits legislation throwing open Hindu religious institutions of a public character to all classes and sections of Hindus.

The panchayat resolution. The panchayat is a local authority and therefore "State" under article 12, so this is a straightforward violation of articles 15(2) and 17 by the State as well.

The hereditary titles. Forbidden by article 18(1). They are neither military nor academic; they are hereditary, which is the paradigm of what the article abolishes; and they are to be used before the name, which is the very use the national awards cases treat as decisive.

The foreign decoration. The officer holds an office of profit under the State, so clauses (3) and (4) apply and the President's consent is required. If he is a citizen, clause (2) forbids acceptance of a title from a foreign State outright, so the answer depends on whether what he accepted is a title or a decoration.

Quick revision

• Article 17 abolishes untouchability, forbids its practice in any form, and makes enforcement of a disability arising out of it an offence punishable in accordance with law.
• Four unusual features: it abolishes rather than guarantees; it binds private persons; it is absolute with no exception or restriction; and it directs its own criminal law, which article 35 reserves to Parliament.
• "Untouchability" is undefined and is read in its historical sense, not literally or figuratively.
• The legislation: the Protection of Civil Rights Act 1955 and the Scheduled Castes and Scheduled Tribes (Prevention of Atrocities) Act 1989. Article 15(2) and article 25(2)(b) work alongside.
• Article 18(1) forbids the State to confer a title, excepting military and academic distinctions; (2) forbids a citizen to accept a foreign title; (3) and (4) require the President's consent for holders of office under the State.
• National awards are not titles, provided they are not used as prefixes or suffixes to the name. Recognition is permitted; conferring a status is not.

Test yourself

1. Name four features that make article 17 unlike any other article in Part III.

2. Why is "untouchability" not defined, and how is it construed?

3. Which two provisions of Part III bind private persons, and what do they cover between them?

4. What are the two exceptions in article 18(1), and what is the purpose of the article?

5. Are the national awards titles? Give the reasoning and the decisive practical qualification.

A fundamental duty, which binds nobody in law. Article 51A(e) makes it the duty of every citizen to promote harmony and the spirit of common brotherhood, and to renounce practices derogatory to the dignity of women. Not enforceable, but repeatedly used as an aid to construction and as a reason for reading a statute in a particular way.

Representation, which is a reservation of seats. Articles 243D and 243T for local government, in force and working; articles 330A, 332A and 334A for Parliament and the State Assemblies, enacted and not commenced. See [Women in Local Government] and [The 106th Amendment, and Why It Has Not Started].

Three things about the scheme worth noticing

Almost all of it is enabling or directive, not mandatory. Article 15(3) permits; the directive principles direct without enforceability; the fundamental duty exhorts. The only hard guarantees are the prohibitions in articles 15(1) and 16(2), which forbid the State to discriminate but do not require it to do anything.

It says almost nothing about private conduct. Article 15(2) reaches private discrimination in access to shops, restaurants, hotels, places of public entertainment and public resort, and article 17 reaches untouchability. Neither reaches discrimination in employment, in the family, or in property, which is where most of what women complain of happens. The gap has been filled by legislation and, in one celebrated instance, by judicial guidelines: see [Sexual Harassment at the Workplace].

Personal law is largely outside it. Marriage, divorce, maintenance, guardianship, adoption and succession are governed by the personal law of each community, and whether personal law is "law" for article 13 is unresolved. This is the single largest limitation on the constitutional scheme and it is [Women, Religion and Personal Law].

Where the legislation sits

An answer that names the statutes is stronger than one that stops at the Constitution, because MU's question asks whether the measures are satisfactory and most of the measures are statutory.

Equal remuneration; maternity benefit, now extended in duration and covering creche facilities; the prohibition of dowry; the prohibition of child marriage; protection from domestic violence, a civil law giving a right to residence; sexual harassment at the workplace, enacted in 2013 after the Court's guidelines; the prohibition of pre-natal sex determination; and the amendment of the Hindu Succession Act in 2005 making a daughter a coparcener by birth.

Each of these is ordinary legislation, alterable by ordinary majority. That is the point to make: the constitutional scheme is thin, and the substance is statutory.

Is it satisfactory? How to answer MU's question

The question invites a critical evaluation, so the answer needs a structure rather than a list.

What the scheme does well. Sex is a prohibited ground in both the general and the employment article. The enabling clause is drafted widely enough to authorise almost any protective measure. The directive principles supply an interpretive resource that courts have used repeatedly. And the local government reservations have produced a genuine change in political participation.

Where it is thin. The guarantees are negative. Nothing in Part III requires the State to act, and a State that does nothing for women violates no fundamental right. The reach into private conduct is minimal. Personal law, which governs most of family life, is largely untouched. And article 15(3) is an enabling power whose exercise no woman can compel.

Where it is contradictory. Protective provisions can entrench the disadvantage they address. A rule excluding women from night work, or from certain occupations, is a special provision under article 15(3) and also a restriction on their employment. The line between protection and exclusion is the hardest question in the topic and it runs through [Women at Work].

What has changed recently. The 106th Amendment 2023 is the largest constitutional measure ever enacted for women, and it is not in force, because article 334A defers it to a delimitation after the next census. Any evaluation written today has to say that.

A worked example

MU's stem: "Critically evaluate whether Constitutional measures for Women Empowerment are satisfactory."

Do not begin with a list of articles. Begin with the structure: what kind of provisions are they, and what kind of obligation do they create.

Paragraph one. The scheme has five groups: guarantees, an enabling power, directive principles, a fundamental duty, and reservation of seats. Name them with their articles.

Paragraph two. Note that only the first group is mandatory, and that it is negative: it forbids the State to discriminate and requires nothing of it.

Paragraph three. Show the reach into private conduct and its limits: article 15(2) and article 17 bind private persons, but neither reaches employment or the family.

Paragraph four. Personal law, and the unresolved question whether it is "law" under article 13. This is where the scheme is weakest, and Shayara Bano shows both what a court can do and how narrow the route is.

Paragraph five. The statutory superstructure, named, with the point that it is ordinary law.

Paragraph six. The contradiction: protection can exclude, and Nergesh Meerza is the illustration both ways.

Paragraph seven. The 106th Amendment: the largest measure, not in force, and why.

Conclusion. The constitutional measures are adequate as a source of POWER and inadequate as a source of OBLIGATION. Almost everything that has actually been achieved was achieved by statute or by judicial direction filling a legislative vacuum, and the one great constitutional measure enacted in this generation is waiting on a census. That is a defensible answer, it engages with the word "critically", and every step of it is traceable to a provision.

Quick revision

• Five groups: equality (14, 15(1), 16(1) and (2)); enabling power (15(3)); directive principles (39(a), (d), (e), 42, 44); fundamental duty (51A(e)); representation (243D, 243T, 330A, 332A, 334A).
• Only the guarantees are mandatory, and they are negative: no duty to act.
• Private conduct is reached only by article 15(2) and article 17, neither of which covers employment or the family.
• Personal law is largely outside the scheme, and whether it is "law" for article 13 is unresolved.
• The substance is statutory: equal remuneration, maternity benefit, dowry, child marriage, domestic violence, sexual harassment, pre-natal diagnostic techniques, and the 2005 succession amendment.
• The contradiction to name: a protective provision can also be an exclusion.
• The 106th Amendment is the largest constitutional measure and it has not commenced.

Test yourself

1. Set out the five groups of provisions with their articles, and say what kind of obligation each creates.

2. Why is it accurate to say the constitutional guarantees for women are negative?

3. Which provisions of Part III reach private conduct, and what do they fail to reach?

4. Name six statutes that make up the substantive law, and say why their statutory character matters.

5. Explain the contradiction between protection and exclusion, with an example.

6. Answer MU's question in seven paragraphs, as outlined above.

It is an exception to the whole article, including clause (2). That is significant, because clause (2) binds private persons. A law requiring a private establishment to provide separately for women is within article 15(3).

It is enabling only. A woman has no fundamental right to a special provision, and a court will not direct the State to make one. What a court can do is test a provision that has been made.

The limit: in favour, never against

The clause authorises provision for women. It cannot be used to justify a rule that burdens them.

The distinction sounds obvious and is not, because a great many restrictions on women have historically been defended as protective: rules excluding them from night work, from underground work, from certain occupations, from certain hours, from certain postings. Each was framed as a benefit and each operated as an exclusion.

The test to apply is whether the provision enlarges the choices available to women or removes choices from them. A creche, a maternity benefit, a reserved seat and a relaxed age limit enlarge. A prohibition on employment after dark removes, and it is not saved by article 15(3) merely because its authors thought it kind.

Facts. Air India v. Nergesh Meerza, AIR 1981 SC 1829, (1981) 4 SCC 335, decided 28 August 1981. Air hostesses retired at thirty-five, or on marriage within four years of joining, or on first pregnancy, whichever occurred earlier, with the Managing Director holding a discretion to extend service; male cabin crew were not on comparable terms.

Held. The provisions were a clear case of official arbitrariness. The words "or on first pregnancy whichever occurs earlier" are unconstitutional, void and violative of article 14, and stand deleted. The Managing Director's unguided option to extend was also struck down, because it allowed him to discriminate between one air hostess and another; extensions must be granted as a matter of course for ten years where she is medically fit. Being severable, the offending parts alone were struck down.

Why it matters. It shows the limit in operation: a rule cannot be defended as a special provision for women when its effect is to end their employment. And read it honestly, because it cuts both ways. The Court did not strike down the different retirement age as such, and its treatment of air hostesses and male cabin crew as distinct classes, so that no comparison arose, is the part of the judgment most criticised. A modern court would probably ask whether the classification was itself sex-based.

Article 15(3) and public employment

A recurring puzzle. Article 15(3) is in article 15; public employment is governed by article 16; and article 16 contains no clause corresponding to 15(3). May the State reserve posts for women?

The orthodox answer is yes, and the route matters.

Not by article 16(4), which speaks of a "backward class of citizens". Women are not a backward class, and treating them as one would misdescribe both the group and the provision.

By article 16(1) read with article 15(3). In Indra Sawhney the Court held that article 16(4) is not exhaustive of the whole subject of reservation: "Reservations can also be provided under Clause (1) of Article 16", which is not confined to preferences, concessions or exemptions alone, provided they are adjusted so as not to exceed the level of representation prescribed for backward classes. Since article 16(1) is an application of the equality principle, and article 15(3) confirms that provision for women is consistent with that principle, a reservation for women is sustained under article 16(1).

In practice such reservations are usually made horizontally: the seats for women are taken from within each vertical category rather than in addition to them, so that a woman of a reserved category is counted in her own category. See [Reservation in Public Employment] for the vertical and horizontal distinction.

A worked example

A State makes three rules for its transport undertaking. Rule 1 reserves thirty per cent of conductor posts for women, applied horizontally. Rule 2 provides that no woman shall be posted on night shifts. Rule 3 provides that a woman conductor shall be entitled to twenty-six weeks of maternity leave and a creche at every depot. A male candidate challenges Rule 1; a woman conductor challenges Rule 2; the undertaking says all three are special provisions under article 15(3).

Rule 1. Valid. Reservation for women in public employment is sustained under article 16(1) read with article 15(3), on Indra Sawhney's holding that article 16(4) is not exhaustive of the subject of reservation. Being horizontal, it does not add to the vertical percentages and does not disturb the ceiling. The male candidate's complaint that he was treated differently is admitted and is not the question; the question is whether the provision is a permissible special provision, and thirty per cent in a service where women are severely under-represented is defensible.

Rule 2. Not saved. It is framed as protection and operates as exclusion: it removes from women a set of postings, and with it overtime, allowances and the experience that leads to promotion. Article 15(3) authorises provision for women and this provision is against them. The proper approach is to secure safe conditions for night work, not to forbid the work.

Rule 3. Valid, and squarely within article 15(3), reinforced by article 42's direction on maternity relief and just and humane conditions of work. It enlarges rather than removes.

The test that decides all three. Does the provision enlarge the choices available to women or remove them?

Quick revision

• Article 15(3): nothing in article 15 shall prevent the State making any special provision for women and children.
• Better regarded as declaratory of what article 14 permits than as an exception, on the same reasoning by which Indra Sawhney treats article 16(4) as an instance of 16(1).
• "Special provision" is wider than reservation; it is an exception to the whole article including clause (2), which binds private persons; and it is enabling only.
• The limit: provision FOR women, never against. Test whether it enlarges or removes choices.
• Nergesh Meerza: termination on first pregnancy is official arbitrariness and void; an unguided power to extend is void. But the different retirement age was not struck down, and the class analysis is criticised.
• Reservation for women in public employment rests on article 16(1) with article 15(3), not on 16(4), because Indra Sawhney holds 16(4) is not exhaustive of the subject. Usually made horizontally.

Test yourself

1. Why is it better to describe article 15(3) as declaratory than as an exception? What difference does the choice make?

2. Give four features of the clause, including the significance of its applying to clause (2).

3. State the test for distinguishing a protective provision from an exclusion, and give an example of each.

4. What was struck down in Nergesh Meerza, what was not, and what is the standard criticism of it?

5. On what provision does a reservation for women in public employment rest, and why not on article 16(4)?

Read the case honestly, because MU's examiners know its weaknesses. The Court did not strike down the lower retirement age. It held that air hostesses and male cabin crew formed separate classes with different recruitment, promotion and conditions, so that no comparison arose between them, and it is that classification which is criticised: if the class is defined by the job women hold, then a comparison between the sexes can always be avoided by defining the job narrowly enough. A modern court would ask whether the classification is itself sex-based.

Retirement age and the comparator problem

The general position now is that a service condition fixing a lower retirement age for women, without a reason connected to the work, is discrimination on the ground of sex and is bad.

The difficulty is the comparator. Article 16(2) forbids discrimination on the ground of sex, which requires comparing the treatment of a woman with that of a man in the same position. Where the employer has created a separate cadre consisting only of women, there is no man in the same position, and the comparison collapses. That is precisely what happened in Nergesh Meerza.

The answer that has developed is to ask whether the cadre itself is sex-based. If the only reason a job is held exclusively by women is a policy of employing only women in it, then the cadre is a product of the discrimination complained of and cannot be used to justify it. An answer that spots this point is doing the analysis rather than reciting the case.

Night work and hazardous work

The classic protective restrictions, and the modern trend is against them.

The old position. Legislation prohibited the employment of women at night and in underground mines, and such prohibitions were defended as special provisions under article 15(3) and as giving effect to article 39(e), the health and strength of workers.

The objection. A blanket prohibition removes employment, overtime and the experience on which promotion depends, and it does so on an assumption about safety that the State could address directly by regulating conditions instead. It is protection that operates as exclusion, which is the limit on article 15(3) set out in [Special Provision for Women].

The modern approach. Where such prohibitions have been challenged, the tendency has been to strike them down or read them down, and to substitute conditions: adequate security, transport, sufficient numbers, consent. Legislation has moved the same way, permitting night work subject to safeguards.

The proposition to state. The constitutional answer to a risk faced by women at work is to remove the risk, not to remove the women.

The armed forces

The recurring modern example, and it is worth naming even though the leading decisions are in the case bank's queue and are not cited here.

The pattern of the litigation has been consistent: women officers were granted short service commissions but denied permanent commission, and denied command appointments, on grounds of physiological differences, the composition of the rank and file, and operational conditions. The courts have rejected those grounds as stereotypes rather than reasons, and have held that a blanket exclusion based on sex, unsupported by evidence about the individual, offends articles 14 and 16.

Article 33 is the provision to notice here: it permits Parliament to restrict or abrogate the application of the fundamental rights to members of the armed forces. So the constitutional position in this field is not that the rights apply in full, but that any restriction must come from a law made under article 33, and cannot be assumed from the nature of the service.

Equal pay, maternity and benefits

Equal pay for equal work is not itself a fundamental right. It is a directive principle in article 39(d), and it has been given effect by legislation. Its constitutional force comes from being read with articles 14 and 16: paying two people differently for the same work, with no difference in duties, responsibility or qualification, is discrimination without a differentia. So the claim is framed under article 14, with article 39(d) as the interpretive support.

Maternity relief is directed by article 42 in terms, and is provided by statute, extended in 2017 to twenty-six weeks with creche facilities in larger establishments. A denial of maternity benefit to a woman in public employment is challengeable under article 14 read with articles 42 and 39(e).

The deeper point for an answer. Maternity benefit is the one place where the Constitution requires the State to treat women differently in order to treat them equally. Requiring a pregnant woman to be treated identically with everyone else would end her employment; the special provision is what makes her participation possible. That is substantive equality in its clearest form, and it is the answer to anyone who says article 15(3) is a departure from equality.

A worked example

A State police service provides: (a) that the maximum age for recruitment of women is twenty-five and of men thirty; (b) that women officers shall not be posted to night patrol duty; (c) that a woman officer who marries shall inform the department and may be transferred to her spouse's district; and (d) that a woman officer shall be entitled to twenty-six weeks of maternity leave. A woman officer challenges (a), (b) and (c).

(a) The age limit. Directly sex-based and to her disadvantage: it shortens the period in which she may apply. Article 15(3) does not save it because it is not a provision for women. No reason connected to the work is apparent. It fails under article 16(2) and, independently, as arbitrary on Royappa.

(b) Night patrol. Framed as protection, operating as exclusion: it removes duties that carry allowances and count for promotion. The constitutional answer to the risk is to secure the conditions, not to withdraw the posting. It fails.

(c) Transfer on marriage. Look carefully, because this one is a genuine question. If the rule entitles her to seek a transfer, it enlarges her choices and is a permissible special provision. If it requires her to be transferred, it removes a choice and treats her career as subordinate to her spouse's, which is the stereotype the equality guarantee exists to defeat. The drafting decides it: "may be transferred" on her request is good, "shall be transferred" is not.

(d) Maternity leave. Valid, and required rather than merely permitted, being supported by article 42 and article 39(e) and being the paradigm of a provision that makes equal participation possible.

Quick revision

• Four recurring questions: entry, continuance, the conditions of the work, and pay and benefits.
• Nergesh Meerza: termination on first pregnancy is official arbitrariness and void; the unguided extension power is void; but the retirement age was NOT struck down and the separate-class reasoning is criticised.
• The comparator problem: where a cadre is all-female there is no male comparator. Ask whether the cadre itself is sex-based.
• Night and hazardous work: the answer to a risk is to remove the risk, not the women. Protection that excludes is outside article 15(3).
• Armed forces: exclusions based on stereotype fail, and any restriction must come from a law under article 33.
• Equal pay is article 39(d), not a fundamental right; the claim is framed under article 14 with 39(d) in support.
• Maternity relief under article 42 is where the Constitution requires different treatment in order to achieve equal participation.

Test yourself

1. Name the four recurring questions in this area and give an example of each.

2. What exactly did Nergesh Meerza strike down, and what is the criticism of what it did not?

3. Explain the comparator problem and the answer that has developed to it.

4. Why is a blanket prohibition on night work not saved by article 15(3)?

5. What is the significance of article 33 in the armed forces cases?

6. Is equal pay for equal work a fundamental right? How is the claim framed?

What the Court did. It held that the fundamental rights identified above were being violated, that the State was under an obligation to secure them, and that in the absence of enacted law the Court would lay down guidelines and norms to be observed at all workplaces, which would be binding and enforceable in law until suitable legislation is enacted, being the law declared by the Court under article 141.

What the guidelines contained. A definition of sexual harassment covering physical contact and advances, a demand or request for sexual favours, sexually coloured remarks, showing pornography, and any other unwelcome physical, verbal or non-verbal conduct of a sexual nature. A duty on the employer to prevent or deter such acts and to provide procedures for resolution, settlement and prosecution. A Complaints Committee headed by a woman, with not less than half its members women, and including a third party such as a non-governmental organisation familiar with the issue, to guard against undue pressure from senior levels. Provision for criminal proceedings, disciplinary action, and protection against victimisation of the complainant and witnesses.

Where the content came from. The Court drew on international conventions and norms, in particular the Convention on the Elimination of All Forms of Discrimination against Women, holding that in the absence of domestic law occupying the field, international conventions consistent with the fundamental rights and in harmony with their spirit must be read into them to enlarge their meaning and content. That reasoning connects this chapter to article 51(c) and to the treaty power in article 253, dealt with in [When Parliament May Legislate on a State Subject].

The report of Vishaka could not be opened from the sources this book is built from. It is named without a citation and none should be invented. The account above is the standard one and should be checked against the report before it is relied on in writing.

The constitutional objection, and the answer

This is what makes the case an LL.M. topic rather than a labour law topic.

The objection. A court laid down a detailed code of conduct, with definitions, machinery, composition of committees and procedures. That is legislation. Under the separation of powers the Court's function is to decide disputes and to interpret law, not to write it, and article 141 makes the law declared by the Supreme Court binding, which is not the same as a power to enact.

The answer. The Court did not create the right; the right was in Part III already. What was missing was the machinery to enforce it, and article 32 obliges the Court to provide an effective remedy for the violation of a fundamental right. Guidelines that fill a vacuum, expressly stated to operate only until the legislature acts, are a remedy rather than a statute. The Court also had the express warrant of article 253 and the international obligations India had assumed.

The honest assessment. The answer is a good one and it does not fully dispose of the objection. The guidelines had every characteristic of legislation except the process, and the legislature took sixteen years to replace them, during which a judicial code governed conduct throughout the country. Whether that is judicial activism vindicated or judicial overreach tolerated is exactly the debate in [Judicial Activism, Restraint and Overreach], and Vishaka is the strongest example on the vindication side precisely because the vacuum was real and the legislature's inaction was long.

The 2013 Act

The Sexual Harassment of Women at Workplace (Prevention, Prohibition and Redressal) Act 2013 replaced the guidelines. Its structure follows them and its differences are examinable.

It applies to all workplaces, organised and unorganised, public and private, and to a woman of any age whether employed or not, so a visitor or a customer is covered. It requires an Internal Committee in every workplace with ten or more workers, presided over by a woman employed at a senior level, with not less than half the members women and one member from a non-governmental organisation. It provides for a Local Committee at district level for establishments with fewer than ten workers and for complaints against the employer himself. It permits conciliation at the woman's request, but not a monetary settlement as the basis of it. It requires inquiry within ninety days and action within sixty days of the report. It protects the complainant, including by transfer or leave during the inquiry. And it penalises the employer for failing to constitute a committee or to comply.

The provision most criticised is section 14, which permits action against a woman for a false or malicious complaint. The objection is that it deters complaints, since a complaint that fails for want of proof may be portrayed as false, and the Act's own proviso that mere inability to substantiate is not enough does not fully meet it in practice.

A worked example

A woman engaged as a consultant, not an employee, complains that a senior officer of the organisation has made repeated sexually coloured remarks. The organisation has forty workers but has constituted no Internal Committee. It says she is not an employee, that the remarks were jokes, and that in any event there is no evidence.

Is she covered? Yes. The Act protects an aggrieved woman in relation to a workplace, whether or not she is employed there, so a consultant, a visitor, a client or a domestic worker is within it. Her status is not the test; the workplace is.

Is the conduct within the definition? Sexually coloured remarks are expressly within the definition adopted in Vishaka and carried into the Act. Whether they were meant as jokes is not the test; the test is whether the conduct was unwelcome.

What of the missing committee? The organisation has forty workers, so an Internal Committee is mandatory, and the failure is itself a contravention for which the employer is penalised. She may complain to the Local Committee at district level.

And the constitutional dimension, which is what the examiner wants? Her rights under articles 14, 15, 19(1)(g) and 21 are engaged. If the organisation is "State" under article 12 she may proceed under article 32 or 226 directly. If it is not, the statutory machinery is her route, and article 226 may still lie if the body discharges a public function, on the reasoning in [Private Bodies and Public Functions].

On the evidence point. The inquiry is not a criminal trial; the standard is the civil one, and the committee has the powers of a civil court for specified purposes.

Quick revision

• The constitutional foundation is four articles: 14, 15, 19(1)(g) and 21. Naming 19(1)(g) is what shows the wrong is to the right to work.
• Vishaka: no statute existed; the Court laid down binding guidelines under article 141 to operate until legislation; drew on CEDAW, reading international conventions consistent with the fundamental rights into them.
• The guidelines: a definition, an employer's duty to prevent, a Complaints Committee headed by a woman with a majority of women and a third-party member, and protection against victimisation.
• The objection is that this was legislation; the answer is that the right existed and only the remedy was missing, and article 32 requires an effective remedy.
• The 2013 Act: all workplaces, any aggrieved woman, Internal Committee where ten or more workers, Local Committee otherwise, conciliation without monetary settlement, ninety days for inquiry, penalties on the employer.
• Section 14 on false or malicious complaints is the most criticised provision.

Test yourself

1. Which four fundamental rights are engaged by sexual harassment at work, and why does article 19(1)(g) matter to the analysis?

2. What was the legal vacuum in Vishaka, and on what basis did the Court say its guidelines would bind?

3. Where did the content of the guidelines come from, and what does that show about the use of international law?

4. State the separation-of-powers objection to Vishaka and the best answer to it.

5. Give five features of the 2013 Act, and say which provision is most criticised and why.

Route one, construe the personal law. Hold that the practice complained of is not part of the religion at all, so no question of protection arises. This avoids the constitutional question entirely.

Route two, test the statute. Where a statute gives effect to or recognises the practice, the statute is law beyond argument, and it can be tested under article 14. This is how manifest arbitrariness enters the field.

Route three, use article 25's own limits. Article 25(1) is expressly subject to public order, morality and health and to the other provisions of Part III. Those last words are decisive: freedom of religion is subordinated to the other fundamental rights, including equality, within article 25 itself. And article 25(2)(b) permits laws for social welfare and reform.

Route four, the essential religious practices doctrine. Only a practice essential to the religion attracts the protection of article 25. If the practice is not essential, it may be regulated freely. See [Essential Religious Practices].

Triple talaq

Facts. Shayara Bano v. Union of India, decided 22 August 2017 by five judges. The petitioner's husband pronounced talaq three times at once. She sought a declaration that talaq-e-biddat, the instantaneous and irrevocable triple talaq, is void and that the practice is unconstitutional as violating articles 14, 15 and 21. The respondents said it was personal law protected by article 25 and recognised by section 2 of the Muslim Personal Law (Shariat) Application Act 1937.

Held. The order of the Court records: "In view of the different opinions recorded, by a majority of 3:2 the practice of talaq-e-biddat, triple talaq, is set aside."

The three in the majority did not agree on the route. Kurian Joseph J. took route one: the practice is not part of Islamic law at all, being contrary to the Quran, so nothing protected is affected. Nariman and Lalit JJ. took route two: the 1937 Act is a statute, it recognises the practice, and a rule permitting a man to dissolve a marriage instantly and irrevocably at his whim is manifestly arbitrary and void under article 14. Khehar C.J. and Nazeer J. dissented on route four reasoning: the practice is an essential part of the religion, protected by article 25, and the remedy lies with Parliament, which they would have directed to legislate.

Why it matters. It is the leading modern case, and it is a case about method as much as outcome. Cite it precisely: the result is clear, the ratio is not, and a student who attributes "manifest arbitrariness" to the Court rather than to two of its members is overstating.

What followed. Parliament enacted legislation in 2019 declaring the pronouncement void and making it an offence, which raised a separate question about criminalising a civil wrong that the Court had already declared ineffective.

Entry to places of worship

Facts. Indian Young Lawyers Association v. State of Kerala, decided 28 September 2018 by five judges. Women between the ages of ten and fifty were excluded from the Sabarimala temple, an exclusion resting on custom and given effect by Rule 3(b) of the Kerala Hindu Places of Public Worship (Authorisation of Entry) Rules 1965. The referred questions were whether an exclusion based on a biological factor exclusive to women is discrimination offending articles 14, 15 and 17 and unprotected by "morality" in articles 25 and 26; whether it is an essential religious practice; whether the temple is a religious denomination under article 26; and whether Rule 3(b) is ultra vires the parent Act.

Held. By 4:1 the writ petition was allowed and the exclusion held unconstitutional. Indu Malhotra J. dissented, holding that what constitutes an essential religious practice is for the religious community to decide and that courts should not ordinarily intervene, and that the petitioners, not being devotees, lacked standing.

Its authority is qualified and an answer must say so. A review was entertained, and in November 2019 larger questions about the relationship between articles 25 and 26 and the other fundamental rights were referred to a Bench of nine judges. So the reasoning is under reconsideration, and presenting Sabarimala as settled overstates the position.

Why it matters. It is the clearest modern confrontation between equality and religious practice, and Indu Malhotra J's dissent is the strongest statement of the case for judicial restraint in this field: that a court is not equipped to decide what a religion requires, and that the right in article 25 belongs to the believer rather than to the judge.

The unfinished business

Article 44 directs the State to endeavour to secure a uniform civil code. It is a directive principle, unenforceable, and it is the constitutional answer the framers actually provided to the problem this chapter describes. See [A Uniform Civil Code].

Codification has been the practical route: the Hindu law statutes of 1955 and 1956, the amendment of 2005 making a daughter a coparcener by birth, and the special provision for civil marriage. Codified law is law beyond argument, so equality applies to it directly, which is why reform by codification sidesteps the threshold question altogether.

The point to make in an answer. Every advance in this field has come by avoiding the constitutional question rather than by answering it. That is a defensible judicial strategy and it leaves the largest question in the topic exactly where it was in 1950.

A worked example

A community's personal law provides that a daughter inherits half the share of a son. A daughter challenges the rule as violating articles 14 and 15(1). There is no statute codifying the rule, but the courts of the State apply it as the personal law of the community.

Route one, is it "law" under article 13? Unresolved. If it is, the rule discriminates on the ground of sex and is void. She should plead it, but she cannot rely on it as settled.

Route two, is there a statute to attack? Here there is none, which weakens her considerably. Had a statute recognised or given effect to the rule, she could have attacked the statute as manifestly arbitrary on the Nariman and Lalit reasoning in Shayara Bano, without needing the threshold question answered.

Route three, article 25 and its limits. She may argue that the community's freedom under article 25(1) is expressly subject to the other provisions of Part III, so it cannot be used to defeat article 15(1); and that article 25(2)(b) permits legislation for social welfare and reform, which shows that reform of religious practice was contemplated.

Route four, essential religious practice. She should argue that a rule of inheritance is a secular matter regulated by religion rather than a practice essential to the religion, so it does not attract article 25 at all. This is often the strongest route, and it is the one Kurian Joseph J. used in substance in Shayara Bano.

What the State can do that she cannot. Legislate. Codification would end the argument, and article 44 contemplates it.

Advice. Plead the threshold question, but win on route four if you can.

Quick revision

• The threshold question: is personal law "law" under article 13? Unresolved, and decisive.
• The arguments: article 13(3)(a) does not name it and article 44 provided separately; against, that it is applied by State courts and is given force by statute.
• Four routes used instead: construe the personal law, test the statute, use article 25's own limits, or apply the essential religious practices doctrine.
• Article 25(1) is subject to public order, morality and health AND to the other provisions of Part III. Article 25(2)(b) permits social welfare and reform.
• Shayara Bano: 3:2, triple talaq set aside, but on three different reasons. Nariman and Lalit JJ. used manifest arbitrariness on the 1937 Act.
• Sabarimala: 4:1, exclusion held unconstitutional, Indu Malhotra J. dissenting. Under reconsideration by nine judges since November 2019.
• Every advance has come by avoiding the constitutional question. Codification is the route that sidesteps it entirely.

Test yourself

1. State the threshold question and explain why it is decisive.

2. Give the argument each way on whether personal law is "law" under article 13.

3. Name the four routes courts have used instead, and match each to a judgment or an opinion.

4. Why do the closing words of article 25(1) matter so much in this topic?

5. Set out the three different reasons in the Shayara Bano majority and say why the distinction matters for citation.

6. What is the present authority of Sabarimala, and why must an answer qualify it?

Offices of chairperson. Clause (4) of each article extends the one-third to the offices of Chairperson at each level, which matters because the reservation of ordinary seats without the offices would leave the presiding positions untouched.

Rotation. The reserved seats are allotted by rotation to different constituencies. This is the most consequential design choice in the whole scheme and it is discussed below.

Every level. Article 243D applies at village, intermediate and district level, so the reservation runs through all three tiers.

Rotation, and the argument about it

What it does. A constituency reserved for women at one election is not reserved at the next; the reservation moves. The object is to prevent the reservation from permanently disenfranchising the men of one constituency while permanently benefiting one set of women, and to spread the opportunity.

The objection, and it is a serious one. Rotation means that a woman elected from a reserved constituency generally cannot seek re-election from it, because it will not be reserved next time and she must contest against the incumbent political networks she was elected to bypass. So the scheme systematically prevents the accumulation of incumbency, experience and a political base, which are precisely the things that turn a seat into a career.

The consequence often observed. A high turnover of women representatives, many serving a single term, and a corresponding difficulty in building the political capital that would let them compete for unreserved seats later. The related criticism is the phenomenon of the proxy representative, where a woman is elected and a male relative exercises the function.

The answer. Rotation is a transitional device, and the alternative, permanent reservation of particular constituencies, would create a different and worse distortion. The proxy problem is real but it is a problem of political culture that the reservation is slowly changing rather than an argument against having it.

How to use this in an answer. MU's question asks whether political reservation is part of grass root democracy. The strongest answer says yes, and then says that the design of the reservation determines how much democracy it actually produces, with rotation as the illustration.

What is known about its effect

Be careful here, and say what is a legal proposition and what is not.

The legal position is that a third of seats and a third of chairperson offices at every level of local government are reserved, that States may and many do raise it to a half, and that the reservation extends within the seats reserved for the Scheduled Castes and Scheduled Tribes.

The empirical claims commonly made, that reserved panchayats spend differently, that women's reported willingness to stand rises after a constituency has been reserved, and that the proxy phenomenon declines with successive terms, come from research rather than from law. An answer may cite them as findings, and should not present them as constitutional propositions.

The comparison to make

An examiner asking about grass root democracy is inviting a comparison, and the comparison is with the 106th Amendment.

Local government. Enacted 1992, in force 1993, one-third with States free to raise it, rotation, applies at every tier and to chairperson offices, operating for over thirty years.

Parliament and the State Assemblies. Enacted 2023, one-third, rotation after each delimitation, and not in force, because article 334A defers commencement to a delimitation following the first census after the amendment's commencement.

The point. The Constitution has been willing to reserve seats for women where the offices are numerous, local and relatively low in political value, and has taken thirty years longer to do it where the offices are few, national and highly contested. That observation is fair, it is supported by the dates, and it is the kind of critical remark MU's "critically evaluate" questions are looking for.

A worked example

A State law provides that in its panchayats, fifty per cent of directly elected seats shall be reserved for women; that the reservation shall be rotated after every two elections rather than every one; and that the office of Chairperson at district level shall not be reserved for women because the office is too important to be filled by rotation. A voter challenges all three provisions.

Fifty per cent. Valid. Article 243D(3) requires "not less than one-third", which is a floor. A State may reserve more, and many have.

Rotation after two elections. Valid. The article requires that reserved seats "may be allotted by rotation to different constituencies" and does not prescribe the interval. A longer interval is arguably better policy, since it allows an incumbent one re-election, and nothing in the article forbids it.

Excluding the district Chairperson. Invalid. Article 243D(4) requires that not less than one-third of the offices of Chairpersons in the Panchayats at each level shall be reserved for women. "Each level" includes the district level, and the State's reason, that the office is too important, is precisely the reason the clause exists.

A further point worth making. The first two provisions show how much room the article leaves a State, and the third shows where it leaves none. An answer that distinguishes the mandatory floor from the discretionary detail is doing the article rather than describing it.

Quick revision

• Articles 243D and 243T, from the 73rd and 74th Amendments 1992, in force 1993.
• Not less than one-third of directly elected seats reserved for women, at every tier; a floor, and States may raise it, many to one-half.
• The reservation operates WITHIN the seats reserved for the Scheduled Castes and Scheduled Tribes as well as the general seats, so it is horizontal.
• Not less than one-third of the offices of Chairperson at each level.
• Seats are allotted BY ROTATION, which is the most consequential design choice and the main criticism, because it prevents incumbency and a political base.
• The proxy representative is the other standard criticism.
• This is the ONLY political reservation for women actually in force in India; the 106th Amendment has not commenced.

Test yourself

1. Set out what articles 243D and 243T each require, in four points.

2. Why is the one-third described as horizontal rather than vertical, and which clause says so?

3. Explain rotation, the reason for it, and the strongest objection to it.

4. May a State reserve more than one-third? May it reserve none of the chairperson offices? Give the reasoning for each.

5. Compare the local government reservation with the 106th Amendment on four points, and say what the comparison suggests.

Unpack the conditions. Three things must happen, in order, before a single seat is reserved.

One, a census must be taken after the commencement of the Amendment Act, and its relevant figures published.

Two, an exercise of delimitation must be undertaken for the purpose of the reservation, after those figures are published.

Three, only then do the provisions come into effect.

And a fourth point on duration. The reservation ceases on the expiration of fifteen years from such commencement, subject to clause (2), under which the seats continue till such date as Parliament may by law determine.

Clause (3) provides that rotation of the reserved seats shall take effect after each subsequent exercise of delimitation as Parliament may determine. Clause (4) provides that nothing in the article affects any representation in the existing Houses until their dissolution.

Why nothing has happened

Three facts, in sequence, and they are what a current answer must contain.

The census. The census due in 2021 was not taken. Until a census is taken and its relevant figures published, condition one is not satisfied and the amendment cannot commence.

Delimitation is frozen. Article 82 requires readjustment of the allocation of seats in the House of the People after each census, but the allocation has been frozen by successive amendments, most recently until the first census taken after 2026. So the delimitation the reservation waits for is itself tied to a census that has not been taken and a freeze that has not lifted.

The Bill that would have moved it was defeated. The Constitution (One Hundred and Thirty-first Amendment) Bill, 2026 was introduced in the Lok Sabha on 16 April 2026 and was defeated the following day, 17 April 2026, having failed to obtain the special majority article 368 requires. Its object was to implement the one-third reservation for women in the Lok Sabha and in the State and Union territory Assemblies.

So the position as this book is written is that articles 330A and 332A are part of the Constitution and are not in operation, and the route by which they would have been brought into operation sooner has been rejected by Parliament.

Do not write that India has a women's reservation in Parliament. Write that it has enacted one which has not commenced.

The connection to federalism

This is the part that distinguishes an LL.M. answer, and it links topic 4 back to topic 1.

The delimitation on which the reservation waits is not a technical exercise. Readjusting the allocation of seats to population, after a freeze of half a century, will transfer seats from States whose population grew slowly to States whose population grew quickly. The States that would lose relative weight argue that they are being penalised for having succeeded at a national population policy, and they treat the freeze as a bargain that should not be reopened.

The consequence is that the women's reservation has been tied to the most contested federal question in the country. A measure with near-universal political support cannot commence until a measure with deep regional opposition is carried out. That is a drafting choice, and it is a fair subject for criticism: the reservation could have been brought into force on the existing constituencies without waiting for delimitation at all.

See [Cooperative and Competitive Federalism] for the delimitation question in its own right.

Evaluating it, which is what MU asks

The case that it is a real advance. It is the largest constitutional measure ever enacted for women. It reaches the House of the People and every State Assembly. It operates within the Scheduled Caste and Scheduled Tribe seats, so it does not benefit only women of the general category. And it carries a fifteen-year horizon with provision for extension, which answers the objection that such measures never end.

The case that it is not. It is not in force and there is no date on which it will be. Its commencement is conditioned on two events, neither within the control of those it benefits. It contains no sub-quota for women of the other backward classes. Rotation, on the experience of local government, prevents the accumulation of incumbency, and is here tied to delimitation exercises that occur once in decades. And a reservation that ceases after fifteen years from commencement, when commencement is itself indefinite, is a promise whose duration is measured from a date nobody can name.

The conclusion that the material supports. As an answer to MU's question whether the constitutional measures are satisfactory, the 106th Amendment is the strongest evidence for both sides: it shows that the Constitution is capable of a large and genuine measure for women, and it shows that such a measure can be enacted in a form that does nothing at all.

A worked example

A candidate is asked in April 2026: "Discuss the constitutional provisions for reservation of seats for women in legislatures."

What a weak answer says. That the 106th Amendment reserves one-third of seats in the Lok Sabha and State Assemblies, describes articles 330A and 332A, and concludes that women's representation will now increase.

What is wrong with it. It states as operative a provision that is not in force, and it omits article 334A, which is the only part of the amendment currently doing anything.

What a good answer does.

One. States what articles 330A and 332A provide, including that the reservation is horizontal within the Scheduled Caste and Scheduled Tribe seats.

Two. Sets out article 334A in full, unpacking the three conditions, and says plainly that the amendment has not commenced.

Three. Gives the three facts: no census since 2011, delimitation frozen until after the first census taken after 2026, and the 131st Amendment Bill defeated on 17 April 2026.

Four. Draws the federalism connection: the reservation is tied to the delimitation question, which is contested between the States.

Five. Contrasts it with articles 243D and 243T, which are in force and have operated since 1993.

Six. Evaluates, giving both cases as above.

Why this scores. Every proposition is traceable to the text or to a dated event, and the answer engages with the question actually set rather than the announcement that was made in 2023.

Quick revision

• The 106th Amendment 2023 inserted articles 330A, 332A and 334A, and amended article 239AA for Delhi.
• One-third of seats, as nearly as may be, in the House of the People and every State Assembly, INCLUDING within the seats reserved for the Scheduled Castes and Scheduled Tribes. No sub-quota for the other backward classes.
• Article 334A: commencement waits for a census taken after the Amendment Act, publication of its figures, and then a delimitation undertaken for the purpose. Duration fifteen years from commencement, subject to Parliament.
• Nothing has commenced: the 2021 census was not taken, delimitation is frozen until after the first census taken after 2026, and the 131st Amendment Bill was defeated on 17 April 2026.
• The reservation is therefore tied to the most contested federal question in the country.
• Contrast articles 243D and 243T, in force since 1993.

Test yourself

1. What do articles 330A and 332A provide, and in what sense is the reservation horizontal?

2. Set out the three conditions in article 334A that must be satisfied before the reservation commences.

3. Give the three facts that explain why nothing has happened, with dates.

4. Why is the women's reservation tied to a federalism dispute, and could it have been drafted otherwise?

5. Contrast the 106th Amendment with articles 243D and 243T on four points.

6. Using this amendment, argue both that the constitutional measures for women are satisfactory and that they are not.

Held. The Act and Order were void as violating article 19(1)(a) and not saved by article 19(2). The freedom of speech and expression includes the freedom of the press. For propagating his ideas a citizen has the right to publish, to disseminate and to circulate them, and the right extends not merely to the matter he is entitled to circulate but also to the volume of circulation. The Act restrained that aspect and was directed against circulation. Article 19(2) does not permit the State to abridge the right in the interests of the general public: the grounds in clause (2) are exhaustive, and helping smaller newspapers is not among them.

Why it matters. Two propositions of the first importance. Circulation is part of the right, so a measure that reduces readership restricts speech even if it censors nothing. And a benevolent object is irrelevant unless it falls within article 19(2).

The right not to be taxed into silence.

Facts. Indian Express Newspapers (Bombay) Pvt Ltd v. Union of India, AIR 1986 SC 515, (1985) 2 SCR 287, decided 6 December 1984. Customs duty was imposed on imported newsprint, and newspapers said a tax which raises the cost of production and reduces circulation restricts article 19(1)(a).

Held. The expression "freedom of press" is not used in article 19 but is included in article 19(1)(a). Freedom of the press means freedom from interference from authority which would have the effect of interference with the content and circulation of newspapers. There can be no restriction other than those in article 19(2), and no interference in the name of public interest: even when clause (2) was substituted by the First Amendment in 1951, Parliament did not add the general interest of the public.

Why it matters. It confirms that the grounds are exhaustive, and it establishes that a tax is not automatically outside article 19: the question is whether its effect is on content or circulation. The press is not immune from general taxation, but a tax aimed at, or operating on, circulation is a restriction.

The right of a company's shareholders to speak through it.

Facts. Bennett Coleman & Co v. Union of India, AIR 1973 SC 106, (1972) 2 SCC 788, decided 30 October 1972. The newsprint policy limited the pages a newspaper might publish, prevented a common ownership unit from starting a new paper, and restricted interchange of quota. The Union said this was rationing of a scarce commodity, touching trade rather than speech, and that a company cannot claim article 19(1)(a).

Held. By the majority, the fundamental rights of shareholders as citizens are not lost when they associate to form a company; where their rights as shareholders are impaired by State action they are protected, because the shareholders' rights are necessarily affected if the company's are. The individual rights of editors, directors and shareholders are expressed through their newspapers, so the shareholders' locus standi is beyond challenge and the presence of the company is no bar to relief. On the merits the policy fell: a limit on pages operates directly on circulation and content and cannot be defended as mere rationing.

Why it matters. It solves the standing problem created by article 19 being a citizens' right, and it establishes the effect test: what matters is what a measure does to speech, not how it is labelled.

What the freedom does not include

An honest answer says this too.

No right of access to government information as such, before the Right to Information Act. The right to information was read into article 19(1)(a) later and by a different route, dealt with in [The Right to Information].

No privilege against disclosure of sources as a constitutional matter; what exists is statutory and partial.

No right to be free of general laws. The press is subject to the law of defamation, contempt, obscenity and the criminal law like everybody else, because it has the citizen's right and no more.

No right to a licence, a quota or a subsidy. But once the State chooses to allocate such things, the allocation is subject to article 14 and cannot be used to punish or reward content.

A worked example

A State grants advertising contracts to newspapers out of public funds, and announces that papers which publish material "prejudicial to the State's development narrative" will be removed from the empanelled list. A newspaper so removed challenges the decision.

Is there a right to government advertising? No. There is no constitutional right to a subsidy, a licence or a contract.

Does that dispose of the case? No, and this is the point. Once the State chooses to allocate a benefit, the allocation is State action subject to article 14 and cannot be exercised arbitrarily, on the R.D. Shetty principle in [Arbitrariness]. More importantly, a benefit withdrawn because of the content of what was published is a restriction on speech in effect, whatever its form. That is Bennett Coleman's effect test: the label is trade, the operation is on content.

Is it saved by article 19(2)? The ground would have to be one of the enumerated eight. "Prejudicial to the State's development narrative" is none of them. On Sakal Papers and Indian Express, the State cannot abridge the right in the interests of the general public, and it certainly cannot do so in the interests of its own reputation.

What of circulation? Withdrawal of advertising revenue reduces the paper's capacity to print and circulate. On Sakal Papers the volume of circulation is part of the right.

Answer. The withdrawal is bad. Not because the paper had a right to the advertising, but because the State used a discretionary benefit as a lever on content, which is a restriction that must be justified under article 19(2) and cannot be.

Quick revision

• The Constitution does not mention the press; the freedom is inside article 19(1)(a), and the press has neither more nor less than a citizen.
• The right is a citizens' right, which produced the standing problem in Bennett Coleman.
• Sakal Papers: the right extends to the VOLUME OF CIRCULATION, and article 19(2) does not permit abridgement in the interests of the general public.
• Indian Express: freedom of the press is freedom from interference with CONTENT AND CIRCULATION; the grounds in article 19(2) are exhaustive; a tax may be tested by its effect.
• Bennett Coleman: shareholders do not lose their rights by incorporating; and the EFFECT TEST, a measure operating on speech is a restriction whatever it is called.
• Not included: a right of access to information as such, a source privilege, immunity from general laws, or a right to a licence or subsidy. But allocation of a benefit is subject to article 14 and cannot be used against content.

Test yourself

1. Why is there no separate guarantee for the press, and what four consequences follow?

2. State the two propositions from Sakal Papers, and explain why the second defeats a benevolent object.

3. How did Bennett Coleman solve the problem that article 19 is a citizens' right?

4. What is the effect test, and how does it apply to a tax?

5. Name four things the freedom of the press does NOT include, and state the qualification on the fourth.

Decency or morality. The ground for obscenity law. The standard has shifted from the old test of a tendency to deprave and corrupt those whose minds are open to such influences, towards a community standards test read with contemporary values.

Contempt of court. Note the drafting: the words are "in relation to", not "in the interests of", for the last three grounds.

Defamation. Both civil and criminal.

Incitement to an offence. Added by the First Amendment 1951. Note that it is incitement to an offence, not to any unlawful act.

Reasonableness, and what it has come to mean

The word "reasonable" is the court's entry point, and the standard has developed in three stages.

Stage one, the classical factors. A court asks about the nature of the right, the purpose of the restriction, the extent and urgency of the evil sought to be remedied, the disproportion of the imposition, and the prevailing conditions. The restriction must be reasonable both substantively and procedurally, so an otherwise justifiable restriction imposed without a hearing or without any procedure may fail.

Stage two, direct and inevitable effect. A restriction is tested by what it actually does to the right, not by the form the legislature gave it. This is the effect test from Bennett Coleman.

Stage three, proportionality. The modern formulation, and it is now applied in terms.

Facts. Anuradha Bhasin v. Union of India, decided 10 January 2020. After the constitutional changes of August 2019 in Jammu and Kashmir, internet and telecom services were suspended and movement restricted under section 144 of the Code of Criminal Procedure. The executive editor of a newspaper said print media had come to a halt without the internet.

Held, among fifteen directions: all orders under section 144 and for suspension of telecom services must be published, so that affected persons can challenge them. The freedom of speech and expression and the freedom to practise any profession or carry on any trade or business over the medium of the internet enjoy protection under articles 19(1)(a) and 19(1)(g), and restrictions must satisfy article 19(2) and (6) including the test of proportionality. An order suspending internet services indefinitely is impermissible; suspension may be temporary only, must be proportionate, must not extend beyond the necessary duration, and is subject to judicial review, with review by the Review Committee within seven working days. On section 144: the power may be used on apprehension of danger, but the danger must be in the nature of an emergency; it cannot be used to suppress legitimate expression of opinion or grievance; the order must state material facts to enable judicial review; the Magistrate must balance rights and restrictions on proportionality and apply the least intrusive measure; and repetitive orders are an abuse of power.

Why it matters. It states proportionality as the operative test, adds the least intrusive measure, and turns publication of the order into a condition of its validity in practice, because an unpublished order cannot be challenged.

The four-part proportionality test

Worth setting out, because it is what "reasonable" now means.

Legitimate goal. The measure must pursue an aim falling within article 19(2).

Suitability, or rational connection. The measure must be capable of achieving that aim.

Necessity, or the least restrictive alternative. There must be no less intrusive measure that would achieve the aim equally well. This is what Anuradha Bhasin calls the least intrusive measure.

Balancing. The benefit to the aim must outweigh the harm to the right.

Two distinctions that decide cases

Public order against law and order. Three concentric circles: law and order the largest, public order within it, security of the State at the centre. A restriction justified as necessary for public order must show an effect on the community, not merely on individuals.

Advocacy against incitement. The eighth ground is incitement to an offence, not advocacy of an idea. Expressing a view, however unpopular or wrong, is protected; urging others to commit an offence is not. The line is proximity: the connection between the speech and the apprehended disorder must be direct and immediate, not remote or conjectural, and the standard image is that of a spark in a powder keg.

A worked example

A District Magistrate issues an order under section 144 of the Code of Criminal Procedure prohibiting the publication or circulation of any material relating to a communal incident, for a period of two months, on the ground that it may inflame feelings. The order is not published. A newspaper challenges it.

Requirement one, is there a law? Yes, section 144 is a statutory power, so the first requirement is satisfied. Had the District Magistrate acted by a press note with no statutory basis, the restriction would fail at once.

Requirement two, which ground? Public order, and possibly decency. The State must identify it, and must show that the apprehended effect is on the community and not merely on individuals.

Requirement three, is it reasonable, meaning proportionate? Legitimate goal, yes. Suitability, arguable. Necessity: a blanket prohibition on any material relating to the incident is not the least intrusive measure, since a prohibition on identifying communities, or on inflammatory characterisation, would achieve the aim with far less cost. Balancing: two months of silence on a matter of public importance is a heavy cost.

The Anuradha Bhasin points. The order must state material facts enabling judicial review; it must be published, and it was not; section 144 cannot be used to suppress legitimate expression of opinion or grievance, which reporting a communal incident is; the danger must be in the nature of an emergency, and a two-month prohibition is inconsistent with an emergency; and the Magistrate must apply the least intrusive measure.

Answer. The order fails on necessity, on the requirement to state material facts, on publication, and on duration. Note that it fails on the manner of the restriction rather than on the existence of the power.

Quick revision

• Three requirements: a law, one of the eight grounds, and reasonableness.
• The grounds: sovereignty and integrity of India; security of the State; friendly relations with foreign States; public order; decency or morality; contempt of court; defamation; incitement to an offence. The list is EXHAUSTIVE, and "in the interests of the general public" is not among them.
• Public order, friendly relations and incitement to an offence were added by the First Amendment 1951; sovereignty and integrity by the Sixteenth in 1963.
• Reasonableness has moved from the classical factors, through the direct and inevitable effect test, to PROPORTIONALITY in four parts: legitimate goal, suitability, necessity or least restrictive alternative, and balancing.
• Anuradha Bhasin: proportionality and the least intrusive measure; orders must be published and state material facts; indefinite suspension impermissible; repetitive section 144 orders are an abuse of power.
• Two distinctions: law and order against public order against security of the State; and advocacy against incitement, the test being proximity.

Test yourself

1. State the three requirements a restriction on speech must satisfy, and say which is most often overlooked.

2. List the eight grounds, and say which were added by amendment and when.

3. Why can a restriction not be justified in the interests of the general public? Give two authorities.

4. Set out the four parts of the proportionality test.

5. Explain the three concentric circles, and the difference between advocacy and incitement.

6. Give six of the directions in Anuradha Bhasin that bear on how a restriction must be imposed.

The circulation cases as prior restraint

The two great press cases are usually taught as circulation cases and they are also prior restraint cases, because both operated before publication.

Facts. Sakal Papers (P) Ltd v. Union of India, AIR 1962 SC 305, (1962) 3 SCR 842, decided 25 September 1961. An Act and Order fixed the pages and size a newspaper might publish in relation to its price, and regulated advertisement space, to prevent unfair competition and help smaller papers.

Held. Void as violating article 19(1)(a) and not saved by article 19(2). The freedom includes the freedom of the press; the right extends not merely to the matter a citizen may circulate but to the volume of circulation; and article 19(2) does not permit abridgement in the interests of the general public.

Facts. Bennett Coleman & Co v. Union of India, AIR 1973 SC 106, (1972) 2 SCC 788, decided 30 October 1972. The newsprint policy limited pages, prevented a common ownership unit from starting a new paper, and restricted interchange of quota. The Union called it rationing of a scarce commodity.

Held. By the majority, shareholders do not lose their fundamental rights by incorporating, and their locus standi is beyond challenge. On the merits the policy fell: a restriction on pages operates directly on circulation and content and cannot be defended as mere rationing.

Why the pair matters here. Neither measure censored a word. Both determined, before anything was written, how much could be published. That is prior restraint by resource control, and the answer to it is the effect test: what a measure does to speech decides its character, not what it is called.

Restraint on reporting court proceedings

Facts. Naresh Shridhar Mirajkar v. State of Maharashtra, AIR 1967 SC 1, (1966) 3 SCR 744, decided 3 March 1966 by nine judges. In a defamation suit a witness asked that his evidence not be published, saying his business would suffer. The trial judge orally prohibited publication of the evidence. Journalists, not parties to the suit, moved the Supreme Court under article 32, saying the High Court had no inherent power to make the order, that it violated article 19(1)(a), and that it was amenable to article 32.

Held. By the majority, the order was within the inherent power of the High Court: a court has inherent jurisdiction to hold a trial in camera where the ends of justice clearly and necessarily require it, and the power to prohibit excessive publication of part of the proceedings is a facet of that power. On the constitutional question, Sarkar J. held that if a judicial tribunal makes an order it has jurisdiction to make, applying a law valid in all respects, the order cannot offend a fundamental right; and that the law empowering a court to prohibit publication of its proceedings is in any event protected by article 19(2), relating as it does to contempt of court. Hidayatullah J. dissented, holding that a court which had heard a witness in open court could not suppress publication of that evidence on the witness's request that his business would suffer.

Why it matters. It is the authority for a court's power to restrain reporting, and it locates that power in the contempt ground of article 19(2). The modern practice is narrower than the case might suggest: such orders are treated as exceptional, must be for a limited period, must be the least restrictive means of protecting the administration of justice, and are recognised as postponement orders rather than permanent bans.

Restraint by executive order

Facts. Anuradha Bhasin v. Union of India, decided 10 January 2020. Internet and telecom services were suspended in Jammu and Kashmir and movement restricted under section 144 of the Code of Criminal Procedure; a newspaper editor said print media had halted.

Held, among fifteen directions: orders under section 144 and for suspension of telecom services must be published so they can be challenged; speech and trade over the internet are protected by articles 19(1)(a) and 19(1)(g) and restrictions must meet article 19(2) and (6) including proportionality; indefinite suspension is impermissible and suspension may be temporary only, not beyond the necessary duration, and is judicially reviewable; section 144 may be used on apprehension of danger but the danger must be in the nature of an emergency, cannot be used to suppress legitimate expression of opinion or grievance, must state material facts, must apply the least intrusive measure, and repetitive orders are an abuse of power.

Why it matters. It is the modern law on prior restraint by executive action, and its most useful contribution is procedural: an unpublished order is in practice unchallengeable, so requiring publication is what makes the substantive limits real.

A worked example

A State enacts a law requiring every periodical to submit each issue to a District Magistrate before publication, who may withhold permission if in his opinion the issue is likely to disturb public order. There is no appeal, no time limit for the decision, and no requirement of reasons.

Is there a law? Yes, so the first requirement is met.

Is the ground within article 19(2)? Public order is a named ground, so the second is met in form.

Is it reasonable, meaning proportionate? It fails on every limb.

Necessity. The State has a wide range of less intrusive measures: prosecution after publication, a prohibition confined to specified inflammatory matter, a short order in a genuine emergency under section 144. A standing permission requirement for every issue is the most intrusive measure available, not the least.

Procedural unreasonableness. No time limit means the Magistrate can suppress by delay; no reasons means the decision cannot be challenged; no appeal means there is no correction. A restriction must be reasonable procedurally as well as substantively.

Unguided discretion. "In his opinion likely to disturb public order" states no principle. On the Anwar Ali Sarkar reasoning in [Reasonable Classification], a power to pick who is treated differently with no principle to guide it is bad in itself.

The Anuradha Bhasin points. The order must state material facts and must be published; and the danger must be in the nature of an emergency, which a standing regime by definition is not.

Answer. Void. Note that it fails not because prior restraint is per se unconstitutional in India, but because this prior restraint is disproportionate, procedurally unreasonable and unguided.

Quick revision

• Prior restraint stops publication; subsequent punishment answers for it afterwards. Four reasons the first is worse: it suppresses the lawful, reverses the burden, leaves no record, and is administered by officials.
• India has no per se rule against prior restraint, unlike the United States. It is a restriction and must satisfy article 19(2) and proportionality.
• Sakal Papers and Bennett Coleman are prior restraint by resource control; the answer is the effect test.
• Mirajkar: a court has inherent power to restrain reporting, located in the contempt ground; a judicial order within jurisdiction cannot offend a fundamental right; Hidayatullah J. dissented. Modern practice treats such orders as exceptional and as postponement.
• Anuradha Bhasin: publication of the order, material facts, proportionality, least intrusive measure, no indefinite suspension, no repetitive section 144 orders.
• Pre-censorship of films is permitted; general licensing of the press is not.

Test yourself

1. Distinguish prior restraint from subsequent punishment, and give four reasons the first is treated more strictly.

2. How does the Indian position differ from the American, and what follows?

3. Why are Sakal Papers and Bennett Coleman prior restraint cases, and what test answers them?

4. Where did Mirajkar locate a court's power to restrain reporting, and how is that power exercised today?

5. Which of the Anuradha Bhasin directions are procedural, and why do the procedural ones matter most?

The old position was that it was not. The immunity in article 105(2) covers publication by or under the authority of a House, which is the official record, not a newspaper's report. A newspaper publishing a defamatory speech was therefore exposed, and a person defamed had a remedy against the paper and none against the speaker.

Article 361A answers this. Inserted by the Forty-fourth Amendment in 1978, it protects a person from any civil or criminal proceedings in respect of the publication in a newspaper of a substantially true report of proceedings of either House, subject to two conditions: the publication must not have been made with malice, and the protection does not extend to a report of a sitting held in secret. The article extends the protection to broadcasting by wireless telegraphy as well.

Publishing what the House wanted withheld, or reporting expunged remarks or the proceedings of a committee before it reports. Here article 361A does not help, because the report is not of proceedings the House has published, and the House may treat it as a breach of privilege.

Criticising the House or its members. The most contested. A House may treat a publication reflecting on it or its members as a contempt, and the width of that power is exactly what the absence of a defining law leaves uncertain.

The constitutional question

This belongs in an answer because it is the reason the topic is in the syllabus.

Do the privileges in articles 105(3) and 194(3) prevail over the fundamental rights?

The argument that they do. Article 105(3) is itself a provision of the Constitution, of equal status with article 19, and the privileges it preserves are conferred by the Constitution rather than by ordinary law, so they cannot be tested against Part III as a statute would be.

The argument that they do not. Part III is a limitation on all State power; a House is part of the State; and a power to punish a citizen for what he publishes is precisely what article 19(1)(a) exists to control. Article 194(3) preserving the privileges of a foreign legislature as they stood in 1950 cannot have been intended to create an enclave immune from the rights the same Constitution guarantees.

The position that has emerged is a compromise, and it should be stated as such. The privileges are real and a House may enforce them; but the exercise is not wholly beyond scrutiny, particularly where a House acts against a person who is not a member and where the process affects liberty. Where a House commits a person, the courts have entertained proceedings about the jurisdiction to commit though not about the merits of the House's assessment. The line is jurisdiction, not correctness, and the parallel with judicial orders in [Is the Judiciary "State"?] is exact.

The leading decisions here are in the case bank's queue and are not cited. Do not attribute holdings to unnamed cases.

Why no law has been made

Worth a paragraph, because it explains the state of the topic.

Articles 105(3) and 194(3) contemplate a defining law and none has been enacted, at the Union or in any State. The reasons usually given are that codification would make the privileges justiciable in a way the present position arguably avoids, and that a House is reluctant to fix limits on a power whose value lies partly in its indefiniteness.

The consequence is that a citizen cannot know in advance what will be treated as a contempt of a House, which is the standard objection to any undefined power to punish, and which would be fatal to an ordinary criminal provision under article 20(1) and article 21.

A worked example

A member makes a speech in a State Legislative Assembly accusing a named official of corruption. A newspaper reports the speech accurately the next day, under the headline "Official accused of corruption in Assembly", and adds an editorial calling the Assembly's procedure for such accusations unfair. The official sues the newspaper for defamation, and the Assembly issues a notice of breach of privilege to the editor over the editorial.

The defamation suit over the report. Article 361A protects a person from civil or criminal proceedings in respect of the publication in a newspaper of a substantially true report of the proceedings of a House. The report is accurate, so the protection applies unless the official proves malice, and unless the sitting was secret. Neither appears. The suit fails.

Would the member have been liable? No. Article 194(2) gives absolute immunity for anything said in the House, so the person defamed has no remedy against the speaker either. That is the price the Constitution pays for free debate, and it should be noticed rather than passed over.

The privilege notice over the editorial. Article 361A does not assist, because the editorial is not a report of proceedings; it is comment on the House. Whether criticism of the Assembly's procedure is a contempt depends on privileges that have never been defined, which is the difficulty this chapter describes.

What the editor can argue. That article 19(1)(a) protects criticism of a public institution; that a restriction must fall within article 19(2), and contempt of a legislature is not among the eight grounds, which name only contempt of court; and that the House's jurisdiction to punish a non-member for comment made outside the House is at the least open to examination.

What the House can argue. That article 194(3) is itself a constitutional provision preserving the privileges as they stood in 1950, and that those privileges included the power to punish for contempt committed outside the House.

The honest answer. The report is protected and the editorial is in contested territory, and the contest is unresolved because the defining law article 194(3) contemplates has never been made.

Quick revision

• Articles 105 and 194 confer two distinct things: freedom of speech in the House, absolute for the member, and the powers, privileges and immunities of the House, including the power to punish for contempt.
• Clauses (3) leave the privileges to be defined by law; no such law has been made, so they remain those the House had in 1950.
• Article 361A, Forty-fourth Amendment 1978: no civil or criminal liability for publishing in a newspaper a SUBSTANTIALLY TRUE report of proceedings, unless made with MALICE, and not for a SECRET sitting.
• Article 361A protects reports, not comment, and not the publication of what the House withheld.
• The constitutional question is whether the privileges prevail over Part III; the compromise reached is that a House may enforce them but its jurisdiction to do so is examinable, though not the merits.
• Note that article 19(2) names contempt of COURT and not of a legislature.

Test yourself

1. Distinguish the two things articles 105 and 194 each confer, and say which is absolute.

2. Why do the privileges remain those of 1950, and what is the standard objection to that?

3. What does article 361A protect, and what are its two conditions and one exclusion?

4. Give the argument each way on whether the privileges prevail over Part III, and state the compromise reached.

5. Why does it matter that article 19(2) names contempt of court but not of a legislature?

Who is covered. A public authority is a body constituted by or under the Constitution, by law, or by a government notification, and includes a body owned, controlled or substantially financed by government, and a non-governmental organisation substantially financed by government. Note that this definition is wider than article 12, which matters after the narrowing described in [The Retreat, and What Survived It].

The Information Commissions. A Central Information Commission and a State Commission for each State, with power to require disclosure and to impose penalties on a defaulting officer.

The exemptions, in section 8, cover national security and sovereignty, matters forbidden by a court, breach of privilege of a legislature, commercial confidence and trade secrets, fiduciary relationships, information received in confidence from a foreign government, information endangering life or safety or identifying a source, information impeding investigation or prosecution, Cabinet papers subject to disclosure after a decision, and personal information.

Section 8(2), which is the important one, provides that a public authority may allow access even to information exempt under the Official Secrets Act 1923 or under the exemptions in section 8(1), if the public interest in disclosure outweighs the harm to the protected interests. That is the general public-interest override.

The two changes that alter the answer

The 2019 amendment. The Right to Information (Amendment) Act 2019 removed from the Act itself the fixed term and the salary and conditions of service of the Chief Information Commissioner and the Information Commissioners, at the Centre and in the States, and made them such as may be prescribed by the Central Government. The objection is one of independence: a Commission whose tenure and pay are fixed by the executive it exists to compel is structurally weaker, and the comparison drawn is with the protections article 324 gives the Election Commission, on which see [The Election Commission].

The rewriting of section 8(1)(j), in force 13 November 2025. This is the change most likely to be missed.

The old provision exempted personal information the disclosure of which has no relationship to any public activity or interest, or which would cause unwarranted invasion of the privacy of the individual, unless the Public Information Officer or the appellate authority is satisfied that the larger public interest justifies disclosure. It also carried a proviso that information which cannot be denied to Parliament or a State legislature shall not be denied to any person.

The new provision, substituted by section 44(3) of the Digital Personal Data Protection Act 2023 with effect from 13 November 2025, exempts information which relates to personal information, without the qualifying words about relationship to public activity, without the unwarranted-invasion test, and without the larger-public-interest override built into the clause.

Why it matters. Requests for the assets, qualifications, service records and decisions of public officials were routinely met by weighing the privacy interest against the public interest inside section 8(1)(j) itself. On the new text that weighing is not written into the clause. The counter-argument is that section 8(2) survives, so a public-interest override remains available for the whole of section 8; and that after Justice K.S. Puttaswamy (Retd) v. Union of India, decided 24 August 2017 by nine judges unanimously, in which the right to privacy was held protected as an intrinsic part of article 21 and as part of the freedoms guaranteed by Part III, a blanket exemption for anything called personal information would itself have to satisfy proportionality, since the right it restricts is also a Part III right.

How to write it. State the change, state its date, state the section 8(2) answer, and say that the balance between the right to information and the right to privacy is now the live question in this topic. That is accurate and it is current.

A worked example

A journalist asks a Ministry for the file notings recording why a particular officer was transferred, and for the officer's declared assets. The Ministry refuses, citing section 8(1)(j) as it now stands, and says that after the 2025 amendment personal information is simply exempt.

Is the file noting personal information? Largely not. A record of the reasons for an administrative decision is information about the exercise of public power. That some of it names an individual does not convert it into personal information, and reading the clause otherwise would exempt almost every record.

Are the declared assets personal information? Yes, on any reading, and under the old clause the answer would have turned on whether the larger public interest justified disclosure, which for a public official's declared assets it frequently did.

Does the amendment end the enquiry? No, for two reasons, and both must be given. Section 8(2) permits a public authority to allow access notwithstanding section 8(1) where the public interest in disclosure outweighs the harm to the protected interests, and it was not amended. And the constitutional derivation survives: the right is part of article 19(1)(a), so a refusal is a restriction on speech that must be justified within article 19(2), and "personal information" is not one of the eight grounds.

What of the officer's privacy? Genuine, and after Puttaswamy it is a Part III right, not merely a statutory exemption. So this is a conflict of two constitutional rights, and the answer is proportionality: disclose what the public interest requires and no more, redacting what it does not.

Answer. The notings should be disclosed; the assets require the section 8(2) balance; and a blanket refusal resting on the amended clause alone is bad.

Quick revision

• No express right; read into article 19(1)(a) because expression without information is empty and the citizen is entitled to know what is done in his name.
• Because it sits in article 19(1)(a), restrictions must satisfy article 19(2), which is narrower than a statutory exemption clause. Indian Express: the grounds are exhaustive and public interest is not among them.
• The Act of 2005: suo motu disclosure under section 4, no reasons required for a request, thirty days or forty-eight hours, two appeals, Information Commissions with penalty powers, and a definition of public authority WIDER than article 12.
• Section 8 exemptions, and section 8(2) the general public-interest override, which survives.
• The 2019 amendment moved the Commissioners' tenure and salary to executive prescription: an independence objection.
• Section 8(1)(j) was REWRITTEN with effect from 13 November 2025 by section 44(3) of the DPDP Act 2023, and the new text exempts personal information without the old qualifiers or the internal public-interest override.
• Puttaswamy makes privacy a Part III right, so this is now a conflict of two rights resolved by proportionality.

Test yourself

1. Set out the three steps by which the right to information was derived from article 19(1)(a).

2. Why does the constitutional derivation still matter now that there is a statute?

3. Give five features of the 2005 Act, including one respect in which its coverage is wider than article 12.

4. What did the 2019 amendment change, and what is the objection to it?

5. State the old and the new section 8(1)(j), the date and instrument of the change, and the two answers to a blanket refusal under it.

Why the case matters beyond section 66A. It establishes that the internet is not a lesser medium attracting a lower standard, and it puts vagueness and overbreadth firmly into Indian article 19 doctrine, where they had been used but not systematically.

Lever two: the intermediary

The distinctive feature of online speech is that it is carried by somebody: a platform, a host, a network. The State can therefore regulate the carrier instead of the speaker, which is far more efficient and far less visible.

Safe harbour. Section 79 of the Information Technology Act exempts an intermediary from liability for third-party information it hosts, provided its function is limited to providing access, it does not initiate or select or modify the transmission, and it observes due diligence and any prescribed guidelines. The exemption is lost if the intermediary conspires or abets, or if it fails to remove material on receiving actual knowledge of unlawfulness.

The problem safe harbour creates. If an intermediary loses immunity on receiving a private complaint, its rational course is to remove anything complained of without examining the merits. Regulation of the carrier therefore produces private censorship at the complainant's request, with no adjudication and no remedy for the speaker.

The answer given. In Shreya Singhal the provision was read down so that actual knowledge means knowledge received through a court order or a notification by the appropriate government, not a private complaint. That converts a private takedown regime into a public one.

Blocking. Section 69A permits the Central Government to direct blocking of public access in the interests of the sovereignty and integrity of India, defence, security of the State, friendly relations with foreign States, public order, or for preventing incitement to a cognisable offence relating to those. The grounds track article 19(2), which is why the power survived, and the rules prescribe a committee, an opportunity to be heard and reasons. The standing criticism is confidentiality: where the direction and its reasons are not disclosed, the safeguards cannot be tested, which is the same objection Anuradha Bhasin answered for section 144 by requiring publication.

Lever three: suspending the medium

Facts. Anuradha Bhasin v. Union of India, decided 10 January 2020. After the constitutional changes of August 2019 in Jammu and Kashmir, mobile networks, internet and landline services were discontinued and movement restricted under section 144 of the Code of Criminal Procedure. The executive editor of a newspaper said print media had come to a halt without the internet.

Held, in fifteen directions. Orders under section 144 and for suspension of telecom services must be published, to enable challenge. Speech and expression, and the practice of a profession or trade, over the medium of the internet are protected by articles 19(1)(a) and 19(1)(g), and restrictions must satisfy article 19(2) and (6) including proportionality. An order suspending internet services indefinitely is impermissible under the Temporary Suspension of Telecom Services Rules 2017; suspension may be for a temporary duration only, must be proportionate, must not extend beyond the necessary duration, and is subject to judicial review. Because the Rules provide neither periodic review nor a time limit, the Review Committee must review within seven working days of the previous review. All existing suspension orders were to be reviewed forthwith and those not conforming revoked. On section 144, the danger must be in the nature of an emergency, the power cannot be used to suppress legitimate expression of opinion or grievance, the order must state material facts, the Magistrate must apply the least intrusive measure, and repetitive orders are an abuse of power.

Why it matters. It is the leading authority on shutdowns. Note what it did not do: it did not hold that access to the internet is itself a fundamental right. It held that the rights exercised through the internet are protected, which is a narrower and more defensible proposition, and it is the one to state.

What is genuinely new, and what is not

MU's label asks about new scientific development, so the answer should separate the two.

Not new. The grounds of restriction, the requirement of a law, proportionality, the advocacy and incitement distinction, and the rule that circulation is part of the right. Every one of these came from the print cases and applies unchanged.

New. That speech is carried by a private intermediary whose incentives are not the speaker's, so the State can regulate speech by regulating a company. That the medium can be switched off wholesale, which no government could do to newspapers. That speech is instantaneous, borderless and permanent, so the harm is done before any remedy operates. And that scale changes the calculus: a rule that is proportionate applied to a publisher may be disproportionate applied to millions of individual users.

A worked example

A State Government, apprehending disorder after a court verdict, directs telecom operators to suspend mobile internet in four districts for one week, and separately writes to a social media platform asking it to remove twenty specified posts as offensive. The order to the operators is not published. The platform removes the posts. A user whose post was removed, and a small business unable to trade, both complain.

The suspension. Test it against Anuradha Bhasin. Is it published? No, and that alone is contrary to the direction, because an unpublished order cannot be challenged. Is it temporary? A week is not indefinite, so it clears that hurdle. Is it proportionate and the least intrusive measure? Suspending mobile internet across four districts to prevent disorder that has not occurred requires justification, and narrower measures, suspending in a specific locality, or restricting specified services, must be shown to be inadequate. Is there periodic review within seven working days? For a one-week order the requirement bites almost immediately.

The business. Its complaint is under article 19(1)(g), which Anuradha Bhasin expressly protects over the medium of the internet, and it is a separate right requiring separate justification under article 19(6).

The takedown. The State did not order removal; it asked. On the reading down of section 79 in Shreya Singhal, actual knowledge means a court order or a government notification, so a request that is not a lawful direction does not put the platform at risk, and its removal was voluntary. That is the private censorship problem: the user has lost his speech without any order he can challenge.

What the user can do. If the communication was in substance a direction, he can attack it as a restriction imposed without a law and without the section 69A procedure. If it was genuinely a request, his remedy against the platform depends on whether it is amenable to article 226 as discharging a public function, which is [Private Bodies and Public Functions], and the answer is not obvious.

The lesson. The suspension is the easier challenge because it is State action with a settled framework. The takedown is the harder one because the State has arranged for somebody else to do it.

Quick revision

• Article 19(1)(a) is medium-neutral. Anuradha Bhasin declares speech and trade over the internet protected by articles 19(1)(a) and 19(1)(g), subject to 19(2) and (6) including proportionality.
• Three levers: the speaker, the intermediary, the medium.
• Shreya Singhal struck down section 66A for vagueness and overbreadth and for failing to distinguish discussion and advocacy from incitement; read down section 79 so that actual knowledge means a court order or government notification; struck down the State blocking power and upheld section 69A.
• Section 69A survives because its grounds track article 19(2); the criticism is confidentiality of the directions.
• Anuradha Bhasin: publication, proportionality, no indefinite suspension, temporary only, review within seven working days, least intrusive measure, no repetitive section 144 orders.
• Anuradha Bhasin did NOT hold that access to the internet is a fundamental right; it held that the rights exercised through it are protected.
• What is new is the intermediary, the switch, the speed and the scale, not the doctrine.

Test yourself

1. On what three grounds was section 66A struck down, and why does the case matter beyond that section?

2. Explain safe harbour, the private censorship problem it creates, and how Shreya Singhal answered it.

3. Why did section 69A survive when the State blocking power did not?

4. Give six of the directions in Anuradha Bhasin bearing on internet suspension.

5. What did Anuradha Bhasin NOT decide, and why is the narrower proposition the better one?

6. Separate what is genuinely new about online speech from what the print cases already settled.

Legitimate aim. The law must pursue an object the Constitution permits.

Proportionality. The means must be rationally connected to the aim, must be the least restrictive alternative, and must strike a balance in which the benefit outweighs the harm. This is the same four-part test set out in [The Grounds of Reasonable Restriction], and Anuradha Bhasin applies it in terms to restrictions imposed over the internet.

Procedural safeguards. Commonly added as a fourth: the law must provide against abuse, by review, oversight or a remedy.

What privacy covers

The judgments identify overlapping zones, and giving them is better than treating privacy as a single undifferentiated thing.

Bodily and physical privacy. Integrity of the body, freedom from unwanted search and physical intrusion.

Informational privacy. Control over personal information: what is collected, by whom, for what purpose, and to whom it is disclosed. This is the zone that data protection legislation addresses.

Decisional privacy or autonomy. The freedom to make intimate choices about family, marriage, procreation, and the conduct of one's life.

Privacy of communications. Interception and surveillance.

Spatial privacy. The home, and places where a person reasonably expects not to be observed.

Surveillance

The constitutional position. Interception of communications is an invasion of privacy and must satisfy legality, legitimate aim and proportionality. Statutory power exists, principally under section 5(2) of the Indian Telegraph Act 1885 and section 69 of the Information Technology Act 2000, each with rules prescribing who may authorise, on what grounds, for how long, and with review.

The recurring difficulties, and an answer should name them rather than describe the machinery.

Secrecy defeats the safeguards. An interception the subject never learns of cannot be challenged by him, so the safeguards depend entirely on internal review by the executive. The same objection Anuradha Bhasin met for section 144 by requiring publication cannot be met the same way here, because publication would defeat the purpose.

Executive authorisation. The authorising officer is part of the executive that seeks the interception, not an independent authority. Comparative systems commonly require judicial authorisation, and the absence of it in India is the standard reform proposal.

Mass collection. The old law was framed for the interception of a particular line. Bulk collection and metadata analysis are different in kind, and the proportionality of collecting everything in order to search it later is a question the framework was not designed to answer.

Chilling effect. Surveillance of communications restricts speech even where nothing is prosecuted, because a person who believes he is watched speaks differently. That is why this belongs in the press topic: the interference is with article 19(1)(a) as much as with article 21.

Data protection

The Digital Personal Data Protection Act 2023 is the legislative answer to the informational privacy limb of Puttaswamy. Its scheme rests on consent, notice, purpose limitation, and duties on the entity that determines the purpose and means of processing, with a Data Protection Board to adjudicate and penalties for breach.

Two features that matter constitutionally and are the standard criticisms.

Exemptions for the State. The Act permits the Central Government to exempt instrumentalities of the State from its provisions in specified interests. An exemption power exercisable by the very actor whose processing is most consequential is the point at which a data protection statute stops constraining the principal risk.

The Board's independence. Composition, appointment and terms are largely determined by the Central Government, which raises the same structural objection made about the Information Commissions in [The Right to Information].

And the change to the Right to Information Act. Section 44(3) of the same Act substituted section 8(1)(j) of the Right to Information Act 2005 with effect from 13 November 2025, exempting information which relates to personal information, without the earlier qualifiers about relationship to public activity, unwarranted invasion, or the larger public interest. That is the sharpest current illustration of privacy and the right to information pulling against each other, and it is treated fully in [The Right to Information].

A worked example

A State sets up a system of facial recognition cameras in public places, matched against a database of persons of interest, under an executive order of the Home Department. There is no statute. The order is not published. A civil liberties organisation challenges it, and a journalist adds that reporters covering protests are being identified and later questioned.

Is privacy engaged in a public place? Yes. Spatial privacy is not confined to the home; the question is whether there is a reasonable expectation, and being recorded and identified by an automated system is different in kind from being seen by passers-by.

Requirement one, legality. There is no law, only an executive order. That is fatal, and it is the shortest route to the answer. After Puttaswamy an invasion of privacy requires a law, and an executive order is not one.

Requirement two, legitimate aim. Detection of crime is legitimate, so the challenge does not fail here.

Requirement three, proportionality. Rational connection is arguable. Necessity is the weak point: identifying every person in a public space in order to find a few is the opposite of the least restrictive alternative. Balancing weighs continuous identification of the whole population against a benefit that is not quantified.

Requirement four, safeguards. No published order, no retention limit, no oversight, no remedy.

The journalist's separate point. Identification of reporters at protests, followed by questioning, is a chilling effect on article 19(1)(a), and it is a distinct claim. Puttaswamy's location of privacy in the whole of Part III is what makes it available: the interference is not only with article 21.

Answer. The system fails at the first requirement for want of a law, and would fail on necessity and safeguards even if a law existed.

Quick revision

• Puttaswamy, nine judges, 24 August 2017, unanimous: M P Sharma overruled; Kharak Singh overruled to the extent it denied privacy; privacy protected as an intrinsic part of article 21 AND as part of the freedoms guaranteed by Part III.
• The location in the whole of Part III is what lets privacy inform article 19.
• Restriction requires legality, a legitimate aim, and proportionality; safeguards are commonly added as a fourth.
• Zones: bodily, informational, decisional, communications, spatial.
• Surveillance: statutory power under section 5(2) of the Telegraph Act and section 69 of the IT Act; the difficulties are secrecy, executive authorisation, mass collection, and chilling effect.
• The DPDP Act 2023 answers informational privacy; the criticisms are State exemptions and the Board's independence.
• Section 44(3) of that Act rewrote section 8(1)(j) of the RTI Act with effect from 13 November 2025.

Test yourself

1. State the four terms of the order in Puttaswamy, and explain why the wording of the third matters.

2. Set out the three requirements for a restriction on privacy, and the fourth that is commonly added.

3. Name five zones of privacy and give an example of each.

4. Give four difficulties with the Indian surveillance framework, and say why the Anuradha Bhasin answer of publication cannot be used here.

5. What are the two standard constitutional criticisms of the DPDP Act 2023?

6. Why does a chapter on privacy belong in a topic on the freedom of the press?

Why the distinction is under pressure online. The proximity test was developed for a speaker addressing a crowd, where the connection between words and consequence is observable. Speech circulated to millions, resurfacing months later, reaching an audience the speaker cannot see, does not fit the model. That is a genuine problem and an answer should say so rather than pretend the old test transfers cleanly.

The public order requirement

Because most hate speech prosecutions rest on the public order ground, the limits of that ground do the work.

The three concentric circles. Law and order is the widest; public order a smaller circle within it; security of the State the smallest. An act that disturbs an individual or a few people disturbs law and order; to reach public order it must affect the community at large or the even tempo of the life of the community.

Consequence, not content. The ground is public order, so the State must show a connection between the speech and a disturbance of public order, and the connection must be proximate. The mere fact that a statement is about a community, or is intemperate, does not establish it.

Facts. Sakal Papers (P) Ltd v. Union of India, AIR 1962 SC 305, (1962) 3 SCR 842, decided 25 September 1961. An Act and Order fixed the pages and price of newspapers and the space for advertisements, to prevent unfair competition and help smaller papers.

Held. Void as violating article 19(1)(a) and not saved by article 19(2). The right extends to the volume of circulation, and article 19(2) does not permit the State to abridge the right in the interests of the general public.

Why it is cited here. It supplies the discipline that decides hate speech cases at the threshold: a restriction must be brought within one of the eight grounds. A measure defended as promoting harmony, social cohesion or communal amity, without more, is defended on a ground that does not exist in article 19(2).

The heckler's veto

Worth a paragraph because it recurs and because the answer is settled in principle.

Where speech provokes a hostile audience, the State sometimes prevents the speech rather than controlling the audience, on the footing that public order is threatened. That is the heckler's veto, and it inverts the guarantee: it makes the right to speak depend on the willingness of opponents to tolerate it, so the most objectionable listeners decide what may be said.

The principled answer is that the State's duty is to protect the speaker and maintain order, and that suppression of the speech is permissible only where the disturbance is imminent and cannot be prevented by any other means. That is the least intrusive measure requirement from Anuradha Bhasin v. Union of India, decided 10 January 2020, in which the Court directed that a Magistrate acting under section 144 must balance rights and restrictions on the principles of proportionality and thereafter apply the least intrusive measure, that the power cannot be used to suppress legitimate expression of opinion or grievance, that the order must state material facts, and that repetitive orders are an abuse of power.

Fake news

The problem. False information circulated at scale can cause real harm: panic, violence, damage to health, distortion of elections.

The constitutional difficulty. Falsity is not a ground in article 19(2). There is no ground of untruth. So a restriction on false statements must be brought under public order, decency, defamation, or incitement to an offence, and must satisfy proportionality.

Why a general power to determine truth is dangerous. A body empowered to declare what is false, and to require its removal, is a censor whose jurisdiction is defined by its own conclusions. The objection is not that falsehoods are harmless but that the remedy places the determination of truth in the hands of an authority that is frequently an interested party, particularly where the statement concerns the government. A fact-checking power vested in the executive, with intermediaries obliged to act on its findings, is the sharpest current example, and it has been litigated on exactly this ground.

The framework that survives scrutiny. A restriction confined to a defined harm, imposed by law, with an independent determination, a hearing, reasons and a remedy, and limited to what is necessary. Anything wider fails on the second or third requirement.

Broadcasting

Why it was treated differently. Broadcasting was regulated more heavily than print for two reasons: spectrum is a finite public resource requiring allocation, and broadcast reaches an undifferentiated audience including those who have not chosen it.

The constitutional position. Airwaves are public property and must be controlled and regulated by a public authority in the interests of the public, to prevent monopoly and to secure a diversity of views. That proposition, from the litigation about the telecast of sporting events, is the foundation of Indian broadcasting law, and it cuts both ways: it justifies regulation, and it forbids a government monopoly, because the reason for regulation is diversity and a State monopoly defeats it.

Where it stands now. Spectrum scarcity is a diminishing rationale as distribution moves to the internet, and the case for treating broadcast under a lower standard than print weakens with it. The current regulatory questions are about the classification of streaming services, the extent to which self-regulation satisfies the constitutional requirements, and whether rules made under the Information Technology Act can validly govern content that is not an intermediary function at all.