Where We Left Off
This part ends on 25 June 1975 and does not cross into it. What follows belongs to Part Nine, and the point of stopping here is that by the last page everything Part Nine describes had already been built.
Part Three found the internal check disappearing between 1950 and 1954. This part is what a system without one does when somebody uses it deliberately.
Five findings carry into this part, and one of them is a debt this part pays.
Part Two owes a verdict on the privy purses. They were guaranteed in the constitution as the price of accession, and abolished twenty-four years later. Part Two recorded the promise and said Part Eight would say whether the withdrawal was justified. Chapter Four is that verdict, and the answer separates the substance from the method.
By 1954 there was nobody who could tell the Prime Minister he was wrong. Part Three followed it: Patel died, Tandon was forced out, Ambedkar resigned, Kripalani left, and nothing institutional replaced them. Chapter One argues that this, and not anybody’s judgment of character, is what the Syndicate got wrong in 1966.
When a court blocks you, amend the constitution. Part Three’s Chapter One found that the First Amendment’s durable product was not a rule about speech but a method. Chapters Four and Six are that method used at full extension, and the second time the target is the court itself.
Congress never won half the vote. First-past-the-post converted a plurality into the two-thirds needed to amend. Part Three said the seat share, not the mandate, is what made constitutional amendment easy. This part is the decade in which that arithmetic was used hardest.
The state could plan when it wanted to. Part Seven’s Chapter Five described eight months of preparation before December 1971 and called it the clearest evidence in the series that the Indian state could learn. The same government appears in this part. Competence and restraint are different things and this part is about the second.
One note on where this part stops. It ends on 25 June 1975 and goes no further. The Emergency is Part Nine’s subject and it would be easy to run the two together. Keeping them apart is the point: by the last page of this part, every instrument Part Nine describes already existed, and most of them had already been used.
How to read this book
The Six Boxes
Six coloured boxes run through the series, each doing one job. Here they are with examples from the material ahead.
A Word Box explains a hard word the moment it first appears.
The Syndicate: the informal name for a group of powerful Congress state bosses who between them controlled the party organisation in the mid-1960s.
They were not a faction with a programme. They were the men who could deliver blocks of votes in the party’s internal elections — a Chief Minister here, a state party president there — and they had chosen Prime Ministers before.
Why it matters here: they picked Indira Gandhi in 1966 believing she would be manageable, and Chapter One is about what they had actually misunderstood, which was not her.
An In Real Terms box turns a number too big to picture into something with a body.
The bill to abolish the privy purses passed the Lok Sabha in September 1970 and failed in the Rajya Sabha by one vote — 149 for and 75 against, short of the two-thirds required.
One vote. In a chamber the size of a large school hall.
The following day the President signed an order doing it anyway. That is the whole of Chapter Four in two sentences, and the twenty-four hours between them is the most instructive gap in this part.
A How We Actually Know This box shows the physical evidence and then says what it cannot prove.
This part rests on the best-evidenced material in the whole series, and it is worth saying why before anything else.
Almost every decisive event here happened on the record and in public: a vote in the Rajya Sabha with a published count, a Presidential order in the gazette, an eleven-judge judgment, a thirteen-judge judgment, an appointment announced by the government, three resignations, and a High Court judgment delivered in open court at ten in the morning.
There is very little in Chapters One to Eight that requires anybody’s memoir, and where a memoir is used it is marked.
What it cannot show: intention. That the government appointed a judge who had ruled in its favour is a fact. That it did so because he had is an inference, and Chapter Six treats it as one.
An Argument box appears where serious people disagree, each side at its strongest.
A prosecution brief about a leader who won two enormous election victories has to deal with this first.
Bank nationalisation was one of the most popular acts of any Indian government. Abolishing hereditary payments to former princes was popular. Garibi Hatao produced a landslide in 1971 and the victory over Pakistan produced another wave after it. Every measure this part criticises was cheered by very large numbers of Indians at the time, and a democracy in which a leader does what her voters want and is then condemned for it by writers fifty years later has an odd idea of legitimacy.
Nobody in this part is arguing that the purses should have been kept or the banks left alone. Both objectives are defensible and one is admirable. What is at issue is the machinery: a parliamentary defeat answered by executive order the next day, a court judgment answered by constitutional amendment, and three judges passed over after ruling against the government. None of those was on any ballot, none was popular in itself, and most voters did not know they had happened.
What would settle it: nothing about popularity settles a question about method, which is why constitutions exist at all.
Why people care so much: because the defence of this period in Indian politics is almost entirely a defence of its aims, and the prosecution is almost entirely about its means, and the two arguments have never once met.
A Hidden Assumption box digs out a belief sitting underneath an argument that both sides accept without discussing.
Everybody who writes about these nine years — admirers and critics — assumes that they are a story about one person.
The admiring version: a leader of exceptional will took on entrenched interests and won. The hostile version: a leader of exceptional ruthlessness dismantled Indian democracy. Both are biographies, and both explain the period by reference to a character.
Ask a different question. What would have stopped her?
Work through the list. The party organisation, which Part Three showed had lost its capacity to remove a leader by 1954. Parliament, where first-past-the-post had delivered a two-thirds majority on a minority of votes since 1952. The courts, which Part Three’s Chapter One showed could be overridden by amendment within a year of any decision. The federal structure, which Part Three’s Chapter Seven showed collapsing the first time it was tested in 1959. The press, which had no protection the First Amendment had not already qualified.
Every one of those was already weak, and every one of them was weakened by somebody else, earlier, mostly for good reasons and with wide approval.
Which means the biography explains the timing and not the outcome. A person with unusual determination arrived in a system with no working brakes, and the interesting question is not why she pressed the accelerator but why nobody had noticed for twenty years that there was nothing else fitted.
This is uncomfortable for both camps. It denies the admirers a hero acting against the odds, because the odds were not against her. It denies the critics a villain who broke something, because most of it was already broken. And it makes 1975 a fact about a design rather than about a family, which is a much harder thing to campaign on.
And every chapter closes with a Remember This box: the chapter in the plainest words available.
Part Two’s debt on the privy purses is paid in Chapter Four. Part Three’s finding — that by 1954 nobody could tell the Prime Minister he was wrong — is what Chapter One argues the Syndicate actually misunderstood.
Almost everything decisive here happened on the record and in public: a published vote, a gazetted order, two enormous judgments, three resignations, and a judgment delivered at ten in the morning.
The right question is not why she pressed the accelerator. It is why nobody had noticed for twenty years that nothing else was fitted.
1The Choice of 1966
A group of party bosses chose a leader they believed they could manage. They were wrong about her, and they were wrong about something much larger that nobody has ever held them to account for.
1.1 — The vacancy
Lal Bahadur Shastri died in Tashkent in January 1966, hours after signing the agreement Part Seven describes. He had been Prime Minister for nineteen months.
The succession fell to be decided by the Congress parliamentary party, which in practice meant it fell to be decided by the men who could deliver its votes — the group known as the Syndicate.
Their candidate problem was straightforward. Morarji Desai wanted the job and had wanted it in 1964. He was able, senior, and entirely unmanageable. What the Syndicate wanted was somebody who would leave the party organisation where it was.
1.2 — What they thought they were buying
They chose Indira Gandhi, and the reasoning is not hard to reconstruct.
She had a name that could win votes in every state, which mattered enormously to men whose own authority was regional. She had held office as Information and Broadcasting Minister without making enemies. She had no independent base in any state party, which meant she would need the Syndicate to survive. And she was widely thought to lack the appetite for a fight.
A socialist member of parliament called her the goongi gudiya — the dumb doll. The phrase stuck because a great many people agreed with it.
She defeated Desai in the parliamentary party vote and became Prime Minister.
1.3 — What they had actually misread
Every account of what followed treats the Syndicate’s error as an error about her character. They thought she was pliable; she was not; therefore they miscalculated.
That is true and it is the smaller half of the mistake.
The larger half is that the Syndicate believed it could remove a Prime Minister it had installed. That belief was the entire basis of the arrangement, and it had already stopped being true.
Part Three set out how. Between 1950 and 1954 every figure with an independent base who could contradict the leadership had died, resigned or been defeated, and nothing institutional replaced them. The party organisation still existed — it held meetings, elected presidents, passed resolutions. What it had lost was the capacity to act against a sitting Prime Minister who declined to go, because the only time it had tried, in 1951, the Prime Minister had won and taken the party presidency as well.
So the Syndicate in 1966 was holding an instrument that had not worked for fifteen years, and it had never tested it because there had never been an occasion.
Chapter Two is the occasion.
Parliamentary party: all the members of parliament belonging to one party, sitting together as a body. In the Westminster system it is the group that chooses, and can remove, the leader.
Its power depends on a simple condition: that MPs owe their seats to something other than the leader. If they were elected on the leader’s name, on her ticket, and expect to need both again in five years, the theoretical power to remove her exists on paper and nowhere else.
Why it matters here: in 1966 Congress MPs mostly owed their seats to the state bosses. By 1972 they owed them to her, and Chapter Two is how that changed.
1.4 — 1967, and what it did to the men who chose her
One event between the choice and the split is usually treated as background and is in fact the hinge.
The general election of 1967 went badly. Congress held the centre with a much reduced majority, and lost power in a large number of states — for the first time since independence, opposition and coalition governments took office across much of northern and eastern India.
Consider what that did to the Syndicate specifically. Their authority was not national and did not come from Delhi. It came from controlling state party machinery and state governments. A Chief Minister who is no longer Chief Minister is a man with a reputation and no instrument.
So between January 1966 and the spring of 1967, the group that had installed a Prime Minister on the understanding that they could manage her lost a large part of the base from which they would have had to do it.
She, meanwhile, had spent the campaign travelling the country under a name that still worked, and had emerged with something none of them had: a claim to be the only person who could hold the party’s vote together.
Chapter One’s argument was that the check was already unusable in 1966. 1967 is when it became visibly so, and the people it belonged to were the last to notice.
The nature of that mistake is easier to see in something more ordinary than a party.
Think of a fire door that has been in a building for fifteen years and has never been opened.
It is on the plan. It is painted the right colour. Everybody walking past it every day assumes that if there were a fire, that door would open.
Nobody has pushed it since 1951, when it stuck. The paint has since been applied over the frame twice.
The building is not less safe than it was yesterday. It is exactly as unsafe as it has been for fifteen years, and everybody in it is behaving as though it were safe, which is the actual danger.
Which is why the standard account of 1966 gets the mistake in the wrong place.
Everybody who tells this story assumes that she was underestimated.
The admiring version: they thought she was a doll and she turned out to be formidable. The hostile version: they thought she was harmless and she turned out to be dangerous. Both are stories about a misjudgment of character, and both are flattering to her.
Suppose they had judged her character perfectly. Suppose they had chosen exactly the pliable person they wanted.
The arrangement still fails, because it depends on the party being able to remove a Prime Minister, and the party could not remove a Prime Minister. Any leader who discovered that — and a leader in that office discovers it within a year or two — would have been in the same position. The Syndicate was not relying on her deference. It was relying on a leash it had not looked at since 1951.
This matters beyond one succession, and here is the general form. A check that has not been tested is not known to work. An organisation can carry an unusable safeguard on its books for decades, and everybody continues to behave as though it functions, because behaving as though it functions is what people do when they have never seen it fail.
The pattern runs the length of this series. Part Three’s institutions looked intact in 1954 because nobody had needed them. Part Six’s chain of command looked functional in 1961 because nothing had been ordered that the army would not do. Part Seven’s Article 370 concurrence requirement looked like a safeguard until somebody noticed who chose the state government.
In each case the discovery that the mechanism was gone happened at the exact moment somebody wanted to use it, which is the worst possible time and the only time anybody ever checks.
One thing to hold as this part proceeds. Nothing in this chapter requires anybody in 1966 to have been foolish. The Syndicate were experienced political operators making a reasonable calculation on the information available, and the information available did not include the one fact that mattered, because that fact could only be discovered by testing something nobody had a reason to test.
Shastri died in January 1966. The Syndicate chose Indira Gandhi because she carried a national name, had no state base of her own, and was thought to lack the appetite for a fight. She was called the dumb doll and a great many people agreed.
They were wrong about her, and that is the smaller half. The larger half is that they believed they could remove a Prime Minister, and the party had lost that capacity fifteen years earlier.
A check that has not been tested is not known to work, and everybody keeps behaving as though it does until the moment somebody needs it.
21969: The Split
The oldest political organisation in India broke in half over the election of a ceremonial president. What it was actually about was who owned the party, and the answer settled the next thirty years.
2.1 — The occasion
The President of India is a largely ceremonial office. In 1969 the incumbent died and a successor had to be elected by an electoral college of legislators.
The Congress organisation — the Syndicate — nominated Neelam Sanjiva Reddy as the party’s official candidate. The Prime Minister did not want him.
What she did next is the pivot of the whole period. She let it be known that Congress legislators should vote according to conscience, and supported the independent candidate V.V. Giri, who won.
A “conscience vote” in a party whose official candidate is on the ballot is not a call for reflection. It is an instruction to defeat your own party’s nominee, issued in language that cannot be quoted as an instruction.
The organisation expelled her. She took the larger part of the parliamentary party with her. Congress became two parties: hers, and the organisation’s, which withered.
2.2 — What it was fought about, in public
Publicly the split was framed as a fight between the left and the right of the party. She had nationalised the banks a month earlier, which Chapter Three covers. The Syndicate was cast as conservative, pro-business and obstructive; she was cast as the champion of the poor against entrenched wealth.
That framing was extremely effective. It was also, as a description of what was at stake, close to useless.
Candidate selection: the process by which a party decides who will carry its symbol in each constituency. In Indian usage, giving somebody the ticket.
It is the quietest and most powerful thing a party does. In a constituency where the party reliably wins, the ticket is the election; the vote afterwards is a formality. So whoever controls selection controls the careers of everybody who holds a seat, and therefore controls the votes of the parliamentary party.
Why it matters here: this is what the 1969 split was actually a fight over, and it is almost never described that way.
2.3 — What it was actually about
The Syndicate were not ideologues. They were the men who controlled candidate selection, party funds and state machinery. What they had that she wanted was the ability to decide who got a Congress ticket.
Consider what a party ticket meant. A Congress nomination in most of India in 1969 was worth a seat. Whoever granted it owned the loyalty of the person holding it, and therefore owned the votes of the parliamentary party, and therefore owned the leadership.
The split transferred that. After 1969, and decisively after the landslide of 1971, a Congress member of parliament owed his seat to her name and her nomination rather than to a state boss.
Which is why the check Chapter One described — the parliamentary party’s theoretical power to remove a leader — went from unusable to unimaginable. It had failed to work because MPs owed their seats elsewhere. Now they owed them to the person the power existed to remove.
Think of a shop with a manager and a board of directors who own the building.
The manager cannot be removed by the customers, only by the board. So she goes to the customers and persuades them to trade only with her, wherever she is.
The board still owns the building. It is now an empty building.
Nothing in the company’s rules has changed. Every provision about removing a manager is still printed in the articles. The board simply no longer has anything the manager needs.
Which is not how the split is remembered, by either side.
Everybody who describes the 1969 split assumes that it was ideological.
Her supporters say a progressive leader broke free of a reactionary old guard. Her opponents say a leader used socialist rhetoric to destroy a party that had constrained her. Both are arguing about whether the socialism was sincere, and both accept that the split was about a direction of policy.
It was about candidate selection. That is not a cynical reading; it is a structural one. The Syndicate’s power consisted entirely of controlling who got a Congress ticket, and no amount of agreement on economic policy would have made them give that up, and no amount of disagreement would have mattered if they had.
The evidence is what happened to the policy afterwards. If the split had been about socialism, the socialism would have been the point. Part Four found that the economic framework the split supposedly fought over was still running unchanged in 1985, was diagnosed as failing by the government’s own committees in 1969 — the same year — and was not altered until the country ran out of money in 1991.
So the party split over an economic direction that then went nowhere for twenty-two years, which is a strange thing for an ideological conflict to produce.
Why the ideological framing survives is that it flatters both sides. It lets her admirers describe a battle of principle. It lets her critics describe a betrayal of principle. What it hides is that the fight was about a nomination form, and that the person who won it acquired something no Indian leader had held before or has held since: a party in which every member’s career ran through one office.
Whether the organisation could have prevented any of this is the question its survivors spent the rest of their lives on.
The oldest political organisation in India broke in half. Whether that had to happen is genuinely disputed.
She did not leave; she was expelled. The organisation could have absorbed a Prime Minister who backed a different presidential candidate — parties survive worse every year — and chose instead to make it a question of authority, which is the one question it could not win. A group whose power depended on a leader’s deference should never have staged a confrontation that would test whether the deference existed.
The confrontation was not about a presidential candidate; that was the occasion. The organisation’s entire function was controlling candidate selection, and a Prime Minister with a national name and a direct appeal to voters was going to acquire that control eventually, because MPs go where their next nomination is. There was no arrangement in which both parties kept what they had. The split was the resolution of an incompatibility, not the failure of a negotiation.
What would settle it: whether any serious attempt was made to construct a rule-based selection process in the new party after 1969. The record suggests none was.
Why people care so much: because every Indian party since has been organised the way Congress was after 1969 rather than the way it was before, and nobody has ever argued that this was an improvement.
Whatever its inevitability, the consequence is not in dispute.
The split was triggered by a conscience vote against her own party’s presidential candidate — an instruction to defeat the official nominee, phrased so it could not be quoted as one.
Publicly it was left against right. Actually it was about who controlled the party ticket, because a Congress nomination was worth a seat and whoever granted it owned the holder.
After 1969 and decisively after 1971, every Congress MP owed his seat to her. The power to remove a leader now belonged to people whose careers ran through her office.
3Nationalisation
Fourteen banks taken into public ownership by an ordinance signed on a Saturday. It was hugely popular, the courts struck it down, and it was done again a fortnight later.
3.1 — What was done
In July 1969 the government took the fourteen largest privately owned commercial banks in India into public ownership. It was done by ordinance — an instrument that allows the executive to legislate when Parliament is not sitting — and it was done at speed.
The case for it was not weak and Part Four’s framework helps state it. Credit was the scarcest input in the Indian economy after permission itself. Private banks lent to the industrial houses they were connected to, in the cities where they were. Agriculture, small business and most of the country were served barely at all.
Public ownership was intended to redirect credit — to require lending to agriculture, small industry and the regions the private system ignored, and to put branches where there had never been any.
Ordinance: a law made by the executive when Parliament is not sitting, having the same force as an Act, and lapsing unless Parliament approves it within a set period after it reassembles.
It exists so a government can act on something urgent that cannot wait for a session. The obvious risk is equally simple: the executive decides both what is urgent and when Parliament sits.
Why it matters here: fourteen banks changed ownership by an instrument that required no vote, and Parliament was presented with the result rather than the question.
3.2 — What it achieved
An honest account gives it two things.
Branch expansion was real and enormous. Banking reached rural India in a way it never had, and the number of branches multiplied over the following decade. For a very large number of Indians the first bank they ever saw arrived because of this decision.
Directed lending changed who got credit. Agriculture and small enterprise received a share of formal credit that the previous system had never given them.
And it gives it two costs.
Lending decisions became political. A bank that must lend to a category on instruction, and whose senior appointments run through a ministry, will end up lending to people the ministry favours. Loan waivers announced before elections became a permanent feature of Indian politics and remain one.
Accumulated bad debt. The consequences of forty years of directed lending sat on the books of public banks for decades and are still being worked through.
Before 1969, a farmer in most of rural India who needed money before the harvest had one source: a moneylender, at whatever rate he named, secured on land or on the crop.
A bank branch does not remove that man. It competes with him. It puts a ceiling on what he can charge, because there is now somewhere else to go — and it does that in a village where, a decade earlier, the nearest formal lender was a day’s travel away.
That is what the branch expansion actually delivered, and it is not visible in any argument about ownership. It is visible in interest rates in places that had never had a choice.
3.3 — The pattern, for the second time
The ordinance was challenged and the Supreme Court struck it down.
The government re-enacted it, in a form addressing the court’s objections, within about a fortnight.
Notice the shape, because it is the same one Chapter Four is about and the same one Part Three found in 1951. A court says no; the government does it again in a way the court cannot reach. On this occasion the response was legislative and proportionate — redrafting a law to meet a judgment is exactly what a government is supposed to do when it loses a case.
Hold that, because it is the benchmark. What happens in Chapter Four, fifteen months later, is the same government facing the same kind of obstacle and reaching for something entirely different.
The nationalisation is documented from both ends. The ordinance and the subsequent Act are published law. The judgment striking down the first attempt is a reported decision. Branch numbers, deposit figures and sectoral lending shares are published by the Reserve Bank annually and were collected for operational reasons rather than to settle an argument.
That makes the outputs of this policy unusually easy to check — which is why this chapter can be confident about branch expansion and about directed lending shares.
What it cannot show: the counterfactual. Whether a differently regulated private banking system would have expanded rurally anyway, more slowly or not at all, is not answerable, and anybody who tells you the branch expansion proves the policy has skipped that step.
Which leaves the question India has been arguing about ever since.
Fifty-seven years on this is still argued in India, and unusually the two sides agree on most of the facts.
Before 1969 formal banking in India served the cities and the industrial houses. After it, banks reached villages that had never had one, and agriculture and small business got credit for the first time. Whatever the later problems, tens of millions of people entered the formal financial system because of this, and there is no serious account of how a private system with the incentives it had would have done it.
The problem was real and the diagnosis was correct: credit was misallocated. But ownership was not the only tool. Branch licensing requirements, priority-sector obligations and deposit insurance could have achieved much of the redirection without transferring lending decisions to a system in which they became political — which is what produced the pre-election loan waivers and the bad debt. Part Four’s distinction applies: the state acted as an owner where it could have acted as a regulator.
What would settle it: comparing rural branch growth in countries that used regulation rather than ownership over the same decades. The comparison is messy and has not been done well.
Why people care so much: because it is the founding act of the economic argument that still divides Indian politics, and because both sides need it to have been either wholly right or wholly wrong.
Before leaving this chapter, note what it does not contain. There is no charge here about the response to the court. The government lost a case, read the judgment, redrafted the law and passed it. That is the system working, and it is worth marking because it is the last time in this part that it happens.
Fourteen banks nationalised by ordinance in July 1969. The problem was real: credit went to the cities and the industrial houses and almost nowhere else.
Branch expansion into rural India was enormous and belongs on the credit side. The cost was that lending decisions became political, permanently, and no government since has reversed it.
The court struck the ordinance down and the government re-enacted it properly within a fortnight. That is what a government is supposed to do when it loses a case. Remember it — Chapter Four is the same obstacle answered differently.
4The Privy Purses
Part Two recorded a promise and said this part would say whether breaking it was justified. Here is the verdict, and it separates cleanly into two answers that point in opposite directions.
4.1 — The debt
Part Two’s Chapter Five described what the princely rulers were given in exchange for signing their states into India. Personal privileges, retention of private property, and a privy purse — an annual, tax-free payment, written into the constitution itself at Articles 291 and 362, with the definition of a “Ruler” at Article 366(22).
Part Two also noted that Article 363 excluded the courts from disputes arising out of the accession covenants, and said this: whatever one thinks of hereditary payments to former autocrats, the guarantee was a term of the bargain that produced the country, and it was withdrawn unilaterally twenty-four years later. It recorded that as a forward promise and said Part Eight owed a verdict.
4.2 — September 1970
On 2 September 1970 a constitutional amendment bill was introduced in the Lok Sabha to delete Articles 291 and 362 and clause 22 of Article 366. It passed, 336 votes to 155.
On 5 September it went to the Rajya Sabha, where a constitutional amendment requires a majority of not less than two-thirds of members present and voting. The count was 149 for, 75 against. It fell short by a single vote.
Parliament had said no.
On 6 September 1970 — the next day — the President signed an order in respect of each ruler, exercising the power under Article 366(22) to recognise rulers, and directing that with effect from that date every one of them ceased to be recognised as a ruler.
Since a privy purse was payable to a “Ruler”, and there were now no Rulers, the payments stopped.
Read that sequence once more. A constitutional amendment fails by one vote in the upper house on the fifth of September, and on the sixth of September the same result is achieved by executive order.
Colourable exercise of power: a legal term for using a power you genuinely hold, for a purpose it was not given to you for, in order to achieve something you could not achieve directly.
The doctrine is sometimes summarised as: what you cannot do directly, you cannot do indirectly. The power is real; the use of it is a device.
Why it matters here: the President’s power to recognise rulers existed so that the state would know who its counterparties were. It was used to make them all cease to exist on the same day, in order to stop payments a constitutional amendment had just failed to stop.
It is worth pausing on what that vote of 5 September actually represented before looking at what happened to it.
The two-thirds requirement in the Rajya Sabha exists for one reason: to make constitutional change harder than ordinary change.
On 5 September 1970 that requirement did its job. It is the only occasion in this entire part on which any institution stopped the government doing something it wanted.
It held for approximately eighteen hours.
4.3 — December 1970
The rulers challenged the order. The lead petitioner was Madhav Rao Scindia of Gwalior.
On 15 December 1970 the Supreme Court struck the order down. The bench was eleven judges. The holding was that the President could not derecognise rulers by executive action where the constitution itself provided for their recognition and their purses, and that rights guaranteed by the constitution could be removed only by constitutional amendment.
Two judges dissented: Mitter and A.N. Ray. Remember the second name; he appears again in Chapter Six.
The payments were restored.
4.4 — 1971
Nine days after the judgment the Prime Minister asked the President to dissolve Parliament and call an election. Chapter Five is about that election, which she won with around 352 seats.
With that majority the amendment route was open. The Constitution (Twenty-Sixth Amendment) Act, 1971 deleted Articles 291 and 362, replaced Article 366(22), and abolished the privy purses and the recognition of rulers.
This time it was done properly, and nobody has ever seriously argued otherwise.
4.5 — The verdict Part Two asked for
It comes in three parts and they do not point the same way.
On the substance: abolition was right. Hereditary, tax-free payments from public funds to the descendants of autocrats, in a republic where most people were poor and a great many could not read, cannot be defended on any democratic principle. The rulers had surrendered sovereignty they had held by conquest and inheritance, not by consent, and Part Two’s own Chapter Five established that the populations of those states were asked about accession exactly once. Nothing in this series should be read as a defence of the privy purses.
On the method of 1971: legitimate. A constitutional guarantee removed by constitutional amendment, by a Parliament with the required majority, after an election fought openly on the issue, is the correct procedure. If the government had done this in the first place, this chapter would be two paragraphs long.
On the method of September 1970: indefensible, and eleven judges said so. A parliamentary defeat answered the next day by an executive order using a recognition power as a device is not a close case. It is the thing constitutions exist to prevent, and the Supreme Court’s response was near-unanimous.
So the verdict Part Two is owed is: the objective was right, the 1971 method was right, and the fourteen weeks in between are the charge.
All of which is documented to a standard that makes the verdict above unusually secure.
Every element of this chapter is a public document with a date. The bill and its Lok Sabha division. The Rajya Sabha count of 149 to 75. The Presidential orders, gazetted. The eleven-judge judgment, reported, with the dissents named. The 26th Amendment, in the statute book.
There is no version of this sequence available to anybody. It cannot be softened by a different memoir or hardened by a better one.
What it cannot show: whether the September order was improvised overnight or prepared in advance as a contingency. That distinction matters for how the chapter should be read — an improvisation is a worse decision and a prepared alternative route is a worse plan — and the papers that would answer it are closed.
Which brings the verdict Part Two asked for into contact with the argument it has to survive.
Part Two framed this as a promise given for the country’s existence and taken back. Both cases deserve full statement.
A constitution is not a contract with dead men. The purses were an inducement offered under duress in 1947 to people who had no democratic claim to anything, and the state that offered them was three weeks old. By 1970 India was a republic of poor citizens paying tax to fund the leisure of families whose only qualification was birth. Every generation is entitled to revisit the arrangements of the last, and the correct instrument for that is a constitutional amendment, which is what was ultimately used.
Nobody defends the purses. The problem is what the sequence taught. The rulers gave up something irreversible — their states — for something the state later withdrew. If a guarantee written into the constitution as the price of accession lasts only until a government assembles two-thirds of a chamber, then the honest description of an Indian constitutional guarantee is that it is a promise good until it becomes inconvenient. And the September order made it worse: it showed that even the two-thirds threshold could be bypassed for a few months by using a different power for a purpose it was not given for.
What would settle it: nothing about the purses. What can be tested is whether Indian constitutional guarantees have generally proved durable against determined majorities, and the answer across this series so far is that they have not.
Why people care so much: because the princes are the least sympathetic possible claimants, which makes this the easiest case in which to establish a principle that later applies to everybody.
Notice that both positions treat the constitutional status of the promise as meaning something definite. That is the shared move and it does not survive inspection.
Both sides of that argument assume that a constitutional guarantee guarantees something.
The critics of the abolition say a promise in the constitution was broken. The defenders say the promise was properly amended away. Both treat the constitutional status of the purses as meaning something — either a binding commitment or a revisable one — and they argue about which.
A guarantee is worth exactly the difficulty of removing it, and nothing else. Not the solemnity of the language, not the fact of being written down, not the intention of the people who wrote it.
In India the difficulty of removing a guarantee is a two-thirds majority in both houses. Part Three established that first-past-the-post routinely delivered that to a party holding around 45 per cent of the vote. So the honest valuation of an Indian constitutional guarantee, for most of this period, is: secure against everybody except the government.
Which is why the September order matters more than it looks. It briefly demonstrated something worse — that the two-thirds threshold could be sidestepped altogether for a few months by using an unrelated power as a device — and that eleven judges were required to say otherwise.
The general rule is the one Part Seven reached about Article 370 by a different route. A protection is only as strong as the hardest obstacle standing between a government and its removal. Everything else about it — the words, the placement, the history — is decoration, and arguing about the decoration is what both sides here are doing.
Carry the September order into Chapter Six rather than leaving it here. What it established was not about princes. It was that when an obstacle appears, the first response available is to find a different power and use it for a purpose it was not given for — and that the only thing standing between that response and success is a court willing to say so within three months.
The amendment failed in the Rajya Sabha on 5 September 1970 by one vote — 149 to 75. On 6 September a Presidential order achieved the same result by derecognising every ruler.
Eleven judges struck it down on 15 December 1970. Two dissented: Mitter and A.N. Ray. The purses were then abolished properly by the 26th Amendment in 1971.
Part Two’s verdict: the objective was right, the 1971 method was right, and the fourteen weeks in between are the charge. A guarantee is worth exactly the difficulty of removing it.
5Garibi Hatao
Two words won the largest majority any Indian party had held since 1957. This chapter scores them the way Part One promised: against the government’s own targets.
5.1 — The slogan
The 1971 general election was called early, nine days after the Supreme Court restored the privy purses. It was fought on a slogan of two words: Garibi Hatao — remove poverty.
It worked completely. Congress won around 352 seats. The opposition alliance, campaigning on Indira Hatao — remove Indira — was destroyed.
The slogan’s power was that it went over the heads of everybody in between. It did not ask voters to approve a programme, or to evaluate a record, or to choose between state party bosses. It addressed the poor directly, from the top, and it named the thing they actually wanted.
5.2 — What was promised
Part One set the standard this chapter uses: judge a government against its own stated targets, because nobody can call that standard unfair.
The commitments here are unusually explicit, because Indian governments state them in plan documents. Poverty removal was made a central objective of national planning in this period, with targets attached, in documents the government wrote itself.
Poverty line: an income or expenditure level below which a person is counted as poor. In India it has usually been anchored to the cost of a minimum daily calorie intake.
The important thing about it is that it is drawn. A committee decides what the minimum is, what it costs, and how to adjust for prices and regions. Move the line a small amount and tens of millions of people cross it without eating differently.
Why it matters here: any statement that poverty fell or rose by a given number of points in this period is a statement about a committee’s choices as much as about anybody’s life, and Part Three’s Chapter Seven set out why.
5.3 — What arrived
The record over the following years is not good and it is worth being precise rather than rhetorical about why.
Poverty did not fall substantially in this period. Part Four’s figures apply: the share of the population below the poverty line remained somewhere in the range of forty to fifty per cent through most of the 1950s to 1980s, with only gradual movement. The years covered by this part contain no break in that pattern.
The immediate years after 1971 were worse than what preceded them. The war with Pakistan was expensive. The refugee burden of 1971 had to be paid for. Two poor monsoons followed. Then the oil shock of 1973 hit an economy that imported its oil, and inflation reached levels that had no precedent in independent India — well above twenty per cent at its worst.
By 1974 the country was in visible distress. There was a national railway strike, which was broken. There were agitations in Gujarat and in Bihar. Chapter Eight is what those became.
An honest account has to concede that most of the immediate causes were external: a war India did not start, a monsoon nobody controls, and an oil price set in Vienna. A government cannot be blamed for the weather or for OPEC.
What it can be held to is the gap between what it promised and what it built. And this is where Part Four’s finding arrives with force. In 1969 — the same year as the split, two years before the slogan — the government’s own committees reported that the economic framework was producing the opposite of its stated purpose. The response was not correction. It was more of the same instrument, and a slogan.
Inflation above twenty per cent means that a family which could buy a month’s grain in January can buy about three weeks’ worth by December, on the same wage.
For a household spending most of its income on food — which is most Indian households in 1974 — that is not an economic statistic. It is a week of the month with nothing in it, arriving every month, getting longer.
And it falls hardest on exactly the people the slogan of 1971 had addressed, which is why the political consequences arrived so fast and from the direction they did.
Scoring the promise therefore requires care about what can and cannot be established.
Two kinds of evidence apply here and they are of very different quality.
Good: the plan documents, which state objectives and targets in the government’s own words; the price series, which is collected monthly for operational reasons; and the electoral results. None of these was produced to settle this argument.
Weak: the poverty ratio itself, for the reason in the box above. Different series and different lines produce materially different numbers for the same years.
This chapter therefore rests its verdict on the first kind and describes the second as a range. Anybody quoting a precise fall or rise in poverty for these years is reporting a methodology.
What none of it can show: what would have happened without the slogan. The 1971 mandate changed what governments had to promise, and there is no counterfactual India in which it did not.
With those limits stated, the promise can be scored.
Scoring a poverty promise is harder than it looks, and Part Three explained why: the line moves.
Judge it against what came before rather than against the words. Before 1971 poverty was something Indian governments discussed as a consequence of growth arriving slowly. After it, no Indian party could contest an election without a direct promise to the poor, and every welfare programme of the following fifty years — food security, employment guarantees, direct transfers — sits inside a political space this slogan created. That is a real achievement and it is invisible if you only count poverty ratios in the four years that followed.
A promise is not a politics; it is a commitment with a date. The government’s own committees had told it two years earlier that the economic framework was producing concentration rather than dispersal, and nothing was changed. No large redistributive instrument was constructed in these years. What was constructed was a direct relationship between a leader and the poor which required no party, no state government and no delivery — which is exactly why it survived the failure to deliver.
What would settle it: the poverty series, which is contested precisely because Part Three’s finding about moving lines applies to every number in it.
Why people care so much: because this is the origin of the argument about whether Indian welfare politics is redistribution or theatre, and both answers have been available from the same evidence for fifty years.
One further consequence belongs to this chapter and runs into the next. A leader who speaks directly to the poor over the heads of her party does not need the party — and a leader who does not need the party has removed the last group of people with both a reason and a means to say no to her. Chapter Two took the organisation apart. This is the electoral demonstration that it was not going to be missed.
Garibi Hatao won around 352 seats in 1971. Its power was that it went over the heads of the party, the state bosses and the programme, and spoke directly from a leader to the poor.
Poverty stayed in the range of forty to fifty per cent. The years after 1971 were worse — a war, two bad monsoons, and an oil shock producing inflation above twenty per cent. Most of those causes were external and this part concedes it.
What is not external: the government’s own committees had reported in 1969 that the economic framework was failing, and the response was a slogan rather than a correction.
6The Courts, 1967 to 1973
Six years of escalation ending on a single morning, when the Bar of the Supreme Court went on strike for the only time in its history.
6.1 — The question
One question runs through all of this and it is worth stating before the cases. Can Parliament amend any part of the constitution, including the fundamental rights?
Part Three’s Chapter One found the Supreme Court answering yes in 1951, in Shankari Prasad, and said that reading stood for over two decades. It is what made the method of that chapter possible: a court reading the constitution inconveniently could be overridden by amendment.
In 1967, in Golaknath, the Court changed its mind. It held that Parliament could not amend the fundamental rights at all.
That was a genuine constraint, and it arrived at a government about to nationalise banks and abolish privy purses — both of which required interfering with property rights.
6.2 — The answer
The response came in a sequence of amendments passed with the majority won in 1971.
The Twenty-Fourth Amendment asserted that Parliament could amend any provision of the constitution, including the fundamental rights, and that the President was obliged to assent to such a bill. It was aimed directly at Golaknath.
The Twenty-Fifth Amendment curtailed the right to property and inserted Article 31C, which protected certain laws from challenge on fundamental rights grounds.
The Twenty-Sixth Amendment abolished the privy purses, as Chapter Four described.
The Twenty-Ninth Amendment placed Kerala land reform legislation in the Ninth Schedule — the device Part Three’s Chapter Two examined, now with a long history of drift behind it.
Taken together the message was unambiguous: the constitution belongs to Parliament, and Parliament belongs to us.
Set out together, the sequence is easier to read than any description of it.
| Amendment | Year | What it did | What it was answering |
|---|---|---|---|
| Twenty-Fourth | 1971 | Asserted that Parliament may amend any provision including the fundamental rights, and obliged the President to assent. | Golaknath, 1967. |
| Twenty-Fifth | 1971 | Curtailed the right to property and inserted Article 31C, shielding certain laws from fundamental rights challenge. | The property judgments that had obstructed nationalisation. |
| Twenty-Sixth | 1971 | Deleted Articles 291 and 362, replaced Article 366(22), abolished the privy purses. | The eleven-judge privy purse judgment of December 1970. |
| Twenty-Ninth | 1972 | Placed Kerala land reform legislation in the Ninth Schedule. | Challenges to state land reform. |
Read the last column. Three of the four were answering a court. That is the method Part Three identified in 1951, running at four times the speed and with the full majority of 1971 behind it.
6.3 — Thirteen judges
Those amendments were challenged, and the challenge became Kesavananda Bharati v. State of Kerala.
It was heard by a bench of thirteen judges — the largest in the Supreme Court’s history, before or since — constituted by Chief Justice S.M. Sikri. The hearing ran sixty-eight days, the longest in the Court’s history. The judgment, delivered on 24 April 1973, ran to over seven hundred pages.
By seven votes to six, the Court held that Parliament may amend any part of the constitution but may not damage or destroy its basic structure.
The judges in the majority did not agree on a definitive list of what the basic structure contained. Different opinions named the supremacy of the constitution, the republican and democratic form of government, secularism, the separation of powers, federalism, the dignity of the individual, and the unity of the country.
The doctrine has held ever since. It is the most consequential judgment in Indian constitutional history, and it exists by one vote.
Basic structure doctrine: the rule that Parliament may amend the constitution but may not alter its essential features.
Its oddity is that the constitution does not say this anywhere. The Court derived it from what a constitution is — reasoning that a power to amend cannot include a power to abolish, because a document that can be replaced by amendment is not a constitution but a resolution.
Its weakness is the obvious one: thirteen judges could not agree on what is in the basic structure, and no list has ever been settled.
Why it matters here: it is the only check in this entire part that was tested and held, and Chapter Nine is about what that is worth.
6.4 — 26 April 1973
Chief Justice Sikri was due to retire on 25 April 1973, the day after the judgment. His successor had not been announced.
By convention observed for the whole of the Court’s existence, the Chief Justiceship went to the senior-most judge. That was Justice J.M. Shelat, followed by Justice A.N. Grover and Justice K.S. Hegde.
All three had been in the majority in Kesavananda.
On 26 April 1973 the government appointed Justice A.N. Ray as Chief Justice of India, over all three of them. Ray had been in the minority in Kesavananda, arguing for an unlimited power of amendment. He had also been one of the two dissenters in the privy purse case in December 1970.
Shelat, Hegde and Grover resigned.
The Bar of the Supreme Court went on strike. It is the only occasion in the Court’s history on which it has done so. Senior members of the Bar and former judges issued a joint statement describing the action as a blatant and outrageous attempt at undermining the independence and impartiality of the judiciary.
A member of parliament summarised it in a sentence that has outlived every official explanation: the boy who wrote the best essay got the prize.
Everything in section 6.4 is public and dated: the judgment of 24 April, Sikri’s retirement on the 25th, the appointment on the 26th, the seniority list, the three resignations, and the Bar’s strike, which was reported at the time.
The government’s own justification is also on the record and should be stated fairly: officials argued that judicial elevation should turn on merit and judicial philosophy and on alignment with the country’s social and economic goals, rather than on mechanical seniority. That is a coherent position and several countries appoint judges that way.
What this cannot show: motive. That the three judges passed over were in the majority and the judge appointed was in the minority is a fact. That the appointment was made because of that is an inference — a very strong one, drawn at the time by the Bar and by most subsequent scholarship, but an inference.
This chapter states it as one, which is why the argument box below gives the government’s case its best form rather than dismissing it.
Before the government’s justification, it is worth registering what the reaction actually was, because it came from people with nothing to gain by it.
Three men gave up the highest office available in their profession, on two days’ notice, with nothing to gain and their pensions and reputations to lose.
Judges do not do this. The whole disposition of the office is towards continuity, and a judge who resigns in protest forfeits the ability to decide anything ever again — which is the only power he has.
And the Bar, which is to say the working lawyers whose livelihoods depend on the courts sitting, went on strike. They have done it once in seventy-six years.
Whatever the government’s justification was worth, those are the people who knew most about it, and that is what they did.
The government nevertheless had an argument, and it deserves to be stated properly rather than waved past.
The government had a real argument and it is almost never stated properly in Indian accounts.
Seniority is a convention, not a rule, and the constitution nowhere requires it. Appointing the senior-most judge automatically is a way of avoiding a decision, not of making a good one, and it hands the office to whoever happens to be oldest. Many democracies appoint chief justices for their judicial outlook, openly. And there is a democratic argument underneath: an unelected court had just given itself the power to strike down constitutional amendments passed by an elected Parliament, on the basis of a doctrine found nowhere in the text, by a single vote.
Every word of that would be worth taking seriously if the appointment had been made at any other time or on any other basis. It was made two days after a judgment, over the three judges who had decided it against the government, in favour of the judge who had decided it for them, having also decided the privy purse case for them thirty months earlier. A government entitled to select for judicial philosophy is a government able to tell every remaining judge what philosophy pays — which is a lesson delivered not to three men but to the whole bench, and Part Nine records what the bench did in 1976.
What would settle it: the file recording how the appointment was decided. Not released.
Why people care so much: because Indian governments of every party have since appointed and transferred judges with an eye to outcomes, and every one of them has cited a version of the first argument above.
Both sides of that argument, though, share a description of the contest that does not match what happened in it.
Everybody who writes about these six years assumes that the fight was between Parliament and the courts.
The standard framing is institutional: an elected legislature asserting its supremacy against an unelected judiciary asserting constitutional limits. Both sides use it. Defenders say Parliament was defending democracy; critics say the Court was defending the constitution.
Parliament was not a party to this fight. It was an instrument in it.
Look at what Parliament actually did across the whole sequence. It passed the amendments the government wrote, with the majority Chapter Five describes, delivered by members whose seats Chapter Two showed running through one office. At no point in these six years did Parliament as a body take a position that the government did not want, with a single exception — the Rajya Sabha’s one-vote refusal in September 1970 — and Chapter Four records what happened to that within twenty-four hours.
So the contest was not between two institutions. It was between a government and every check on it, and Parliament was one of the checks that had already stopped functioning as one.
This matters because the institutional framing supplies the government’s best defence and it supplies it for free. “The elected branch versus the unelected branch” is a genuinely serious argument in constitutional theory, and it is the argument the second side of the box above has to answer. But it only applies where the elected branch is actually deciding something.
The test is simple and this series has applied it before. Name an occasion in these nine years when Parliament stopped the government doing something it wanted. There is one. It lasted a day.
The last word on this chapter should be a fair one. The doctrine that emerged from these six years is the single most important protection in Indian constitutional law, and it exists because a court under pressure produced it by one vote. Whatever else the confrontation of 1967 to 1973 did, it left India with a limit on amendment that no government since has been able to remove — including, as Chapter Nine records, in circumstances where a government tried.
Golaknath in 1967 held that Parliament could not amend fundamental rights. The 24th, 25th, 26th and 29th Amendments answered it, using the majority won in 1971.
Kesavananda Bharati: thirteen judges, sixty-eight days, judgment 24 April 1973, seven to six. Parliament may amend anything but may not damage the basic structure. It is the only check in this part that was tested and held.
On 26 April 1973 the government appointed A.N. Ray Chief Justice over the three senior judges who had decided against it. All three resigned. The Bar struck, for the only time in its history.
7The Machinery
An inventory, taken on 11 June 1975. Everything Part Nine describes already existed by this date, and almost none of it was new.
7.1 — Detention without trial
Part Two recorded a promise. Detention without trial was justified in 1950 by circumstances everybody could see — an armed insurrection, an assassination, millions in camps — and Part Two said the test of an emergency power is what happens when the emergency ends. Part Three tracked it through the 1950s: the conditions disappeared and the power was renewed.
The Preventive Detention Act of 1950 was renewed repeatedly and lapsed in 1969.
In 1971 it was replaced by the Maintenance of Internal Security Act, generally called MISA. It permitted detention without charge or trial on the executive’s satisfaction that a person might act in a manner prejudicial to the security of the state, public order or the maintenance of essential supplies.
So the power that Part Two’s Chapter Nine described as an emergency measure in 1950 had, by 1971, been continuously available for twenty-one years, across a period containing no insurrection of the kind that justified it.
7.2 — Article 356
Part Three established the instrument and its first political use, against the elected Kerala government in 1959, and said every later use stood on that one.
Between 1966 and 1975 the provision was used against elected state governments repeatedly — not a handful of times but dozens, across most of the country, and overwhelmingly against governments the centre opposed.
The exact figure is a matter of published record. Every use is a formal act with a date, a proclamation and a parliamentary approval, and anybody with a fortnight could compile the list. Nobody has published one and this series has now met that pattern often enough to name it: an unmade count, using material anybody could reach.
7.3 — What else was on the shelf
Take the inventory as it stood on 11 June 1975.
| Instrument | Origin | Status by June 1975 |
|---|---|---|
| Detention without trial | Article 22 of the constitution; the Act of 1950 | In force as MISA since 1971. Twenty-one years of continuous availability. |
| Dismissal of state governments | Section 93 of the Government of India Act 1935, carried into Article 356 | Used dozens of times since 1966, mostly against opponents. |
| Emergency provisions | Inherited from the 1935 Act, carried into Part XVIII | Unused for internal disturbance. Available. |
| Amendment of the constitution | Article 368, and the precedent of 1951 | Used four times in 1971–72 alone, including to reverse two Supreme Court decisions. |
| The Ninth Schedule | The First Amendment, 1951 | Grown from thirteen entries to a long list, still unreviewable. |
| Restriction of speech | Article 19(2) as amended in 1951 — public order, friendly relations, incitement | Available and long-established. |
| Appointment of judges | Convention of seniority, no rule | Convention broken in April 1973. Three resignations. Bench aware. |
| Control of the party | Candidate selection after 1969 | Complete. Every Congress MP’s career ran through one office. |
Look down the origin column. Six of the eight predate 1966 entirely, and four of them predate independence. Nothing on this list was invented in the period this part covers, with the partial exception of the last two.
7.4 — 1974
Two events of 1974 belong here because they show the machinery in ordinary use.
In May 1974 a national railway strike involving something in the order of a million workers was broken. Large numbers of workers and union leaders were arrested and held, the detention powers described above being available for exactly this.
In the same month India conducted its first nuclear test.
Set the two side by side and the picture of the state in 1974 is complete: capable of a nuclear device and of arresting a million-strong strike, and unable to do anything about the price of grain.
The inventory in section 7.3 is entirely composed of published law and published acts. Each instrument can be traced to a statute or a constitutional article with a date, and each use of Article 356 is a proclamation laid before Parliament.
That is why this chapter can be categorical where much of this series has to hedge. There is no interpretation in the table; it is a list of powers and the dates they became available.
What it cannot show: how often the powers were used in ways that never reached a record. A detention order is documented. A conversation in which a Chief Minister is told what will happen if he does not comply is not, and there is no honest way to estimate how much of the machinery’s effect ran through the second kind.
That table invites an objection which is the strongest available against this whole part, and it should be put rather than avoided.
7.5 — The ledger for this part
Part Five committed every subsequent part to carry the running record of which states were affected by its material. This part is mostly about Delhi, so the entry is short and the shortness is itself informative.
| Where | What happened |
|---|---|
| Kerala | Its land reform legislation was placed in the Ninth Schedule by the 29th Amendment in 1972, and the challenge to that amendment became Kesavananda Bharati. The state Part Three showed being dismissed in 1959 is the one whose litigation produced the doctrine that limited amendment. |
| Gujarat | The Navnirman agitation of 1974 brought down the state government. The assembly result of 12 June 1975 gave the state to the Janata Morcha on the same morning as the Allahabad judgment. |
| Bihar | The centre of the movement led by Jayaprakash Narayan from 1974. The state Part Five’s ledger records as the largest single casualty of central economic policy is the state where the political challenge came from. |
| Uttar Pradesh | Rae Bareli, and the Allahabad High Court. The judgment that unseated a Prime Minister came from a High Court in the state that had elected her. |
| States generally | Article 356 used dozens of times between 1966 and 1975, overwhelmingly against governments the centre opposed. The count is public and has never been compiled. |
The shortness of that table is worth noticing. Parts Five and Seven had long ledgers because their material was about what the centre did to places. This part is about what the centre did to itself, and the states appear mainly as the place where the resistance came from.
Chapter Seven’s table is the whole of this part’s structural claim, so the objection to it should be put at full strength.
Every state needs emergency powers, a way to take over a collapsed provincial administration, and a procedure for amending its constitution. Britain has detention powers, America has emergency statutes, and every federation can intervene in a failed state government. What matters is restraint in use, not the existence of the tool — and a series that indicts the toolbox is really indicting the possibility of government.
Nobody argues for a state with no emergency powers. The objection is to powers with no expiry, no automatic review, and no threshold that anybody outside the government checks. Every item in that table shares those three properties, and the pattern of this whole series is that such a power is used, then normalised, then relied on. Detention without trial was justified by an insurrection in 1950 and was still in force twenty-one years later, through a period containing no insurrection, because nothing in its design required anybody ever to justify it again.
What would settle it: comparing India’s emergency statutes with those of comparable federations on the specific question of sunset clauses and mandatory review. The comparison has not been made systematically.
Why people care so much: because every government of every party has used every instrument in that table, and a finding about the design rather than the user implicates all of them.
The inventory therefore is not a list of things that were done to India in these nine years. It is a list of things that were already there.
Detention without trial had been continuously available for twenty-one years by 1971, when MISA replaced the 1950 Act. Part Two’s warning about emergency powers that never lapse arrives here.
Article 356 was used against elected state governments dozens of times between 1966 and 1975, mostly against opponents. The exact count is public and nobody has compiled it.
Of the eight instruments on the shelf in June 1975, six predate 1966 and four predate independence. Almost nothing here was invented by the government that used it.
8The Twelfth of June
At ten in the morning a judge in Allahabad read out a decision on two narrow findings. Thirteen days later the constitution was suspended.
8.1 — Courtroom 24
Raj Narain had contested Rae Bareli against Indira Gandhi in 1971 and lost by a large margin. He filed an election petition in the Allahabad High Court alleging corrupt practices under the Representation of the People Act.
The case was heard by Justice Jagmohanlal Sinha. Argument concluded in May 1975.
At ten o’clock on 12 June 1975 he delivered judgment. The petition was allowed. The election of Indira Nehru Gandhi was declared void and she was disqualified from holding elected office for six years.
It was the first time in India that a sitting Prime Minister had been unseated by judicial order.
8.2 — What was actually found
This is where most accounts, on both sides, go wrong, and getting it right is necessary for the argument that follows.
Raj Narain had made a long list of allegations. Most of them were rejected for want of proof. The claims about liquor and blankets being distributed, about Air Force aircraft being used for campaign travel, and about exceeding the expenditure limit all failed.
Two findings succeeded.
The first concerned Yashpal Kapoor, an Officer on Special Duty in the Prime Minister’s Secretariat who acted in her campaign. He had submitted his resignation on 13 January 1971 and the notification relieving him was dated 25 January. She had held herself out as a candidate from late December 1970. Under the Act as it then stood, a candidate obtaining the assistance of a government servant in that period commits a corrupt practice.
The second concerned state officials. The District Magistrate of Rae Bareli, the Superintendent of Police and the state Home Secretary arranged rostrums, loudspeakers, barricades and police for her election meetings in February 1971.
Both are real. Both are technical. Neither involves money changing hands, votes being stolen, or anybody being coerced.
And the disqualification of six years was the maximum available under the law rather than the only outcome. The judge had the discretion to void the election with lesser strictures, and did not use it.
Election petition: a legal challenge to the validity of a particular election, heard by a High Court under the Representation of the People Act.
It is not an appeal against the result in the ordinary sense. It alleges that specified corrupt practices occurred — a defined statutory list including bribery, undue influence, appeals to religion, exceeding the expenditure limit, and obtaining the assistance of government servants.
If proved, the court may declare the election void, and may in addition disqualify the person from contesting for up to six years.
Why it matters here: this was an ordinary civil proceeding of a kind decided routinely across India, brought by a defeated candidate, under a statute Parliament had written.
8.3 — The same day
Two other things happened on 12 June 1975.
Her close adviser D.P. Dhar died.
And the result of the Gujarat assembly election came in. Congress took 75 of 182 seats. The Janata Morcha coalition took 86 and formed the government.
So on one morning the Prime Minister lost her seat in court, lost a major state at the polls, and lost one of the few people whose advice she took.
8.4 — Thirteen days
She did not resign.
An appeal went to the Supreme Court. On 24 June 1975, Justice V.R. Krishna Iyer granted a conditional stay: she could continue as Prime Minister pending the appeal, but could not vote in Parliament or draw the salary of a member.
Outside the courts, the opposition movement led by Jayaprakash Narayan escalated. At a rally in Delhi on 25 June he called on the police and the armed forces not to obey orders that were illegal or immoral.
That call was, and remains, the single strongest item the government had. Asking the security forces to judge for themselves which orders to obey is not an ordinary act of opposition, and there is no democracy in which it would be treated as one.
On the night of 25 June 1975 a proclamation of emergency was issued under Article 352 on the ground of internal disturbance.
This part stops here.
The judgment is treated in India as heroic and as a technicality, depending entirely on the politics of the person describing it. Both readings have something.
Two findings survived out of many, and both were procedural. A civil servant’s resignation was in train and completed within twelve days of the disputed period. District officials arranging barricades and loudspeakers for a Prime Minister’s public meeting is what district officials did for every Prime Minister. Neither finding suggests a single vote was affected. And the six-year disqualification was the maximum the statute allowed, imposed where the judge had discretion to do less — for conduct that, if the law were applied equally, would have unseated a very large number of Indian politicians.
The judge threw out every allegation he could not prove, which is what a court is for and is why the judgment is credible. What survived, survived on the evidence. The law was clear, the facts were found, and the penalty was the one Parliament had legislated. A rule that applies to everyone else and not to the Prime Minister is not a rule. And the argument that others did the same thing is an argument for prosecuting them, not for excusing her.
What would settle it: a comparison of how similar findings under section 123(7) were penalised in other election petitions of the period. The judgments exist and nobody has assembled them.
Why people care so much: because if the judgment was disproportionate then the Emergency has a provocation, and if it was routine then it does not.
The interval between the judgment and the proclamation is usually described as though it were a scramble. It was not.
Thirteen days is not a long time to make the largest decision in a country’s constitutional history.
It is also not a short one. It is longer than the 1971 war lasted. It is long enough for a cabinet to meet several times, for legal opinions to be taken, and for the Supreme Court to hear an application and grant a stay — which it did, on the eleventh of the thirteen days, in her favour.
So the proclamation was not issued in a panic on the day of the judgment. It was issued after a court had already given her what she needed to stay in office.
Part Nine begins on the night of 25 June. Everything after that sentence belongs there.
At ten in the morning on 12 June 1975 Justice Jagmohanlal Sinha declared the Prime Minister’s election void and disqualified her for six years — the first time a sitting Indian Prime Minister was unseated by a court.
Most allegations were rejected. Two survived, both technical: a civil servant assisting before his release was gazetted, and district officials arranging rostrums and loudspeakers. No votes were shown to have been affected.
The same day Congress lost Gujarat. On 24 June the Supreme Court granted a conditional stay letting her remain Prime Minister. On 25 June the Emergency was proclaimed. This part stops here.
9What Would Have Prevented It
The defence at full strength, four moments where something could have held, and the reason the standard account of June 1975 has the causation backwards.
9.1 — Four arguments for the defence
These are the strongest arguments a serious defender of this period would make. The third is very strong and this part concedes it.
One. The crisis was real. Inflation above twenty per cent. A war and a refugee burden paid for out of an economy that could not afford either. Two poor monsoons. A railway strike involving a million workers. Agitations in two states, one of which had just brought down a government. And a national opposition leader publicly asking the police and the army to decide for themselves which orders to obey. A government facing that combination is not facing normal politics.
Two. Every objective in this part was popular or defensible or both. Nationalising banks brought banking to rural India. Abolishing hereditary payments to princes was right. Addressing poverty directly was overdue. This part has said so in every chapter and does not resile from it.
Three. The one check that mattered held. This is the strong one. Kesavananda was decided against the government and it stood. The government appointed a Chief Justice who had opposed the doctrine — and in November 1975, with the Emergency running, a bench under that same Chief Justice used that same doctrine to strike down a constitutional amendment which had sought to place the Prime Minister’s own election beyond judicial reach. The basic structure doctrine was tested at the worst possible moment, by the least sympathetic bench, and it held.
Four. She called an election in 1977 and lost and left. That belongs to Part Nine and it is flagged here because it is the single most important fact any defender of this period has, and no account of these nine years should be written as though it did not happen.
9.2 — Four moments
Part One requires an alternative that somebody proposed at the time and that was actually available. There were four.
1969: the party could have held. The Syndicate expelled her and lost the parliamentary party. Chapter Two argues this was already lost by then, because the ticket had moved. The alternative was earlier — the organisation could have tested its authority in 1966 or 1967, while state bosses still controlled selection. It did not, because there was no occasion, which is Chapter One’s finding about untested checks.
September 1970: the President could have declined. The office is largely ceremonial and refusal would have been extraordinary. It was also available, and the order was so plainly a device that eleven judges said so within three months. The occupant was V.V. Giri, elected the previous year in the contest that split the party, with her support. Chapter Two’s consequence arrives here, on a specific day, in a signature.
April 1973: the three judges could have stayed. This is the uncomfortable one and it belongs in the list. Shelat, Grover and Hegde resigned on principle, and the resignation was honourable. It also removed three basic-structure judges from the Supreme Court at the moment the doctrine most needed defending. Whether staying would have served the constitution better than leaving is genuinely arguable, and the men themselves plainly thought not.
June 1975: she could have stepped aside for the appeal. This is the strongest of the four and it was proposed at the time, inside her own party. Resign, install a trusted colleague as Prime Minister, pursue the appeal, and return when it succeeded — which it did, on 7 November 1975. The cost was a few months out of office. It was done in India before and has been done since. It required no new law, no amendment and no risk to anybody but her.
Those four are not of equal evidential standing, and it matters which is which.
Part One’s test asks whether an alternative was genuinely available at the time, not merely visible afterwards. For three of the four, availability is a matter of record rather than of speculation.
The party in 1966 and 1967: the organisation existed, held meetings and elected officers. Whether it could have acted is arguable; that it was there is not.
The President’s signature in September 1970: the office holds the power to withhold assent and to return advice for reconsideration. It was not exercised. The eleven-judge judgment three months later establishes that the order was not one a President was obliged to sign.
Stepping aside in June 1975: the mechanism is trivial and has been used in India before and since. A Prime Minister resigns, a colleague is elected leader of the parliamentary party, and the appeal proceeds. It required nobody’s permission.
What the record does not establish is the fourth item and the crucial one for that last alternative: whether anybody senior actually proposed it. That rests on memoirs written afterwards by people with reason to say they had counselled restraint. Grade: Alleged.
So the charge in section 9.2 is strongest on availability and weakest on advocacy — which is the opposite of how it is usually made.
9.3 — What I would have to retract
Three things would break this part.
If the record showed that the September 1970 order was drafted before the Rajya Sabha vote as a contingency approved on legal advice, Chapter Four’s charge softens considerably — from a device improvised in twenty-four hours to a planned alternative route.
If the appointment file for April 1973 showed that A.N. Ray was under consideration before 24 April, Chapter Six’s inference weakens substantially and the government’s merit argument gets much stronger.
If it could be shown that no senior figure proposed stepping aside in June 1975, then the fourth alternative above fails Part One’s test and this part loses its best charge.
All three depend on files the Indian government holds. That is now the fourth part in a row ending on that sentence.
Everybody who writes about June 1975 assumes that the Emergency was caused by the Allahabad judgment.
The hostile version: a Prime Minister facing disqualification suspended the constitution to keep her job. The sympathetic version: a technical judgment in a moment of national crisis forced a decision no leader should have faced. Both are stories in which a court ruling on 12 June produces a proclamation on 25 June.
Apply a test. Name an instrument used after 25 June 1975 that did not already exist before 12 June.
Detention without trial: available continuously since 1950, in force as MISA since 1971. The dismissal of state governments: used dozens of times since 1966. The emergency provisions themselves: in the constitution since 1950, inherited from a colonial statute. Censorship powers: available under Article 19(2) as amended in 1951. Constitutional amendment to reverse a court: established as a method in 1951 and used four times in 1971 and 1972. A judiciary that had watched three judges passed over: since April 1973. A party in which no career survived opposition: since 1969.
Every single one predates the judgment. Most predate the government.
Which means the judgment did not cause the Emergency. It supplied a date. What produced the Emergency was a set of instruments assembled over twenty-five years by several governments, and the removal, one at a time, of everything that might have stopped somebody using them together.
This is uncomfortable for both camps and that is the point. It denies her critics the clean story of one woman’s response to one humiliation. It denies her defenders the extenuating one about an intolerable provocation. And it makes the question for Part Nine not “why did she do it” but the harder one: why, on 26 June 1975, did nothing stop her?
That is the case against this part, and the third defence in particular is one I think simply wins on its own terms. What none of the four reaches is the list in section 9.2, because every item on it was cheap, available, and refused.
Four defences: the crisis was real; the objectives were popular or right; the basic structure doctrine held even under the Chief Justice appointed to break it; and she called an election in 1977 and lost and left.
Four moments: the party in 1966, the President’s signature in September 1970, the three resignations in April 1973, and stepping aside for the appeal in June 1975. The last was proposed inside her own party and cost nothing but a few months.
Name an instrument used after 25 June that did not exist before 12 June. There is none. The judgment supplied a date, not a cause.
10An Honest List of What We Do Not Know
This is the best-evidenced part in the series so far, which makes the short list of gaps unusually revealing about what kind of thing is never recorded.
10.1 — Genuinely unknown
Five things in this part cannot be established, and every one of them is about a decision rather than an event. The events are all documented; what is missing is why.
Why the September 1970 order was made. Whether it was improvised in the twenty-four hours after the Rajya Sabha vote, or was a prepared contingency, or was made on legal advice that it would hold, is not known.
Why it is unknown: Cabinet and Law Ministry papers are closed. This is the single gap that would most change Chapter Four, and it could move the charge in either direction.
How the April 1973 appointment was decided. That three judges were passed over is a fact. That it happened because of Kesavananda is an inference drawn at the time by the Bar and by almost all subsequent scholarship.
Why it is unknown: the appointment file has not been released. Chapter Six states the inference as an inference and gives the government’s own justification in its own terms.
Whether stepping aside was seriously proposed in June 1975. Chapter Nine’s strongest alternative rests on it having been raised inside the party.
Why it is unknown: it rests on memoirs by people with positions to defend, written afterwards, in a period when everybody had reason to say they had counselled restraint. Grade: Alleged, and Chapter Nine says so.
How Article 356 was used, in full. Every proclamation is a public document laid before Parliament. Nobody has compiled the series.
Why it is unknown: an unmade count, not a closed file. This is the third such item in three parts and the pattern is now worth naming in its own right: the most damaging material in Indian political history is frequently not hidden but unassembled.
What poverty actually did. Chapter Five scores Garibi Hatao against the government’s own targets and finds no break in the trend.
Why it is unknown with precision: Part Three’s finding applies. The poverty line is drawn by a committee, has been redrawn more than once, and each redrawing changes the number without changing anybody’s life. Any confident percentage for this period is a methodological choice presented as a measurement.
10.2 — Solid
Every vote, order, judgment and appointment in this part. The Lok Sabha vote of 336 to 155 and the Rajya Sabha vote of 149 to 75. The Presidential order of 6 September 1970. The eleven-judge judgment of 15 December 1970. The 24th, 25th, 26th and 29th Amendments. Kesavananda Bharati — thirteen judges, sixty-eight days, judgment 24 April 1973, seven to six. The appointment of 26 April 1973 and the three resignations. The Bar’s strike. The judgment of 12 June 1975 and the conditional stay of 24 June.
The 1969 split and the 1971 result. The conscience vote, the expulsion, the division of the party, and the seat count.
The findings in the Allahabad judgment, including which allegations were rejected. The judgment is reported and can be read.
The inventory in Chapter Seven. Every instrument is traceable to a statute or a constitutional article with a date.
Bank nationalisation and its measurable outputs. Branch expansion and sectoral lending shares are published annually by the Reserve Bank and were collected operationally.
The economic conditions of 1973 and 1974. The oil shock, the inflation, the railway strike and the nuclear test.
10.3 — What this part deliberately does not establish
One thing has been left alone throughout and it should be named rather than left as an absence.
This part does not establish intention. It does not claim that anybody set out in 1966 to arrive at 25 June 1975, or that the sequence was designed.
That restraint is not politeness. It is that the argument does not need it and is stronger without it. Chapter Nine’s test — name an instrument used after 25 June that did not exist before 12 June — works whether or not anybody planned anything. So does Chapter One’s finding about untested checks, and Chapter Four’s about what a guarantee is worth.
A series that has to prove a conspiracy is hostage to a document that will never be released. A series that shows a machine, a set of dates and an absence of brakes is not.
If a reader finishes this part believing that one person is the explanation, I have written it badly. The person supplied the will. Everything else had been lying about for twenty-five years, put there by people who mostly meant well.
Genuinely unknown: why the September 1970 order was made, how the April 1973 appointment was decided, whether stepping aside was proposed in June 1975, the full Article 356 count, and what poverty actually did.
Solid: every vote, order, judgment, amendment, appointment and resignation in this part — this is the best-evidenced material in the series.
This part deliberately does not establish intention, and is stronger without it. The person supplied the will. Everything else had been lying about for twenty-five years, put there by people who mostly meant well.
Sources & further reading — Part 8
Timeline, 1966–1975
Dates only. The arguments are in the chapters.
| Date | What happened |
|---|---|
| Jan 1966 | Shastri dies at Tashkent. The Syndicate chooses Indira Gandhi over Morarji Desai, believing she will be manageable. |
| Jun 1966 | The rupee is devalued. Part Four’s Chapter Eight: economically defensible, politically catastrophic. |
| 1967 | Golaknath. The Supreme Court holds that Parliament cannot amend fundamental rights. Congress loses power in many states in the general election. |
| Jul 1969 | Fourteen banks nationalised by ordinance. The Supreme Court strikes it down; it is re-enacted properly within about a fortnight. |
| Aug 1969 | A conscience vote in the presidential election defeats the official Congress candidate. V.V. Giri is elected. |
| Nov 1969 | She is expelled from the party organisation and takes the larger part of the parliamentary party with her. Congress splits. |
| 2 Sep 1970 | The privy purse amendment passes the Lok Sabha, 336 to 155. |
| 5 Sep 1970 | It fails in the Rajya Sabha by one vote: 149 for, 75 against. |
| 6 Sep 1970 | The President signs orders derecognising every ruler. The privy purses stop. |
| 15 Dec 1970 | An eleven-judge bench strikes down the order. Mitter and Ray dissent. Payments are restored. Nine days later Parliament is dissolved. |
| Mar 1971 | Garibi Hatao. Congress wins around 352 seats. |
| 1971 | The Maintenance of Internal Security Act replaces the lapsed preventive detention regime. |
| Dec 1971 | The war with Pakistan. Part Seven’s Chapter Five. |
| 1971 | The 24th Amendment asserts Parliament’s power to amend fundamental rights; the 25th curtails the right to property and inserts Article 31C; the 26th abolishes the privy purses. |
| Jul 1972 | Simla. Part Seven’s Chapter Six. |
| 1972 | The 29th Amendment places Kerala land reform laws in the Ninth Schedule. |
| 1973 | The oil shock. Inflation climbs past twenty per cent over the following year. |
| 24 Apr 1973 | Kesavananda Bharati. Thirteen judges, sixty-eight days, seven to six. Parliament may amend anything but may not damage the basic structure. |
| 25 Apr 1973 | Chief Justice Sikri retires. No successor has been announced. |
| 26 Apr 1973 | A.N. Ray is appointed Chief Justice over Shelat, Grover and Hegde, all in the Kesavananda majority. All three resign. The Bar strikes for the only time in its history. |
| May 1974 | A national railway strike involving around a million workers is broken, with large-scale detentions. India conducts its first nuclear test in the same month. |
| 1974–75 | Agitations in Gujarat and Bihar. The movement led by Jayaprakash Narayan grows. |
| 12 Jun 1975 | At ten in the morning, Justice Jagmohanlal Sinha declares her election void and disqualifies her for six years. The same day, Congress loses Gujarat and D.P. Dhar dies. |
| 24 Jun 1975 | The Supreme Court grants a conditional stay: she may remain Prime Minister but may not vote in Parliament. |
| 25 Jun 1975 | At a rally in Delhi, Jayaprakash Narayan calls on the police and armed forces not to obey illegal orders. That night, a proclamation of emergency is issued on the ground of internal disturbance. |
Part Nine starts here. It takes the twenty-one months from the night of 25 June 1975, and it is the largest single item in this series.
Glossary
Every hard word used in this part, in plain English.
| Word | What it means |
|---|---|
| Basic structure doctrine | The rule from Kesavananda Bharati that Parliament may amend the constitution but may not alter its essential features. Not written anywhere in the constitution; derived from what a constitution is. |
| Colourable exercise of power | Using a power you genuinely hold, for a purpose it was not given for, to achieve something you cannot do directly. What you cannot do directly, you cannot do indirectly. |
| Conscience vote | A vote in which members are released from party discipline. In August 1969 it functioned as an instruction to defeat the party’s own candidate, phrased so it could not be quoted as one. |
| Corrupt practice | Under the Representation of the People Act 1951, a defined category of electoral wrongdoing — including obtaining the assistance of a government servant — which can void an election. |
| MISA | The Maintenance of Internal Security Act 1971, permitting detention without charge or trial on the executive’s satisfaction that a person might act prejudicially. |
| Ordinance | An instrument allowing the executive to legislate when Parliament is not in session. Used to nationalise the banks in July 1969. |
| Parliamentary party | All the MPs of one party sitting as a body — the group that in theory chooses and removes the leader. Its power depends on members owing their seats to something other than the leader. |
| Privy purse | The annual, tax-free payment guaranteed to former princely rulers by Articles 291 and 362 as part of the terms of accession. Abolished by the 26th Amendment in 1971. |
| Supersession | Appointing a junior judge over more senior ones, breaking the seniority convention. Done on 26 April 1973 and again later in the decade. |
| The Syndicate | The informal group of Congress state bosses who controlled the party organisation in the mid-1960s and chose the Prime Minister in 1966. |
What Comes Next
Part Nine — The Emergency
Part Nine takes the twenty-one months from the night of 25 June 1975. It is the largest single item in this series. It asks:
- Part Two recorded detention without trial as an emergency measure of 1950 and said Part Nine owed the account of what it became. How many people were held, for how long, and how do we know?
- The press was censored. What was the mechanism, who administered it, and which newspapers complied, resisted or were closed?
- The 42nd Amendment is the largest single alteration ever made to the Indian constitution. What did it actually do, clause by clause?
- The sterilisation programme has a number attached to it that everybody repeats. Where does that number come from, and what does the evidence actually support?
- In ADM Jabalpur the Supreme Court held that the right to life could not be enforced during an emergency. Who wrote it, who dissented, and what happened to the dissenter?
- She held an election in 1977 and lost and left. Why did she call it, and what does the answer do to every argument in Part Eight?
Questions, not topics. If any of them is answered badly, you will be able to tell.