The Ghosts of Emergency Past — II: The Clause That Stopped Time

The bill that became the Forty-second Amendment passed the Lok Sabha at 6.02 in the evening on 2 November 1976, by 366 votes to four, seven months after the FCR Bill’s single afternoon described in Part 1. Twenty-one opposition MPs sat in preventive detention, the CPI voted aye, most of the rest boycotted, and four members voted no.

The amendment is remembered as the “Mini-Constitution”, for the Preamble it rewrote and the courts it caged. Sections 15 and 16 did something quieter. Every constitutional reference to “the last preceding census” would, until the first census taken after the year 2000, “be construed as a reference to the 1971 census”. Parliament directed the Constitution to treat 1971 as the permanent present.

The freeze arrived as population policy. Defending the year’s sterilisation spending the following afternoon, a Congress member explained: “we have even frozen the delimitation of constituencies till 2002 so as to make family planning a national objective.” April’s National Population Policy had set the two side by side, compulsory sterilization if a state wished and no lost seats for restraint; in the amendment’s year sterilizations roughly doubled, and the northern states competing for Sanjay Gandhi’s favour became the vasectomy belt.

This is the delimitation freeze’s birth certificate. Whoever defends the freeze defends a clause drafted as cover for the most coercive population campaign in the country’s history; whoever attacks it attacks the South’s only reward for half a century of restraint. Neither side’s hands are clean. As with the FCRA, this should not be a partisan issue.

THE TEMPORARY THAT BECAME PERMANENT
The lesson of Part 1 was that powers built for one government survive it. The power here is the clock: deciding when, and whether, the count of the people becomes the map of their representation. In 1976 the Emergency chamber seized it. In 2001 the power passed to a man the Emergency had jailed. Vajpayee’s government extended the freeze anyway; the 84th Amendment changed, in essence, one word: “2026” for “2000”. No Emergency was required. Any two-thirds majority to end the freeze must include the states it advantages, so the coalition can never be assembled: “an entrenched unamendable rule”, as one constitutional scholar calls it.

The freeze has inverted its justification: built to protect states from punishment, it now administers it. On the Carnegie Endowment’s figures, eight per cent of Lok Sabha seats sit in the wrong states, and a Bihar MP answers to twice a Kerala MP’s population. A rule frozen to reward restraint now watches southern governments legislate for the opposite. Since August 2024, Andhra Pradesh and Telangana have repealed their two-child bars for local elections, Chandrababu Naidu, who wrote the original bar and says so, proposes one pointing the other way, and M.K. Stalin has urged Tamil newlyweds to have more children, for the seats.

The symmetry with 1976 is exact: the Emergency’s knife suppressed births in the North, the statutes of 2024-26 solicit them in the South, and both rest on one conceit, that the state can compose the census it will be measured by. Sterilization collapsed when the Emergency lifted, “family planning” became “family welfare”, and the districts where coercion was heaviest show fertility rising from 1980 onward. The North grew fastest where the state cut deepest; the South fell below replacement with no coercion at all. The second attempt is eighteen months old and cannot succeed. Seats follow shares of the count, and no statute of 2024 moves the South’s share in thirty months; Japan and China, Naidu’s own warnings, are the ledger of lavish pro-natalism failing. What survived the first failure, and will preside over the second, is the machinery: the freeze, on its clock.

DELIMITATION BY NOTIFICATION
The government’s April answer was a package: a ceiling raised from 550 to 850, a new Delimitation Act, and the women’s reservation of 2023 finally in operation. The seat arithmetic was fought over loudly; the machinery went unexamined. It enlarges executive control over the map in three ways.

First, it cuts the link between census and map. The rule today is simple: count first, then redraw, on the count just taken. Even 1976 and 2001 left that rule alone. The Amendment Bill hands both choices to a simple majority acting by ordinary law: when to redraw, and which census to use.

Second, the Delimitation Bill leaves the schedule to the executive. The Central Government “may, from time to time, by notification, constitute a Commission”, under no trigger and no deadline, and the applicable census is whatever was last published on the day the Commission is constituted: the executive picks the data by picking the day it acts.

Third, the map is closed to review. The Commission’s orders “shall not be called in question in any court”.

Section 16 of the 42nd Amendment itself provided that readjustment “shall take effect from such date as the President may, by order, specify”; the executive clock was wound in 1976, and every successor has kept it running. The 2002 delimitation left out four northeastern states by executive deferment orders; the deferment was rescinded in 2020, and Assam was redrawn in 2023, by the Election Commission under a different statute, on the 2001 census; the Supreme Court declined to stay it. Deferral by notification, revival by notification, route-shopping, antique data: all have a field test and a survival record.

THE VOTES THAT SETTLED NOTHING
On 17 April the amendment fell: 298 in favour, 230 against, fifty-four short of two-thirds, the first constitutional amendment to fail in twelve years. The chamber of 1976 passed its amendment with the opposition in jail; the chamber of 2026, with its opposition free, defeated one.

The guarantee that no state would lose its share of seats appears nowhere in the Bill. The Prime Minister gave it at a Kerala rally; the Home Minister gave it as a quip, offering to return within the hour with an amendment. He never returned. The assurance lived in the speech; the discretion lives in the statute.

This monsoon the government returned with 324 assured votes, still short, and held the legislation back rather than lose twice, in a session that debated only one of its twelve bills. The Emergency chamber legislated without an opposition. This one legislates without a conversation.

THE PUBLICATION-DAY TIME BOMB
Neither vote settled the decisive thing.

The freeze holds only until “the relevant figures for the first census taken after the year 2026 have been published”; the census is under way, reference date 1 March 2027. On publication, the Constitution’s default reasserts itself: delimitation on the new census, by ordinary law, on a simple majority. The courts have declined both to compel delimitation before publication and to accelerate the women’s reservation that waits upon it.

That is the timebomb, aimed at the South, and April’s defeat armed it. The defeated Bill grew the House so every state gained; the default is zero-sum inside the existing ceiling, so Tamil Nadu falls from 39 to 32, Kerala from 20 to 15, and Uttar Pradesh rises to 89. The South cheered a defeat that left it worse off. The government’s shortcut needed a supermajority and failed; the South’s protection, a fresh amendment extending the freeze, needs the same supermajority; the default needs only a publication date, controlled by the Registrar General under a Home Ministry that postponed the last census six years.

In 1976, stopping the clock took theatre: an amendment, an opposition in jail. The theatre left behind the ghost this series is named for, a clause that outlived the Emergency and every government since and decided who represents whom for fifty years after its crisis had ended. The haunting now needs nothing from anyone. The reckoning it deferred for half a century will arrive exactly as it was postponed, without the consent of the governed.

THE ROADS OTHERS TOOK
The United States froze after the 1920 census, when the House refused to reapportion for a decade; the 1929 cure removed Congress’s discretion, and every census since has translated into seats automatically. Washington also pays no state a share of any pool: federal money arrives as per-person program money, with Medicaid’s match rising as a state’s income falls, so a bigger count buys proportional dollars and never a slice of a rival’s. Britain, after blocking three boundary reviews, removed Parliament from the final step in 2020; the next review succeeded. Canada wrote the bargain down: by statute no province falls below its seat floor while the House grows around it, which is Shah’s promise, in force forty years. The bad roads are mapped too. Australia pooled its money centrally and has spent ninety years buying off its donor state, most recently with a guaranteed 70-cent floor one of its own economists calls the century’s worst policy. Nigeria ties both seats and oil revenue to headcount and has not completed an accepted census since 2006: where the count decides the money, every region fights the count. The 2026 package chose discretion where others built automation, an oral promise where others wrote floors, and left untouched the fiscal stakes that Nigeria proves fatal.

WHY THE CLOCK IS WORTH A WAR
The remaining question is why a headcount inspires such fear; the answer is fiscal. On the standard proxy, since India publishes no contribution figures, industrialised states recover roughly 35 to 55 paise per rupee contributed, Uttar Pradesh receives about two rupees, Bihar approaches three, and Karnataka’s Chief Minister claims thirteen paise. For decades an unwritten bargain compensated: the South financed the North and stayed overrepresented. Headcount delimitation tears up one side and keeps the other: taxation with shrinking representation.

The grievance is real. The remedy the South has reached for attacks the seat map instead of the money. Telangana proposes weighting fiscal contribution in delimitation itself; name that for what it is, a contribution-weighted franchise. No democracy gives an industrialist more votes than his driver; a state boundary does not launder the idea. A person in Patna is worth exactly a person in Palakkad. On this narrow point the government’s slogan is correct: one person, one vote, one value.

The point is not that the South should submit. It is that every remedy on offer answers the same premise: Delhi collects the money, and a Lok Sabha majority directs it. The seat map is worth a demographic war only in proportion to what the seats control.

One remedy has gone unproposed: shrink the pool. Canada’s provinces raise close to half of all government revenue, at rates they set themselves; India’s states collect about a third, at rates a Union-chaired council largely fixes, with no income tax of their own, and depend on Delhi’s devolution for much of what they spend. Rebalance that. Raise the states’ share of the divisible pot, and cap the cesses and surcharges that Article 270 keeps outside it, grown from a tenth of gross tax revenue to a quarter in a decade, which shrinks headline devolution of 41 per cent to an effective 30. A Lok Sabha that controlled less could be apportioned honestly: a census would be a count, not a raid, and a state could watch its headcount decline the way it watches its rainfall.

1976, AGAIN
The 1976 chamber gave India two machines in one year: the FCRA, a grip on civil society, and the 42nd Amendment, a clock over the electoral map. Both were built in the same jailed silence, outlived their builders, and were inherited, gratefully, by governments sworn against their origins.

The freeze will now end because the calendar requires nobody’s consent. India can fight over the map, with no answer both democratic and federal at these stakes, or shrink the stakes and return the money; the fear will leave with it.

The lesson of Part 1 was that the machinery remains when the government changes. The lesson of Part 2 is its corollary. The only machinery that cannot be abused is the machinery that was made smaller.