Eight Hundred and Fifty Seats — What Delimitation Actually Decides

Pritiraj BrahmaFounding Partner, Pritiraj & Partners · 8 August 2026

The Delimitation Bill, 2026 proposes a 850-member Lok Sabha with 272 seats reserved for women. The seat count is the least consequential thing about it.

India is arguing about how many seats each state gets. The harder questions — who draws the boundaries, on what data, and what happens to the assemblies — are being decided almost without argument.

The Delimitation Bill, 2026, introduced in the Lok Sabha on 16 April 2026 by the Minister of Law and Justice, Arjun Ram Meghwal, proposes to raise the strength of the House from 543 to 850 — 815 members elected from the states and 35 from the Union territories. In the debate that followed, the Union Home Minister set out the government's arithmetic: under a 50 per cent expansion model, the southern states' 129 seats would rise to 195, leaving their share of the House at roughly 24 per cent, effectively unchanged. The Law Minister told the House that the results of Census 2027 are not expected before the end of 2027 or the beginning of 2028; the Bill therefore provides for a Delimitation Commission to be constituted ahead of the census, in order to shorten the path to women's reservation, of which 272 seats would be set aside.

That is the entire public argument as it currently stands: how many seats, and for whom. It is the least consequential part of the exercise.

Five findings anchor this analysis:

1. The "loss of seats" fear was always about share, not count. Every published scenario that holds the House at 543 shrinks the South. Analysis circulated by ISAS, drawing on projected 2026 population, finds that reapportioning 543 seats would give Uttar Pradesh eleven more and Bihar ten more, while Tamil Nadu and Kerala each lose eight. Expanding to 848 seats does not fix the imbalance; it changes its expression — Uttar Pradesh gains 63 and Bihar 39, while Tamil Nadu gains ten and Kerala none. Absolute counts rise, relative weight falls.
2. An expansion model buys peace on the count and concedes the ratio. The government's own figures make this explicit: the South's seats rise by 66, its share stays near 24 per cent. That is a defensible political settlement, and it should be argued for on those terms rather than presented as though nothing has moved.
3. The sequencing is the substantive reform. Under Articles 82 and 170, delimitation follows a census. The Bill decouples the constitution of the Commission from the availability of the 2027 count. That is a real departure, made for a real reason — women's reservation is otherwise held hostage to enumeration timelines — but it changes what the Commission works from, and that question deserves more scrutiny than the seat table.
4. The assembly-level consequence is larger than the parliamentary one, and unexamined. Article 170 governs state legislative assemblies. A national delimitation is also a redrawing of every assembly map in the country, with a women's quota applied within it. Almost the entire public debate has been about 543 versus 850. Almost none of it has been about the several thousand assembly constituencies that will be redrawn on the same data at the same time.
5. The freeze being lifted is 55 years old. The three delimitations between 1951 and 1971 took the House from 494 to 543 seats; the count has been frozen since 1972 on 1971 census data. India is not correcting a decade of drift. It is correcting more than half a century of it, in one motion, against a population that has roughly tripled.

The data question nobody is asking

Delimitation is an exercise in dividing a number. Everything downstream — how many seats a state receives, where the lines fall within it, which constituencies are reserved — is a function of which population figure is treated as authoritative.

India's position is unusual. The 2021 census was postponed; enumeration is now scheduled across 2026 and 2027, with house-listing in 2026, the population count in February 2027, and a reference date of 1 March 2027. The Law Minister's own statement is that results will not be available until late 2027 at the earliest. A Commission constituted before that must work from something else: 2011 census figures, projections from the Technical Group on Population Projections, or administrative registers such as electoral rolls.

Each of those choices produces a different map, and each carries a different failure mode. The 2011 base is fifteen years stale and understates exactly the urban growth that most needs representation. Projections are defensible in aggregate and unreliable at the district level, which is the level at which boundaries are actually drawn. Electoral rolls measure enrolment, not population, and enrolment quality varies by state — which means using them rewards the states that manage their rolls best and penalises those with large young or migrant populations.

None of these is disqualifying. All of them are choices with distributional consequences, and a government that wants this settlement to hold should state its choice, publish the underlying dataset, and defend it before the Commission begins work rather than after its draft orders are published. Delimitation orders, once laid, are not justiciable. The moment for contest is now.

Why the migration question will surface next

There is a second-order effect that the seat table conceals. Representation is being reallocated toward states with higher population growth. But a significant part of the working-age population of those states does not live there. Internal migration moves people from the states that gain seats to the states whose relative weight declines — and India's electoral system registers a voter where they are enrolled, not where they were born.

The consequence is a slow mismatch: destination states absorb the fiscal and service burden of migrant populations while their relative political weight falls, and origin states gain weight partly on the strength of people who no longer live within their boundaries and cannot conveniently vote there. This is not an argument against delimitation. It is an argument for taking the enrolment and portability agenda seriously in the same window, because the alternative is a representation settlement that is already drifting from reality on the day it is notified.

What the women's quota actually requires

Reserving 272 of 850 seats is the most consequential representational change India has attempted since universal franchise. It is also an operational problem that has been treated as a drafting problem.

Three questions decide whether the quota produces durable political careers or a rotating carousel of one-term members. First, is reservation fixed for the life of a delimitation, or rotated each cycle? Rotation destroys incumbency — the single most valuable asset a legislator builds — and the evidence from panchayat-level rotation is that it depresses the incentive to invest in a constituency. Second, how does the women's quota interact with existing Scheduled Caste and Scheduled Tribe reservation within the same map? Third, what happens in the assemblies, where the same reservation applies to a far larger number of seats with far thinner candidate pipelines?

Parties that treat this as a compliance exercise in 2029 will field placeholders. Parties that treat it as a recruitment problem starting now — identifying, funding, and giving real work to two to three hundred women candidates across a decade — will hold those seats for a generation. The pipeline takes longer to build than the law takes to pass.

What we would do

For a state government or a serious political organisation, the next eighteen months are the entire window. Five things are worth doing now.

1. Build your own map before the Commission builds it. Any state with competent statistical capacity can model its likely seat allocation and internal boundary redraw under each plausible population base. States that arrive at the Commission with published, defensible numbers will shape the draft. States that arrive with grievance will be answered with the Commission's numbers.
2. Contest the data basis, not the outcome. The outcome follows arithmetically from the base. Argument about the base is technical, respectable, and effective; argument about the outcome after the base is settled is merely loud.
3. Prepare the assembly case separately. The parliamentary map will absorb the national argument. The assembly redraw is where a state's own political geography — urban versus rural weight, district balance, reserved-seat placement — will actually be decided, largely unwatched.
4. Audit your electoral rolls now. If enrolment data enters the process in any form, roll quality becomes seat count. This is unglamorous administrative work with a direct constitutional payoff.
5. Start the candidate pipeline. Two hundred and seventy-two seats is not a quota to be filled at nomination. It is a decade-long recruitment problem, and the organisations that begin in 2026 will be the ones with credible names in 2029.

India is about to redraw the map of its own representation for the first time in fifty-five years. The seat count is the headline. The base data, the assembly redraw, and the reservation design are the settlement. It would be characteristic, and costly, to argue only about the headline until the settlement is beyond argument.