India's Delimitation Dilemma

ELI5/TLDR

Delimitation is the boring word for redrawing political boundaries — how many seats each state gets in Parliament, and how those seats are sliced up inside each state. India is supposed to redo this every ten years after the census, but in 1976, during the Emergency, the government froze the map to the 1971 census numbers and never properly unfroze it. Fifty-five years later, the South has slowed its population growth, the North has not, and so a vote in Bihar today carries roughly half the weight of a vote in Kerala. Modi’s government tried to ram through three bills in April 2026 to unfreeze the map and bolt women’s reservation onto it. The bills failed. India is now one census away from the most consequential political fight of this generation.

The Full Story

Milan Vaishnav at Carnegie sits down with constitutional economist Shruti Rajagopalan for an hour-plus on what they both think is “the political bargain that will set the tone for all other political bargains to come — constitutional, federal, and fiscal.” It is one of those conversations where two people who have written the canonical papers on a topic finally compare notes, and come away admitting the puzzle is bigger than either of them expected.

The constitutional setup

The mechanics first. Article 81 of the Constitution says Lok Sabha seats are to be distributed across states in proportion to population, with the ratio of MP-to-citizen “as practicable as possible” the same everywhere. Article 82 says this should be redone after every decennial census. A delimitation commission — usually chaired by a former Supreme Court judge, with the Election Commissioner and Lok Sabha Speaker’s appointees — does the actual nuts and bolts. Its orders have the force of law and cannot be questioned in court. The current ceiling on the lower house is 550 seats; 543 are filled.

There is also a quota architecture stitched on top. About a quarter of all seats are reserved for Scheduled Castes and Scheduled Tribes, allocated to constituencies with the highest SC/ST populations. And in 2023, Parliament passed the 106th Amendment reserving one-third of all Lok Sabha and state assembly seats for women — but tied the start of that quota to the next delimitation, which is tied to the next census, which has not happened.

The freeze

Here is where the wheel comes off. In 1976, in the middle of the Emergency, Indira Gandhi’s government passed the 42nd Amendment — a wide-ranging power grab that pulled authority away from the judiciary and the states and handed it to the Prime Minister’s Office. One of its many provisions: freeze delimitation to the 1971 census. The freeze was supposed to last 25 years. The thinking was that poorer northern states would catch up — economically and demographically — and the gap would close on its own.

It did not close. By 2001, when the Vajpayee government had to revisit the question, liberalisation had widened the gap further. Southern and western states grew faster economically and dropped fertility further. So Vajpayee — whose third-term coalition leaned heavily on the DMK, Telugu Desam, Shiv Sena — kicked the can again. The freeze was extended to 2026. As a small concession, intra-state boundaries and SC/ST reservations would be recalibrated using the 2001 numbers. Inter-state shares would stay locked at 1971.

“We are apportioning our parliamentary seats based on a census which is 55 years old, which is frankly just extraordinary.” — Shruti Rajagopalan

How lopsided things have actually become

Vaishnav uses the Samuels-Snyder Index — a standard political-science measure of malapportionment that captures the share of legislative seats “out of place” relative to population share. India scores 8 percent. That puts it on the worse end of the federal-democracy spectrum, between Argentina (the worst) and Belgium (the best). The two most underrepresented states are Uttar Pradesh and Bihar. The two most overrepresented are Tamil Nadu and Kerala.

In raw human terms: an MP in Bihar represents roughly 3.1 million people. An MP in Kerala represents about 1.75 million. Almost double the weight, sitting in the same Parliament, voting on the same bills.

Rajagopalan adds a wrinkle the headline number misses. The malapportionment is concentrated in the lower house — the directly-elected one. In most federal democracies it is the upper house that gets the deliberate skew (the US Senate gives every state two seats regardless of size). India’s Rajya Sabha is also proportional to population, which means India scores best in the world (12 percent) on upper-house malapportionment. The Council of States is not really a council of states. Both houses of Parliament essentially mirror the same demographics, and the system has no built-in counterweight to majoritarian outcomes.

There is also a fertility wrinkle. Poorer groups — by class, caste, and youth — saw their fertility rates drop later. So the malapportionment penalises not just the South but also SCs, STs, and young Indians disproportionately, all of whom are over-represented in the under-counted northern states.

What Modi tried in April 2026

In late April, the government called a special session of Parliament with very little notice and tabled three bills.

The 131st Constitution Amendment Bill would have raised the Lok Sabha ceiling from 550 to 850 seats — 835 for states, 15 for union territories. It would have removed the constitutional language tying delimitation to the most recent census, replacing it with a freshly-empowered Delimitation Commission that could decide which census to use. A companion Delimitation Commission Bill specified the 2011 census. A third bill handled the union territories. The whole package was bolted to women’s reservation — the 1/3 quota would only kick in once this delimitation was complete.

The constitutional amendment failed the two-thirds threshold. The other two bills, which only need simple majority, were never put to vote because the constitutional one fell.

Rajagopalan reads the bills carefully and flags two quiet provisions that almost slipped past. First, the bill shifted the choice of census from the Constitution itself (hard to change) to a statute (changeable by simple majority) — meaning the most consequential political question would now be settled by ordinary politics, not constitutional principle. Second, the new delimitation orders would explicitly be immune from court challenge: “shall not be called in question in any court.”

“We’ve basically really switched who makes this decision from a constitutional mandate to more of a political decision-making.” — Shruti Rajagopalan

The puzzle of intent

This is where both speakers admit they cannot read the government’s mind. The bill said one thing — population-based reapportionment, full stop. The Home Minister and PM said the opposite, in speeches: don’t worry, we will not touch state-wise shares. Amit Shah at one point offered to leave the chamber and come back in an hour with an amendment hard-coding that promise.

Vaishnav and Rajagopalan walk through the candidate explanations. One: pure trolling. The government knew it lacked a two-thirds majority, knew the bill would fail, and used the failure to paint the opposition as anti-women’s-reservation. Two: narrative-building heading into Tamil Nadu and West Bengal state elections, with women as the wider tent the BJP can pitch in southern states where caste and language pitches do not land. Three: real intent to reapportion, with the empowered Delimitation Commission acting as a politically pliable instrument that the executive could lean on later. Rajagopalan notes the third option is “constitutionally and legally dubious” — which is exactly why the bill embedded the no-court-challenge clause. They expected to be sued.

Neither finds any of these fully convincing on its own. The most persuasive composite is that the government was preserving negotiating flexibility — testing how hard the South would push back, whether allies like Chandrababu Naidu would stay (he did), whether the opposition would hang together. Useful intelligence, even from a defeat.

The numbers, if the freeze actually breaks

Vaishnav, working with Louise Tillin and Andy Rubinas, has done the arithmetic on what reapportionment would actually look like. Two baselines: the 2011 census, or projected 2026 numbers.

Using 2011 with the existing 543 seats: Tamil Nadu loses 7, Kerala loses 5, Andhra Pradesh loses 3. UP gains 8, Bihar gains 6, Rajasthan gains 5. Only Assam and Chhattisgarh stay put.

Using 2026 projections, the numbers grow. Tamil Nadu loses 10. Kerala loses 6. Andhra loses 5. UP, Bihar, Rajasthan together gain 26 seats. Twenty-one seats migrate just from the three southern states.

Why the gainers are silent

A small but telling observation from Rajagopalan: the chief ministers of UP, Bihar, MP, Rajasthan have not spent much oxygen building the “we are getting ripped off” narrative. Why? Because the states that would gain seats are also the ones receiving the largest fiscal transfers from the South. India’s Finance Commission devolves 41 percent of the divisible tax pool to states, and the horizontal split between states uses weights that lean hard toward poverty and population. The same demographic facts that under-represent the North in Parliament over-allocate fiscal transfers to it. A loud “one person, one vote” campaign would invite a louder “one taxpayer, one rupee” counter-campaign. So the gainers stay quiet.

The conversation closes with the actual options. Extend the freeze again — kick to 2051, the lazy default. Move to full population-based reapportionment, either at current size or expanded. The “moderated expansion” idea Vaishnav has written about — expand the Lok Sabha just enough that no state loses an absolute seat. The math comes to about 775 seats. Hybrid formulas: degressive proportionality (more populous states get more seats but with diminishing returns), floors below which no state falls, blended census formulas to smooth the transition.

Rajagopalan keeps pulling the conversation back to the fiscal layer. The malapportionment is a constitutional problem nested inside an economic problem. India in 1976 was socialist but not yet a welfare state — it lacked the revenue to be one. After 1991 it became one, and the political stakes of “who controls Parliament” exploded because Parliament now decides who gets the welfare transfers. If the union devolved 90 percent of revenue to states instead of 41, the South would care a lot less about losing seats. The constitutional bargain is unsolvable until the fiscal bargain is renegotiated.

She also flags the Rajya Sabha — which everyone ignores in this conversation but which is the obvious lever. Make the upper house an actual council of states. Give each state equal weight (with some adjustment for joint sittings, and with constitutional locks on creating new states by voice vote, which India has done before). Reinstate the domicile requirement for Rajya Sabha MPs (gutted by a 2003 amendment). And — the reform Rajagopalan wants most — make money bills require Rajya Sabha passage. Right now a money bill bypasses the upper house entirely, which is why the government rammed Aadhaar through as a money bill. Without that, no fiscal-federal bargain has teeth.

The 10th Schedule (anti-defection law) sits behind all of this. MPs cannot defy a party whip without losing their seats, so even reformed, better-apportioned representatives cannot actually vote their states’ interests in Parliament. The bottleneck is upstream of seat counts.

Key Takeaways

Delimitation = redistricting. Two linked steps: apportionment (how many seats each state gets) and boundary-drawing (how those seats are carved up inside the state). Article 81 sets the proportional rule; Article 82 sets the post-census timing.
The 1976 freeze was an Emergency-era 42nd Amendment provision. It locked Lok Sabha state shares to the 1971 census. The 44th Amendment undid most of the 42nd’s damage but left this in. Vajpayee’s third government extended it 25 more years in 2001. We are now at the 55-year mark.
A vote in Bihar is worth roughly half a vote in Kerala. 3.1 million citizens per Bihar MP versus 1.75 million per Kerala MP.
India scores 8 percent on the Samuels-Snyder Index of malapportionment. Worse end of the federal-democracy spectrum (Argentina worst, Belgium best). The skew sits in the lower house, which is unusual — most countries skew the upper house deliberately.
The Rajya Sabha is the least-malapportioned upper house in any federal democracy (12 percent). Both Indian houses mirror population, so there is no constitutional counterweight to majoritarian outcomes — unlike the US, where the Senate’s two-per-state design balances the House.
The Lok Sabha is constitutionally capped at 550 seats; 543 are filled. The cap can only be raised by constitutional amendment.
Modi’s April 2026 push: three bills. 131st Amendment to raise the cap to 850 seats, a Delimitation Commission Bill specifying the 2011 census, and a UT bill. The constitutional amendment failed the 2/3 threshold; the other two never reached a vote.
The hidden moves in the bill: shifting the census choice from the Constitution to a statute (changeable by simple majority), and explicit immunity from court challenge.
Women’s reservation (106th Amendment, 2023) was deliberately bolted to delimitation. This was likely a backroom deal — without that linkage, the constitutional amendment would not have passed. The men in Parliament wanted a delay mechanism.
If reapportionment happens at current size (543 seats) using 2026 projections: Tamil Nadu -10, Kerala -6, Andhra -5; UP, Bihar, Rajasthan gain 26 between them.
Moderated expansion option: ~775 seats, set so no state loses any absolute seats. Same math the 1971 delimitation used.
Why the would-be winners (UP, Bihar) are politically silent: they receive the bulk of fiscal transfers from the South. Loud demand for “one person, one vote” would invite a counter-demand for “one taxpayer, one rupee.”
41 percent of the divisible tax pool goes to states. The horizontal split uses poverty and population weights — the same demographics that drive the malapportionment, mirrored in the fiscal formula.
The fiscal bargain and the constitutional bargain are joined. If India devolved revenue closer to 90/10 like other federations, the seat fight would lose most of its temperature.
The Rajya Sabha is broken in three specific ways: it mirrors population instead of representing states; the 2003 amendment killed the domicile requirement; and money bills bypass it entirely (Article 109), which the government exploited to pass Aadhaar.
The 10th Schedule (anti-defection law, 1985) means MPs cannot vote against their party whip without losing their seats. Even better-apportioned MPs cannot actually represent state interests inside Parliament.
Census 2027 is the trigger. It formally started 1 April 2026. Once results are published, the existing constitutional framework — population-based, intra and inter-state — will fire automatically unless the government negotiates a new compromise first.

Claude’s Take

This is the cleanest treatment of delimitation I have seen in a single hour. Vaishnav and Rajagopalan are both deep on the substance, they push each other in real time, and neither tries to dunk on the BJP — which would be the easy move. The asymmetry is left to do its own work, which is the only honest way to discuss this.

The thing they get most right is treating the constitutional question as inseparable from the fiscal one. Almost every popular take on delimitation frames it as North-versus-South culture war. It isn’t. It is a fight about who controls the welfare-transfer machinery in a country that became a serious welfare state only after 1991. Once you see that, the BJP’s strange behaviour — pushing a bill the home minister immediately walked back in speeches — starts to make more sense as a positioning move rather than a coherent policy push. Rajagopalan’s framing, that we have been in a constitutionally dubious place for 50 years already and the only real change is that this generation gets to make the next compromise instead of inheriting one, is the kind of dry honesty you don’t get on most podcasts.

The one thing the conversation doesn’t fully reckon with is what happens if the freeze just gets extended again — which Vaishnav notes is probably the most likely outcome. They treat it as the boring default but don’t dwell on what the cost is of running a democracy where a 60-year-old census still apportions political voice. Some of that is probably running out of time; some is that the answer is uncomfortable for everyone.

Score: 9. The half-point I’d shave off is that the audience-question section toward the end gets a bit scattered, and the Rajya Sabha reform menu deserved a full episode of its own. But this is essential listening on the most important political question India will face in the next five years.