The Feasibility Audit

1What exactly are we auditing?

Before you can test a proposal you have to know what it is. Part 1 showed that four different places share the name Punjab. A separate state built on each of them is a completely different object, with different rivers, different borders and a different population. So the first job is to pick one — and I pick the one that helps the proposal most.

The three candidate territories

Version C deserves one further sentence because it is quietly fatal, and because supporters and opponents both avoid it. Pakistani Punjab holds well over a hundred million people and effectively no Sikhs. A state combining both Punjabs would be a Muslim-majority country in which Sikhs were a tiny minority. It would not be Khalistan in any sense its supporters intend.

The other assumptions I am giving away

Every one of those is unrealistic in the direction that helps. If the audit fails with all five granted, the failure is structural — it is about the ground, the water and the arithmetic, not about politics or bad luck.

2Water: the chapter that ends the argument

Punjab means five rivers. Here is the fact that decides this entire part: not one of the three rivers Punjab now has begins in Punjab, and neither of the two dams that control them stands in Punjab. A separate Punjab would own the fields and not the tap.

Where the water actually comes from

Part 1 explained the rivers. Part 8 explained the allocation fight. This chapter is about something narrower and more brutal: physical control.

Read the third column again. Bhakra is in Himachal Pradesh. Pong is in Himachal Pradesh.

These are not minor structures. They are the reason Punjab has water in the dry months at all. They store the monsoon and the snowmelt and release it across the year. Without regulated release, the rivers are seasonal — heavy in summer flood, thin the rest of the time. That was Punjab's condition for most of its history and it is why Part 1 called the five rivers "the machine that made Punjab, and the trap inside it."

The demonstration that removes all doubt

Here is where the audit stops being theoretical.

In April 2025, following the Pahalgam attack, India placed the Indus Waters Treaty in abeyance. India's position, restated repeatedly through 2026, is that the treaty will remain in abeyance until Pakistan credibly ends support for cross-border terrorism, and that it cannot function in its present form. Pakistan's position is that the treaty remains fully operational and that abeyance is not a thing the treaty provides for.

The legal proceedings went against India. In 2026 the court of arbitration held that the treaty does not permit unilateral abeyance and reaffirmed its own jurisdiction; a further award concerned limits on India's water-control ability. India rejected the awards and called the court illegally constituted.

I am not adjudicating that dispute here. I am extracting one finding from it, and it is the most important finding in this part.

This is not a prediction that India would cut the water off. It is a statement about who would hold the decision. In an audit, that is the whole point: a critical input controlled entirely by another party is not a risk to be managed, it is a dependency that defines the enterprise.

The fallback is already failing

A supporter might answer: forget the rivers, Punjab runs on tubewells anyway. That is substantially true and it is the worst possible answer.

A stage of extraction of 156% means Punjab takes out roughly half again as much water each year as nature puts back. That is not agriculture. That is a withdrawal from a savings account with no deposits.

The slight improvement since 2023 is real and I record it honestly — it is what expanded canal irrigation does. But note what that improvement depends on: canal water, which comes from the dams in Himachal. The fallback improves only when the dependency deepens.

The steelman on water, and the answer

The answers, in order.

One. The comparison to Egypt and Bangladesh cuts the wrong way. Both are examples of downstream states in permanent, unresolved anxiety about upstream construction, and neither has ever managed to compel an upstream neighbour to do anything. They are illustrations of the problem, not solutions to it.

Two. Leverage over Pakistan is not leverage over India. The upstream card is played against the wrong player. A state can be simultaneously a bully to its south and helpless to its north — and that describes exactly the position on offer.

Three, and this is decisive. The appeal to international law was tested in 2025 and 2026 and failed in front of us. A court of arbitration ruled, and the upstream state rejected the ruling and continued. That is the single most recent, most relevant and most damaging precedent available, and it happened on these exact rivers.

3Landlocked: the two roads out

Punjab has no coast. It never has. For a farming economy that must move tens of millions of tonnes of grain, that means everything depends on the road to a port — and there are exactly two, one through the country you have just left and one through the country that armed your insurgency.

The geography, stated flatly

Present Indian Punjab is a landlocked block of about 50,362 square kilometres. Its neighbours are Jammu and Kashmir to the north, Himachal Pradesh to the north-east, Haryana to the south-east, Rajasthan to the south-west, and Pakistan to the west.

Four of those five are India. The fifth is Pakistan.

So the audit finding on supply lines is short. A landlocked Punjab's access to the sea would run either through the state it seceded from, or through the state that spent a decade supplying weapons to its insurgency — and the second route has been closed for seven years already, while both parties were still notionally at peace.

Why this is worse than it sounds

Landlocked is not automatically fatal. What matters is the quality of the neighbours.

Landlocked and rich, landlocked and poor

The variable is not the sea. It is whether your neighbour wants you to succeed.

The border itself

One further item belongs here rather than in Chapter 7. Punjab's western frontier with Pakistan is an international border of roughly 550 kilometres — the figure commonly cited for the Punjab sector, though different sources measure the line differently.

Today that border is held by the Indian Army and the Border Security Force, at Indian expense. On day one of independence it becomes a border a new state must hold at its own expense, on the far side of which is the country whose intelligence service, on the documented record of Part 10, armed the insurgency that killed thousands of Punjabis.

4The books: what Punjab earns and what it owes

This is the least emotional chapter in the whole work. It is a state's balance sheet, taken from its own budget documents, and it says that Punjab today does not cover its own running costs.

The headline numbers

All figures are from the Punjab Budget for 2026-27, presented on 8 March 2026, as analysed by PRS Legislative Research, and from the Punjab Economic Survey 2025-26.

Three of those lines matter more than the rest, and I want to take them one at a time in plain words.

The line that ends the chapter

Now the item that an audit exists to find. Punjab's revenue receipts for 2026-27 are estimated at Rs 1,26,190 crore. Of that:

Roughly thirty-one paise in every rupee Punjab spends arrives from Delhi.

On the first day of independence, that thirty-one paise stops. Not gradually — on the day. Simultaneously, a series of costs that Delhi currently carries arrive on Punjab's books: defence, foreign affairs, the currency, the central share of major schemes, and debt servicing on whatever share of India's liabilities is negotiated across.

So the arithmetic of day one is: revenue falls by about a third, and expenditure rises. Starting from a position that already runs a revenue deficit and carries debt at 45% of output.

The subsidy that cannot be cut and cannot be kept

One item deserves separate treatment because it sits exactly where the economics and the water meet.

Fourteen lakh tubewells running on free electricity is precisely the mechanism producing the 156% extraction rate in Chapter 2. The subsidy and the water crisis are the same policy seen from two directions.

The steelman on the economy, and the answer

The answers.

One. The per capita comparison is correct and is the proposal's best number. But Punjab's lead over the Indian average is now small and shrinking — a few percent, on slower growth. The trend line crosses. That is a fact about the last two decades and it is Part 13's subject.

Two. The claim that the transfer is Punjab's own money returning is partly true and does not survive contact with the day-one problem. Even if every rupee of it is Punjab's by right, the mechanism that collects and returns it would cease to exist. A new state would have to build a revenue administration capable of collecting what the Government of India currently collects, immediately, during a secession.

Three. Bangladesh is a serious analogy and I answer it properly in Chapter 9, where it belongs. The short version: Bangladesh had a port.

5The grain trap: who buys the harvest

Punjab's agriculture is not a market. It is a purchasing arrangement with a single customer, at a price that customer sets, for a crop chosen to suit that customer's needs. The customer is the Government of India, and on the first day of independence it stops being obliged to buy anything.

The size of the arrangement

In the 2025-26 wheat season, India procured 256.31 lakh tonnes of wheat at Minimum Support Price. Punjab supplied 103.89 lakh tonnes of it — roughly two fifths of the national total, from a state with about 2% of India's population and 1.5% of its land.

In the 2026 season Punjab was reported as set to procure about 132 lakh tonnes against a target of 122 lakh tonnes. The Minimum Support Price for wheat in 2025-26 was Rs 2,585 per quintal. The national MSP outflow for wheat that season was around Rs 62,156 crore across 21.03 lakh farmers.

Paddy runs on the same machinery, with the same buyer, at a separately announced MSP.

What happens on day one

I want to be careful here, because this is a place where it is easy to overclaim. I am not saying Punjabi wheat is bad or that nobody would buy it. I am making a narrower and firmer point about cost structure.

Punjab's farm economy has been built, over sixty years, around a guaranteed administered price. The tubewell, the free power, the input costs, the land prices and the farm debt are all calibrated to it. An economy calibrated to a guaranteed price does not adapt instantly to a variable world price, and it certainly does not do so while also absorbing a 31% revenue collapse and building a water treaty from nothing.

6Who is inside: the demographic problem nobody will name

Khalistan means land of the Khalsa. At the 2011 Census, about 38 in every 100 people in Punjab were Hindu. This chapter is about that sentence, and about the fact that in fifty-five years of argument nobody has published an answer to it.

The numbers

And the state is not uniform. The picture by region matters more than the state total.

The question, asked directly

A state whose name and constitutional premise are religious would contain, on day one, over eleven million people who do not belong to that religion, concentrated in a whole region and a majority of at least one border district.

So: what is the proposed status of Punjab's Hindus in Khalistan?

I have looked. There is no published answer. There is no draft constitution, no citizenship provision, no minority rights guarantee, no proposed franchise. Part 11 established that the movement has produced passports, stamps, currency and a flag over forty-six years. It has not produced a clause on who counts as a citizen.

The other half of the demography

The problem runs in the opposite direction too, and it is the one supporters find hardest.

Large Sikh populations live outside Punjab — in Delhi, Haryana, Rajasthan, Uttar Pradesh, Maharashtra and elsewhere across India. Part 10's central atrocity, the pogrom of November 1984, did not happen in Punjab. It happened in Delhi.

A Sikh state carved out of Punjab does nothing for those communities. It would, on any realistic reading, make their position considerably worse: they would become a minority identified with a hostile neighbouring state, in a country that had just lost territory to it.

That is not a rhetorical point. It is the arithmetic. A state that cannot protect the specific people whose killing is its central grievance has a design fault at the level of purpose.

7Who holds the border

Punjab currently has one of the best-defended frontiers in the world and pays nothing for it. A separate Punjab would have to build an army from zero, on a border with a nuclear-armed state, using soldiers it would first have to ask to change sides.

The cost that appears from nowhere

Defence, foreign affairs and currency are the three functions a state cannot outsource and cannot avoid. Punjab currently performs none of them. They are carried by the Union of India and do not appear anywhere in the Punjab budget analysed in Chapter 4.

None of these can be phased in. A state is either defending its border on day one or it is not.

The soldiers

Here is where an honest audit has to record a genuinely disputed number.

Now apply that to the audit.

The professional military manpower a new Punjabi state would need is currently employed by the Indian Army, has served in it in some cases for generations, and in many families across several. A secession would require a very large number of those men to change allegiance — not as a matter of sentiment but as a matter of staffing a defence force.

Part 6 examined exactly this question at Ghadar in 1914-15, when a diaspora revolutionary movement assumed Punjabi soldiers would come over. They did not; many families of serving soldiers turned the revolutionaries in. Part 10 recorded what happened after Blue Star, when some units did mutiny: the mutinies were localised, were suppressed, and did not spread.

The steelman on defence, and the answer

The answer is in two parts, and the second is the one that matters.

First, the Israel comparison omits the decisive variables: a coastline, external great-power support, a population arriving with nowhere else to go, and no dependency on the adversary for water or food purchase. Punjab has the reverse of all four.

Second, and this is the sharper point: the argument that Punjabi soldiers are the backbone of the Indian Army is simultaneously the argument that Punjab is deeply integrated into the Indian state at its most sensitive level. You cannot invoke that integration as a source of strength for separation. It is the strongest single piece of evidence against the premise that Sikhs cannot be safely Indian — which is the proposition Part 11 identified as the movement's necessary foundation.

8The precedent: the only Punjabi state that ever existed

There is exactly one case study available. Part 5 examined it in detail. A sovereign Punjabi state existed, it was strong, it was well run, it was rich — and it destroyed itself in the ten years after its founder died, before any foreign army defeated it.

What the record shows

Part 5 audited Ranjit Singh's empire honestly, including the parts that flatter it. It was militarily formidable. It was not a theocracy — it employed Muslims and Hindus, Europeans and Americans, at the highest levels. It governed a genuinely plural population with a lighter hand than most contemporaries. On its own terms it worked.

And then Ranjit Singh died in 1839, and within ten years the state was gone.

Part 5 called this "the audit: what the Sikh Empire proves about Khalistan," and deferred the conclusion to here. This is the conclusion.

The finding

The single most common argument for Khalistan's viability is that a Punjabi sovereign state has existed before and can exist again. That argument is available, and it cuts the way its users do not intend.

The proposed state's governing institution, unexamined

One more item, which is uncomfortable and belongs in an audit.

Any religiously defined Punjabi state implies a relationship between religious authority and civil government. Part 4 traced the doctrine of Miri-Piri — temporal and spiritual authority together — and noted the question it leaves open. Part 8 examined the Anandpur Sahib Resolution, which was a document about federal autonomy rather than theocracy.

But there is a live institution here: the Shiromani Gurdwara Parbandhak Committee, the elected body that manages the historic gurdwaras. Its elections have been irregular. Its control has been the subject of prolonged political capture. Part 11 showed the same pattern in North American gurdwara committees, where control of a management committee delivers both money and legitimacy.

An audit does not need to decide whether religious authority in a Sikh state is desirable. It needs to observe that the question of what body holds ultimate authority has not been answered, that the most likely candidate institution has a contested governance record, and that the historical precedent in Chapter 8 failed precisely on this class of question.

9The case against this chapter, at full strength

Rule 3 of this work says every side gets its best case, not its worst. This chapter is the strongest argument against everything I have just written, put the way its cleverest defender would put it. I have tried to write it well enough that it hurts.

My answers

That is the best case and parts of it land. I am going to answer all four, and concede where I should.

On point one — the new-state objection. Partly conceded, and it is the strongest of the four.

It is true that feasibility audits have a poor record against determined national movements. It is true that Singapore, Israel and Bangladesh all existed against the arithmetic.

But look closely at the examples, because they are not interchangeable.

So the concession is this: a feasibility audit cannot prove a state will fail, and I do not claim it can. What it can do is identify which constraints are binding. In every successful precedent offered, the binding constraint was solved by a specific asset — a port, an ocean, an alliance, a demographic supermajority. Punjab has none of the four, and its binding constraint — upstream water controlled by the adversary — is not one any of the four faced.

On point two — the wrong counterfactual. Substantially conceded.

This is a fair hit and I accept it. I have audited the proposal and not the status quo. The status quo is genuinely failing, and Part 13 audits it with the same instruments and no more mercy. If Part 13 finds that Punjab inside India is on a trajectory to ruin, that finding stands alongside this one and does not cancel it.

But two things can both be unworkable. "The current arrangement is failing" is an argument for changing the current arrangement. It is not, by itself, an argument that any particular alternative would work — and the alternative on offer makes the water position worse, not better.

On point three — dependency as the product of subjugation. Partly conceded, and then answered.

The historical claim is largely right. Punjab's cropping pattern was a national policy choice. The industrial thinness is partly a licensing legacy. Part 8 documented the water decisions. I am not disputing the mechanism.

But the audit answer is about physics, not history. Bhakra is in Bilaspur and Pong is in Kangra because of where the Himalayan foothills are. No colonial policy put them there. If every extractive decision Delhi ever made were reversed tomorrow, the Sutlej would still rise in Tibet and enter India through Himachal. That constraint is geological, it predates the Indian state, and it would outlast it.

On point four — that this is a counsel of submission. Rejected, and here is why.

The objection assumes that if secession fails the audit, the grievance has no remedy. That is the movement's own framing and it is false. It treats one proposed remedy as the only one.

Chapter 10 sets out what the audit actually implies, and it is not submission. It is that a grievance this well documented deserves a remedy that would work, and that the one on offer would leave Punjab poorer, thirstier and less safe than it is now — which is not a defence of Delhi. It is the reason to demand something better than a flag.

10The verdict

The audit is finished. Here is what it found, stated once, without hedging, and then here is exactly what that finding does and does not license anyone to say.

The ledger

The verdict, stated once

I want to be precise about the strength of that claim, because precision is the only thing that makes it worth anything.

I am not saying Khalistan is impossible. Chapter 9 conceded that feasibility audits cannot prove impossibility and have a poor record trying.

I am saying something narrower and firmer: the proposal has a binding constraint it cannot solve, and it has never attempted to solve it. The binding constraint is that the water is controlled from outside. Everything a Punjabi state would need to survive — irrigation, drinking water, the crop, the export earnings, the tax base — passes through that single gate, and the gate is in Himachal Pradesh.

In fifty-five years of pamphlets, declarations, passports, flags, referendums and lobbying, nobody has published a page explaining how a separate Punjab gets water. That is the finding of this part.

What this does not mean

Rule 5 of this work says nobody is only a hero or only a villain, and it applies to conclusions as much as to people. So here is the disciplined statement of what the audit does not license.

Reading the verdict correctly

The sentence this part exists to make possible

Fifty-five years of argument have run on a bad trade. One side has said: the grievance is a fabrication, so the demand is treason. The other has said: the grievance is real, so the demand must be right.

Both sides need those two things welded together. This part is the chisel.

11Scorecard: what Part 12 proves

Timeline: how the constraints were built

This part has no narrative of events. It has a structure — and the structure was assembled by a series of decisions, most of them taken without Punjab's consent, and most of them now irreversible.

What happens in Part 13

Chapter 9's second objection was the one I conceded most fully: you audited the proposal and not the status quo. Part 13 audits the status quo, with the same instruments and no more mercy.

It begins where Part 10 ended — with a decade of capital flight, halted investment and a lost generation — and runs to the present.

• The Green Revolution's invoice, now due. The aquifer at 156% extraction. The wheat–paddy cycle nobody can break. Stubble burning and what it actually costs.
• Farm debt and farm suicides — the numbers, the disputes about the numbers, and what the data will and will not support.
• The drug crisis — evidence against moral panic, and the difference between the two.
• The emptying. Punjab's outward migration, and the sharp reversal since 2024 as Canada tightened. Roughly 1.8 lakh students went to Canada from Punjab in 2023; India's share of Canadian student intake fell from 51.6% in 2023 to 33.6% in 2024 to 8.1% in 2025. An economy built on the assumption that the young leave has just had that assumption withdrawn.
• Politics after 1995 — the Akali collapse, the AAP sweep, and the 2027 election.
• The final ledger for the whole series. Every major claim across thirteen parts, in one table, tagged.

Part 12 asked whether leaving would work. Part 13 asks the question that actually determines Punjab's future, which is what happens if nothing changes.