What Pakistan Is For
On the military, the political elite, and a state that has become an instrument of its owners.
I.
The police post that was overrun in Ziarat on the night of July 6 was guarding a dam. Not a garrison, not a cantonment, not a nuclear facility. A dam. The Mangi Dam project has been under construction in the hills east of Quetta for years, and when it is finished it is supposed to pipe water into a provincial capital whose aquifer has been failing since the 1990s, a city where families with money sink private bores four hundred feet and families without money buy tankers from men who sell them what the bores took. The policemen at that post were not defending the Pakistani state’s sovereignty over Balochistan. They were defending a construction site that might one day allow the residents of Quetta to turn on a tap.
Eighteen policemen and eleven soldiers were killed across Balochistan between July 4 and July 8, along with four civilians, according to the military’s own spokesman, Lieutenant General Ahmad Sharif Chaudhry, who put the total at forty-two and said security forces had killed fifty-four fighters in response. By the end of that week the announced figure for insurgents killed had risen to seventy-five. The Prime Minister flew to Quetta and told the families that the deaths had not been in vain. Two weeks later, in Tank district in Khyber Pakhtunkhwa, four suicide bombers drove an explosives-laden vehicle into the perimeter wall of a joint army and police checkpost at Manji Khel and killed twelve soldiers, two policemen, and a retired forest department official. The Pakistan Institute for Conflict and Security Studies counted 2,166 deaths from militancy and counter-militancy operations in the first six months of this year. The Center for Research and Security Studies found that in the second quarter alone, Khyber Pakhtunkhwa accounted for 475 of 774 deaths nationally, a jump of fifty-three percent over the previous three months, which is how a province becomes the worst-affected in the country while the national figures are described as improving.
Here is the fact worth sitting with. The Mangi Dam has been under construction, in one form or another, for the better part of a decade. In November of last year, the Parliament of Pakistan passed the Twenty-Seventh Constitutional Amendment, fifty-nine clauses restructuring the command of the armed forces, the architecture of the judiciary, and the reach of criminal law, from introduction in the Senate to presidential assent, in four days. The Senate cleared it on the tenth. The National Assembly cleared it on the thirteenth, 234 votes to four. President Asif Ali Zardari signed it the same day.
The Pakistani state is not incapable. It is extremely capable, at speed, when the thing being built is a wall around the people who own it.
II.
The standard defense of Pakistan’s military establishment has always contained an implicit bargain: the civilians are corrupt and incompetent, the generals are disciplined and clean, and if the country would only stop pretending at democracy, it could get on with the business of building roads and power plants and industrial capacity. It is worth noticing that this argument has already been tested, at length, under favorable conditions, and that we know the result.
Ayub Khan took power in October 1958 and held it for a decade. He had American money, Cold War strategic value, a compliant bureaucracy inherited from the Raj, and no parliament to answer to. The growth figures of the 1960s were the best in Pakistan’s history and among the best in Asia: real output expanding at roughly six percent a year, large-scale manufacturing faster than that, the Green Revolution arriving in the Punjab with tube wells and Mexican wheat. Mangla Dam was completed in 1967. Tarbela was begun. Islamabad was laid out by the Greek planner Constantinos Doxiadis on a grid of numbered sectors, and South Korean planners came to study the Pakistani five-year plans. This was the decade the establishment still points to when it wants to argue that the country works better under a uniform.
Then, in April 1968, in the last months of the celebration the government had branded the Decade of Development, the chief economist of the Planning Commission stood up at a meeting in Karachi and said what the growth had actually produced. Mahbub ul Haq’s finding was that twenty-two industrial family groups controlled around two thirds of the country’s industrial assets, four fifths of its banking, and the bulk of its insurance. He had helped design the model. He was describing its output. The growth had been real and the concentration had been the point: the plans had deliberately favored capital accumulation in a small number of hands on the theory that the wealth would be reinvested rather than consumed. The wealth was reinvested. It was reinvested in the families that already held it.
Three years later the country broke in half. The immediate cause was the refusal to transfer power to the Awami League after it won 160 of 300 general seats in the December 1970 election. The deeper cause was the arithmetic of the preceding twenty-three years, in which the eastern wing supplied the majority of the population and the bulk of the export earnings from jute and received a minority of the development spending, the military commissions, and the civil service appointments. When the army moved on Dhaka on the night of March 25, 1971, it was not defending the country against a foreign enemy. It was defending a distribution. The commission that Pakistan appointed to inquire into the defeat, under Chief Justice Hamoodur Rahman, delivered its findings in 1972 and a supplementary report in 1974. The state suppressed both for nearly three decades.
So the case for military stewardship has had its trial. Given a decade of unbroken authority, foreign patronage, and a growth rate the country has never matched since, the establishment produced an economy owned by twenty-two families and a war that cost it the majority of its citizens. The failure was never technical. The generals could build. The question was always what, and for whom.
III.
To understand why Pakistan’s ruling arrangement behaves as it does, it helps to abandon the idea that the arrangement was ever interrupted. Pakistanis are taught a periodization of coups: 1958, 1977, 1999, with civilian interregnums in between. The periodization is misleading. What the country has experienced since 1947 is closer to a continuous occupation with variable costuming.
The founding conditions set it. Muhammad Ali Jinnah died in September 1948, thirteen months after the state existed. Liaquat Ali Khan was assassinated in Rawalpindi in October 1951. The first Constituent Assembly took nine years to produce a constitution, which the President and the army chief abrogated two years later. In that vacuum, the two institutions that had come through Partition intact and organized, the army and the civil bureaucracy, absorbed the functions that elected politicians were too fragmented to hold. By 1954 Pakistan had signed a mutual defense assistance agreement with Washington and joined the alliance structures that would supply it with equipment and cash for the next twenty years, which meant that the army acquired an independent foreign revenue stream before the state had acquired a constitution.
The territorial logic followed. Kalat acceded to Pakistan in March 1948 under conditions the Baloch have contested ever since, and the first armed revolt began within months. There have been four more since, in 1958, in 1962, in the mid-1970s when Zulfikar Ali Bhutto dismissed the elected provincial government and sent in the army, and the one that has been running since 2004. In 1955 the government dissolved the provinces of the western wing into a single administrative unit, One Unit, for the express purpose of manufacturing parity with the Bengali majority in the east. The Sui gas field in Dera Bugti district began commercial production in the mid-1950s and piped gas to Karachi, Multan, and Lahore. Large parts of the district that sits on it were still cooking on wood a half century later.
And in the early 1950s, from a fund the departing British had left for the families of Indian Army veterans, the Fauji Foundation was set up as a charitable trust. It is now the largest business group in Pakistan.
IV.
There is a habit among foreign analysts of describing the Pakistan Army’s commercial interests as a distortion of its professional role, as though a well-intentioned institution had wandered into cement and fertilizer by accident. It is more accurate to say that the commercial interests are the institution’s actual business and the professional role is the collateral it borrows against.
When the government was asked in the National Assembly to account for it, the figure that came back was that the army runs business entities worth in the region of forty billion dollars, making it by a wide margin the largest commercial house in a country where a quarter of the population cannot reliably eat. There are roughly fifty projects, industrial units, and housing colonies under the administrative control of the Fauji Foundation, the Shaheen Foundation, the Bahria Foundation, the Army Welfare Trust, and the Defence Housing Authorities. The Fauji group alone holds Askari Bank, Fauji Fertilizer, Fauji Cement, Fauji Foods, and a stake in Mari Petroleum. Nine other military-linked entities are individually valued above a billion dollars. The Defence Housing Authorities have made the armed forces the largest land developer in the country. These enterprises enjoy tax treatment, access to state land, and regulatory shelter that no civilian competitor can obtain, and their boards are staffed by the retired officers whose careers they are designed to conclude.
The land is the older instrument and the more important one. The British allotted irrigated land in the canal colonies of the Punjab to loyal recruits, and the practice was inherited rather than abolished. A Pakistani officer today retires with a rank, a pension, a plot, and frequently a seat on a board or an ambassadorship. Under the Special Investment Facilitation Council, the hybrid civil-military body created in 2023 to bypass the ordinary machinery of government, the state identified some 4.8 million acres of public land for lease to a military-run company for corporate agriculture under the Green Pakistan Initiative. A 2,250-acre corporate agriculture park was inaugurated at Pirowal in the Punjab. The land is described as barren. Much of it is grazed, farmed at subsistence level, or claimed under customary tenure by people who will not be consulted and cannot litigate.
This is the structure that explains why Pakistan’s economy does not grow the way its neighbors’ economies grow. It is not a shortage of talent or capital or entrepreneurial appetite. It is that the most profitable activity available to anyone in Pakistan is proximity to the institution that writes the rules, and the second most profitable is real estate, and actual production of tradable goods runs a distant third against competitors who pay full tax, full land price, and full regulatory cost. A country cannot industrialize when its largest conglomerate is exempt from the discipline that makes conglomerates efficient. Ayesha Siddiqa gave this economy its name two decades ago and the intervening years have only made the accounting easier.
V.
Look at this year’s numbers as the government presents them and you will find a recovery. GDP grew 3.7 percent in fiscal 2026, the fastest in four years. Inflation came down into single digits. The current account was close to balanced. Reserves rebuilt faster than the Fund projected. The KSE-100 index rose 18.4 percent between July and March, and the Finance Minister, Muhammad Aurangzeb, tabled a budget in June crediting the country’s improved international standing to the performance of the armed forces.
Now look at the same year through a microscope.
Total public debt stood at Rs83.28 trillion at the end of March, Rs57.57 trillion of it domestic and Rs25.72 trillion external. Debt servicing consumes Rs8.054 trillion of a federal budget with a total outlay of Rs18.771 trillion, which is to say that before a single school is built or a single doctor is hired, forty-three rupees out of every hundred the federal government spends have already gone to the holders of its paper. The national poverty headcount, by the government’s own Economic Survey, rose to 28.9 percent in fiscal 2024-25 from 21.9 percent in fiscal 2018-19. That is roughly twenty-seven million additional people pushed below the line in six years, bringing the total to around seventy million. Rural poverty went from 28.2 percent to 36.2 percent. Urban poverty went from 11 percent to 17.4 percent. The Fund’s own country report records a poverty rate of 25.3 percent in fiscal 2024 against 18.3 percent two years earlier and notes, in the flat register of these documents, that health and education outcomes remain very weak relative to comparable lower-middle-income countries.
Against that, the allocations. Defence services received Rs3 trillion in the fiscal 2026-27 budget, an increase of 17.65 percent over the previous year’s original allocation, with the largest line items being Rs967.5 billion for personnel costs and Rs925.8 billion for physical assets, the latter up about thirty-nine percent. The federal health allocation, covering tertiary care, cancer treatment, and emergency response across the country, was Rs25.1 billion. Higher education received Rs46 billion, with a further Rs65 billion for the Higher Education Commission. Basic education and skills received Rs22 billion. Science and technology received Rs3.6 billion.
Twenty-five million Pakistani children are out of school, the second-largest such population on earth, more than two years after the government declared a national education emergency. Adult literacy stands at 63 percent, and at 54 percent among women. A quarter of households report moderate or severe food insecurity, with Balochistan and Sindh worst affected and the lowest income quintile carrying roughly five times the risk of the highest.
Set the two columns beside each other. Defence services: Rs3 trillion. Federal health: Rs25.1 billion. That ratio is not the residue of a hard fiscal year. It is a statement of what the state is for, published annually, in a document, with a foreword.
VI.
Now to the part that people find hardest to say plainly, because saying it plainly sounds like an accusation of malice, when what it actually describes is something closer to arithmetic.
An arrangement of this kind cannot survive consent. If the ninety-eight percent of Pakistanis who are outside it were permitted a free press, an uncaptured judiciary, an election whose result was allowed to stand, and a parliament with genuine line-item authority over the defense budget and the military’s commercial holdings, the arrangement would end. Not gradually. Within a decade. The land would be audited. Agricultural income and urban real estate would be taxed at rates that reflect their share of national wealth. The Rs3 trillion would face a committee that could reduce it. The foundations would lose their exemptions. The retired officers would lose their plots and their boards. The Baloch would demand royalties and the Pashtun would demand an accounting for two decades of displacement and the Kashmiris would demand the seats.
Every one of these reforms is individually affordable. Collectively they are fatal. That is the whole of it. The people who run Pakistan are not withholding water, schools, hospitals, and courts from the population out of sadism or even indifference. They are withholding them because a population that has water, schools, hospitals, and courts is a population that starts asking who owns the rest, and there is no answer to that question that the owners can survive.
Which leaves violence. Not as a failure of policy but as the operating cost of the policy. The state cannot argue its way to legitimacy in Balochistan, so it holds Balochistan by force. It cannot answer the tribal districts for a war that displaced hundreds of thousands of them twice in five years, so it garrisons them. It cannot explain to Muzaffarabad why a territory that has been governed from Islamabad for seventy-eight years still has no constitutional standing, so it bans the committee that asked. The violence is not evidence that the system is breaking down. It is evidence that the system is working exactly as designed, and being paid for on schedule.
VII.
Balochistan is where the design is least disguised. The province holds Reko Diq’s copper and gold, the Saindak deposits, the Sui gas that heats Punjabi kitchens, the coastline at Gwadar that Beijing paid for, and the largest concentration of extreme poverty in the country. Its people are treated as a security variable in the extraction of assets located beneath them. The Ziarat policemen died guarding a dam. The dam is guarded because water infrastructure in Balochistan is a legitimate military target for an insurgency that regards every state project as a further installment of the occupation, and it is regarded that way because for seventy-eight years the state’s projects there have gone out of the province.
On June 22 of this year, an anti-terrorism court sentenced Dr. Mahrang Baloch and Sibghatullah Shahji to life imprisonment. Mahrang is a physician. She founded the Baloch Yakjehti Committee, which does one thing: it documents enforced disappearances and demands that the disappeared be produced before a court. That is the entirety of its program. It is a demand that the state obey its own constitution. She led the long march from Kech to Islamabad in December 2023, and the national gathering at Gwadar in July 2024, and she was nominated for the Nobel Peace Prize shortly before her arrest in March 2025. She has been in solitary confinement in Block 9 of Hudda Jail in Quetta since. Her trial was moved out of open court into the prison itself and conducted by video link, with journalists and civil society denied access. She and Shahji boycotted it. United Nations experts called the sentence a travesty following an unfair trial and the misuse of counter-terrorism law to suppress peaceful protest.
The Commission of Inquiry on Enforced Disappearances, a body the state created itself, had recorded 10,285 cases by 2024 and left more than two thousand pending. The Baloch Yakjehti Committee documented over 1,200 new cases in 2025 alone. The commission has spent years without a chairperson at various points, cannot compel testimony, and cannot punish anyone.
Consider the sequence the state has constructed. A young woman who could have left for a hospital job in Karachi instead builds a lawful, non-violent, constitutionally framed movement to recover missing men. She marches a thousand miles rather than take up a weapon. The state does not answer her. It arrests her, tries her inside a jail, and gives her two life sentences. Then it holds press conferences about the Balochistan Liberation Army. Every eighteen-year-old in Panjgur and Kech watched that sequence. The state has just explained to them, with unusual clarity, what happens to Baloch who choose the peaceful route and what does not happen to those who do not.
The insurgency the army is fighting in Balochistan is, to a degree the army will never concede, a product the army manufactures.
VIII.
If Balochistan can be dismissed by Islamabad as separatism, Azad Jammu and Kashmir cannot, and that is what makes it the clarifying case of this year.
The Jammu Kashmir Joint Awami Action Committee is not a separatist front. It is a coalition of traders, lawyers, students, transporters, and shopkeepers, and it began by demanding cheaper wheat flour and lower electricity tariffs in a territory that generates hydropower for the rest of the country. Its charter expanded into representation, taxation, employment, healthcare, and education. In other words, the least revolutionary agenda imaginable: municipal, material, and constitutional. Its central grievance this year has been the twelve seats in the AJK Legislative Assembly reserved for refugees resident elsewhere in Pakistan, seats that give Islamabad a permanent bloc inside a legislature it already controls by other means.
The authorities banned the committee under anti-terrorism law. Protests began in early June. Rawalakot went under curfew. The internet was cut. Rangers and Frontier Corps contingents were deployed into a territory whose distinct status Pakistan cites at the United Nations every year. By the third week of June at least twenty-four people were dead; by July the count had passed thirty. Thousands have been camped in an open field outside Rawalakot for more than a month. Elections to the assembly were scheduled for July 27.
The federal budget passed in June allocated Rs146 billion to Azad Jammu and Kashmir. The federal health allocation for the entire country was Rs25.1 billion. Read those two numbers together and the nature of the transfer becomes visible: Islamabad will spend on Kashmir, generously, because Kashmir is a claim it needs. It will not spend on Kashmiris, because Kashmiris asking for representation are an inconvenience to the claim.
Pakistani officials condemn India’s conduct in Srinagar with internet shutdowns and mass detentions. They did so this year while running a curfew in Rawalakot, an internet blackout in Muzaffarabad, and a proscription order against a traders’ association.
IX.
In Khyber Pakhtunkhwa the bill being paid is for a policy the establishment designed and has never renounced. Strategic depth, the doctrine of cultivating armed proxies in Afghanistan to secure the western flank against India, was pursued from the 1980s onward and produced, among other things, the ecosystem from which the Tehreek-e-Taliban Pakistan grew. The state fought that organization from 2009 to 2016 in operations that emptied Swat, South Waziristan, and North Waziristan of their populations, sent hundreds of thousands into camps and rented rooms in Bannu and Dera Ismail Khan, and were declared victories. In 2021 the Taliban returned to Kabul and Islamabad’s establishment celebrated it in public. The TTP came back over the border with them.
Fatalities in Khyber Pakhtunkhwa rose fifty-three percent between the first and second quarters of this year. The province now accounts for three out of every five violence-related deaths in Pakistan. In July the provincial chief minister, Sohail Afridi, approved military operations across eight districts: Tank, Khyber, Kurram, Bajaur, Kohat, Lakki Marwat, Bannu, and Waziristan. A day later two policemen were killed by a roadside device on the Tank-Jandola road. Eleven days after that, the Manji Khel post was destroyed.
The men in that post were policemen and soldiers from families in the same districts. The Pashtun Tahafuz Movement, which formed to demand the clearing of landmines, an end to disappearances, and accountability for extrajudicial killings in exactly those districts, has been treated as a security threat, its leaders jailed and its gatherings banned. Once again the constitutional route was closed while the armed route was left open, and the results are counted every Tuesday by think tanks in Islamabad.
X.
The most consequential thing that has happened to Pakistan in this decade was not an attack. It was the Twenty-Seventh Amendment, and the number that matters most about it is not 234. It is four.
Four members of the National Assembly voted against a package that made the army chief the constitutional Chief of Defence Forces with command over the navy and the air force, abolished the office of the Chairman of the Joint Chiefs of Staff Committee, placed the Strategic Plans Division and therefore the nuclear arsenal under a single officer, created a Federal Constitutional Court that sits above the Supreme Court on constitutional and national security matters, granted lifetime immunity from criminal proceedings to the President and to any officer holding five-star rank, and set the threshold for removing the Chief of Defence Forces at a two-thirds parliamentary majority. A Prime Minister of Pakistan can be removed by a simple majority. The Chief of Defence Forces cannot. Chatham House described the amendment, accurately, as a step toward authoritarian rule and as the dismantling of what remained of civilian authority.
Nawaz Sharif was in the chamber. So was Bilawal Bhutto-Zardari. So was the Prime Minister, Shehbaz Sharif. Asif Ali Zardari, whose own immunity was extended by the same text, signed it.
This is the point at which the conventional Pakistani narrative of a military that seizes power from unwilling civilians stops being usable. Nobody seized anything in November 2025. The political parties handed over the state by roll call, on the record, in exchange for the ministries, the development funds, the contracts, the protection from prosecution, and the tacit understanding that whichever of them is currently favored will be permitted to govern. Nawaz Sharif was removed by the establishment in 1993 and again in 2017 and jailed. Zulfikar Ali Bhutto was hanged. Benazir Bhutto was dismissed twice. Imran Khan, who arrived on the establishment’s shoulders, has been in prison since 2023 and his party’s mandate was set aside in February 2024. Every one of these families learned the same lesson from that history, and the lesson they learned was not that the arrangement should be dismantled. It was that one should be inside it.
So there are not two elites in Pakistan contending for control of a state. There is one elite with two departments. The uniformed department owns the land, the conglomerates, the security budget, and the veto. The civilian department manages the tax code, the privatization pipeline, the provincial patronage, and the international paperwork, including the letters of intent to the Fund. Both departments are compensated. Neither audits the other. When the arrangement is threatened from below, in Quetta or Rawalakot or Bannu, both departments describe the threat in identical language and the Frontier Corps is deployed with the signature of an elected chief minister.
XI.
What does a state look like at the end of this road?
It looks like an economy that satisfies every benchmark its creditors set and none of the needs of its citizens: a stock market up eighteen percent while poverty climbs seven points in six years. It looks like a budget in which debt service and defense together consume more than half of federal spending while the entire federal health allocation would not build four modern hospitals. It looks like twenty-five million children out of school in a country whose ruling institution has identified 4.8 million acres for corporate farming. It looks like a physician serving two life sentences in solitary confinement for asking a court to produce the disappeared. It looks like curfew in a territory whose freedom Pakistan raises annually at the United Nations. It looks like a hundred and eight people killed in a single week in July, and a national leadership whose principal legislative achievement of the year was to make one man unprosecutable for the rest of his life.
None of this is collapse. Collapse implies loss of control. Pakistan’s rulers have more control today than at any point since 1971, and their institutions are richer, more legally insulated, and less accountable than they have ever been. What has failed is not the state’s capacity. What has failed is the proposition, offered in 1947 and never honored, that the state existed for the people inside it.
There is a question that this year leaves open and that nobody in Rawalpindi or Islamabad appears willing to consider. The last time this arrangement was pressed to its logical end, it produced a population that had been outvoted, out-spent, and finally shot at, and that population left, taking a majority of the country with it. The generals of 1971 also believed they had the situation in hand. They also believed the problem was foreign-sponsored miscreants. They also had figures showing insurgent casualties.
They were wrong about what a people will absorb. There is no evidence that anything has been learned since, and considerable evidence that the arrangement has simply grown better at surviving the discovery.