For more than a decade, the same story has resurfaced in security analyses, think-tank papers, and countless opinion pieces: Pakistan’s religious seminaries, or madrasas, were systematically expanded in Pashtun-dominated regions, chiefly Khyber Pakhtunkhwa and northern Balochistan, to radicalise an entire ethnic community. The idea fits neatly into explanations of border militancy and the fallout from post-9/11 geopolitics. It feels plausible. Yet when placed side by side with the government’s own numbers, the narrative quietly collapses.
The most up-to-date official count, presented to the Senate on 19 December 2024 and still the benchmark well into 2026, records 17,738 registered madrasas across Pakistan, enrolling a total of 2,249,520 students. These figures were supplied by the Ministry of Federal Education and Professional Training through the Directorate General of Religious Education in direct response to a parliamentary question. No significant revisions have appeared in Senate records or ministry statements since.
The provincial distribution immediately undercuts any notion of ethnic targeting. Punjab accounts for 10,012 madrasas and 664,065 students. Khyber Pakhtunkhwa reports 4,005 madrasas but a strikingly higher enrolment of 1,283,024 students. Sindh has 2,416 madrasas with 188,182 students. Balochistan trails with 575 madrasas and 71,815 students. Azad Jammu and Kashmir contributes 445 madrasas and 26,787 students, while Islamabad has roughly 255 madrasas enrolling 11,301 students, and Gilgit-Baltistan lists about 30 madrasas with 4,346 students.
The contrast between Punjab and KP looks puzzling at first: more than twice as many madrasas in Punjab, yet less than half the students. Some might suspect incomplete records or deliberate under-reporting. But the explanation lies not in clerical error but in the very different character of the institutions themselves.
In Punjab, madrasas are typically small, neighbourhood-based operations, often attached to local mosques, serving children from the immediate community with modest daily attendance. Punjab’s enormous population — over 120 million according to the 2023 census — naturally supports thousands of these low-enrolment setups, averaging around 66 students per institution. In Khyber Pakhtunkhwa, the landscape is dominated by larger residential dar-ul-ulooms, prominent boarding seminaries located in places like Swat, Mardan, Nowshera, and Dera Ismail Khan.
These institutions draw students from a much wider area: the southern districts of KP, the northern fringes of Punjab, and even other provinces. They offer not only religious education but also free accommodation, meals, and a structured daily routine, making them especially appealing to families in remote villages, economically strained households, or areas with scarce government schools. As a result, KP’s average enrolment reaches roughly 320 students per madrasa, nearly five times that of Punjab.
This difference is well recognised by those familiar with the sector. Historical Deobandi scholarly networks, which took deeper root in the northwest, have contributed to the presence of larger, more established boarding schools there. But the primary drivers are practical and socioeconomic, not part of any coordinated ethnic project.
The Senate figures cover only registered madrasas, formalised under registration drives that intensified after 2019 and were reinforced by the Societies Registration Amendment Act of 2024. The fuller picture emerges from Pakistan’s first Economic Census, conducted in 2023–24 and released in 2025 by the Pakistan Bureau of Statistics, which enumerated roughly 36,000 madrasas in total, including unregistered ones, outnumbering the country’s factories at around 23,000 and illustrating how deeply embedded these institutions are in everyday community life. Even here, Punjab leads in absolute numbers, in line with its demographic weight.
Looking back for historical perspective, a 2018 study by the Karachi-based Majlis-e-Ilmi Foundation documented 37,517 madrasas enrolling 4.59 million students, with Punjab at 13,798, Sindh at 10,033, KP at 3,579, and Balochistan at 2,959.
Pashtun-majority regions accounted for less than 18 per cent of the total. The long-term pattern remains consistent: no evidence of a disproportionate “madrasa belt” engineered in KP or Balochistan.
The broader education crisis provides essential context.
According to the Pakistan Institute of Education’s 2023–24 report, echoed in analyses through 2025, Pakistan has around 25.1 million out-of-school children — roughly 35 per cent of those aged 5 to 16 — though some recalculations using the 2023 census population base suggest a range of 20 to 22 million. Punjab bears the largest absolute burden at approximately 9.6 to 11 million, followed by Sindh at 7.8 to 8 million, KP at 3.6 to 4.9 million, and Balochistan at 2.9 to 3.4 million, with Balochistan recording the highest rate of exclusion, up to 69 to 70 per cent. Across the country, madrasas frequently serve as accessible, low- or no-cost alternatives in places where state schools are distant, overcrowded, or simply absent. That role is nationwide, not ethnically targeted.
This is not to suggest the madrasa sector is beyond criticism.
Questions about curriculum content, financial transparency, and occasional links to extremist networks remain valid, and successive governments have pursued registration, audits, and mainstreaming initiatives with varying degrees of success. Extremism in Pakistan draws nourishment from many wells: geopolitical pressures, economic deprivation, digital propaganda, urban mosques, and even elite educational institutions. Reducing the entire madrasa phenomenon to a supposed Pashtun-focused radicalisation scheme ignores both the data and the complexity on the ground.
The evidence, from Senate disclosures to Pakistan Bureau of Statistics censuses, reveals a straightforward reality: madrasas exist throughout Pakistan, most numerous and entrenched where population density and educational gaps are greatest. Punjab leads in sheer count; Khyber Pakhtunkhwa stands out for its larger boarding institutions that serve students from multiple regions. The pattern is structural, rooted in history, geography, and need, not sinister design.
As long as public debate clings to ethnic stereotypes instead of these facts, it will distract from the urgent work ahead — consistent regulation, meaningful curriculum integration, and genuine efforts to close educational inequalities across every province. The numbers have been clear for years. The conversation should finally catch up.