A hydroelectric dam, a constitution with no word for them, and a peace accord still unimplemented after nearly three decades — how the Chittagong Hill Tracts became one of Asia's least visible ethnic conflicts, and why the government in Dhaka keeps changing without the policy ever changing.
In the southeastern corner of Bangladesh, wedged between India and Myanmar, a range of forested hills covers about a tenth of the country's land but holds barely one percent of its people. The Chittagong Hill Tracts (CHT) are home to eleven ethnolinguistic communities, the Chakma, Marma, Tripura, Tanchangya, Mro, Bawm, Khyang, Khumi, Chak, Pankhua, and Lushai, who since the 1970s have adopted a shared political name for themselves: Jumma, after the swidden farming practice, jhum, that shaped their way of life for generations. They are mostly Buddhist, with Hindu, Christian, and animist minorities, and they speak languages unrelated to Bengali.
That single fact, a population that is not Bengali, in a state whose founding idea was that all its citizens are, explains almost everything that follows.
British administrators annexed the Hill Tracts in 1860 and, by 1900, had designated the region an "Excluded Area": outsiders could not settle there or acquire land without the consent of a traditional Circle Chief and the Deputy Commissioner. The protection wasn't humanitarian in intent, it was about frontier stability, but it worked. In 1947, at Partition, non-indigenous residents made up only about 2.5 percent of the region's population.
That legal wall came down in stages. In 1964, under Pakistan's Ayub Khan government, the CHT's protected status was abolished and courts subsequently confirmed that any citizen could move there freely. The demographic floodgates were opened fifteen years before anyone walked through them.
The first wave arrived in the form of a dam. The Karnaphuli Hydroelectric Project, universally known as the Kaptai Dam, was built between 1957 and 1962 with American financing and engineering, at the height of Cold War competition for influence in the region. It remains Bangladesh's only hydroelectric plant, and its power was routed to the port city of Chittagong, not to the hills it flooded.
The reservoir submerged roughly 655 square kilometers, including about 40 percent of the region's cultivable land, precisely the fertile valley floor that had sustained settled agriculture. Government figures, broadly accepted even by indigenous organizations, put the number displaced at around 100,000 people, about 18,000 families, roughly 70 percent of them Chakma: close to a quarter of the entire region's population, uprooted by a single project. Many were resettled onto land that was itself later flooded as the reservoir filled, displacing the same families twice. The original town of Rangamati, the residence of the Chakma Raja, and multiple Buddhist temples disappeared under the water permanently.
Between 1964 and 1966, an estimated 40,000 to 65,000 people walked out of the flooded valleys into what is now Mizoram and Tripura in India. Their descendants are, decades later, still fighting citizenship cases in Indian courts.
The 1971 Liberation War was fought and won under the banner of Bengali linguistic nationalism, a powerful answer to Pakistan's religious nationalism, and one that left no room in its vocabulary for a population that was neither Bengali nor Muslim.
In February 1972, M. N. Larma, the sole Hill Tracts representative in the Constituent Assembly, brought four modest demands to Prime Minister Sheikh Mujibur Rahman: regional autonomy, protection of the pre-1900 land rules, constitutional recognition of the region's indigenous communities, and a bar on further in-migration. All four were rejected. Article 6 of the new constitution declared that the citizens of Bangladesh "shall be known as Bengalees." On the floor of the Assembly, Larma made the objection as precisely as it has ever been made: he was a Chakma and a Bangladeshi, he said, but not a Bengali, and a national identity cannot be assigned to people who do not hold it.
Within days, the Parbatya Chattagram Jana Samhati Samiti (PCJSS) was founded in anticipation of exactly this outcome. Its armed wing, the Shanti Bahini, followed. The sequence matters: armed struggle in the Hill Tracts began as a response to the closing of every constitutional path, not the other way around.
The most consequential single intervention in the region's modern history came under the Ziaur Rahman government. Between roughly 1979 and 1985, the state facilitated the relocation of an estimated 400,000 landless Bengali households from the plains into the Hill Tracts, offering land, cash, food rations, and security protection. The land handed out was classified as state or forest land, a classification made possible because indigenous land tenure in the CHT was customary and largely undocumented, and the administration simply declined to recognize possession that wasn't backed by paper title.
No decree of expropriation was needed. Applying an ordinary documentary standard of ownership to a population that had never had reason to generate documents was enough.
The demographic result is stark and, unlike almost everything else in this story, not seriously disputed. In 1947 the indigenous population of the Hill Tracts stood at roughly 97.5 percent. By the 2022 national census, the three hill districts, Khagrachari, Rangamati, and Bandarban, were reported at 1,842,815 people, essentially split down the middle: just over half Bengali, just under half indigenous. No comparable demographic inversion has occurred anywhere else in Bangladesh.
Alongside settlement came militarization. A civilian development board was placed under the command of the regional army commander, and dispersed indigenous homesteads were consolidated into monitored "cluster villages", a counterinsurgency technique with a long pedigree in Malaya and Vietnam, whose local effect was to sever farmers from the swidden plots that sustained them.
Anyone trying to understand this conflict runs immediately into a second problem: nobody agrees on the numbers, and the disagreement is not marginal. No independent, region-wide casualty count has ever been conducted.
Amnesty International reported that hundreds of villagers were killed in a single incident at Logang in April 1992. The Bangladesh government's own inquiry, whose full findings were never released in Bengali, put the toll at twelve. A thirty-fold discrepancy, over a single afternoon, in a single village
That is not a margin of error. It is the signature of a conflict in which the body count has itself become an object of dispute. A pattern recurs across the documented mass-casualty incidents of the 1980s and 1990s: investigation without publication, and publication without prosecution. The Chittagong Hill Tracts Commission, an independent monitoring body, found that most such incidents were never properly investigated at all, and that not one individual has ever been convicted in a Bangladeshi court for participating in any of them.
The case of Kalpana Chakma is the starkest illustration of how that impunity actually works. Chakma, a 23-year-old organizer with the Hill Women's Federation, was abducted at gunpoint from her home on the night of 12 June 1996, hours before a national election. Her brothers, who escaped the same abduction, identified three of the men involved, including a serving army officer. The investigation ran for twenty-eight years. In April 2024, a court in Rangamati accepted a police report clearing every named suspect, who, the family says, were never even questioned.
The Chittagong Hill Tracts Accord, signed on 2 December 1997 between the Awami League government and the PCJSS, ended more than two decades of armed conflict. It was a genuine achievement: a Ministry of CHT Affairs was created, a regional council was established under indigenous leadership, roughly 12,000 refugee families were repatriated from India, and nearly 1,947 PCJSS fighters surrendered their weapons and demobilized.
The land question was left unresolved, and land is the actual grievance. The Land Dispute Resolution Commission created to adjudicate competing claims has never issued a single ruling, because the government has never written the procedural rules the Commission needs to function, despite draft rules being submitted in 2017. Roughly 22,000 land complaints sit untouched. This is not a matter of contested jurisprudence or scarce resources. It requires no parliamentary vote, no negotiation, and no budget, only a decision to publish a set of rules that already exist in draft form. It has not been made, across four different governments of three distinct political characters.
The removal of the Awami League government in August 2024, after a mass uprising, created an unusual opportunity to test a specific question. The Awami League had signed the 1997 Accord and governed for fifteen of the following twenty-seven years; it was also the party most often blamed, in the advocacy literature, for failing to implement it. If non-implementation was a partisan failure, its removal should have changed something. If it was structural, built into the logic of the state rather than the politics of any one party, it should not have.
The evidence available so far supports the second reading. Rights monitors recorded 268 incidents of human rights violations in the Hill Tracts in 2025 under the interim administration, including eight killings and 26 cases of sexual violence, and estimated that roughly two-thirds of the Accord's core provisions remained unimplemented. In September 2025, a twelve-year-old Marma girl was raped in Khagrachari; her family says police pressured them to describe the attackers as unidentified rather than naming their community. Protests followed in Khagrachari and Dhaka. Security forces opened fire on demonstrators on 27 September; at least three people were killed.
The party that won Bangladesh's February 2026 election with a two-thirds majority, the BNP, opposed the 1997 Accord at the time it was signed. In January 2026, its chairman told an audience at Chittagong University that hill communities should be treated as Bangladeshis without special quotas, not as indigenous peoples with distinct rights. Human rights groups warned that removing existing protections risks reigniting the conflict the Accord was meant to end.
It would be a disservice to this issue to pretend the government's position has no substance. Bangladesh is one of the few states in South Asia to have negotiated and implemented any peace accord with an insurgent movement, disarming it without mass prosecutions. Literacy in the hill districts now exceeds 63 percent, up from a much lower baseline. The security threat from armed groups, including a newer faction, the Kuki-Chin National Front, active since 2022, is real, and no state would withdraw its security presence from an unstable border region.
None of that, however, explains why 22,000 land claims remain unadjudicated for want of rules that require no legislation to write. It explains why maximalist demands for full restitution are not realistic. It does not explain the specific, decades-long pattern of administrative non-performance that the record shows.
The measures that would constitute genuine progress are, notably, modest and specific: publish the Land Commission's procedural rules; release the long-suppressed inquiry reports into the Logang and Naniarchar incidents; reopen the Kalpana Chakma case and question the men who were named by eyewitnesses in 1996; recognize the Hill Tracts population's status under international instruments Bangladesh already ratified half a century ago. None of these steps would displace a single settler or move a single soldier. That they remain undone, across a military dictatorship, an elected secular-nationalist government, a nationalist opposition party, a reformist interim administration, a mass uprising, and a constitutional referendum, is itself the finding. The instrument has changed hands repeatedly. The outcome has not moved.