Two Colonisations, One Border: What the Data Actually Says About Bengal’s Post-1971 Demographic Story

Check BB’s personal anecdote on the Northeast.

As mentioned, the Special Intensive Revision of electoral rolls in West Bengal has removed 9.1 million names, 2.7 million of them under contested adjudication. The argument has compressed itself into two bad positions. Either every deletion is disenfranchisement, or every deletion is a Bangladeshi caught. Neither is true, and the census of the last hundred years tells a more specific story than either side wants.

The received wisdom on Northeast India collapses into one sentence: Bangladeshi Muslims are flooding across the border. The received wisdom is partly wrong and mostly incomplete. A narrower reading of the census, focused on the actual border districts, tells a stranger story. There are two demographic colonisations in eastern India, not one. They run in opposite religious directions, and Indian statecraft has treated them as opposites: one ratified, one criminalised.

Tripura: the Hindu Bengali takeover

Tripura*: 1951: 71% Hindu, 7% Muslim (with ~21% still counted under tribal religions separately).

Today: 83% Hindu, 9% Muslim.

In 1941, tribals were 50% of Tripura. By 2011, 32%. Partition and the 1971 war did the rest.

The population that replaced the Kokborok, Reang, and Jamatia is Bengali Hindu, not Muslim. Tripura’s Muslim share today is 9%, below the all-India average. The Northeast state most transformed by Partition and 1971 is the one that became a Hindu Bengali colony.

South Tripura district is the cleanest data point. The ST share there is 17%. The Bengali Hindu majority there is overwhelmingly composed of descendants of refugees who crossed between 1947 and 1971.

*Note on Tripura: the Hindu figure jumped partly because tribals were reclassified as Hindu between 1951 and 1971 in the census. Real Hindu Bengali influx adds on top of that statistical shift.

Assam border districts: the Muslim Bengali case

Assam: 1951: 72% Hindu, 25% Muslim.

Today: 61% Hindu, 34% Muslim.

Dhubri, the westernmost district on the Bangladesh border, was 64% Muslim in 1971. Today, 80%. Karimganj moved from 52% to 56%. Barpeta from 59% to 71%. Goalpara sits at 58%.

These rises are real. They also sit on a historical base that predates Bangladesh. The Bengali Muslim presence in lower Assam was seeded by British colonial labour-settlement policy after 1901, which moved East Bengali peasants into Assamese floodplains for revenue extraction. Post-1971 growth is partly natural increase, with Assam’s Muslim TFR at 3.6 against Hindu 1.9, and partly continued informal migration.

The NRC finally told on the narrative. The National Register of Citizens published in August 2019 excluded 1.9 million persons out of 33 million applicants. The political class had promised between 5 and 20 million. The gap between the promise and the register is the size of the story.

West Bengal: the quiet case, and the one that breaks the thesis

West Bengal: 1951: 78% Hindu, 20% Muslim.

Today: 71% Hindu, 27% Muslim.

Murshidabad is 66% Muslim. Malda 51%. Uttar Dinajpur 50%. These are Bangladesh-border districts, and they are where West Bengal’s statewide Muslim share grew from 20% in 1951 to 27% today.

But Murshidabad was already 55% Muslim in 1951. This is not a post-1971 colonisation. This is the pre-Partition geography of eastern Bengal, redrawn by Radcliffe, completing itself through demographic inertia. The migration component sits on top of a Muslim-majority base inherited in 1947, with higher TFR doing most of the arithmetic work.

The cleaner thesis is not “Muslim Bangladeshis colonising West Bengal.” The cleaner thesis is that West Bengal’s border districts were never fully un-colonised in 1947, and the Partition boundary was drawn across a live demographic gradient that continues to drift east to west.

The live SIR argument sits on top of this geography. The adjudication category in West Bengal flagged 2.7 million names, including the Nobel laureate Amartya Sen, using an algorithm that penalised standard Bengali transliteration variance and multi-generational surname drift. The algorithm is a bad instrument. That does not make the underlying question illegitimate. A ruling party fifteen years in office has an incentive structure to pad its rolls that does not disappear because the audit is clumsy, and comparable SIR exercises in Tamil Nadu (7.3 million deletions) and Bihar generated no equivalent outcry. The thread argument that flaws in the instrument invalidate the exercise is not the same argument as the thread argument that there is no demographic question to answer. Both can be right, and in Bengal they are.

The post-1971 political economy

The war generated roughly 10 million refugees in 1971. Official Indian and Bangladeshi figures claim the bulk returned after December 1971. The residual, largely Hindu and largely absorbed into West Bengal and Tripura, was never counted properly.

The Indira-Mujib Treaty (1972) fixed 25 March 1971 as the cutoff date for legitimate residence. The Assam Accord (1985) preserved the same cutoff for Assam after six years of agitation and the Nellie massacre (1983). The Illegal Migrants (Determination by Tribunal) Act (1983) made Bangladeshi identification in Assam operationally impossible, and was struck down by the Supreme Court in Sarbananda Sonowal v. Union of India (2005).

The Citizenship Amendment Act (2019) codifies the asymmetry: non-Muslim arrivals from Bangladesh, Pakistan, and Afghanistan before 2014 are fast-tracked to citizenship; Muslim arrivals remain deportable. This is the Tripura settlement written into statute and exported to the rest of eastern India.

What the data supports, and what it does not

1. Tripura was demographically colonised by Bengali Hindus, with state sanction. The tribal share fell from half to a third in seventy years. This is the clearest case of a recognised ethnic group becoming a minority on its own soil in post-independence India.
2. Assam’s border districts saw real Muslim demographic growth, but smaller than the political narrative claims. The NRC is the hard ceiling on the infiltration claim, and it came in an order of magnitude lower than the promised number.
3. West Bengal’s border districts are mostly pre-1947 geography, not post-1971 invasion. The statewide Muslim share rose roughly 7 points in sixty years, which is unusual but not catastrophic, and TFR carries a large share of it.
4. The “Northeast demographic threat” narrative is Assam plus Tripura generalised outward. Meghalaya, Nagaland, Mizoram, and Arunachal Pradesh are insulated by the Inner Line Permit and the Sixth Schedule. Their demographics have barely moved.

The precedent

Indian statecraft ratified one colonisation and criminalised the other. Bengali Hindus who colonised Tripura are citizens, voters, and the political majority of the state. Bengali Muslims who settled Dhubri under British revenue policy, and their descendants, are serially re-litigated as foreigners. Both movements are real. Only one gets called colonisation. The other gets called refuge, and then, when the refuge is Muslim, infiltration.

The precedent for the rest of South Asia is sharper than the NE India debate usually admits. A border drawn across a demographic gradient does not freeze the gradient. It registers it, criminalises half of it, and rewrites the rest into citizenship law. Tripura is what that settlement looks like seventy years in.

The Clean Count Captures a Lagging Indicator

The SIR argument in West Bengal is a lagging indicator of this longer story. The Bengal boundary was drawn across a demographic gradient in 1947 and started leaking in both directions within months.

Seventy-nine years on, the state is counting what moved, with an algorithm not fit for Bengali nominal conventions, administered by a body that has political reasons to hurry, contested by a state government that has political reasons to obstruct.

A cleaner count is worth having. A cleaner count will also disappoint both sides of the current argument. The Assam NRC (2019) is the precedent for that disappointment, and neither side has absorbed it.