It would also have shifted 8-10 seats towards Muslim-majority Lower Assam in a state whose modern history was forged by exactly this kind of demographic anxiety.
There is a relatively simple sentence in the Election Commission of India’s final delimitation order for Assam, published in August 2023, that merits far closer examination than it initially received. It reads: “All Assembly and Parliamentary constituencies in the State have been delimited based on the 2001 Census as provided under Articles 170 and 82 of the Constitution.”
That sentence, dry as government paperwork, simultaneously serves as a legal justification, a foreboding of political outcome, and a historical choice.
What it does not address, and what remains a fair subject of inquiry, is why the 2011 Census — publicly accessible and officially published by the Government of India — was not used for Assam when the same ECI used it for the redrawing of Jammu and Kashmir’s constituencies after the abrogation of Article 370. It also does not explain why a state whose boundaries were last drawn in 1976 was being redrawn in 2023 using data from 2001 — a 22-year-old source — when more reliable and recent information did exist. And it leaves one to honestly wonder what the 2011 dataset would have produced had it been applied instead.
This article attempts to do what that order did not: to map, as thoroughly as the available evidence permits, what a delimitation based on the 2011 Census might have looked like in Assam. To read where seats would have shifted, which communities would have gained representation and which would have lost it, what the political consequences would have been — and then to ask whether such a delimitation would have introduced a different and perhaps more dangerous set of complications.
Why did Assam use the 2001 Census for delimitation instead of the 2011 Census?
Assam’s 2023 delimitation was based on the 2001 Census, and not the 2011 Census, mainly because of the legal framework under which it was carried out. It was conducted under the Delimitation Act of 2002, a law passed by Parliament that required constituency boundaries across India to be redrawn using only 2001 Census data and nothing else. This itself followed the 87th Constitutional Amendment, which made the 2001 Census the official population base for all future delimitation exercises, while the 84th Amendment froze the total number of Lok Sabha and Assembly seats until after 2026.
Assam was supposed to undergo delimitation along with the rest of the country in the 2000s. However, in 2008, the process was postponed by the President under Section 10A of the Delimitation Act due to concerns over the sensitive political climate in the state. The process was not cancelled; it was simply put on hold.
When the Centre lifted this deferral in 2020, it did not pass a new law or begin a fresh delimitation exercise. Instead, it simply restarted the old process that had been paused under the same Delimitation Act of 2002. Since that law specifically required the use of only 2001 Census figures, the authorities had no legal option but to use 2001 data unless Parliament amended the law. No such amendment was made.
This is also why Assam’s case differed from Jammu and Kashmir, which used the 2011 Census for its delimitation. J&K was able to do so because its delimitation took place under a completely separate law — the Jammu and Kashmir Reorganisation Act, 2019, passed after Article 370 was abrogated. That new law specifically allowed delimitation on the basis of the 2011 Census, giving J&K a fresh legal framework separate from the older national delimitation cycle.
How has Assam changed between 2001 and 2011?
To understand the gap between 2001 and 2011, it helps to compare two sets of numbers side by side.
In 2001, Assam had a population of 2.67 crore. By 2011, this had grown to approximately 3.12 crore, an increase of around 45 lakh people. This comes to a decadal growth rate of around 17.07 per cent. As per the 2001 Census, the state’s 126 Assembly seats represent approximately 2.11 lakh people each. On a 2011 basis, that figure rises to approximately 2.48 lakh per Assembly seat.
To dig deeper, one would need to see how evenly or unevenly those additional 45 lakh people were distributed. That is precisely where the political significance of the gap lies. The central fact driving this difference is simple: Assam has not grown evenly.
Dhubri district grew by 24.44 per cent between 2001 and 2011, from 15.66 lakh to 19.49 lakh. Morigaon grew by 23.34 per cent, Goalpara by 22.64 per cent, Nagaon by 22 per cent, and Darrang by 22.19 per cent. These are districts located in Lower and Central Assam, where Muslim populations are large and growing rapidly.
On the other side, Sivasagar grew by 9.44 per cent, Jorhat by 9.31 per cent, Kokrajhar by 5.21 per cent, and Dibrugarh by 11.92 per cent. These are districts in Upper Assam, where Ahom and various other indigenous communities and tribes are dominant, and where population growth has been considerably slower.
In terms of religious breakdown, the Muslim population share in Assam rose from 30.9 per cent in 2001 to 34.22 per cent in 2011 — an increase of 3.3 percentage points in a single decade, the largest such increase recorded in any Indian state during that period. In 2001, six districts had Muslim majorities. By 2011, three more had crossed that threshold, bringing the total to nine: Dhubri (79.67 per cent), Barpeta (70.74 per cent), Darrang (64.34 per cent), Hailakandi (60.31 per cent), Goalpara (57.52 per cent), Karimganj (56.36 per cent), Nagaon (55.36 per cent), Morigaon (52.56 per cent), and Bongaigaon (50.22 per cent).
It would not be a stretch to say that the decision to use 2001 data over 2011, and the absence of any constitutional amendment, was a decision to redraw Assam before the impact of those demographic changes — and before the demographic centre of gravity had shifted to Lower Assam.
What would the map have looked like under 2011?
This is the analytical core of the article, and it requires considerable thought, since delimitation is not a simple mathematical exercise. It involves geographic contiguity, compactness, administrative boundaries, and reservation requirements — none of which can be simulated with precision from census data alone. What can be accurately traced, however, is differential growth: which districts and communities would have received proportionally more representation under a 2011-based exercise.
The core principle of delimitation is “one person, one vote, one value”: the idea that constituencies should have roughly equal populations. The average population per seat under the 2011 Census would be approximately 2.48 lakh; under the 2001 Census, it was approximately 2.11 lakh.
Consider the contrast between Dhubri and Kokrajhar. Dhubri had 19.49 lakh people in 2011; Kokrajhar had 8.87 lakh. Both were given five Assembly seats in the current delimitation. This means each seat in Dhubri represents 3.9 lakh people while each seat in Kokrajhar represents 1.77 lakh — a ratio of more than 2:1. Addressing a disparity as stark as this is precisely what a 2011-based delimitation would have been uncomfortably required to do.
If applied, Dhubri would have received eight seats instead of five. Dhubri has a nearly 80 per cent Muslim population; an increase in seats would almost certainly have resulted in a net gain for Muslim-dominated districts.
Nagaon, the state’s most populous district, was given nine Assembly seats under the current delimitation. Under 2011 data, it would have been closer to eleven. Additional seats in a district that is 55.36 per cent Muslim would likely have added to Muslim representation.
Barpeta was given six seats under this delimitation, down from the previous eight. The district, with a 70.74 per cent Muslim population, would deserve around seven seats on 2011 figures. The actual order gave it six, plus one of those reserved for Scheduled Castes. This effectively reduced Muslim electoral influence to just four or five competitive seats, down from the eight that population data alone would have suggested.
On the other side, Kokrajhar has a per-seat population well below the state average, yet was given five seats. On strict population-proportion logic, it should have had closer to 3.6. The Bodoland region’s seat count was increased from 16 to 19. The direction of the 2023 delimitation is visible here: it moved seats towards the slow-growing, tribal-dominated Bodoland region and away from the fast-growing Lower Assam districts.
Aggregating across regions, the likely redistribution under the 2011 Census would have been as below:
