Did the Special Intensive Review (SIR) in West Bengal, with its unique advantage of judicial adjudication, lead to gerrymandering of the state’s constituencies ahead of the Assembly elections?
This question arises because areas or constituencies with larger numbers of Muslim voters suffered higher deletions during the segregation process. Since Muslims have a low probability of voting for the BJP and a high probability of voting for the Trinamool Congress (TMC), their omission will help the BJP. But how does one answer the larger question of electoral results and gerrymandering? The best way is to use a three-step process.
These pages were the first to point out that SIR should not lead to a decline in the absolute number of voters. On August 2, 2025, while analyzing SIR data for Bihar, the first state to undergo this process this cycle, we argued that the number of post-SIR voters has declined significantly compared to pre-SIR voters but is still larger than the number of actual voters in the previous pre-SIR elections in Bihar.
People omitted by the SIR are likely people who have died, emigrated, or are registered in more than one place. We also rightly predicted a statistical increase in voter turnout in the state. Bihar turnout data for 2025 has borne out this assessment. Turnout figures from Kerala and Assam, which underwent SIR and a special review, also support this argument. In Bengal, the final number of voters is still much higher than the number of actual voters who voted in the 2024 Lok Sabha elections, the last pre-SIR election. However, for voter turnout not to drop in the next state election, turnout would have to rise from 81.7% in 2021 to 88.9% in 2026 — a state record.
Read also | At 8.4 million, the SIR in Uttar Pradesh sees the highest addition of voters from the recruitment rolls
What about changes at the AC level in the number of voters in the state?
HT reported on Thursday that the deletion of voters by the AC during the adjudication process in West Bengal was significantly more skewed than the deletions under the SIR before the adjudication. This is driven by trends in two types of local councils where one community is more politically prominent than others.
The 67 electoral committees that elected at least one Muslim MLA since 2011 had a higher share of deletions in judicial dismissal than their share in pre-separation or even pre-SIR list. This trend also applies to the 16 advisory councils identified by Hizb ut-Tahrir as seats dominated by the Matua community, a conglomeration of Dalit groups who largely moved from what is now Bangladesh. These are assemblies in which the Matua Mahasangha, the supreme body of the sect, has a strong presence or the local legislator is from the community. Nine out of 16 ACs are reserved for Scheduled Castes (SC). To be sure, both groups – the 67 Muslim councils and 16 autonomous councils in Matua – are likely to have voters from other communities as well.
The number of voters in 2026 fell below their number in 2024 in three states: Metiaboroz in South 24 Parganas district, and Lalgola and Samsirganj in Murshidabad district. TMC led Metiaburuz AC in the 2024 Lok Sabha elections and led Congress in the other elections.
More importantly, many MLAs are about to see a decline in the absolute number of votes cast in the 2026 elections, compared to the 2024 Lok Sabha elections. In 96 seats, the difference between the last tally of electors and the 2024 voters is less than 10%. Another way to look at this question is to simulate the required turnout which will ensure that votes cast in 2026 are not lower compared to the 2024 Lok Sabha. This number should be higher than 90% for 121 air conditioning units. 29 of them are Muslim dominated and the other 15 are Matua dominated as shown above.
So, is there gerrymandering?
It is impossible to answer this question without compromising statistical integrity, unless one has a complete religious analysis of the pre- and post-SIR list of Bengal. However, the argument can be made on the basis of probability by assuming that larger deletions in Muslim and Matua communities entail the possibility of deleting a greater number of Muslims and Matua than in other communities who also vote there.
There is another statistical test that can be applied. Were Muslim-dominated municipal councils and Matua among those with a higher turnout in 2024? The data show that a greater share of Muslim-dominated municipal councils were in the bottom half of municipal councils sorted by turnout, while the opposite was the case for most Matua-dominated municipal councils.
On the basis of this, one can come up with two arguments – one that the pre-SIR electoral list in Muslim-dominated municipal councils was more likely to have dead leaves than municipal councils in Matua, or two, that the greater probability of electing a Muslim made Muslim voters less willing to go out and vote. With the current elections being held against the backdrop of high deletion rates in Muslim advisory committees, Muslim voters may be more willing to exercise their votes than last time. We’ll know more when seat turnout data is released.
