Control All, Delete

SIR exclusions in West Bengal were connected to AITC lead and Muslim population

Former sex workers collect forms from polling officials, as part of the Special Intensive Revision exercise, in the Sonagachi area in Kolkata. Dibyangshu Sarkar/AFP/Getty Images
01 June, 2026

We’re glad this article found its way to you. If you’re not a subscriber, we’d love for you to consider subscribing—your support helps make this journalism possible. Either way, we hope you enjoy the read. Click to subscribe: subscribing

The Bharatiya Janata Party’s landslide victory in the West Bengal assembly election, held in April, has been clouded by allegations of electoral malpractice. At a press conference on 5 May, a day after the election results showed the BJP winning 207 of the 294 seats, Mamata Banerjee, who had been chief minister for the past 15 years, refused to follow convention and voluntarily resign. “Why should I step down?” she said. “We have not lost. The mandate has been looted.”

While she also alleged that violence and intimidation by paramilitary forces had affected the counting process, Banerjee’s claim of being “defeated not by public mandate but by conspiracy” took aim at the Election Commission of India’s Special Intensive Revision, which had resulted in over 9 million names being removed from the state’s voter rolls. According to data compiled by the SABAR Institute, a Kolkata-based research group that painstakingly processed millions of documents published by the ECI during the SIR process, the number of deletions exceeded the margin of victory in 102 of the 207 seats won by the BJP.

As the political scientist Gilles Verniers argues in this issue of The Caravan, the SIR makes it impossible to reach any definite conclusions about the verdict thrown up by West Bengal. Several analyses published in the aftermath of the election have attempted to quantify the SIR’s impact on the result, estimating how many excess seats the BJP gained from the deletions. Any such exercise “is, at best, incomplete and, at worst, meaningless,” Verniers writes. “We can only count the votes that were cast. We cannot recover the votes that were never allowed to be cast.” While we cannot reliably conclude whether the BJP would have won the election even without the SIR, we can study the deletions themselves. After examining the SABAR Institute’s data, I found some troubling patterns that raise questions about the ECI’s impartiality during the process.

The SIR was first carried out in Bihar, a few months before the 2025 assembly election, in which the BJP-led National Democratic Alliance won more than eighty percent of the seats. The ECI’s booth-level officers—most of whom are employees of the state government, such as schoolteachers and anganwadi workers—first examined whether someone’s name appeared in the voter rolls from 2003, when the SIR was last conducted. If it did not, the person had to furnish one of 11 documents to prove their date and place of birth, as well as those of one or both of their parents, depending on when they were born.

Thanks for reading till the end. If you valued this piece, and you're already a subscriber, consider contributing to keep us afloat—so more readers can access work like this. Click to make a contribution: Contribute