A Record Turnout, Read as a Protest
On the day West Bengal's first phase of voting closed, turnout crossed 92 percent — by most tallies the highest ever recorded in the state. Political camps read the number in opposite directions: the Trinamool Congress called it a rebuke to the Special Intensive Revision (SIR) of electoral rolls; the Bharatiya Janata Party called it a rejection of Mamata Banerjee's fifteen years in power. Sabir Ahmed, a Kolkata-based researcher who has spent over a decade studying West Bengal's electoral rolls, sees something more specific in the numbers: an angry, organised response from precisely the districts where the SIR cut deepest.
Ahmed points to Murshidabad, a Muslim-majority district that lost large numbers of names in the revision. There, he says, migrant workers travelling home from Delhi, Pune and Kerala paid inflated fares — flights that normally cost a couple of thousand rupees selling for fifteen thousand — rather than lose their chance to vote, driven partly by rumours that skipping this election could cost them their names a second time. "They wanted to participate in the election process," he says of the returning migrants. In Shamsherganj, one of the constituencies with the heaviest deletions, turnout by early afternoon had already crossed 60 percent among both men and women. To Ahmed, that is less an electoral swing than "a political society" insisting on being heard.
Inside the Special Intensive Revision
The SIR set out to re-verify West Bengal's electoral rolls from scratch, tracing each voter's lineage back to the 2002 rolls. Ahmed's research began by mapping where the exercise flagged people as unlinked to that base year. In Muslim-majority pockets of Murshidabad — Farakka, Bhagawangola, Shamsherganj — the "unmapped" share was under 2 percent, he says, while it ran to 10 or even 15 percent in the Matua belt, settled by families who migrated from East Bengal. That contrast, he argues, embarrassed the Election Commission, which did not want the exercise to look like it was singling out Muslims. What followed, in his account, was a shift to document-by-document scrutiny using artificial intelligence — and the emergence of a new category with no basis in the SIR manual: "logical discrepancy."
Ahmed illustrates the category with his own name — Sabir Ahmed, which a schoolteacher once rendered with an extra letter — and with the dozens of ways Mohammad, Sheikh or Shabbir can be transliterated across documents. Large families drew scrutiny too. "Why do you have six siblings?" was, in his telling, treated as a red flag, though nothing in the constitution limits family size. In Bhabanipur, Mamata Banerjee's own constituency and roughly 20 percent Muslim, Ahmed says the share of logical-discrepancy cases came in far higher than that — over 40 percent. In Metiabruz, a roughly 60-percent-Muslim area, he puts the discrepancy share at 90 percent. "This clearly indicates this is a targeted intervention," he says, arguing that probing voter rolls for Muslim names is trivial once artificial intelligence is used to flag them.
The Weight of "Under Adjudication"
Names flagged for logical discrepancy were pushed into a category called "under adjudication," requiring citizens to appear before a hearing with documentary proof — passports, birth certificates, decades of paperwork. Ahmed offers a family example: both his brother and his father were placed under adjudication despite holding passports, which themselves require rigorous prior verification. His father's name was eventually restored to the rolls; his brother's was not. He says the hearings were often conducted by judicial officers posted from outside Bengal who could not read the Bengali-language documents in front of them, working through a caseload of more than one crore citizens in roughly two weeks — a pace he calls "not humanly possible."
The numbers, in his account, kept growing. An initial figure near 90 lakh flagged names was followed by a further 27 lakh deletions after adjudication, taking cumulative deletions to roughly 91 lakh voters — a total Ahmed says includes some genuinely deceased people, but far more than that can explain. In Nandigram, the constituency Shubhendu Adhikari won by a margin of around a thousand votes, Ahmed's team crowdsourced deleted names from residents and found the deleted list was overwhelmingly Muslim in a constituency that is roughly a quarter Muslim; journalists searching for non-Muslim names among the deleted found barely over two hundred in the entire seat. He says his organisation has published similar raw, constituency-level breakdowns for Murshidabad, north Bengal and Kolkata neighbourhoods like Bhabanipur, Jorasanko and the Port area, all showing the same skew.
Compounding this, Ahmed says, is a translation problem: voter rolls outside Kolkata are compiled in Bengali, and Arabic, Persian or Urdu-origin names lose their form when transliterated into Bengali and back into English — Abdul Jabbar rendered as "A Jabbar," Santu Pal as "Sabzi Pal." He says his team has kept screenshots of dozens of such errors as evidence of the process's unreliability, alongside cases of living people marked dead, residents marked "migrated," and voters deleted over duplicate entries they had no chance to contest before the fact.
Costs Beyond the Ballot
Ahmed frames the SIR's consequences as extending well past voting day. Because welfare entitlements are linked to the voter card, he says being placed under adjudication has, in practice, frozen land transactions and suspended access to government schemes for affected families. A separate strand of his organisation's research has looked at the exercise's toll on mental health: he describes people arriving at their legal-aid centre in tears, and a young man from Murshidabad who told him he could no longer sleep, afraid his family would be sent to a detention camp after his parents' names dropped off the rolls even as his own remained.
His team has also tried to price the exercise in economic terms — photocopying, travel between offices, a day's wages lost by daily-wage workers navigating the hearings — arriving at an estimate running into thousands of crores of rupees spent by citizens simply to prove they belong on a list they were already on, alongside an estimated 1.16 crore hours of lost school time as booth-level officers, many of them teachers, were redirected to the exercise. "The state is dealing with its own citizens in a very crude manner," he says.
A Slow, Unequal Appeal
Ahmed is scathing about the appeals process that followed. The tribunal set up to hear challenges was named after Shyama Prasad Mukherjee, a choice he calls symbolically loaded given the numbers involved — of the lakhs excluded, he says barely under 200 names had been restored through the appellate route by the time of the interview. The application process itself, he argues, favours the already privileged: appeals must be filed online with a 1,000-word cap, a serious barrier in a country with wide digital disparity, while offline filers were not even issued receipts. Groups with the least documentation to begin with — homeless residents, sex workers, transgender people, families displaced by river erosion, people whose only past address proof was an electricity pole the Election Commission had once agreed to accept — are, in his account, the ones least equipped to fight their way back onto the rolls.
Echoes of Assam
Ahmed repeatedly draws the comparison to Assam's National Register of Citizens, calling the Bengal exercise "a backdoor NRC" — a process that, whatever its stated purpose, ends in citizens having to prove they belong to the only country they have ever known. He notes that Assam's NRC ultimately excluded large numbers of Hindus as well as Muslims, which he says is one reason it came to be seen as a failed exercise, and that Bengal's Matua community — of East Bengal origin like many Muslim voters affected here — has at least a legal route to citizenship under the CAA that does not exist for Muslims left off the rolls. He says his own analysis of Election Commission data turned up no confirmed case of a Rohingya or Bangladeshi national voting fraudulently in Bengal, even as politicians invoked exactly that narrative to justify the revision. "These are our own citizens, our own voters," he says.
What Comes After the Count
Ahmed's closing argument is that the fight over deleted voters should not end with the election result. He wants the restoration of wrongly deleted names to be the first priority of whichever government forms next — and says civil society intends to press that demand regardless of outcome. He also appeals for Bengal's intellectuals and civil society to carry the issue to a national audience, arguing the pattern he has documented is not confined to one state. Whatever the scoreboard shows after counting, he suggests, the deleted voters' cases will still be open, and so will the question of who is willing to keep fighting them.
Discussion
0Readers only · Comments are reviewed by the desk before they appear
The journalism is free for everyone — create a free account to join the discussion.