The BJP was founded in West Bengal in 1951 as its predecessor, the Jana Sangh. Shyama Prasad Mukherjee, who started it, was a Calcutta man.

It took seventy-five years for his party to win his home state. The closest historical comparison is 2011, when the Left Front fell to the TMC after thirty-four years. Bengal’s electorate, when it turns, turns hard.

The headline number obscures a more uncomfortable one. The seat Mamata Banerjee herself won this election: zero.

She lost Bhabanipur, the south Kolkata constituency she has represented since 2011, to Suvendu Adhikari. Adhikari is the man she lost to in Nandigram in 2021. He moved to Kolkata to contest her this time and won by a margin; the count is still being finalised.

Only two Bengal chief ministers have lost their own seats before. Prafulla Chandra Sen in 1967. Buddhadeb Bhattacharjee in 2011. Banerjee is the third.

By Tuesday morning, Banerjee was telling reporters the election had been rigged. She refused to resign.

The 18th Legislative Assembly had reached the end of its five-year term anyway. On May 7, Governor R.N. Ravi dissolved it.

Banerjee’s office ended by the calendar, not by her concession. She has still not formally accepted the result.

The first story is anti-incumbency. After fifteen years, cut money and syndicate operations had stopped looking like rough edges of a party machine and started looking like how the state worked.

Then came the bigger wounds. The RG Kar rape and murder in August 2024 became a campaign-long issue, especially after the government’s initial advisory asking women doctors to avoid night duties.

Sandeshkhali came earlier. The TMC strongman Shahjahan Sheikh was accused of land grabs and sexual exploitation in early 2024. The grievance built up over the year.

By polling day, the two grievances had names on the ballot. In Panihati, the BJP candidate Ratna Debnath won her seat. Debnath is the mother of the doctor killed at RG Kar.

In Hingalganj, the Sandeshkhali survivor Rekha Patra won as a BJP candidate. The campaign had moved from abstract complaint to specific people, and voters knew who they were.

The second story is the SIR. Between October 2025 and February 2026, the Election Commission removed about 9 million voters from the West Bengal rolls.

That is close to twelve per cent of the state’s seventy-six million electors. The size of the operation was not in dispute. Its motivation was.

The result was probably wide enough to have happened on a clean roll. It would certainly have looked different on one of them.

Anyone who tells you which factor mattered more is guessing beyond what the data can show. Form-20 vote-share data is not yet public. Community-wise deletion breakdowns are still being compiled.

The honest answer is that both mattered. Which one mattered more is the argument over the next two years.

Below the statewide average, the deletion rate varied widely. The southwestern coast, where Adhikari is from, lost about 3.4% of its voters.

That coast is also where the BJP was already winning Lok Sabha seats. Whatever the SIR was correcting, it found less to correct there.

The industrial belt around Asansol, Burdwan, and Birbhum lost about 10%. The border districts touching Bangladesh lost about the same. Urban Kolkata lost roughly 14.

If the SIR was working as designed, you would expect this. Stable rural districts lose few names. Dense urban populations and migrant border districts lose many.

If it were working as a partisan tool, you would also expect this. The parts of Bengal where the BJP was not already winning were the ones with the most names to delete.

The geography is consistent with both readings. On its own, it cannot tell you which was operating.

The Election Commission’s argument is that some constituencies had decades of dead, duplicate, and shifted voters that earlier revisions had missed. The last comprehensive revision of the Bengal roll was in 2002. Twenty-four years of drift had accumulated.

The SIR used 2002 as its baseline. Voters had to trace themselves to an entry on that roll or supply documents linking them to it.

The Commission also added a new category, “Under Adjudication,” for names whose paperwork was contested. Three months of work were meant to clear two decades of backlog.

The opposition’s argument is that 37 per cent off any roll, for any reason, requires more due process than a three-month adjudication window allowed.

About three million appeals were filed against deletions. The Calcutta High Court was asked to assign 150 district judges to the appellate tribunals. Some appeals were heard before polling day. Most were not.

The Supreme Court, hearing the matter through April, declined to grant interim voting rights to those whose appeals were pending. Both positions had merit. Both were argued hard.

Heavy deletions follow the Bangladesh border. Maldaha Dakshin lost 16.5 per cent, Jangipur 14.5, Bangaon 11. Cooch Behar and the Alipurduar tea belt lost between 8 and 11 per cent. The pattern hugs the frontier.

Home Minister Amit Shah and the BJP campaign argued these districts had been padded for years with illegal Bangladeshi migrants, improperly added under the TMC governments. The SIR, they said, was correcting a problem previous Commissions had been unwilling to touch.

That argument is older than this election. Concerns about cross-border voter cards in Bengal go back to the 1990s and have been raised by parties across the spectrum.

The TMC’s reading is different. A large share of the deleted names belonged to Bengali-speaking Indian Muslims whose paperwork was inconsistent for reasons unrelated to their religion.

Marriage. Migration to other states for work. Parents who never bothered with school records.

CNN’s reporting from Murshidabad found a Kargil veteran who lost his vote because of a fifteen-year age gap between him and his mother on his birth certificate. The example is real and not unique.

Some share of the deletions probably removed people who should not have been on the roll. Some shares may have removed eligible Indian citizens. Without Form-20, the proportion is unknown.

The cleanest test of whether the SIR was a partisan instrument is what it did to people the BJP needed on the rolls. The Matua community is one such group.

Scheduled Caste Hindu refugees who came from East Bengal in waves starting in 1947, the Matuas have been one of the BJP’s most reliable vote banks since 2019. The community organised around the CAA's citizenship promise.

The BJP won fourteen of fifteen Matua-majority seats in 2021. The community’s loyalty had a number attached to it.

In April 2026, Newslaundry, The Quint, and The Federal all reported from the Matua heartland in North 24 Parganas. Thousands of Matua names had come off under the SIR’s “legacy linkage” rules.

The rules required voters to verify their identity on the 2002 electoral roll. Many Matuas had migrated to West Bengal after 2002. Their names came off.

The BJP appears to have won most of the Matua seats anyway. The broader point still stands: a revision that deleted Hindu refugees who were core BJP voters, Muslim TMC voters in the border districts, and long-resident urban voters in central Kolkata was indiscriminate enough to damage every party’s coalition.

Whether the remaining asymmetry was a feature or a bug of the design is what the courts are still hearing.

Central Kolkata is not on any border. It is 600 kilometres of water and rail line to the nearest crossing into Bangladesh.

But Chowrangee, Jorasanko, Bhabanipur, Ballygunge, Beleghata, and Entally all lost between 20 and 38 per cent of their voters in this revision. None of those neighbourhoods is migrant-heavy by any definition.

The Diplomat, drawing on the Sabar Institute's analysis, argued that the SIR’s documentation requirements were difficult for the population that actually lives in these neighbourhoods. Women who change their names after marriage. Families with several children. Migrants who came after 2002 from anywhere.

The Election Commission’s position is that messy urban paperwork is exactly what an intensive revision is for. Both can be true at once.

A quarter of central Kolkata, a TMC bastion since 2011, walked into the polling booth with fewer voters on the roll than it had on Diwali.

Bhabanipur lost about a fifth of its electorate. It also flipped to the BJP for the first time. The two facts sit side by side and cannot be neatly separated.

This does not prove that the deletions decided these seats. The deleted voters, had they remained on the roll, might have voted anyway. They might not have voted at all. There is no way to know with the data we have.

But the multiples close off one easy reading of the result. In these twelve constituencies, the size of the electorate’s removal cannot be cleanly bracketed from the size of the win.

Form-20 vote-share data, when published, will enable a more rigorous analysis. For now, the ratios sit on the record.

Satgachia and Chowrangee show the cycle in miniature. Satgachia: TMC won by 35,390 votes in 2021, and lost by 401 in 2026. The seat had 17,783 net deletions in between.

Chowrangee: TMC won both times. But the margin halved, from 45,344 to 22,002, while the constituency lost 74,118 names. Same map, two different shapes of result, one electorate cut by a third.

Two things probably happened together. Voters still on the rolls were unusually motivated this cycle, on every side.

Anti-incumbency drove BJP voters out. Anxiety about the next SIR drove TMC voters out. Matuas worried about citizenship and turned up to defend it.

Some Muslim and refugee voters feared that not voting might be used against them in future revisions, or to deny them government benefits. That fear may have done its own work at the booth.

The percentage also shifted because the denominator did. With about eleven per cent of the rolls gone, the same number of actual ballots produces a higher turnout percentage.

About 6.3 crore ballots were counted. That is around three million more than in 2021. More people voted in this election than in the last one. They just voted from a different list.

The seat tally does not capture the strategic weight of what changed hands on May 4. West Bengal is the state with the Chicken’s Neck.

The Siliguri Corridor is a strip of land in north Bengal, twenty-two kilometres wide at its narrowest. It is the only land bridge between mainland India and the eight northeastern states.

About forty-five million Indians live east of it. Bhutan is to the north, Nepal to the west, and Bangladesh to the south. The Chumbi Valley, which belongs to China, sits on the ridgeline above.

The Lowy Institute notes that a Chinese advance of about 130 kilometres through Chumbi would, in theory, sever Bhutan and the Indian northeast from Delhi.

A state government aligned with the centre on border security will manage that responsibility differently from one in confrontation with Delhi. Whether differently is better or worse depends on your priors.

For fifteen years, every conversation about the Siliguri Corridor had Mamata Banerjee on one end of it. That is no longer true.

Dhaka is watching for two specific reasons. The first is the Teesta River.

A water-sharing agreement between India and Bangladesh has been blocked for the better part of fifteen years. Banerjee refused to give up the share of water from her northern districts that they depend on for irrigation.

Her position was defensible for a chief minister whose farmers needed that water. It also meant Bangladesh, which loses an estimated 1.5 million tons of rice production each year to Teesta shortages, could not get the deal it had been seeking since 2011.

With Banerjee out, that question is back on the table. A Delhi-Kolkata alignment may finally make what was politically impossible for 15 years possible.

The second reason is harder. The Diplomat reports that between May 2025 and January 2026, Indian authorities forcibly pushed 2,479 people across the border into Bangladesh.

Border Guard Bangladesh said 120 of them were Indian citizens. A Bengal under a BJP government, with the SIR positioned as an anti-infiltration tool, will probably not slow that practice.

Whether you read push-ins as a legitimate response to infiltration or as a humanitarian concern depends on which side of the border you are reading from. And which community you belong to.

In the 2024 general election, Modi’s BJP fell short of an outright majority and had to govern with allies. The opposition INDIA alliance helped engineer that outcome.

Outside Congress, the alliance’s two principal pillars were the DMK in Tamil Nadu and the TMC in West Bengal. Both pillars went down on the same day.

Bengal fell to the BJP. Tamil Nadu fell to a film-star outfit, the Tamilaga Vettri Kazhagam, which has no clear position on Modi. Congress won Kerala. The BJP returned to Assam.

The map for the 2026 cycle does not look like the one for 2024. The opposition that hemmed in Modi in 2024 is harder to find now.

Whether that is good or bad for Indian democracy is a question with at least three reasonable answers. Stable governance, vigorous opposition, and pluralism in regional power centres are all goods. Each is partly served and partly damaged by what just happened.

The country will argue about it for the next three years.

The cleanest reading is also the most uncomfortable. The BJP won by a margin large enough to claim a real mandate.

The conditions of the election were also unusual enough that the size of that mandate will be argued about for years. Both statements are true at the same time.

The TMC had a real anti-incumbency problem after fifteen years. It found voters in 2026 in a way it had not in 2021. Cut money, RG Kar, Sandeshkhali, the night-duty advisory: none of these was a BJP invention. They were genuine grievances that a long-incumbent party had failed to address.

The BJP organised those grievances into a coherent campaign, Adhikari ran it, and the voters who remained on the rolls returned the verdict.

At the same time, 9 million voters were removed from the rolls before any of them voted. The deletions hit BJP-supporting Matuas, TMC-supporting Muslims in the border districts, and long-resident urban voters in central Kolkata.

About three million appeals are still pending in the tribunals. The cases are still in the Supreme Court. Both things were true on the same day.

Seat counts are from the Election Commission of India’s party-wise results page for West Bengal, last updated 9:45 PM IST on May 4, 2026. Falta goes to a re-poll on May 21, with results on May 24.

Form-20 vote-share data was not public when this was written, so the story does not make vote-share claims. AC-level deletion data is from the Sabar Institute’s compiled SIR dataset covering all 294 ACs, with state totals confirmed against the Election Commission’s published voter list comparing 7.66 crore (October 2025) and 7.04 crore (28 February 2026).

Anti-incumbency and Matua reporting draws on The Federal, The Wire, Newslaundry, The Quint, and CNN’s Murshidabad dispatch. Diplomatic context draws on The Diplomat, the Lowy Institute, Chatham House, Outlook India, and South Asian Voices.

This story does not try to settle which factor mattered more. The data needed to settle it, particularly Form-20, full Supreme Court adjudication, and community-wise deletion analysis, is not yet available.

The story instead tries to lay out what each side argues, what the data shows and does not, and what the result looks like once both stories are taken seriously.