The 1.08 Crore Question

In May 2023, Karnataka's 224 constituencies elected the state assembly: the Indian National Congress (INC) won 135 seats, the Bharatiya Janata Party (BJP) 66, the Janata Dal (Secular) (JD(S)) 19, and four went to others. In April and May 2024 the same 224 assembly segments voted again in the Lok Sabha election, and the Election Commission of India (ECI) published electors and votes polled for every one of them. In 2026, ECI ran a Special Intensive Revision (SIR) of the electoral rolls - a periodic, legally mandated re-verification in which every elector is sent an enumeration form to confirm they still live where the roll says they do. Forms that cannot be delivered, collected, or verified are marked "uncollectable" (ASDDO) and removed outright; electors who survive that first cut can still be issued a formal notice to produce further documents, and lose their place on the final roll if they don't (Exhibit 07).

This analysis compares ECI's own 2024 Lok Sabha turnout against ECI's own 2026 SIR numbers, constituency by constituency, to see what changed. Every turnout figure on this page is the 2024 Lok Sabha figure, counted within each assembly constituency. The 2023 assembly results (winner, party, margin) are used only to ask whether the pattern follows the political contest. It draws on the Commission's published statistical reports, its SIR summary tables, its constituency-level reason codes for removal, and its notice-service records - no outside or unofficial data was used. The full working dataset is in the ledger table below.

ECI's 2023 assembly-election data puts Karnataka's electorate at 53,131,579 (5.31 crore). By the 2024 Lok Sabha election it had grown to 54,772,332 (5.48 crore), and the 2026 SIR drew its base list from 5.54 crore - 6.60 lakh more than in 2024. The SIR then marked 10,796,415 names (1.08 crore) "uncollectable" (ASDDO) (19.5% of its own base), leaving a draft roll of 44,635,899 (4.46 crore). Statewide, that is 15.5% above the 38,643,669 people (3.86 crore) who voted in the 2024 Lok Sabha election - a statewide margin of 59.92 lakh between the SIR Draft Roll and 2024 voter turnout. That looks comfortable. Exhibit 03 shows it is anything but comfortable in Bengaluru.

Bengaluru's deletion rate is 45.1%, against 13.6% across the other 196 constituencies - more than three times higher. Its notice rate is even further apart: 42.2% of Bengaluru's draft roll has been served a notice, against 5.1% elsewhere. And its 2024 turnout was 54.8%, against 74.1% in the rest of the state. The 28 highest deletion rates in Karnataka are all Bengaluru seats: the lowest of them (Yelahanka, 36.0%) is higher than the highest rate anywhere else in the state (Gulbarga Dakshin, 34.0%).

The headroom tells the same story from the other side. Outside Bengaluru, the SIR Draft Roll is still 17.6% above the number of people who voted in 2024. Inside Bengaluru it is only 2.7% above - and, as Exhibit 05 shows, below it in 13 seats.

Across all 224 seats, the correlation between a constituency's 2024 Lok Sabha turnout and its 2026 deletion rate is r = −0.88. That is not just Bengaluru's low-turnout, high-deletion cluster carrying the number: leaving Bengaluru out entirely, the correlation is still r = −0.75, and each 10 points of lower turnout goes with about 6.5 points more deletion. Using the 2023 assembly-election turnout instead gives an even tighter fit (r = −0.92) - the constituencies that deleted the most in 2026 were already the ones that voted the least in 2023 and 2024. That is a pattern equally consistent with a demographic explanation (large migrant and renting populations, who are least likely to vote where they are registered, and most likely to have moved) or an administrative one.

The pattern is not confined to Bengaluru. Outside it, 21 constituencies exceed a 20% deletion rate, and almost all are seats centred on a district city: Gulbarga, Mysuru, Raichur, Vijayapura, Ballari, Tumakuru, Bidar, Belagavi, Hubballi-Dharwad, Davanagere and Shivamogga. At the other end, the lowest deletion rates in the state are coastal and Malnad seats such as Karkala (5.0%), Kaup and Belthangady, where turnout runs close to 80%.

That does not, by itself, indicate an error. Some voters will legitimately leave the roll between elections - they die, move, or are found duplicated elsewhere. But it creates a straightforward reconciliation question: what happened to the people who demonstrably voted in 2024? In Rajajinagar the draft roll is 5.1% smaller than the number of people who voted there in 2024.

Statewide, the largest single reason is "permanently shifted" (60.6% of deletions), followed by "untraceable or absent" (17.6%), death (15.2%) and "already registered" elsewhere (6.6%). But the mix changes completely at the Bengaluru boundary. In Bengaluru, 69.8% of deletions are people recorded as having shifted away and a further 22.0% as untraceable or absent - deaths are only 6.8%. In the rest of Karnataka, 21.7% of deletions are deaths and 10.6% are duplicates: closer to the roll clean-up the SIR is designed to be.

There is also a difference in who is being removed. Women make up 45.3% of Bengaluru's deleted names, but 53.1% of those deleted elsewhere - consistent with women who marry and move away being counted as shifted in rural constituencies, while Bengaluru's deletions are dominated by working-age migrants.

SIR is a two-stage filter. Stage one marks unreachable forms "uncollectable" (ASDDO) and removes them from the draft roll - that is everything in Exhibits 01-06. Stage two runs on whoever is left: a surviving elector can be flagged either as No Mapping (their record could not be linked to the roll of the previous SIR) or for an anomaly (a discrepancy in the record, such as an age gap with a listed parent or a name mismatch), and issued a formal notice demanding proof. An elector who does not respond is removed from the final roll when it is published, even though they were never marked "uncollectable" (ASDDO) at all.

Karnataka's SIR database records 43.81 lakh such notices - 9.8% of everyone who survived the initial cull, or 7.9% of the original 2026 SIR base. Add that to the 19.5% already marked uncollectable (ASDDO), and 27.4% of Karnataka's 5.54 crore-name 2026 base is currently either struck off outright or fighting to stay on.

In Bengaluru the figure is far starker: 68.3% of its SIR base is deleted or under notice, against 17.9% in the rest of the state. Bengaluru's notices are overwhelmingly No Mapping (74.1% of them, against 28.5% elsewhere) - people whose records could not be tied back to the last roll - consistent with a large, mobile population whose registrations have moved between constituencies.

Congress won 135 of Karnataka's 224 seats in 2023, the BJP 66, JD(S) 19, and four others. Split by 2023 winner, BJP-held seats have a much higher average deletion rate than INC-held ones - but that is geography, not politics: 16 of the BJP's seats are in Bengaluru, against 12 of the INC's. These are reported as plain distributions, not as a claim that either party benefited.

Margin of victory

In Delhi, the closer a race, the more its roll was gutted. Karnataka shows no such gradient. The correlation between a constituency's 2023 winning margin and its 2026 deletion rate is r = +0.12 statewide and r = −0.02 outside Bengaluru - effectively zero. Seats decided by under 1% lost a mean of 17.6% of their SIR base; seats decided by 20% or more lost 20.3%. Once turnout is in the picture, margin adds nothing.

Scale is a different question from correlation. The five closest contests in Karnataka in 2023 - Jayanagar (16 votes), Gandhinagar (105), Sringeri (201), Malur (248) and Kumta (676) - were decided by a combined 1,246 votes. Those same five constituencies recorded a combined 248,626 "uncollectable" (ASDDO) names in the 2026 SIR: 200 times the margin that decided all five races. Across the state, 173 of 224 constituencies have more names deleted than their 2023 winning margin (198 once notices are counted), and 61 seats were decided by under 5%.