Election Commission ignored mismatch between votes ‘cast’ and votes ‘counted’: Reporters’ Collective

• The Election Commission maintains an internal Register of Voters (Form 17A) and an Account of Votes Recorded (Form 17C). The two counts must match before results are declared

• ECI does not reveal the counts publicly. In 2024 when ADR asked for the booth-wise records entered in Form 17C, the Commission declined

• The Reporters’ Collective investigation by Sumedha Mittal, Venkatesh Nayak and Firquan Ameen accessed internal reports related to 63 parliamentary constituencies and 467 assembly constituencies in the three states of Maharashtra, Haryana and Jammu and Kashmir. Mismatches were found in 52 constituencies

The Election Commission keeps two sets of records at polling booths of every voter who walks in. Three polling officers are involved. One checks the eligibility and identity of the voter. The second records the details in Form 17A and issues a slip. The third checks the slip and allows the voter to cast his vote. Votes are required to be cast in the same order as recorded in 17A.

When the polling ends, the presiding officer pushes on the ‘close’ button on the control unit of the EVM which shows the number of votes cast, which is then recorded in Form 17C. The two counts should ordinarily match.

However, what happens if a voter walks into a booth but cast his votes more than once? What if the voter walks into a booth but refuses to cast his vote? What if a voter walks into a booth and is prevented from voting by officials for breaching rules and the sanctity of the secret ballot? Such cases, it is believed, are rare and comprise a miniscule percentage of votes. However, there does not appear to be any provision to record such cases and identify the voters involved.

An investigative report by The Reporters’ Collective published on Tuesday, 10 November 2026, however, reveals that in 19 of the seats analysed, the Register of Voters actually showed ‘Nil’ indicating that nobody voted. In some constituencies, on the contrary, the entire voting population was recorded as having voted in the Register. Clerical errors or not, the shoddy recordkeeping, says the report, raises questions on the conduct of elections and the quality of supervision.

In both Badkhal assembly constituency in Haryana and in the North Mumbai Lok Sabha constituency, for example, the Register of voters mentioned the number of registered voters (3.3 lakh and 18.11 lakh respectively) as having voted. Votes counted were 1.60 lakh in Badkhal and 10.33 lakh in Mumbai North. On the other hand, in two assembly constituencies, Kaithal in Haryana and Srinagar in Jammu and Kashmir won by the Congress and the National Conference respectively, the records suggest that more votes were counted than the votes cast.

When the Collective contacted the Election Commission for its version, the ECI treated the queries as grievances and asked the Collective to share the data in strict condfidence. It assured that the data would be kept confidential and would not be shared with anyone. The Collective has responded by publishing the questions it had asked and the official reply it received. It would rather share the data publicly, it has maintained.

The Election Commission has previously explained that differences between votes polled and votes counted can occur if mock-poll data are not properly cleared before voting begins. Mismatches between machine records and Form 17C records at a polling station too can occur in some cases but generally the votes may not affect the final result if the margin exceeds the number of disputed votes.

The investigation by The Collective does not prove rigging but does raise questions related to recordkeeping, conduct of elections and accountability.