The EVM Debate in India — What Each Side Actually Claims
Every Indian election produces a fresh round of allegations about Electronic Voting Machines (EVMs). The 'EVMs are rigged' camp wants paper ballots back. The 'EVMs are foolproof' camp dismisses all concerns as conspiracy. Both framings miss the actual technical question. Here is what each side claims, and what the evidence supports.
What an Indian EVM is
An Indian EVM is a standalone two-piece machine: a Ballot Unit (with the candidate buttons) and a Control Unit (where the count accumulates). The machines run firmware burned to a one-time-programmable chip, are not networked, and store votes in a serial-write memory that cannot be wirelessly accessed. Since 2017, every Polling Station also has a Voter-Verifiable Paper Audit Trail (VVPAT) printer that produces a paper slip the voter can verify before it drops into a sealed compartment.
What the critics claim
Three categories of claim. First, hardware tampering — that machines can be physically modified before polling day to favour a particular candidate. Second, software tampering — that the firmware itself can be altered or that compromised machines can be substituted. Third, manipulation of the VVPAT-EVM count reconciliation — that VVPAT slips are not actually counted in enough constituencies to verify the EVM count.
What the ECI claims
First, EVMs are not networked, so remote tampering is impossible. Second, the firmware is burned to a chip that cannot be reprogrammed without physically replacing the chip, and chip-level manipulation requires factory access. Third, every EVM is sealed and the seals are checked before polling by candidates' agents. Fourth, the VVPAT reconciliation is done in five randomly-selected booths per assembly constituency.
Where the evidence actually sits
On the hardware/software side: the ECI's claims are technically defensible. The architecture is conservative by design (no networking, OTP firmware, physical seals). No public demonstration of an in-the-wild EVM manipulation that survives ECI's seal-checking has been produced. Some demonstrations have been made on prototype machines or simulated environments, which are not the same as production machines.
On the VVPAT reconciliation side: the critics have a stronger point. The 2019 Supreme Court ruling (N Chandrababu Naidu v Union of India) increased VVPAT reconciliation from one booth per constituency to five. Critics argue that five booths is a tiny statistical sample given the scale (a Lok Sabha constituency has roughly 2,000 booths). The April 2024 Supreme Court ruling (Association for Democratic Reforms) declined to mandate 100% VVPAT counting, but did mandate that any mismatch in the reconciliation would trigger 100% recount for that constituency.
The proportionate position
Neither 'paper ballots back' nor 'EVMs are perfect'. The proportionate reform is: expand VVPAT reconciliation to a statistically rigorous sample (10–15% of booths per constituency), publish booth-level VVPAT vs EVM count data, and make all chip-level audits independently verifiable. None of this requires returning to paper ballots, and all of it would close the credibility gap.
Further reading
Electronic voting in India on Wikipedia: en.wikipedia.org/wiki/Electronic_voting_in_India. The N Chandrababu Naidu v Union of India judgment (2019) and the Association for Democratic Reforms v ECI judgment (April 2024) are both on the Supreme Court website.
“Main bhi cockroach. Tum bhi banno.”