Can We Trust EVMs? MP Election Vote Count Shows Huge Discrepancies

Despite the huge disparity between the two sets of vote counts from the same EVMs, the EC seems unconcerned.

(Photo: Shruti Mathur/The Quint)

Till now, nobody has been able to demonstrate whether Electronic Voting Machines (EVMs) can be hacked or tampered with. Now the question is, can EVMs malfunction?

As we know, on polling day, ‘a mock poll demo’ is conducted on EVMs, in the presence of the polling agents of various candidates, to ensure that the EVMs are functioning properly. Experts say about 3-10% of EVMs fail on that polling day.

So there is every likelihood of a certain percentage of EVMs again malfunctioning due to random causes during the long interregnum between the date of polling and the date of counting, an electoral expert told The Quint.

The data in The Quint’s possession will compel you to rethink – can we trust EVMs and all the accompanying processes?

The Election Commission (EC) had uploaded on its website two sets of data after the Madhya Pradesh Assembly polls concluded on 28 November 2018.

The first set of data refers to the total number of people who cast their votes in the EVMs on polling day. This data was uploaded sometime between the polling day and result day. The second set of data refers to the numbers of votes counted in the EVMs on result day. When we tallied the numbers of both sets, we found that there is a mismatch in 204 of 230 constituencies in Madhya Pradesh.

The graphics below show the difference between the two data sets in the top 10 selected constituencies. The comparison between the data shows that the Assembly Constituency-wise discrepancies ranged from a maximum surplus of +2605 to maximum deficit of -1831.

The top five MP constituencies where surplus votes were recorded on result day. (Graphic: Aroop Mishra/The Quint)

The top five MP constituencies where deficit votes were recorded on result day. (Graphic: Aroop Mishra/The Quint)

The Process of Counting Votes On EVMs

The Electronic Voting Machine (EVM) comprises of a Control Unit which is with the Presiding Officer, a Ballot Unit on which a voter votes for his/her candidate, the VVPAT Unit and VSDU. (Image: Shruti Mathur/ The Quint)

Most of us are not aware of the process of vote counting. The Quint spoke to a couple of government employees who were deployed on election duty to find out exactly what happens.

On polling day, the Presiding Officer is in charge of the Control Unit attached to the ballot unit. The voter can vote only after the ballot button on the Control Unit is pressed by the Presiding Officer. On polling day, this is the procedure laid down:

• The Presiding Officer needs to keep a record of the number of votes polled every 2 to 3 hours.
• After the polling is over, the Presiding Officer has to record the number of votes polled in the EVM machine. He has to write it on a piece of paper, sign next to the number of total votes, and paste it on the EVM machine before sealing it.
• He has to tally the total votes in the EVM with the number of voters who have signed next to their names in the register.
• The Presiding Officer has to hand over to each polling agent a copy of the account of votes recorded.

• On counting day, the total votes recorded in a particular control unit is tallied with the numbers recorded and if there is any discrepancy, this can be pointed out by the Counting Agents.

‘Human Error’?

The Quint contacted the Election Commission to find out the reason for the discrepancies in the number of votes counted on polling day and the number of votes counted on result day. In response, the EC said:

“For quick estimate of turnout, provisional data of votes casted collected within 24 hours of poll day 28 Nov & placed on CEO website. The votes polled are votes counted in EVMs on 11 Dec. There is only one figure of votes polled i.e, votes counted in EVM. Hence there is no discrepancy.”  

Election Commission

Maintaining that the first set of figures is unreliable, the EC dismissed the perceived discrepancy as “human error”.

An EVM expert, who had been a part of election duty, said on condition of anonymity:

“It cannot be a human error because there are clear rules laid down by the EC which are to be followed strictly. But it is difficult to explain what led to such huge discrepancies in the two numbers.”

EVM Expert

While the EC’s website says,

“The Presiding officer has to hand over to each polling agent a copy of the account of votes recorded. At the time of counting of votes, the total votes recorded in a particular control unit is tallied with this account and if there is any discrepancy, this can be pointed out by the Counting Agents.”

The EC says that the EVM count of the result day should be taken as final count. But experts say that the Presiding Officers are more reliable because they submit the voter turnout figures every hour and submit a final figure at the end of polling.

Were the counting agents aware of the discrepancies in the EVM votes?

According to the EC, the first set of data was uploaded within 24 hours after the polling was over in the Madhya Pradesh.

Interestingly, the EC hadn’t uploaded this first set of data in other Assembly elections like Uttar Pradesh and Chhattisgarh.

Following the EC’s reply denying any discrepancy in the tallies, we sent them another set of questions. We asked:

Q1. The data shows that the numbers were divided into three sections – male, female and third gender, which means some amount of time was devoted to preparing it. Then how is it that in some constituencies, the discrepancy is over 2,500 votes, while in some it’s nil?

The above data was uploaded by the EC after the polling was over on 28 November 2018.(Screenshot: Election Commission)

Q2. Why did the CEC upload incorrect data on their website? What was the rush?

The EC has not replied to these questions. It is not clear why the EC is maintaining opacity on a matter of such grave concern.

Scenarios in which the EVM count could go wrong:

• The Presiding Officers of 204 constituencies forgot to clear their mock polling data before beginning the real count on polling day. But this still can’t explain the negative figures.
• EVMs are malfunctioning because of the gap between the polling day and the counting day – experts say EVMs can sometimes malfunction when left unused.

Why Is EC So Unconcerned?

While the EC appears unconcerned with the huge discrepancy in the EVM vote count in MP, electoral experts say that there are carefully laid down procedures for counting the votes; wherever there is a discrepancy between the Presiding Officers’ figures of votes polled and the EVM count in any polling station, 100% of the VVPAT slips for that polling station must be counted, they said.

The difference in the numbers in the two sets of data is so glaring, yet the country’s election authority is brushing it under the rug by attributing it to a ‘quick count’ instead of ordering an inquiry.

We spoke to several former chief election commissioners, electoral and EVM experts but none of them could give a clear answer on what exactly has led to these discrepancies.

Rather, most of them were surprised and insisted that an investigation into the matter is needed.