Archive for March, 2017

Electronic Voting Machines (EVMs) used in Indian Elections

March 26, 2017

In the recent elections to five provincial states of India, Electronic Voting Machines (EVMs) were used for all the voters. After the massive victory of central ruling party, BJP, there were many baseless complaints about EVMs, raised by responsible politicians, including sitting and former chief ministers. Though they did not provide any proof of rigging of the machine, their allegation surely creates some doubts in at least some small sections of the people. There are many features in the EVM which prevents any type of mechanical, electronic or network based fraud. The actual Balloting Unit (BU) is only a slave unit to the main Control Unit (CU). The micro programmed chip which resides in the CU is manufactured and programmed abroad. Once programmed, it cannot be altered in any way. The CU does not have any remote input and hence cannot be controlled from remote by any network device. There are many tests done at the booth level before the machine is put to use. These tests are witnessed and approved by the booth agents of the contestants, before the machine is finally sealed off securely and thereafter, will be always under the watchful eyes of police, election officials and the agents of contestants till the results are downloaded and declared. All these features are generally accepted and agreed upon by all the politicians and the voting public. However a few doubts are expressed in the following areas.

The Doubts:
The program inside the CU could be biasetowards one particular button. This doubt is easily answered, as button numbers of candidates of different political parties are different at different constituencies and doing an en-mass biasing of buttons is not a choice at all. In addition a mock polling test is conducted one hour before the polling in front of the agents and officials, with more than 50 votes polled at random and the results shown to all. After this test the CU is sealed and secured. There is a demo of this mock polling, available on U-tube as below:

The following are some of the questions answered by Election Commission in their website:

Q11. Can booth – capturing be prevented by the use of EVMs?
Ans. By booth-capturing, if one means, taking away or damaging of ballot boxes or ballot papers, this evil cannot be prevented by the use of EVMs as EVMs can also be forcibly taken away or damaged by miscreants. But if one looks at booth capturing as a case of miscreants intimidating the polling personnel and stamping the ballot papers on the symbol and escaping in a matter of minutes, this can be prevented by the use of EVMs. The EVMs are programmed in such a way that the machines will record only five votes in a minute. As recording of votes has necessarily to be through Control Unit and , whatever be the number of miscreants they can record votes only at the rate of 5 per minute. In the case of ballot papers, the miscreants can distribute all the 1000 odd ballot papers assigned to a polling station, among themselves, stamp them, stuff them into the ballot boxes and run away before the police reinforcements reach. In half- an –hour, the miscreants can record only a maximum of 150 votes by which time, chances are the police reinforcement would have arrived. Further, the presiding Officer or one of the Polling Officers can always press the “close” button as soon as they see some intruders inside the polling station. It will not be possible to record any vote when once the ‘close’ button is pressed and this will frustrate the efforts of the booth-capturers.

Q21. Is it possible to program the EVMs in such a way that initially, say upto 100 votes, votes will be recorded exactly in the same way as the `blue buttons’ are pressed, but thereafter, votes will be recorded only in favour of one particular candidate irrespective of whether the `blue button’ against that candidate or any other candidate is pressed?
Ans. The microchip used in EVMs is sealed at the time of import. It cannot be opened and neither any rewriting of program can be done by anyone without damaging the chip. There is, therefore, absolutely no chance of programming the EVMs in a particular way to select any particular candidate or political party.

Q24. In the conventional system, it will be possible to know the total number of votes polled at any particular point of time. In EVMs ‘Result’ portion is sealed and will be opened only at the time of counting. How can the total number of votes polled be known on the date of poll?
Ans. In addition to the ‘Result’ button, there is a ‘total’ button on EVMs. By pressing this button the total number of votes polled upto the time of pressing the button will be displayed without indicating the candidate-wise tally.

Q28. In the conventional system, before the commencement of poll, the Presiding Officer shows to the polling agents present that the ballot box to be used in the polling station is empty. Is there any such provision to satisfy the polling agents that there are no hidden votes already recorded in the EVMs?
Ans. Yes
Before the commencement of poll, the Presiding Officer demonstrates to the polling agents present that there are no hidden votes already recorded in the machine by pressing the result button. Thereafter, he will conduct a mock poll by asking the polling agents to record their votes and will take the result to satisfy them that the result shown is strictly according to the choice recorded by them. Thereafter, the Presiding Officer will press the clear button to clear the result of the mock poll before commencing the actual poll.

Q29. How can one rule out the possibility of recording further votes at any time after close of the poll and before the commencement of counting by interested parties?
Ans. As soon as the last voter has voted, the Polling Officer in-charge of the Control Unit will press the ‘Close’ Button. Thereafter, the EVM will not accept any vote. Further, after the close of poll, the is disconnected from the Control Unit and kept separately. Votes can be recorded only through the . Again the Presiding officer, at the close of the poll, will hand over to each polling agent present an account of votes recorded. At the time of counting of votes, the total will be tallied with this account and if there is any discrepancy, this will be pointed out by the Counting Agents.

My Suggestions:
In addition to all the above features, to improve the voter confidence, Supreme Court had ordered a system called Voter Verifiable Paper Audit Trail (VVPAT). Under this system, the Balloting Unit, BU, includes a sealed printer with a viewing window and sealed box underneath where the paper strip falls after Voter verifies his voting preference as printed therein. This will help in actual counting, in case of any major contest about the result. Though it is not clear how this system will be used when it comes to wider use, my suggestion is the following:
These voting slips should actually be counted as a sample in some random booths to check whether it follows the general trend of voting pattern in the respective areas of polling. For example we may decide on actual slip counting, in 10 polling booths in a bye-election, in 30 polling booths throughout the state in a state assembly election and in 100 polling booths throughout the country in case of parliamentary elections. In case of any major discrepancy in voting trends, more PU’s may be examined for further investigation. In case of any legal demands, full counting of the slips may also be ordered by the court. This would definitely improve the confidence of voters and candidates on the EVM procedures.
Here is my another suggestion to increase the voter confidence: it is to allow for mock-polling within the actual polling period also. Two such intermediate mock polls may be allowed at the times chosen by the agents. During this mock poll, the times of the commencement and end of mock poll, and the actual voting pattern may all be noted and stored in the control unit (CU), for later verification at the time counting . During these intermediate mock polls, say for about 15 minutes each, actual polling may be stopped and resumed immediately afterwards. The CU programme should include a feature to find the ‘RESULT’ between different times also. This will help the agents at the time of final counting, to check and verify whether the intermediate mock poll results tally with what is already noted down. This timed counting feature will also help to detect other polling frauds, if any, found in future.

With this feature, the results when downloaded will be interpreted as in the example below:
Total votes polled – 1,25,372.
This may be verified against actual votes polled added with total intermediate mock poll votes polled.
Intermediate mock poll results are first retrieved from CU, as below:
1) From 11.30AM to 11.45AM – 57 Votes – Party-wise : A – 25, B – 18, C-4, D-10
2) From 3.15 PM to 3.30 PM – 66 Votes – Party-wise : A – 20, B – 22, C-11, D-13
Total – Mock Votes – 123 – Party-wise : A – 45, B – 40, C-15, D-23
Verified with the actual Test data and found correct by Election officials and agents.
Actual votes polled – 1,25,372 – 123 Mock Votes = 1,25,249
Party wise votes:
A – 36,253 – 45 Mock Votes = 36,208
B – 42,117 – 40 Mock Votes = 42,077
C – 26,318 – 15 Mock Votes = 26,303
D – 20684 – 23 Mock Votes = 20 661
Total – 1,25,372 – 123 Mock Votes = 1,25,249
Accordingly B will be declared as the winner.

It is sincerely hoped this will convince all the political parties about the use of EVMs. More than that, voters will not have any doubt on the election process. What is required is for the machine to be redesigned in a way that you may get the results in a time tagged manner. As far I could see, the counting procedure does not seem to include any facility for printing of the result. Such a facility may also be considered, as manual noting down of the result form the machine display is susceptible to human errors and mischief. It is now for the Election Commission and EVM designers to take up the issue. Long live Indian Democracy.

Post Script:

As I published the above blog on EVMs on 28th March 2017, I was disturbed to know from the media about the recent fiasco of EVM/VVPAT in MP, while testing and demonstration. VVAPAT was apparently found to print the same election symbol irrespective of the button pressed on Balloting Unit. The Election Commission has not only not offered any explanation for the fiasco, they stayed away from assuring the voting public any credible investigation. This definitely reduces the confidence of Voters like me on these machines. I also understand the concerns of politicians. However the solution is not to discard the EVMs and to go back to cumbersome Paper ballots. On the contrary we should find ways of improving the reliability and tamper-proof quality of EVMs. None of the politicians have suggested any solution but only blamed Election Commission. Response of EC is also far from satisfactory. While it is okay for normal politicians to immediately ask for debunking of the EVMs (with or without VVPAT), I thought, we can do something better than all of them. We can suggest solutions to make these machines better. As a voter I want the EVMs to continue with more improvements as we go along.

I have already suggested two solutions as above. The first suggestion is as explained in earlier paragraphs to allow for intermediate mock polls at random times during polling. My second suggestion was to decide on the operation VVPAT machines and declare openly about the mode of its usage, post completion of counting. Now my third suggestion is to improve the operation of VVPAT itself. Actually, the news came about the VVPAT fiasco, as I was posting this suggestion on VVPAT.

VVPAT machine is connected at one end to Balloting Unit (BU) and at the other end to Control Unit (CU). When voting button is pressed, the BU prompts the VVPAT printer to print the slip, which goes into the box after verification by the voter, (if at all he is smart enough to verify and report in case of any mismatch). VVPAT also sends this voting info to CU for recording it as a vote cast. Here is the catch. As a voter I prefer to verify what is recorded in CU and not what was registered in BU/VVPAT. Hence my suggestion is: Connect BU directly to CU as existing in the normal EVM. Connect VVPAT to CU and let VVPAT print what is recorded in CU after the vote is cast. Additionally, VVPAT can have another output to BU to blink the vote-recorded button LED for 6 Secs. This way voter will find it easy to verify the functioning of EVM, thus boosting his confidence level. Having a VVPAT machine in between Balloting Unit and Control Unit, introduces a source of tamper and hence must be avoided.

Presently VVPAT machine is connected as below:







The voting process will be:

– Press the required button in BU

– Look for the VVPAT to printout your selection

– Check whether your selection is shown correctly in the printout

– Look for the printout slip to get cut properly and falls down in the sealed box below.

Now you may leave the booth.

Suggested change is as below:









When connected like this, flow of voting information reverses. VVPAT will be able to now print what is registered in the Control Unit instead of what is selected in the Ballot Unit. We can use the ballot button lamp more intelligently as per the voting process described below:

– Press the required button in BU

– The blue LED behind the button will light up

– The CU will register your vote and communicate the same to VVPAT and the BU

– The Blue LED will start blinking six times to indicate your choice as recorded by the Control Unit

– At the same time VVPAT will print out the selection as recorded by CU.

– You may verify both as above and leave the booth.

I request the media and other powerful NGO for democracy to take up this suggestion seriously and improve the reliability and tamper proof nature of EVM and VVPAT machine. Let us make our electoral democracy fair and just.