Scrap EVMs and make voting, counting, and sealing of the ballot boxes more robust

A summary:       

Normal and rigged modes of EVM, switchable with secret signal. The current VVPAT system has several rectifiable problems. Solution for problems associated with current VVPAT. Problems with ballots getting into the hands of voters, and remedies . Providing voters with alternate ways of voting for a candidate:   Improvements in the counting process .Using the locking mechanism of ballot boxes properly. Improving transparency of the Election Commission officials . An idea for future consideration:         

Appendix-I: Various ways of manipulating EVMs

————————–

Scrap EVMs and make voting, counting and sealing of the ballot boxes more robust

A brief summary:

Concession on the part of ECI, to the use of paper trail, and acceptance of ‘supremacy of the VVPAT (Voter Verifiable Paper Audit Trail) slip counts over electronic tally of the ballots cast’ is acceptance of the fact that paper trails are more trustworthy than electronic counts. This is inherent in a rule stated on ECI’s website: “Rule 56D(4)(b) of the Conduct of Election (Amendment) Rules, 2013 provides for the primacy of the VVPAT slip count over the electronic tally of ballots cast and calculated on the EVMs.” There is no change in this rule so far as I know.

Just to prevent the delay in announcing the result of an election by a couple of days, relying on less trustworthy method of electronic counting votes via EVMs does not make sense compared to the more trustworthy hand count of a paper trail from VVPAT. It does not make sense to spend crores of rupees using a less trustworthy machine to decide the result of elections that decide the fate of 140 crores of people, and which is opposed by crores of people. It does not make sense to contribute to the lack of trust in the election process, the very foundation of democracy, in India. 

Despite hundreds, if not thousands, of requests to scrap EVMs, the ECI continues to insist on verifying only a random sample EVMs by counting their corresponding VVPAT slips (ballots). It can be shown (as has been argued in the Appendix of this article) that relying on verification of just a sample of EVMs is not good enough. There are ways of rigging the software that can avoid detection of manipulations. All of the VVPAT slips must be counted, for which the ECI is not ready, as yet.

However, if the ECI relents and allows all of the VVPAT slips to be counted, and results of the two kinds of vote counts yield contradictory results, it may lead to terrible riots in the street, with each group claiming victory by one of the two methods of vote counting. Eventually, the party or coalition that is stronger fighting is likely to emerge victorious – does not seem to augur well for peace-loving people who do not (or cannot) hoard up caches of deadly weapons.

If a rogue party has managed (with the help of loads of money) to get rigged software printed on the chips, they would know that they are winning the election with the help of manipulated counting of the EVMs and that without rigging they must have, (or may have), lost it. With that kind of foreknowledge, would they allow the hand counting of paper trails to be finished? Never, their jubilant supporters, celebrating the victory of their party based on the electronic count, will ransack the manual counting of the paper ballots. Probably they would make a bonfire of the paper ballots. They may claim that somehow the paper ballots, or the counts, have been manipulated. 

Therefore, given that electronic count is less reliable than hand count of VVPAT slips, we must scrap the use of EVMs in future elections. 

We can, and should, improve the VVPAT machine. The multiple red lights should be replaced by a single inexpensive mobile phone, which can serve the purpose of a monitor, to display the candidate’s profile (candidate’s name, party name & symbol, serial number assigned by the ECI, etc.). The voting process should be made a two-step process: (1) When a voter presses an appropriate key, s/he should see the candidate’s profile on a smartphone. The voter should then press the blinking PRINT button. Thereupon a screen capture should be sent to the printer, and the printer should print the chosen candidate’s profile. (2) When the voter is satisfied with the ballot just printed, s/he would go ahead and press the CAST-the-VOTE button. This would cause the ballot to get cut off and drop in the ballot box. 

We can also easily create the possibility of canceling one’s vote, particularly for the Presiding Officer (PO) to be able to test the correctness of the VVPAT machine. The canceled votes would have the word “CANCELLED” written at the top of a canceled vote, and for that reason, it would be a bit longer than regular votes, hence they can be easily weaned out of the heap of the ballots. 

The counting of votes can be expedited with the use of sorting and counting machines, and recounting in an adjacent room, as explained later.

Finally, our current locking system of ballot boxes (as well as the locking of the control units of the EVMs) with the help of paper tapes, duly signed by various poll agents [PAs], appears to be foolproof. But in fact, the key for such an elegant locking mechanism (paper tapes with the signature of poll agents present) is left hanging in the lock and never checked at the counting time to verify whether the seal/paper tape is the original one or not. Currently, the poll agents (PAs) do not take pictures of the paper tape and hand a full-scale printout of the same over to their party’s headquarters, to be given to appropriate counting agents (CAs). In the absence of a copy of the original paper tape, the counting agents (CAs) can do nothing but simply watch as one of the counting officers (COs) pulls out the paper tape and throws the same into a garbage bin. Thus, the CAs do not have a clue whether the ballot box has been opened before they are opened for counting of votes, or replaced by another box. That is the reason why, particularly the rogue parties are interested in getting the Returning Officers (ROs) and other top officials of their 

Without a copy of the original key, the CAs can do nothing but be silent spectators of ballot boxes being opened. The first thing that the CAs should do is to make sure that the paper tape taken out of a ballot box is indeed the original one. After the ballots are dumped on a table, the first thing that the COs need to do is: count the canceled and uncanceled votes and verify these counts against the report regarding the same written by the PO and countersigned by the PAs of various parties and inserted into the ballot box. Dropping the PO’s report into the box is the last thing to be done before its slot (for dropping the ballots into the box) is closed off when a spring-loaded self-locking shutter is pulled into place. After the slot is closed off, nothing can go into the box and nothing can be brought out. This way, the parties can rest assured that the ballot boxes cannot be tampered with before the counting time, no matter who is the RO, or any other election official.

For proper utilization of the paper tapes, all parties as well as Election Commission’s officers need to be properly trained, with explanations regarding proper use of the paper tapes. They are simple but elegant locking mechanisms if only they are properly used. 

More detailed discussion: Normal and rigged modes of EVM, switchable with secret signal

EVM chips can be designed to work in two modes: (1) normal mode, and (2) abnormal mode. In the abnormal mode, the EVM can be made to do tricky things depending upon the kind of signals it receives by secret codes broadcast to it via radio waves like those used by mobile phones. The EVM can be designed to work in normal mode initially so that no amount of testing can reveal that it can also work in abnormal modes to favor candidates of a rogue party. But on the receipt of secret codes, it can do tricky things to favor a rogue party. 

To make the process of counting votes faster Electronic Voting Machine (EVMs) were constructed and manufactured on a large scale in the first decade of 2000. Soon people realized that electronic counting could not be relied upon. This is the gist of the video and texts on the website: https://indiaevm.org/, entitled “India’s EVMs are Vulnerable to Fraud Hari K. Prasad, J. Alex Halderman, Rop Gonggrijp”

Very capable technical Expert, Mr. Hari Prasad, Prof. Halderman from Michigan University, USA,  and Dr. Gonggrijp from Holland had undertaken thorough research to meet the challenge of the ECI to hack their EVM, which the ECI had claimed to be hack-proof. Realizing the importance of their research, an ECI or government official had made one of the EVMs, used in elections in India, available to them. 

These experts showed that by replacing or adding some of the chips in the EVM it can be hacked to change the outcome of the election. In their demonstration, as documented in the video and associated texts at the above website, they conclusively proved that EVM software can be rigged to change the results of an election, make a losing party winner, and vice-versa.

The ECI was so perturbed by the excellent research and demonstration of the above-mentioned group of experts on EVM, that the ECI charged them with theft of the EVM, even though the experts openly claimed that the EVM they were demonstrating had been used in Indian elections, but, apprehensive of any repercussion against the official who made the EVM for their research, they did not want to divulge the name of the official who had made the EVM available to them for research purpose.

The ECI got Mr. Prasad arrested, and jailed, and they hounded the two professors to make a quick exit from India, lest they too would be arrested. Instead of charging them for the theft, they should have been decorated and rewarded for their research to save democracy in India. This was proudly announced by the ECI way back in 2010 and it continues to be the stand of the then Chief Election Commissioner (CEC), even today even after he has retired from government service. (Watch “Ex CEC S.Y. Quraishi ने Election commission पर क्यों खड़े किए सवाल ?” on YouTube https://youtu.be/o41fAPbFmY4?si=bdJrZy0xPNacFw31)

I wish the interviewer had asked of Dr. Quraishi: What is more important? Is it the supposed theft of a decrepit EVM (to research it to prove that EVMs are not safe machines to be used for elections), or saving democracy in India and around the world, by proving that EVMs are not trustworthy machines? I wish the interviewer had pointed out to Dr. Quraishi other things like the possibility of delay in manipulation of the vote counts (with the help of secret signals to be sent to EVMs) until after vote counts on randomly selected EVMs have been verified by counting corresponding VVPAT slip counts, to evade detection of possible manipulations. I wish the interviewer had asked some other technical questions showing the vulnerability of EVMs.

Further, I have heard some people suggesting that the ECI should allow the EVM software to be examined by independent “experts”. First of all, it is the so-called experts who have brought in the EVM. In the voting device, we need something that ordinary people can understand without the intervention of the so-called “experts”. Nobody should be asked to rely on the trustworthiness of anyone, expert or otherwise. Most importantly, even if we do have trustworthy software, is it going to be very difficult for a person with unfathomable moneybags to bribe the company that prints the software onto chips to print a rigged software on the chips and throw away the original software that they were approved by the experts? 

Once a program/software is imprinted on a chip, no one can read it. The only way to test it is by giving it different kinds of input and looking at the output. If the output happens to be as expected, the machine has to be certified as good. But, as has been pointed out earlier, rigged software can be made to work in (1) normal mode, and (2) rigged mode. The modes can be switched from one to the other with the help of secret codes sent to the machine via radio waves like the ones used by mobiles. In the normal mode, the machine is supposed to work as a normal unrigged good machine would, and no amount of testing would reveal that it has been rigged. Only the people who know the secret codes can show the tricky things that the rigged EVMs can do. 

So, people should not fall into the trap of getting the software examined by “independent experts”. We must stay away from electronic counting of the votes, period. And that the VVPAT system should be made more robust.

In the following, we will show that (1) the use of VVPAT can be continued, (2) the counting of the paper slips can be improved as outlined below. And, (3) the paper-tape sealing of the ballot boxes can be improved.

Current VVPAT system has a number of rectifiable problems

Currently, a voter presses a button with the candidate’s profile (name of the candidate, affiliated party name, party symbol, and any serial number provided by the Election Commission of India), right above the vote button. All of the vote buttons are arranged in a row, 16 of them on one voting module. If the number of candidates is more than 16, more than one voting unit would be needed. As a voter enters a poll booth, s/he sees one or more voting modules, and s/he finds the button corresponding to his/her favorite candidate on the voting module and presses the same. A red light right above the vote button is supposed to light up, indicating that a vote has been cast in favor of the intended candidate. But the voter does not know for sure where the vote went. A vote slip/ballot gets printed behind a little glass window and stays there for 7 seconds before vanishing (cut off automatically from the paper roll to drop into the ballot box). The voter is supposed to go in a hurry and check the printed ballot to make sure that the ballot is indeed in favor of the candidate s/he intended to vote for. 

Secondly, even if the voter sees that the candidate’s ballot (candidate’s profile) printed was different from the one s/he intended to vote for, it will be difficult for him/her to prove to the Presiding Officer (PO) that s/he intended to vote for a certain candidate but the ballot that got printed was in favor of some other candidate. The PO will have to cast some test votes and see whether it is printing the correct ballot or not. At the end of the test voting, if the VVPAT seems to be working flawlessly, the complainant voter will have to pay a fine of Rs. 10,000/- or something like that. Besides, the voter will have to face the humiliation of raising a false alarm. For all these reasons most of the voters seldom challenge the PO to subject the EVM to test-voting. Further, in the case of test voting, the PO and the voter who forced the test votes had to face the ordeal of filling out Form 17A and 17C, so that they may be taken into account at the time of vote counting. On account of these kinds of complications, the PO is generally very reticent to the idea of casting test votes, and the voter would be even more reluctant to complain on account of the stipulated fine, despite his/her having noticed that the ballot was indeed in favor of a candidate s/he did not intend to vote for.

BTW, the rigged/rogue software may have been programmed to vote incorrectly only for alternate voting. So, the voter’s next vote (test-vote) would correctly print his/her intended candidate’s ballot, proving the candidate’s complaint to be wrong, and thus liable for the stipulated fine, and voting would continue on the rigged VVPAT-EVM combination. On the other hand, if the software in one EVM is rigged, in all likelihood all of the machines are rigged. If it is conclusively proved that all machines have rigged software, the whole election will have to be canceled and we will have to wait until new EVMs have been manufactured and distributed all over the country. 

Moreover, the glass window used to be clear glass until 2017, and the ballot paper used to stay behind the window for 15 seconds for the voter to observe it to his/her satisfaction. Since 2017 the clear glass has been replaced by tinted black glass through which one cannot see anything unless the object behind it is illuminated with light behind the black glass. And the duration for which the ballot can be seen has been cut down to 7 seconds. Recently one of the activist scientists/engineers has demonstrated how votes can be manipulated behind the tinted glass, with the help of suitable software. (Watch “Ex CEC S.Y. Quraishi ने Election commission पर क्यों खड़े किए सवाल ?” on YouTube https://youtu.be/o41fAPbFmY4?si=bdJrZy0xPNacFw31 )

To this day I cannot comprehend the rationale behind the use of black tinted glass over the printed VVPAT slips. I would love to see the reasons given by the ECI for the use of tinted glass. As someone said in Hindi: “Daal men jaoor kuchh kaalaa hotaa hogaa kaale glass ke pichhe.”

Solution for problems associated with current VVPAT

These kinds of problems with the current VVPAT can be alleviated with the following two-step voting process, including the replacement of the red light above each of the keys corresponding to various candidates with a monitor (possibly a simple smartphone with the ability of screen capture that will get printed as ballot slip in the printer, as discussed below) with suggested audio instructions playing while the voter is voting.

In the following explanation of the new improved voting process, there will be audio instructions playing most of the time after a voter enters a polling booth until s/he comes out after casting his/her vote. The suggested texts of the audio have been printed in bold fonts. 

As soon as a voter goes into the polling booth, a green light near the door should change to red indicating that someone is currently voting in that booth. At the same time, an audio recording should start playing with something like the following message: Welcome to this polling booth. As soon as you are ready to vote, please press the blinking START button. The audio may restart if the voter does not press the START button within a few seconds. 

When the voter presses the START button, the audio will say: Thank you, now you can go ahead and press the key corresponding to the candidate’s party symbol and name. When the voter presses the key of his/her choice, a more detailed profile of the voter should be displayed on the monitor or the mobile phone. Possibly the ECI may prescribe the format of the first page of the official flyers of candidates running in an election. It may include the candidate’s photo, in addition to the candidate’s name, party name & symbol, and serial number assigned by the ECI. It may also include a bar code and QR code, the purpose of which will be discussed below. The bar code and QR code may expedite the process of counting with the help of sorting machines. The details on the first page would be considered to be the official profile of the candidate. This official profile is the one that should be displayed on the smartphone (functioning as a monitor) when a voter presses the corresponding key. When the voter presses the PRINT button, a smaller-scaled copy of the official profile of the candidate should get printed as his/her ballot. The second page may contain other details of the candidate like his/her qualifications, experience, manifesto, endorsements, etc.

At this point, the audio should say: Thank you for making your choice. Are you satisfied that the profile displayed does correspond to your intended candidate? If not, press the START button again and try pressing your candidate’s button again. If the displayed profile does not correspond to the desired candidate a couple of times, there may be something wrong. Possibly the labels on or above the keys have been glued incorrectly. At this stage, the voter should complain to the Presiding Officer (PO) and should alert the polling agents as well regarding the problem he/she is facing. On the other hand, if the monitor or cell phone display does correspond to the intended candidate, the audio should say: Thanks for making your selection. Now, please go ahead and press the blinking PRINT button.

As soon as the voter presses the PRINT button, the printer should start churning and print a screen capture of the profile seen on the monitor. The audio should then be saying: Thanks for trying to print the profile of the candidate of your choice. Now, please take a look at the printout behind the glass window near the sign YOUR BALLOT is printing here. After the printer has finished printing the ballot, the sign would change to: YOUR BALLOT has been printed, as can be seen behind the glass window here, where the CAST-the-VOTE button is blinking.

Then the audio would continue with the instruction: “Please take a good look at the ballot printed. If the ballot printed is for your intended candidate, please press the button CAST-the-VOTE”. The audio should broadcast this repeatedly until the voter presses the button CAST-the-VOTE. Needless to say, we would ban the use of tinted glass over the printout of the ballot. Only clear glasses will be used.

As soon as the voter presses the CAST-the-VOTE, the printed ballot should get cut off and drop into the ballot box. The audio should then start saying: Thank you very much for casting your vote. Many freedom fighters had to spill a lot of their blood and sweat for you to be granted this privilege to choose their own government. Celebrate that privilege and enjoy this glorious day. 

Now you can pick up your things and let the next voter come in and vote. The audio can repeat this last sentence until the voter has left. 

After the voter leaves the polling booth, the audio can stop and a light outside the booth would turn green from red, indicating that it is ready for the next voter to come in. As soon as the next voter goes into the polling booth, the green light should change to a red light, and the voting process (inside the booth) would start all over again with the welcome sentence. 

In the above voting process we did not see the need for canceling any vote because the voter can visually see the intended candidate’s profile on the monitor (it’s no longer just a red light), which will continue to be displayed for the voter to see and compare with the profile printed on the ballot paper. So, no crooked designer/creator of the VVPAT section would dare design it such that the theft would be immediately caught. However, a vote-cancelling mechanism should be made available so that the PO may be able to resort to a test vote and cancel the same without the necessity of lengthy documentation process of filling out Form 17A & 17B. This test vote and cancellation of the same can be provided, as follows:

The PO may be provided with a CANCEL-the-VOTE switch/button. While a ballot printout can still be seen behind the clear-glass window, on pressing the CANCEL-the-VOTE button, the word CANCELLED should get printed right above the ballot already printed, without severing this CANCELLED portion from the ballot already printed. After the word CANCELLED has been printed, it should be severed from the paper roll. The two together (ballot for an unintended candidate [or the test-vote] and the label CANCELLED) will be longer than usual ballot length, hence easy to be weaned out from a heap of regular ballots during the counting process. With this possibility of cancellation of a vote before it is cast, a voter may even be able to change his/her mind just before casting his/her vote. 

At this point allow me to mention that some people have proposed getting the ballot into the hand of the voter and then the voter would drop that ballot into the ballot box. There is one little problem here that these people are overlooking. A voter may be given some money by a rogue agent to vote for a specific candidate, and then instead of dropping the same into the ballot box, bring the ballot out and give it to the rogue agent, and collect the rest of the vote-selling price. After the rogue agent collects a bunch of ballots in favor of the rogue candidate s/he favors, the agent may hand those ballots into the hands of a more trusted voter who is inclined to vote for the rogue candidate already, requesting the latter to drop all of the votes into the ballot box. To stop this kind of cheating one has to figure out a way to prevent the ballots from being taken outside the polling booth for being sold to rogue agents.

To stop ballots from being taken out of a poll booth one may resort to a mechanism that is used to prevent taking a book out of a library without getting it properly checked out at the checkout desk. The mechanism used by the libraries comes in the form of strips made of ferromagnetic material that is not magnetized at first. In that state it will respond to a low-frequency magnetic field by producing harmonics of the driving frequency that can be picked up in the theft detector pedestal at the door of the poll booth and an alarm will ring. To disable it, it is magnetized far into the saturation area. Perhaps, one can mix a fine powder of ferrite in a thin paint-like liquid/glue and paint a thin stripe of the same on the ballot paper. This will prevent the ballots from being taken out of the poll booths, thus preventing ballot theft and buying & selling of votes. However, I am not sure how much ferrite would be needed and how practical this would be.

Problems with ballots getting into the hands of voters, and remedies

The access of a voter to the paper ballot would require major modification of the ballot paper to prevent it from being smuggled out of the voting booth and being sold to rogue agents. In the two-step process described above, when the ballot paper does not get severed from the paper roll “automatically,” there is ample time in the hands of the voter to be satisfied that the printed ballot is indeed for the candidate the voter intended to vote for. 

For the sake of argument, in the unlikely case, if, on inspection of the printed ballot, it is found to be of an unintended candidate, i.e., the profile on the ballot is different from what is still being displayed right above the voting module, the voter would have two options: (A) Report to the PO that the intended candidate, still displaying on the monitor, is different from the what is printed on the ballot. At the same time, the voter should, with permission of the PO, alert all polling agents of the clearcut fraudulent ballot printing. (B) If somehow, the voter wants to keep his/her vote confidential even from the PO, s/he should be able to simply request the PO to cancel and try voting again. The PO should cancel a vote when requested, without asking any question. If in the next voting ballot printout is the same as the intended one, s/he can press the CAST-the-VOTE button, and forget about the mismatch the previous time. However, if the voter repeatedly finds a mismatch between the ballot printout and the display of the intended candidate, the PO will have to tell the voter that s/he (the PO) will have to observe the mismatch by himself/herself before canceling a vote again. At this point, all of the polling agents will need to be alerted regarding the mismatch between the printed ballot and the displayed ballot. The mismatch will be right there for the whole world to see. There will be no need to fill out Forms 17A and 17B. The voting will have to be stopped, at least in that polling booth and the matter thoroughly investigated. 

I said, “In the unlikely case,” because given the robust system described above no one is going to have the gumption to do that kind of mischief by rigging the software or hardware. It would indicate something is wrong with this machine or all machines. In any case, if the PO wants to cast a few test votes, there should be a provision for “test voting”. 

Thus the two-step process is: (A) the voter can see the ballot of the candidate s/he intends to vote for on the monitor, and then only s/he can press the PRINT button; (B) then go over to the printer and after verifying that the printed ballot is indeed for the intended candidate just like the candidate-profile (ballot) displayed of the monitor, the voter would press the “CAST-the-VOTE” button. The printed ballot would then get cut off and drop into the ballot box. The ballot would not get automatically cut off from the paper roll and would not drop into the ballot box automatically. 

Providing voters with alternate ways of voting for a candidate:

There is just one more little modification that can make it easy for the ECI to accommodate many more candidates than just 16 or so. The modification is that aside from the piano-type keyboard voting unit, the voter may be provided with two more alternatives to cause his/her favorite candidate’s profile to be displayed on the smartphone serving the purpose of the monitor. 

(a) The voter should be able to type in the candidate’s Serial number on a keypad. This addition will allow for the accommodation of a much larger number of candidates to be chosen from. 

(b) Further, a QR code may also be made a part of the candidate profile. The QR code can simplify getting one’s favorite candidate’s profile displayed on the monitor. All that a voter will need to do is to bring his/her favorite candidate’s flyer with the QR code into the poll booth and place the same at a designated place (frame) in front of the smartphone serving the purpose of a monitor. When the camera of the smartphone detects a QR code in front of its camera, it will pull out the profile of the candidate from its memory and display the same on its screen. Thus the same smartphone can serve the purpose of the camera to see the QR code on a candidate’s flyer as well as display the candidate’s profile. 

When the voter is satisfied that the displayed profile is of the intended candidate, s/he can press the PRINT button to print the same profile onto the ballot paper. The smartphone can then screen-capture the ballot on its screen and transmit the same to the printer for printing. If the voter is satisfied that the printed ballot is one meant for his/her favorite candidate, (still on display on the smartphone monitor,) s/he will press the CAST-the-VOTE.

Further, if a bar code could be made a part of a candidate’s flyer, sorting the ballot papers may be expedited, as explained later.

Improvements in the counting process

This completes the voting process. Next, we will see how the counting process can be improved. 

If the ballots are made on somewhat sturdy paper similar to money bills, they may be amenable to counting by machines like those available to bank tellers. The ballots will need to be oriented (orientation may be necessary particularly if machines are to be used for sorting with the help of bar codes on each ballot), sorted, and finally, counted, with spot checking by manual counting, or full manual counting.

The bank tellers always count the money bills by hand and then run them through the machine to verify their hand counts. These machines are usually taken help for instant verification of the manual ballot counts.

One big advantage of the alternative ways of inputting candidates’ info into the system is that when the number of candidates is more than what the keyboard type of input system can accommodate or deal with, the keypads may come in handy. The QR code too can easily facilitate a large number of candidates running in an election, as well as candidates for different elected bodies. 

The other advantage of the keypad system is that it would be easy for the ECI to set up a mock polling booth in every panchayat office for the people to come and get familiar with the voting process. The whole VVPAT setup can be left in the panchayat office so that they can make use of the same for their own electoral processes as well as for any referendum or initiatives, making direct democracy possible at the local level. 

When the machine is being used for training the voters, instead of real ballot paper they could arrange for projecting the favorite candidate’s flyer/profile at the place where that VVPAT slip was to be printed and displayed for the voter to see and be satisfied that the ballot has been correctly printed. If the voter is satisfied with it (the projected ballot), pressing the CAST-the-VOTE key near the mock printer would cause the mock ballot to vanish (as if it has been cut and made to drop into the ballot box). Otherwise, pressing the CANCEL button will print the word “CANCELLED” at the top of the un-detached unintended ballot. Ultimately, when the voter is satisfied that the ballot already printed has been canceled s/he can press the CAST-the-VOTE button. Doing so will cause the unintended ballot together with the word “CANCELLED” will vanish, as if they were cut and dropped into the ballot box. In this kind of system, a voter will get trained to vote without wasting ballot papers.  

A computer too should be provided in the panchayat office so that a voter, as well as a candidate, should be able to type in her/his name or birthday or other particulars and search for their names and other particulars, and make sure that they are in the system and that the particulars about them are correct. If not, they should be able to complain to the appropriate authorities and get the mistake(s) rectified. When a voter knows the ID number of the candidate s/he intends to vote for, the candidate’s profile should be displayed on the monitor/smartphone. The voter can then go on and finish the mock polling exercise. Alternatively, if the voter has the official flyer of the candidate s/he intends to vote for, the flyer should be placed in front of the smartphone. This too should cause the profile of his/her candidate to be displayed on the smartphone. 

Further, they should also be able to run training videos for the prospective polling agents as well as counting agents, and also, prospective  Counting Officers (COs) as well as Presiding Officers (POs) (and any other officers involved in the election process). In these training videos, the locking mechanism for the ballot boxes, as well as control units (if electronic counting has not been abolished yet), should be shown. Proper training of voters, candidates, and polling & counting agents cannot be over-emphasized. 

We will now pay attention to the locking mechanism of the ballot boxes and control units (if electronic counting has not been abolished yet.)

We have observed that a counting officer (CO) brings out a ballot box, showing that the box-opening hole is still covered by unpunctured paper tape. S/he asks the counting agents if the paper tape is still un-punctured. The counting agents (CAs) speak in unison that it is not punctured. The CO then proceeds to puncture the paper tape with a finger and then uses the finger to push or pull the spring-loaded latch inside the ballot box to open the same. S/he then pulls out the paper tape and goes ahead to drop the paper tape in a waste basket. None of the CAs ask for the paper tape to make sure whether the tape is the original duly signed paper tape signed by various polling agents, or not. None of the CAs have been given copies of the duly signed original paper tape. 

In the absence of the verification, of whether the paper tape is the original one bearing the original signatures of the CAs present at the polling booths, the paper tape locking mechanism is like an elegant lock that appears to be locked but has never really been properly locked; it just looks locked. With the polling agents (PAs) signing the paper tape, the key has been prepared. But the PAs never take a copy of the key and give the same to CAs for comparing it with the keys (paper tapes) taken out of the ballot box to make sure that the paper tapes are indeed the original ones signed by the poll agents at the booths. During the days that pass by between voting and counting, or during their transport from the polling booths to the strong rooms, the ballot boxes may have been replaced with new boxes with similar outside insignia. It is only if the paper tapes pulled out of the box are compared with copies of the original paper tapes that one can be sure that the box is the same original box checked and locked in the presence of various polling agents, and that it has not been tampered with or replaced.

To rectify the above-mentioned lacuna, the poll agents as well as counting agents have to be properly trained. They have to be trained to understand the necessity of taking pictures of the duly signed paper tape,  getting the same printed, and giving the printouts to their respective party headquarters. Then it will be the duty of the head-quarter officials to duly distribute the copies of the paper tapes to proper counting agents. Also, the necessity of comparing the tapes pulled out of the ballot box and control unit against photocopies of the tapes supplied to the CAs via respective party headquarters cannot be overemphasized. 

Using the locking mechanism of ballot boxes properly

To use the locking mechanism properly, the polling and counting agents as well as officers have to be instructed/trained appropriately, as shown below:

1. Check the poll boxes to make sure they do not already have any votes in them. 
2. Put your (polling agents’) signatures on the paper tape that is to be used for sealing the ballot box. 
3. Get a carbon copy of the paper tape, or if there are too many polling agents and not enough carbon copies, take a picture of the duly signed tape in your smartphone. The next day, the PAs who took pictures of the paper tapes should get the pictures printed to the size of the original tape. These pictures and carbon copies (from those lucky enough to get them) must be handed over to respective party headquarters (PHQs). The PHQs must provide pictures of the paper tapes (or the carbon copies) to appropriate Counting Agents (CAs) on the counting day(s). 
4. Make sure that the Presiding Officer (PO) inserts the paper tape properly in appropriate slots, so that the paper tape covers the “opening hole” on the box, and then shuts the lid of the ballot box, which gets locked with the help of a spring-loaded locking mechanism. Make sure that the box is indeed locked internally and cannot be opened. After the box is locked, it cannot be opened without puncturing the paper tape and inserting a finger into the “opening hole”  to open the spring-loaded mechanism that locks the box lid in place. 
5. The box should have a slot at the top through which the ballots/paper slips can drop in after they are cut off when the voter presses the CAST-the-VOTE switch next to the printer.
6. The poll agents then have to sit there, keeping a watch, making sure that nothing fishy happens during the voting phase. 
7. A Close of Voting Period Form (CVPF) should be filled by the PO and signed by him/her and counter-signed by the polling agents present at the close of the voting period. The CVPF should be the last thing dropped into the ballot box. The CVPF should contain the number of voters who have voted and the number of canceled votes. 
8. There must be a provision in the box to pull a self-locking shutter to cover the slot so that nothing can be dropped into the same after the box’s shutter is closed, nothing can be taken out as well. This provision may be in the form of a spring-loaded shutter that can be pulled to shut the slot, and the spring-loaded mechanism would prevent it from opening without opening the box by perforating the paper tape already inserted. The polling agents should make sure that after they countersign the CVPF form, and drop it into the box, the box’s slot is shuttered and locked with the self-locking mechanism.
9. The polling agents should get a printout of the pictures of the duly signed and countersigned CVPF form, and hand the printouts over to their respective Party Head Quarters (PHQs). Thus there are two printouts to be handed over to the PHQs: (a) Printouts of the duly signed paper tapes by the morning polling agents, and (b) copies of the CVPF form duly signed by POs and countersigned by the polling agents present in the evening at vote closing time. 
10. When the transport bus comes to pick up the vote box, if there is enough room, the polling agent should hop in and accompany the ballot box to the strong room. 
11. The poll agents (PAs) should make sure that their respective PHQs give the paper tape copies (or carbon copies) and copies of CVPFs to the counting agent (CAs) on the counting day. 
12. At the counting time, the Counting Agents (CAs) should be ready with the pictures of appropriate duly signed original paper tapes for sealing the ballot boxes, to compare them with the original tapes, as they are pulled out of the ballot boxes by the counting Officers. They should also compare the CVPF forms taken out of the box with the copies of the pictures of duly signed CVOF forms taken by their polling agents at the close of the voting time.
13. The CAs should insist on taking a close look at the paper tapes as well as the CVPF, as they are pulled out of the ballot boxes by COs, and they (the CAs) should compare them with the appropriate photocopies given to them by their party headquarters. If they do not match, the box is considered as tampered or changed and the ballots in them cannot be relied upon. New voting must be ordered in such cases, for the concerned polling centers. In case of large-scale tampering, constituency-wide re-polling has to be ordered.

After the CAs make sure that the paper tape that was used for locking the ballot box was the original tape, and hence the box has not been tampered with or replaced, the contents of the ballot boxes can be dumped over a counting table. Counting officers will first fish out the CVPF form and hand it over to the CAs of various parties to compare the same with their copies in the hands of CAs. If the original paper tape and CVPF match with their copies in the hands of CAs, we can be sure the ballot boxes have not been tampered with. If not, the boxes have been tampered with and votes at least at the concerned booth have to be canceled and re-polling ordered. If any tampering is found, perhaps re-polling of the whole constituency has to be ordered. 

After the paper tape and CVPF are found to match the copies of the same in the hands of the CAs, we can be reasonably sure that the boxes have not been tampered with. We can then wean out the canceled votes and count the valid votes as well as canceled votes and make sure they tally with the numbers given on the CVPF.

The polling booth’s valid ballots are now ready to be mixed with those of other booths so that one would not be able to tell which colonies/communities voted or did not vote for whom. This is done to prevent bias in the minds of elected officials against communities that did not vote in favor of the elected officials. 

After the valid votes of a sufficient number of booths have been mixed, each of the ballots may be properly oriented (topside up) so that mechanical sorting and counting machines, if available, may be used for help and verification. 

After orienting the ballots properly, they may be sorted mechanically (with the help of a bar code on each ballot), and spot-checked manually. 

The sorted ballots may then be counted manually and verified mechanically a couple of times, or vice-versa, as the bank tellers do. 

Votes for different candidates may be bundled separately and their votes may be added together to make sure that the total tallies with the original counts before orienting, sorting, and counting. If they do not tally, the counting will have to be repeated manually.

Besides, the counted and bundled ballots can be sent over to another room (without letting them know of counts for the various bundles) to count the ballots for various candidates all over again. The counting agents in the second room would simply have to do recounting of the correctly sorted and oriented ballots. They would not have to orient and sort all of the ballots, saving a lot of time.

If everything appears to be AOK, the bundles have to be sent over to RO. The RO will keep count of each of the bundles under the hawk eyes of the CAs of various parties. ROs,  however highly designated, should be taken to be above any kind of suspicion. At the hands of COs, small-scale pilferage may take place. But at the hands of ROs, if improprieties take place, they would more like be like robberies, and they do happen. 

That’s why concerned governments pay special attention to the appointments of friendly ROs. That’s why at least at the RO level it is crucially important to have two ROs from the opposing sides, one from the incumbent government side, while the other one should be picked by the opposition. In case of conflicting opinions of the two ROs, the matter may be referred to a judicial bench.

Improving transparency of the Election Commission officials

All individuals, including government officials, vote for one party or the other. They can surely be supposed to lean in favor of their party. Therefore among the officials involved in elections (i.e., among the ECI officials) parties on the ruling side and those in the opposition, should be equally represented.

In the USA, for all practical purposes, they have only two parties. So, in the USA, they employ Republicans and Democrats equally (on a 50-50 basis) at all levels of the hierarchy of their Election Commission. Even the temporary election workers are chosen from the two parties on a 50-50 basis. In India, we can consider the ruling coalition as one party, and those outside the ruling coalition as the other party, and have them nominate their representatives in the election commission. We know that the two coalitions change from one election to the other. So, the employees of the Election Commission may have to be changed from one election to the next. We need to do the best we can for the sake of transparency in the Commission. 

They can be paired up (with one from each side of the divide, ruling side/coalition, and opposition side/coalition) for various duties so that no one can manipulate by himself or herself. Just because someone is a permanent employee of ECI or the government, one cannot be assumed to be unbiased. All are biased, firstly in favor of their caste, then the alliance their caste is a part of. The importance of pairing up opposite sides increases, with higher hierarchies. Even at the highest level there should be equal representation from the incumbent government and the opposition side. If the disagreement between the two sides cannot be resolved amicably, the matter will have to be decided by the judiciary, different judges at different times, so that different biases of different judges may, perhaps cancel each other.

An idea for future consideration:

All individuals, including government officials, vote for one party or the other. They can surely be supposed to lean in favor of their party. Therefore among the officials involved in elections (i.e., among the ECI officials) parties on the ruling side and those in the opposition, should be equally represented.

In the USA, for all practical purposes, they have only two parties. So, in the USA, they employ Republicans and Democrats equally (on a 50-50 basis) at all levels of the hierarchy of their Election Commission. Even the temporary election workers are chosen from the two parties on a 50-50 basis. In India, we can consider the ruling coalition as one party, and those outside the ruling coalition as the other party, and have them nominate their representatives in the election commission. We know that the two coalitions change from one election to the other. So, the employees of the Election Commission may have to be changed from one election to the next. We need to do the best we can for the sake of transparency in the Commission. 

They can be paired up (with one from each side of the divide, ruling side/coalition, and opposition side/coalition) for various duties so that no one can manipulate by himself or herself. Just because someone is a permanent employee of ECI or the government, one cannot be assumed to be unbiased. All are biased, firstly in favor of their caste, then the alliance their caste is a part of. The importance of pairing up opposite sides increases, with higher hierarchies. Even at the highest level there should be equal representation from the incumbent government and the opposition side. If the disagreement between the two sides cannot be resolved amicably, the matter will have to be decided by the judiciary. 

Appendix-I: Various ways of manipulating EVMs

Everything that can be done with the help of adding the remote-controlled chips or hardware, can be done by including programs with the same capability into the original chip in EVM. So, all one needs to do is to have a chip printing/etching company print a rigged program on the chip in place of the original program that was supposed to be printed on the chips. The rigged software may be able to work in at least two different modes: (1) a Normal mode, (2) an abnormal mode.) In the abnormal mode, it may be able to do different things per the kind of secret codes/signals sent to it via radio waves like the ones used by our mobile phone, as outlined below.

There are innumerable ways one can think of rigging them. Rigging of the EVM may be of two types: (1) EVM may be programmed to favor a rogue candidate “during the voting period,” or (2) it may be programmed to favor a rogue candidate “after voting has finished”. The second type of rigging may be found to be easier to do without being caught. So, we have described the second type of rigging.

In the abnormal mode, it can be made to do unthinkable things. For example, at the close of the voting time, an EVM may be made to broadcast at a certain secret frequency, its location, its serial number, and votes in the bins of all candidates. The rogue party which got the rigged software printed on the chips would receive the information broadcast by various EVMs at their closing time, and get a panoramic view of the voting over the whole country. If it is losing, or not winning by as large a margin as it would have liked, it can pick and choose some of the EVMs at particular locations with the help of their serial numbers broadcast earlier by those EVMs for manipulation of the kind outlined below.

The rogue party would wait until any randomly selected EVMs have been checked by comparing their electronic counts for various parties and comparing the same with corresponding VVPAT slips’ counts in favor of those parties. After such random testing has finished, the rogue party may send secret signals on a secret frequency to a selected few EVMs to manipulate vote counts in them in such a way that half of the votes in all candidates’ vote bins are taken out and added to the bin of the rogue party’s candidate. After the maldistributed votes for various candidates have been read and recorded by various observers, the software may reset the vote counts in the bins to their original count. Such a resetting would prevent the detection of rigging via checking a random selection of the EVMs after all counting has finished. Such a rigging can only be caught only by full counting of all VVPAT slips. 

The election process in India is stretched over a month or so anyway. So, most would not mind an additional delay of a couple of days for the sake of avoiding electronic count via untrustworthy EVMs. But somehow the ECI does not want to count all of the VVPAT slips. They want to count only a random sample of the electronic counts that have been verified counting the ballots in corresponding boxes of VVPAT slips. However, it is easy to see that the rogues may wait to manipulate the machine counts (for example, by shifting half of the votes for all candidates to a favored rogue candidate) until random samples have been checked for their correct counting. 

Random checks even after all of the counting is finished can be made to avoid detection if all of the manipulated machines have been reset back to their original counts just before they are turned off. Thus, with random checks on the EVMs either at the beginning of the counting period, or after the counting has finished, or both, a properly designed rogue software can avoid detection of its manipulation. 

The only way to be sure about detecting electronic count manipulation is to count all of the VVPAT slips. 

Dr. Satinath Choudhary, Retired from college teaching, but active as an activist for worthy causes.

Email: [email protected];