Electronic Voting Machines: Are They Safe?

Right since its first trial in 1982 in a bye-election in Kerala, India, doubts have been expressed by stakeholders like contesting candidates and political parties as to whether Electronic Voting Machines are safe and tamper-proof. Even though use of EVMs was made universal from the 2004 Parliamentary General Elections, as recently as on 20 June 2008, on the eve of municipal elections in the state of West Bengal, political rivals clashed over the issue of testing EVMs and a 12-hour strike was called in Burdwan by three political parties. After the 23 February 2008 Assembly Elections in the state of Tripura, the Tripura Congress demanded a National Police probe on suspected EVM manipulations. On 9 June 2008, a delegation of Trinamool Congress urged the Election Commission of India not to use EVMs in the 2009 Parliamentary Elections. In respect of the 2007 Manipur elections, an application was filed under the Right to Information Act to get precise data on the specific EVMs used. Television talk shows and long essays in newspapers have been keeping the debate alive: are EVMs truly safe and immune from high tech rigging?

The concerns expressed through these discussions and debates are varied. One line of argument compares Assembly polls that used EVMs with Panchayat elections conducted with paper ballots and concludes that better results were obtained by opposition parties where EVMs were not used. Another line of argument goes on to state that it is theoretically possible to reprogramme the resident software to favour one particular political party only. Then there are those who refer to introduction of upgraded versions of the machine to hint that the earlier designs were faulty. Yet another class of doubts stem from the controversies over voting technology in America's presidential election in 2000 and alludes to a wide variety of literature available in the public domain doubting the Diebold-based touch-screen voting machines used in the US. The last group bases its objections on the apparently very forceful logic that if indeed India’s EVMs are so good, why aren’t these adopted by the international community? Why do almost all advanced democratic countries still rely on the good old ballot papers?

A basic reason for a large part of the debate surrounding use of EVMs is that insides of an EVM cannot be seen like an empty ballot box. The transparency with EVMs is demonstrated by conducting a mock poll in each polling booth before the voting actually starts. In this mock poll phase, each polling agent of contesting candidates presses few buttons of the EVM after which the votes recorded are counted and tallied. This is repeated if necessary and after satisfying all agents, the contents of the machine are cleared a final time, the cover concealing the ‘clear’ and ‘results’ buttons is closed and the agents are invited to put their seals on the lock.

But isn’t it possible in theory at least, so the argument continues, to so programme the software of the chip inside an EVM that after the first 100 votes or so – to take account of the mock poll - all subsequent votes could be recorded in favour of only one political party? Even without looking at the software technicalities, it is easy to see why this would not be possible by noting that ballot papers are pasted on its balloting unit alongside the buttons while the sequence of names written on the ballot paper is arranged alphabetically in three groups of recognised political parties, registered parties and independents. This sequence is determined after the list of validly nominated candidates is finalized by the Returning Officer, usually only a few days ahead of the poll, and is thus distinctly different in different constituencies. In other words, since it is not possible that candidates of party X appear only at the, say, fifth button everywhere, even a hypothetically rigged set of EVMs cannot ensure party X’s victory everywhere. Besides, the commissioning of the EVMs is done in full view of the agents of the candidates and the whole operation is video-graphed. For added transparency, the EVMs are assigned randomly to different constituencies and the serial numbers of the machines are carefully recorded for reference. After the ‘candidate setting’, the EVMs are placed in a dedicated strong room with 24x7 armed guards under the watchful vigil of polling agents till their deployment in voting booths.

In a much publicized event soon after the General Assembly Elections in West Bengal in 2006, a public demonstration on how EVMs could be manipulated was organised at the massive Science City auditorium in Kolkata by a group of people who alleged that the just concluded elections used rigged EVMs. It turned out to be a damp squib where only a Power-Point presentation projected few dummy EVM models and some hypothetical scenarios of hi-tech rigging of EVMs. “We never said that EVMs had been actually tampered with” remarked one prominent organizer later. “We merely pointed out that there is a theoretical possibility that this could be done.” In a 2005 case in the High Court of Bombay, even a computer expert of the petitioner challenging the use of EVMs admitted to the Court that no one could hack into the EVM system because the hardware and software used in EVMs were not known.

Nobody has been able as yet to prove that EVMs can be doctored or tampered with.

The Election Commission of India goes to extraordinary lengths to ensure that the secrecy of the hardware and software in the EVM is maintained. The machines are manufactured only by two public sector enterprises, one of which is under the Ministry of Defence and the other with the Department of Electronics, Government of India. The micro-controller chips are non-erasable and are encoded in machine language that cannot be understood by human beings. The encoding and chip manufacturing is done not in India but in a foreign country and its details are a business secret. The memory is non-volatile and even if the battery is removed, no data is lost. If the connecting cable is damaged or cut, the letters “LE” (Link Error) are displayed. If the memory device is removed, the micro controller detects it and declares that the machine is in an error state. The encrypted code and data is unchangeable and indelible by anybody, even by the manufacturer. Little wonder that the Election Commission has been challenging anybody to show them how to rig these EVMs.

It is therefore no surprise that the Supreme Court, High Courts, Government Standing Committees and Expert Technical Groups have over the years subjected the EVMs to the strictest scrutiny and found them to be totally secure and safe. The Goswami Committee on Electoral Reforms that was set up in 1990 included a report of a technical sub-committee consisting of experts from Defence Research and Development Organisation, Indian Institute of Technology, Delhi and Electronics Research and Development Centre, Trivandrum in which the tamper-proof functioning of the EVMs was fully established. In a Supreme Court case of 2004, the Court examined a scientist who was a co-designer of the EVM as an expert court witness; in its judgment, the Court observed that the EVM was a great invention, a national pride. The same year saw the nationwide roll-out of EVMs during the General Parliamentary Elections 2004 involving 670 million electors.

To understand why the EVMs used in India are quite different from the touch screen voting machines used in other countries like the USA, it is necessary to look at the origins and raison-d’être of the Indian EVMs. The design of the EVM was mandated in India purely by its own needs: to capture and reflect accurately the voters’ intentions, to save money and to make it difficult for miscreants to stuff ballot boxes with bogus unused ballot papers in the dying hours of the poll day. In the paper ballot era, ambiguous or multiple markings on ballots resulted in invalid votes and it was not unusual to find that the number of invalid votes were greater than the margin of victory. Complaints of deliberate counting errors by allegedly biased counting officials were also common and led to frequent recounting pleas and court cases. Reduction in consumption of thousands of tonnes of paper on economical and environmental grounds was also a strong motivation. Because each vote occupied a prescribed predetermined time interval in the EVM, quick loading of false votes were not possible. Above all, the initial design was consciously made to be as close to the system of marking on ballot papers as had been evolved from the third General Elections in 1962 so that the average voter had no difficulty in exercising her franchise.

The Indian EVMs are thus strictly stand-alone devices that record people’s voting intentions accurately and store it in its non-volatile memory till they are counted. The EVMs still have to be physically carried under armed police escort to the counting centres for machine-by-machine result counting in presence of agents of candidates, much like the ballot paper days. In contrast, in all other countries where e-voting is attempted through voting machines, the emphasis is on networking, centralized data transfers through public communication channels and for enabling remote voting by people who are away from their polling areas on the date of poll. In such a scenario, in spite of various levels of security firewalls, there is a larger possibility of hacking from anywhere in the network pipeline. The concerns for using EVMs are thus overwhelmingly different in countries other than India where no networks is attempted, nor any non-resident voting permitted, except for a very few government employees who are entitled to vote by post. Little wonder, therefore, not many advanced countries are keen to adopt the India’s stand-alone EVMs. To extrapolate the controversies over EVMs in other countries to the Indian context would thus be rather fallacious.

Why then do controversies continue? A survey conducted by the Centre for Study of Developing Society in 1999 indicated that 90.6% of the public preferred the EVMs in the 1998 assembly polls spread over two states where EVMs were used. It can be safely presumed that this figure has gone up now. Even so, few skeptics possibly remain. It is necessary to also win them over by publishing white papers and launching a carefully crafted awareness programme. For the just-defeated crest-fallen political candidate, however, the EVM can be a scapegoat to nurse his wounded ego: we think of the legal fight instituted in the Kerala High Court by a contesting candidate challenging the first ever use of EVMs in 1982. He won the election and lost the case. His political rival lost the election and filed an appeal.

It has been said that compared to the shiny touch-screen Diebold-system based US voting machine, the Indian EVM is rather ‘simplistic’. Perhaps. But it is reliable and robust. Born in India, it understands its needs. It can be easily transported in its plain plastic carry-case to the remotest villages through the heat and dust of rural roads. It can be set up by a village school teacher acting as a polling officer on the election day, helping him by displaying messages whenever the cables are incorrectly connected or the batteries run low. It works without electrical power or complex networks of connectivity. It is friendly even to literacy-challenged voter. It can accurately record and count our votes. It can be trusted to keep our votes safe and secret. It may not be much to look at but it is difficult not to be a bit proud of this rather lovable little machine.