The e-voting challenge Published in The News International, April 13, 2018
Discussions on e-voting have re-emerged as the election draws closer and the National Database and Registration Authority (Nadra) gears up to present its plans about the three-tiered Internet Voting System for overseas Pakistanis, this week. Since I came across this subject in my capacity as a federal minister in the 2013 caretaker government, I am sharing a few insights for the planning currently underway.
E-voting for overseas Pakistanis was a subject of much discussion during the initial days of the 2013 caretaker government. The Supreme Court had issued a directive in response to a petition about the establishment of an e-voting system for overseas Pakistanis and had directed that this be expedited. Within this context, the Election Commission of Pakistan (ECP) convened a meeting of government functionaries for the process to be facilitated, which is how I and other ministers were invited to the table.
After the first meeting it became clear (at least to me) that it would be practically impossible to turn around the task in eight weeks due to time limitations. Although Nadra had developed a customised application and had demonstrated its use during presentations at the ECP, several steps were needed to be completed before the software could actually be deployed. Perhaps the most important was that it needed testing in real life situations overseas, along with quality and integrity audits, which are time consuming.
The procurement process of the hardware, on which the application was to run, was yet to be initialised. To ensure the system functions in Pakistan’s missions abroad, human resource was needed to be hired and trained, and their visas and travel arranged. The Ministry of Foreign Affairs had been pursuing permissions from the nine countries that the Supreme Court wanted this system to be functioning in. There were huge implications for Pakistan’s missions abroad in terms of physical requirements, awareness creation and compliance with local procedures. Eight weeks was nowhere near enough time to develop a foolproof system. A half-baked attempt could have been catastrophic. I volunteered to appear before the Supreme Court to explain the situation.
I narrate this story because as per the Supreme Court’s decision, e-voting for overseas Pakistanis is to be implemented this time round. I would like to flag two points in this regard. First, there are differences between the 2013 and 2018 Pakistani e-voting pilot systems. In 2013, Electronic Voting Machines (EVMs) were developed for deployment in Pakistan’s missions abroad but the project could not come to fruition. The EVMs were later used in the NA-4 by-elections in October 2017, and important lessons were learnt about their limitations. However, the system is not going to be used for the upcoming 2018 elections.
In contrast, the 2018 envisaged e-voting system allows voting from remote locations using internet, rather than at polling stations, and is being developed for overseas Pakistanis. Although each option has its own inherent challenges, one key lesson is that every new system needs adequate time and planning for conceptualisation, deployment, integrity checks, piloting and ploughing back lessons to refine the model. Other supporting factors – procedural, administrative and legislative – need to be taken into account when planning, as technology is only one cog in the process chain.
When I stepped down from my role as federal minister in 2013, I had reiterated in my Handover Papers the need for the next government to take this matter up in a timely manner and examine implications carefully. I am not privy to details as to how the matter was dealt with over the last five years, but if a demonstration is being given now, the process must be in its pre-pilot stages.
Second, there are important lessons to be learnt from the e-voting experience of Estonia. Estonia became the first country in the world to offer internet voting nationally – in local elections in 2005 and later for parliamentary elections in 2007, which is when 30 percent Estonians voted through the internet. Lessons from the Estonian example are critically important since Pakistan’s infrastructure of digital identities is somewhat similar to the Estonian model.
In 2016, an independent assessment of the procedural components of the Estonian Internet Voting System was conducted by the Cyber Studies programme of the University of Oxford. The evaluation highlighted the importance of the Estonian experience in conducting electronic elections for the last eleven years. However, while praising the system, the evaluation highlighted the unique circumstances under which Estonia has been successful and questioned the ability of the systems’ procedural controls against sophisticated cyber-attacks. The threat of cyber security is so real and pressing that in Brazil, the Brazilian Supreme Electoral Court has gone so far as organising ‘hacking competitions’ to create additional confidence in the technology. Recognising security challenges, several European countries including the UK, Germany, Netherlands and Norway cancelled e-voting systems or have decided against its large-scale use.
On balance, internet voting is feasible, which if effectively deployed, could improve accessibility for voters, especially those that are overseas and others for whom access is an issue, such as the disabled. It could also potentially save costs and time. However, in practice, the most important thing is that there needs to be a high-level of confidence in the system. Success depends on the expertise and experience of the team. There must be adequate time for the system – and not just the technology – to develop and undergo rigorous integrity checks.
Designing effective operational and procedural controls and safeguards against the risk of tampering is fundamental for the success of the programme. Estonia is grappling with this issue after two decades of having commenced on this journey. Pakistan’s institutions mandated with e-voting should ascertain where they stand in this process, and the implications a fallout would have on the credibility of the election. More broadly, they should assess how this relates to the provisions of ‘secrecy of ballot’ and ‘anonymity of voters’, which are enshrined in our constitution. As democracies around the world struggle to deal with hackers, fake news and twitter bots, would now be an optimal time to use limited resources on rolling out a lightly tested internet voting system?