Pakistan's Internet Voting Experiment
PPakistan’s Internet Voting Experiment
Hina Binte Haq , Ronan McDermott , and Syed Taha Ali School of Electrical Engineering and Computer Sciences (SEECS), NationalUniversity of Sciences and Technology (NUST), Islamabad, Pakistan. { hhaq.dphd18seecs, taha.ali } @seecs.edu.pk MCDIS [email protected]
Abstract.
Pakistan recently conducted small-scale trials of a remoteInternet voting system for overseas citizens. In this contribution, we re-port on the experience: we document the unique combination of socio-political, legal, and institutional factors motivating this exercise. We de-scribe the system and it’s reported vulnerabilities, and we also highlightnew issues pertaining to materiality. If this system is deployed in thenext general elections —as seems likely —this development would con-stitute the largest enfranchised diaspora in the world. Our goal in thispaper, therefore, is to provide comprehensive insight into Pakistan’s ex-periment with Internet voting, emphasize outstanding challenges, andidentify directions for future research.
Keywords:
Internet Voting · Overseas Voters · Pakistan.
This paper has been accepted for publication in TUT Press proceedings ofE-Vote-ID 2019. E-Voting.CC GmbH is the owner of the copyright.
Pakistan recently piloted a remote Internet voting system for overseas citizens.This system, i-Voting , was indigenously developed and originally scheduled forfull-scale deployment in the General Elections of July, 2018. However, theseplans were deferred after a third-party technical audit of the system uncoverednumerous vulnerabilities and security concerns. i-Voting was deployed shortlyafter on a trial basis: in bye-elections, first in October, 2018, spanning 35 con-stituencies, and next in December, 2018 in 1 constituency. Some 7,538 votes werecast (7,461 in October and 77 in December) using this system and these weredeclared binding and incorporated into the final results.It is widely expected that this pilot is a prelude to full-scale deploymentin the General Elections of 2023. Since Pakistan currently has over 8 millionoverseas citizens [1], this may well be the largest deployment of Internet votingin the world. It is therefore essential to document and study this experiment.In this paper, we make the following contributions: also referred to as iVoting, iVOTE, IVoting Currently, the law restricts use of such systems to bye-elections. a r X i v : . [ c s . C Y ] J u l HB Haq, R McDermott et al.
1. We report on the deployment: we document the public debate on Internetvoting and the legislative and political process to facilitate it. We describethe i-Voting system and we report on the pilot exercise.2. We describe the various risks involved in this modality of voting and sum-marize key findings of the technical audit. We examine the materiality, andtherefore the potential political significance of overseas voting.3. We highlight the unorthodox combination of unique political and social fac-tors in Pakistan that have resulted in this exercise and we discuss variousparticular challenges that may arise as a result.This paper holds relevance for election stakeholders including governments,political parties, election administrators, political scientists, researchers, andtechnologists. Pakistan’s Internet voting experience may also prove instructivefor other countries, particularly in the developing world, where governments areseverely limited in terms of financial resources, technical expertise, and infras-tructure to undertake such critical large-scale projects.
Organization of Government
Pakistan has a parliamentary form of govern-ment with bicameral legislature, comprising a Senate (upper house) with 104members and a National Assembly (lower house) with 342 members. Each ofthe four large provinces have a unicameral legislature, consisting of a ProvincialAssembly. The electoral system is the first-past-the-post system under universaladult suffrage. Members of the National Assembly and Provincial Assembly areelected by representation in electoral districts (referred to as seats or constituen-cies ). The number of seats in each administrative division is listed in Table 2.General elections are conducted every five years and are overseen by the ElectionCommission of Pakistan (ECP), which is an independent and autonomous bodyas defined in the Constitution of Pakistan.
Body Total Federal Baluch- Federally Khyber Punjab SindhSeats Capital -istan Administered Pakhtun-Islamabad Tribal Areas -khwaNational
272 3 16 12 39 141 61
AssemblyProvincial
577 - 51 - 99 297 130
Assembly
Table 1.
National and Provincial Assembly Seats [2]
Overseas Pakistanis and the Right to Vote
Pakistan, has over 8 millionoverseas citizens [1] which comprises the sixth largest diaspora in the world[3]. Overseas citizens are actively engaged in the socioeconomic well-being ofthe country and every year send home remittances worth approximately US$19billion, which accounts for around 5% of Pakistan’s GDP [4]. Federally Administered Tribal Areas (FATA) and Federal Capital Islamabad areadministrative divisions in addition to the four provinces, included in the contestedelections and comprise National Assembly seats only.akistan’s Internet Voting Experiment 3
Article 17 of the Constitution of Pakistan grants all adult citizens the fun-damental right to vote [5]. This article has generally been interpreted to ac-knowledge that this right extends to all Pakistani citizens, irrespective of placeof residence. Overseas Pakistanis have raised calls for enfranchisement and facil-itation of their voting rights since the first general elections of 1970 [6].The earliest constitutional petition filed to facilitate overseas voters was in1993 by a British-Pakistani student and the Supreme Court of Pakistan referredit to the government and the ECP for consideration [7]. After a hiatus of almosttwo decades, more petitions followed in quick succession: in 2011, Dr. Arif Alvi,Secretary General of Pakistan Tehreek-e-Insaaf (PTI), a popular political partypetitioned the Supreme court in this regard; in 2014, the Islamabad High Courtwas likewise petitioned by a concerned overseas citizen, and in June 2015, bythe Chairman of PTI, Mr. Imran Khan.The judiciary, while addressing this grievance, has upheld this fundamentalright of overseas citizens on multiple occasions [8] [9] [10], ruling that this rightcannot be denied on technical grounds, and it has repeatedly directed the ECPto make the necessary logistical arrangements. We discuss these attempts next.
Efforts by the Election Commission and Parliament
In 2012, in one ofthe earliest statements on the subject, the ECP dismissed the possibility of over-seas citizens’ participation in the General Elections of 2013, citing logistics andbudgeting issues [11]. However, by 2015, the ECP had established a Directoratefor Overseas Voting in its Secretariat, which conducted mock overseas votingexercises using postal ballots and voting via telephone [12].These trials were unsuccessful. The reasons were clarified in a study com-missioned by the ECP: “We find that any remote voting solution using cur-rently available technology whether postal, internet, telephone, or proxy willlack the necessary electoral integrity checks to preserve the credibility of an elec-tion result.”
Commenting on the feasibility of other modalities, the report stated: “...given the size and dispersal of the Pakistani diaspora, coupled with the lim-ited official resources available in-country and abroad, any significant in-personvoting operation would be expensive and logistically challenging” [13].In July of 2014, the Parliamentary Committee for Electoral Reforms wasconstituted with Finance Minister Ishaq Dar as chair. A sub-committee wasformed in January, 2016 by MP Dr. Arif Alvi (one of the petitioners for overseasvoting mentioned earlier and Secretary General of PTI) to devise a mechanismfor overseas voting [14] and in March, 2017, it proposed remote Internet voting asa potential solution. Consequently, the committee authored the Elections Act of2017, which authorized the ECP “to conduct pilot projects for voting by OverseasPakistanis in bye-elections” [7].The ECP subsequently requested the National Database Registration Au-thority (NADRA) to build a system. NADRA is an independent and autonomousagency working under the Ministry of Interior and tasked with managing gov-ernment databases and issuing national identity cards to citizens. However thesystem did not materialize: in June 2017, ECP again contacted NADRA, but
HB Haq, R McDermott et al.
NADRA expressed its regrets at not having a solution available [12].
The Supreme Court Intervenes
Interest in overseas voting peaked again inthe six months leading up to the General Elections of July, 2018. The SupremeCourt of Pakistan consolidated 16 similar constitutional petitions and resumedhearings on the issue [15]. It sought reports from the ECP and NADRA overnon-compliance of Section 94 of the Elections Act (regarding pilot projects foroverseas voters) [16]. In an attempt to break the deadlock, the Supreme courtdirected NADRA to develop an Internet voting system. NADRA informed thecourt that it would require 4 months to build a system and would cost Rs.150 mil-lion (approximately US$ 1.36 million) [17]. However, the Court ordered NADRAto present it in 10 weeks [18].The new system, i-Voting, was unveiled on April 12, 2018 at a public sessionconvened by the Supreme Court [19]. The audience included members of variouspolitical parties, IT experts from Pakistani universities, concerned citizens, andmembers of the media. It was here that IT experts aired serious security concernsregarding this system and pointed out that similar systems had been demonstra-bly attacked and were being phased out in developed countries. The SupremeCourt concluded the session by forming an Internet Voting Task Force (IVTF)to audit the system and assess its suitability for deployment in the forthcominggeneral elections of July, 2018.
Here we describe NADRA’s i-Voting system and summarize the IVTF findings.
System Architecture
The i-Voting system conforms to the traditional designof Internet voting solutions, where a centralized database is used to store andtabulate votes, which voters access using a Web portal.More specifically, as depicted in Fig. 1, a central datacenter hosts the overseasvotes database and an application and email Server. These servers interface witha webserver hosting the i-Voting Web portal (Fig. 2) and NADRA databases (forverification of voter information). The ECP can monitor the system using anadministrative portal. Load balancing and backup arrangements are deployedas well as standard security solutions including firewalls, intrusion detectionmechanisms, and mitigation of distributed denial of service (DDoS) attacks.
Voter Registration
The registration process is depicted in Fig. 3 [20]. To enrol,a user must possess his/her passport, a National Identity Card for OverseasPakistanis (NICOP), and a valid e-mail address. The ECP announces a publicRegistration Phase during which prospective voters enter their basic details intothe system. A confirmation email with a PIN is then sent to the voter. To confirmthe account, the user enters the PIN and solves a CAPTCHA.Now the user logs in to the system and provides further details of his NICOPand passport. He also answers two randomly chosen questions pertaining to his akistan’s Internet Voting Experiment 5 identity after which he is successfully registered. The system allows a maximumof 3 answer attempts, failing which the NICOP number is restricted.
Vote Casting and Preparation of Results
Prior to polling day, each regis-tered voter is emailed containing a unique passcode (which acts as a one-timepassword), enabling him to log on to his i-Voting account and cast a vote for hisrespective National Assembly and/or Provincial Assembly seat (Fig. 3).When polling concludes, the ECP tabulates the votes via the Reporting Por-tal and dispatches the tally to concerned officials for consolidation of results.
Internet Voting Task Force
The Internet Voting Task Force (IVTF) wasgiven a time window of 3 weeks to assess the security of the i-Voting system.The team comprised of IT and security specialists and academic researchers [21].They conducted a high level security analysis of the system, examined the code,and mounted some typical attacks. The results were written up in a report andsubmitted to the Supreme Court. Their key findings are as follows:1. i-Voting does not provide ballot secrecy, a fundamental right defined in theConstitution of Pakistan. This further opens up the possibility of vote buyingand coercion of overseas voters.2. A key security vulnerability allowed overseas voters to bypass their nativeconstituencies and cast votes for any two seats of their choice in the country.3. The IVTF successfully launched impersonation attacks, enabling them tosend fake emails purportedly from the ECP to direct voters to fake websites.4. i-Voting avails the services of a leading DDoS mitigation solution, a measurewhich researchers have recently demonstrated can potentially compromiseballot secrecy and election integrity [22].5. The system employs certain third-party security components (such as text-based CAPTCHAs) which are obsolete and demonstrably insecure.The IVTF also raised other critical non-security concerns: i-Voting lackedverifiability, fail-safe, or redundancy mechanisms. There were no security poli-cies or procedural controls defined to protect critical security processes frominsider attacks. No usability studies or trials had been conducted for the system.Futhermore, the system was built in an ad-hoc manner with key documentationmissing. For instance, there was no documented Solution Requirements Speci-fication (SRS) or documentation pertaining to key operational processes (suchas administration, hosting, responsibility of critical components), which limitedassessment for certain important security attacks.
Fig. 1. i-Voting: System Architecture HB Haq, R McDermott et al.
Fig. 2. iVoting: Interface
The IVTF therefore strongly argued against the deployment of i-Voting inthe upcoming General Elections of 2018. Their report stated that this would be “a hasty step with grave consequences”
The report also emphasized that “manyof these security vulnerabilities are not specific to iVOTE [sic] but are inherentto this particular model of Internet voting systems” [23].The report also made various recommendations to facilitate overseas voters.We discuss these in Sec. 6 and 7.
The Supreme Court of Pakistan revisited the matter of Internet voting after thegeneral elections in August, 2018, and ruled: “Based on these representationswe prima facie find the mechanism of I-Voting (sic) to be safe, reliable andeffective for being utilized in a pilot project. We are sanguine that the aforesaidproposed rules shall be incorporated in the Election Rules, 2017 to enable overseasPakistanis to exercise their right of vote in the forthcoming bye-elections.”
Thecourt further stipulated that votes cast using i-Voting not be added to the finaltally until the ECP is satisfied with regards to their “technical efficacy, secrecy,and security” . In case of any dispute the ECP was authorized to exclude theseoverseas votes from the official results [10].
Fig. 3.
Voter Registrationakistan’s Internet Voting Experiment 7
Fig. 4.
Vote Casting
The ECP consequently amended the Election Rules to accommodate therequirements of Internet voting. NADRA implemented certain technical recom-mendations of the IVTF and trained ECP officials to administer the system.The ECP launched a media campaign for voter awareness and published detailedguides and video tutorials for the i-Voting system. A dedicated support centerwas also set up to provide telephone and email assistance [12]. First Pilot (Bye-Elections - 14 October, 2018)
Bye-elections were heldfor 35 constituencies (11 National Assembly and 24 Provincial Assembly seats).The total overseas Pakistanis eligible to participate in these polls numbered asignificant 631,909. However, out of these only 7,419 citizens (1.17%) actuallyregistered to vote using the new system. On the day of the elections, a total of6,146 voters of these citizens cast their votes [12].ECP later reported that on the day of the polls the system successfullywithstood 7,476 DDoS attempts. The top 5 countries by voter count were theUnited Arab Emirates (1,654), Saudi Arabia (1,451), the United Kingdom (752),Canada (328), and the United States (298). The pilot project cost approximatelyRs. 95 million (0.67 million USD approximately) [12].The trial was smooth and uneventful. In its own report, the ECP attributedthe low turnout to the short time frame within which the system was deployedand advertised. The ECP also cited key issues which echoed the concerns of theInternet Voting Task Force (discussed in Sec. 3), in that the system violatesballot secrecy, enables voter coercion, lacks auditability, and may be vulnerableto state-level cyberattacks.
Second Pilot (Bye-Elections - 13 December, 2018)
As many as 4,667overseas Pakistanis from more than ten countries were eligible to vote for oneProvincial Assembly seat [24]. However, only 77 overseas Pakistanis registeredto vote. The ECP has not released any further details about this trial [25].
In this section we undertake a basic post-election analysis to examine the po-tential impact of overseas votes on final results. No details have been published on what specific changes were made. No details of these attacks have been released to the public. HB Haq, R McDermott et al.
The leading Pakistani citizen observation group, Free and Fair Election Net-work (FAFEN) has conducted an analysis of 2018 General Election results [26]and has determined that, in a significant number of constituencies, the Marginof Victory (MoV) is less than the number of invalid votes. A similar analysis, butcomparing Margin of Victory with number of Eligible Overseas Voters would beuseful to highlight the materiality . There is no publicly available voter regis-tration or population data which shows how many registered voters are actuallyOverseas Pakistanis for all constituencies. The only exception to this is for theconstituencies where by elections were conducted using i-Voting in October andDecember 2018. The average percentage of eligible overseas voters, as a per-centage of total voters in the October 2018 By-elections is 6.88% ( Table 2).With this assumption, we estimate the number for eligible overseas voters forindividual constituencies under scrutiny.We then calculate an estimated Overseas Pakistani Voters value for eachcontested constituency in the October 2018 By elections that was also contestedin the July 2018 General Elections. We now compare these Estimated OverseasVoters (EOV) value with the margin of victory and flag where the MoV is less.We do this both for the October 2018 By-elections and the July 2018 GeneralElections. We may describe the number of Overseas Pakistani Voters in thesecases as material to the outcome of the election .As Table 3 shows, in ten of twenty-seven by-election races, Overseas PakistaniVoters had the potential to be material. This grows to thirteen of twenty-sevenfor General Election races. In five races, both Bye-Election and General Electionsaw Margin of Victory less than estimated Overseas Pakistani Voters. It is, webelieve, reasonable to assume that, in competitive future General Elections atleast one in five races may be decided by votes cast by overseas Pakistani voters.This places the integrity of and trust in any internet voting solutions deployedby the Elections Commission of Pakistan into very sharp focus. This is positivein the sense that overseas Pakistanis can feel their votes count. At the same timeit necessitates election integrity checks so that this right is not misused. In this section we further examine outstanding issues arising from this experi-ment and we attempt to contextualize these by examining the unique politicaland institutional factors that motivated these pilot projects. Materiality in this context refers to the theoretical scenario where all possible over-seas votes are cast, and all are cast for the second place or losing candidate, theoutcome of the election might have been different/might be different. The public domain sources for Table 2 and 3 are no longer available on ECP Website.The documents will be made available at a URL which will be cited later (to retainanonymity).akistan’s Internet Voting Experiment 9
Description Number
Total Registered Voters in bye-election 9,185,705 []Total Eligible NICOP 734,777 []Non-Machine Readable Passports ( 14%) 102,868 []Estimated Total Overseas Pakistani Voters 631,909 []Eligible Overseas Pakistani Voters as % of Voters Registered 6.88% []
Table 2.
Calculating Estimated Percentage of Overseas Pakistani Voters
Fig. 5.
Twitter users Posting Screenshot of Vote
In delivering one right (the right of overseas Pakistanis to vote), the solution risksundermining another (the right to secrecy). The i-Voting system does not complywith Article 226 of the Constitution of Pakistan [27] and the Elections Act 2017,Section 81 [28], that impose ballot secrecy. Being a remote voting modality,there is no mechanism to prevent an individual from revealing their vote toothers. Similarly, certain event logging software (specially on a shared/publicdevice) can secretly capture the choice of a voter.Electoral offences were committed by voters unintentionally, by posting screenshots on social media, as shown in Fig 5. The tweeters seem to be unaware theiractions are electoral offences, and being outside the jurisdiction of Pakistan it isunclear, how such offenders can be brought to justice .In addition, to the lack of secrecy for the voter at the client end, the low levelsof participation in pilots also mean that, in some cases, typically PA seats (PB-35, PP-165, PP-292 from October 2018 bye-elections [12]), the voter’s choice isrevealed. The usual solution to this problem (mixing votes from multiple ballot The exceptions in Section 81 are not to the secrecy requirement, rather to require-ment of casting a vote by inserting paper ballot into a ballot box. Section 178 of the Elections Act 2017 elaborates the offences relating to ballot se-crecy.0 HB Haq, R McDermott et al.
Constit Total Estimated General Elections Bye Elections Bothuency Registered Overseas July,2018 October,2018 ElectionsVoters Voters (EOV) MoV MoV < EOV MoV MoV < EOV Mov < EOV
PB-40 76,173 5,240 13,345 - 9,141 - -NA-53 313,141 21,541 48,763 - 18,630 - -NA-35 582,785 40,091 7,001 Yes 23,455 - -PK-3 146,180 10,056 5,550 Yes 1,163 - -PK-7 155,719 10,712 5,825 Yes 334 - -PK-44 202,601 13,937 10,857 Yes 1,630 - -PK-53 153,352 10,549 6,729 Yes 61 - -PK-61 139,517 9,597 4,593 Yes 5,247 - -PK-64 160,728 11,056 18,579 - 13,215 - -PK-97 155,032 10,665 16,461 - 10,172 - -NA-56 640,133 44,036 64,490 - 41,593 Yes -NA-63 371,713 25,571 35,979 - 26,292 - -NA-65 553,289 38,062 51,963 - 68,591 - -NA-69 469,177 32,275 73,172 - 50,803 - -NA-124 535,172 36,815 65,287 - 47,533 - -NA-131 365,677 25,155 756 Yes 10,031 Yes YesPP-3 224,755 15,461 37,008 - 227 Yes -PP-27 312,370 21,488 1,766 Yes 656 Yes YesPP-118 222,190 15,285 548 Yes 5,189 Yes YesPP-164 137,906 9,486 20,870 - 7,561 Yes -PP-165 132,077 9,085 20,372 - 5,742 Yes -PP-201 232,120 15,968 17,297 - 7,024 Yes -PP-222 196,858 13,542 11,446 Yes 6,083 Yes YesPP-261 187,510 12,899 9,371 Yes 14,261 - -PP-272 167,467 11,520 5,390 Yes 8,899 Yes YesPP-292 141,297 9,720 253 Yes 10,692 - -NA-243 402,731 27,704 67,291 - 21,601 - -
Table 3.
Materiality of Overseas Voters in Selected Constituencies, General and Bye-Elections 2018 boxes or polling stations) is not available in the i-Voting context or could onlybe implemented at the cost of further erosion of already minimal transparency.As the IVTF report points out, some jurisdictions [29] allow a voter to waivetheir right to secrecy. This is not a solution to the problem, as any voter (orparty or candidate) to assert their constitutional and legal rights to secrecy forthe system to be challenged.Almost half of the diaspora, over 4 million Pakistanis reside in the MiddleEast, and about a quarter (over 1.5 million) reside in Europe [30]. A bulk of thediaspora specifically in the Middle East are labourers. The ECP itself recognizesthe risk of vote buying and coercion when it speaks of the ”kafeel” abusingcustody of passports [12]. The ILO describes this system as placing migrantworkers in ”a position of vulnerability and have very little leverage to negotiatewith employers, given the significant power imbalance embedded within the em- Sponsor for a migrant worker International Labour Organizationakistan’s Internet Voting Experiment 11 ployment relationship. Common grievances expressed by migrant workers includerestrictions on free movement, confiscation of passports, delayed or non-paymentof salaries, long working hours, untreated medical needs, and violence all condi-tions that can give rise to situations of forced labour and human trafficking” [31].It is reasonable to assume that anyone who will treat migrant workers in thismanner will not hesitate to exploit their votes for political or financial benefit.Migrant workers are bound to face difficulty to independently use the i-VotingSystem. This could pave the way for coercion, vote buying and compromisesecrecy if vote casting is aided by a computer literate party. Thus, usability testsneed to be conducted to receive direct input from real users. It might be arguedthat low usability, was a primary reason of the low registration turnout, whereonly 1.17 % [12] of the total eligible overseas voters successfully registered withthe i-Voting system. Further, there seem to be no special accessibility featuresincorporated to address the needs of voters with disabilities.
The process of registration on the i-Voting platform [32] is entirely out of ECP’scontrol, relying as it does on a verification method conducted and adjudicatedby a computer programme. Potential overseas voters are quizzed with questions,whose answers are considered ”secret” . Common sense dictates that, despite thefamilial/personal nature of these questions, the answers will not be known ex-clusively to the voter. As a consequence, ECP cannot guarantee that the voterregistered via the online platform [32] is indeed the eligible citizen, or an impos-tor. Other key mechanisms to protect the integrity of the electoral rolls - thepublic display of and claims/objections on the draft electoral rolls - are omittedfrom the online process. Political parties, observers, and voters themselves, arenot given the access to these electoral rolls to allow for the scrutiny that wouldcontribute to stakeholder confidence in the electoral rolls.Furthermore, the mechanism within i-Voting to “lock” an identity followingrepeated incorrect answers or CAPTCHA verification may be used for votersuppression - a sort of denial-of-service attack, albeit on a vote-by-vote basis.Voters may not know the answers to all the questions they might be asked(where, for example, an 18 year old was registered and a parent provided theinformation). Corrupt or partisan Presiding Officers could merely strike out thenames of legitimate voters saying that they had registered online for i-Voting.
On election day in polling stations across Pakistan, a long list of integrity mech-anisms are in place, arising from the Constitution, the Elections Act 2017 andthe Election Rules 2017, as amended . In the i-Voting system these fourteen The election is conducted in full view of polling staff, party/candidate agents andobservers, who first-hand witness the integrity checks in place: verify ballot boxesare initially empty, identify voters on arrival, ink their fingers (to prevent multiple2 HB Haq, R McDermott et al. separate mechanisms are missing with consequences for electoral integrity. Theexclusion of party/candidate agents and citizen observers from the i-Voting pro-cess is compounded by the inherent absence of any verifiability mechanism orpossibility to audit the i-Voting system - by design - ”In order to ensure thatVoting is kept secret, all data was encrypted and no audit trail of voting waskept by the system” [12].ECP may exclude based on its ”opinion” as to whether the ”technical efficacy,secrecy and security of the voting has not been maintained” [12]. It is not knownhow ECP informs that opinion, or whether it has the required access to the i-Voting system. Given the 2018 recourse to establishing the IVTF, it seems likelythat ECP lacks the technical capacity to properly arrive at an informed opinion.Given the likely materiality of votes cast by overseas Pakistanis in a significantproportion of contests, we may expect many electoral disputes to centre aroundthe integrity of i-Voting system.Specifically, in a developing country like Pakistan, where the democratic pro-cess is at an inflection point, and the mechanisms to investigate and resolve elec-toral disputes, are still very fragile, electoral improprieties or even the impressionof such can potentially lead to political deadlock and turmoil. An indication ofthis is the PILDAT (Pakistan Institute of Legislative Development and Trans-parency) report on the perception of pre-poll fairness which notes that ”theinternet-based OP voting may also be a major instrument of rigging in 2018General Election” [33]. Diverse stakeholders including the ECP itself [12], andprominent mainstream political parties expressed similar reservations [34] [35].Lack of auditability features means there is no evidence if results are challengedthrough an election petition. These, coupled with questions over the capacityand willingness of the judiciary, raises concerns about the resolution of electoraldisputes around internet voting.
Concerns have also been raised regarding the threat model on which i-Voting isbased. The recent controversy of foreign interference in US elections hints that adeveloping country like Pakistan may also be at risk. ECP’s report on the pilotdeployment highlights this concern: “[adversaries] did not materially interferemerely to put us off track. When the system is finalized and put into practice inthe next elections, shall we be able to counter/control cyber attacks[?]” [12].Furthermore, whereas the high number of DDoS attacks (around 7476 on bye-elections day), posed an outage threat, the use of a DDoS mitigation service, asdemonstrated recently by Culnane et al. [22], introduces a new attack vector.The mitigation service is in a position to decrypt incoming traffic, thereby ableto compromise ballot secrecy and potentially even alter the content. The IVTF voting), vote casting in secrecy behind a screen, placing the ballot paper into thetransparent ballot box, the Presiding officer conducting the count and disseminatingthe results form to all stakeholders, and packaging all ballot papers (valid, invalid,challenged, spoiled) separately in tamper-evident envelopes.akistan’s Internet Voting Experiment 13 audit highlighted this concern in their report and pointed out that the serversemployed by the DDoS mitigation service were all based overseas and beyondcontrol of Pakistani authorities.
We see the Supreme Court at the forefront, driving the institutions to deliver avoting solution to overseas Pakistanis. Here we try to make sense of the uniquepredicament and examine the various factors that led to this situation. Thejudiciary has time and again reiterated the ECP to roll out a voting mechanismfor overseas Pakistanis, but to no avail. A concrete step in this direction waslong overdue and it had to take the Supreme Court to push it through, giventhe institutional inertia within the government.The Supreme Court of Pakistan frequently takes Government ministries andother public bodies to task for not fulfilling their obligations [36]. Whetherthrough judicial activism (using suo moto powers) or responding to petitionsfrom interested parties, it often gets involved in the technical specifics of cases.Its jurisdiction “is not limited to mere procedural technicalities as it enjoys cer-tain inherent powers to do complete justice in any case” [37]. A vivid example ofthe Supreme Court’s ambition beyond procedural technicalities is the fund es-tablished in July 2018 to raise money to build dams. This fund currently exceedsten billion Pakistani rupees (approximately 71 million US $) [38].Cognizant of this deviation, the Honourable Chief Justice of Pakistan, in-quired whether it was the job of the Supreme Court to give the right to voteto overseas Pakistanis? [16] In the matter of how best to enfranchise overseasPakistanis the Supreme Court initially directed the elections management bodyto develop an internet voting system and then later mandated the use of thissystem in binding political bye-elections. In doing so, the Supreme Court dis-missed unambiguous and dire warnings from the IVTF about the hazards ofthe proposed system as mere ”technical and security apprehensions” , and thatthe report was ”generally positive and encouraging” [10]. While the IVTF re-port clearly says ”Hopefully, this discussion thus far demonstrates to the readerwhy internet voting is recognized by security experts to be a controversial andrisky undertaking” , and it concludes by asserting ”We would, therefore, urge allstakeholders to exercise extreme caution in approaching the question of internetvoting” [23]. This disconnect has received media recognition [39].The ECP itself was very reluctant to adopt this modality of overseas voting.Recently, when the matter was taken up in the Senate, in May 2019, SenatorJaved Abbasi recollected that ”the ECP had convinced political parties that thesystem should not be introduced in Pakistan, but could not convince the SupremeCourt” , at which an ECP representative expressed his dismay that while theECP tried to dissuade the Supreme Court, no political party supported theECP in Supreme Court [40]. The absence of a broad political consensus on theuse of i-Voting to enfranchise the diaspora does not bode well for the future.Neither the Elections Commission of Pakistan, nor the developers of the i-Voting system challenged the Supreme Court’s interpretation of the IVTF find- ings or recommendations. Since there is no higher court than the Supreme Court,no appeal is possible. If a future election is decided on votes cast by overseasPakistanis, via the i-Voting system, and the result is challenged —which seemshighly likely, given both the deficiencies and materiality described earlier in thispaper —it will be interesting to see if the electoral dispute resolution processends up in the same court.
Pakistan’s experiment for the October 2018 by elections was the largest de-ployment of Internet voting in a binding election, anywhere in the world. Therecommendations of both the IVTF report, and ECP’s own report on the Octo-ber 2018 pilot exercise are comprehensive and we endorse these. Going beyondthese, and comparing the Pakistani experience with other countries who are fur-ther along the internet voting pathway, we would highlight two vital priorities.First, transparency: ECP and NADRA succeeded in delivering a working pro-totype system in the short time available, but the details of the process were,and remain, opaque. Stakeholder acceptance cannot be assured in future with-out meaningful transparency and greater consultation, having regard to votersecrecy. Second, capacity building across all stakeholders, starting with, and ledby ECP (such as establishing a dedicated R&D cell within the ECP), to delivercompetent national ownership and informed policymaking. It seems likely thatescalation from pilots in bye-elections to full-scale use of internet voting for theenormous Pakistani diaspora will happen in 2023. The issues highlighted in thispaper should receive urgent attention by all Pakistani stakeholders.
References
1. PRIO. Pakistan as a return migrant destination. https://opf.org.pk/media/1410/pakistan-as-a-return-migration-destination.pdf , 2015.2. ECP. Final List of Constituencies (Final Delimitation 2018). .3. Riaz Haq. India tops labour export, Pakistan ranks6th. https://en.dailypakistan.com.pk/opinion/blog/international-migrants-day-india-tops-labor-export-pakistan-ranks-6th/ ,December 2016.4. Muzaffar Rizvi. Pakistan remittances may hit $22 billion in 2018-19. , October 2018.5. Article: 17 Freedom of association — The Constitution of Pakistan, 1973. https://pakistanconstitutionlaw.com/article-17-freedom-of-association/ , 1973.6. Ahmed Bilal Mehboob. Voting from abroad. , May 2017.7. Ahmed Bilal Mehboob. Arranging Voting byOverseas Pakistanis. https://pildat.org/blog/arranging-voting-by-overseas-pakistanis-when-and-whose-job-is-it-2 ,April 2018.akistan’s Internet Voting Experiment 158. Const.p.39and90of2011-dt-3-5-2013. .9. Shahid Rao. Ihc refers matter to ECP. , December 2014.10. Const.p. 74 2015. .11. Qaiser Butt. Election preparations: Overseas Pakistanis un-likely to vote in polls. https://tribune.com.pk/story/435109/election-preparations-overseas-pakistanis-unlikely-to-vote-in-polls/ ,Sept 2012.12. Report on I/-Voting Pilot Test. https://ecp.gov.pk/documents/ivotingreport.pdf , October 2018.13. Ahmed Bilal Mehboob. Voting from abroad. , May 2017.14. Furkan Ali. Apathy of electoral reform. ,May 2016.15. Haseeb Bhatti. SC accepts 16 petitions regarding overseas Pakistanis. , January 2018.16. Correspondent. CJP asks whether parliament will give votingrights to overseas Pakistanis. http://dunyanews.tv/en/Pakistan/423769-CJP-Justice-Saqib-Nisar-parliament-voting-rights-overseas-Pakistanis ,January 2018.17. Monitoring Report. NADRA to develop Internet voting sys-tem for expats. , January2018.18. Hasnaat Malik. Overseas Pakistanis suffrage: SC orders NADRA to de-velop voting software in 10-weeks. https://tribune.com.pk/story/1621120/1-software-overseas-pakistanis-vote-elections-will-ready-april-nadra-tells-cjp/ ,January 2018.19. Supreme Court Press Release. SC hears constitution petitions regarding right ofvote to overseas Pakistanis in general elections. .20. iVoting- Registration Closed. .21. TOR and ECP order. https://ecp.gov.pk/TOR\%20and\%20ECP\%20order.pdf ,April 2018.22. Chris Culnane, Mark Eldridge, Aleksander Essex, and Vanessa Teague. Trust im-plications of DDoS protection in online elections. In
International Joint Conferenceon Electronic Voting , pages 127–145. Springer, 2017.23. Ivtf Report Executive Version. , May 2018.24. FAFEN. By-Election Report - PP-168 Lahore. http://fafen.org/wp-content/uploads/2018/12/By-Election-Report-PP-168-Lahore-XXV.pdf , Dec 2018.25. Rana Bilal. PTI defeats PML-N by close margin in Lahore’s PP-168 by-poll. , December 2018.26. General Election Observation 2018: Key Findings and Analysis. http://fafen.org/fafen-general-election-observation-2018-result-assessment-and-analysis/ .27. Constitution of Pakistan, part VIII: Chapter 2: Electoral Laws and Conductof Elections. .6 HB Haq, R McDermott et al.28. Elections act, 2017. .29. Mike Orcutt. Internet Voting leaves out a Cornerstone of Democ-racy: The Secret Ballot. ,August 2016.30. Qadeer Tanoli. 2.43 million Pakistanis workingin Europe. https://tribune.com.pk/story/1391730/overseas-workforce-2-43-million-pakistanis-working-europe/ , April2017.31. ILO. Employer-Migrant Worker Relationships in the Middle East. , May 2017.32. i-Voting - Registration Closed. .33. PILDAT. Scorecard on Perceptions of Pre-poll Fair-ness. https://pildat.org/wp-content/uploads/2018/05/PILDATScorecardonPerceptionofPre-PollFairness_May-2018.pdf , May 2018.34. Iftikhar A. Khan. No haste in introducing i-voting for expats. https://epaper.dawn.com/DetailImage.php?StoryImage=12_09_2018_005_008 , Sept 2018.35. Staff Reporter. i-Voting being pushed to rig polls,says raza rabbani. , Sept 2018.36. Ikram Junaidi. CDA promises proper waste disposal to Supreme Court. , Jan 2014.37. Syeda Saima Shabbir. Judicial Activism Shaping the Future of Pakistan. https://ssrn.com/abstract=2209067 , January 2013.38. Supreme Court of Pakistan Diamer Basha and Mohmand Dam Fund. .39. Waqas Younis. Risky voting. , October2018.40. Ikram Junaidi. Electronic Voting not appropriate for Pakistan, senate body told.