Establishing a Legal Framework for E-voting in Canada
"Clearly, in a democratic society, the right to vote as expressed in s. 3 must be given a content commensurate with those values embodied in a democratic state" (Haig v. Canada, 1031).
Since the introduction of the secret ballot in Canada in 1874, arguably little has changed with the actual act of voting that involved a voter marking a symbol next to the candidate whom they wish to represent them in Parliament.
This is not to say that election regulations and legislations themselves have remained stationary. Electoral legislation has been changed to extend the franchise to women and other groups, lower the voting age, include party names on ballots, allow voters to register at the polls in rural and then urban areas, provide additional voting days and accommodate disabilities. Some of these were part of global trends, while others were uniquely Canadian decisions either spurred on by the electorate, or in some cases, such as prisoner voting, mandated by the courts based on Charter rights.
The push for Internet voting and other electoral technology (collectively called electronic voting or e-voting) is arguably a continuation of attempts to make elections more accessible. Instantaneous interaction with centralized computers has the potential to facilitate voting from home or abroad, and allows electoral officials to consider alternative methods of casting votes and transmitting them to the electoral authority. In an online system, a voter should be able to securely and privately visit a website, enter a user identification number and password, and cast a ballot.
Any change in how Canadians vote will require a comprehensive legal framework. The main purpose of this paper is to provide assessment to electoral decision makers in Canada considering implementing remote Internet voting in some capacity in a future federal election. This paper also references other electronic technology used in a controlled setting, albeit in a limited way.
The goal of this paper is to recommend a legal framework for e-voting in Canadian federal electoral events. This paper was commissioned by Canada's electoral authority, Elections Canada, an independent, non-partisan agency that reports directly to Parliament. Its mandate includes conducting federal electoral events and administering the Canada Elections Act, which would be a critical component of an e-voting legal framework.
Our research consisted of conducting extensive case studies of other jurisdictions' legal frameworks and experiences, which were then synthesized to include the most pertinent details. The legal framework for e-voting will span various legal and institutional instruments, including legislation, regulations and policies.
The goal of this paper is not to advocate the benefits or underestimate the risks of e-voting, but rather to present options for regulations and legislation that could mitigate and manage the risks and increase voter trust in a new system. Where possible, this paper recommends ways to make e-voting as secure and accurate as current means of voting.
Canada is not alone in exploring the use of Internet voting. Other countries, including Estonia, Switzerland, Norway, Australia and France, have already begun testing it at various levels of government, as well as developing regulations and legislation for its implementation. Even within our borders, a number of municipalities and political parties have experimented with Internet voting, and several provinces are also exploring the potential.
Our research consisted of reviews of academic and stakeholder commentary, observer reports from the conduct of elections and auditor reports, as well as a close look at naysayers and groups who remain suspicious of the risks of election results. Where regulations and legislation existed, this paper compares them with what currently exists in Canada. This paper is not limited to the circumstances surrounding Internet voting alone, but also reviews other uses of technology in the voting process, to look at how best practices under different contexts have been used.
Much can be learned about the legal framework and practical results of Internet voting taking a comparative law approach.
Law reform in Canada often involves studying laws that have been enacted and implemented in other countries, and adapting and adopting them. There are many advantages to this approach. Comparative exercises help to identify new policy concepts. They can provide a framework for thinking through the merits and demerits of various options. The text of legislation in other jurisdictions can provide inspiration and guidance on how to translate concepts into legal language.
Legislation in other jurisdictions can provide indispensable lessons for Canadian policy makers. Many apparently promising legal ideas turn out to have unexpected and adverse consequences. The subjects of legislation may find surprising ways to avoid or overcome their strictures. There can be no laboratory experiments in public policy making; the only way to test ideas is through societal practice.
Sceptics of Internet voting, however, might argue that comparative exercises must be viewed with caution. There is the risk in comparative exercises that "a little knowledge is a dangerous thing." An observer might focus on a particular law without appreciating the larger legal context in which it occurs. For example, it is true that Estonia has used Internet voting, but it must be recognized that its voting system is proportional representation, not like Canada's first-past-the-post system (where the candidate with the most votes wins). The difference of a few votes, or even a few dozen votes, might not have a substantial impact on the outcome of an Estonian election; seats are allocated on the basis of a percentage of votes achieved throughout the entire country. On the other hand, in a plurality system such as used by Canada, a single vote in a single constituency could determine the balance of power in the House of Commons.Footnote 1
As another example of how Canadian elections can be closely contested, in the 2011 federal election, 5 seats were decided by fewer than 100 votes (Funke, n.d). In highly contested votes, it is foreseeable that the number of electronic voters would in many constituencies exceed the margin of victory from traditional balloting. The practical implications of e-voting would intensify the need to ensure that instructions to voters and accessibility and reliability of the technology are intrinsically sound and adequately explained to the public.
Similarly, in examining the Estonian experience, one might overlook that part of its success is due to its reliance on the existence of a national identity card, for which there is no counterpart in Canada. Another risk in comparative exercises can be a misguided belief that the "law on the books" is the law in reality. The official legal code, for example, might be selectively enforced or not at all.
There is also the risk of overlooking the geographical, social or economic differences between the countries compared. One particular risk in the Canadian context is overlooking the issue of scale. A system that might work in a geographically and demographically small jurisdiction might run into unexpected difficulties if extended, without careful thought, to a much larger one. Currently, there is limited use of Internet voting at a national or even state level in a major election. The four most predominant deployments so far have been in Estonia at the national level, with a population about the size of Manitoba; France, where 240,000 overseas votes were cast; (Scytl 2012) and Switzerland, which limits e-voting to 10 percent of its voters. Only Australia's New South Wales state, which is roughly the size of Quebec, is closely analogous to Canada in using online voting in a parliamentary election. However, the state printed each Internet vote and manually counted the 44,000 votes cast online across 93 legislative districts. Additionally, Internet voting was limited to distance and disabled voters (Brightwell 2011).
In conducting the comparative part of this study, special emphasis was placed on jurisdictions such as Norway and Estonia, where there is extensive access to the legal standards used, and on reports on how e-voting actually worked in practice. We also drew on European as well as domestic sources of law and guidance.
This study has attempted to keep in mind as well that when the stakes are higher, as in national-level elections, there might be much greater incentive for misconduct.
The ability to alter the balance of party power in Parliament might be a big enough prize to attract earnest tampering efforts by persons motivated by commercial or ideological interests to affect the results of an election. Additionally, it may also attract those with no direct interest in the results of an election, but who want to disrupt the election.
The advantage of our methodology of extensively setting out principles, standards and procedures include the following:
- The discipline of formally and publicly documenting guiding norms will encourage more precision in thinking through the real-world application of those norms.
- The existence of a well-documented set of guiding norms will assist parliamentarians in evaluating a proposed test of Internet voting and facilitate public input into that process and the promotion of public trust in the process.
- The documentation will facilitate the eventual legislation of changes to the Canada Elections Act or its accompanying regulations should Parliament wish to continue testing Internet voting or to adopt it as an integral part of the national voting system.
There are limits, however, to how much can or should be placed in a set of guiding norms. Some discretion is likely desirable to select among different technical options, including the hardware for receiving and tabulating votes and outside experts to test and certify systems, and to react to problems as they arise.
Structure of This Paper
This paper contains two major parts: "Background, Context and Literature Review" and "Main Findings and Recommendations."
The first part contains an overview of how e-voting is used in both an uncontrolled environment and a controlled environment. This part provides a brief overview of both the benefits and concerns raised with introducing e-voting. This is followed by an overview of Canada's constitutional right to vote and an introduction to the notion of functional equivalence, which requires synthesizing key attributes from the current legislation and applying them to e-voting so that the e-voting rules function in an equivalent way to paper-based voting. Next, there is a summary of the experiences in other jurisdictions with e-voting as well as a brief overview of some of the major international organizations that have conducted work on e-voting. The major findings of these groups are also summarized as an appendix. Lastly, there is a summary of the key values and attributes that are important to conducting elections, whether by computer or paper.
The second part surveys specific components of legal frameworks for e-voting and makes recommendations on what should be included in a Canadian framework. This is introduced by a brief discussion about the form of the framework, such as comparing legislation passed by Parliament with regulations and policy issued by an electoral authority. The discussion and recommendations are grouped into nine topics:
- When are voters eligible and under what circumstances can they access e-voting?
- What measures are taken to ensure the e-voting system is transparent?
- What major roles and responsibilities are required to administer e-voting?
- What contingency plans should be in place for worst-case scenarios?
- What electoral offences should be created for e-voting?
- What technological standards and consultation are needed?
- What types of testing are needed to ensure the integrity of the vote?
- What unique considerations are required for controlled voting?
- Is there a need to restructure the electoral authority?
Return to source of Footnote 1 For example, in Canada, a 2005 non-confidence motion was decided by a single vote in Parliament. If that Member of Parliament had been elected by one vote, then the vote in the constituency would have effectively determined the outcome of that motion.