Principles

This draft strategy examines the desirability and technical feasibility of electronic voting (e-voting), and considers how it could be implemented for general elections and referenda. The proposed approach is cautious, and aimed at ensuring that high levels of public and political confidence in the electoral system are maintained. 

Voting methods currently in use are working well for most people and enjoy high levels of public confidence. However, there is an existing demand for e-voting from sections of the community for whom paper-based ballots or the need to attend polling places result in accessibility difficulties or an unsatisfactory voting experience, for example a lack of independence and privacy. 

Looking to the future, there are indications that the demand for e-voting is likely to grow. There is a risk that voters with a strong preference for e-voting may not vote at all if the choice is unavailable to them. It is important to note, however, that e-voting is only part of the solution to diminishing voter turnout. The introduction of e-voting would be a highly visible demonstration of the viability of electronic delivery of government services. Its introduction would support Government and state sector priorities for the next decade, including the Digital Strategy and the Disability Strategy. 

E-voting will not suit all voters. It will be inaccessible for some. There should be no pressure on voters to change. Current methods of voting should be maintained for the foreseeable future, and e-voting should only be an additional and optional means of casting a ballot.

The costs of e-voting will exceed savings until the volume of e-voters builds up. Under a cautious step-by-step approach, this is not expected until several cycles of elections have been completed. A wide range of non-financial benefits have been identified however

Proposal

The  draft strategy suggests that e-voting should not be made widely available until three or more cycles of pilots at elections have been completed. This enables a prudent step-by-step path of learning and development to be followed. In this way, increasing demand for a wider range of methods to cast a ballot can be met while maintaining democratic principles and a trustworthy electoral system, and ensuring the integrity of individual votes.

Small scale, carefully controlled pilots that would test specific elements of e-voting solutions in real electoral environments, is proposed. The greatest benefits would be gained by enabling electronic voting from unsupervised locations such as home, work, or public Internet facilities. Telephone and Internet voting methods are favoured for pilots, and both could be piloted at the same time. 

Unsupervised remote voting raises particular challenges for: voter privacy and the secrecy of the vote; the exposure of the voter to undue influence or coercion; system security and integrity; and the ability of voters to be confident that their vote has been received and counted as intended. These challenges can be met, with the proposed solutions to be tested in trials and pilots.

Online voting also requires a higher level of authentication of voters’ identity compared to in-person voting as a different level of risk applies. E-voting could potentially be susceptible to large scale electoral fraud or attempts to disrupt elections. Such crime may originate from outside New Zealand. Use of the Internet for voting raises a number of risks which are particular to the nature of the Internet and personal computers (such as security weaknesses and highly coordinated ‘denial of service’ attacks which result in web sites becoming unavailable to users). These problems have been well documented in other e-voting analyses and are likely to affect public trust. The draft strategy takes account of these risks, with a package of mitigations suggested.

To gain maximum value, it is suggested that the initial pilots be designed to improve the accessibility of voting to voters who are currently disadvantaged by the paper-based system, such as blind and vision-impaired voters, and voters with other disabilities; and to test the various aspects of an e-voting system that would enable its secure and reliable extension in the future. After evaluation and subject to satisfactory outcomes, the scale and scope of pilots would increase over the course of three or four elections, thereby: facilitating access to voting; enabling cost efficiencies in the electoral system; and, as scale increases, improving choice and convenience to a wider base of voters. The approach should be as open to scrutiny as possible, including publication of system details and the outcomes of pilots. It is suggested that independent e-voting observers be appointed.

The option of electronic voting in polling places (such as the ‘kiosks’ used widely in the USA and parts of Europe) is not favoured. The costs would be high and the benefits, compared to the well-functioning paper-based system, would be minimal.

The Chief Electoral Officer must have the authority to suspend or cancel e-voting pilots at any time if the integrity of the election is threatened or if public confidence in election outcomes could be lost.Two important features of the proposed e-voting system are designed to mitigate key risks. The first is that e-voting be undertaken in the advance voting period – usually about 17 days ending on the day before polling day. This would ensure e-voters are not under time pressure and can choose an opportunity convenient to them when they are likely to be in private. Advance e-voting also provides administrators opportunities to manage any technical, security or process issues that may arise, with minimal impact on e-voters. It is unlikely that an e-voter would inadvertently be deprived of their vote in the event of any problem with the e-voting system because advance voting would allow them to vote conventionally on election day.

If an e-voter is not satisfied that they can cast a secret electronic vote of their own free choice, they can cast a traditional ballot at a polling place, and that vote would over-ride any previous e-vote. The democratic principle of everyone having the same number of votes is maintained in that everyone has the same number of votes counted.

The usability of e-voting systems is just as important to the success of e-voting as other challenges, and must be balanced carefully against the complexity that can arise from risk management measures. Pilots should include assessments of user-friendliness, e-voter satisfaction, and wider public attitudes. E-voting on a moderate to large scale should not proceed until there is well justified public confidence in all aspects of the system.

Implementation

E-voting will not be possible for the 2008 election, and there are significant risks with implementation in 2011. Further legal and policy work (including public consultation) and enabling legislation are required in advance of the introduction of e-voting. The preparation of a detailed business case and cost estimates is also advocated.

Two implementation timelines are included, commencing either in 2011 or 2014. The 2014 timeline would enable the successful implementation of e-voting, without compromising the delivery of high quality election services in 2008, when the general election must be held. The 2011 timeline has been included in the draft strategy to illustrate the operational risks to the Chief Electoral Office. These would be exacerbated if a by-election or a citizens initiated referendum were also to be held in 2008/early 2009.