Technology and the Voting Process
VIII. Implications of the New Technologies for Electoral Administration
A. Effect on criteria for effective electoral administration
Our analysis shows that adoption of the new electronic technologies in the electoral process can have a generally positive impact on the seventeen criteria for electoral administration presented in Section III. Fundamental to support for technology-based voting is the presence and functionality of a relational database. This database would hold information on registered voters, their electoral districts and the candidates. Elections Canada currently maintains the foundation for this data base in the National Register of Electors.
Our approach to assessing the organizational and administrative impacts involved interviews with Elections Canada officials to gain an understanding of the issues which we then presented to technology experts for their views.
1. Democracy – one eligible voter can cast one vote
If any of the various voting options were offered on election day, Elections Canada would require an on-line, real time database of information across the options to ensure that no person voted more than once. Alternatively, if different options were provided on different days (e.g., advance poll), this would remove the necessity for real time information.
2. Accuracy – the final vote count is reflective of the intent of voters
Technology options can provide electors with the opportunity to confirm that their vote is recorded as they intended. Electors can change or cancel their recorded vote prior to it being cast.
3. Security – measures in place to protect the integrity of the process
Technology-based security controls are mature and widely used in many industries. Security controls are also an area where there is continuing active advancement to support emerging applications of technology.
Use of unique Personal Identification Numbers:
- will link to an elector (and therefore, electoral district);
- will be canceled upon casting of a vote; and
- will deny access to system after a specified number of attempts with an incorrect PIN.
Telephone voting access from a particular telephone number can be blocked if the security system detects a specified number of attempts to vote with non-valid PIN.
Connection to the voting system by Internet or telephone can be disrupted at the end of a specified time period to prevent blocking of lines to electors.
Firewall protection of the database will be a dynamic issue to be addressed and updated periodically.
4. Secrecy – no vote can be traced to the voter
The act of casting a vote can be linked to an elector in order to mark the elector as having voted. After confirming and casting the vote, it will be "detached" from the elector and placed for tabulation with no retraceable path.
5. Verifiability/auditability – the vote results can be verified after the count
Voting results can be recorded electronically. These can be re-accessed electronically or presented as transcripts. Counts of electors and any spoiled or incomplete ballots can be compiled during the poll.
6. Privacy/confidentiality – information collected on electors is used for elections purposes only
Use of and access to stored data must be covered by legislation.
7. Transparency – process is open to outside scrutiny
Agents of candidates can view updates of electors who have voted using new technology such as computer screens or printouts.
8. Accessibility – the specific needs of eligible electors are taken into account so that none are disenfranchised
Use of technology can expand access to numerous groups of electors who have transportation, mobility or availability of time issues. It would probably be several years however, before a technology based voting option could be selected to support 100 percent of the electorate.
9. Neutrality – electoral processes do not favour one candidate or party over another
Adoption of technology options would be for the convenience of the electors not the candidates.
An exploration of technology options does offer the opportunity to present candidates for selection with such additions as party colours or candidate photographs. Adoption of such opportunities may support the multi-lingual requirements.
10. Simplicity – the voting processes do not make voting unduly complicated
A large proportion of the electorate is familiar with the use of many technologies; such as, banking terminals, automated phone services for such tasks as stopping or starting newspaper delivery, voice mail, remote control access devices for televisions, security access, credit and debit cards. Administration of the electoral system will be simplified through automation of many tasks.
11. Flexibility – the voting process can handle a variety of ballot styles and counting formats
The use of technology provides sufficient flexibility to allow for different options for voting as well as different ballot styles and counting formats (e.g., telephone voting, voting using a touch screen kiosk, voting by computer).
12. Scalability – the voting process can be scaled to handle large and small electoral events
Hardware capacity and speed is increasing exponentially and indications are it will continue to do so. Current technology will support a database large enough to support a federal election.
13. Recoverability – the voting process provides for duplication of systems to prevent data loss
Storage technologies will retain data in a compressed format; therefore, storage of duplicate data does not require a doubling of storage capacity. Transcripts of data can be prepared and stored.
14. Mobility – the voting process allows votes to be cast from locations other than the traditional polling station
The use of technology based voting options can eliminate the need for an elector to attend a polling station. Electors could cast their votes from any site where telephone or Internet access is available. This could permit Canadians to vote from overseas locations.
15. Speed of count – results can be reported quickly
The count of the cast votes could be ongoing at a national or site basis. Transmission of site counts could be periodic or at the close of polls. Compilation of transmitted results on the central data base could provide results in minutes.
16. Cost effectiveness – the voting process is effective and economical
Adoption of technology can decrease the time involved in the heavily manual electoral administration processes. Use of the telephone or Internet for electoral process relies on technology available to the electorate. Elections Canada could provide the infrastructure to support the process. The planned use of acquired technology, in addition to federal elections, should be a factor in determining the cost effectiveness of the acquisition. Options for multiple use of the technologies should be investigated and considered.
17. Technical durability – the voting process allows the basic electoral infrastructure to be reasonably insulated from obsolescence
Elections Canada would have to carefully evaluate the acquisition of supporting technologies in an effort to determine what will be most viable for the future. Periodic migration to emerging database technologies should be anticipated. Keeping the database current with system versions will help ensure the data integrity. Acquisition and maintenance of specialized end use terminals must be carefully considered if these may not be used for another four years. This does present an opportunity for partnerships with provincial and municipal electoral bodies.
B. Legal considerations
In our work on this project we developed a list showing the sections of the Canada Elections Act that would have to be amended to allow for electronic voting (please see Appendix B). On reflection, we have concluded that amendments to the Act as listed would serve to complicate an already difficult and complicated piece of legislation. In fact, the Act is so intrinsically linked to the current process that changes would need to be sweeping and widespread. The better course, if Parliament were considering amendments to allow for the possibility of electronic voting, would be to write a new schedule of the Act to permit electronic voting for whatever purposes, and under whatever terms and conditions, Parliament deemed appropriate. This was the course followed in the case of special voting rules.
It is a "law" of information technology that as the power and efficiency of computing systems increases, costs drop, often dramatically. Thus, the degree to which cost is a factor in the potential application of the new technologies to the voting process (and it is in almost all cases) changes with every passing month. And if experience is any guide, what might appear an impossible costly solution to a problem today (in the area of security or privacy) is likely to be resolved cheaply within only a couple of years. So cost is a factor, but it is a relative factor, and one that is constantly changing. The key point is that technology as it applies to voting is becoming cheaper, not more expensive.
A second important consideration, and one that helps to keep the entire issue of technology and voting in proper context, is that the different voting technologies are each relevant to meeting the accessibility needs of different groups in the Canadian population. A particular technological solution for voting by persons with disabilities might appear expensive until one realizes that current measures to ensure accessibility can also be very costly. An average cost-per-elector at the federal level of, $1.50Footnote 22 for the conduct of a federal election may include costs of ten times that for members of specific groups where accessibility is already an issue. Technological solutions aimed at the same group should be assessed in that context.
D. Opportunities for partnering
One question which must be answered is whether or not an investment at the federal level in new voting technologies could be recouped through cooperation with other jurisdictions in Canada or abroad. While it would be presumptuous to think of recovering significant portions of the federal cost, it would also be only reasonable to think that some provincial, municipal or foreign jurisdictions would be prepared to share some development costs, or (in Canada) to share the operation of data bases or other infrastructure. This is already the case with some aspects of the National Register of Electors. Furthermore, such cooperation within Canada would reflect and reinforce the notion of the "single taxpayer" and provide value-for-money in the use of tax dollars for the purposes of conducting federal, provincial and local elections.
Beyond the issue of cost, there already have been significant benefits to Canada from leadership and partnering in electoral administration. These benefits would be amplified in areas such as electronic voting where leadership in technology and related administration could pay off in everything from influence at the governmental level to broader business development.
Return to source of Footnote 22 The cost per elector of polling activities for the 1993 and 1997 General Elections was $1.45 (adjusted for inflation 1997 dollars) and $1.54 (estimated) respectively. These cost estimates are comprised of fees and payments made to deputy returning officers, poll clerks, central poll supervisors, information officers, landlords (rental of polling stations) and printers (printing of ballots) for advance, mobile and ordinary polls. Also included are cost estimates for the stock required at polling stations (e.g., ballot paper, guides/instruction manuals, forms, signs, voting compartments, ballot boxes and stationary). All other costs related to the preparation and delivery of an election are excluded (e.g., elector registration, Special Voting Rules, and candidates and party reimbursements).