Technology will continue to transform the way elections are conducted. Whether that statement is seen as promise or threat depends upon a number of factors. There are differing opinions as to the criteria for a successful election, varying levels of understanding and acceptance of different technologies, debates about what are reasonable costs for conducting an election and competing demands for credibility and convenience.

The International Foundation for Electoral Systems (IFES) has extensive experience dealing with these issues and researched the effective use of technology in the election process through a series of case studies. This study will be published in Summer 2010 and through analysis of each case will seek to determine whether the introduction of particular set of technologies led to a net positive or negative impact on the election process, and whether any positive impact provided improvements significant enough to justify the costs. The study uses these findings to develop a set of common factors that may help to provide “best case” guidelines for maximizing the chance of successful implementation of new technologies into election processes. The goal of the study is to provide Election Management Bodies (EMBs) and other critical stakeholders with information and data that they can utilize to guide their decision-making on technology implementation. While this publication is not necessarily an exhaustive survey of technology implementation in the election process, IFES sees it as a vital first step in providing guidelines based on commonly accepted principles used by electoral experts around the world.

Four key concerns guide the introduction, implementation, and management of technology in the election process:

1. Efficiency versus transparency
2. Confluence of technology, elections, and funding cycles
3. Planning for the technology life cycle
4. Development of criteria for adopting appropriate technology

Efficiency vs. Transparency

The use of technology in the election process is a subject that often provokes contentious debate, and this debate often concerns the role of technology in introducing efficiencies into many facets of the election process versus its role in possibly reducing the level of transparency demanded of credible electoral processes. Technology is well-suited to solving problems of efficiency, and election management has no shortage of such problems. Without doubt, many election management processes have been significantly improved through the application of technology. Large scale data entry, management of voter registration data, production of ballots and logistical planning are examples of problems where application of technology has made major advances. However, there are some electoral processes that demand transparency (i.e. vote counting and tabulation) be given even greater emphasis than efficiency. The application of appropriate technology for elections requires careful deliberation to determine when this is the case, followed by even more careful deliberation to decide whether there are tools that can help promote transparency.

The implications of this philosophical distinction may have major, far-reaching and long-lasting impacts upon the practice of democracy. One approach to technology focuses on centralized control, dependency upon a technological elite and application of manufacturing principles of uniformity. In many domains, it is desirable to create “black box” components that focus on inputs and outputs and require no knowledge of the inner logic; such components lead to efficient development and deployment. In conducting voting, counting and tabulation, however, the “black box” approach is patently undemocratic, taking power away from the many (election officials, party agents, observers, media), and putting into the hands of the few.

Proponents of black box voting, counting and tabulating systems argue the systems are trustworthy. A common argument in the last decade to procure and deploy direct recording “direct recording electronic,” (DRE) voting machines was that the machines are similar to automated teller machines used by banks. Not to go unnoticed, the largest vendor providing such machines was Diebold, whose core business provides such ATMs to banks; “if we can trust these machines with our money, surely we can trust them to accurately record our votes.”¹ The strongest argument against black box systems in elections is not whether or not they are accurate or trustworthy, but whether they provide the required level of transparency to empower stakeholders to observe and confirm fairness and accuracy of the process. While we should seek to ensure all components of an election are trustworthy, it is even more important to ensure that all are accountable.

Technology, Elections, and Funding Cycles: When to Introduce New Technology

One important and nearly universal guideline relates to appropriate timing for introduction of new technology. The concept of the electoral cycle has been advanced by International IDEA and the European Commission to move away from a focus on planning for a single election and to promote more sustainable election planning. There are a number of visual representations of the cycle, but all divide election planning into three distinct phases: Pre-Electoral Period, Electoral Period and Post-Electoral Period.

Whenever possible, new technology should be introduced during the Post-Electoral Period. EMBs typically have fewer demands on their time during this period. Also, by beginning during the Post-Electoral Period, the EMB allows more time for feasibility studies, pilot projects, modification of procedures and training, all of which are necessary to maximize the success of a new technology.

Planning for the Technology Life Cycle

For at least 30 years, businesses have planned technology infrastructure with an understanding that the cost of new technologies does not end with initial procurement. There are costs for deployment, training, maintenance, upgrades, security and all technologies have an anticipated life expectancy after which they become obsolete and require disposal. Technology Lifecycle Management has a number of implications for election technology planning.

With most countries conducting national elections on either a four or five year cycle, it is important to recognize that anticipated life expectancy means that any new technology will be used for a limited number of elections.²A quick review of changes in the Windows operating system illustrates the speed with which new technologies become obsolete and are replaced by newer technologies:

This rapid evolution of the Windows operating system means that most computer systems from a decade ago, if still operational, are running on operating systems that Microsoft considers obsolete. The official policy of Microsoft is to provide mainstream support for 5 years, followed by extended support for 5 years. The difference is that “Microsoft will not accept requests for warranty support, design changes, or new features during the Extended Support Phase.”³

These ongoing changes are not limited to the operating system. Since 1990, personal data storage devices have evolved from floppy disks, through CD-ROM, DVD, USB flash drives and online storage. Communication has gone from dial-up services to proprietary services such as Compuserve and America Online to the internet, with broadband data transmission via DSL, cable, fibre optic, satellite and mobile phone networks. Peter Wolf points out that even by relying on standards it is difficult to find any longer term stability; the ANSI standards for biometric systems has gone through four significant versions since 1993. Overlaying technology timeline with an election timeline, Wolf produces the following graphic illustration of how much change occurs over a few election periods:

The graphic above illustrates that the decision-making process employed by EMBs should not only be taking into account technical considerations for the election cycle in which the technology is expected to be utilized, but long-term considerations related to technological evolution as well.

Technology Lifecycle Management (TLM) takes a broad planning view over the design, procurement, deployment, management and disposal of all elements in the organization’s technology infrastructure. TLM can provide a realistic estimate of total cost of ownership, training needs and deployment schedules – and can assist election administrators with the difficult task of coordinating introduction of technology within the election timeline. Equally important, TLM provides a tool for anticipating budgeting requirements necessary to ensure sustainability of the new technology.

The different stages involved in Technology Lifecycle Management include:

1. Assessment and identification of organizational mission, objectives and policies for determination of appropriate technology.
2. Procurement of technology, including feasibility studies, pilot projects, specifications and vendor evaluations.
3. Deployment of systems and training of end users.
4. Maintenance, repair and necessary upgrades, including ongoing helpdesk services and technical support.
5. Plan for proper disposal, including provision of security for all data stored on any media

This broad planning perspective can help avoid rushing to implement new technologies during critical election periods, plan ahead for funding requirements and anticipate staffing needs.

Criteria for Determining Appropriate Technology

Nearly every discussion of technology and elections at some point includes reference to the concept of appropriate technology, often as a way of justifying application of a particular tool. Through overuse, the concept has been stretched nearly beyond recognition and used to support almost every imaginable approach to addressing election related issues. It seems that the concept of “appropriate technology” has become bendable enough that it can serve whatever purpose the user wants.

One counterbalance against this elasticity of meaning is to go back to the origins of the concept. E.F. Schumacher is credited with initiating the discussion of appropriate technology in his book, “Small is Beautiful: Economics as if People Mattered.”⁴ Schumacher criticized international development practices of the 1960s as an attempt to export models of management and technologies that may not have been appropriate for the countries they were being foisted upon. One example he gave was in agriculture, where industrialized nations tried to convince non-industrialized nations that it is preferable to import machinery that will allow a single farmer to manage crops on a thousand acres. Schumacher questions the impact upon the thousand farmers who had previously owned one acre each and were now unemployed. Hence, the title of his book counters the claim that bigger is better, that centralized is more efficient, that minimal human involvement is an advance over maximum human involvement.

There is a growing consensus in identifying which basic principles should be applied to all election processes; and technological processes should not be exempt from these standards. A good list of standards for helping to determine “appropriateness” might include:

Accessibility – Should be able to be maintained by ordinary mortals without long-term, hands-on support of a team of engineers
Security – Should provide protection from unauthorized access, and from loss due to natural disaster or malice
Accountability – Must ensure all data entry and modification is done according to legal guidelines, and with a clear record of who did what
Auditability – Must include capability to verify that the output of the technology is logically consistent with the input
Transparency – Must provide for a clear understanding by stakeholders of how the technology works, and the implications for election process

General Recommendations on the Way Forward

There is a risk in trying to prescribe a one-size-fits-all methodology for implementing effective and sustainable technology; every country has a unique set of problems and cannot necessarily import the solution of another country. But there is even greater risk in introducing new election technologies without some clear and careful methodology to ensure the technology meets the requirements. So, with the disclaimer that this is not a comprehensive list and with an invitation for further refinement, the following steps are proposed as part of the way forward:

1. Identify the problem
This may sound like an obvious first step, but countless projects have begun with a solution in mind before the problem is fully identified. One example will illustrate the extent to which this happens. Many countries have at least considered introducing a biometric system as part of voter registration, and in many countries this type of technology can be a valuable election management tool. However, the technology should only be considered if there is a problem that the technology can help to address. In the author’s experience working with EMBs, some have turned to biometrics to solve everything from underage voters to foreign voters to deceased persons on the list – none of which can be addressed by a biometric solution. In cases where there is a problem with duplicate registrations, EMBs have attempted to solve the problem with a biometric system before determining whether existing laws and procedures are being enforced or whether the laws provide any mechanism for removing persons who are identified as duplicates. The problem-identification step should be able to clearly identify a difference between what should be and what is. It should further include a clear understanding of why things should be the way they are envisioned. Without this clear understanding, there is a very real danger that any solution introduced will solve a wrong or nonexistent problem.

2. Invite broad discussion from stakeholders and implementers
Even if a real problem has been clearly identified and there is a good, affordable and easily available technology that can address the problem, it is still important to invite discussion from political parties, civil society and electoral staff who will be responsible for implementing the solution. Often these discussions can provide valuable additional information about the scope of the problem and direction to effective implementation. At a minimum, discussions may be able to broaden acceptance of the final approach. If it is impossible to get endorsement from stakeholders and implementers there is a strong probability that the problem has not been adequately understood, and an even stronger probability that the solution will not be accepted.

3. Consider whether there is a need to modify the legal and/or procedural framework
Effective technology often requires a change in procedures. A thorough “requirements analysis” process will address not only the tools needed, but also the systems and processes that must be reengineered in order to shape an effective solution. If the underlying issue involves procedures that result in a flawed implementation of a particular facet of the election process, then introducing technological solutions without addressing flawed procedures will not necessarily lead to a more efficient process. In many cases, legal regulations or statutes may also need to be addressed so the technological solution complies with electoral and constitutional mandates.

4. Provide for required staffing
One problem that confronts many EMBs, especially in the developing nation context, is the difficulty of recruiting and retaining good, qualified technology staff. In order to compete with the private sector, EMBs should evaluate whether there is a significant disparity between their pay scales and those offered in the private sector. If so, it may be necessary to find ways to adjust the pay scale in order to be competitive; otherwise it will be difficult to recruit technology staff, and any staff recruited and trained will be repeatedly tempted by outside offers.

There are inducements beyond salary that can help to recruit and retain staff, and the EMBs that are most successful have been able to institute one or more of the following measures:

• Obtain a waiver from Public Service/Civil Service payment constraints for mission-critical elections IT personnel
• Provide opportunities for Information Communication Technology (ICT) personnel to travel overseas on secondment to international Non-Governmental Organizations (NGOs), the United Nations, Organization for Security and Cooperation in Europe (OSCE), other Electoral Management Bodies (EMBs) et cetera, and be remunerated accordingly
• Allow lengthy leaves of absence for relevant study or personal development projects
• Seek joint funding support for IT salaries from international donors

If it is not possible to retain adequate in-house staff, the EMB should carefully consider whether there will be adequate funds to pay for continued support from external vendors.

5. If possible, start small
Perhaps more than any other measure, a decision to start with small steps and then develop incrementally can help avoid major failures in technology projects. Some initial steps that can help build a foundation include:

Feasibility tests
Whenever possible any new technology should be tested in small trials that involve representative samples both of the staff who will be required to use the technology and the persons who will be affected by it.

Pilot testing
A pilot test in two or three communities can provide valuable “lessons learned” that can be built into procedures and training manuals before rolling out a new system. In some cases, the pilot may reveal weaknesses that cannot be adequately addressed.

Build infrastructure
The cost of computers and communication networks continues to drop at a rapid pace. Providing adequate tools for staff both at headquarters and at field offices can have a major impact on the organization’s ability to effectively manage elections. Over the past decade, it is surprising how many EMBs have rolled out large, expensive systems for voter registration while staff responsible for accounting, logistical planning, supply and fleet management, and human resource management limp along with outdated computers and software, or worse, without any computers at all.

6. Provide adequate time
Any project that must be implemented on an overly aggressive schedule is a project that has a very high risk of failure. If there is not adequate time to carefully define needs, discuss possible solutions with stakeholders, conduct feasibility tests and pilots and reevaluate the approach repeatedly throughout this planning cycle, then there is probably not enough time to do an effective job in implementing the technology.

7. If you have never done it before, don’t test it during a major election 
This is related to the “start small” principle listed above. Elections should not be used as a testing ground for new, unproven technologies, nor as a place for election staff to get their first introduction to a technology. How can you expect to make a knowledgeable decision about a handwriting recognition system if you have never experimented with one, or with a biometric system if you have never taken a fingerprint or installed an automated fingerprint identification system (AFIS) software package?  If no one within the EMB has any experience with a proposed technology it is especially important to start with a series of feasibility tests and pilots to gradually gain familiarity with the systems to be capable of guiding a responsible decision-making process.

8. Do not introduce technology to compensate for poor procedures
One of the most important steps you can take to maximize the probability of a successful technology system is to first do everything possible to address the problem without resorting to technology. Once you have a sound legal framework, a good set of organization procedures and well-trained staff, if the problem is not already solved, then you at least have the best possible basis for introducing technology. Bill Gates said it clearly: “Automation applied to an efficient operation will magnify the efficiency. Automation applied to an inefficient operation will magnify the chaos.”

9. The Common Sense Test
To a great extent implementing effective and sustainable technology is a matter of common sense, yet the number of failed or marginally successful election technology projects is an indication that common sense has failed to serve as a guide. Before buying a new car, a new watch, even a new jacket, most of us run through a set of questions in our minds. Does it meet my needs? Does it fit? Can I really afford it? Is it easy and relatively inexpensive to maintain? Can I have it delivered by the time I need it? Will it last long enough to justify the price? If the answer is no, then common sense dictates this is not an effective, appropriate or sustainable purchase. By exercising just this much discretion when planning for implementing election technology this common sense test would significantly improve EMBs’ chances of implementing effective, appropriate and sustainable technology.

1. Whether such machines are, indeed, trustworthy is a subject that has already seen much debate. That we trust them to manage our money is a fallacious argument, since bank accounts are subject to a number of accounting and auditing processes that do not depend on the ATM.
2. A comparison of Technology Lifecycle and Election Cycle was presented by Peter Wolf, International IDEA, at EC-UNDP-IDEA Joint Thematic Workshop on The Use of Information and Communication Technologies in Electoral Processes in November 2009. We are indebted to him for introducing this concept as well as for the analysis of technology change since 1984.
3. Microsoft Support Lifecycle website, Support
4. E.F. Schumacher. “Small is Beautiful: Economics as if People Mattered.” Blond and Briggs, Ltd. London. 1973.

International Foundation for Electoral Systems
Contact:
Michael Yard, Technology Analyst
myard@ifes.org