Language:
Search
 
Internet Voting
State of Geneva
Switzerland

The Problem

“No taxation without representation”, shouted the American revolutionaries who achieved the US independence in 1776. They protested at the fact that although they were paying taxes, they were excluded from political life. The 1789 French Revolution reversed this motto by linking the voting right to the payment of a tax: the “no representation without taxation” principle was born. This “suffrage censitaire”, known in English as “poll tax”, would dominate the early stages of the Western democracy.

The poll tax has been gradually abolished over the 19th century. France, for example, cancelled it from its Constitution in the 1850’s, Belgium in 1893 and Canada in 1898. In the United States, the 24th Amendment to the Constitution, adopted in 1964, made it illegal.

Census suffrage was dead. Or was it?

The low turnout common to many modern democracies doesn’t simply reflect a diminishing interest for public issues and a retreat into the private life, but also an inadequacy between the access channels to vote and the state of the society. The expression “invisible poll tax”, borrowed to Daniel Gaxie, aptly describes this situation.

Turnout figures and trends in many developed countries seem to suggest that there are still barriers preventing the access to political rights or simply to the polls. Whether registration procedures, scarcity or short opening times of polling stations or the presence of uniformed officials on the voting premises, “invisible” obstacles still exist in the eye of many a voter. As a result, access to vote isn’t as easy as one would imagine in our era of universal suffrage.

In Geneva, the introduction of postal voting has led to an increase of 20 percentage points of turnout over the 1995-2003 period, to reach an average of 55%. Turnout has stabilized since but hasn’t decreased. The gains are being kept.

It seems reasonable to conclude from this experience that postal voting has broken an invisible barrier preventing voters to cast their ballot. This barrier was not so much the scarcity of polling stations (in Geneva, they were 70 for some 200’000 registered voters and at least one in each village or neighborhood) nor restricted opening times of these. Bringing the vote directly into the voters’ home through postal voting appears to be the real barrier breaking factor that triggered the turnout increase. It can be described as a qualitative change that made a difference for voters.

In 2001, the Swiss federal government launched an internet voting pilot project. The government wanted to capitalize the success of postal voting by offering a one-stop voting solution. With postal voting, indeed, one has to drop her ballot in a mailbox.

The 8 e-enabled official ballots conducted in geneva since 2003 have shown that internet voting is the favorite channel for citizens who do not otherwise vote. For the September 2004 ballot, which we investigated more thourghly, five out of nine voters describing themselves as systematic or quasi-systematic abstainer chose internet to cast their ballot. In other words, while 21.7% of all voters used internet, 55.5% of the usual abstainers who voted on that occasion did so, more than twice as much. Similarly, the online turnout of citizens who vote only occasionally - 30.8% - is higher than the global value of internet turnout, 21.7%.

In summary, users of eVoting come mostly from the files of the abstainers. They declare that eVoting has the potential to mobilize them. Once they’ve tried it, they stick to it. It can reasonably be said that internet voting lowers the access threshold to vote for a group of persons who otherwise would not participate in polls.

Solution and Key Benefits

 What is the initiative about? (the solution)
In Geneva, eight official ballots have been organized offering voters the possibility to cast their vote online. No other state or country has such a long record of internet ballots. Far from demonstrating the doubts over electronic voting that can be observed elsewhere, voters used this new possibility in percentages ranging from 22% to 28%, much more than expected.

It must be emphasized that the tendency to vote online has been very similar among municipalities and ballots. The issues at stake, whether municipal, cantonal or federal, have had no effect on online turnout, while they impacted the overall turnout.

A phone survey was organized on a sample of 1014 registered voters in the days following the September 2004 ballot. The sample composition reflected the registered voters’ splitting into abstainers and effective voters, and, in this second group, among polling station voters, postal voters and online voters. All municipalities involved in this ballot had already access at least once in the past few months to online voting.

As far as socio-demographic variables are concerned, the findings of this survey - the first of its kind in any of the countries experimenting with online voting - confirm the socio-demographic results that the online surveys conducted in previous ballots had already revealed. These can be summarized as follows:
- Postal voting is appreciated by voters over 50 years of age, while internet voting is strong among voters under 50.
- Women tend to prefer postal voting; men prefer online voting. The difference between genders is however weak among the younger generation.
- The use of internet voting is positively correlated to the level of education, while education level does not affect postal voting nor polling station voting.

Is there a way to qualify this group, a common characteristic that makes them a group and not only an aggregation of individuals?

To try answering these questions, a set of “ICT variables” were defined. These were the respondents’ self-assessed IT skills, their internet utilization frequency, their confidence in online information, online communication, online transactions and in the internet voting procedure, the place of internet access (home, office or other) and the type of internet connection (dial-up or broadband).

The multivariate analysis shows that the subjective elements in the voters’ relation to internet, e.g. their self-assessed IT skills, frequency of internet use, confidence in internet communications and in the procedure of internet voting, as well as their type of internet connection, are the predominant drivers explaining the use of internet voting. The statistical analysis reveals indeed that this model explains to a far larger extent the choice of online voting than any other model.

Here, the divide is not between the “internet access have” and “have not”, but between “computer skills have” and “have not”. This divide is not correlated to socio-demographic variables, but to subjective feeling of ease and trust with ICTs.

While postal voting geographically displaced the vote and introduced it into voters’ home, internet voting created a paradigm shift and reached the intimacy of those who feel confident with ICTs. Internet voting appeals to a subjective dimension of life, which is at the same time private, emotional and characterized by confidence in one’s own skills and the technology. On the contrary, postal voting requires no skills and the confidence in the postal service has never been an issue.

Actors and Stakeholders

 Who proposed the solution, who implemented it and who were the stakeholders?
2000: launch of the project, conception, analysis, search for partnership.
2001: development and internal testing.
2002: first external test with student as a blank vote (2000 student did votes), bulding the operational system.
January 2003, Anières (1250 registered voters), municipal referendum, 44% of ballots cast online. This first official internet ballot ever in Switzerland is made possible by a legal provision enabling testing new voting methods.
November 2003, Cologny (2500 registered voters), municipal referendum, 29% of ballots cast online;
April 2004, Carouge (9050 registered voters), municipal referendum, 26%of ballots cast online;
June 2004, Meyrin, (9180 registered voters), two municipal initiatives, 22% of ballots cast online;
September 2004, four municipalities (22'000 registered voters) voting on four federal referendums and two cantonal ones, 21,8% of ballots cast online;
October 2004, Vandoeuvres (1382 registered voters), municipal referendum, 32% of ballots cast online;
November 2004, eight municipalities (40'000 registered voters), three federal referendums and two cantonal ones, 22,4% of ballots cast online;
April 2005, fourteen municipalities (88'000 voters), eight cantonal referendums, 20,35% of ballots cast online.
2005-2006: development of a new secure channel (a new encryption layer) that comes on top of the SSL secure transaction between the voters’ PC and the voting server.
June 2006: The Geneva State parliament validates the report on internet voting.
November 2006: the Geneva technical School teachers and students elect their board online. It is the first online biding election in Geneva.
November 2006: The Geneva State parliament starts dealing with a legal amendment that would make internet voting a regular way of ballot casting.

(a) Strategies

 Describe how and when the initiative was implemented by answering these questions
 a.      What were the strategies used to implement the initiative? In no more than 500 words, provide a summary of the main objectives and strategies of the initiative, how they were established and by whom.
Turnout and citizen inclusion are two recurring issues in the contemporary European societies. While referendums can draw a large number of citizens to the polls, depending on how much do the question asked and the outcome of the ballot directly affect the individuals, turnout at elections tend to be low over the years, and can even be on a decreasing trend.

Inside this general pattern, one can observe further inequalities or "bias" affecting citizen participation:
- Older people are overrepresented among voters (people aged 60 years and more make out some 11-12% of the population, but 20 to 30% of voters);
- Younger people are underrepresented among voters (people aged 18- to 29 years of age make 10% of the population, but 7%-8% of the voters);
- Swiss citizen living abroad are disenfranchised: out of some 600'000 citizens living outside Switzerland, only 90'0000 have registered as voters. Postal service is so poor in many countries, including some European countries, that their ballots are received too late to be counted. The Swiss living abroad are de facto disenfranchised;
- Finally, disabled voters, especially the blind ones, cannot vote by themselves today. ICTs can provide them a way of reaching independence it this field too.

The problem is thus to boost participation, notably by engaging younger voters, in order not only to have a wider share of citizens coming to the ballot, but also a more representative share.

(b) Implementation

 b.      What were the key development and implementation steps and the chronology? No more than 500 words
The specific objectives are:

"public objectives"
- increase turnout among youg voters (mid- to long-term objective)
- increase turnout for elections (long-term objective)
- start with at least 10% of voters voting online (immediate)

"State internal objectives"
- Keep pace with the technical evolution in users' PCs (immediate). Our application is suited for a number of browser operating under windows, Mac OS or Linux. Yet, the software and OS are constantly changing: since we first offered eVoting in 2003, Firefox appeared and became the second most widely used browser, Mac released Mac OS 10 and OS 11 and Microsoft released Windows XP SP2. All these novelties imposed us to redesign the user's interface in order to provide the same functions, security and service with these new programs than with the ones existing at the date of the eVoting first launch.
- Provide a simple and easily accessible application (immediate). For details, see above "Policy context and strategy".
- Fully secure the application and conceive procedures able to prevent any wrongdoing (immediate). We have to implement such requirements as "one citizen one vote", fully anonymous vote, no identity theft, backups to prevent any problem with the electronic ballot box, implementing the Electoral commission's mandate and controls in this new context, configure all hardware and software in such a way that only the request consistent with a normal voting procedure receive an answer by the system. Any other request must be ignored by the system, in order to prevent any tampering and any leak of information that would give a hacker clues on our components or software.

The main problems are:
- Voters' anonymity: the system contains no names. Voters are registered under a number that changes for every ballot. Ballots are counted in a random order to avoid matching of entry and counting order.
- One man-one vote rule over three voting channels: there's only one voters' register. And the pin code on the voting card is hidden behind a rubber film. If we receive a card whose pin code is visible, we carefully check for any previous vote.
- PC-server connection: now, the PC server connection is secured using SSL 128, to make eVote broadly accessible. We are developing a secure channel that will use a second encryption layer, based on java.
- eBallot box security: all software and hardware elements are configured not to react to any command that is not compatible with a normal voting procedure. The system is only online during ballots. Ethical hacking has helped us optimise our protections. Ethical hacking always failed breaking into the system.

(c) Overcoming Obstacles

 c.      What were the main obstacles encountered? How were they overcome? No more than 500 words
In 2000, the Swiss Confederation launched an eVoting pilot. Geneva was chosen to develop an internet voting application together with Neuchâtel and Zurich. The electronic voting application is not meant to replace any of the existing voting channels (postal voting or polling station), but constitutes a third way of casting a ballot. Therefore, we didn’t reengineer the voting process, but we rather transferred the postal voting process on the internet.

Since any change to electoral rules and processes is a very touchy matter, we first dealt with referendums and initiative, leaving elections aside for a further stage. And, in order to build confidence, we gave started by implementing internet voting in small communities. We have then progressively increased the size of the electorate who could vote online.

Finally, after we successfully organised 9 ballots (8 referendums and an election), we submitted the parliament a legal amendment aimed at making internet voting a regular way of casting a ballot in Geneva.

Several studies have shown that in Switzerland the eServices provided by the public service were underused, because they were not known. Based on this finding and in line with our objective to increase turnout, we have devised a policy of promoting our eVoting application. We have called a public meeting in each municipality to present eVoting when it was introduce the first time.

The States installed PC in the town halls and main shopping malls, to allow citizens who do not have an internet access to vote online. There, trained staff answered questions and explained the interface's use. This presence proved useful especially for older voters and women.

We set up an helpdesk for the duration of the ballots (ballot period lasts from 2 to 3 weeks, according to the type of ballot: municipal, cantonal or federal). The helpdesk proved useful however the simplicity of the problems asked proved the userfriendliness of the application.

All voters receive by post a brochure explaining how to cast a ballot online and answering the most frequent questions on vote and online voting. The brochure is sent together with the voting card and the voting material, as was the case before eVoting.

The turnout online has reached beyond our expectations. While the very first figure (Anières, January 2003) is not representative of the short-term eVoting potential (people in Anières were very much aware of the sort of World Premiere they were), it is worth noting that the percentage of online ballots has never gone under 20%, twice as much as our expectations:

There has never been a problem, not any attack during ballots. The fact that the eVoting application is online only a few weeks a year and the fact that we have it evolving on a steady basis makes it difficult for hackers to accumulate sufficient knowledge on it to break it.

To comply with the Electoral commission's control, we have designed a test ballot box where members of the commission send unofficial votes during the official ballots. Before counting the eBallots, we count the content of this special box, whose content we know, since the Electoral commission writes down the content of the ballots it sends. We can thus ensure that the system functions properly and does not introduce a bias in reading the ballots. The eBallot box is locked by the Electoral commission. It can not be open without them.

(d) Use of Resources

 d.      What resources were used for the initiative and what were its key benefits? In no more than 500 words, specify what were the financial, technical and human resources’ costs associated with this initiative. Describe how resources were mobilized
The success of internet voting and the pace of change in the end users software has led us to reinforce the heart of the system and to free us form any dependency vis-à-vis a given service or hardware provider.

Although few has changed for the users of the eVoting system (we do not provide a link anymore to the voting web site, in order to avoid man-in-the-middle attacks, the invisble part of the system has considerably evolved over the last two years.

We also have the experience of a constantly evolving universe, namely the users' PC. The browsers and OS have changed over the years, think for example of the Microsoft Service Pack 2 for Windows XP, which obliged us to change some feature of the application in order not to be blocked by SP2's security features.

As of the November 2006 election, our voting platform has become technologically neutral. While we were depending on an HP server, running on an HP proprietary secure OS, we now operate on standard server running open source OS. This makes it easier for us to scale up our system to better accommodate the sheer number of votes we receive.

Yet, the major changes our application has provoked are changes in the citizens’ voting pattern. When internet voting is made available, between one voter out of five and one out of four votes online. These votes are cast late, compared to the postal ballot, which has been available since 1995. While postal ballots are evenly distributed over the 3 weeks voting period, half the electronic ballots are concentrated during the last week. This reinforces the effect of late campaigning by political parties and makes it more interesting for them to keep some resources for last minute ads.

Sustainability and Transferability

  Is the initiative sustainable and transferable?
Since our application is technically neutral, it can be replicated anywhere. Yet, in any country the organisation and counting of votes is strongly regulated by laws and our application should be adapted for each given context.

We are currently looking for an industrial partner who would have the commercial network and contacts to ensure the successful selling of our application, while engineers from the State of Geneva would supervise the necessary developments and adaptations.

The long-term sustainability of the application is guaranteed. We have already shown that it can evolve to follow the pace of the hardware and client software developments.

Lessons Learned

 What are the impact of your initiative and the lessons learned?
We have shown by our project that eVoting does not need be complicated nor expensive. Actually, we have done exactly the opposite of what is often presented as the must for eServices, the reengineering of processes. We have not reengineered the vote, we have transferred all stages of postal voting on the internet and we have applied the specific safety measure that the Net requires. The goal was to increase participation among the citizens. This comes at a cost, not so much financial, but in term of workload. We have to manage three voting channels whereas there were two not long ago, at, although we have developed an application to automatically consolidate the results, we must initialise, control and maintain a third voting channel.

The Swiss direct democracy and the federal structure of the State have proven to be facilitating factors. Swiss MPs are used to see their decision challenged by referendums. They do not fear a network-based voting system. The network is a good image of the very political system in Switzerland. It is however easy to see that in a more hierarchically organised country, such a voting system could generate anxieties, would it be only because it opens the door to a form of direct democracy: how could you refuse people the right to challenge the parliament’s decision when online polls is becoming ubiquitous and when the State has developed a secure and anonymous way of casting an online ballot?

If we look at the users behaviour, we see that the adoption rate has exceeded the project's team expectations. For 7 out of 8 official eEnabled ballots we have organised, the turnout on internet has been contained in the 20%-25% range. For the first ever eEnabled ballot in Switzerland (January 2003) it peaked at 44%.

The sociopolitical studies we have commissioned (www.geneve.ch/evoting/english/doc/rapports/rapport_26sept_english_final.pdf), have yield interesting data.

The findings are:
- The differences in internet turnout between municipalities are not statistically significant. The share of users of internet voting is consistent among the different municipalities, whether urban or rural.
- Online voting is performed home for more than 80% of the voters.
- Nine out of ten voters who voted at least once online remain trustful to this voting system.
- Internet voting is the favorite channel for citizens who do not otherwise vote.
- While the younger voters aged 18 to 40 years represent 14% of the voters casting a ballot, they represent 23% of the internet voting users. The use of eVoting peaks for the 29-39 years age group and decline linearly after that along the age line.
- At any age, there's a 0.5 to 1 percentage point difference between men and women for eVoting use. Men use it more. But the general pattern of eVoting use over the age is similar for both genders.
- There is no political tendency among the eVoting users, they have always voted with the majority.
- Although we have offered eVoting only in the framework of referendums, citizens do not see any restriction to its use and would be welcoming at 95% its use in elections.
- The multivariate analysis of the answers gathered shows that the subjective elements in the voters’ relation to internet is the predominant drivers explaining the use of internet voting.

Voting online depends upon the attitude toward novelty and technical matters. The socio-demographic factors are not relevant. The use of internet voting could almost be considered the indicator of a lifestyle in which information technologies play a pivotal part. On the opposite, it is in no way an indicator of social rank.

Contact Information

Institution Name:   State of Geneva
Institution Type:   Government Agency  
Contact Person:   Michel CHEVALLIER
Title:   Project head of communication  
Telephone/ Fax:   +41223272049
Institution's / Project's Website:   +41223270411
E-mail:   michel.chevallier@etat.ge.ch  
Address:   Chancellerie d'Etat
Postal Code:   1211
City:   Geneva 3
State/Province:  
Country:   Switzerland

          Go Back
Print friendly Page