A democratic society can only function when the actors in that democracy, the citizen/voters, have trust in the mechanisms of that democracy.
First, let me say that I fully endorse the use of online technologies to improve access to information for voters. That said, as I've written before, I think Online Voting is a fundamentally flawed idea - my former colleague at Catalyst IT, Stephen Judd, has also written a very thoughtful piece on online voting which I thoroughly recommend. Nigel McNie provides another thorough and very readable assessment of Online Voting in his submission to an NZ parliamentary select committee in June 2015.
We are currently seeing a series of efforts, world wide and here in NZ, at both a local and national level, to introduce "online voting". The rationale for doing so is based on these assumptions:
- it will help to arrest the falling voter turn out, particularly among the younger demographic,
- it will will increase access to voting to people who find traditional voting practices challenging due to disabilities, and
- it will reduce the cost of the democratic process at both the local and national levels.
I think that these assumptions are deeply flawed and based on naïve extrapolation by people who are insufficiently informed to make a decision. Moreover, the idea that online voting can be undertaken in a way which does not dissolve all trust in the mechanisms of democracy (thereby possibly causing a full collapse of the democratic institution) is pure fantasy.
Keep in mind, my purpose in this post is not to identify or enumerate the technical barriers to trustworthy online voting which are too many to count. I'm primarily looking at the conceptual problems with it here.
Addressing the Assumptions
First, the idea that online voting will somehow increase voter participation is largely unsupported by evidence. This is because the problem of disaffected voters is not about convenience of voting. We adopted postal voting for local body elections in NZ based on precisely the same logic. To my knowledge voter turn out has not increased substantially. I suspect we'll find that far more uninformed votes are being submitted, however. Therein lies the problem.
There are many in decision making roles within government who think that providing "convenience" for younger voters who're used to doing "secure things" like online banking or even completing their NZ census forms online, will overcome their apathy. To them, this just makes sense - after all, if it's safe to doing banking or fill in the census online, how could it not be safe to vote online. That simply demonstrates the lack of fundamental understanding of voting: we don't want more people voting... we want more engaged and informed people voting. Quantity and quality. We want eligible voters to understand and value their democratic right, which comes with responsibility to be informed.
Also, if we focus on online voting to attract young disenfranchised voters, will we address that apathy? More importantly, will we affect the confidence in the democratic process of those traditionally more engaged voters, namely our older voters, many of whom will never willingly use online tools for voting (or much else)? I suspect we could be undermining one group in a vain and unwarranted hope to enfranchise another.
Second, will online voting allow more people who are disenfranchised by the normal physical voting process to vote? I think it is very much open to debate. The success of online solutions for users with disabilities is highly contentious: while there have been some successes, there is often a fundamental disconnect between the needs of disabled users and those designing solutions on their behalf. In the case of voting, there is definitely an issue with the traditional paper ballot and, for example, voters with vision disabilities. To address that, a rather clever telephone voting solution has been developed which has both received a very positive response from the blind community, is seen as far more credible by security experts I've spoken to, and sidesteps many of the usability issues inherent in online voting.
Last, how can councils or the NZ government possibly justify their hope that online voting will reduce the cost of elections or referenda? Introducing a new, unproven method for voting, while having to continue offering the existing methods in perpetuity due to the incomplete adoption of computing technologies by the voting population can only add to the cost. A cost reduction is dependent on 100% adoption of online tools by eligible voters, which is unrealistic, even if we're prepared to wait for a generation. It is a hopeless pipe dream to think online voting will somehow reduce the cost of voting in the short or medium term.
Low Stakes, High Stakes?
I perceive a pervasive impression within governments, both local and national, that online voting should be tested in "low stakes" contexts, like local body elections, where the incentives to engage in fraud are relatively low - based on the idea that nobody really cares that much about the outcome, and that there's not much money or influence at stake. I see this as a dangerous mentality. It suggests that, should the approach be successful in low stakes situations, it can then be applied to increasingly higher stakes situations. That creates an implicit contradiction: showing that the risk is acceptable low in a low-stake elections specifically does not provide any useful insight into the risk for high-stake elections precisely because the stakes are different. So it's a basically flawed extrapolation.
Democracy Cannot Survive No-Confidence
Today we have postal voting for local (low-stakes) elections. Postal voting introduces a number of risks to the voting process that allow for both small and large-scale fraud. From stand-over tactics by members of a household where one family member coerces other family members to vote a certain way, to large scale fraud at the points of vote collection, e.g. post boxes and postal depots. Similarly, most countries offer absentee votes via postal ballot, which use quaint security measures like written signatures (unauthenticated and trivially spoofed) to determine authenticity. These risks were deemed acceptable due to the potential cost reductions.
Many people cite the move to postal voting, with greater risk of fraud (it's a very real risk), as justification for pursuing online voting. I personally think we should roll back postal voting.
While banks can afford to tolerate a certain amount of fraud in their commercial dealings, as can, it would seem, those who use and trust the services of banks, I think the idea of extrapolating our confidence in online banking to online voting is based on a flawed premise: that online banking and online voting are somehow equivalent or even related problems. The key difference: when dealing with a bank, both you and your bank need to know the identity of each party (so called "man-in-the-middle" attacks and common "phishing schemes" demonstrate that most people trust that pretty blindly, but that's for a different blog post). A vote in a democratic election, however, has a different condition: that the identity of the voter cannot easily be associated with their vote. This allows voters to vote according to their conscience without fear of reprisal. If votes were not anonymous, a malevolent element could first execute fraud in an election, and then, having gained power, identify all those who voted against them and engage in reprisals. This is not hypothetical, it has happened countless times elsewhere in the world.
I think democracy is only as good as the confidence its voters have in the mechanisms of democracy. Just like a good reputation - which takes many years to build, and minutes to ruin - confidence in the democratic process can be eroded in the blink of an eye, and would take years, perhaps generations, to rebuild. We cannot afford to have an "acceptable risk" approach to our democratic processes - yes there is always risk, however, we must work to ensure that the risk is always limited in scale. Moving elections online makes limiting liability inherently impossible, because complete exposure is a single security vulnerability away (and we software developers know that there are always vulnerabilities waiting to be found).
I hope I have provided a considered case against online voting. Unfortunately, some will consider the negativity of my conclusion more significant and damning than the substance of my arguments. To counter that, I would like to offer what I see as an alternative approach to addressing the greater goals which some think online voting could help them achieve.
As someone who is focused on the creative capacity of self-organising communities of people, I have tremendous confidence in the wisdom of communities (which is very different from the wisdom of "crowds"). Communities share ownership of a commons on which everyone depends. I think the answer to voter engagement and motivating voters to inform themselves lies in peer pressure from communities.
Instead of throwing millions of dollars at developing online voting systems (and it will take many tens of millions to do anything even remotely credible) I would argue that a much more cost effective way of improving both the quantity and quality of votes cast in both local and national elections would be for the government to fund hui around each local polling station and ensuring that all voters are able to take part (perhaps requiring changes to labour laws). These hui should be accessible (from a disabled access point of view) and child friendly. There should be local community members present, especially elders with sufficient mana to focus those gathered on the topic of voting.
I think it would be possible for these groups to cover the issue and/or candidates, and provide some direction to the available resources (some of which might be online) provided to inform voters. Then groups could discuss these themselves and educate one another in a congenial environment. I'd love the community voting tradition to become a celebrated part of NZ culture, where people get together with their local communities, across all demographics, to work out how their vote can affect them and those around them.