eDemocracyBlog.com

In a previous post I summarised the arguments against eVoting which a series of speakers in a recent Hansard Society debate had put forward.

In this follow-up, I intend to provide some balance by responding to their criticisms and putting forward some other points in favour of eVoting.

Just as some background, I wouldn't call myself an evangelist for eVoting (I've previously called its attempted implementation by the government a "failed policy") but I do think that it has significant potential, indeed may become a necessity at a point in the not-too-distant future.

Security of the process

To begin with Jason Kitcat's criticisms of eVoting, his fundamental objection seemed to be that eVoting could not be secure, verifiable and anonymous - and without this there could be no public trust in this vital element of the democratic process.

On the issue of security, it is hard not to get the feeling that the argument is being made that eVoting should be unbreachably secure in both theory and practice. This is, of course, an impossible goal.

More importantly, it sets a far higher bar for eVoting than it does for voting by postal ballot or even in person at the polling station, where no identification is required and where fraud is not just possible but is uncovered with alarming regularity.

To mention just one situation I've had experience of, the lag between people moving houses and the updating of the electoral register means multiple out-dated polling cards can be delivered to one property which the new owner could easily use to cast repeat votes.

Related to the security point was Kitcat's comment that delivering PINs to anyone wanting to vote electronically would create a further threat to security. Yet banks generally seem able to handle the process.

Kitcat also said eVoting could enable "ballot stuffing on a massive scale" which the need to photocopy and complete postal ballots makes more difficult. But for a would-be fraudster it should be far harder to get hold of a large number of PINs than it is to get hold of a blank ballot paper and photocopy it.

And again this overlooks the point that the use of PINs when casting a vote would actually be a huge step forward in security given the total lack of identification currently required.

As for the possibility of somehow hacking into the system and creating false voting records, it may be possible that details of voters can be held separately from the details of votes, and then matched again during the counting process with each voter told how their vote was registered so that they can report if it was changed without their permission. Similarly, any vote which did not have a corresponding voter could also be investigated as suspicious. These are just thoughts off the top of my head, and I'm sure far more secure systems can be imagined by those who are experts in this field.

Kitcat's other main points related to the tension between being anonymous and yet verifiable.

I think Andy Williamson made a telling point that wasn't rebutted when he noted that banks manage to verify cash machine transactions without ever knowing the cardholder's PIN.

It is also worth pointing out that the current paper-based balloting system is not anonymous either, so again this would seem to be a case of making demands of eVoting which are not equally applied to the existing system.

If the same requirements were applied equally, the real issue would be over who had access to the key which matches the voter's ID number with the ID number of their vote. Looked at this way, it hardly seems like a difficult technical issue at all.

Another question is whether any system can be both anonymous and verifiable anyway? If it is genuinely anonymous then who is to tell whether any ballot was cast by a legitimate voter rather than, say, dumped into the ballot box by a corrupt council employee before it is sealed?

None of the above, thanks

There is also another issue with verifiability, which relates to Tom Harris's comments on being able to go to an election night count and watch the ballots stack up and check any that are called into question because the intention is unclear.

With eVoting, there is no opportunity to be unclear in your voting intention. The vote is for a candidate or not. There are no scribbled messages on the ballot paper, no smiley faces drawn next to a candidate instead of an X.

So at least part of the issue with verifiability is therefore removed from the equation, or at least is moved from the counting phase to the point at which the vote is actually cast.

However, I do think this also highlights what for me is the biggest case against eVoting, and an argument which no one at the event made.

If I wanted to abstain in a paper-based election, I could spoil my ballot paper or hand it in blank. With an eVoting mechanism I would be required to select a candidate

To me, this means that eVoting has to be accompanied by a 'none of the above' option so that it does not de facto remove an existing democratic freedom.

This option terrifies most players in the political system; something which is pretty telling about their confidence in their abilities and arguments.

Personally, though, I doubt that such an option would attract as many votes as people seem to fear. Most of those who currently vote seem likely to continue voting for parties as they have in the past, while more of the support for a 'none of the above' option would quite possibly come from those who have not voted in the past.

Verifiability

However, before we leave the issue of verifiability, it is also important to note that this cannot be an objection in principle to eVoting.

This brings us on to points made by both Kitcat and Jenny Watson, that the software that runs eVoting is currently a metaphorical 'black box', which voters therefore cannot have confidence in.

Yet a proactive Electoral Commission, rather than the disinterested one we seem to have, might have pulled its finger out and set up a panel of experts to which firms could submit their software for approval for use in UK elections. They would also have established processes for ensuring that the software which is approved is the same that is used on polling day, perhaps with snapshots of the code in use being taken at random points in the day and software comparison tools being used to identify any places in the code which do not match that which was approved. This need not be a time-consuming process.

Of course, the Electoral Commission has done nothing of the sort. As I've noted in a previous post, the Commission seems to have had little constructive discussion with the government about eVoting since 2007.

And Williamson made a useful point about "the social appropriation of technology" and its role in public confidence in a system. People do not need to know how something works, or even be entirely confident in its security and privacy policies, in order to use it in their millions. I could perhaps mention Facebook at this point.

Barriers to voting

Another advantage of eVoting addresses the concerns expressed by both Watson and Harris about those who do not know how to vote.

Watson suggested that eVoting would be of no help to such people, ignoring the potential for on screen information providing instructions for voting in multiple languages.

And Harris then appeared, in my view, to contradict himself by moving on to argue that there should be a mystique about voting which eVoting would undermine.

I would suggest that it just this mystique, to whatever extent it exists, which creates a barrier to entry for voters who are unsure about the process, and therefore helps to perpetuate the exact problem of non-voters which Harris says he is concerned about.

The ability to browse a website like any other, click buttons and make a choice in the surroundings of your own home rather than with a queue of people watching in an unfamiliar location is far less likely to deter nervous would-be voters.

It costs too much

To briefly address the issue of cost, it certainly is the case that eVoting pilots have had a higher cost-per-vote than traditional mechanisms.

But this seems largely due to the fact that they have been piloted in local elections with relatively low levels of turnout. Given that eVoting should scale relatively easily from a few thousand voters to tens of thousands, the cost-per-vote in a general election could be expected to be significantly lower.

In addition, those councils which trialed eVoting would have been making individual purchases of the software and hardware. If eVoting was rolled out nationally, then there would be much more scope for bulk purchasing, driving the costs lower still by making use of economies of scale for suppliers and councils alike.

Having criticised the cost of eVoting, Kitcat went on to suggest the idea of weekend voting or a public holiday on election day. The cost of weekend voting was estimated during a government consultation at an extra £58m (see page 40 here), due to extra overtime payments for staff, etc.

Presumably asking those people to work on what might be a national holiday for the rest of the country would cost at least as much, while the cost of a public holiday in terms of lost economic productivity has been estimated at £6bn by the CBI - which makes the cost of an eVoting system seem almost trifling in comparison.

The same consultation found, by the way, opposition to weekend voting but support for eVoting - 53 per cent favoured retaining polling day on a weekday but 62 per cent supported greater access to eVoting or further investigation of the issue.

(Update: I received an email pointing out that the eVoting pilots also included one-off development costs - another reason why citing the cost-per-vote of the pilot schemes is misleading)

Keeping track of votes

Another criticism related to the difficulty of running a multi-channel election (one where voting can be done at a polling station, over the internet, by postal ballot, by SMS, etc).

The backbone of such a system would need to be a nationally accessible electoral register which can be updated when a vote is cast by any means.

The Co-ordinated Online Register of Electors (CORE) project to standardise local electronic electoral registers and make them fully interoperable was announced in 2004, but has been making slow progress since.

This system remains a pre-requisite for any multi-channel elections, yet for such an important piece of democratic infrastructure it attracts very little attention.

But were such infrastructure established, it would mean staff in polling stations ticking a box on a PDA rather than writing it down on a piece of paper.

Then, whichever means a voter uses to cast their ballot, the national register would be instantly updated to prevent them voting again.

It would also make it easy for people to check if someone else has used their vote (thereby adding security), while those who did seek to vote multiple times could either be flagged up for investigation or perhaps given the opportunity to overwrite their previous vote.

What's the strategy?

Watson, meanwhile, said existing fragmentation should be addressed before eVoting is considered, and that a strategy for electoral modernisation should be put in place.

Apart from the fact that the Commission has taken little proactive interest in developing such a strategy despite having called for one, there is no reason why eVoting should not constitute a major part of such a medium-term plan.

And Watson also suggested that eVoting has risks - although her organisation's reports have not, to my knowledge, flagged up any cases of abuse or fraud during eVoting trials. They have, though, identified any rush to implement eVoting as a problem. But this is a simple lesson to learn from the pilots, and just goes to make the point that eVoting should be part of an integrated strategy for electoral modernisation.

The turnout diversion

Another point made during the debate was that eVoting will not increase turnout.

I think this is something of a red herring. While it might be beneficial and could, as noted above, lower the barrier to entry for some voters, it is not the main point for introducing an eVoting system.

I think the biggest argument is convenience; the opposite of Harris's point that voting should be "inconvenient" in order to make it special.

Having millions of people detour out of their way to their local polling station and having to stand in line is a waste of time on a huge scale. This time could be spent more productively doing other things if people could cast their ballot with a few clicks in a time and at a place that is convenient for them.

So I would not disagree with Kitcat that the key to higher turnout lies with our political culture and the voting system, not the electoral administration system.

But this is not an argument against eVoting.

Make the connection

In other remarks, Watson admitted that the existing system is "close to breaking point" but seemed unable to think of any way in which technology could help despite the fact that it scales much more effectively than a paper-based system.

Watson also raised the important point of access to polling stations for people with disabilities, yet again seemed unable to think of any way in which eVoting might be of benefit.

She added that there should be no eVoting till an electoral modernisation strategy and individual registration are in place. But the coalition government seems intent on speeding up the introduction of individual registration, making the development of strategy a matter of political will.

And Harris mentioned his support for the 'save general election night' campaign without noting how eVoting could have helped by reducing the number of votes which need to be counted manually.

A few more positives

There are also some other arguments in favour of eVoting which weren't brought up during the debate.

One is the ability to provide information, perhaps in a standardised and comparable way, about the political parties, the candidates and their policies at the point of voting, or at least just a click away.

Given the anachronistic bar on information in polling stations and the tradition of not campaigning on election day itself, it seems to me that eVoting offers a chance to do things in a more useful way.

Then there is the potential to improve turnout through email reminders to vote on polling day itself and during the run-up to it.

And there could be improved feedback to voters, even if just a thank-you email from the returning officer. But that could be taken further to perhaps include information about how many other people have voted online in their constituency and news about the result to provide positive reinforcement for the process of voting.

Think big

Finally, there is a much broader picture to consider.

eVoting is just a small part of the change that technology is bringing to the structure of the state.

It is part of a wider move to more direct engagement through both eDemocracy and eGovernment, in which voters become more embedded in the state and more engaged with the services that are provided to them, controlling these through choices that are increasingly likely to be made online.

So does this mean eVoting can be done simply and effectively? Nope, there will be arguments over cost and most likely technological glitches along the way.

But should it go ahead with care and determination? Yes, it has to.

eVoting, Hansard Society, participation