(with thanks to Liam Bell for feedback on the first draft.)
On Tuesday 12 June, I was at a workshop organised by the Scottish Government (SG) to consider how lessons from e-voting, as used in participatory budgeting (PB), can inform SG’s investigations of how e-voting might be used in Scottish elections. A key theme of the discussion was how a system that would initially be used for PB voting could be set up so that it would ‘naturally’ evolve to be used for other voting ‘use-cases’.
It’s not my aim to describe what others said at the meeting: a scribe took detailed notes and I hope that a transcript will be published soon. However, I hope this post will describe my thinking, and how it’s evolved a bit since I last wrote about e-voting.
What are the problems?
I still think there is little actual dysfunctionality in the paper-based systems that are used in UK elections. There are high logistical and environmental costs, and the systems do not suit people who have mobility issues, visual impairments or whose lives otherwise prevent them from going to polling places. However, there are, for example, proxy voting, postal voting and statutory aids for people with visual impairments. Paper voting doesn’t need electricity. If a venue is unavailable, the back seat of a car can become a polling place.
As far as I can see, the issues that e-voting might tackle are (1) increasing turnout, (2) facilitating reliable and rapid results (including fewer spoilt ballots), (3) reducing the costs of paper ballots: the logistics and large financial costs of printing and distributing ballot papers, then safeguarding them, then counting and storing them. However, I understand that turnout is mostly a factor of people believing voting makes a difference, and people being used to voting. I believe (admittedly without research) that the UK’s systems are reliable, and that counting of paper votes can be verified by the vast majority of the UK’s population. (Remember that justice must be seen to be done.) I don’t mind all-night election parties, and I think it’s good that many people are involved in running our democracy. So the only issues that remain are the financial, environmental and logistical costs of paper voting.
What can be learnt from PB?
Most of the following is based on my experience of staffing a ballot-box at Leith Chooses (LC) earlier this year, and a very informative conversation with an Edinburgh Council (official who was very involved with administering that PB event’s voting systems.
I think it’s important to first acknowledge that PB is different to elections. Firstly, in the case of LC, the electorate was everyone aged 6 or more who lived, worked, studied or volunteered in Leith. So the electoral roll wasn’t a guide to who could vote. For example, homeless people who happened to stay in Leith the night before voting day would presumably have been eligible, even if most of the time they stay elsewhere. I was eligible to vote because I work for two Leith community councils, even though I live elsewhere. As far as I know, elections use either residence or membership eligibility criteria, not a mixture.
Secondly, PB (in my limited experience) doesn’t apply the rules too strictly. I couldn’t always be 100% sure that each voter put only 4 paper votes in the ballot box, or that they didn’t vote more than once for any project. Occasionally people didn’t want to vote for 4 projects. (The e-voting system insisted on 4 votes per ‘pot’.) I believe that most people would want elections to be much more rigorous.
Thirdly, none of the above mattered too much: key LC aims included engagement and education. That is, the organisers wanted people to be involved with making choices about their community, and to learn what can be done. A third key aim was inclusion, and hence minimising barriers to participation. So people could vote online via computers in local venues, thus eliminating any need for them to have their own IT kit, and there was a 2-week e-voting period after the in-person voting day. As far as I could see, all of these aims were successfully achieved. So I think PB is a great way of showing people that their votes count, and getting people used to taking part in direct democracy. However, elections are about indirect (representative) democracy.
Fourthly, PB processes can involve idea-generation and refinement steps before voting takes place. As I understand the UK system, parties usually set out their ideas in manifestos, and voters can only vote for the most appealing ‘package’. It can be hard to get single-issue measures through, even if the vast majority of representatives support a measure.
Fifthly, PB can be about achieving certain themes, and about deciding how predetermined amounts of money should be spent. In the case of LC, people voted for various projects, and the most popular four in each ‘pot’ won funding. In other systems, each voter can ‘spend’ money up to the predetermined limit. Then the system outputs a final result, presumably based on the most popular choices. In contrast, elections are about individuals’ beliefs about who would be the most advantageous (or least disadvantageous) local representative and/or governing party. Such decisions are about more than money.
So the lessons from PB are, as far as I can see, that voting makes positive differences, that it generates community engagement, that it can be run by amateurs (in the positive sense of that word) with support from local authority staff and democracy workers, and that e-voting and paper voting can be successfully combined. I’m not sure LC significantly reduced environmental or logistical costs. (It relied on volunteers to reduce financial costs.) The paper voting system used paper, and was in addition to the e-voting system. However, the single paper ‘polling place’ eliminated the need to safeguard and transport multiple ballot boxes, although there were voter support events. Their costs might be seen as analogous to those of multiple polling places.
New e-voting systems?
One of my concerns about e-voting is that voting must be separate from proving eligibility. I’ve softened my position here: there are systems that provide this separation, and that allow voters to check that their votes are recorded accurately, without others being able to know how they vote.
Current proposals for e-voting systems appear to be based on implementations of blockchain. There’s an example here. As I understand it, blockchain makes it hard to very falsify transactions, because records are distributed so any attack would need to hit all records simultaneously. That is, there is no central ‘honeypot’ server to attack. Also each record is based on the previous records in the chain, so any attempt to falsify records would need to dig far back in time, on every device that stores records. Because blockchain systems can be open-source, anyone can (in theory) check and improve the software. (In practice, I suspect only a very limited number have the necessary expertise. Still, this is better than systems being opaque to everyone but ‘establishment high priests’ of e-democracy.) Because everything is encrypted (or at least hashed (Wikipedia)), privacy issues should be minimised. The verifiability/anonymity issue may be solved by using ‘zero-knowledge proofs’ (Wikipedia). Because voting systems based on blockchain would be purely electronic, there should be much smaller environmental and logistical costs. Elimination of physical polling places should minimise policing costs. Finally, individual blockchain-based identities could reduce eligibility and ‘personation’ issues (even though the latter are currently rare). So will blockchain solve all the problems?
I don’t believe so. I believe it can provide efficient technical solutions, but there are social factors to consider too.
- Low turnout is not solely due to where we currently vote. As Leah Lockhart has pointed out, if someone feels disaffected, disenfranchised and generally thinks politicians are clowns then no amount of technology is going to help that.
- Not everyone has IT kit, so we will still need public polling places, even if they contain computers rather than ballot boxes.
- Not everyone trusts IT-based solutions, nor can everyone use IT-based solutions, so I think paper-voting (including postal and proxy voting) will continue to be needed in addition to e-voting. However, I presume that as soon as someone has e-voted, proved their identity at a paper polling-place or their postal vote has been recorded, it should be possible to amend records everywhere instantly to prevent them from voting again.
- Some delays in updating eligibility are unavoidable. As I understand it, blockchain e-voting would be based on everyone having a smart ‘contract’ which can be annotated to prove whether the owner fulfils eligibility criteria. For example, my contract would know my age and that I live in Edinburgh’s West End, so that I am eligible to vote in West End Community Council elections and any PB events they run, as well as in elections for my council ward, MP and MSPs. My contract might even know that I have UK/Australian dual nationality, potentially enabling me to vote in Australian elections. In addition, my contract might know that I have diabetes and so do not have to pay VAT on my blood-sugar sensors. My contract might know I work at Edinburgh Napier University, and maybe how much I am paid – thus possibly helping to automate tax returns. So far, so good, but…
- There is no conceptual barrier to any state issuing everyone born in its jurisdiction with individual blockchain contracts, ready to record data that proves eligibility to vote, receive state benefits etc in that jurisdiction. The UK already does that with NHS numbers, NI numbers, Unique Taxpayer References and the like. However, what about people who were not born (or recorded as born) in the jurisdiction? It should be possible to issue a contract to everyone entering the jurisdiction regularly, even if they do not claim citizenship. (For example, my father retained his Australian citizenship despite living in the UK for decades, but became eligible to vote in UK elections, had an NI number and paid tax normally.) However, the delays in dealing with the Windrush people and the mess of Universal Credit do not give me faith in the UK’s current abilities.
- As far as I know, prisoners are not allowed to vote, so the bureaucracy of sentencing must include amending their contracts appropriately, and the systems must not allow prisoners to amend their contracts to regain the right to vote until they have been released. More bureaucracy: more potential delays.
- I work for three Edinburgh Community Councils, two of which pay me by cheque. I doubt they would be able to update a blockchain contract to show how much they pay me, or that I work in their areas and so am eligible to vote in PB events they run. In any case, contracts would need to deal with people who work in several places.
- It’s not my role to update my contract with my medical details. Firstly, I’m not qualified. (I’m a doctor of chemistry, not a medical doctor.) Secondly, what about people who falsely state that they have illnesses that they don’t have, no matter whether this is due to mistakes on their part or mendacity (e.g. trying to get extra paid sick leave)? So amending medical records should be done by relevant medical professionals. Given my experience of how slow the NHS can be, there will be delays and inaccuracies from time to time. Technological solutions will speed up access to data, aid epidemiology and other good things, but won’t be 100% accurate all of the time.
- I travel from time to time. I want more than just a blockchain contract to prove my eligibilities. I want something physical that is easy to read and act on, avoiding any technical incompatibilities. (I wear a MedicAlert chain and wristbands saying I have diabetes, and carry a letter from my doctor saying I should be allowed to carry needles on aeroplanes.)
None of the above should stop appropriate modernisation of our voting systems. I would love it if e-voting was as easy as internet banking. I recognise that some of my concerns are edge-cases, but any voting system must take full account of the vagaries of human life. So I would love to contribute to research and testing that truly kicks the tyres of any proposed system.