Not sure if I have linked here to my LSE book review about electronic voting, so here it is.

Thus:

The heart of the book is the authors’ emphasis on sensible risk analysis. Above all, they punch on the nose the odious “precautionary principle” – the superficially appealing but in fact spurious idea that nothing new should be accepted unless it can be proved to do no harm.

The authors do a fine job in insisting that it makes no sense to apply that argument to proposed reforms while not applying it to the current state of affairs too. In other words, those who object to (say) electronic voting on the grounds that it increases the risk of abuse need to compare those supposed risks with the actual risks of abuse in the current system. Advantages of electronic voting likewise need to be given fair wind: research shows (they say) that electronic voting systems reduce ambiguous or “spoilt “votes, thereby increasing effective participation rates and (arguably) the legitimacy of the result.

The book also looks at the cost of elections. Classic paper-based voting arrangements are labour-intensive and expensive. Electronic voting cuts costs significantly once the system is set up and reduces the risk of human error. What price democracy?

The book concludes with recommendations for looking at voting system reform in the USA, and hopes for “a thoughtful, fact-based, rational discussion …of election administration and voting technology”. Given the very complexity of the policy and philosophical issues this book itself describes, this is a forlorn hope.

My main concern is that the book does not do justice to what I would call the “law of large numbers”. The very fact that old-fashioned voting systems are so labour-intensive means that it is almost impossible to cheat on a scale that matters without that cheating being obvious.

The supreme risk of electronic voting is that those who program the system or malevolent hackers can infiltrate it secretly to influence the outcome. Such abuse need not involve crudely forcing a victory for one party or the other: it might be enough to “nudge” the results so that over time advantages for one party accumulate.

The book quotes the US National Academies of Science: “Democracies derive their legitimacy from elections that the people collectively can trust”. Are voters ever likely to trust an e-voting procedure which might be systemically distorted without anyone knowing?

Haha prescient words, what with them thar pesky Russians hacking the elections and all. Or not, as the case may be.

NB:  There is an important distinction to understand, folks, between electronic voting and electronic counting.

An electronic voting system means voters pressing a button or touch a screen to register their vote automatically. This is attractive to local authorities wanting to save money on running elections. But it is fraught with operational and conceptual difficulties – how to make the process secret, transparent and safe from manipulation either by external hackers or by malevolent insiders programming/running the electronic system?

An electronic counting system of the sort I saw in Russia in Nizhny Novgorod is simpler. The citizen votes as usual on a paper ballot then inserts the ballot paper into the ballot box via an electronic ‘reader’. The votes are counted automatically, but the paper votes are there as a back-up in case the result is contested.

The great advantage of the paper-based voting system is that it is clear, simple and in principle reliable. Ordinary people can see what is happening and understand it. Mistakes in counting are unlikely to make a difference. But it is amazingly labour intensive and therefore expensive.

Electronic systems for voting are accurate and fast but much less transparent. Plus an electrical blip of some sort might change the result without anyone knowing.

As the OSCE report on the latest Russian elections sensibly noted:

Two types of new voting technologies were used during these elections. The first was a ballot scanning system called “KOIB”, the second was an electronic voting system “KEG”, based on touch-screen machines. Both systems were used on a moderate scale.

PEC members in most of the regions observed received training on the use of new voting technologies. The practice of publicly testing both systems on or immediately prior to election day can potentially help build trust in e-enabled voting. However, the absence of provisions for random mandatory manual recounts of the processed ballots is of concern. In addition, transparency in the design and functioning of both systems is insufficient as both types of technologies are based on proprietary software not open to public scrutiny.

Touch screen voting machines were equipped with an embedded printer giving voters the possibility to verify their vote whilst voting. Although this enhanced the verifiability of the process, the fact that votes were printed consecutively on one strip of paper created the potential for the violation of the secrecy of the vote.

No special conclusions. But be very reluctant to move to e-voting if it’s ever offered. The transparency and security issues are completely different and not easy for mere humans to follow.