July 08, 2020

The Digital State, Part 2

A Roman coin from around 60 BC depicting a voter putting a ballot into a container.

by Costa Vayenas

This is the second in the series on the paradigm shift to a digital state. In Part 1, published here, I pointed out that the state is becoming increasingly digital and that no pillar of government is going to remain untouched.

In this article, I focus on one area of citizen-government interaction that is being digitized, namely the technology of elections.

The 2020 accelerator

The coronavirus has been a massive accelerator of all things digital involving the state. In the area of public health, 28 governments were able to rapidly introduce COVID-19 apps. [1] Tracking the virus at sewage plants has also given governments a one-week city by city heads-up on where outbreaks will be confirmed next. [2]

Beyond digital public health initiatives, dozens of governments around the world either postponed elections in 2020 and/or changed the election process to “remote” (i.e. permitting voting from home). [3]

New kinds of ballots

Over thousands of years, the ballot has been many different things: a show of hands, the spoken word, light and dark pebbles, engravings on bits of clay, entries on small wax tablets, olive leaves, pieces of bronze, papyrus, parchment, vellum, marbles, the lifting of swords, and paper. [4]

In our age, a number of state entities have experimented with e-voting. The US was one of the first countries to experiment with voting by email, at first just for the military abroad. At the time, it didn’t seem to matter that the votes were not really secret – someone still had to print them out at the other end. More recently, West Virginia became the first US State to permit the military to vote via mobile phones using blockchain. [5]

There is a lot of trial and error involved. I remember how impressed I was by the Norwegians, who explained why one of their electronic elections nearly turned into a disaster as a result of an error in the encryption code. [6] If this type of thing could nearly happen in Norway, a wealthy country with a technocratic administration, then we might well ask what could all go wrong if e-voting were launched in less well-prepared countries that have fewer resources and lower standards of transparency and governance?

Voting technology is not neutral – it can shift the outcome of elections

That technology influences how the people instruct their representatives is as old as recorded history. My choice of the Roman coin above illustrates one of the ways that voting took place two thousand years ago.

There was a time when elections were not conducted by secret ballot, neither in ancient Greece or Rome, nor in the United Kingdom or in the United States. In the UK, the secret ballot was only introduced as late as 1872 and later in the US. Before the introduction of secret ballots, one method of voting was for voters to verbally tell an appointed election official at the polling station who their vote was for. The official then had to record that and tally the results.

When society decided it was time to switch to a secret ballot, a different voting technology was required. The new method delivered its objective – a secret ballot – but it also changed election outcomes. In one study, voter turnout in the US declined by around seven percentage points after the introduction of the secret ballot. [7] A possible explanation for this decline is that some voters had only gone to the polls as a result of pressure from their landlord or their employer, or as a result of gifts from a particular candidate. The introduction of the secret ballot also discriminated against those who could not read or write, thus keeping them away from the polls.

Another effect of the introduction of the secret ballot, noted after Rome introduced it in 139 BC, was that gifts from candidates were now redistributed to specific tribes or regions. Sometimes the gifts were provided after the regional result became available, thereby confirming the quid pro quo. [8]

The Swiss experience: Postal voting changes the numbers

What about the impact of postal voting on election outcomes? In Switzerland, every eligible citizen, regardless of where they are registered in the world, automatically receives a ballot in the post, complete with a self-addressed return envelope, whether they want it or not. A long-term study that looked at the impact on voter turnout following the introduction of postal voting in Switzerland found that it resulted in an increased voter turnout of around four percentage points. [9] At first glance, that is a surprisingly low number considering that the ballot arrives at your home without you requesting it, and includes a return envelope addressed to the ballot-counting authority. How much easier can it be than that? But that is also a reminder of how uninterested the average person is in politics – they don’t even want to open the envelope. Getting people to vote is hard work. On the other hand, a permanent increase of four percentage points in voter turnout shows that more convenience does make a difference. And given how narrow some election wins are, four percentage points can make a huge difference in some places. Seen from that perspective, a permanent increase of four percentage points is a big deal in politics.

The type of ballot matters, therefore

So, if you change the ballot or tweak the election process (who can vote, when, where and how?), you change the voter turnout and that may change the election outcome.

As the state goes digital in a big way, voting from one’s mobile phone from home in one’s pyjamas will inevitably have an impact in certain jurisdictions if the voting choices are considered relevant. Evidence from different jurisdictions is that the introduction of e-voting resulted in different rates of usage and different trends. It was considered a flop when introduced in a limited way in France for French residents living abroad, but was better used by Swiss residents overseas in their elections. In Estonia, e-voting has resulted in more persons voting from more different places (i.e. also when they were outside the country) – a result which is more in line with the idea that if something is relevant, then people will probably use it. But voters are not going to be persuaded to use their new devices to vote for candidates and choices that do not appear relevant to them.

Many aspects play a role in this new remote voting value-chain, including how user-friendly the technology is. A legally-recognized digital ID will be a major component of a more user-friendly system. This is indeed one of the foundational technologies that has made e-voting widely used in Estonia.

New rights created

There is also evidence that new voting technology has created important new freedoms: eligible citizens have acquired new freedoms in terms of when they can vote (anytime, day or night, from the time early votes are accepted) and where they can vote (from anywhere). This technology has even created the new freedom (which exists in Estonia) for voters to change their vote, as often as they wish, switching it from one candidate to another, until the polls close. The reasons for such a feature are two-fold: first, because the technology makes it possible (that’s what technological innovation is all about, after all – if the app offers the feature, some will use it), and second, because it makes it less likely for coercion to occur at home or at work since voters can keep changing their preferences from anywhere at any time. All three of these new freedoms were triggered by a new voting technology – a paper-based system would not make it easy to keep a polling booth open 24/7 or allow voters to change their vote after they had cast it or allowed voters to vote from anywhere.

Concluding thoughts

E-voting (assuming honest software, good governance and the appropriate checks and balances) [10] potentially represents a major upgrade of democracy. It will become cheaper and easier to formally obtain the people’s consent. The “transaction costs”, to use the academic jargon, will fall dramatically – both for the voter and for the state. The implications of this efficiency gain will differ from country to country. But its real disruptive potential lies elsewhere: It opens the possibility, unforeseen by the founding fathers of old constitutions, that it would become logistically much easier to consult the people on more things more frequently. Once that realization dawns in more places, someone, somewhere is going to become the first to want to try out such a disruptive thought. If the ballot is a button on your mobile phone, should the app only be used for general elections every few years? What about the in-between years? Should the people not be consulted more often on more things? If not, why not?

The growing availability and use of technology enabling electronic voting will, over time, open up a Pandora’s box. It will challenge the centuries-old customs of a representative system that was built in the days when the majority of the voters were labouring in the fields and knew very little about the policy debates taking place in a faraway legislature.

The availability of new technology by which the people can choose to make their will known, combined with the growing complexity faced by modern-day legislatures, is hastening the decline of the representative system as we’ve known it.

References

[1] https://en.wikipedia.org/wiki/COVID-19_apps

[2] https://uk.reuters.com/article/us-health-coronavirus-sewage/swiss-researchers-see-sewage-as-early-warning-sign-for-covid-flares-idUKKBN22C1QD

[3] https://www.ncsl.org/research/elections-and-campaigns/state-action-on-covid-19-and-elections.aspx

[4] Krimmer, Robert. 2012. The Evolution of E-voting: Why Voting Technology is Used and How It Affects Democracy, Tallinn University of Technology Doctoral Theses Series I: Social Sciences, No. 19, Tut Press, pp. 16-17.

[5] https://www.technologyreview.com/2019/04/18/136042/west-virginia-will-allow-blockchain-voting-in-the-2020-election-thats-a-risky/

[6] Bull, Christian et al. 2016. “The Imperfect Storm”. Paper presented at the International Conference on Electronic Voting, “E-Vote-ID 2016”, 18–21 October 2016, Lochau/Bregenz, Austria.

[7] Heckelman, J.C. 1995. “The Effect of the Secret Ballot on Voter Turnout Rates”, Public Choice (1995) 82: 107. doi:10.1007/BF01047732

[8] Yakobson, Alexander. 1995. “Secret Ballot and Its Effects in the Late Roman Republic”, Hermes, Vol. 123, No. 4, pp. 426-442.

[9] Luechinger, Simon et al. July 2006. “The Impact of Postal Voting on Participation: Evidence for Switzerland”, Working Paper No. 297, Institute for Empirical Research in Economics, University of Zurich, http://www.econ.uzh.ch/static/wp_iew/iewwp297.pdf.

[10] Apart from IT security concerns (see the MIT Technology Review in footnote 5 above), there is already a build-up of case law on the issue in many countries, including in Argentina, Australia, Brazil, Estonia, France, Finland, Germany, India, Mexico, Switzerland, the United States and Venezuela. Source: Driza Mauer, Ardita and Barrat, Jordi (eds.). 2015. E-Voting Case Law: A Comparative Analysis. Routledge.

The legal challenges come from three main angles: first, there is the claim that the technology might not ensure the secrecy of each individual’s choice (like when you mark your cross on a piece of paper privately behind a curtain at a polling booth), second, that what happens on the inside of the computer is not verifiable by ordinary members of the public (like when election observers can see how the ballot boxes are emptied onto counting tables to be counted and the people doing the counting can be monitored), and third, that each vote might not be given equal treatment if different types of ballots and different mechanisms are used for recording, transmitting, sorting, counting and storing votes in different constituencies. This latter point becomes especially important in very tight races and when recounts are needed, and a tiny number of votes can make the difference.