The Guardian APRIL 6, 2018 - by Brian Eno


Political parties, in league with tech firms, target voters in a few marginal seats to sway national results. This isn't democracy.

Amid all the coverage of the Cambridge Analytica scandal, the real danger has hardly been mentioned: an electoral system that is wide open to manipulation.

Because the UK uses first past the post, our general election actually means six hundred and fifty local elections - each one to elect one MP. A handful of these local battles tend to be close contests, hard fought by political parties, but most are not. Come election time, most constituencies have no prospect of changing hands. This means that the few marginal seats, and the small number of swing voters, who live in them decide our government.

To understand how tiny this key group of voters can be, consider that we would now have a Conservative majority government if just five hundred and thirty-three people had voted differently in just nine constituencies last year. It's the marginals that matter, and so the ruthless targeting of them is an irresistible temptation. All the political parties have been doing it for years. Candidate spending in marginal constituencies was 37.5% higher on average than in safe seats, without taking into account the volunteer activists bussed into these key seats. Momentum reports that more than a hundred-thousand people used their My Nearest Marginal to find where to campaign for Labour.

Labour may be "for the many", but when it comes to elections, its approach - like all parties - is to target the few. This is the rational strategy in a first-past-the-post system, ignoring the point that our votes are supposed to have equal value, wherever we live.

Now, in the new world of online campaigning, and with the ability of companies such as Facebook to produce detailed profiles of its users, first past the post is more vulnerable than ever. Make Votes Matter has been drawing attention to a surprisingly candid "success story".

Facebook explains how in 2015 it "provided the Conservatives with the means to reach the people that mattered: key demographics in marginal constituencies", using micro-targeting to avoid "wasted impressions".

Facebook claims to have reached 80.6% of all Facebook users in the target constituencies. By speaking to "the right people, over and over again", less than thirty-seven percent of the popular vote was turned into fifty-one percent of the seats, and therefore a Conservative majority. If you want to win elections in our archaic system, you can do it by buying data about marginal voters.

When the world's most primitive voting system is targeted by the world's most sophisticated data outfits, democracy doesn't stand a chance. The Cambridge Analytica scandal shows just how out of control the data side of the equation has become. Whether it broke the law or not is missing the point: what it was setting out to achieve can be done - and is done - entirely within the law. Putting one company out of business doesn't solve the problem, but changing our electoral system would.

With proportional representation, the share of seats each party wins reflects the share of the vote they receive. There are tried-and-tested systems of PR in use across the world - and in the UK's devolved assemblies - that keep a local constituency link and give voters far more power to decide who will represent them.

The basic intention of all forms of PR is the same: to ensure that parliament reflects the people as fairly as possible. It avoids the clearly unfair results that frequently occur under first past the post - a party with a minority of the votes ending up with a majority of the seats, or a party that gets millions of votes winning no representation at all.

Every vote counts with PR, not just the marginal ones, so it takes millions of votes to change the final result. In Germany's federal elections last November, the CDU won ninety-three more seats than the Social Democrats. Because they use PR, the only way to change this outcome would be to persuade 2.9 million people to vote differently. There's no possibility - and therefore no temptation - to target a couple of thousand votes in a handful of marginal seats.

This isn't a radical new innovation. Around 85% of developed countries already use some form of PR. We are the outliers; the exception. PR would make our elections far more resilient to the rapid evolution of big data and micro-targeting; it would free political parties from a joyless arms race in marginal constituencies, and it would make everyone's vote matter.