By John Naughton

Investigations into the political role of data analytics are welcome, but they won’t explain why people voted for Trump or leaving the EU.

Last May the UK information commissioner, Elizabeth Denham, launched a formal investigation into the use of data analytics for political purposes. This involved an initial exploration of what went on in the Brexit referendum campaign but potentially also in others. And given the global nature of digital data, the investigation also involved an investigation into how “companies operating internationally” were using the personal data of UK citizens for political purposes.

The overall goal of this inquiry was to understand how personal information was used in political campaigns. The commissioner was concerned about the “invisible processing” of citizens’ private data by algorithms that carry out data-matching and profiling – which of course is what Google and Facebook do for a living. Their automated engines were originally built to facilitate the targeting of commercial messages at their users. But what became clear in 2016 is that those same engines had been “weaponised” by political actors to deliver targeted political and ideological messages, and that’s a very different game. “When the purpose for using these techniques is related to the democratic process,” wrote Denham, “the case for a high standard of transparency is very strong.”

Quite so. What the commissioner may not have appreciated when she embarked upon her inquiry, though, is that transparency is about as popular with these companies – not to mention the political actors who exploit their automated systems – as garlic is with Count Dracula. Her investigation has covered more than 30 organisations – including political parties and campaigns, data companies and social media platforms, together with a mysterious Canadian outfit, AggregateIQ, whose name continually crops up in connection with this stuff.

Denham’s progress report – published last week on her office’s website – suggests it hasn’t exactly been plain sailing. “A number of organisations,” she reports, “have freely cooperated with us, answered our questions and engaged with the investigation. But others are making it difficult. In some instances we have been unable to obtain the specific details of work that contributed to the referendum campaign and I will be using every available legal tool and working with authorities overseas to seek answers on behalf of UK citizens.”

And some outfits have been obstructive – or “failed to be as comprehensive as I believe they need to be in answering our questions”, as the commissioner politely puts it – forcing her to invoke statutory powers to demand information from data processors. Her office has to date served four formal information notices as part of the investigation, including one to – guess who! – Ukip, which has now appealed against the commissioner’s demands to the Information Rights Tribunal. Cue battalions of m’learned friends and a long war of legal attrition.

Political campaigning has always been a dirty and corrupt business, polluted by dark money and the unscrupulous use of whatever communications media happens to be dominant at any particular time. The main lesson of the 2016 Brexit and Trump campaigns is that social media and data analytics are now the tools of choice for manipulating public opinion and influencing elections. At the moment, they have several advantages over old-fashioned broadcast TV. First, a lot of the money spent on them in political campaigning can be hidden from regulators, partly because it involves exploiting troves of personal data accumulated over decades by social media companies (and is therefore outside any particular campaign window). Second, unlike older media, different (and possibly contradictory) messages can be delivered to individuals, and no one will ever know. And third, they are fantastically cost-effective. Investigations in the US, for example, suggest that messages crafted by Russian agencies, which spent about $30,000 on Facebook, may have reached up to 126 million users. That’s an awful lot of bangs per buck and perhaps explains why the Leave campaign spent 98% of its communications budget on data analytics.

It seems obvious now that the weaponisation of social media played some role in both the Brexit referendum and the US election. What’s much less clear, however, is whether it was critical in determining the outcome. Personally, I’m sceptical. Our current obsession with digital technology as the trigger for these political earthquakes may actually be a kind of displacement activity. What we’re overlooking is that none of this would have happened if our ruling elites had noticed what four decades of globalisation and neoliberal economics had done to the life chances of many of our fellow citizens.

Nearly four million people in the UK voted for Ukip in 2015, for example, and got just one MP for their trouble. So when David Cameron presented them with a chance to give the neoliberal order a good kicking, they hardly needed their Facebook feeds to tell them what to do. I hope the information commissioner does succeed in unearthing the role of data analytics in Brexit. But even if she does, she’ll only have retrieved one piece of the jigsaw.