Predata Signals Offer a Fresh Perspective on the Brexit Campaign

Crossed lines: polls show the Brexit race neck and neck, but Predata offers a different way to view the campaign.

Richard Laurent

With UK polls unreliable in recent years and the pundits hedging their bets, signals drawing on metadata from YouTube, Twitter and Wikipedia offer a fresh perspective on how the Brexit race is unfolding.

If this is how the world ends, T.S. Eliot got it wrong: there’s both a bang and a whimper. Last week saw the start of official campaigning ahead of the June 23 referendum to decide whether Britain will leave the European Union. Speeches by two major figures in the campaign offered a poignant study in contrasts. Boris Johnson launched the Vote Leave campaign in Manchester with a typically exuberant, compulsively quotable performance in which he urged voters to embrace “freedom” and dismissed Prime Minister David Cameron and other “Bremainers” as “the Gerald Ratners of modern politics” — a reference to a prominent jeweler who once admitted on British TV his products were “crap.” Labour leader Jeremy Corbyn, meanwhile, took to the stage at Senate House in London and threw off years of inveterate Euroscepticism, and months of dodging the question on where his referendum allegiances lie, to officially embrace the Remain cause. Corbyn began his speech with a long rehearsal of the Labour left’s traditional gripes against the EU, and mostly read off his notes with head bowed; if this was an endorsement, it was offered with all the vim of wet lettuce. The contrast between the two speeches seemed to capture the choice facing the British people: the insurrectionist flash and dazzle of the Out campaign versus the dowdy, dad-like pragmatism of the In campaign’s case for the status quo.

Scottish voters were presented with a similar choice in the run-up to the independence referendum of 2014: the pro-independence movement had romance and the lion’s share of the headlines on its side, while the No campaign was left to push the dull-but-sensible option of leaving things the same. Polls predicted the result would go down to the wire; but despite the polls and the blaze of constant publicity the Yes side was able to call on throughout the campaign, the final result was a comfortable victory for the status quo. Polls ahead of the 2015 general election in the UK predicted a similarly nail-biting finish, with many putting the Conservatives and Labour on level pegging or Labour slightly ahead. Again the polls were wrong, and the Tories completed a thumping victory.

Leery perhaps of repeating the mistakes of the recent past, many commentators in the UK are approaching their prognostications for the Brexit referendum with an abundance of caution: the emerging consensus is that despite the Out campaign’s roster of high-wattage personalities and romantic message of freedom, the boring option will win out and Britain will still be part of the EU come June 24. There is significant divergence among the polls: online polls put the race neck and neck, while phone polls show a significant lead for the In campaign.

With UK polls historically unreliable and the pundits hedging their bets, new data points can offer a fresh perspective on how the Brexit race is unfolding. At Predata, we have built three signals that capture levels of interest online in the campaign. The first is a “Britain In” signal, which includes the most popular videos on YouTube for the In camp (some drawn from the channel of the official Britain Stronger in Europe campaign, others from neutral media sources featuring interviews or speeches given by prominent Bremainers) and the leading Twitter accounts advancing the case for staying. The second is a “Britain Out” signal, which pulls in equivalent material from the same sites for the Leave campaign. There is a similar number of sources used to generate each of these signals. The final signal is a general index for the Brexit topic as a whole, combining the sources for both the Britain In and Out signals as well as a number of neutral, information-only web pages that provide context on the referendum (e.g. the Wikipedia pages on Brexit in various European languages). Together, these signals draw on four primary sources: Twitter, YouTube, Wikipedia and the comments sections of leading news websites. The Britain In and Britain Out signals are composed exclusively of sources which clearly advance one respective side of the debate.

These signals, which are scaled from 0 to 100 and generated by combining page views with the level of contestation on a given webpage (roughly, how argumentative the conversation on that page is, whether via comments, replies, edits, or revisions), offer a way to understand where the interest in different components of the Brexit debate is greatest online and how that interest is shifting as the referendum date approaches. They do not, importantly, indicate how the final vote is likely to play out, substitute for polls, or track intentions or the sentiment of the British voting public (since the metadata we use to generate the signals is pulled from the open internet; anyone can look at or contribute to the webpages we monitor). But what these signals do offer is a gauge of how volatile the conversation online around each campaign is.

A referendum on membership of the EU was one of David Cameron’s main campaign promises in the 2015 election. He announced the June 23 date for the referendum in late February. Most of the sources used to generate our three signals — “Brexit,” “Britain In” and “Britain Out” — were created earlier this year. But other sources stretch back to 2015 or earlier. Over the last three months, the three signals have behaved as follows (the color key is on the right).

And here is a visualization of the signals over the last 30 days.

Digging into the substance of the signals — which pages have driven the various spikes in activity over the last few months — and how they correlate to each other, a few early conclusions can be drawn.

1. Britain Out dominates the debate online, but its lead is narrowing

Leading figures on the Eurosceptic side, such as Nigel Farage and George Galloway, have been calling for Britain to leave the EU for years, so it’s no surprise there’s a rich store of material online attesting to their commitment to the cause. (“I have dedicated 20 years of my life not just to getting this referendum, but to winning this referendum,” Farage said in February.) The Leave campaign has traditionally dominated the Brexit debate online; over the last six months, the Britain Out signal has a correlation of 86% with the broader Brexit signal, versus 74% for the Britain In signal. That correlation remains strong for the period covering February 20 (when the referendum was announced) to today.

In the last month, however, the relationship of the two campaigns to the broader Brexit topic has reversed: Britain In now has a correlation of 81% with Brexit, while Britain Out has fallen to 62%.

Correlations over the last month of the Brexit signal, Britain In signal, Britain Out signal, and GBP vs. USD exchange rate

Britain Out dominated the early skirmishes, but over the last month the In campaign has begun to organize and occupy a greater share of the chatter online. This can be seen clearly in the shift in what’s driving volatility in our overall Brexit signal: for the first three months of the year, signal activity was driven mainly by videos of Farage and Galloway laying out the case for Brexit, some of them from as far back as 2013. In recent weeks, however, the Britain Stronger in Europe campaign’s content — on both Twitter and YouTube — has emerged as the main focus of Brexit activity online.

2. The campaign is bad for the pound, but Britain Out is worse than Britain In

Both campaigns have a negative historical correlation to the value of the pound, but Britain Out’s negative relationship is more pronounced, at -72% over the last two months and -47% over the last month. The more volatile the Britain Out signal has been, in other words, the more the pound has weakened versus the US dollar; note, however, that this relationship is narrowing, suggesting perhaps that much of the negative effect of the Brexit case on the pound has already been priced in. The Britain In signal is similar directionally and the negative correlation to the pound-dollar exchange rate is also tightening, to -2% in the last month from -13% in the last two months and -35% since the beginning of the year. Should this trend continue, further increases in activity online around the In campaign will continue to turn more supportive for the pound.

3. Britain Stronger in Europe has a better organized online campaign than Vote Leave

After weeks of deliberation, the UK Electoral Commission last Wednesday designated Britain Stronger in Europe and Vote Leave as the official campaigns for the In and Out sides, respectively. This was an important moment in the referendum as each official campaign will be able to spend up to £7 million, versus a £700,000 spending cap for non-official campaigns. Britain Stronger in Europe was unchallenged for its designation on the In side, but Vote Leave had several challengers, most notably Grassroots Out, a coalition of Tory and UK Independence Party interests.

This division perhaps explains why the Out campaign’s presence, especially on YouTube, is so thin in comparison to that of Britain Stronger in Europe (BSE video count vs. Vote Leave: 23–9). By uniting and organizing earlier around one group, Britain Stronger in Europe has been able to release more content faster and has displayed greater consistency in its messaging, in which, perhaps predictably, themes of fear (loss of global influence, the negative impact of Brexit on the National Health Service, the risk of an economic downturn) cloud out the occasional ray of sunshine (the EU is good for British science and innovation).

4. Much of the interest online in the Britain Out case is centered on Nigel Farage

A close investigation of the history of the Britain Out signal shows a handful of videos of UKIP leader Nigel Farage — including his performance at the Brexit debate at Oxford earlier this year, as well as various appearances on the campaign trail over the last two months — regularly feature among the top drivers of activity online for the Out side. Farage was leading the charge for Grassroots Out to win designation as the official Out campaign; the victory of Vote Leave means he will presumably not now feature on official campaign literature.

Vote Leave organizers reportedly intend to place less of an emphasis throughout the campaign on Farage’s pet issue of immigration, which they fear has the potential to alienate centrist and undecided voters. In this sense, the official designation decision already appears to have had a decisive impact on how the case for Brexit will be framed. But Predata’s signals suggest that what we might call “the Farage case for leaving” — which puts a heavy emphasis on risks to Britain associated with the free movement of people throughout the EU, along with more routine concerns over loss of sovereignty and law-making power, the unaccountability of Brussels bureaucrats, and so on — remains a key driver of discussion and interest online in the Britain Out cause.

Predata will continue to add to and monitor these signals as more material comes online and the campaign intensifies. In the coming weeks, we will look more closely at how our signals relate to the major polls and betting markets associated with the referendum, as well as the parallels and lessons to be drawn from shifts in online activity leading up to the 2014 Scottish independence vote. | twitter: @predataofficial |