To brexit or not to brexit?

Not many people know that for the Global Positioning System to produce the incredibly precise temporal and spatial measurements (yes GPS is first and foremost a clock…) that it does we needed to build receivers capable of picking up extremely faint radio signals emitted by satellites floating in space, from among an enormous amount of signals much stronger and much more numerous. We are not nearly as good when it comes to picking up economic or social signals able to predict socioeconomic events. We are in particular far from predicting the voter. Why am I talking about GPS? Because I am looking at what we know about the BREXIT referendum and I am amazed at how much we know about things far far away from our lives and how little we know about our backdoor. Perhaps we should be shifting our priorities when it comes to where we invest our tax payer money…

Less than a month before the BREXIT referendum takes place in the UK (June 23) the Financial Times Brexit poll tracker is calling it 47% to 40% for staying with 12% undecideds (which practically means that the polls are useless at this time). The Economist has it broken down along some demographics showing among other things that the old are mostly brexiteers. According to the bookmakers as of this writing 40% of the bets are on leaving and 60% on staying. On skybet you need to place 7 £ on staying to earn 1 £ while placing 1 £ on leaving will earn you 4 £. Clearly those that bet think the remain outcome is more likely.

In Google search it is a whole different ballgame. For the first time after February 20th 2016, at which point the date of the referendum was set, its salience matched and surpassed its February levels (see Figure below). The share of searches for leave has dropped from 75% in February to (at times) below 60% in May but leave-searches are still remarkably and persistently ahead. The searches are still about getting informed and clearly do not identify voter intend. This means that “remaineers” may still be included in searches of the form “what are the consequences of leaving the EU?” and hence we are just picking the fact that more people want to know about the ramifications of leaving than about those of staying. This would be quite natural since the unknown poses more questions than the status quo.

As we approach the date of the referendum -and provided that brexiteers and remaineers will polarize and demonstrate- we might get a different signal which is more related to voter intend.


This entry was posted in Big Data, nowcasting, politics, Social Media and tagged . Bookmark the permalink.