Bayes’s theorem, Brexit and the world economy

This is something you won’t hear from either Brexit or Remain camps:

Statistical theory even allows us to quantify how expectations about the US presidential election should shift if Brexit wins in Britain. Suppose, for the sake of simplicity, that we start by giving equal credibility to opinion polls showing Brexit and Trump with almost 50% support and expert opinions, which gave them only a 25% chance. Now suppose that Brexit wins. A statistical formula called Bayes’ theorem then shows that belief in opinion polls would increase from 50% to 67%, while the credibility of expert opinion would fall from 50% to 33%.

Since bond markets hate uncertainty, you can imagine what happens next.

The triumph of hope over experience

This morning’s Observer column:

There is something irresistibly comical about the spectacle of two CEOs announcing a friendly takeover. The two chaps (for they are still generally chaps) stand side by side, grinning into the cameras. The proud new owner explains what a great outfit his latest acquisition is, how pleased he is with the deal, extols the “synergies” that will magically materialise once the marriage is consummated and expresses his undying admiration for the poor schmuck who is now his latest subordinate.

The schmuck, for his part, declares his undying admiration for his new boss and his deep respect for the gigantic organisation into whose maw he is about to disappear. He, too, is “incredibly excited” by the new horizons that are now open to him and his colleagues. The marriage is a very good deal for both organisations – a win-win outcome no less. The fact that he omits to mention how much he has personally made from the deal is tactfully overlooked by his admiring media audience.

Last week’s announcement of Microsoft’s acquisition of LinkedIn followed this script to the letter…

Read on