Lessons learned from the Trump victory
The free-rider problem (1): from an economics point of view, voting in large democracies suffers from a problem … each individual voter’s vote is highly unlikely to influence the outcome. This problem is often cited as a reason for low voter turnout, and is a reason why Australia has compulsory voting. The less cited problem is that it allows voters to vote without accounting for the consequences of their action. The protest vote is not usually considered a big problem, but, yesterday, individual voting without accounting for the collective consequences has produced an outcome with global outcomes that will last for years to come.
Free rider problem (2): Despite access to exceptional quantities of real time data and unprecedented computing power pollsters appear to have been completely wrong. There is an important lesson to be learned. Clearly scope to do polling better. One reason why they didn’t is that information tends to be free (even if it’s used to sell ads, once the polling numbers are released, they can be shared freely). So polling firms will tend to under invest in the resources needed to do it well.
This has real implications. First, it influences people’s behaviour. Some people might have stayed home and not voted because they figured it was a sure thing. But it also has resulted in markets being very wrong. Markets used the information from polls to value assets that have turned out to be very wrongly priced. It looks like stock markets will be down 5 per cent globally. Likely trillions of dollars. The value of better information is huge, but the free rider problem meant markets failed to provide it. The parallel with the Brexit vote is obvious.