Pollsters should shut up shop after this election disaster

5 November 2020

12:50 AM

5 November 2020

12:50 AM

High Street chains are closing all the time. The restaurant chains are shutting their doors. The gyms are going out of business and it doesn’t look as if all the airlines will survive. There have been lots of different industries that have been wiped out this year. But now we can add one more to the list: the polling companies.

It remains to be seen who won the US Presidential election. Donald Trump and Joe Biden may fight it out for a few more days yet. But it is absolutely clear who lost. The people who are paid plenty of money to talk to voters and work out what they are thinking.

It was meant to be a ‘Blue Wave’, meaning the Democrats would take both the White House and Congress, probably in a landslide. The forecasts based on polling predicted that with between 80 per cent and 90 per cent certainty. FiveThirtyEight predicted that if the election was held 100 times, Trump would win only ten times. The stock market was so convinced of the result it had more or less given up thinking about it, beyond working out which companies would benefit most from a Biden victory. And yet that isn’t how it has turned out. It is neck and neck, with Trump doing far better than anyone predicted.

That is hardly the first time it has happened. The polls got the 2016 presidential election wrong. And the Brexit referendum in the UK. True, they still get the occasional call right – the British election of 2019 for example – but flipping a coin will give you the right answer as well sometimes, especially in a two-horse race. Poling was meant to be a bit better than that. Investment banks, newspapers, and political parties paid big bucks to know what was going to happen. They weren’t paying for someone to flip a coin; they could have done that themselves.

That is a big change. A decade ago, polling was so accurate it was hardly ever worth staying up for the result. Now it has started to fail on a regular basis. Why? We can debate the reasons endlessly. Maybe voters are lying because they don’t want to own up to their real opinions. Maybe the pollsters are asking the wrong people. Maybe social media is shifting opinion at the last moment. Who knows. The major polling companies have done their own inquests and tried to fix the way they are collecting information. The trouble is, so far it doesn’t seem to be working.

That has two consequences. First, the companies will be in big trouble. Private and corporate polling was a lucrative gig: the banks and hedge funds in particular can make big money by calling an election right, and will pay handsomely for an edge. And elections results will be far more volatile. Instinctive, gut politicians will thrive in a world where no one knows what the voters might decide. Focus group obsessed moderates will find themselves regularly defeated. Without accurate polls, elections will be a lot more interesting – but also a lot harder to navigate.

Got something to add? Join the discussion and comment below.

Show comments