First it was the UK general election, then the EU referendum and now Donald Trump winning the US Presidential election. The polls and bookmakers were wrong on all three (but, to be fair, the bookmakers always priced their book for the Conservatives to win the most seats in the General Election, they just did not think they would achieve an overall majority).
Historically, the bookmakers have had a better run of political prediction than the polls. Apart from recent times they have managed to forecast the winner of every general election since Edward Heath (7/1) beat Harold Wilson in 1970.
What has gone wrong in recent year? Let’s deal with the polls first. It is mostly media companies – both broadcast and print – which pay the polling companies to produce a poll on who will win an election or referendum. Bias is all too prevalent amongst the media. The Daily Mail and Daily Telegraph favour the Conservative Party and Brexit. The Daily Mirror and The Guardian favour Labour and Remain. Then there is a third group, the Liberal left, which does not exactly favour one political party or another, and tends to support those whose manifesto aligns with their views at a particular time. In this camp I would place The Financial Times, The Economist, Gina Miller, and mostly all the television networks. They were all Remain. All were opposed to Trump.
By the time a newspaper has been printed the news is old hat. Everything has already been picked up on social media, news channels and internet messaging. So why buy a newspaper? For the opinion. Newspapers are more often than not purchased because the leader writers express a view and explain issues that are complex. But opinions are biased. They have to be otherwise it would be just information. So in an era of instant news newspapers have become more biased than at any time I can remember.
With all this influence I believe a modicum of bias has crept into polling. As a predictor of future outcomes for the purpose of betting they are not worth the paper they have been printed on. Party activists I talk to in the UK very often tell me when they knock door to door the reaction they get is very different to what the polls are telling them.
But what about the bookmakers?
They have undoubtedly been influenced by the media, and the polls. They have crossed the delicate line from bookmaker to becoming punter. All along Trump proclaimed he was going to win big, ‘Brexit times three’ he said. No doubt he was boosted by the reaction his people were getting from the streets, and also their private polls. The press were certainly in favour of Hillary Clinton, the democratic candidate. Around 240 daily publications endorsed Clinton but Trump only received 20 endorsements. But wanting someone to win does mean it will become a reality.
Paddy Power, having well and truly crossed the line from bookmaker to punter, advertised “A woman’s place is in the House! [With the White House in the background], for once we agree Donald. We’re paying out early on a Hillary victory”.
Clever stuff – Paddy Power paid out on both Hillary and The Donald.
Other bookmakers also lost money. William Hill told the Financial Times (FT) that 70% of total money received was on Clinton, but 70% of individual bets placed were on Trump. Hills made a six figure loss. How come with all that liquidity?
Hills took a position and thought Hillary the winner for the book and Trump the loser because they thought he could not win. Just because Brits might not vote for him it clearly does not mean Americans would not.
When Trump announced his candidacy he was 150/1. When the polls opened on Tuesday he was priced at 10/1 in a two candidate race. The man in the street could not resist a binary bet at 10/1 and they bet Trump with their £10 and £20 bets.
According to the FT, Bet365 lost GB£ 4million on the election, Skybet £1 million.
The lesson for bookmakers from this debacle is this: leave your own political opinions at home, ignore the media (the electorate certainly does) and bet to figures.