If you want to know the election winner, look to the betting markets

Updated June 1, 2015. This was the most polled election in British history, and most projections based on the polls suggested that Labour would finish with most seats in the House of Commons. But there is another way to predict elections, by looking at the bets made by people gambling on them. The betting markets were saying for months that the Conservatives would win a lot more seats than Labour at the election of May, 2015. So where should we be looking for our best estimate of what is actually going to happen in an election, to the polls or to the markets? It’s a question that we have been considering actively in the UK for nearly 30 years. We can trace the question to July 4, 1985, for that is the day that the political betting markets finally came of age in this country. A by-election was taking place in the constituency of Brecon and Radnor, a semi-rural corner of Wales, and at the time the key players, according to both the betting markets and the opinion polls, were the Labour and Liberal candidates. Ladbrokes were making the Liberal the odds-on favourite. But on the very morning of the election a poll by MORI gave the Labour candidate a commanding 18% lead. Meanwhile, down at your local office of Ladbrokes the Liberal stubbornly persisted as the solid odds-on favourite. So we had the bookmaker saying black and the pollster white, or more strictly yellow and red. And who won? It turned out to be the Liberal, and of course anyone who ignored the pollster and followed the money. Since then, the betting markets have called it correctly in every single UK general election. While this may be a surprise to many, it will be much less so to those who had followed the history of political betting markets in the US, which correctly predicted (according to a famous study) almost every single US Presidential election between 1868 and 1940. In only one year, 1916, did the candidate favoured in the betting the month before the election end up losing, and that in a very tight race, and even then the market got it right on the day. The power of the betting markets to assimilate the collective knowledge and wisdom of those willing to back their judgement with money has only increased in recent years as the volume of money wagered has risen dramatically, the betting exchanges alone seeing tens of millions of pounds trading on a single election. Indeed, in 2004 one betting exchange actually hit the jackpot when their market favourite won every single state in that year’s election. This is like a tipster calling the winner of 50 football matches in a row simply by naming the favourite. The power of the markets has been repeated in every Presidential election since. Two weeks before the 2005 UK general election, buoyed already by the track record of the markets in forecasting UK elections, and that 2004 prediction miracle, I was sufficiently confident, when asked by The Economist, to call the winner and the seat majority in the 2005 UK General Election over two weeks out. My prediction of a 60-seat majority for the Labour Party, repeated in an interview on the BBC Today programme, was challenged in a BBC World Service debate by a leading pollster, who wanted to bet me that his figure of a Labour majority of over 100 was a better estimate. I declined the bet and saved him some money. The Labour majority was 66 seats. The assumption here is that the collective wisdom of many people is greater than the conclusions of a few. Those myriad people feed in the best information and analysis they can because their own financial rewards depend directly upon this. And it really is a case of ‘follow the money’ because those who know the most, and are best able to process the available information, tend to bet the most. Moreover, the lower the transaction costs (the betting public do not pay tax on their bets in the UK) and information costs (in never more plentiful supply due to the Internet) the more efficient we might expect betting markets to become in translating information today into forecasts of tomorrow. For these reasons, modern betting markets are likely to provide better forecasts than they have done ever before. This is not to say that betting markets are always exactly right, and pollsters are always hopelessly wrong, but when they diverge the overwhelming weight of evidence suggests that it is to the betting markets that we should turn. The same happened in 2010, where the betting markets for weeks were strongly predicting a hung parliament, while the polls swung from at one point having the Liberal Democrats in the lead to in another case putting the Tories a whole twelve points up on election day.

Indeed, in an event as big and recent as the 2012 US presidential election, no less a name than Gallup called it for Mitt Romney, and the national polling average had the candidates essentially tied. Meanwhile, Barack Obama was very short odds-on to win.

Last year’s Scottish referendum is another example. While the polls had it very tight, and with more than one poll calling it for independence, the betting markets were always pointing solidly to a No. The mismatch between the polls and the result echoed the 1995 Quebec separation referendum in Canada. There the final polling showed ‘YES to separation’ with a 6% lead. In the event, ‘No to separation’ won by 1%. We happen to know that one very large trader in the Scottish referendum markets had this, among other things, very much in mind, when he placed his £900,000 to win a net £193,000. The point is that people who bet in significant sums on an election outcome will usually have access to all the polling evidence and their actions take into account past experience of how good different pollsters are, what tends to happen to those who are undecided when they actually vote, and even sophisticated consideration of what might drive the agenda between the dates of the latest polling surveys and polling day itself. All of this is captured in the markets but not in the polls. To some this is magic. For example, I was at a conference in the US when an American delegate, totally dumbfounded that we are allowed to bet on this type of thing in the UK, and that we would anyway be mad enough to do so, asked me who would win the Greek election. I only needed to spend 30 seconds on my iPhone to tell her that Syriza were sure things. She couldn’t believe a market could tell me that. The same thing happened when I announced two weeks before the recent election in Israel that Netanyahu was already past the post. The polls pointed the other way, but the global prediction markets had Benjamin Netanyahu as a clear and strong favourite to win. On declaring victory, he declared that he had won against the odds. Not so. He had in fact won against the polls. To be fair to the opinion polls, they were onside in the Greek election, as they were in the French and Australian elections. That is good. The real question, though, is whom to believe when they diverge. In those cases, there is very solid evidence, derived from the interrogation of huge data sets of polls and betting trades going back many years, much of which I have undertaken myself, that the markets prevail. To be still fairer to the pollsters, they are not claiming to be producing a forecast. They are measuring a snapshot of opinion, though we are have to be careful about his ‘snapshot defence’, as I term it, as sometimes this can be used as cover for a poor methodology. In any case, those inhabiting the betting markets are certainly trying to produce a forecast, so we would to that extent hope that they would be better at it. Moreover, the polls are used by those trading the markets to improve their forecasts, so they are potentially a valuable input. But they are only one input. Those betting in the markets have access to so much other information as well, including informed political analysis, statistical modelling, canvass returns, and so on. I say that the polls are potentially a valuable input. The most recent election in the UK, on May 7, 2015, demonstrated that this is certainly not always the case. In that election, the polls, even on the day, showed it neck and neck in vote share, consistent with the Labour Party winning significantly more seats. The betting markets, meanwhile, always had the Conservatives well ahead in the seats tally, and were unmoved even by last minute polls indicating a late swing to Labour. Take a comparison of the polls and betting markets three days before polling. The polls were indicating most seats for Labour while the betting markets had the Conservatives as short as 6 to 1 on favourites to win most seats, i.e. £6 to win £1. That’s confidence. In other words, those who put their money on the line were pretty sure what was going to happen all along. Since then we have had the ‘gay marriage’ referedum in Ireland. The polls were pointing to a decisive 70%-30% win for YES. The betting markets had the line at 60-40. In the event, YES won by 62-38. In conclusion, there is a growing belief that betting markets will become more than just a major part of our future. Properly used they will, more importantly, be able to tell us what that future is likely to be. We seem therefore to have created, almost by accident, a ‘high-tech’ crystal ball that taps into the accumulated expertise of mankind and makes it available to all. In this sense, the market is like one mind that combines the collective wisdom of everybody. In this brave new world of prediction markets, it seems only sensible to make the most of it.