The power of prediction markets have been illustrated many times over the years, but a quite visceral example was provided by a team from the University of Cambridge, who highlight how betting markets were much quicker to predict the outcome of the Brexit referendum than financial markets.
The findings, which were shared in a recently published study, showed that the betting markets predicted the outcome roughly an hour before the financial markets.
The researchers compared activity on the Betfair betting markets with the sterling-dollar exchange rate after polls had closed at 10pm. The team suggest that both markets were pretty slow to react to whatever data was available, with early bets providing profitable in both markets as a result.
First signs of movement
The markets began to shift after midnight, but it was the betting markets that were first out of the gate, with movement towards a Leave result at around 3am. The currency markets didn’t really begin to adjust until around 4am however, and at 4.40am the BBC and other news outlets began to predict a victory for the Leave campaign.
The researchers believe their results highlight the value of prediction markets in forecasting complex events, such as elections.
“Clearly, punters trading on Betfair are a different group of people to those dealing in FX for international finance. It looks like the gamblers had a better sense that Leave could win, or that it could at least go either way,” the authors say. “Our findings suggest that participants across both markets suffered a behavioural bias as the results unfolded. Initially, both traders and gamblers could not believe the UK was voting to leave the EU, but this disbelief lingered far longer in the city.”
That’s not to say that the betting markets were perfect however, and the researchers developed their own model that they believe could have predicted the outcome at around 1.30am.
Public information
Central to the model is the so called ‘efficient market hypothesis’, which supposes that information in the public domain is widely acted upon, therefore providing no benefit to people. This is an assumption that the Brexit vote didn’t seem to support however, as both betting and financial markets seemed slow to respond to information in the public domain.
Of course, you could argue that the unsociable hour may have contributed, but data from Betfair revealed that some 182,000 individual bets were still made between 10am on June 23rd and 5am on June 24th, whereas 88,000 trades were made on the GBP futures market during the same timeframe. Indeed, the volume of bets broke Betfair records for trading on a political event.
“Prediction markets such as betting exchanges are an ‘incentive compatible’ way to elicit the private opinions of participants, as people are putting their money where their mouth is, whereas what they tell pollsters can be cheap talk,” the authors conclude. “Prediction markets could in theory be used to help value or price financial assets during events such as major votes. This is an area I will be focusing on for future research.”
