Man versus machine: computer-driven hedge funds win on Brexit night

 

As hedge fund manager Buford Scott sat at home, watching the TV in shock as it emerged Britain had voted to leave the European Union, his computer-based trading models were quietly boosting his business by 1.5 percent.

Scott's algorithm-driven fund, and others like it, beat hedge funds run by humans on Brexit night, not because the computers had correctly predicted the results of the June 23 referendum, but because they followed trends already in place.

Computer-driven funds placed bets on safe-haven assets like the yen and gold, which had performed well in the six months leading up to the vote, against riskier currencies like the Mexican peso - a play yielding a more than 7 percent return on June 24.

Hedge funds on average were down 0.18 percent on the day of June 24, while trend-following machine-based strategies gained 0.71 percent, according to data from industry tracker Hedge Fund Research (HFR).

Some of the biggest hedge funds' computer-run strategies - also known as systematic or quantitative - made huge gains in the aftermath of the vote for Brexit.

British systematic firm Winton Capital, which manages more than $30 billion, made a gain of 3.1 percent on June 24, while Greenwich, Connecticut-based AQR's $13.3 billion computer strategy returned 5.2 percent - about $700 million.

Polls had shown the Leave and Remain camps neck-and-neck before the referendum. But partly on the view that undecided voters tend to opt for the status quo, bookmakers and betting exchanges had largely predicted Remain would win comfortably. Most players in financial markets agreed.

"Almost no one was betting on the actual outcome of Brexit," said GAM portfolio manager Anthony Lawler in London, a hedge fund investor whose firm recently bought systematic investment manager Cantab Capital Partners, which made 4.1 percent in the year to June 27.

"Man chose to reduce risk given the unknown. Machines chose to be long safe-haven assets because that trend was already in place. Machines won."

This does not necessarily mean machines would tend to outperform during big global events - a Remain vote might well have seen humans come out top - or during unexpected twists. Machine models were down 2.42 percent in 2015 while hedge funds on average were down 1.12 percent, according to HFR.

But in this case machines fared better because, with no preconceived ideas about the result, they followed the same strategies as previously, strategies which a Leave vote happened to favour.

The few non-systematic hedge fund managers who did make money in the immediate referendum aftermath were highly contrarian, such as Crispin Odey, a Brexit backer who made 15 percent by betting his $10.2 billion fund's assets on a fall in financial markets and a surge in safe-haven gold.


source

Reason: