Humans Beating Robots Most Since ’08 as Trends Shift: Currencies

Humans are proving more adept than computers in reacting to the Federal Reserve’s mixed messages on when policy makers will reduce their unprecedented stimulus.

Currency funds that use computer models for trading decisions made 0.7 percent this year through June, compared with 2.3 percent for those that don’t, the biggest margin since 2008, according to data today from Parker Global Strategies LLC. Hong Kong-based Ortus Capital Management Ltd.’s $1.1 billion computer-model fund lost 13.8 percent in the first half, while the FX Concepts Global Currency Fund, which employs a similar strategy, fell 3.3 percent, Barclay Hedge Ltd. data show.

Diverging policies at the world’s central banks helped send measures of currency and interest-rate volatility to the highest levels in at least a year, creating a disadvantage for computer algorithms that rely on established trading patterns and correlations. While the Fed signals it may reduce the amount of money it injects into the economy, central bankers in Europe and Japan say their economies still need help.

“Central banks have become the insider traders of the currency market, which is a paradigm shift that systematic traders cannot pick up as well as fundamental traders,” Richard Olsen, the founder of Olsen Ltd., who has designed currency-trading models since 1986, said yesterday in a phone interview from Zurich.

Relative Returns

Computerized systematic funds, also known as quants, have grown in popularity since the 1990s and are now more common than the human, discretionary method of trading currencies. Before this year, they have historically been more profitable.

Of 43 currency funds tracked by Parker Global, 27 are driven by computer models. They returned an average 10.7 percent a year since the Stamford, Connecticut-based consulting firm started compiling the data in 1986, compared with 8.6 percent for discretionary funds.

Systematic funds were “wrong-footed” in the second quarter by the Fed’s signals that it would start scaling back monetary easing, according to Caio Natividade, the head of foreign-exchange and commodity quantitative research at Deutsche Bank AG, the world’s biggest currency trader.

Carry Trades

After a positive start to the year, quant funds added to bets in the carry trade, in which investors borrow cheaply in one currency to buy assets in other currencies of economies with higher interest rates, he said.

“Most systematic funds are large in size,” Natividade said in a phone interview from London on July 19. “When there’s a significant change in market sentiment, these funds can’t adjust positions that quickly.”

Fed Chairman Ben S. Bernanke and other policy makers have made what markets took to be conflicting announcements on whether the U.S. central bank will pare the $85 billion of monthly bond purchases it uses to boost the economy.

European Central Bank President Mario Draghi, in contrast, said July 4 he foresees maintaining stimulus measures for an “extended period.” The Bank of Japan earlier this month stuck with an April pledge to expand the monetary base by 60 to 70 trillion yen ($704 billion) per year.

The policy debate whipped up volatility and caused the euro-dollar rate, which tracks the most-traded currency pair, to alternate between gains and losses each month from March through June, oscillating in a range of $1.2746 to $1.3417.

Rising Volatility

JPMorgan Chase & Co.’s Global FX Volatility Index jumped to a one-year high of 11.96 percent on June 24, from 8.07 percent at the end of last year, and was at 9.34 percent today. Deutsche Bank AG’s G-10 FX Carry Basket has lost almost 8 percent since peaking at a 4 1/2-year high of 125.37 April 11.

The Bank of America Merrill Lynch Option Volatility Estimate Move Index, which measure the projected pace of swings in U.S. government debt yields, climbed to 117.89 on July 5, the highest level since 2010.

Quant funds use computer models based on inputs such as price history and correlations, and typically execute buy or sell orders automatically at certain trigger points.

While discretionary fund managers may use technical analysis in their research, they commonly rely on economic indicators to dictate their trading strategies.

‘Independent Mind’

To Olsen, the advantage of computerized systems is that their decisions aren’t handicapped by human emotions.

“A human being doesn’t have an independent mind,” Olsen said. “A trader who sits on a profit evaluates the economic, political and market situations differently to a trader who sits on a loss. A trader who has a hangover from a long night of alcohol will also look at it differently.”

Axel Merk, the president and founder of Merk Investments LLC in Mountain View, California, said he changed the approach of a quant fund his firm manages to incorporate more human judgment after it lost 8.4 percent in 2011.

The approach paid off, with the $23 million Merk Absolute Return Currency Fund gaining 8.5 percent over the past year, beating 82 percent of its competitors, Bloomberg data show.

“No matter how great your back-testing is, if policy makers change the rules all the time, your back-testing is going to fail,” Merk said in a July 18 phone interview. “That environment will persist because policy makers will continue to be very active in the markets.”

Correlation Breakdown

The breakdown of long-held asset correlations is proving problematic for many computer-based funds.

The Mexican peso, which used to be a barometer of risk sentiment, is no longer moving in tandem with the Standard & Poor’s 500 Index of stocks. The 60-day correlation declined to 0.4, the lowest since 2009, according to data compiled by Bloomberg. A reading of 1 indicates the two move in lockstep.

“In an environment where we see more choppy price action, it can be more problematic for a lot of those funds,” Adrian Owens, a money manager in London at GAM U.K. Ltd., which oversees $53 billion, said in a phone interview on July 16. “If you look for a big trend, you get chopped up.”

Owens, whose GAM Star Discretionary Fund has gained 12 percent this year, said he’s betting that the British pound will rally against the euro as the U.K. economy shows “signs of life,” while the 17-nation euro area needs a weaker currency to export its way out of a record-long recession.

The euro strengthened 5.9 percent this year to 86.06 pence at 12:07 p.m. in New York, touching a four-month high of 87.11 pence on July 17.

Probability Theories

Quant investing dates back to the 1960s, when Edward Thorp, a mathematics professor at the Massachusetts Institute of Technology, used probability theories to bet, first on blackjack and then on Wall Street. Banks and hedge funds such as London-based Man Group Plc and Winton Capital Management Ltd. popularized their use over the past two decades, expanding trading from stocks to bonds and currencies.

Ortus’s quant fund is headed for its second year of declines, following a 17.3 percent loss in 2012, according to Fairfield, Iowa-based Barclay Hedge, which tracks hedge funds’ performance. Its assets fell to $1.1 billion in June, from $3.1 billion in 2011. Officials in the investor-relations department didn’t return phone calls and an e-mail seeking comment.

Assets of the FX Concepts Global Currency Fund, run by John Taylor’s FX Concepts LLC in New York, shrank to $624 million in June, from $3.7 billion in 2008, when it returned 11 percent. This year’s losses followed a less than 1 percent gain in 2012 and a decline of 19 percent in 2011, data compiled by Barclay Hedge show.

“Political rather than economic” factors tend to move markets now, Robert Savage, the chief strategist at FX Concepts, said in a July 22 phone interview from New York. “This is a hard business whether you’re on a gut-checking basis or computer-model basis. Either way, no one’s making money easily.”

Before it's here, it's on the Bloomberg Terminal.