Com­mod­ity traders forced to shut up shop by ma­chines

Irish Independent - Business Week - - INTERVIEW - Eric On­stad

HOCFINGER” made his name and his money by tak­ing bold bets on co­coa mar­kets. But af­ter nearly four decades of trad­ing, some­times win­ning, some­times los­ing, An­thony Ward threw in the towel.

Mr Ward blames the rise of com­puter-driven funds and high-fre­quency trad­ing for forc­ing him and some other well-known com­modi­ties in­vestors to close their hedge funds and look for op­por­tu­ni­ties where ma­chines can’t make a dif­fer­ence.

While com­put­erised trad­ing is not new, Mr Ward and oth­ers ar­gue its steady rise has reached a tip­ping point that is dis­tort­ing prices and cre­at­ing un­cer­tainty not only for in­vestors, but for choco­late firms, car­mak­ers and oth­ers who rely on com­modi­ties.

It was in Jan­uary 2016, af­ter a slide in co­coa prices, that Mr Ward de­cided the days of tra­di­tional com­mod­ity in­vestors do­ing well from tak­ing po­si­tions based on fun­da­men­tals such as sup­ply and de­mand may be num­bered.

“It was just too big, too quick, too dra­matic. And com­pletely against the fun­da­men­tals,” Mr Ward told Reuters.

Com­mod­ity mar­kets fell across the board that month af­ter weak fac­tory data in China raised fears of lower de­mand from the world’s top con­sumer of raw ma­te­ri­als.

Ward blamed the slide in co­coa on what he re­garded as mis­placed sell­ing by com­puter-driven funds re­act­ing to the Chi­nese data, given China has scant im­pact on the co­coa mar­ket.

“The ac­tual fun­da­men­tals in co­coa were ex­traor­di­nar­ily bullish in Jan­uary 2016. We were fore­cast­ing the largest har­mat­tan in his­tory, which is ex­actly what hap­pened,” he said.

His pre­dic­tion that a hot, har­mat­tan wind from the Sa­hara desert would hit har­vests in Ivory Coast and Ghana and drive co­coa prices higher did come to pass – but not be­fore the fund had been forced to cut its losses when the mar­ket slumped.

At the end of 2017, Mr Ward closed the CC+ hedge fund that had in­vested in co­coa and cof­fee mar­kets for years.

And at the end of Jan­uary, com­mod­ity hedge fund Jami­son Cap­i­tal Part­ners run by Stephen Jami­son closed. He told in­vestors that ma­chine learn­ing and ar­ti­fi­cial in­tel­li­gence had elim­i­nated short-term trad­ing op­por­tu­ni­ties, while com­modi­ties did not of­fer ob­vi­ous ben­e­fits in the long term.

Also in 2017, renowned oil trader An­drew Hall, who earned $100m (€81m) in 2008, called time on his main Asten­beck Com­modi­ties

Fund II.

He had said in an ear­lier let­ter to in­vestors that ex­treme volatil­ity caused by “non-tra­di­tional in­vestors and al­go­rith­mic trad­ing” made it dif­fi­cult to hold onto long-term po­si­tions when the mar­ket moved against them.

In 2016, Michael Farmer, found­ing part­ner of the Red Kite fund that spe­cialises in cop­per, also ac­cused high-fre­quency traders us­ing su­per-fast com­put­ers of dis­tort­ing the mar­ket and get­ting an un­fair ad­van­tage.

Other in­vestors have taken refuge in re­lated sec­tors or left com­modi­ties al­to­gether, ex­as­per­ated by the au­to­mated trad­ing that drives about half of US com­mod­ity fu­tures trad­ing.

A study by the US Com­mod­ity Fu­tures Trad­ing Com­mis­sion last year showed com­put­erised trad­ing on the world’s largest fu­tures ex­change, CME Group, ac­counted for 49pc of the vol­ume in agri­cul­ture con­tracts and 58pc for some en­ergy con­tracts.

At the same time, data from in­dus­try tracker Hedge Fund Re­search shows the av­er­age hedge fund re­turned 8.64pc in 2017 but com­mod­ity hedge funds eked out re­turns of just 0.43pc. Mr Ward es­ti­mates that while in the past au­to­mated trad­ing would dis­tort the mar­ket by 10pc to 15pc from prices jus­ti­fied by fun­da­men­tals – which he said was ir­ri­tat­ing but of­ten man­age­able – it can now reach 25pc to 30pc.

Al­go­rith­mic, or sys­tem­atic, funds use com­put­ers to make de­ci­sions af­ter pro­cess­ing vast amounts of data, or trade on sig­nals such as mar­ket mo­men­tum or when prices hit key lev­els.

Those who run au­to­mated funds ar­gue that they in­ject much-needed liq­uid­ity while cap­tur­ing the dy­nam­ics of the mar­ket more ef­fi­ciently than tra­di­tional trad­ing strate­gies.

Mr Farmer and oth­ers say, how­ever, that it is un­fair for ex­changes to al­low high­fre­quency trad­ing (HFT) groups to have co-lo­ca­tion plat­forms, al­low­ing them to put su­per-com­put­ers in the same data cen­tre as the ex­change servers.

They say that gives HFTs the tiny ad­van­tage they need to jump ahead of in­com­ing or­ders, ef­fec­tively pig­gy­back­ing. Tra­di­tional in­vestors say this ex­ac­er­bates mar­ket moves and in turn makes it more costly for them to take out hedges when price moves go against them.

Sys­tem­atic fund man­agers see the rise of their sec­tor as a part of a trend that is trans­form­ing not only fi­nan­cial mar­kets but wider so­ci­ety with the ad­vent of ar­ti­fi­cial in­tel­li­gence, ro­bots and ma­chine learn­ing.

“I don’t feel too sorry (for tra­di­tional fund man­agers),” said An­thony Lawler, co­head of GAM Sys­tem­atic, the quan­ti­ta­tive part of

Swiss money man­ager GAM Hold­ing, which had as­sets un­der man­age­ment (AUM) of SFR148.4bn at the end of Septem­ber.

“In­for­ma­tion, which used to be ex­pen­sive, dif­fi­cult to get, not eas­ily shared, is now ubiq­ui­tous,” Lawler said.

“That means it’s much more dif­fi­cult to have an in­for­ma­tion edge and ad­van­tage to the player who can digest and an­a­lyse the data the quick­est.”

GAM Sys­tem­atic, which had $4.3bn of as­sets at the end of Septem­ber, re­gards com­modi­ties as one of many mar­kets it mon­i­tors for op­por­tu­ni­ties by crunch­ing data such as weather fore­casts and ship­ping data that shows how full a ves­sel is.

He ac­knowl­edged that al­go­rithms can lead to mis­takes and over­gen­er­al­i­sa­tion, but said that should just open up op­por­tu­ni­ties for tra­di­tional fund man­agers.

Sys­tem­atic funds also make de­ci­sions based on the struc­ture and tech­ni­cal lev­els of mar­kets. Some an­a­lysts ar­gue that mar­kets ab­sorb fun­da­men­tal in­for­ma­tion be­fore many an­a­lysts are aware of it and au­to­mated funds are sim­ply more ef­fi­cient.

“I used to do a lot of fun­da­men­tal anal­y­sis when I was in eq­ui­ties and I re­alised it doesn’t re­ally mat­ter, it was a waste of time,” said Guy Wolf, global head of mar­ket an­a­lyt­ics at com­modi­ties bro­ker Marex Spec­tron.

“The truth is fun­da­men­tal anal­y­sis is ef­fec­tively the work of a his­to­rian, seek­ing to pro­vide ex­pla­na­tion for what has al­ready oc­curred,” he said.

Some fund man­agers and an­a­lysts say com­put­erised trad­ing has been am­pli­fied in com­modi­ties be­cause other par­tic­i­pants such as banks and pen­sion funds have cut ex­po­sure to the sec­tor.

Ever the risk taker, Mr Ward still plans to trade co­coa and cof­fee – but only with his own money - while look­ing for op­por­tu­ni­ties in the un­der­ly­ing phys­i­cal com­modi­ties mar­kets. And he’s keep­ing a close eye on com­put­erised trad­ing.

“In the end the whole thing will blow up be­cause when there’s a big prob­lem, a black swan event, they won’t be able to get out. It will be very messy,” said Mr Ward. (Reuters)

Newspapers in English

Newspapers from Ireland

© PressReader. All rights reserved.