Ro­bots pick the stocks

The National - News - Business - - The Life - Nis­hant Ku­mar

Since he left his job as a hedge fund man­ager last year, Muhammed Ye­sil­hark has spent his days teach­ing com­put­ers to pick stocks like he did in his 12year ca­reer.

By the end of the year, the former head of Euro­pean eq­ui­ties at Carmignac Ges­tion, who pre­vi­ously traded for the bil­lion­aire Steven Co­hen and the hedge fund York Cap­i­tal Man­age­ment, says the al­go­rithms will be ready to help run money for clients of his own in­vest­ment firm, Q2Q Cap­i­tal. Mr Ye­sil­hark, who will run the fund from Dubai and Lon­don, said he has raised US$100 mil­lion and aims to gather $800m in to­tal from 40 in­vestors. The com­puter mod­els he de­signed are meant to ac­cel­er­ate idea gen­er­a­tion for Mr Ye­sil­hark’s small in­vest­ment team and learn from mis­takes. More than 40 of them cur­rently trawl through troves of data, scan real-time trad­ing pat­terns and read com­pany re­leases us­ing nat­u­ral language pro­cess­ing tech­niques to spot the per­fect trad­ing op­por­tu­nity. It is an area that re­quires a heavy in­vest­ment of time and money and is dom­i­nated by top firms, in­clud­ing Mr Co­hen’s Point72 As­set Man­age­ment.

“He has cer­tainly got a nice pedigree and the idea is a great one, but the dif­fi­culty is in the im­ple­men­ta­tion of it,” says Ja­cob Sch­midt, the chief in­vest­ment an­a­lyst at NLP Fi­nan­cial Man­age­ment, which in­vests in hedge funds in­clud­ing quan­ti­ta­tive strate­gies. “It’s a ques­tion of money and man­power. All the big guys are do­ing it. They have huge teams.” Mr Co­hen’s Point72 As­set Man­age­ment is test­ing mod­els that mimic trades of its port­fo­lio man­agers. He is also ex­per­i­ment­ing with au­tomat­ing the work of its ex­e­cu­tion traders, who place buy and sell or­ders with bro­kers on be­half of money man­agers. Ray Dalio’s Bridge­wa­ter As­so­ciates is de­vel­op­ing al­go­rithms based on em­ployee data to help au­to­mate de­ci­sion-making at the world’s big­gest hedge fund.

“More than 50 per cent of what we do is repet­i­tive,” says Mr Ye­sil­hark, a mar­ried fa­ther of two. “The best way to solve the prob­lem is tech­nol­ogy.”

Us­ing com­put­ers to au­to­mate some of the in­vest­ing process is a broader trend among hedge funds, which are look­ing for a high-tech an­swer to years of un­der­per­for­mance and dwin­dling fees. Some firms are even turn­ing to com­put­ers to iden­tify trad­ing sig­nals them­selves and im­prove on their own, us­ing artificial in­tel­li­gence.

Such funds, known by the acro­nym AI, have yet to prove them­selves in the long run and on a broader scale. The Eureka­hedge AI Hedge Fund In­dex, which tracks 12 of these money pools, has out­per­formed hedge fund peers since 2013 but trailed the S&P 500 In­dex.

The trend to­wards artificial in­tel­li­gence is “real but it de­pends on the re­sources that one has to spend on tech­nol­ogy and hire the best peo­ple to re­tain an edge”, says Michele Ge­sualdi, who over­sees $3 bil­lion as the chief in­vest­ment of­fi­cer at Kairos In­vest­ment Man­age­ment, which in­vests in hedge funds. “It’s dif­fi­cult for a young firm to do this and com­pete.”

Mr Ye­sil­hark says his mod­els, while less am­bi­tious, have shown en­cour­ag­ing per­for­mance in the short span he has used them to find trades. Since Septem­ber, when the money man­ager started test­ing them with his own cap­i­tal, they re­turned 38 per cent, helped by win­ning bets on Syn­genta, the seeds multi­na­tional Mon­santo and the bat­tery maker Blue So­lu­tions. He de­clines to pro­vide de­tails of his mod­els.

“They never sleep, do not go to toi­let, have no girl­friends or boyfriends,” says the 37-yearold, who was born and raised in Ger­many. “I think artificial in­tel­li­gence and in­vest­ing is a big op­por­tu­nity.”

Mr Ye­sil­hark joined Carmignac in Jan­uary 2014 and left in March 2016. The Carmignac Port­fo­lio Grande Europe fund that he co-man­aged gained 10 per cent in 2014 and lost 1.4 per cent in 2015, ac­cord­ing to Carmignac’s web­site. He also co-man­aged the Carmignac Euro-En­trepreneurs fund, which gained 10 per cent in 2014 and al­most 12 per cent in 2015, ac­cord­ing to its web­site

The idea to clone his mind was sparked by his trad­ing days at SAC Global In­vestors, Mr Co­hen’s pre­vi­ous in­vest­ment firm, where Mr Ye­sil­hark was a part­ner be­fore mov­ing to Carmignac Ges­tion. Once his mod­els, which can au­to­mate as much as 90 per cent of the de­ci­sion-making, de­tect a po­ten­tial trade, Mr Ye­sil­hark makes the fi­nal call on whether to buy, sell or re­ject it. His new anal­y­sis is cod­i­fied again and the machine learns not to re­peat the same mis­take. The fund will not charge a man­age­ment fee but a fixed fee for ex­penses, as well as take a 25 per cent cut of the re­turns gen­er­ated. The per­for­mance fee will sup­port the bonus pool and a char­i­ta­ble foun­da­tion that Mr Ye­sil­hark has set up to back not­for-profit or­gan­i­sa­tions. The money man­ager says he will not be part of the firm’s bonus pool, us­ing it as an in­cen­tive for his staff to work for the firm and help him hire the best tal­ent in the in­dus­try. He ex­pects to make money solely from re­turns on his own in­vest­ment.

“This in­dus­try is ripe for dis­rup­tion in a ma­jor way,” he says.

An­toine An­to­niol / Bloomberg

Muhammed Ye­sil­hark, former head of Euro­pean eq­ui­ties at Carmignac Ges­tion, teaches com­put­ers to pick stocks.

Newspapers in English

Newspapers from UAE

© PressReader. All rights reserved.