Daily Management Review

LCH: Computer-controlled hedge funds lead the game


Artificial intelligence has once again demonstrated its power. Hedge funds based on algorithmic trading system are leaders of efficiency ratings.

Top 20 hedge fund manager list, compiled by LCH investments (invests in other hedge funds and included in Edmund de Rothschild Group), features DE Shaw, Citadel and Two Sigma. All three companies use so-called "systematic strategy" (a common term for decision-making strategies with minimal human factor). Nonzero predictability of market fluctuations is detected quantitatively by statistical processing of large volumes of observations on market instruments’ behavior.

LCH says that the three newcomers, together with Pure Alpha that also employs a systematic element, have earned about $ 90 billion over the past 10 years.

LCH’s decision to include systematic hedge funds in the rating demonstrates prevalence of modern quantitative trading strategies in hedge funds’ activities. The industry is forgetting exclusive use of discretionary accounting methods of systematic trading and analyzing large amounts of data. "The results clearly illustrate increasing technological possibilities of investment systems" - says Rick Sopher, chairman of LCH.

According to The Wall Street Journal, Bridgewater hedge fund (won first place in the rating) also employs an automated control system based on artificial intelligence technology. The fund’s owner Ray Dalio predicts that the system will soon perform a variety of tasks from short-term investment to hiring and optimizing operations.

The hedge fund’s founder of predicts that artificial intelligence will take three of four administrative decisions in the near future. David Ferrucci, who previously worked at IBM, is in charge of software development.

Hedge funds based on algorithmic trading system experienced a reputational hit in 2007 as many popular strategies did not work and brought losses to investors. However, quant hedge funds were the best to adapt to the new reality after the 2008 crisis, whereas their somehow old-fashioned colleagues were struggling to win the market. Inflow of funds to algorithmic trading system hasn’t seized for the seventh year in a row. In 2009, it amounted to $ 408 billion, and numbered $ 879 billion in 2016 (Hedge Fund Research’s data).

Not all operations in such organizations are performed by computers. For example, hedge fund Two Sigma Investments ($ 24 billion under management) employs dozens of non-financial professionals - astrophysicists, immunologists, linguists. They process information from news feeds, financial reports, weather reports, and social networks. Algorithms are written on the basis of data that predict movement of quotations.

Text-mining algorithms read financial and weather reports (required for trading of commodity derivatives and commodity sector stocks). The goal is to translate this information into trading signals and to become first to detect signals of changes. Two Sigma said that the forecast accuracy is not 100%, yet it is still enough to make a profit.

For example, in order to understand reasonability of buying shares of a major retailer, the algorithm checks whether the company’s quotes stay within average, and see if management keeps holding their papers. All this is added by a conclusion of risk management, which checks if the company’s risk does not exceed permissible level. After all examination, the computer makes a deal. In order to monitor data 24 hours a day, Two Sigma has more than 100 teraflops of computing power capable of performing more than 100 trillion calculations and 11 petabytes of disk space.

source: ft.com