The TRADE 68 - Q2 2021 | Page 76

Goldman Sachs

Goldman Sachs Electronic Trading ( GSET ) offers benchmark strategies , including peg , TWAP and VWAP , plus portfolio algos ( APEX ), across global equities markets . The firm currently has an average of 530 institutional buy-side clients per week actively using its algorithms across EMEA . Goldman Sachs has recently expanded its offering to include its full algo suit for MENA markets , which are typically complex to trade . The New York-headquartered investment bank received 42 responses from hedge fund managers in this year ’ s survey , 30 % of which represent clients managing US $ 10 billion or more in AUM . Sixty percent of the hedge fund respondents indicate that they trade at least half of their portfolio value algorithmically . Despite Goldman Sachs increasing its overall average score by four basis points this year to 5.73 - up from 5.69 in 2020 - it fell slightly short of the survey average by 0.09 points , underperforming in aspects of service such as anonymity ( -0.29), customer support ( -0.21) and reduce market impact ( -0.20). Nevertheless , the bank outperformed the survey benchmarks across four key categories , including cost and customisation features . Year-on-year score increases were achieved in eight categories , with the largest improvements recorded in increase trader productivity (+ 0.34 ) and breadth of dark or alternative liquidity sources (+ 0.26 ). Marginal declines were observed in a number of service areas , the most notable being reduced market impact and customer support , both scoring 12 basis points behind the category benchmarks and the latter representing the most significant year-on-year fall for the firm . Moving forward , Goldman Sachs is building out its new strategic trading platform – Atlas - which serves as the backbone of GSET ’ s trading infrastructure , providing flexible , modular architecture upon which products can drive consistent performance improvement across clients ’ execution benchmarks . AXIS – the first algo on the new platform – was launched at the end of 2020 , with additional algorithms scheduled to go live this year .