A Goldman Sachs sign is seen on at the company’s post on the floor of the New York Stock Exchange.
Brendan McDermid | Reuters
Goldman Sachs is taking aim at the world’s biggest quant hedge funds.
The investment bank recently approved a three-year plan to spend more than $100 million to overhaul its stock trading platform, according to Mike Blum, a Goldman partner and chief technology officer of electronic trading.
The project, named Atlas after the Greek God, is meant to accelerate the shift Goldman has been making since realizing in 2014 it was falling behind in Wall Street’s equities technological arms race. Huge quant funds like Renaissance Technologies and Two Sigma are among the most demanding of clients from a technology perspective, and competitors including Morgan Stanley and J.P. Morgan have been jockeying to serve these money managers.
“With this investment we’re trying to tackle the quantitative hedge fund space and do so front-to-back to create a seamless experience for our clients, and just try to get as efficient as they are at doing their jobs,” said Blum, a 25-year electronic trading veteran who joined Goldman in 2017.
Stock trading has become increasingly cutthroat as firms including Goldman, Morgan Stanley and J.P. Morgan have been winning a larger share of a shrinking pie. These three banks made $11.4 billion in stock trading revenue so far this year, 14% lower than 2018. All three have been investing money in the latest electronic trading technology as more of the world’s volumes shifts to screens.
Going after the big quants
With Atlas, Goldman is targeting more than a dozen quant funds, meant to represent a range of trading styles, on the theory that if they can satisfy this group, they can satisfy any type of hedge fund or asset manager client.
Source: Goldman Sachs
The project takes the technology in its core trading platform, designed for speed and reliability, and extends it to 32 markets around the world and in other functions, including clearing and settling trades, allocating stock, lending shares and trade reporting, Blum said.
“As we learned the quantitative client base and what their demands and needs were, we decided to take the technology and basically turned it into a framework that can be used to solve lots of different problems,” said Blum.
For the entirety of his career, which began at tiny South Carolina-based Automated Trading Desk in 1993, Blum has been working on and improving his approach to electronic trading.
The design philosophy that he and other e-trading pioneers have landed on: Using so-called microservices to break complicated problems into easy-to-solve ones, he said. With that, as well as a developer strategy known as event sourcing, Goldman’s system has become fast, customizable and reliable.
One goal was to improve the firm’s performance on trade orders that last a microsecond (a millionth of a second) to about 30 seconds, shorter-lived orders that the bank may have struggled with in the past, Blum said. In some instances, trades that took hundreds of milliseconds will happen faster than 100 microseconds, he said.
Some quants also send hundreds or thousands of trades simultaneously, and Atlas will be able to compress the time the bank takes to digests those orders, he said.
“As we roll it out globally, they should absolutely see a massive speed improvement,” Blum said. “They should see quality of execution go up, not just because of speed, but we’re completely rewriting our algorithms, we’ve brought in more researchers, more quants to improve our algos.”