(Bloomberg) -- Goldman Sachs, once a skeptic of speed for its own sake in equity markets, is doubling down on automated trading.
Long a leading destination for hedge funds where humans pick what to trade, Goldman Sachs hasn't fared as well with quantitative managers, who rely on computer strategies to make buy and sell decisions. Now the Wall Street firm says it's trying to attract more quant funds with a sales effort aimed at achieving a top-three ranking.
Goldman Sachs has a full suite of services to offer quants after upgrading its technology, according to the partners leading the effort, Raj Mahajan and Jeff Nedelman.
"What we found is that that there are real basis points of execution quality at risk," Mahajan, hired in 2015 as head of the equity electronic-execution business, said in an interview last month at the firm's New York headquarters. "We've been investing heavily to ensure that our clients capture them."
That's a big selling point for many quant funds. Speed and efficiency have also become more important for traditional money managers like T. Rowe Price and Fidelity Investments as they look to shave costs and beat back the threat from low-cost index and exchange-traded funds.
Goldman Sachs, which is the lead prime broker to just one of the top three quant funds, is wagering on machines just as some computer-driven money managers run into trouble. BlackRock's main quantitative hedge-fund strategies were on track for losses last year, with four of the five set for their worst returns on record, according to a monthly client update with results through November. The main quant fund at Leda Braga's Systematica Investments lost 11% last year.
And competition is heating up. In late 2014, Virtu Financial, one of a number of upstarts using speedy algorithms to make trades, created a service to execute trades for T. Rowe Price.
"The buy side is the traditional client of these big banks, whether it be quant funds or mutual funds," said Rich Repetto, an analyst at Sandler O'Neill & Partners who covers exchanges. "But now everyone is trying to sell their algos to them."
It's taken a few years for Goldman Sachs to get to this point. In 2015, the same year it hired Mahajan, the firm acquired Pantor Engineering AB to help improve its connectivity to stock exchanges, with much of that work finished last year. Meantime, in a second initiative to improve speed, Goldman Sachs bolstered the plumbing that lets more than 130 systems digest information needed for compliance, risk and auditing on every trade.
In 2014, Gary Cohn, Goldman Sachs's recently departed president, called in a Wall Street Journal op-ed for the industry and regulators to improve the market's structure to reduce the risks from splintered markets and faster trading speeds. It was the same year Michael Lewis's "Flash Boys" spurred a public outcry.
The new strategy signals an acknowledgment that while conditions haven't changed all that much since the firm spoke out against speed, Goldman Sachs has to continue investing in automation. The plan is twofold: First, persuade risk-averse quant funds to move their business to Goldman Sachs by convincing them the firm has improved technology. Second, work with existing clients who are already using strategies carried out by humans — like fundamental-equity mutual funds, or market-neutral hedge funds — as they explore the quant space, Mahajan said.
Among the bank's offerings are a selection of stocks that can be sold short. With inventory tight, the supply Goldman Sachs amassed as one of the top prime brokers to long-short and market-neutral hedge funds is an attractive asset, Mahajan said.
The bank also offers access to trading flow that's attractive to quant managers, Nedelman said, because it typically gets matched among clients internally rather than pushed onto an outside exchange that charges additional fees. That could allow Goldman Sachs to reduce or eliminate the costs for clients. On any given day, the firm sees between 8% and 12% of U.S. equity market volume, according to Mahajan.
For other clients, the company has asked its sales people to act as advisers on how best to allocate trades, said Nedelman, who ran equities sales for the Americas until September, when he took over as global head of prime services. They've also given salespeople analytics tools that monitor trading orders in real time.
Goldman Sachs wouldn't say how many competitors it needs to leapfrog to meet its market-share target, and industry metrics are difficult to come by. That means there may be no good way to independently gauge the firm's success.
By one measure, it's losing ground. Morgan Stanley, the No. 1 bank by equity-trading revenue, widened the gap with No. 2 Goldman Sachs through the first nine months of 2016, bringing in $766 million more than its rival. In 2015, Morgan Stanley posted $8.1 billion in revenue to Goldman Sachs's $7.8 billion. Goldman Sachs was the leader for 2013.
The two biggest quant funds rely on Goldman Sachs's competitors. Two Sigma, with $40 billion in assets, uses Deutsche Bank AG, according to regulatory filings. Renaissance Technologies, the $36 billion fund founded by Jim Simons, counts on Deutsche Bank, JPMorgan Chase and Morgan Stanley as its main prime brokers.
Goldman Sachs's largest such client is Winton Capital Management, a London-based firm that manages $33 billion, filings show. The bank also counts Engineers Gate, a quant prop-trading firm that doesn't take outside money, as a client.
"Those kind of clients are looking for us to be important to them in the space where we might not have been," said Nedelman.
For quants, reliability and ease of use is paramount, according to John Feng, a managing director at Greenwich Associates. Quant funds are often reluctant to move their business to a new firm for fear of computer glitches or other operational mistakes. As a result, Goldman Sachs faces a bigger hurdle but also a greater opportunity, he said.
"The entry barrier for opening the door could be a little higher for those types of funds," Feng said. "It could also mean that once you are in, the long-term payoff may be well worth the effort."