At the peak of the market volatility in March and April, Goldman Sachs Electronic Trading (GSET) saw volumes 135% higher than the average trading day, with market data volumes processed by the platform almost tripling overnight.
Several years of intense investment across the GSET franchise was about to be verified by arguably its most rigorous test, as one of the greatest liquidity events unfolded in global markets.
From an eerily empty Goldman Sachs office building in the heart of the City of London, Daniel Mallinson, head of EMEA execution services, alongside co-heads of EMEA electronic trading, Alex Harman and David Cornish, tell The TRADE that throughout the COVID-19 crisis GSET has gained market share and broken several records multiple times over.
“The resilience and reliability of our technology stack meant we had the ability to process those extraordinary volumes consistently during that period, and that was certainly a huge factor in us taking market share,” Mallinson explains. “The consolidation amongst the top five brokers was exaggerated during the extreme volatility. I believe a lot of that, in contrast to prior crises, was reflective of banks being able to provide a resilient operating model.”
As the crisis unfolded and volatility peaked, reports of gaps in liquidity and a lack of reactiveness from sales traders soon began to surface from buy-side traders. Some banks and brokers had no choice but to switch off entirely and pull back from the market, leaving many traders frantically adjusting strategies and seeking liquidity on-screen.
Similarly, the GSET team were conscious of the lack of liquidity in the marketplace and the difficulty clients were having in executing their flow.
“Clients were much more vocal in what they wanted and needed from us instead of relying on the systematic aspect of the relationship,” Mallinson adds. “The ability to stay connected, whether that was on the telephone, Zoom, Skype or Microsoft Teams, was unbelievably important.”
Specifically, Mallinson notes that as volatility spiked clients experienced a decline in provision of principal liquidity. Conversely, over the last 12 months, Goldman Sachs has systematically provided $300 billion in principal liquidity globally, saving clients an estimated $50 million in execution costs in aggregate.
“We over communicated our willingness to consistently be there for clients across the full spectrum of product,” he adds. “We didn’t go into the volatile environment with 80% of the franchise available, and say the other 20% will be back when things are easier for us. Everything remained available.”
Resource and investment
Over the past three years, Goldman Sachs has doubled its market share in European execution. It now trades one of every seven dollars in the marketplace.
Without a doubt, the scale of the Goldman Sachs franchise will have lent itself to the challenging period the industry has just been through, and proved to be a significant advantage during the crisis. The other advantage is resource and investment.
Since 2015, GSET has been completely overhauled. Harman, who has been with Goldman Sachs for the past seven years, says the plan was to revamp the entire platform and front-to-back infrastructure, including exchange gateways and operations, not just the smart order router (SOR) and algorithm suite.
Although it began with the SOR, which has evolved into a low-latent, technology-led routing system that was pivotal to GSET after the European liquidity landscape was transformed by MiFID II. Upon its introduction in 2018, MiFID II saw the number of venues shift from around 15 to now more than 50 in operation across the region.
“It wasn’t just the number of venues increasing, it’s the increased complexity of the venue types we now have in Europe,” Harman explains. “For example, periodic auctions, electronic liquidity provider (ELP) systematic internalisers (SIs). All of these mechanisms are reliant on data, speed, smart technology, and the ability to connect quickly.
“We’ve broken a number of records recently, and clients are encouraged by our progress. GSET has broken market share records in nine out of the last 10 quarters, and the number of clients trading each week has grown 87% over the past two years. The recognition at The TRADE’s Leaders in Trading 2019 awards in November makes a difference too. The investments made in the years prior were truly tested during this period, and we didn’t let any clients down. It has further validated our progress.”
Cornish, who also heads up EMEA quantitative services distribution in Europe at Goldman Sachs, adds that in terms of the market share gains seen in 2019, more than half came from quantitative hedge funds.
He joined GSET around two years ago to build out the quant segment, an area where Goldman Sachs has historically underperformed compared to its peers. Similar to quantitative funds seeking direct use of GSET’s SOR, he says the development of Goldman Sachs’ Algorithmic Portfolio Execution offering (APEX) has been another area of interest for quants.
“Our APEX portfolio algorithm was built predominantly for asset managers and passives, but it is already being used by a number of sophisticated quant funds that are looking at it as a potential differentiator to the more traditional volume weighted average price (VWAP) strategy,” Cornish says. “It’s that osmosis effect of investment.
“Another area is in the post-trade. Something that is considered essential for the systematic community is resiliency and straight through processing (STP). Our strategic operations platform has been built out and we now see 99.9% STP rates, with massively hectic days seeing north of 99%. If you go back 10 years and adjust it to the volume we have today, we would have seen around 60%. It’s a huge difference in our capabilities that will benefit all clients that do business with us.”
The GSET overhaul has not been all about investment in the platform and infrastructure. It has also focused on boosting human capital and the team behind GSET has grown exponentially in recent years.
The number of sales traders, for example, has increased 250% since 2014 amid a huge rethink as to how each team can add value to the clients they service.
GSET has created new and dedicated divisions for areas such as algo consulting, market structure, product, acquisition, and marketing.
The algo consulting team, which the co-heads of electronic trading agree has been key, forms part of the wider coverage team working closely with Goldman Sachs’ institutional clients, who themselves, have expanded their quant and research teams.
As part of efforts to boost market insights for clients, in January 2019 Goldman Sachs hired 15-year Morgan Stanley veteran and former head of market structure, Ellie Beasley, to lead the new-look market structure team.
Another high-profile hire has been industry veteran Michael Steliaros, previously the global head of quant trading at Bank of America, who has been working hard on developing the execution research team after joining Goldman Sachs in 2017.
Providing market insights has been pivotal for the team during the crisis and particularly at the height of the volatility. Mallinson says that GSET published more than 50 different pieces of liquidity analysis between February and the end of June this year, delving into various aspects of the landscape such as spreads and top of book data. He adds that the team managed to get the analyses to clients swiftly as events were unfolding.
“Gone are the days of having one sales trader as the only relationship for the institutional and hedge fund accounts,” Harman explains. “Now, it’s four or five key people across execution that can really be there for clients to give them essential resources and help them execute efficiently. Along with their execution coverage, the market structure, index research, and execution analytics teams are all now essential relationships and services for our clients.”
With each of these newly-developed teams coming together, the focus on front-to-back services comes more clearly into focus. All of the teams that form GSET are coached on the fact that no two clients use the platform in the same way. Despite the emphasis on the front-office, combining all of the resources and teams to provide a more bespoke service to each client has been an important part of the new GSET strategy.
As GSET focused on providing timely access to market insights along with continued provision of liquidity at the height of the volatility, market participants experienced an unprecedented surge in execution costs. Estimates from GSET revealed that the cost of execution for clients rocketed fivefold in some cases during the most volatile periods.
“I don’t think we have ever seen a fivefold increase in the costs of execution over such a short period,” Mallinson says. “Suddenly the traders, who are scrupulously measured by their own organisations, were having to justify that increase in cost.
“Our liquidity access enabled us to actually be one of the cheapest providers. All of the content that the team provided also helped to explain why costs were going up week-by-week, and how Europe costs compared to the US and other markets. Switching from the objective to the subjective side of the business, I would say that our ability to hand-hold our clients through that period was of great importance.”
Upon outlining some of the execution trends, Harman says that GSET noted a shift in the way market participants were trading as the environment became increasingly volatile. There was, of course, an uptick in volumes as it became more difficult to trade and market participants became more urgent. Long duration-type strategies began to decline, as more urgent liquidity seeking strategies increased.
On an average day, the GSET platform trades around 35% in short duration liquidity seeking strategies, and 45% in long duration strategies. During the more volatile period, however, Harman says this was essentially flipped on its head.
Harman also notes interesting findings across the ELP SI landscape. He says GSET holds one of the largest market shares and connectivity to ELP SIs across the street.
In a typical environment, ELPs typically quote three or four times larger on the touch side compared to the available liquidity in the lit market. GSET noticed that despite ELP touch size quotes slightly decreasing at the peak of the volatility, that decline was minor compared to the sharp decrease in available touch liquidity in the lit market.
That resulted in touch liquidity quoted by the ELPs being five or six times larger than what was available on the lit touch. Harman outlines that having connectivity to and available liquidity in the SOR was crucial during the most intense days of trading, especially as urgency to execute was elevated.
“In [times of] high volatility the bid-offer touch size decreases and that was a big problem for the street,” Harman explains. “In the STOXX Europe 600 in the second quarter, for example, it declined from about $55,000 on average at the touch, to around $18,000. “We found that whenever we wanted to use the ELP SIs, we could get more liquidity. For our clients, trying to find that liquidity is all about who has the connectivity, and those important bilateral relationships with the ELP SIs were absolutely key for GSET.”
Additionally, GSET noted an uptick in clients using its stock clustering data service, which is a stock classification data source that feeds information to the platform’s algorithms to educate on how best to trade a stock. That data can now be fed through an application programming interface (API), meaning that clients using tools such as algo wheels can plug the data directly into their own system.
Data that fed algo wheels when they were first developed was fairly basic – mainly ADV, notional or a static list of securities. Clients are now incorporating the Goldman Sachs clustering data into their platforms to enhance decision making for what should or should not be in an algo wheel, or which execution style to use.
This has been available to Goldman’s clients in Asia for three years and those based in Europe for around 18 months. Take up of that data service has been enormous over the past year, Harman Harman explains. In the past 12 months, more than 60 institutional and hedge fund clients are now utilising the API for their algo wheel decision making.
“All of the investment, and this underpins everything, is forward-looking,” Mallinson says. “It’s not designed to be a band-aid solution to a problem that we might not need next week. It has all been designed to accommodate what our clients are genuinely going to need in the coming years. That has been a key part of GSET’s market share win. The forward-looking view is focused in a way that when the market structure does evolve, our product is ready to accommodate the new environment.”
With this forward-looking view in mind, GSET also targeted its capabilities in helping clients trade emerging markets. While a clear area of opportunity for many market participants and brokers, emerging markets are still considered difficult to trade.
The focus on emerging markets at GSET was client-driven, and the aim was to ensure that clients could trade emerging markets just as easily as developed markets.
The business treats the emerging markets platform and developed markets platform as though it were a single platform, providing front-to-back capabilities that allows users to navigate those markets considered more challenging.
“Our emerging markets franchise now has the same market share rank as our developed markets franchise, which is a huge win for us relative to where we would have been two years ago,” Cornish explains.
“Being able to provide the full-suite of services is incredibly important. It is not just about the execution platform – it’s also about stock loan, synthetic, foreign exchange, settlement, and operations. We are piecing all of that together, automating it, and making clients feel comfortable that they can trade Kuwait just as easily as they can trade France.”
Place your bets
Despite the market share gains and progress that GSET has made in recent years, Mallinson emphasises that the job is nowhere near complete. Given that Goldman Sachs has a unique vantage point on the world due to its scale and resources, an important aspect of the ongoing investment in GSET is placing bets on what its clients will need in the future.
Similar to emerging markets, which remains a significant bet for GSET, the team sees opportunities in exchange traded funds (ETFs) strategies. Like the US markets, clients in Europe are seeking more ways to trade ETFs on-exchange, alongside the more typical request for quote (RFQ) strategy.
With this in mind, GSET is preparing to launch new strategies this quarter for clients to trade ETFs more effectively both on-exchange and via principal through its algorithms.
Perhaps unsurprisingly, the closing auction is another GSET wager. Closing auctions have increasingly become the most important part of the trading day. The percentage of full-day volume executed at the close has soared from around 15% around five years ago, to nowadays often more than 25%.
Goldman Sachs was one of the first banks to go live with a systematic product for its passive clients that guarantees market on close, known as GMOC. It has been live for around eight years in the US and five years in Europe.
Building further on this, GSET has developed an advanced target close strategy and a portfolio algo focused on the close benchmark amid increased demand from institutional clients for basket-type strategies.
“It is our job to make these bets based on what we think our clients are going to need in the future. We want to build around those bets and advise our clients on exactly why that’s the bet we are making,” Mallinson says. “We make no secret of what those bets are – we are fully open. It’s as important to our culture as it is to our business strategy.
“From a top down point of view, principal liquidity is a big bet for us. A vast amount of that principal liquidity is linked to the products and strategies we are preparing to roll out.”
For quantitative funds, Goldman Sachs is spending upwards of $100 million on developing a new high-speed equities trading platform, known as Atlas, which has already inspired the development of other products under the GSET franchise.
For example, with the continued focus on short duration strategies, the team recently rolled out a new short duration algorithm, named AXIS, based on the new Atlas platform.
“Clients have been asking how they can execute in a tightly defined duration passively,” Cornish explains. “AXIS is a new signal-driven strategy optimised for that short time horizon, which is really interesting to quants, as well as a host of other client types. It is all about the increased use of alpha trading signals and reducing signalling and market impact by leveraging the raw processing power and speed of our Atlas platform.”
While there has been a clear move by GSET to maximise the opportunity laid out by the global developments this year, the recent growth in market share and increasing engagement with asset managers and quantitative clients has, in many ways, been years in the making.
Mallinson, Cornish, and Harman are confident they have made serious headway in the resurgence of GSET, and interestingly, the current operating environment certainly lends itself to further consolidation and further market share gains in the near future.
“I don’t think I’ve ever seen an opportunity as good as this for running an execution platform – it’s hugely exciting,” Mallinson concludes. “We are achieving market shares that we could only have dreamt of in recent years and I see no reason why that can’t be replicated across our entire client base.”