Capturing liquidity at Jefferies

Ben Springett, European head of electronic and program trading at Jefferies, tells The TRADE about new algo strategies the bank has rolled out and the impact the market volatility in 2020 has had on the ELP SI landscape. 

Buy-side traders ranked Jefferies the standout dealer in equity coverage during the most volatile period of 2020 as part of a US flash study from Coalition Greenwich that queried brokers’ performance in providing liquidity, hedging solutions and market insights.

Coming out on top of major market players and rivals such as Morgan Stanley, Goldman Sachs and JP Morgan, the results reflected Jefferies’ significant expansion in recent years, particularly in the electronic trading arena. Since the appointment of Bill Bell, a former global head of equities electronic distribution at Barclays, as the investment bank’s global head of electronic trading in 2014, Jefferies has been building out its electronic platform and gaining market share.

Ben Springett, European head of electronic and program trading at Jefferies, tells The TRADE that the platform built out by the team is flexible, modular and able to adapt to market developments. As the number of trading venues and order types has increased over the years, he adds that client customisations and tools such as adaptive algorithms have become even more important.

“Our focus is the long-only and hedge fund clients, particularly those that have quantitative assessment of their trading,” Springett says. “Being measured based on our capabilities and performance as a relatively newer entrant and as a firm that has been taking market share from existing players has been beneficial for us.”

Algo wheels and automation of flow is a significant part of how Jefferies interacts with its clients. The bank has worked on expanding its algo stack in response to growing demand from the buy-side. It has rolled out adaptive algorithms, for example, that provide greater flexibility and are responsive to both the current trading environment and their own success.

Using a dynamic scoring system that determines a rank of venues a client may want to use based on factors such as toxicity and information leakage risk, the adaptive algo trades with venues considered benign to capture sufficient liquidity but can expand its search for liquidity according to the urgency of the order.

Another such strategy developed by Jefferies targets the close, the most important liquidity event of the trading day. The adaptive algo, known as TOUCHDOWN, behaves like a trader looking to take advantage of the liquidity and volume at the close without moving the stock too far. It is responsive to the indicative price and size, monitoring the impact on the market to optimise trading outcomes.

“When we receive an order from a client, we have an objective around capturing liquidity,” Springett explains. “We are all about liquidity capture, but we also care about reversion and adverse selection associated with those trades. Having that more dynamic approach to trading has been beneficial to the performance of those strategies and it has been well received by clients. It will continue to evolve and continue to be a focus for us.”

Put your money where your mouth is

As part of developing its electronic platform, Jefferies has integrated more Level 3 data to crunch within its quantitative processes, algo construction and performance analysis to better inform trading decisions. Level 3 data goes beyond top of book and full order book depth, providing information on individual orders at certain price points.

Capturing, tracking, and manipulating that data, however, requires significant processing power. Jefferies has teamed up with analytics provider BMLL Technologies and deploys cloud computing capabilities to carry out detailed analysis, allowing the team to understand on a granular level what happens to the order book when they trade passively or aggressively.

“We’ve been doing that analysis for a very long time, but it’s now at a far greater level of granularity,” Springett says. “We can see which orders move when we cross the spread, for example, rather than knowing that some orders move. We can track those and see where they came from. It enables much smarter decisions and using that within the algo constructions process is a huge priority for our team.”

On the program trading side, Jefferies has also recently developed PT Match, a new product that has been piloted with a number of clients and is expected to be rolled out more widely over the course of this year. PT Match allows traders to send a basket into an internal crossing engine that is matched off automatically against liquidity within the pool. Orders that can’t be matched are returned to the client and remain anonymous, giving traders the confidence to continue working the basket with the program trading desk or in some other way.

PT Match was born out of conversations with clients on the program trading side, which Springett says has worked the same way for the last 20 years or so. The buy-side receives vast amounts of information from brokers on busy sectors or countries, or average daily volume (ADV) profiles on any given morning and they need to decide which broker to send a basket to. It’s considered a balancing act for brokers who can’t share too much information on client order flow but are looking to attract liquidity for clients.

“It’s a put your money where your mouth is type of arrangement and I think that’s why clients like it because they do wrestle with that decision on where to place baskets,” Springett says. “The beauty of PT Match is that it becomes self-fulfilling. As you bring more liquidity onto the platform, we can offer higher fill rates to the next orders, and then the next orders and so on. The advantage to clients is that it’s fully anonymous and there isn’t a requirement to send the basket to a sales trader to see what crosses they can produce.”

Permanent impact

Looking back at 2020, Springett says the volatility effect during that period encouraged a change in execution approach for many buy-side traders, which coincided with the relocation to working from home and various technological limitations.

As the immediacy of liquidity became more highly in demand for clients, he notes more use of high-touch trading and capital within that space, alongside increased adoption of urgent liquidity seeking algos on the low-touch side. There was a more even distribution of volume across the trading day without the usual spike at the end of the day as traders looked to execute quickly amid intense volatility.   

“As orders were hitting the buy-side trading desks, they were looking to get them done as fast as possible as opportunity risk is much greater in that scenario,” Springett explains. “Trading is a trade-off between impact, cost and opportunity risk. If you ramp up opportunity risk, then you shorten your trade duration and are less sensitive to 10, 15 or 20 basis points worth of impact because the stock could drop in the next hour.”

Data tracked by Jefferies also revealed a decline in systematic internaliser (SI) share of market volume overall, which is yet to recover. SI volume is considered opaque in terms of the type of trade and provider, but drawing a picture of the landscape Springett witnessed a pull-back from electronic liquidity provider (ELP) SIs that were willing to provide quotes to Jefferies’ algorithms during the volatility. The impact was a significant reduction in the number of ELP SIs that Jefferies now trades with, declining from eight prior to the pandemic to just two.

ELPs work with risk-driven models that have certain tolerance for value at risk and as volatility increases smaller positions can generate the same value at risk, leading to a fall in quote sizes. Springett adds the volatility caused ELP SIs to not only vastly reduce their size, it also caused some to suspend operations in the space entirely.

“I think that is permanent,” he explains. “It was an interesting period because prior to that we were still very much in the post-MiFID II SI growth phase and they were clamouring for business. We were connected to eight ELP SI providers before, but with the volatility a number of them have pulled out of the market completely, or meaningfully reduced the amount of business they are doing.”

Elsewhere, Springett highlights an evolution in high- and low-touch trading with the emergence of a more hybrid model gaining traction with some buy-side clients. While there are those that still desire a distinct line between the businesses, others are seeking more information from the low-touch trading desk on their orders which was previously more aligned with the high-touch service.

There are a number of different buckets that clients fall into with regards to their willingness to engage in those conversations around the types of crosses that Jefferies may look to do. If the team has two algo orders, or a high-touch order and an algo order, for example, clients may be open to having a conversation rather than just meeting somewhere in the dark.

“Increasingly, we are seeing a more hybrid high- and low-touch model come through,” Springett says. “There’s a lot of different flavours of where we could end up with that. For us, one of our big focuses across the floor at the moment is figuring out the different workflows and permutations you could have under these scenarios and building out a mid-touch service for the clients that would value it. We think it’s going to be a very important part of our service provision as the business evolves over the coming year or so.”