Share page with AddThis

Magazine Issues » July-August 2018

BUY-SIDE TRADING: All systems algo

FormulasThe world of the buy-side trader is changing. JP Morgan AM has seen its headcount fall, but trade volumes rise – and more managers may develop their own algos, finds Nicholas Pratt.

When the first Markets in Financial Instruments Directive (MiFID) was introduced in 2008, it was expected to usher in a new era for buy-side equity traders where the playing field was levelled, fees were drastically reduced and investors were the ultimate beneficiaries. The directive was also expected to encourage more asset managers to take a greater role in their trading practices and to engage in more algorithmic trading via a roster of new and alternative execution venues.

In reality, very little changed for buy-side traders. Many of the new execution venues ended up either being acquired by the same stock exchanges they were supposed to compete with (such as Turquoise and the London Stock Exchange), or else consolidated among each other (Bats and Chi-X).

Meanwhile, the quality of execution did not necessarily improve, in part because of the relationship between asset managers and their brokers. The suspicion in the market was that buy-side traders felt pressured to use certain brokers’ algos to maintain access to research, IPOs and order flow – all things that have now been separated under MiFID II’s unbundling requirements.

However, ten years on, research suggests MiFID II may achieve what its predecessor could not. Agency broker ITG’s ‘Global Cost Review’ showed that asset managers paid brokers less for research fees and reduced execution rates by close to 30% in the last quarter of 2017. This demonstrates that MiFID II’s unbundling requirements are starting to bite and prompting more asset managers to use algorithms for trading.

“MiFID II has been very positive for buy-side traders because it has removed the conflict of interest that came with prior needs to pay for research,” says Duncan Higgins, managing director and head of electronic sales at ITG.

He adds: “There is now a regulatory rule that requires a focus on execution performance and justifies the need for new technology and analytical tools. They have the levers they need to invest to achieve best execution. Furthermore, MiFID II requires that best execution policies are continually assessed and updated, so I think we will continue to see positive developments without any immediate need for MiFID III.”

For some of the larger asset managers with better-equipped trading desks, MiFID II has had far less impact.

Reduced headcount
“MiFID II has only really changed the best-execution requirements, it has not led to any profound change or driven anyone to electronic trading,” says Kristian West, global head of equity trading and data science at JP Morgan Asset Management (JPMAM).

JPMAM launched an automated trading platform in 2009 and more recently incorporated machine-learning technology to help create real-time ‘actionable analytics’ on another platform. “With MiFID II, you have to have an expected outcome and that takes away a lot of the subjectivity from traders,” says West. “Our platform (Spectrum) takes all of our historical trading data and makes a trading recommendation based on this information.”

The greater use of electronic trading and its internally developed execution analytics has made JPMAM much more selective about the brokers it engages with, says West. “Our commission rates have more than halved in the last five years and our broker selection has become concentrated, but they are much more detailed relationships.”

It has also had a profound impact on JPMAM’s trading desk and its dealers, he adds. “Our trader headcount is down 34% but the volume of trading has gone up – per head it has tripled in the last five years. That direction of travel is likely to continue. On both the buy and sell-side, the order management process is becoming a technology solution and a trader getting involved is becoming the exception.”

As a consequence of these changes, JPMAM has divided its technology team into three functions: order management (building the tools), analytics (managing the data) and automated (executing the orders). “Given the industry’s direction of travel, the vast majority of order flow will be automated in the future. For the first time, we have more people not trading than trading on our desks,” says West.

These changes have not gone unnoticed by senior portfolio managers, he says. “Previously the exchanges between traders and PMs [portfolio managers] were sporadic and based around certain special stocks or situations. With advances in technology now, PMs can raise and execute orders at the touch of a button without any interaction. As a result, that engagement has migrated over time to one that is systematic and technology-enabled.”

As demonstrated by JPMAM’s more concentrated selection, greater use of algorithmic trading will also have implications for brokers. “Any bank or broker executing a trade has to be the best available,” says John Adam, senior director, portfolio management and trading solutions at FactSet, which provides trading data to asset managers. “As a generalised broker, if you’re not in the top tier, you’re going to be very lonely.

“It is a good time to be a large broker because you are also better able to handle the impact of the buy-side paying for research out of their own P&L rather than via commissions.

For many second-tier brokers, they will face the classic dilemma in certain sectors – get big, get a niche or get out.”

In this new, automated and low-touch trading environment, brokers will have to reinvent themselves, says Dan Bassett, head of electronic trading services at fund management software firm Linedata. “They need to provide other services beyond execution like analytics, artificial intelligence and machine learning, corporate access, IPOs and securities lending if they are to remain of value.”

Meanwhile the buy-side traders have morphed into a new role that is half quantitative analyst and half technologist as their use of algorithms moves beyond execution and into areas such as order routing, where they are using their own data to design algorithms that decide where orders get traded, says Bassett.

Brokers are still required to provide access to the exchanges and other points of liquidity but in the future, their role in the trading process could continue to diminish along with any associated fees, despite the fact many have spent millions on their algorithmic trading engines in order to preserve their relationships with asset managers.

In-house algos
Furthermore, brokers will find it increasingly difficult to sell their algorithms and other execution tools to the buy-side, because there is more competition for fewer sales opportunities as asset managers look to consolidate the number of algorithmic providers they use, says Curtis Pfeiffer, chief business officer, Pragma Securities.

“They are looking at a variety of factors - transparency, global reach, asset coverage, execution quality and performance. If you are using 25 providers, they can’t all have equal performance.

“It is also harder to track the performance of 25 algorithm suites as each suite probably has six to ten algorithms. By consolidating, asset managers will then be able to have more meaningful conversations with their remaining providers and be able to customise their algorithms to the point where they are more like in-house designed algos,” says Pfeiffer.

It is a trend that will take some time to properly play out but in the longer term, it would not be a surprise to see more asset managers developing their own algos in-house, he adds. It would give them greater control of the algorithms’ behavior and makes it easier to track execution performance by having direct access to the trading data.

However, the necessary investment in technology may pose some competitive barriers to smaller firms in the short term, says Massimo Labella, head of multi-asset execution sales at GPP, a prime broker. Prime brokers such as GPP are aiming to assist these smaller firms by providing access to an aggregation of algorithms that would otherwise be unavailable to them.

Greater use of algorithmic trading and MiFID II’s unbundling may have changed the relationship between buy-side traders and brokers but only to a small degree, says Labella.

“The market share of the largest brokers is still much the same as it always was but I think this is mostly due to inertia and things will start to change.”

Technology, rather than regulation, will prove to be the great leveller, says Labella, particularly as the cost of trading tools comes down dramatically over the next few years.

However, the difference in the quality of algorithmic execution will also come down significantly in that time.

This raises an important question: if firms are spending more on technology to assess a constantly diminishing gap in execution quality, are they any better off than they were ten years ago or are they simply spending more on their trading process for no real material gain?

Labella is unequivocal in his response. “Buy-side firms are better off now. Certainly the commission compression has been evident over the last ten years and even more so since the introduction of MiFID II. Execution management system providers are multi-broker, which gives asset managers an easier way to route orders to the broker they believe best serves them. In general, technology advances have been incredible and made processes more efficient.”

©2018 funds europe