This post first appeared on TabbForum - Where Capital Markets Speak
Protect and Serve an Advancing Front Office
Both buy-side and sell-side market participants have been investing heavily in algorithmic trading processes and technologies. The opportunities presented by algorithmic (algo) trading from off-the-shelf components to highly optimized machine learning routines and AI augmented trade flow is expanding the algo frontier. What was initially a high-frequency trading play has evolved to integrate operational efficiency with customer service. The technological innovations that are advancing in the front office must be extended to the middle and back office to ensure that the risk stewards of the organization can efficiently protect and serve the best interests of the business.
Sell and Buy-Side Advantage
Bank holding companies and asset managers have developed sophisticated computer models that provide better pricing, execution, routing, and hedging tools for both their clients and for their own trading desks. The buy-side has led the algo driven frontier with systematic trading firms like AQR, MAN AHL, and Point72’s Cubist Systematic Strategies. According to a recent Wall Street Journal article, about 31% of all investing is done by these kinds of investors.[1]
Sell-side bank holding companies also benefit from algo trading in at least two ways:
First, regulations prohibit banks from proprietary trading and curtails their ability to gain directly through market activities. Algo trading enables them to eke out profits from ever tightening bid-offer spreads by rapidly trading on infinite combinations of assets, price, volume, and market indicators.
Second, algos enable the banks to serve and satisfy their buy-side and institutional clients faster by providing new and automated products that are tailored to their clients’ investment strategies. This generates new revenue on services while improving client retention.
Move Beyond Defense
If properly designed and continually monitored, a machine can price, execute, and route an asset that is aligned to a client’s priorities faster than a human and with minimal error. But if algos are not properly designed and tested to handle abrupt market changes and regulatory transformation, they can lead to larger losses and regulatory fines at a faster rate than a human could do on their own. Traditionally, there are three lines of defense to mitigate risk like these.
The front office is the first line of defense. They build and run the algos with “circuit breakers” that pause operations in case of market stresses or unknown market behaviors. Algos are “encased” with limits and parameters that specify the quantum of financial risks within which they are allowed to operate. Algos are also continuously monitored by “algo traders,” who can take control as conditions require.
The risk and compliance teams are the second line of defense. They are the risk stewards whose job is to challenge the front office, ensuring that the businesses are operating within the bank’s approved risk appetite.
Auditors are the third line of defense who provide independent assurances that the first and second lines of defense are managing and monitoring risk appropriately.
As algos take over the trading floor, risk stewards are saddled with new challenges that can’t be met with a traditional defensive posture. A proactive approach is necessary that extends front office innovations to the middle and back office so that the risk stewards have the data and tools they need to effectively challenge and advise the front desk.
Engage the Clients and the Risk Stewards
Even with recent technological innovations, human relationships continue to drive the business, especially as client expectations for transparency continue to rise. Some frank and honest discussions between the sell-side and their buy-side clients reveal nuances that improve algo design and increase share of wallet, while other conversations weed out accounts that don’t meet minimum revenue requirements. In each case, the front office has the tools and the data to support either type of conversation and to defend decisions that are challenged. This level of engagement needs to be extended to the risk stewards, both to protect the bank and to meet regulatory requirements.
Risk stewards are expected to be independent from the business. Not only do they require independent knowledge of the markets, they also need the tools to independently monitor, evaluate, and challenge front-office activity. The technology gap between the front-, middle- and back-office systems should be narrowed as there is room for improvement on systems that traditionally operate in an old-school batch file world.
In order to have clear visibility of activity and to better advise the front desk, the risk stewards require three things:
Their own set of real-time risk and compliance algos that independently monitor trading.
The ability to replay and scrutinize suspect trading scenarios at any point of time.
Transparency into algo design, controls, and testing to ensure they are operating within approved practices.
Risk Needs Automation Too
Here is a conceptual view of how the points described above might be covered. First, while the front desk has their pricing, hedging, and execution algos running on live market data, the risk stewards would be afforded their own set of trade surveillance, market risk, credit risk, and permitted products algos that run on the same live market data. The machine learning routines would monitor trading activity in real time, analyzing and evaluating each event and alerting the risk stewards as needed.
Potential technologies that could enhance the aforementioned alerting mechanism would be artificial intelligence (AI) capabilities such as Natural Language Generation (NLG) that provides a clear and concise narrative of the issue and recommendations for next best action. Additionally, Robotic Processing Automation (RPA) could be used to automate data collection and escalation procedures.
At the point of alert, the risk stewards would use next generation big data tools and platforms that provide on-demand data access, modeling, and analytics. This would enable them to replay suspect events that could occur at any time and to evaluate those events against other real-time or archived data sources. The sheer amount of data generated by the algos is the biggest challenge of the expanding frontier, so an economic balance must be maintained between immediate access and archived storage of real time generated data.
The new levels of insight obtained above would improve algo design because the data results could be peer reviewed faster and more often. But an additional proactive level of improvement could be had by enabling the front desk and the risk stewards to share a machine learning platform that facilitates a collaborative approach to building, reviewing, testing, iterating, and deploying algos into production.
Magnified Risk and Missed Opportunities
The automated nature of algo trading can improve business outcomes but magnify risks. If risk stewards do not have the right tools to effectively oversee a business that is running faster and faster, operational losses, regulatory fines, and oft-occurring reputational risk could increase. There could also be missed opportunities because the invaluable subject matter expertise of the risk stewards would be left off the table.
Front office automation reduces cost and affords better customer service. This is how the business runs nowadays and it can’t be slowed down to accommodate outmoded risk and compliance technologies. The next generation front-office technologies that are being presented to clients also need to be extended to the middle and back office so that the risk stewards can more efficiently and effectively protect and serve the best interests of the organization
Paul Lashmet is a business integration architect with expertise in orchestrating global strategic programs across the financial business landscape. His company, North Castle Integration, creates opportunities for business integration, optimization, and new revenue streams by matching business challenges with fit for purpose innovative technologies in artificial intelligence and data integration.
[1] The Wall Street Journal. “Point72’s Head Quant Is Leaving the Firm” (March 2020)