Balancing Regulatory Obligations and Operational Efficiency

As an expert participant in a thought-provoking debate at the APAC 1LoD conference, I had the privilege of engaging in a lively discussion on ‘Leveraging surveillance technology and AI to increase efficiency’ in the realm of trade surveillance.

Together with three other distinguished experts, we delved into the intricacies of this rapidly evolving landscape, shedding light on the challenges and opportunities that lie ahead for financial institutions.

One of the key focal points of our discussion was the existence of two primary technology approaches to trade surveillance: scenarios and models. These approaches, both readily available through various vendors, offer distinct functionalities and implementation strategies. Scenario-based surveillance necessitates risk-specific configurations, with each potential risk requiring its own dedicated detection setup. Conversely, model-based surveillance adopts a more holistic approach, detecting a range of risks simultaneously.

While both approaches have their inherent benefits and challenges, it became evident during our debate that transitioning between them is not a trivial undertaking. Surveillance teams must be meticulously set up, trained, and staffed differently based on the chosen approach. This implies that pivoting from one approach to the other can be a time-consuming, costly, and overall complex endeavour. Organizations must carefully assess their unique requirements and available resources before committing to a specific surveillance strategy.

For me, one of the surprising elements was the poll results on this topic, where less than 20% of the audience confirmed that their surveillance team had already deployed AI in their trade surveillance program. With several vendors in the market who have specifically developed trade surveillance tools with AI models to detect risk, this demonstrates how novel this approach still is and how little of the market share they have claimed so far. A further surprise was that 33% of respondents confirmed they are not even looking at AI at this stage as part of their program or strategy for improving trade surveillance. This leaves about 50% of the market in the process of looking, assessing and considering the role AI might (or might not) play in their program.

Another much-discussed topic that emerged was the proliferation of trading venues and the formidable challenges teams encounter in ensuring comprehensive and accurate coverage of these venues within their trade surveillance programs. As the number of trading platforms continues to expand, surveillance teams find themselves grappling with the intricacies of data integration and normalization across multiple venues, striving to maintain a robust oversight framework that encompasses both internal teams and their data compliance but also vendors who are not incentivised by the compliance data needs from their venue.

Interestingly, our debate also ventured into the critical distinction between "risk acceptance" and a "risk-based approach" when banks evaluate their trade surveillance coverage, especially in the context of new venues. Often these new venues start as supporting low volumes, a small number of traders and may not have clear revenue contributions to the business established yet. So the discussions about ‘risk acceptance’ are opened, as the business seeks to trade where surveillance is not deployed. The concept of "risk acceptance" raises pertinent questions to address immediately, such as the nature of the risk being accepted, the individual or entity accepting it, and the extent to which the accepted risk deviates from the bank's risk appetite and regulatory obligations.

For me, "risk acceptance" should be regarded as an interim process for capturing elements that fall outside the scope of the established surveillance program and are not yet integrated into the business-as-usual (BAU) operations. There may be valid justifications for this, such as the introduction of a new trading desk, low trade volumes, or the high cost of integrating a specific platform. In such cases, banks may be assessing the long-term viability and revenue potential of the business before fully aligning it with BAU processes. However, it is crucial to emphasize that "risk acceptance" should not imply a complete absence of controls, surveillance, oversight or be misconstrued as a "risk exception." Instead, it should be viewed as a transitional step towards a longer-term strategy that may involve evaluating additional headcount, deploying new technologies, and allocating appropriate budgets.

On the other hand, a "risk-based approach" entails banks meticulously assessing risks against available budgets and resources, ensuring that the highest-risk areas receive the necessary attention and resources for comprehensive coverage. Lower-risk areas may be subject to a more streamlined resource allocation and differentiated surveillance strategies. This approach allows banks to optimize their surveillance efforts based on risk, while maintaining a complete risk management framework.

Another critical aspect of our discussion revolved around the pervasive issue of false positives in trade surveillance. The staggeringly high volumes of false positives can significantly impact the cost and time efficiencies of surveillance programs, as well as the human behaviour and performance of analysts tasked with identifying and closing these erroneous alerts. When the majority of an analyst's workload consists of dealing with false positives, it can lead to fatigue, reduced motivation, and potential oversights of genuine risk events.

Addressing the false positive problem is paramount to enhancing the efficiency and effectiveness of trade surveillance. By implementing advanced analytics, machine learning techniques, and fine-tuning detection algorithms, banks can significantly reduce the volume of false positives generated by their surveillance systems. This, in turn, allows analysts to focus their efforts on investigating high-priority, potentially suspicious activities, ultimately improving the overall efficacy of the surveillance program.

However, it is crucial to strike a delicate balance when tackling false positives. While reducing their volume is essential, banks must also ensure that they do not inadvertently suppress genuine risk events or compromise their ability to meet regulatory expectations regarding risk coverage and detection. This requires a continuous process of calibration, testing, and validation to maintain the integrity and robustness of the surveillance system.

As the debate unfolded, it became increasingly apparent that the landscape of trade surveillance is fraught with complexities and constantly evolving challenges. Teams are navigating a myriad of considerations, from selecting the most suitable technology approach to ensuring comprehensive venue coverage, maintaining data integrity, and optimizing false positive management. The stakes are high, now with the looming spectre of significant fines focused directly on trade surveillance programs and the imperative to meet regulatory requirements.

However, amidst these challenges, there are also abundant opportunities for innovation and collaboration. The potential for developing integrated surveillance strategies and the use of advanced analytics holds immense promise for enhancing efficiency and effectiveness in detecting market abuse. Moreover, the sharing of best practices and lessons learned among industry participants, such as we all participate in here on the 1LoD platform, can foster a more resilient and robust surveillance ecosystem.

In conclusion, our debate on leveraging surveillance technology and AI to increase efficiency at the 1LoD conference provided a thought-provoking exploration of the current state of trade surveillance and the intricacies that financial institutions must navigate. As the industry continues to evolve, organizations must remain agile, adaptable, and proactive in their approach to surveillance and risk management. By embracing innovation, fostering collaboration, and maintaining a keen focus on balancing risk, technology, and efficiency, banks can not only meet their regulatory obligations but also contribute to the overall integrity and fairness of the financial markets.

If you are grappling with this, or related surveillance issues, feel free to reach out to me for an informal discussion about how I may assist.