The FCA’s latest Market Watch presents findings from a peer review of firms’ testing of automated surveillance models. Undertaken in 2023, the peer review process assessed how investment firms review their surveillance tools and processes to ensure that they are accurately capturing potentially suspicious trading activity that could be classified as market abuse.
This assessment was not intended to analyse the surveillance systems themselves, but rather the testing around these systems to ensure that they were functioning as anticipated.
Specifically, the regulator examined “the frequency and methods used by nine investment banks to test the efficacy of their client order front running models”.
Mixed performance from firms
From performing this analysis, the FCA found that, while most firms did have formal procedures in place to ensure the accuracy of surveillance models, “the remainder had no formal process or a semi-formalised process”.
Firms with a proper review process in place examined:
- Parameter calibration
- Model logic
- Model code
- Data comprehensiveness and accuracy
Some firms opted for a “risk-based approach” to surveillance where the frequency and nature of testing was dependent on the relative risk of the relevant market abuse type.
The importance of robust surveillance controls
As regulators enforce penalties for market abuse infringements with ever-increasing rigour, static ‘check-box’ surveillance systems are becoming increasingly ineffective against the impact of digital acceleration, artificial intelligence, and highly dynamic market conditions. Firms are expected not only to have a surveillance system in place, but to rigorously test its configuration to ensure that alerts are being generated as anticipated.
In particular, firms are expected to have nuanced and flexible controls in place which are capable of dynamically adjusting to various data fields and market abuse topologies.
In the newsletter, the FCA specially stated that surveillance systems and processes needed to account for variables such as:
- Assets traded
- Actors involved
- Trading methods
- Venues accessed
The regulator found that, at present, “firms can take further steps in these areas”, suggesting that there is a lack of dynamism and flexibility in some of the systems that firms currently employ to detect instances of market abuse.
How to avoid surveillance failures
Market Watch 79 finishes with a series of suggestions for firms looking to avoid potential trade surveillance pitfalls.
Specifically, they recommend firms implement processes around data governance, model testing and model implementation/amendment.
eflow is able to provide solutions for all of these issues. Our automated data reconciliation, enrichment and health checks ensure that both market and trade data are always aligned and accurate.
eflow’s TZTS trade surveillance solution allows users to configure dynamic parameters to automatically adjust to variables and external factors. It also provides users with the option to fine-tune their platform’s configuration in an independent sandbox, which can then be promoted seamlessly to a live environment. This allows for flexible and robust model testing to ensure test parameters are calibrated to ensure there are no gaps in their alerts, while simultaneously reducing false positives.
eflow’s support team also takes part in active monitoring to ensure client systems are being utilised correctly. If any suspected issues are detected, eflow will contact the user and help them ensure their system is calibrated to effectively capture instances of suspected market abuse.
For more information on TZTS, click here (link) or book a consultation with the eflow team (link).
