FCA publishes review of algorithmic trading controls

The FCA has published the results of its multi-firm review of principal trading firms’ compliance with MiFID Regulatory Technical Standards 6, which imposes requirements on investment firms engaged in algorithmic trading.

Governance

The FCA found that firms’ self-assessment processes had generally improved, with some firms having their assessments reviewed by external auditors. However, it maintained that more detail is needed in some areas, including in respect of certain outdated policies and unclear processes.

Technical knowledge among compliance staff was varied – some had strong expertise and provided challenge to trading processes where appropriate, but others lacked knowledge and has limited ability to challenge trading behaviours.

Most firms maintained complete algorithmic inventories capturing key details, but the FCA highlighted that some inventories did not specific individuals who were approved to operate the relevant algorithm.

Development and testing

In terms of conformance and simulation testing, most firms complied with RTS 6, but procedures varied slightly. Some firms clearly defined scenarios requiring conformance testing, and had dedicated significant resources to robust simulation testing with a variety of stress-scenarios. However, others had poorly defined testing procedures and record keeping, or did not consider a wide enough range of market scenarios in their stress testing.

Firms generally took a conservative approach in deploying algorithms. Good practice included using pilot trades to test functionality over a robust governance framework. The FCA noted that some firms lacked formal documentation procedures for deployment, with unclear ownership of key elements.

Risk controls

All firms had adequate pre-trade controls, which were clearly defined and calibrated according to algorithm type and asset class. However, ownership of pre- and post-trade controls was in some cases poorly defined and documented.

Market abuse surveillance

Many firms used in-house surveillance systems, with some customised to trading types. However, some firms had not updated their systems, and lacked formalised procedures for investigating alerts.