FINRA recently released Regulatory Notice 15-09 (the “Notice”) to provide guidance on supervision/control practices for member firms and market participants using algorithmic trading strategies (including high frequency trading). According to the Notice, algorithmic strategies account for a substantial portion of trading activity in U.S. markets. While member firms engaging in algorithmic strategies are subject to SEC and FINRA rules, FINRA’s staff has noted that several firms lack appropriate supervisory procedures/controls concerning trading algorithms. In an effort to reduce any future occurrences of issues, FINRA developed effective practice initiatives designed to require firms engaged in algorithmic trading to be “trained, educated, and accountable for their role in equity trading.” FINRA’s best practices are summarized as follows:
- General Risk Assessment and Response: As member firms use of algorithmic strategy has increased, the potential of such strategies to impact market/firm stability has also grown. Accordingly, firms need to take a “holistic review” of their trading activity. Firms should consider engaging in support and controls through a cross-disciplinary committee which would assess and react to evolving risks.
- Software/Code Development and Implementation: Firms need to focus on development of algorithmic strategies and should consider monitoring activity to assure that development and change procedures are followed. Firms should consider maintaining (at a minimum) a basic summary of strategies which would enable supervisory, compliance, and regulatory staff to understand the algorithm without a need to resort to code review.
- Software Testing and System Validation: Firms should test strategies prior to being put into production – firms should consider implementing date integrity, accuracy, and workflow validation testing processes.
- Trading Systems: Firms should develop policies and procedures which include review of trading activity after an algorithmic strategy is in place or has been changed, and periodically monitor those controls.
- Compliance: Firms should consider developing compliance monitoring tools and periodically evaluate those tools. Firms should also implement trainings for supervisory and compliance staff concerning policies and procedures regarding algorithmic strategies.