Emerging challenges for algorithmic trading governance and controls

More conduct risk with less human intervention?

More conduct risk with less human intervention?


Also on home.kpmg

Man looking at a tablet

Triumphant success of trading algorithms

Algorithmic trading, which has already become an integral part of equities as well as FX markets, is now successfully conquering the world of fixed income and commodities. This is driven by a greater demand for speed and transparency by market participants as well as some of the European regulations implicitly encouraging electronic execution to meet the transparency and best execution requirements.

According to the most recent studies conducted by the CFTC and SIFMA, the percentage of electronic execution on these markets has gone up on average by 30 percentage points over the last 5 years and continues to increase. Across fixed income, the increase has been most dramatic for all standardised and fungible bonds (i.e. high class sovereigns). For example, almost 70% of all US Treasuries is now executed electronically. At the same time, corporate bonds are slightly lagging behind due to the lack of fungibility and more bespoke characteristics.

The increase has also been rather significant for commodities markets. The CFTC numbers demonstrate that the share of automated orders across energy and metals futures has raised by 20 percentage points reaching the level of 80% in 2019.

Can we say that less human intervention in algorithmic trading means less conduct risk associated with this type of activity? As the history of ‘flash crashes’ shows, the risk does not seem to disappear. The risk has transformed, becoming more complex and sophisticated in its nature but it remains as true as ever.

How to measure your conduct risk?

How can businesses guard against conduct risk in an automated trading environment? Removing human traders from the equation eliminates behavioural risks directly associated with trade inception, but it does not completely remove the risk to conduct outcomes expected from banks. As the FCA representative mentioned in one of its most recent speeches “the FCA cannot prosecute a computer, but we can seek to prosecute the people who provided the governance over that computer”. To better understand the regulatory expectations, it is worth looking at 5 Conduct Questions Programme , where the FCA asks the following questions?

  1. What proactive steps do you take as a firm to identify the conduct risks inherent within your business?
  2. How do you encourage the individuals who work in front, middle, back office, control and support functions to feel and be responsible for managing the conduct of their business?
  3. What support (broadly defined) does the firm put in place to enable those who work for it to improve the conduct of their business or function?
  4. How does the Board and ExCo (or appropriate senior management) gain oversight of the conduct of business within their organisation and, equally importantly, how does the Board or ExCo consider the conduct implications of the strategic decisions that they make?
  5. Has the firm assessed whether there are any other activities that it undertakes that could undermine strategies put in place to improve conduct?

Algorithmic trading systems add extra considerations and risks to all of these questions. All new and existing algorithms should be properly assessed and reviewed before they interact with the marketplace. To demonstrate this we can already observe a few trends in best practices around conduct risk management applied by large market players:

  • Strengthening monitoring and surveillance systems to account for specific risks associated with the algorithmic trading
  • Analysing how a set of strategies rather than an individual algorithm may create market disruption and lead to disorderly trading
  • Taking into account wide set of possible market manipulation and disruption scenarios during the design and testing phase as well as sufficient challenging of this by all control functions
  • Conducting detailed market abuse trainings for all relevant functions involved in designing, testing and trading
  • Strengthening all control functions by hiring skilled algo trading specialists and upskilling existing staff

Expectations are high and it is time to act now

Historically, algorithmic trading has evolved at a faster pace than the corresponding regulation. However, now the regulators across the globe have already caught up with various technological developments and continue to scrutinise the assessment of all risks and controls in the industry. Following MiFID II implementation and subsequent additional guidance from regulators (PRA Supervisory Statement, FCA report on algorithmic trading compliance), firms are finalising their second self-assessment of trading algorithms and preparing for 2020. This does not mean, however, that conduct risk assessment and oversight is a one-off or one time in a year process. Given the pace of change in algorithmic trading, the approach to algorithmic trading governance and controls should also keep pace, covering newly added asset classes and involve all departments of your firm.

 How can we help? Download our brochure here.

© 2021 KPMG LLP a UK limited liability partnership and a member firm of the KPMG global organisation of independent member firms affiliated with KPMG International Limited, a private English company limited by guarantee. All rights reserved.

For more detail about the structure of the KPMG global organisation please visit https://home.kpmg/governance.

Connect with us


Want to do business with KPMG?


loading image Request for proposal

Save, Curate and Share

Save what resonates, curate a library of information, and share content with your network of contacts.

Sign up today