MiFID II, FICC and Market Abuse Regulation

The MiFID II / MAR regulations to be implemented in January 2017, require firms to extend market surveillance capabilities to fixed income, currencies and commodities. While most firms have systems in place covering equity trading, FICC asset classes such as fixed income, currencies and commodities require different treatment. Firms considering extending their existing equity systems as some form of MiFID II / MAR interim solution should be aware this may be sub-optimal given the more aggressive stance of the regulators and the size of the fines the regulators are now seeking to impose. MiFID II / MAR surveillance should be specifically constructed for the FICC markets to meet the specific characteristics of the asset classes being traded in this area.

Organisations trading in FICC markets need to ensure market abuse and cross-instrument suspicious behaviour can be detected. This requires automated surveillance systems designed to cover specific asset classes particularly in those asset classes where there is no central order book, such as OTC trading. Thus alert logic needs to be constructed for each instrument type as different types of transactions can lead lead to different types of market abuse.

Further, it is a regulatory requirement that firms detail why specific alert parameters have been selected and the number of parameters used for each alert instance. These may be based on the timeframe of an evaluation window, frequency of order, unusual order size and so forth. Firms need to calibrate these alerts accordingly. For example, frequent but low margin trades executed on nominal amounts over an extended period by a single counter-party may require scrutiny given a series of suspiciously timed and accumulating pattern of trading activity.

Alert parameters need to be configured from a central console enabling different parameters for different types of instruments to be specified. Further diagnostic tooling should be provided that enables specific client, trader or algorithms to be examined.

Once alerts are calibrated recurring behavioural patterns need to be identified. This requires a machine based approach as it will require related and potentially related alerts to be correlated. Visualisation tools will add value, as this will enable related alerts to be correlated as event streams rather than isolated activities.

Finally any MiFID II / MAR technology should be able to integrate seamlessly with the existing infrastructure. This should be achieved without requiring changes to the messages, installing agents on data repositories or impacting the production environment in any manner.

← Return to News