TECHNOLOGY AND AI IN CAPITAL MARKETS TRADE SURVEILLANCE
R&C: What are the common challenges capital markets firms face with respect to trade surveillance? Historically, how have they tried to address those challenges?
Fagone: Heightened regulatory expectations and focus, coupled with an increasingly complex operating environment that includes, for example, multi-leg orders with various counterparties across multiple product types or exchanges, increasingly strain the capabilities of surveillance programmes. A couple of the common challenges we see are high alert volumes and data disconnects. In terms of high alert volumes, rules-based surveillance patterns tend to generate an unmanageable volume of alerts that require manual follow up and disposition. This white noise can not only make it very difficult to properly flag suspicious trading activity, but can also mask emerging threats. Firms should make use of advanced machine learning (ML) techniques to tune and optimise parameters and establish thresholds and move away from rule-based surveillances. As for data disconnects, data quality and availability challenges dramatically impact the integrity of trade surveillances and the corresponding volume and integrity of the resulting alerts. Disconnects between expectations and data reality can lead to a false sense of security and significant, unknown surveillance gaps. Firms should develop and implement a data governance framework to help monitor the data quality, including establishing data sources, lineage, controls and change management protocols.
Jan-Mar 2020 Issue
KPMG