Remember me

Register  |   Lost password?

The Trading Mesh

Roundtable Report - Trade Surveillance - Separating the Spoof From the Truth

Thu, 08 Jun 2017 08:47:35 GMT           

As regulatory scrutiny intensifies, could machine-learning and other artificial intelligence-based (AI) techniques play a greater role in the trade surveillance strategies of financial market participants? Will it usurp or simply complement existing rules-based approaches, helping firms to refine and target their compliance efforts more effectively?

This was the key focus of a roundtable event for industry practitioners held in late May, hosted by Certeco and The Realization Group at the London Capital Club, the third in a series focused on the introduction of the Europe-wide Market Abuse Regulation (MAR).

Whereas previous legislation centered on the exchange-traded activities of top-tier banks, MAR extends trade surveillance requirements to a much wider spread of firms and across a much more comprehensive range of markets. At the same time, market participants are adjusting to the increased reporting and best execution obligations of MiFID II, as well as the Financial Conduct Authority’s Senior Managers Regime, which places personal responsibility on individual executives for aspects of firm’s operations, including trade surveillance.

It is not clear as yet whether MAR (effective since July 2016) will herald a sea-change in the authorities’ approach to prosecuting market abuse. But it was noted that the number of enforcement cases had risen recently in the UK and regulators across Europe were taking a greater interest in the effectiveness of firms’ trade surveillance systems and controls.

In this context, it was pointed out that many firms’ compliance efforts are fragmented, with dedicated teams working separately on KYC and fraud, for example. While existing rules-based trade surveillance tools and techniques will remain crucial to flagging potential breaches of compliance, it was suggested that AI could play an important future role in helping firms coordinate monitoring activities on a holistic, enterprise-wide basis, i.e. across diverse data sources and formats, as they adjusted to the wider scope of MAR and other regulatory obligations.

Roundtable participants agreed that AI-based tools offered the potential to help firms target their trade surveillance efforts more effectively, for example by identifying patterns across different sources of data or prioritising the trade monitoring alerts most likely to require immediate attention. But for many firms only recently covered by MAR, the development of a risk-based approach to market abuse compliance is currently at a much more basic level.

Certeco business development director Nick Gordon said many smaller firms are initially looking at ‘out-of-the-box’ solutions to kick-start the process of monitoring their trading activities, predicting the need for ‘quality assurance’ services which help such firms match surveillance tools with their particular needs. At the other end of the scale, he noted, tier one firms were aiming to achieve more coordinated surveillance over all trading activity - voice-based and electronic - while looking to use AI and Behavioural systems to reduce Compliance workload and increase the surveillance capability.

AI may not play a significant role in the efforts of smaller firms to comply with MAR in the short term but its potential contribution to the future of trade surveillance - and compliance more broadly - was underlined by a number of roundtable participants. The established ability of AI-based tools make recommendations and observations (if not necessarily to explain them) based on consumption of structured and unstructured data sources is already being used, even to check for evidence of abuse in voice-brokered markets, where manual reply would be prohibitively time-consuming. It is still a challenge to monitor every communication within a financial institution that might reveal non-compliance, but AI also has preventative applications, roundtable participants noted.

As market abuse is effectively a deviation from the normal, optimal behaviour of a trader or portfolio manager, the ability of AI to identify changing habits, preferences or performance across multiple measures or media - e.g. a dip in trading profitability, changes in use of language or channels for work-based communication, or a switch to different asset classes - could flag the issues that could eventually lead to abuse, thereby potentially reducing actual incidence.

The fundamental challenge is how to understand and change human behaviour in order to prevent abuse and can the technology help drive this.

As well as monitoring and shaping behaviour to aid optimal performance, enhanced trade surveillance tools can support trading efficiency in other ways too. A number of trading professionals argued that real-time AI-based trade monitoring solutions could bring further business benefits if they could draw on multiple data sources to help traders better understand and respond to fast-moving market dynamics in order to hone execution tactics.

As noted by Certeco’s Gordon, this move of trade surveillance software onto the trading desk - and into the first line of defence against abuse - depends on delivery of real-time alerts but offers significant benefits in terms of compliance cost-efficiencies. This was also stated by a couple of the traders present at the table, it would be seen as a clear benefit to the business. While compliance officers saw the benefit in seeing the surveillance being undertaken before hitting the second line of defence.

Rules-based trade surveillance tools have served the industry well and are achieving a level of sophistication and maturity that is helping long-established users to meet the majority of their regulatory requirements effectively. But there is no room for complacency. As regulatory needs intensify, trade surveillance capabilities must further evolve. Rather than sending out even more alerts, tools must help firms identify the biggest and most immediate risks, whilst also monitoring behaviours in ways that both detect and prevent abuse in an effective and cost neutral approach.

As such, AI will not replace existing rules-based trade surveillance capabilities, but augment and reinforce them, becoming an integral part of the ongoing fight against market abuse, potentially taking an ever-larger role as data analytics capabilities develop. The increased regulatory emphasis on individual responsibility may be a particular driver of AI’s use in trade surveillance efforts as it encourages market participants to invest in solutions that will support the exercise of judgement in support of investment and trading goals - rather than just identifying abuse after the fact.

The increased reporting and surveillance requirements in today’s financial markets are clearly an administrative overhead at a time of generally rising costs and squeezed margins. But if handled cost effectively through the intelligent application of technology innovation they will also boost market confidence, with a beneficial knock-on impact on participation rates, liquidity levels and market efficiency.

 

This was the third in a series of round tables discussing issues around trade surveillance.

 

Discussion hosted by The Realization Group, led by Certeco, Nasdaq, Sybenetix and Charles Russell Speechlys, on 25th May, 2017