The 2016 market abuse regulation rules are a real test for surveillance functions in financial services. The new generation of smart, real-time data and analytics systems have a major role to play in compliance. But, there’s still no substitute for human judgment when keeping an eye out for trading malpractice.
Since the Market Abuse Regulation (MAR) went live on 3 July 2016, the surveillance functions of many banks have struggled with how best to implement the new rules. The challenge has been made even harder by a lack of quality data, immature technology and the application of the regulation across a wide range of markets.
Financial services firms don’t have a problem with the aim of the regulation: they warmly welcome effective controls to detect and prevent misconduct. And, across the board, they have been investing heavily in communications and order-monitoring surveillance systems. These can even detect possible market abuse without trades taking place. But, nevertheless, communications surveillance systems are still relatively inefficient. Mainstream audio surveillance solutions, for example, can be prohibitively expensive when used across the whole organisation. An explosion of new communications channels is making the workload even heavier.
On 12 November 2014, the Financial Conduct Authority fined five large banks $1.7 billion for failing to control business practices in their spot FX trading operations. That effectively ended any argument that it was “not within scope” for regulators to uphold standards of market conduct.
At the same time, those traders trying to manipulate the market have become ever more inventive. Layering, spoofing, smoking, stuffing, flashing: the slang they give to an armoury of pre-trade strategies hints at this creativity. Electronic trading certainly makes it easier to capture relevant data and design quantitative tests to spot these schemes. But in the world of manual (or voice) trading, surveillance functions are often blind to pre-trade activities. In markets such as fixed income and commodities, quotes or indications of interest have not been recorded in a way that supplies the structured data that surveillance technology needs.
The regulators are well aware of this shortcoming. On 12 November 2014, the Financial Conduct Authority fined five large banks $1.7 billion for failing to control business practices in their spot FX trading operations. That effectively ended any argument that it was “not within scope” for regulators to uphold standards of market conduct. Today, surveillance functions need to monitor activities related to a core portfolio of financial instruments. But, they also need to consider the wider context in which they operate. FX spot, physical commodities, storage space in warehouses, financing transactions and even freight arrangements may be outside the strict regulatory perimeter. Yet any surveillance function ought to keep a close eye on them when monitoring market conduct.
Machine learning and other technologies are starting to revolutionise surveillance. Artificial Intelligence (AI) is not, as yet, intelligent enough. The ingenuity of those who try to fly under the radar, across a wide variety of markets, means sound professional judgment is as much in demand as ever. Former traders are highly sought-after as surveillance experts. They can assess scenarios as they develop, analyse alerts, interpret trading patterns and identify conflicts of interests.
Even with cognitive technology and expert human judgment, how do surveillance functions know if their controls are working? To put it another way: if you don’t identify any serious market abuse, is that because the controls are so good they deter bad behaviour – or so weak that nothing gets found out?
MAR offers some practical examples of scenarios to consider when thinking about this point. Other bodies, such as the Fixed Income, Currencies and Commodities Markets Standards Board (FMSB), has also published useful guidance on the surveillance of FX markets that could be applied to surveillance functions monitoring other markets.
Ultimately, there’s no substitute for a good understanding of the culture, habits and behaviours of the trading floor, built up through rigorous monitoring – and expert human judgment.
Find out more about how KPMG can help with successful MAR implementation.
Lucas Ocelewicz is a Director in Risk Consulting at KPMG in the UK, specialising in trading risk and wholesale conduct. He has investigated and supported banks in remediation of major scandals such as LIBOR and FX as well as one of the largest incidents of unauthorised trading.