Behavioural analysis in compliance in financial markets: a feasibility survey

Chris Hamblin

22 August 2017

The phrase that the software firm actually uses is "behavioral analytics," which it defines as the measurement of anomalies from an individual’s 'normal' behaviour or that of his peer group "across multiple risk factors." More and more, compliance officers are using analytical software to counteract "conduct-related threats" such as collusion, unauthorised trading, benchmark manipulation and rogue trading. These crimes are difficult to prevent because they do not follow a predictable pattern that analysts can 'model.' Such threats — along with weak trade controls and failures to detect suspicious activity — can spark off eight-figure financial losses, substantial legal fees and irreversible reputational damage for a firm.

Nevertheless, the emergence of 'Big Data' and advanced analysis, including behavioral analysis, is bringing risk-based surveillance within the reach of financial institutions at last. By looking at vast amounts of data and measuring deviations from normal behaviour, firms can identify risky entities — including traders, investment advisors and accounts.

NICE Actimize worked in tandem with Infosurv earlier this year to conduct a survey on the subject. Its objectives were to quantify the adoption of behavioral analysis in surveillance 'programmes' (a word it uses throughout the report to describe various plans of action) by financial service firms; to identify the problems and 'use cases' (a term it only uses once) that behavioral analysis might solve; to understand its worth to firms; and to identify the data, entities and risk factors that firms intend to analyse when conducting it. 'Intend' is the operative word here because most of them have yet to do it.

Sixty individuals from financial institutions all over the world, including broker/dealers and investment banks, participated in its anonymous survey. Their geographical spread was reasonably even, with 22% residing in the Asia-Pacific region, 36% in Europe, 14% in North America, 9% in the rest of the world, and 19% from global organisations. Nearly one-third - 31% - of them worked in compliance; 14% worked in risk; and 19% worked in IT. When it came to the asset classes on which they were 'focusing,' the figures were 24% equities; 15% fixed income; 14% FX; 9% futures; 10% options; 11% swaps; and 17% other.

Main findings

42% expect to start analysing behaviour in the next 12 months.
79% believe that doing so will help their compliance efforts.
84% say that behavioral analysis is one of their top five priorities for the tax year 2017-18.
82% think that behavioural analysis can uncover hidden threats that model-based analysis is not able to detect.
77% believe that if they uncover "hidden threats" they will be able to reduce the likelihood that regulators will fine or censure them.
82% believe that if they detect these "hidden threats" they will be able to reduce their operational risks and financial losses.
91% believe that if they analyse behaviour, while also analysing things in a model-based way, they will be able to investigate things more quickly.
73% believe that if they use behavioural and traditional, model-based, analysis together they will assess 'alerts' more accurately. Later on, the survey mentions market surveillance alerts, so it is likely that they mean these.

NICE Actimize has stated: "Firms will seek an integrated surveillance solution – combining model-based analytics and behavioral analytics – in order to improve effectiveness, while reducing the total cost of compliance."

Good regulation

It also believes that various regulatory initiatives "seem to favour an approach based on analyzing anomalies from normal behaviour."

The UK's Fair and Effective Markets Review (FEMR), it argues, suggests that mitigating risk by monitoring an individual’s conduct and identifying 'vulnerabilities.'
The US Financial Industry Regulatory Authority (FINRA) likes the idea of mining trade data for deviations from normal trading patterns.
The Fixed Income, Currencies and Commodities (FICC) Markets Standard Board in London endorses systems to prevent various kinds of transaction-based conduct.
The Monetary Authority of Singapore (MAS) recommends advanced analysis to firms, the better to detect complex patterns of market abuse.

When asked what their behavioral analysis was going to concentrate on, 26% of respondents said traders, 24% said accounts, 18% said brokers, 12% said advisors, 9% said legal entities, and 11% said 'other.'