Skip to main content

Detection of Unauthorised Trading

By 28 Feb, 2018June 14th, 2018Case Studies, Surveillance


Swiss Investment Bank


Front-office; Risk; Technology

Asset Class





12 months

The Ask

Following recent cases of unauthorized trading, the client established a new unit to provide early detection of behaviours indicative of potential unauthorized trading in order to avoid financial and reputational loses.

Using statistical models applied to Key Risk Indicators, trading activity and historical series, the client expects to identify traders whose outlier behaviours differ to their peers. Once the outliers are detected, the unit will conduct individual reviews of the trading activity, ultimately reporting any suspicious activity.

The client asked GreySpark to define additional behavioural indicators and conduct an independent review of the model in addition to using the analytics of the model’s results.


To fulfil the client’s need for an experienced cross-asset specialist, GreySpark deployed a Senior Consultant with significant products, processes, and Operational Risk experience in Global Banking.

  1. The Consultant analysed the statistical model used to detect outliers in trading behaviour and proposed improvements to the algorithms and quantitative and holistic indicators.
  2. The same investigations were then conducted over trading behaviours.
  3. The Consultant engaged with the front-office supervisors and trade management units to share his findings about the traders’ activities. These conversations helped him receive feedback and structure his search.


GreySpark delivered:

  • Trader-review reports;
  • System requirements;
  • An analytics report of the model’s weekly results;
  • A new requirements definition to improve the model and tools used by the unit; and
  • A new requirements and process definition to conduct trader reviews.

GreySpark Delivered Benefits

  • Given the high degree of specialization needed by the client, GreySpark provided the necessary expertise across multiple fields to deliver a successful assessment of the algorithms used in the model.
  • The client has duplicated the number of trader reviews conducted per week.
  • GreySpark’s recommend changes to the algorithm has helped the unit run more efficiently and take better control of their processes at-hand.

Do you have a similar project that you need help with?