Skip to main navigationSkip to main content
The University of Southampton
Centre for Risk Research

CRR Researchers Use Behavioural Analytics to Manage Risk in the City

Published: 9 October 2015

The Centre for Risk Research has been undertaking a Knowledge Transfer Partnership (KTP) with Star Financial Systems since 2012. The project has focused on using behavioural analytics to provide innovative ways for managing financial risks.

Since retail-sector derivative companies (e.g. those offering spread betting, CFDs, FX, etc. to retail clients) prefer only to hedge the trades of successful traders, the ability to accurately predict which traders are likely to make profitable trades is very important for their risk management. However, as Dr. Fraser-Mackenzie, the lead researcher on the project points out, “an important early finding was that a client’s past profitability is not a good predictor of that client’s future profitability.” As such, hedging strategies cannot be based solely on clients’ historical profitability.


Consequently, work on the project has examined individual account data in order to identify a number of behavioural traits that could better predict a client’s future profitability. Drawing from industry knowledge and recent advances in cognitive psychology these traits are being developed in predictive models to classify traders, thereby helping to develop more effective hedging strategies.

Star Financial Systems

Subsequent work has focused on using behavioural analytics to improve risk management decisions. A major problem for companies that hedge their client’s risk is that the company can be exposed to the risk of being over/under-hedged if there are sudden changes in the underlying client book. Consequently, the KTP, jointly supervised by Professor Ming-Chien Sung, Professor Johnnie Johnson and Dr. Tiejun Ma, has involved developing a cognitively intuitive computer application that, again based on behavioural analytics, simulates how each client will respond over a range of different market conditions. Each client’s “stress”, measured by an index, is modelled over a range of market conditions. The simulations, therefore, reveal to the company those price points that they may be at risk of being over/under hedged, enabling the company to stay ahead of these risks.


The project continues until early summer next year.

Privacy Settings