Risk CDT – Predictive Modeling for Interbank Markets

PLEASE APPLY ONLINE TO THE SCHOOL OF ENGINEERING, PROVIDING THE PROJECT TITLE, NAME OF THE PRIMARY SUPERVISOR AND SELECT THE PROGRAMME CODE "EGPR" (PHD - SCHOOL OF ENGINEERING)

This is a project within the multi-disciplinary EPSRC and ESRC Centre for Doctoral Training (CDT) on Quantification and Management of Risk & Uncertainty in Complex Systems & Environments, within the Institute for Risk and Uncertainty. The studentship is granted for 4 years and includes, in the first year, a Master in Decision Making under Risk & Uncertainty. The project includes extensive collaboration with prime industry to build an optimal basis for employability.

Motivation:
International corporations that deal in multiple currencies need to manage foreign exchange (Forex, FX) risk, i.e., to protect against adverse currency moves. To do this cost effectively requires the use of a combination of financial instruments such as options and forwards as well as the use of FX spot markets. Often corporations do not have the internal expertise in finance and trading to effectively do this, and so they resort to unsophisticated and sub-optimal methods.

Teramark technologies helps such large corporations manage foreign exchange risk. They do this by combining econometric predictive models of currency prices with sophisticated machine learning models that drive their decision making, along with their capability to hold their own FX-inventory positions in the market. This means that in many cases they can fill their clients’ upcoming transactions directly from their own book without having to go to the market, yielding a higher level of price stability and less liquidity impact.

Objectives:
The main aim of this project is to research with Teramark Technologies how to replace and/or augment their existing econometric simulation models by other approaches, such as agent-based modelling with specific shock and scenario exploration capabilities. This would give the opportunity to include effects of certain events and more information from other sources in the simulation of market dynamics.
The project would involve the following steps:
• build agent-based representations of relevant markets that capture key players and can be used to build predictive models via simulation;
• using these representations, develop a range of candidate predictive models of short-term currency price movements using Machine Learning techniques (e.g. Reinforcement Learning);
• validate the predictive models using custom datasets provided by Teramark, and benchmark these models against their existing econometric models.
If the developed techniques show good potential then there is a possibility to incorporate this into Teramark’s production system.
A successful candidate will have good programming and problem-solving skills and a keen interest in Machine Learning, and in particular its application to financial markets.