Top of page
Main Content

New methods for modelling and forecasting risk

This project develops and assesses risk measures and risk forecasting. It assesses why customary measures failed in the financial crisis and develops new and better techniques.

The project is unique in terms of the scope and range of methods to applied and tested. This is of value to investors, institutions and regulators alike.

Research questions:

  • Which existing data sources and modelling techniques provide the most accurate and timely indicators of risk?
  • How can these risk indicators be improved through new innovative methods?

Research Methods:

Support Vector Machines (SVM) for regression, Merton-KMV structural modelling, realised volatility models, quantile regressions and CAViaR, portfolio spillover GARCH, Conditional Value at Risk (CVaR) and EVT (Extreme Value Theory).

Researchers:

Funding body:

Australian Research Council, Discovery Projects grant.

Timeline: 

January 2011 - December 2013

Skip to top of page