Sydney Financial Mathematics Workshop
Sponsored by
National Australia Bank and
Q-Group Australia
Analysis of Monte-Carlo simulations for measuring
credit risk exposure using historical correlation
Adam Kucera and Volf Frishling (CBA)
| Time: | 5:15--7:00 pm |
| Date: | Monday 17th April 2000 |
| Venue: | NAB Offices, Conference Room, Level 19,
255 George St, Sydney |
Abstract
Tens of thousands of open positions spread over thousands of instruments
create unusual difficulties for risk management. The main thrust of our
talk will centre on three primary issues:
-
Identifying relevant time series and instruments
- Time series increments versus time series values.
- Investigate the impact of using various time lengths
-
The simulation of large multi-factor portfolios
- Principal component decomposition of the correlation matrices.
- Simulation of the parsimonious dominant components.
- Computational issues arising when dealing with large matrices
-
Investigating the explanatory power of the results
- Determining the explanatory powers of the principal components.
- Separating the signal from the noise in the market data
We will present the results from real market data as used by CBA. Time
allowing, we will discuss a number of possible future directions in this
field.
Please feel free to bring this to the attention of interested colleagues.