Seminar by Dr Chi Truong
- Topic: Internal loss models from external loss data
- Speaker: Dr Chi Truong, Postdoctoral Research Fellow, Department of Applied Finance and Actuarial Studies, Macquarie University
- Venue:E4A 623
- When: 31st May, 2013, (Fri)
- Time: 11.30am - 12.30pm (Morning Tea from 11am)
This seminar describes work that has been done in conjunction with, and under the sponsorship of, the CBA. The aim of the project is to extract information from losses from a variety of banks around the world, consolidated into the so—called ORX database, for the purpose of more accurately predicting possible loss outcomes for CBA. This process is called the development of an "internal model" of losses and internal models are used to set aside regulatory capital.
Development of internal models for operational risk regulatory capital under the Advanced Measurement Approach of Basel II requires the use of both internal and external data – data from other banks. External loss data must be scaled to make loss relevant to any specific bank. Knowledge of the factors that significantly affect the loss size helps banks to improve operational risk management. This research aims to: 1) identify key drivers for operational loss severity using the ORX database; 2) identify inter-relationships or co-relationships between key drivers; and 3) estimate the operational risk loss severity distribution for identified risk categories utilizing the ORX data.
Flexible distributions are used within Generalized Additive Models of Location, Scale and Shape (GAMLSS) framework, allowing the examination of the impacts of explanatory variables on any distributional parameters. We introduce new measures to check the goodness of fit and the explanatory power of the model with higher weight assigned to the upper tail. The model is applied to find the drivers for operational risk in Mortgage mis-selling and Transaction capture.
The views expressed are those of the authors and do not necessarily reflect those of the Commonwealth Bank of Australia.