issues
search
kwoongbae
/
risk-management-papers
studying risk and regulation (risk management, solvency II)
1
stars
0
forks
source link
A piecewise-defined severity distribution-based loss distribution approach to estimate operational risk: evidence from chinese national commercial banks
#11
Closed
kwoongbae
closed
7 months ago
kwoongbae
commented
7 months ago
https://www.worldscientific.com/doi/abs/10.1142/S0219622009003727
Li, J., Feng, J., & Chen, J. (2009).
Contribution
전통적인 LDA는 body에, GPD를 이용한 EVT는 tail에 적용
this paper applied GPD for large loss data, and for parameters estimation, MLE, MM, PWMB, and PWMU are used.
Value at Risk (VaR), expected shortfall (ES), economic capital (EC)
kwoongbae
commented
7 months ago
Introduction
AMA: Advanced Measurement Approach
AMA에는 "top-down"방식과 "bottom-up"방식이 존재. 그리고 "bottom-up"은 (1) process approach와 "actuarial approach"로 나뉘는데 LDA가 (2)에 해당됨.
전통적인 LDA에서는 로그노말, 웨이블, 지수분포를 severity 분포 추정에 많이 사용하지만, low-frequency&high-severity data에는 피팅이 잘 안됨 --> 따라서 EVT로 tail부분을 추정함.
본 논문은 the use of different severity distributions for high-frequency&low-severity data와 low-frequency&high-severity data에 다른 분포를 적용함.
kwoongbae
commented
7 months ago
frequency distribution
포아송분포는 natural way to describe the number of unusual incidents in a unit of time or space하므로, 본 논문은 보아송 분포를 가정하였음.
포아송 분포만 사용하였으며, 추가적으로 통계적 계산은 하지 않음 (gostat.. 등을 진행하지 않음)
Contribution