Details
| Original language | English |
|---|---|
| Pages (from-to) | 197-214 |
| Number of pages | 18 |
| Journal | Astin bulletin |
| Volume | 29 |
| Issue number | 2 |
| Publication status | E-pub ahead of print - 29 Aug 2014 |
Abstract
Numerical evaluation of compound distributions is one of the central numerical tasks in insurance mathematics. Two widely used techniques are Panjer recursion and transform methods. Many authors have pointed out that aliasing errors imply the need to consider the whole distribution if transform methods are used, a potential drawback especially for heavy-tailed distributions. We investigate the magnitude of aliasing errors and show that this problem can be solved by a suitable change of measure.
Keywords
- aliasing, change of measure, Fourier transformation, random sums, ruin probabilities, Total claim size distribution
ASJC Scopus subject areas
- Business, Management and Accounting(all)
- Accounting
- Economics, Econometrics and Finance(all)
- Finance
- Economics, Econometrics and Finance(all)
- Economics and Econometrics
Cite this
- Standard
- Harvard
- Apa
- Vancouver
- BibTeX
- RIS
In: Astin bulletin, Vol. 29, No. 2, 29.08.2014, p. 197-214.
Research output: Contribution to journal › Article › Research › peer review
}
TY - JOUR
T1 - Computation of compound distributions i
T2 - Aliasing errors and exponential tilting
AU - Grübel, Rudolf
AU - Hermesmeier, Renate
PY - 2014/8/29
Y1 - 2014/8/29
N2 - Numerical evaluation of compound distributions is one of the central numerical tasks in insurance mathematics. Two widely used techniques are Panjer recursion and transform methods. Many authors have pointed out that aliasing errors imply the need to consider the whole distribution if transform methods are used, a potential drawback especially for heavy-tailed distributions. We investigate the magnitude of aliasing errors and show that this problem can be solved by a suitable change of measure.
AB - Numerical evaluation of compound distributions is one of the central numerical tasks in insurance mathematics. Two widely used techniques are Panjer recursion and transform methods. Many authors have pointed out that aliasing errors imply the need to consider the whole distribution if transform methods are used, a potential drawback especially for heavy-tailed distributions. We investigate the magnitude of aliasing errors and show that this problem can be solved by a suitable change of measure.
KW - aliasing
KW - change of measure
KW - Fourier transformation
KW - random sums
KW - ruin probabilities
KW - Total claim size distribution
UR - http://www.scopus.com/inward/record.url?scp=85011437009&partnerID=8YFLogxK
U2 - 10.2143/AST.29.2.504611
DO - 10.2143/AST.29.2.504611
M3 - Article
AN - SCOPUS:85011437009
VL - 29
SP - 197
EP - 214
JO - Astin bulletin
JF - Astin bulletin
SN - 0515-0361
IS - 2
ER -