Finite and infinite time ruin probabilities in a stochastic economic environment
Let (A1,B1,L1),(A2,B2,L2),... be a sequence of independent and identically distributed random vectors. For , denoteYn=B1+A1B2+A1A2B3+...+A1...An-1Bn+A1...AnLn.For M>0, define the time of ruin by TM=inf{n Yn>M} (TM=+[infinity], if Yn[less-than-or-equals, slant]M for n=1,2,...). We are interested in the ruin probabilities for large M. Our objective is to give reasons for the crude estimates P(TM[less-than-or-equals, slant]x log M)[approximate]M-R(x) and P(TM<[infinity])[approximate]M-w where x>0 is fixed and R(x) and w are positive parameters. We also prove an asymptotic equivalence P(TM<[infinity])~CM-w with a strictly positive constant C. Similar results are obtained in an analogous continuous time model.
Year of publication: |
2001
|
---|---|
Authors: | Nyrhinen, Harri |
Published in: |
Stochastic Processes and their Applications. - Elsevier, ISSN 0304-4149. - Vol. 92.2001, 2, p. 265-285
|
Publisher: |
Elsevier |
Keywords: | Insurance mathematics Ruin problem Level-crossing probability Stochastic discounting Large deviations theory |
Saved in:
Saved in favorites
Similar items by person
-
Nyrhinen, Harri, (2007)
-
Upper bounds of the Gärtner-Ellis theorem for the sequences of random variables
Nyrhinen, Harri, (2005)
-
On the ruin probabilities in a general economic environment
Nyrhinen, Harri, (1999)
- More ...