Ruin Probabilities for Levy Processes with Mixed-Exponential Negative Jumps

From MaRDI portal





Assume that \( X \) is a Lévy risk process with mixed-exponential negative jumps, so that its Lévy measure \( \Pi \) is given by \[ \Pi (dy) = \begin{cases} \pi (dy) , & y>0, \\ \lambda\sum_{k=1}^{n}a_k b_k e^{b_k y} dy , & y<0.\end{cases} \] Here \(\pi\) is an arbitrary Lévy measure on \( (0,\infty)\), \(0<b_1< \dots <b_n\), \(a_k>0 \) and \(\sum_{k=1}^n a_k =1 \). Denote by \(k(p)\) the Laplace exponent of \(X\), i.e. \( k(p) = \frac{1}{t}\log Ee^{pX_t}\) and assume \(k'(0-)>0 \). Further, let \(\tau(\gamma)\) be an exponential r.v. with parameter \(\gamma >0\). The author's main result claims that a) the process \(X\) drifts to infinity for \(t \rightarrow \infty \); b) the ruin probability of \(X\), \( R(x)= P(\inf \{ x+X_t : t\in [0, \tau (\gamma))\} \leq 0)\), \(x\geq 0 \), is given by \(R(x)= \sum_{j=1}^{n+1} A_j e^{-p_j x}\). Here \(-p_1, \dots , -p_{n+1}\) are the negative roots of the equation \(k(p)=\gamma \). The constants \(A_j , j=1\dots n+1 \), depending on \(p_j\) and \(b_k , k=1\dots n \), are explicitly found.




Cited in
(19)








This page was built for publication: Ruin Probabilities for Levy Processes with Mixed-Exponential Negative Jumps

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4830844)