Tail behavior of some shot noise processes

The author presents a mathematical model (shot noise process) that can represent financial model in which so-called ``heavy tails can appear. A shot noise process is \((X(t)=\sum_{T_i\leq t}R_i(t-T_i)\); \(t\in\mathbb{R})\). \((T_i\), \(i\in\mathbb{N})\) are the jump times of a point process on the real line; the stochastic processes \((R_i(t), t\in\mathbb{R}_+)\) are the response processes. Let \(X_{\gamma,d,m}(t)=\sum_{T_i\leq t}Z_if_{\gamma,d,m}(t-T_i)\), where \((Z_i)_{i\in\mathbb{N}}\) is a sequence of i.i.d. random variables, and where \(\gamma>0\), \(d>0\), \(m\geq 1\), \[ \begin{aligned} f_{\gamma,d,m}(x) & =j^\gamma\cos\left(\pi{j\over m}\right);\quad x\in\bigl(j-(j+1)^{-(d+1)},\;j+(j+1)^{-(d+1)}\bigr),\quad j\geq 1,\\ & =0\text{ otherwise.}\end{aligned} \] \((T_i\), \(i\in\mathbb{N})\) are the jump times of a Poisson process with intensity \(\lambda\). Theorem 3.1 gives, under some conditions, the heavy tail property of \(X\); part (i) states that \[ P\bigl(X_{\gamma,d,m}(t)>x\bigr)\sim c_{d,\gamma,m}\lambda EZ_1x^{-d/\gamma} \] when \(x\to\infty\), where \(c_{d,\gamma,m}={1\over 2\gamma m}\Gamma(d/\gamma)\sum^{2m-1}_{i=0}|\cos\pi{i\over m}|^{d/\gamma}\). For the entire collection see [Zbl 0901.00010].

Recommendations

Cited in

(14)

This page was built for publication: Tail behavior of some shot noise processes

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4247117)