Intensity process and compensator: A new filtration expansion approach and the Jeulin-Yor theorem (Q2476401): Difference between revisions

Let \((\Omega,{\mathfrak I},P)\) be a probability space, \(\mathbb{F}= ({\mathfrak I}_t)_{t\geq 0}\) any filtration and \(\tau\) an arbitrary nonnegative random variable. Let \(\mathbb{G}= (G_t)_{t\geq 0}\) be any filtration expansion of \(\mathbb{F}\). An increasing measurable process \((A_t)_{t\geq 0}\), such that \(A_0= 0\), \(\mathbf{1}_{\{\tau\leq t\}}- A_t\) is a \(\mathbb{G}\)-martingale and \(A\) is \(\mathbb{G}\)-predictable is called \(\mathbb{G}\)-compensator of \(\tau\). The intensity process \((\lambda_t)_{t\geq 0}\) of \(\tau\) is defined as the Radon-Nikodým derivative \(({dA_t\over dt})_{t\geq 0}\), provided that \(A\) is a.s. absolutely continuous with respect to the Lebesgue measure. In the present paper the authors study the problem of existence and computation of intensity process of \(\tau\). This problem is motivated by problem from credit risk, when \((X_t)_{t\geq 0}\) be a continuous-time time-homogeneous strong Markov process and \(\tau\) is its first hitting time of a Borel subset of the state space. The authors give some examples to use the obtained results.