A diffusion-type process with a given joint law for the terminal level and supremum at an independent exponential time (Q645596)

The paper presents a weak solution \((X_t)_{t \in [0,\infty)}\) to the stochastic differential equation \[ X_t = x_0 + \int_0^t \sigma(X_s,M_s) dW_s, \] where \(x_0 \in \mathbb R\) is a constant, \(\sigma : \mathbb R \times \mathbb R \rightarrow (0,\infty)\) is a bounded continuous function, \((W_t)_{t \in [0,\infty)}\) is a standard Brownian motion and \(M_s = \sup_{0 \leq r \leq s}X_r\). The solution \(X\) is constructed from an auxiliary Brownian motion by a suitable time change. Under the additional assumption that \(\sigma\) is Lipschitz-continuous with respect to its second argument, excursion theory is used to derive an expression for the probability density of \((X_{\xi_{\lambda}},M_{\xi_{\lambda}})\), where \(\xi_{\lambda}\) is an exponentially distributed random variable with parameter \(\lambda>0\), independent of \(X\). Further, the paper discusses how \(\sigma\) can be chosen so that the law of \((X_{\xi_{\lambda}},M_{\xi_{\lambda}})\) equals \(\mu\) for a given probability measure \(\mu\) that is supported on \(\{ (x',m') \in \mathbb R^2 : x' \leq m', x_0 \leq m' \}\), absolutely continuous with respect to the Lebesgue measure, and satisfies certain regularity conditions. This provides an alternative solution to the problem of constructing a martingale, the terminal value and running maximum of which follow a given law, studied by \textit{L. G. C. Rogers} [Probab. Theory Relat. Fields 95, 451--466 (1993; Zbl 0794.60042)]