Archimedean survival processes (Q1941420)

The authors construct a family of multivariate stochastic processes called Archimedean survival processes (ASPs) defined over a finite time horizon with terminal value having multivariate \(l_1\)-norm summetric distribution so that the terminal value has an Archimedean survival copula. These ASPs are \(n\)-dimensional gamma random bridges (GRBs) with each marginal process as GRB. The authors generalize the ASPs to Liouville processes which are Markov processes whose increments have multivariate Liouville distributions. The ARPs are applied to interpolate dependence structures when using Archimedean copulas in discrete time models and the Liouville processes are applied to intraday forecasting of realized variance. Several characterizations of the law of ARPs are also derived.