Title: Representations of linear continuous time models with mixed frequency data
Authors: Michael Thornton - University of York (United Kingdom) [presenting]
Abstract: Methods to derive an exact representation for observations generated at mixed frequencies by an underlying continuous time ARMA (CARMA) process are derived. The data may be stocks, flows or a mixture of the two. The method involves reconstructing a discrete time linear process, with first and second properties identical to the underlying model, from a state space model and avoids treating mixed frequency data as a problem of missing or aggregated observations.