This paper presents a new framework for coping with problems often encountered when modeling seasonal high frequency data containing both flow and stock variables. The idea is to apply a multivariate weekly representation of a daily periodic model and to exploit the possible cointegration and...