Common Periodic Correlation Features and the Interaction of Stocks and Flows in Daily Airport Data

Abstract

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 common feature properties of the variables in order to obtain a more parsimonious model representation. We introduce the notion of common periodic correlations, which are common features that co-vary - possibly with a phase shift - across the different days of the week and possibly also across weeks. The paper also suggests a way of modelling the dynamic interaction of stock and flow variables within a periodic setting that is similar to the concept of multicointegration among integrated variables. The proposed modelling framework is applied to a data set of daily arrivals and departures in the airport of Mallorca.

Keywords: Periodic autoregression, seasonality, high frequency data, cointegration, multicointegration, common features

JEL Classification: C12, C22, C32