Multivariate and intervention analysis of Maltese tourism time series

Abstract

Using monthly data of Tourism survey from 2003 to 2012, in this study we apply multivariate and intervention analysis to determine relations between the variables under study and other external events. The first part of the thesis involves the fitting of a VAR model, through the specification and estimation process, followed by assessing the fit by model adequacy checks. The model is used for determining its adequacy for forecasting purposes, and for understanding better the relationships between its variables via impulse response analysis. In the latter, this thesis examines the impulse response function, which describes the response over time of each variable in the VAR to a one-time shock in any given variable while keeping all others constant. Intervention analysis, which we cover in the second part of this thesis, is an important method for analyzing the temporary or long-lasting effects of sudden events on time series data. In an intervention model, the input series is an indicator variable that contains discrete values that flag the occurrence of an event affecting the response series. The reason why this is relevant within the context of our study is because we would like to understand the effect of the advent of low cost airlines on the tourism time series variables under study. Intervention methods are quasi-experimental in nature and the validity of modelling with these methods depends upon assumptions about the timing of the intervention and the response of the process to it.