Testing Alternative Dynamic Systems for Modelling Tourism Demand

The goals in this paper are to contribute an empirical study of tourism demand dynamics, and to point out areas where the scrutiny of relationships between theoretical and empirical considerations are likely to produce new insights in this area of research. A flexible general form of a Dynamic Almost Ideal Demand System (DAIDS) is derived to analyse the UK tourism demand for its geographically proximate neighbours Portugal, Spain and France, in the period 1969-1997. Nested within the general dynamic structure are Deaton and Muellbauer’s static AIDS model itself, the partial adjustment model and the auto-regressive distributed lag model, which are tested against the general dynamic alternative. The empirical results obtained show that DAIDS is a data coherent and theoretically consistent model, providing evidence of the robustness of this methodology to conduct tourism demand analysis in a temporal context. Moreover, the dynamic model offers statistically strong evidence on the inadequacy of the orthodox static AIDS and the other restricted models to reconcile consistently data and theory within their formulations. Estimates for tourism price and expenditure elasticities are obtained, permitting a comparative analysis of the relative magnitudes and statistical relevance of long and short run sensitivity of the UK tourism demand to changes in its determinants.