Monthly, annual and quarterly frequencies: a comparison of models for tourism in Sardinia and bounded rationality

This paper constructs and estimates the demand for international tourism for the Italian Province of Sassari. The sample period under estimation is from 1972 to 1995. Three dynamic models are estimated at monthly, annual and quarterly data frequencies. Similarities and differences are explored amongst the three models, using recently developed econometric techniques. A "pre-modelling" data analysis is undertaken for the economic series of interest. By adopting the LSE "general-to-specific" methodology, dynamic estimations are run. A full range of diagnostic tests is provided. Short and long run income elasticities, negativity and substitutability are tested on the light of economic theory. On balance, evidence is found that the monthly and quarterly models present homogenous results in terms of seasonal and long run unit roots. Annual data show different and perhaps misleading results.