Persistence in the Short- and Long-Term Tourist Arrivals to Australia

Abstract

This study examines the persistence in the international monthly tourist arrivals to Australia by using a variety of models based on fractional integration and seasonal autoregressions. The results based on disaggregated monthly data indicate that the series are on average mean reverting, though highly persistent, and present intense seasonal patterns. A forecasting performance of several competing models was also conducted, and it was concluded that the models based on long-range dependence outperform other more standard ones based on nonseasonal and seasonal unit roots. The model based on long memory at zero and the seasonal frequencies seems to be the most accurate in this context. More detailed analysis of the results is also derived.