Effect of seasonality treatment on the forecasting performance of tourism demand models

Shujie Shen, Gang Li, Haiyan Song

Research output: Journal article publicationJournal articleAcademic researchpeer-review

33 Citations (Scopus)


This study provides a comprehensive comparison of the performance of the commonly used econometric and time-series models in forecasting seasonal tourism demand. The empirical study is carried out based on the demand for outbound leisure tourism by UK residents to seven destination countries: Australia, Canada, France, Greece, Italy, Spain and the USA. In the modelling exercise, the seasonality of the data is treated using the deterministic seasonal dummies, seasonal unit root test techniques and the unobservable component method. The empirical results suggest that no single forecasting technique is superior to the others in all situations. As far as overall forecast accuracy is concerned, the Johansen maximum likelihood error correction model outperforms the other models. The time-series models also show superior performance in dealing with seasonality. However, the time-varying parameter model performs relatively poorly in forecasting seasonal tourism demand. This empirical evidence suggests that the methods of seasonality treatment affect the forecasting performance of the models and that the pre-test for seasonal unit roots is necessary and can improve forecast accuracy.
Original languageEnglish
Pages (from-to)693-708
Number of pages16
JournalTourism Economics
Issue number4
Publication statusPublished - 1 Jan 2009


  • Econometric model
  • Forecasting
  • Seasonal unit roots
  • Seasonality
  • Time-series model
  • Tourism demand

ASJC Scopus subject areas

  • Geography, Planning and Development
  • Tourism, Leisure and Hospitality Management


Dive into the research topics of 'Effect of seasonality treatment on the forecasting performance of tourism demand models'. Together they form a unique fingerprint.

Cite this