Is the time-varying parameter model the preferred approach to tourism demand forecasting? Statistical evidence

Shujie Shen, Gang Li, Haiyan Song

Research output: Chapter in book / Conference proceedingChapter in an edited book (as author)Academic researchpeer-review

1 Citation (Scopus)

Abstract

This study is one of the first attempts to apply rigorous statistical tests to examine whether a significant difference exists in the forecast accuracy between alternative models in the context of tourism demand forecasting. Two econometric models have been analysed in this study: the traditional regression-based static model and the TVP model. The empirical results show that the TVP model performs well in forecasting the demand for Thai tourism by tourists from seven origin countries. The TVP model outperforms its fixed-parameter counterpart in forecasting the demand for Thai tourism by all of the origin countries at different forecasting horizons. This study provides robust empirical evidence of the superiority of the TVP model. The results suggest that by taking into account the possibility of parameter changes in the demand model, forecast accuracy can be significantly improved. This conclusion is drawn based on the current empirical study, and further examination of this issue using different datasets is recommended.
Original languageEnglish
Title of host publicationAdvances in Tourism Economics
Subtitle of host publicationNew Developments
PublisherPhysica-Verlag HD
Pages107-120
Number of pages14
ISBN (Print)9783790821239
DOIs
Publication statusPublished - 1 Dec 2009

ASJC Scopus subject areas

  • Economics, Econometrics and Finance(all)
  • Business, Management and Accounting(all)

Cite this