Forecasting performance of smooth transition autoregressive (STAR) model on travel and leisure stock index

Umer U. M., Sevil T., SEVİL G.

Journal of Finance and Data Science, vol.5, no.1, pp.12-21, 2019 (Scopus) identifier

  • Publication Type: Article / Article
  • Volume: 5 Issue: 1
  • Publication Date: 2019
  • Doi Number: 10.1016/j.jfds.2018.02.004
  • Journal Name: Journal of Finance and Data Science
  • Journal Indexes: Scopus
  • Page Numbers: pp.12-21
  • Keywords: Nonlinear time-series, Out-of-sample forecasting, Smooth transition autoregressive, Travel and leisure
  • Anadolu University Affiliated: Yes


© 2018 The AuthorsTravel and leisure market records a consecutive robust growth and become among the fastest economic sectors. Numerous studies proposed distinct forecasting models to predict the dynamics of this sector and provide early recommendations for investors and policy makers. In this paper, we compare the forecasting performance of smooth transition autoregressive (STAR) and linear autoregressive (AR) models using the monthly returns of Turkey and FTSE travel and leisure indices from April 1997 to August 2016. MSCI world index used as a proxy of the overall market. The result shows that nonlinear LSTAR model cannot improve the out-of-sample forecast of linear AR model. This finding demonstrates little to be gained from using LSTAR model in the prediction of travel and leisure stock index.