Border Tourism Demand from GMS Countries to Thailand: X12ARIMA and TRAMO/SEAT Model

Aims: To find out suitable tourism forecasting models of each border checkpoints under the concern of seasonal adjustment.
Study Design: Time series study.
Place and Duration of Study: We used secondary time series data from immigration Bureau office 2004-2010.Three GMS countries of origin (Myanmar, Laos PDR and Cambodia) for border tourists to Thailand with 12 immigration border checkpoints, selected on the basis of the number of tourists. The time series from January 2004 to December 2010 that is 84 observations per border checkpoint.
Methodology: The first stage of this methodology applies unit roots test in Thailand border tourism demand and the second stage develops a forecasting model to estimate the number of border tourists’ demand in each border check point by X12ARIMA and TRAMO/SEAT.
Results: The results reveal that there are existences of unit root in data series. And we require X12ARIMA to modeling six from twelve data series and TRAMO/SEAT model also shows the suitable model for each border.
Conclusion: However both X12ARIMA and TRAMO/SEAT are seasonal adjustment models, the results still have difference in some data series. The user should concern about using the most suitable model or use together by weighting the result.