Can Seasonal Unit Root Testing Improve the Forecasting Accuracy of Tourist Arrivals?

Abstract

Tourism has emerged as one of the main industries in many Caribbean countries. Jamaica is no exception: tourist receipts are its largest source of foreign exchange earnings. Because of the importance of the tourism sector to the Jamaican economy, accurate forecasts of tourist arrivals are critical. This paper argues that forecasting accuracy can be improved by the use of seasonal unit root testing. Using six tourist arrival series, from 1968:1 to 2001:3, it was found that unit root testing, applying the HEGY procedure, did help to improve the forecasts at all horizons. The exceptions were those series that exhibited considerable volatility; in these cases the seasonal unit root test did not prove superior at the short horizon. These results, however, may not be robust with respect to data at other frequencies.