FORECASTING U.S. TOURISM DEMAND IN BERMUDA AND THE WEST INDIES

Abstract

This study reviews the literature on the explanation of tourism demand and demonstrates the application of least squares regression, smoothing methods and Box-Jenkins analysis to the problem of estimating future levels of tourism demand in selected Caribbean and Atlantic destination areas. The results support the conclusion that it is possible to make reasonably accurate one- to three-year forecasts of tourism demand by assuming that there are no significant changes in demand conditions. No single technique generates the most accurate forecast for all of the destinations considered, and some approaches generate less accurate results than can be achieved by the use of a "naive" forecast, in, which the most recently observed level of demand is used to predict future levels. However, in part, this latter outcome reflects inadequacies in the data base that must be used for prediction.