Forecasting models for tourism demand in city dominated and coastal areas

The paper uses Structural time series models to forecast the demand for tourism by nationality in coastal and city-dominated Spanish provinces. Intervention variables are introduced to take account of sudden shocks to tourism demand, such as the bombing of Libya and football's World Cup. The model demonstrates the considerable differences in demand by nationality, and in seasonality, which can occur at the subnational level. The Structural model generally provided more accurate forecasts than Box-Jenkins models. The results indicate that, ceteris paribus, the tourism demand growth rates in the Spanish provinces considered are unlikely to revert to their previous high levels.