Tourism Demand Forecasting

Tourism demand forecasting predicts future tourist arrivals, overnight stays, or expenditure from historical data, supporting planning by destinations, airlines, hotels, and policymakers. The field spans two broad model families. Time-series models such as seasonal ARIMA (SARIMA) extrapolate the patterns embedded in the demand series itself — trend, seasonality, and autocorrelation — without explanatory variables. Econometric models such as autoregressive distributed lag models (ADLM) and error-correction models relate demand to drivers like income, relative prices, and exchange rates, allowing both forecasting and policy analysis. Haiyan Song and Gang Li's influential 2008 review in Tourism Management synthesized this literature, documenting the proliferation of methods since 2000 and emphasizing rigorous out-of-sample evaluation. Their work, with Stephen Witt, helped make tourism demand forecasting a methodologically mature subfield.