Leading indicator tourism forecasts

Abstract

Leading indicators have been widely used in general business forecasting situations, but only rarely in a tourism context. In this study leading indicator transfer function (TF) models are developed to generate forecasts of international tourism demand from the UK to six major destinations. The out-of-sample forecasting accuracy is compared with the accuracy of forecasts generated by univariate ARIMA and error correction models (ECMs). The inclusion of a causal input within an ARIMA time series framework (TF model) does not result in an improvement in forecasting performance. The time series models outperform the ECM for short-term forecasting, but the ECM generates more accurate longer-term forecasts. © 2003 Elsevier Science Ltd. All rights reserved.