Abstract
The advantages of error correction models (ECMs) and time varying parameter (TVP) models have been discussed in the tourism forecasting literature. These models are now combined to give a new single-equation model, the time varying parameter error correction model (TVP-ECM), which is applied for the first time in the context of tourism demand forecasting. The empirical study focuses on tourism demand, measured by tourism spending per capita, by UK residents for 5 key Western European destinations. Based on the discussion of how the series considered related to most, the empirical results show that the TVP-ECM can be expected to outperform a number of alternative econometric and time series models in forecasting the demand for tourism. By measuring performance in terms of the accuracy of the forecasts of growth (rates of change) and showing that TVP-ECM performs very well for this as well as conventional assessment of the level of demand in this study, it is suggested that forecasters of tourism demand levels and growth rates can feel comfortable using TVP-ECM given that it is expected to perform well.