Learning and Ability to Pay: developing a Model to Forecast Ski Tourism

The decision whether to use time series or econometric methods to forecast demand is not clear. The literature reviewed only indicates that models should he simple and ideally be able to evolve over time. In 1997 two models were proposed to forecast the numbers from Britain skiing in Europe.The first used it learning curve approach and forecast a stationary market, whilst the second used a Varying Coefficient Model linking sales and ability to pay and forecast a gradually expanding market. This paper reviews the outcomes 1996-2000, the forecast performance of the two models and the stability of the structure of both when updated. It unequivocally suggests that the learning curve approach produced better forecasts. In the penultimate section a model that attempts to combine both approaches is developed. In this context the role of "historic" data is discussed. The paper concludes that the best forecasting approach will depend upon whether the market is stable and that the weight given to data must reflect the information content of that data.