Using high-frequency data and time series models to improve yield management

We show the potential contribution of time series models (TSM) to the analysis of high frequency (less than monthly) time series of economic activity. The evolution of the series is induced by stable patterns of behavior of economic agents; but these patterns are so complex that simple smoothing techniques or subjective forecasting can not consider all underlying factors and TSM are needed if a full efficient analysis is to be carried out. The main ideas are illustrated with an apllication to Spanish daily electricity consumption.