A Friction Model for Describing and Forecasting Price Changes

This paper presents a new friction model for describing the price changes of a product or brand over time and for forecasting both the timing and magnitude of such changes from one period to the next. After a review of the related pricing literature, we present our model and a modified controlled random search procedure for estimating its parameters. The model is applied to describe and forecast the weekly mortgage interest rates for a local bank in the Philadelphia area. Finally, limitations and several potential applications of this methodology are discussed.