Planning, scenarios and Bayesian forecasting models

Planning and forecasting are two related activities in which the construction of scenarios for the future depends on the type of model considered. On a changing environment, it is crucial to select an appropriate methodology or even, in more complex situations, go through the modelling process with more than one methodology. This article briefly presents several forecasting methodologies along with their applications to planning. It also explicitly describes the relation between business planning and forecasting as a decision process oriented for the future which depends on various types of judgment. It shows the importance of classifying forecasting models and of the modelling structure. It also deals with the process of scenario construction. Finally, it points out how the Dynamic Bayesian Models, associated with the ‘Principle of Management by Exception’ can be a very appropriate toll for obtaining scenarios and forecasts. An application of this technique to the Brazilian Economy is presented, using the Wholesale Price Index.