Prévisions de l’inflation et de la croissance en zone CEMAC

The main objective of this study is to look for the best model for forecasting inflation rate and real growth for each CEMAC country. Using AR, VAR and BVAR models, it is clear from our study that forecasts made from Bayesian models have a higher predictive power than those made by classical approaches. However, in the very short term, classical univariate and multivariate models have better results. The forecasts obtained using our models are in most cases similar to those made by the IMF. We also find that the fancharts proposed in our models can contain the majority of forecasts made by the IMF. Since the forecasting exercise is very complex, because it depends on exogenous factors that are sometimes unpredictable, it would be advantageous for the BEAC to add in its projection tools, the fancharts approach in order to put more emphasis on the intervals of credibility instead of focusing only on specific points. This logic used in most central banks has the advantage of providing some flexibility to the conduct of monetary policy.