Les prévisions conditionnelles sont-elles plus précises que les prévisions inconditionnelles dans les projections de croissance et d’inflation en zone CEMAC ?[Should conditional forecasts of inflation and real growth more accurate than unconditional forecasts in CEMAC subregion ?]

This study compares the predictive performance of the conditional forecasting technique against the unconditional technique. The conditional technique consist of taking into account the information available on an endogenous variable over part of the forecast horizon. We develop a Bayesian VAR model with three endogenous, real growth, inflation and monetary growth, in which we condition the evolution of monetary growth by considering three types of scenarios : basic, optimistic and pessimistic. Two main results can be draw from our simulations : (i) the conditional forecasting approach is generally more precise than the unconditional approach ; (ii) the uncertainty around the central forecast is reduced with the conditional forecast technique. These results therefore call on the central bank to adopt the conditional forecasting technique in projections of real growth and inflation ; but also to consider various scenarios on the variable to be conditioned.