Conception et estimation d'un modèle DSGE pour la prévision macroéconomique : un petit modèle d'économie ouverte pour le Cameroun

This thesis aims at analyzing the macroeconomic dynamics of the Cameroonian economy. It begins with a quantitative analysis of the business cycle in Cameroon, based on annual macroeconomic data, especially gathered for this purpose. This preliminary inquiry highlights a number of features that can be accounted for in a new-keynesian modelling framework. A dynamic stochastic general equilibrium (DSGE) model of the new-keynesian family is thus constructed as a mean of describing the salient features of the Cameroonian economy. It has the traditional blocks of new-keynesian DSGE models (Sticky prices and wages, adjustment costs, etc.). But it also accounts for a number of characteristics of the Cameroonian economy that are shown to be influential in the dynamics of the Cameroonian economy (e.g. oil revenues or primary goods exports). The model is then estimated and evaluated, based on a Bayesian approach. Its forecasting performance is also assessed through comparison to the performances of a random walk model, a vector autoregressive (VAR) model and a Bayesian VAR (BVAR) model. It turns out that, at least for short horizons, the DSGE model shows the highest performance. As to macroeconomic fluctuations, the estimated  model suggests that commodity price shocks generate an output expansion, an increase in employment and a fall in inflation. In addition, oil price shocks have a direct impact on marginal costs which increase and provoke a rising in inflation while output and employment tend to fall. Foreign shocks and domestic supply shocks account for a large share of output and investment fluctuations. The evolution of output over the whole sample is dominated by commodity price shocks and oil price shocks as one would expect.