Décoder l’économie pour mieux prévoir : cas de Madadagascar

This paper presents a comparison of econometric models used for economic forecasting, focusing on Madagascar’s Gross Domestic Product (GDP). The models analyzed include OLS, ARIMA, ARIMAX and VAR, each with its advantages and limitations. The OLS model is selected for its robust performance according to AIC and SIC criteria, although a hybrid approach or the integration of modern techniques such as neural networks is recommended for more accurate forecasts. Finally, the paper underlines the importance of foreign trade (exports and imports) in economic dynamics, while stressing the need to contextualize results for informed policy decisions.