Tests de G-causalité et spécification d’un modèle économétrique: Application sur un panel sectoriel marocain

The paper aims to use the Granger causality to deduce the structure of recursive model. Manipulating data from five sectors of Moroccan economy we form causal chain between endogenous variables to build a recursive system. The findings exhibit two group, the first one consists of agriculture, agro-industry and manufacturing sectors where the investment effort determines the balance trade and influences the cash-flow level. Meanwhile, in the second group formed by energy and mines sectors, the balance trade determines the investment effort and influences the cash-flow. The Granger causality justifies the modeling of the system. But, we cannot avoid ex-post the causality and exogeneity tests for the predetermined endogenous variables as Hausman and Holly tests. There tests are running once the model is estimated.