Résumé en anglais

This paper proposes an extension of the Granger (1969) causality definition to panel data models with fixed coefficients. Given the heterogeneity of the data generating process, four definitions of causality relationships are proposed. A procedure of nested tests is then defined and a particular attention is done to finite sample properties of these tests. Finally, an application to the link between financial deepening and economic growth is proposed with a panel of sixteen sub-Saharan countries over the period 1967-1998.