Following the argument that growth models in developing countries shouldrnbe treated in different way from that of the developed).g countries, I tried torn/rnmodel economic growth in Ethiopia using data from 1960/61-1998/99.rnThe dependent variable is real GDP (in logs). The explanatory variablesrn(factors) used are gross fixed capital formation, labor (population adjustedrnfor activity rate), human capital index (calculated using gross enrollmentrnratio, years of schooling adjusted for return, world technology frontier andrnrate of adoption of technology), export, and mean annual rainfall (all inrnlogs). The analysis makes use of the Johansen maximum likelihoodrnrnestimation procedure. The main finding is that growth model in Ethiopia isrnbest specified by inclusion of non-traditional factors, in this case rainfallrnand export. In this co-integration analysis, all the explanatory variablesrnused with the exception of fixed capital formation are statisticallyrnsignificant with expected sign. Applying the Vector Error Correction Modelrn(VECM), it has been found that only rainfall and export among thernvariables used for this study are significant in the short-run.rnAlong with this, a total factor productivity type of method has been used torndetermine the contribution of sectors to growth. It has been found thatrnagriculture has been contributing to growth only half times its share in thernGDP while other sectors have been contributing almost twice their share inrnthe GDP.