The system of cereal grain production in Ethiopia has been stratified into private, cooperative and state farms since the 1975 land reform. The private farms are being gradually replaced by the cooperative farms while the state farms are receiving increased technical and budgetary assistance by the Ethiopian government. lt is, however, not clear if these policies are consistent with the technical characteristics of these three modes of production. This study, therefore, evaluated technical efficiency, impacts of known and latent input factors, and returns to scale parameters for each farm type. Sample data were collected from Ethiopia on five cereal crops, namely, barley, com, sorghum, teff and wheat and several input factors, including labor, land, oxen, traditional farm implements, tractors, machinery services, modern yield·increasing inputs, livestock, education and rainfall over 77 awrajjas for the 1980-1986 production period. A covariance regression model was applied with these data to determine an appropriate functional form between the Cobb-Douglas and translog production functions. The Translog functional form was selected for the analysis on the basis of statistical tests. Results of the analysis suggest that the producer cooperatives collectively appear to have a potential to generate increased gross income per hectare at a declining rate with respect to an equiproportionate increase in all inputs, except land, upon an increasing average cost of production per unit of cereal output. The private and state farms appear to be operating with a close to fixed proportions type of production technology with a constant average cost of production per unit of cereal output per hectare. Moreover, the range of substitutability between input factors tends towards a complementary relationship as the institutional transformations and management techniques of the cereal producing farms shift from the traditional to a more advanced and centrally managed state mode of production. Partial income elasticity parameters suggest that (a) the private sector’s gross income per hectare is most responsive to traditional hand tools, fertilizer, labor, human capital at primary level of education, and rain in August and September; (b) the cooperative sector’s gross income per hectare is most responsive to the use of tractors and September rain; whereas (c) gross income per hectare of the state farms is most responsive to the use of traditional labor, machinery services and rain in June and August. Thus, Ethiopia’s agricultural income production per hectare is likely to be revitalized by: (a) qualitative changes in the traditional inputs, water management, and introduction of modem technical inputs such as fertilizer and farmer education in the private sector; (b) increased traditional labor employment, improved management of water, machinery and modem yield-increasing inputs on the state farms; and (c) a better usage of tractors and collaborative input factors, improved water management, and a substantial increase in capital investment to achieve full employment of the seemingly redundant labor and oxen input factors on the cooperative farms. It seems unlikely that the producer cooperatives will achieve the goal of maximum cereal output per hectare with the most prevalent composition of the redundant traditional input factors which contribute insignificantly at the margin without a major change in the current production techniques and structural policies of the sector. / Ph. D.