INTRODUCTION Yam (Dioscorea spp) has been described as one of the major staple in West and Central Africa where it provides food for over 160 million people (Orkwor et al. 1995). Average statistics shows that the West African yam belt produced 95% of the world’s output of 34 million metric tonnes (mmt) of yam in 2001 and Nigeria alone produced 75% of West African output. The yam tuber is a good source of energy mainly from their carbohydrate contents since it is low in fat and protein. The yam tuber is said to contain pharmacologically active substances such as dioscorine, saponin and sapogenin. Dioscorine, which is the major alkaloid in yam, is medicinally a heart stimulant (Eka 1985). Also, it has been reported that yam is a good source of industrial starch whose quality varies with species. Despite this importance of yam, its production in Nigeria has not been accorded the needed attention (Orkwor and Asiedu 1999). This is reflected in the fall in output percentage growth © Kamla-Raj 2010 J Soc Sci, 24(2): 143-148 (2010) Determinants of Yam Production and Technical Efficiency among Yam Farmers in Benue State, Nigeria J. F. Shehu*, J. T. Iyortyer**, S. I. Mshelia*** and A. A. U. Jongur*** *Department of Agricutural Economics and Extension, Adamawa State University, Mubi, Nigeria **National Business and Technical Examinations Board (NABTEB) Yola, Zonal Office ***Department of Agricultural Economics and Extension, Federal University of Technology, P. M. B. 2076, Yola, Adamawa State, Nigeria, Postal code: 640001 KEYWORDS Inefficiency Effects. Cobb-Douglas. Stochastic Frontier Production. Kwande L.G.A ABSTRACT This paper investigates the determinants of yam production and technical efficiency of yam farmers using stochastic frontier production function which incorporates a model of inefficiency effects. Farm-level data were collected from a sample of 100 yam farmers in Benue State using structured questionnaires. The empirical results indicate that land, seed yam, family labour and fertilizer were the major factors that influence changes in yam output. Farmer-specific variables such as education, membership of association and household size were found to have significant effects on the observed variation in technical efficiency among the yam producers. The technical efficiency of farmers varied from 0.67 to 0.99 with a mean of 0.95. The implication of the study is that efficiency in yam production among the farmers could be increased by 5% through better use of land, seed yam, family labour and fertilizer in the short term given the prevailing state of technology. This could be achieved through policy interventions that would contribute to better access to land, improved seed and fertilizer as well as provision of labour saving technologies to ease farm operation. Also, improved farmer’s educational levels through adult education and literacy campaign would probably increase efficiency in the long term. rate of yam from 42% in 1990 to 16.3% in 2001 despite the increase in land devoted for the production of the crop from 1270 million hectares to 2742 million hectares in the same period (Federal Ministry of Agriculture, FMA 2001). Since increased productivity is directly related to production efficiency, it is imperative to raise productivity of the farmers by helping them reduce technical inefficiencies. Efficiency is concerned with the relative performance of the processes used in transferring given inputs into outputs. Farrel (1957) identified three types of efficiency- technical, allocative and economic. An important assumption relating to efficiency measurement is that firms operate on the outer bound production function, that is, on their efficiency frontier. When firms fail to operate on outer bound production function, they are said to be technically inefficient. The stochastic frontier production function, which is often used for efficiency studies was first independently proposed by Aigner et al. (1977) and Meeseun and van den Broeck (1977). A stochastic frontier production function comprises a production function of the usual regression type with a composite disturbance term equal to the sum of two error components. One error component represents the effects of statistical noise (e.g. Corresponding author: E-mail: fintan_j@yahoo.com; Telephone: +2348065653142