Efficiency has been proven to be an important managerial tool in improving total factor productivity in agriculture. The four sources of economic inefficiency: allocative, technical, pure technical and scale inefficiency of a sample of seventy-six Nepalese rice farmers were examined using the data envelopment analysis (DEA) decomposition method. The inefficiency indices computed by the DEA were then used as the dependent variable in a Tobit (censored) regression model using decision- makers’ attributes as the explanatory variables. The results revealed relatively large inefficiencies among the farms sampled. The average economic, allocative, technical, pure technical and scale inefficiencies were 34%, 13%, 24%, 18%, and 7% respectively. There is also a significant variation in the level of inefficiency across farms. The inefficiencies are attributed to the variations in use intensities of resources such as farm land, seed, human labour, fertilisers and mechanical power. The inefficiencies were explained by farmers’ level of age, education, gender, family labour endowment and risk aversion. Therefore, to be successful the efficiency improvement program must be flexible enough to accommodate the diversity of farmers and their improvement needs. The results also suggest that increasing effort towards educating farmers on best practices can enhance efficiency in Nepalese rice farming. Other policy implications and development strategies are also drawn from the findings.