Purpose – The purpose of this paper is to analyze the technical efficiency (TE) of the rainfed rice farming system in Nigeria. Design/methodology/approach – Special attention is directed to the construction of confidence interval (CI) for individual efficiency estimates using stochastic frontier production model. Findings – The results show that the coefficient of the regression monotonically and significantly increased output level, while returns-to-scale (RTS) of 0.943 suggest decreasing RTS in the study. An average point estimate of TE of 0.669 is obtained from the analysis. CI show that the TE intervals are quite wide, as the assumption that farms originally identified to be on the frontier, or very close to the frontier, by the point estimate may be below the frontier. Age and years of schooling of the farmers are efficiency-increasing policy variables found in the study. Originality/value – Contrary to the previous efficiency studies, the present study demonstrates that economic inferences drawn on the basis of CI seem plausible to provide much more fruitful and insightful information about the TE of farms in the country, rather than the point estimation alone.