Aims: To evaluate the technical efficiency (TE) in selected agricultural sub-sectors and to propose possible policy interventions to the government with the aim of reducing the poverty of farmers in the developing world. Study design: A meta-analysis based on empirical studies conducted by various scientists throughout the developing world. Methodology: Research articles for the meta-analysis were selected using a thorough screening process based on the PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analyses) concept. Selected 94 articles were sub-divided in to three main agriculture sub-sectors for detailed analysis; (a) paddy, other field crops-OFC and vegetables, (b) fruits, and (c) livestock. Mean TE of each crop or livestock type was calculated by averaging the TE values for a particular crop or livestock type across different studies included in this study. Results: TE data presented in the original articles showed a considerable dispersion within a given study. The highest mean TE was recorded in B-onion (0.83±0.15) whereas the lowest was recorded in maize (0.703±0.09) and in soybean (0.705±0.13). The TE of chili cultivation was 0.78 with the greatest variability (standard error of mean [SEM] 0.19) among the crops considered, which signifies the unpredictable nature of the chili cultivation. Mango was found to be the least technically efficient crop among the studied, with a mean TE of 0.596±0.11. Dairy, poultry and aquaculture farming operations were found to be highly technically efficient having mean TE values of 0.80±0.16, 0.89±0.02 and 0.88±0.08 respectively. Conclusion: Findings of this study will lead to several key policy implications including, improvement of the socioeconomic characteristics of farmers, implementation of farmer field schools (FFS) and establishment of a cautious and gradual strategy for expansion of the rural financial institutions.