The paper employed a Cobb-Douglas stochastic production frontier function to estimate of technical efficiency (TE) with 770 observations of the cooperative farmer group (272, 243, and 255 in seasons 1, 2 and 3, respectively), and 900 observations of the non-cooperative farmer group (300, 324, and 276 in seasons 1, 2, and 3, respectively). The findings showed that TE of the cooperative farmer group was greater than that of the non-cooperative farmer group in season 1, while TE of cooperative group was lower than that of non-cooperative group in seasons 2, and 3. More so, the positive determinants of TE of the cooperative farmer group were the land area and plant density in three seasons, while the negative factors were the market access in season 1, the age and credit access in season 2. In the non-cooperative farmer group, the positive determinants were the plant density and land area in three seasons, the market access in season 2, whereas the negative elements were the wrapping bag in seasons 1, and 3, and the classifying sale in season 2. The paper helps more understanding role of TE as well as identify some technical constraint exists. This helps policy and decision makers associated with farming practice offer policy implications in order to improve productivity in mango farming. A difference of the study compared to others on tropical fruits, the paper analysed the TE of the cooperative and non-cooperative farmer group during three seasons instead of only during one or throughout an entire year.

Keywords— Technical efficiency, cooperative, non-cooperative, key mango varieties