Measuring and explaining technical efficiency of dairy farms: a case study of smallholder farms in East Africa

This article measures and explains the technical efficiency (TE) of 371 dairy farms located in 17 districts in East African countries. Three output and 10 input types were used to calculate the efficiency score for each farm. A two-stage analysis was conducted to measure and explain the efficiency scores. Firstly, the efficiency scores were measured by a data envelopment analysis (DEA) approach, which was implemented using a linear programming (LP) method. About 18 per cent of the farms were fully productive, each with an efficiency score of unity, which means this group is currently operating on the production possibility frontier. About 32 per cent of the farms had efficiency scores below 0.25, which means that about a third of the dairy farms would need to expand their dairy production by at least 75 per cent from the current level without any increase in the level of inputs. Secondly, a fractional regression method was used to explain the efficiency scores by relating them to a range of explanatory variables. The findings indicate that technology adoption factors, such as the existence of improved breeds, and feed and fodder innovations (e.g., growing legumes), have positive and statistically significant effects on the level of efficiency. Similarly, zero-grazing seems to have positive and highly significant effects. As far as marketing variables are concerned, selling milk to individual consumers or organisations seems to contribute positively and more significantly to dairy efficiency than other marketing outlets such as traders or chilling plants. Membership of a dairy cooperative has a positive effect but is not statistically significant.