Technical efficiency of smallholder sunflower farmers in Singida, Tanzania: the application of the zero-observations approach

Improving agricultural productivity is crucial for enhancing income and alleviating poverty for the majority of farm households in sub-Saharan Africa. However, fostering productivity requires a good understanding of farm-level technical efficiency (TE). While earlier studies have examined TE across the globe, the current study applies the rarely used zero-observation approach to compare the performance of the Translog and the Cobb-Douglas stochastic frontier models, based on 269 randomly selected smallholder sunflower farmers in Singida, Tanzania. The findings suggest that the Translog stochastic production frontier model fits our data better than the Cobb-Douglas stochastic production frontier model in the context of the zero-observation approach. The zero-observation inputs such as improved sunflower seeds and non-zero observation inputs such as land significantly influence sunflower output and technical efficiency. In addition, using this approach, some sunflower farmers were found to be fully technically efficient. Thus, the findings suggest that the provision of improved sunflower seeds is imperative.