In this paper, farm level technical efficiency of production and its determinants are investigated in a sample of 51 cereal producing farms located in the main cereal production region in Tunisia using a stochastic frontier production model. Empirical findings show that labor input factor appears with a minimal effect on the production. The hypothesis of constant returns to scale is rejected at the 5% level of significance, and returns to scale were found to be decreasing. Moreover, the estimated coefficients in the technical inefficiency model are also as expected. The estimated coefficients of the instruction level of farmer and the rotation, technical variable, are negatives and statistically significant at 5% level, which indicates their positive effect on technical efficiency. In addition, results indicated that estimated technical efficiency of cereal production in the sample varied widely, ranging from 52.63% to 94.62, with a mean value of 77%. This suggests that, on average, cereal producing farmers could increase their production by as much as 23% through more efficient use of production inputs. On a second step, Timmer and Kopp indexes of technical inefficiency were estimated for the same farms using a Cobb–Douglas frontier production function with a composite error term, and a developed relationship between these two indices. Results show that the mean values of the Timmer and Kopp TE indices were over 0.80, but one half of the farms were below 0.80 for the Timmer index and below 0.83 for the Kopp index. The level of inefficiency was found to be related to farm size: small and large farms were shown to be more technically efficient than medium-sized farms. With the given inputs, the production of cereals could be increased by 20% on average through making all farms 100% efficient. Alternatively, inputs could be reduced by 17% on average to produce the same amount of cereal output. Finally, the lower level of efficiency but higher yield and total factor productivity in the medium-sized farms means that more cereals can potentially be produced in these farms. The findings revealed that significant factors related to TFP were age, education level and the share of wheat crops into total cropped area. These results calls for policies aimed at provision of training programs, extensions services. In addition, the encouragement of experienced farmers by applying improved input management on these farms can be recommended alongside appropriate new technologies, especially for wheat farmers.
