AN INTERMEDIATE INPUT MODEL OF DAIRY PRODUCTION USING COMPLEX SURVEY DATA

Agricultural production is often characterised by multiple inputs and multiple outputs to multiple production processes. Where an output from one process is used as an input to another, this output is called an intermediate product. This is a common situation when a farm produces both crops and livestock. The analysis of production efficiency is important for the evaluation of agricultural policy, but until recently, no methods have explicitly included intermediate products. This study applies a non-parametric technique of efficiency measurement which includes intermediate products. The data set is a sample of dairy farms drawn using a complex survey design. The use of non-parametric efficiency measurement and the subsequent application of bootstrapping and kernel density estimation to the results allow inferences to be drawn concerning the whole population from which the sample was drawn. We find that the decomposition of production into subproduction processes reduces the dimensions of problem specification, with the effect that a larger number of variables may be usefully included in the model.