Estimation of a Multi-Input Multi-Output Model of Lot-Fed Beef Cattle in Australia

Most analyses of technical efficiency and productivity focus on the activities of firms or institutions using inputs to produce outputs. In this paper, we report the results of an efficiency analysis that is couched in a different context. We analyse the efficiency with which physical characteristics of individual lot-fed Australian beef cattle are combined with conventional inputs to produce a final product with several attributes that consumers value. Data on 214 animals across seven breeds were used to estimate a stochastic input distance function with multiple inputs and multiple outputs. Estimates were obtained after controlling for differences between breeds, seasons and sex. Because of variations in feeding regimes for different markets, the data set was confined to animals whose meat was destined for the Korean export market. Technical efficiency measures are reported for each animal, along with the estimated mean technical efficiency. The conventional inputs included in the analysis are feed and number of days in a feedlot (reflecting capital, veterinary and supervisory inputs). Physical measures of cattle traits on their entry into the feedlot are age, liveweight, muscle score, eye muscle area, rump fat depth and rib fat depth. Outputs are carcass weight, meat quality (assessed by using sensory scores on tenderness, juiciness, flavour and overall acceptability), marbling and the proportion of meat weight retained after cooking. A high mean technical efficiency was estimated for the cattle but it was found that breeds have significantly different output frontiers and inefficiency levels. These differences are most likely associated with variation in genetic merit between sires within a breed and the different farming and climatic backgrounds of cohorts of animals. Results are reported on the input-output relationships as well as the relationships between the outputs. Important findings in respect of output relationships in this sample are that the proportion of meat weight retained after cooking is highly significantly and positively associated with the meat quality sensory score, and carcass weight is significantly and negatively associated with the meat quality sensory score. No significant relationship was discerned between carcass weight and the proportion of meat weight retained after cooking.