Forecasting N.S.W. Beef Production: An Evaluation of Alternative Techniques

This paper reports on the evaluation of the performance of several forecasting methods used to forecast New South Wales quarterly beef production, one quarter ahead. The forecasting procedures used are a single equation regression model, a Box-Jenkins univariate time series model, a forecasting committee's judgement and a naive model. Absolute accuracy and relative accuracy measures are used to evaluate ex ante forecasts. Although the evaluation gave some mixed results according to the criteria used, it appears that the forecasting committee performed better than the alternative forecasting procedures considered in this study. However, the results indicate that the committee's performance was not much better than that of a naive (no change) model, indicating there is room for improvement