An annual econometric beef model with imposed biological restrictions

Abstract

After reviewing past beef and livestock econometric models, economic theory, and econometric techniques for simultaneous systems and their simulation, an annual econometric system of equations was estimated for the U.S. beef economy. Care was taken in each OLS specification to include relevant economic variables and to have parameter estimates which exhibit strong t-statistics and proper signs. High coefficients of determination were also desirable. In order to incorporate as much known information of the beef industry as possible, long term biological averages were imposed in supply equations at the farm level. Resulting retail demand elasticities were consistent with those of previous studies. The 29 equations were simulated as a simultaneous system generating simulation statistics. These statistics are reasonably good which implies that the system of equations captures the underlying structure of the beef industry fairly well. Key individual equations were subjected to non-nested tests to determine the level of significance that one specification explains more of the variation in the dependent variable than does the alternate specification. This model containing economic and biological information was compared to an established model developed without biological information imposed on its estimation. The previously developed Yanigida model simulated quite well and marginally better than the biologically constrained model in some cases. A recommendation was made for improving the performance of the biologically constrained model by imposing short-term biological averages, possibly from the last 7-10 years,thus eliminating some of the trend these averages have taken over the past 22 years.