Amino acid based soybean component pricing systems
Differences in pricing between systems were calculated and tested using non-parametric statistics. Protein was shown to have flaws as a proxy for price signal transmission, since a wide variety of protein percentages were observed at equal diet costs. Logistic regressions were calculated to approximate the probability that a soybean with a given protein content could outperform expectations based upon that protein content within an amino acid pricing system. These showed that lower protein beans are less likely to outperform protein based expectations, and probabilities improve for such an occurrence when protein levels are above average. Finally, regression equations of the prices in each system versus protein were examined, to construct a method for estimating the difference in value of a soybean sample in any two given pricing systems. The paper then concludes by identifying areas of future research, and speculating on the future structure of the market for soybeans.The history of soybean component pricing was examined as a starting point. The theoretical basis for component pricing was then developed in a macroeconomic context. Finally, a pricing system based upon the amino acid content of soybeans was built. To accomplish this, a data set of 268 soybean samples was used with animal feed diets from two species, broiler chickens and hogs. Diets were constructed at varying life cycle stages to create a total of seven diets. These diets were input into the Brill least cost livestock ration program to obtain variations in diet cost. The resulting variations were then translated into variations in marginal value product in order to price soybeans. This diet cost pricing system was then compared to two simpler pre-existing methods. It was found that the amino acid based pricing system was more accurate and had an effect on virtually all of the samples.