Date of Award
Master of Science
Charles B. Sappington
Darell Mundy, John Brooker
The primary purpose of this study was to develop and test a pre-diction equation for broiler prices eleven weeks in the future using currently known weekly data. The method used to develop this model was to use multiple regres-sion in a modified demand equation for broilers. The modifications were necessary because the values of all explanatory variables had to be known 11 weeks prior to the price being predicted (the dependent vari-able) . Eleven weeks were used since there is usually an eleven week period between egg set and marketing the product. Several modifications of the equation were tried with the follow-ing model being chosen as the one to be used for prediction: PER = 82.364 + .289 Y - 275.77 1/FH -1.246 EST - .107 GST + WD + HD Where: PER = price of broilers in cents per pound Y = average weekly income in dollars for blue-collar workers in the U. S. FH = futures price of hogs, 11 weeks forward, in cents per pound EST = egg set in millions of eggs weekly GST = cold storage in millions of pounds of frozen broiler meat WD = weekly dummy variables with week 52 left out to prevent a singular matrix HD - holiday dummy variables with non-holiday periods left out to prevent a singular matrix The predictions of this model were not acceptable as compared to the futures market estimates for broilers, therefore, the most current known residual, or difference between actual and predicted prices, was added to the predictions. However, since these predictions still had wild fluctuations, the two-week, three-week, and four-week moving averages were taken in order to smooth out the peaks and troughs. These predictions along with the futures market estimates were then compared to the actual cash prices for 1974. The results of these comparisons were that the model's predictions and the futures market estimates were not significantly different from each other but that all of the estimates were significantly different from the actual broiler prices for the same time period. The one central conclusion of this study is that if broiler producers wish to have a reasonable estimate of a price to expect for their product, there is no known vehicle to arrive at expected broiler price more accurate than the report in the newspaper of futures market prices 11 weeks hence.
Asbridge, David Donald, "A regression model for predicting broiler prices. " Master's Thesis, University of Tennessee, 1976.