Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)


This study utilizes the Regression, Exponential Smoothing, Census X-11 and Box-Jenkins techniques to simulate the historical time series on selected livestock and crop prices and quantities in Louisiana. The data period covered was from 1972 through 1983 for the monthly time series, and from 1924 through 1983 for the annual time series. The price and production situation for the selected time series was also reviewed for the period 1972-1983. The variability among and within the selected time series was also evaluated. The methodology in this study provided a blend of economic and statistical frameworks. The economic framework provided the medium for explaining how the time series components--trend, seasonal, cyclical and irregular--contribute to the overall variation in any given time series. The statistical framework reinforced the economic framework quantitatively. It was evident from the review of the agricultural situation in Louisiana for the period 1972-1983 that both cyclical, seasonal and irregular factors have changed with the general trend underlying the agricultural series considered in this analysis. It was also found that over the years, livestock prices in Louisiana have had a larger relative variation than crop prices--except for sweet potatoes. Among the crops, soybean prices were the least variable. By utilizing some measures of accuracy statistics to evaluate the ex-post forecast estimates generated by the forecasting techniques, it was found that the Box-Jenkins technique consistently out-performed the other techniques. For seasonal adjustment purposes, the Census X-11 and exponential smoothing provided better seasonally adjusted estimates than the regression procedure. However, in both the seasonal adjustment and forecasting evaluations, unique contributions were realized from each technique, for some of the selected time series. Combined forecasting provided the minimum mean squared error estimates for at least 94 percent of the monthly and annual data studied.