Price forecasting models for soybeans, cotton, and tobacco in Tennessee

The primary objectives of this study were to formulate the demand and production response functions for soybeans, cotton, and tobacco at the farm level in Tennessee; to formulate the simultaneous equation systems to generate apparent equilibrium quantities and prices for each commodity; and to use simulation techniques to forecast the price changes at the farm level in Tennessee resulting from specified changes in exogenous explanatory variables. The analysis was done using secondary data from the period 1950-1976.

The demand and production response functions were formulated and estimated separately using OLS methods to obtain necessary information for each crop. Given the information from the OLS estimation, a simultaneous equation system was formulated for each crop. Each simultaneous system consisted of three behavioral equations — demand, production, and acreage — and three identities. Both 2SIiS and 3SLS techniques were employed in order to obtain consistent and efficient estimates.

The simultaneous equation systems were tested by evaluating "ex post" forecasts. The Mean Percent Error showed .31, .017, and .56 percent for soybeans, cotton, and tobacco, respectively, which is adequate and acceptable. Therefore, simulation techniques were used to obtain "ex ante" forecasts. Exogenous explanatory variables were projected first for the next decade under the assumption that the values of all exogenous variables would follow past trends. A "base model" was formulated from those projections, and forecasts from this model projected prices of the three crops.

In the simulation analysis, the levels of exogenous variables in the system were changed in order to determine the effects on prices of each commodity. The results showed that soybean prices were influenced significantly by disposable personal income. The analysis also showed that price support rates for cotton and the farm wage rate in Tennessee affected the farm level prices of cotton and tobacco in Tennessee, respectively.

Based upon the assumptions and results of the study, the prices of the three crops at farm level in Tennessee would increase continuously during the next decade. The forecasts showed that soybean prices in 1986 would be $12.05 per bushel, which is 91 percent higher than that in 1976. However, the prices of cotton and tobacco would be $0.81 and $1.73 per pound in 1986, respectively, which are 30 percent and 54 percent higher, respectively, than in 1976, if all of the exogenous variables increase following the past trends. So, soybean prices are expected to increase more rapidly than the prices of the other two crops.