Date of Award:

1993

Document Type:

Thesis

Degree Name:

Master of Science (MS)

Department:

Economics and Finance

Advisor/Chair:

Jay C. Andersen

Abstract

Potatoes are a capital-intensive crop. A farmer who is considering expanding his potato acreage must carefully consider revenue requirements to offset the high costs of raising the crop. A method to forecast annual farm potato prices would be useful not only to the farmer, who is considering potato acreage expansion (or contraction), but also to the potato buyers.

Seven forecasting models were considered: (1) a simultaneous equation model (with five equations); (2) a Box-Jenkins type ARIMA model; (3) an exponential smoothing model; (4) a moving-ave rage model; (5) a trend model; (6) an "opposite" model; and (7) a current. or naive, model.

The results reveal the following three things: (I) The "best" model was the trend model. This model gave the most accurate one-period out-of-sample forecasts of the models tested (as measured by the mean absolute error (MAE), the root mean squared error (RMSE), and Theil's U2 statistics). The simultaneous equation model could be considered as the next best model. (2) The forecast for the average Utah farm potato price for 1992 was about $5.40 per cwt. (3) The average Utah farm potato price for 1993 should be in the $5.51 to $5.95 range (the forecasts from the trend and simultaneous equation models, respectively).

Included in

Economics Commons

Share

COinS