Use of Planting Intentions to Predict Actual Plantings

This study deals with the usefulness of farmers' planting intentions for predicting acres actually planted to six crops: corn, soybeans, sorghum, oats, barley, and spring wheat. The statistical results indicate that the use of both USDA estimates of planting intentions and objective data, such as prices, explained (a) higher proportions of variance in acres planted to oats, barley, and spring wheat than are accounted for by the use of either type of data alone; (b) a higher proportion of variance in corn and sorghum acres planted than does use of prospective plantings alone. The use of prospective plantings and objective data together did a better job of reproducing turning points in acres planted than did the use of objective data alone for each of the crops except oats. A number of economists have studied the forecasting value of various sorts of subjective variables: anticipations, plans, and expectations. Okun [21; 22], Jorgenson, Hunter, and Nadiri [12]; Rippe and Wilkinson [4]; Keezer, Ulin, Greenwald, and Matulis [14]; and Levine [16] have studied the forecasting accuracy of data on anticipated investment expenditures. Juster and Wachtel [13], Burch and Stekler [2], and Mueller [191, have studied the usefulness of consumer anticipations and intentions for forecasting purposes. Brown [1], Lovell and Hirsch [17], Orr [19], and Pashigian [23] have studied the forecasting value of data on anticipated sales and inventories. Evans devoted chapter 15 of his book [3] to use of data on consumer attitudes and buying plans; and investment, sales, and inventory anticipations in forecasting. The findings from these studies are mixed. One intriguing result from some of these works is that a better forecast is made by using subjective data and objective data together than is made by using either alone. PURPOSE For some twenty-five years the U.S. Department of Agriculture has reported March 1 intended plantings of various crops [30], but no systematic studies of the forecasting value of these data have been published.1 The purpose of the research reported here was to investigate the usefulness of the March 1 intended acres for predicting actual acres planted. Consider a person concerned with forecasting the number of acres planted to a crop. Late in year t or early in year t+1, to forecast the number of acres planted in year t+1, he can use such objective data as acres planted and prices in year t and measures of government programs in year t+1. After the March 1 planting intentions are known, can he increase the accuracy of his forecast by use of these subjective data? Should he continue to use the same equation he used earlier, the equation containing only objective data? Should he use a simple regression of actual acres on intended acres to predict actual acres for t+1 from intended acres for t+1? Should he use an equation that contains intended acres and objective data? Our purpose was to provide answers to these questions.