A PLICATION OF Bayesian decision theory to problems facing agriculture has come during a decade of continually increasing interest in the weather [12]. The National Weather Service spends about 70 million dollars annually observing, analyzing, and forecasting weather phenomena. In addition to the usual general forecasts, specialized forecasts of particular value in local areas are used in agricculture as well as other industries [13, p. 86]. In the future forecasts of specific weather phenomena will undoubtedly become increasingly available, either as a public service from the Weather Service or for sale by private consulting firms. The public forecasting service would want to avoid the cost of issuing forecasts that cannot profitably be used by the clientele in the forecast area. Or farmers purchasing private forecasts would want to avoid paying a price that exceeds their value. Thus, as the capability of issuing forecasts increases, it seems reasonable to assume that agricultural economists will be called upon to estimate their value. The Bayesian decision model would seem to be suited to the task of estimating the value of weather forecasts. The model itself has been extensively developed in the literature [5, ch. 5 and 6; 9; 11, ch. 5; 14; 15]. Its use in price forecasting has been demonstrated by Eidman, Dean, and Carter [7] and Bullock and Logan [1]. Carlson [4] illustrated its use for crop disease forecasting. Byerlee [2] and Byerlee and Anderson [3] have used it to evaluate a long-range weather forecast. Estimates of the production function, the prior probability distribution for weather inputs, and the likelihood function are needed to compute the value of a forecast. The production function would include the random weather variables to be forecast as well as a set of inputs under the manager's control. It would be used to compute the profits from alternative actions and suggest the form of the forecast. Unfortunately, when using Bayesian analysis the information needed to estimate the value of the forecast is the same information needed to
[1]
Harold O. Carter,et al.
An Application of Statistical Decision Theory to Commercial Turkey Production
,
1967
.
[2]
Samuel H. Logan,et al.
An Application of Statistical Decision Theory to Cattle Feedlot Marketing
,
1970
.
[3]
W. J. Maunder,et al.
The Value of the Weather
,
1971
.
[4]
Clifford Hildrethi.
BAYESIAN STATISTICIANS AND REMOTE CLIENTS
,
1963
.
[5]
Jock R. Anderson,et al.
Value Of Predictors Of Uncontrolled Factors In Response Functions
,
1969
.
[6]
Lester B. Lave,et al.
THE VALUE OF BETTER WEATHER INFORMATION TO THE RAISIN INDUSTRY
,
1963
.