User Responses to Imperfect Forecasts: Findings from an Experiment with Kentucky Wheat Farmers

Abstract

Abstract Weather and climate forecasts can help agricultural producers improve management choices in anticipation of uncertain growing conditions. Current literature conjectures that the extent to which forecasts are useful depends on their accuracy, that is, the probability with which a forecasted event, such as precipitation, is projected to occur. Too little accuracy can potentially render forecasts effectively useless, even if they convey some form of information. In this study, we collect farmer-based data through a questionnaire and a framed field experiment to test for the existence of an accuracy threshold for forecasts, below which forecasts do not induce any behavioral changes. We do this in the context of a very specific management choice—the timing and amount of nitrogen that Kentucky farmers apply to their wheat in early spring—in response to randomly generated 6-to-10-day forecasts of rainfall conditions. We find that forecasts provide economically significant value to decision-makers only when they depart dramatically from what is normally expected. These results have implications that extend beyond the nitrogen-application decision for winter wheat: if this type of behavior is widespread, at current accuracy levels, other types of forecasts may be of little value to decision-makers and therefore go unheeded.