Economic Losses from Forecasting Error in Agriculture
提出一种估算预测误差经济损失的新方法，替代传统分布滞后模型。以1874-1933年美国堪萨斯州农业为例，发现农民价格预测能力存在时空差异，且与市场改进、技术地理因素及人力资本相关。
This paper develops a new method for estimating economic losses from forecasting error. The method is an alternative to the traditional distributed lag approach to estimating expectations. An application to Kansas agriculture over the period 1874-1933 reveals significant variation in farmers' price forecasting performance across counties and over time. These differences can be explained in terms of market improvements, technological and geographical factors, and county characteristics suggestive of human capital differences. The magnitudes, timing, and cross-sectional pattern of the losses from forecasting error confirm the importance of forecasting skill and allocative ability in American agriculture during the period examined.