How Precise and Efficient are USDA Price Forecasts?
How Precise and Efficient are USDA Price Forecasts?
Evaluation of USDA price forecasts (table 2) focused both on their performance and their ability to accurately reflect uncertainty associated with the forecast. Differently from all other USDA forecasts, price forecasts have historically been published as a range. WASDE provides a judgmental price projection from a balance sheet approach that includes ranges reflecting uncertainty associated with prices in the future (Vogel and Bange, 1999). These price ranges were constructed symmetrically but the distribution of forecast prices is asymmetric and does not represent a constant confidence interval (Isengildina-Massa, and Sharp, 2012). Researchers proposed appropriate methods for evaluating accuracy of these forecasts as intervals (Isengildina, Irwin and Good, 2004), as well as alternative methods for generating more accurate prediction intervals (Isengildina-Massa, Irwin and Good, 2010; Isengildina-Massa et al., 2011; Adjemian, Bruno and Robe, 2020). The biggest issue with these forecasts was the lack of confidence level information to accompany price ranges, rendering them difficult to interpret. However, rather than adopting some of these recommendations and publishing more informative intervals, USDA decided to switch to publishing point price forecasts in May 2019. Given the confusion about what price ranges represented, these ranges have usually been reduced to their midpoints in interpretation and analysis. For example, Hoffman et al. (2015) demonstrated that WASDE corn price forecasts were more accurate than several futures-based benchmarks in 9 out of 16 forecasting periods over 1980-2012. However, combining WASDE and futures-based forecasts reduced forecasts errors by 12-16 percent on average. This study further demonstrated that favorable average trading profits could be generated in some months using WASDE projections, suggesting that WASDE projections of the U.S. corn season-average price provide useful information to the market and could enhance the efficiency of the agricultural sector.