The goal of this paper is to systematically review the literature on United States Department of Agriculture (USDA) forecast evaluation and critically assess their methods and findings. The fundamental characteristics of optimal forecasts are bias, accuracy and efficiency as well as encompassing and informativeness. This review revealed that the findings of these studies can be very different based on the forecasts examined, commodity, sample period, and methodology. Some forecasts performed very well, while others were not very reliable, resulting in forecast specific optimality record. We discuss methodological and empirical contributions of these studies as well as their shortcomings and potential opportunities for future work.

This study examines the accuracy of United States Department of Agriculture (USDA) crop Acreage and Production forecasts for corn, soybeans, and winter wheat relative to their private counterparts over 1970–2019. Our main findings suggest that USDA forecasts often had significantly smaller errors than their private counterparts. The accuracy of both USDA and private forecasts has improved over time, but the accuracy of USDA forecasts has improved more than that of private forecasts, maintaining the USDA’s relative accuracy advantage. The accuracy advantage of Prospective Plantings and Acreage reports highlights the importance of survey-based approaches used for these forecasts.

This study evaluates the farmland price forecasts provided by the Federal Reserve Bank of Chicago's Land Values and Credit Conditions Survey from 1991: quarter 1 (Q1) through 2016: Q1. Prior studies have demonstrated that similar surveys of agricultural bankers provide accurate predictions of the direction of future farm real estate values through qualitative forecast evaluation. This study extends the existing knowledge base by converting the qualitative responses to quantitative expectations. The quantified expectations are then subjected to additional forecast optimality tests, which suggest that the forecasts are unbiased but inefficient.

The recent global financial crisis and the subsequent recession have revitalized the discussion on causal interactions between financial and economic sectors. In this study, I apply the financial stress and the national activity indices–respectively developed by Federal Reserve Banks of Kansas City and Chicago–to investigate the impact of financial uncertainty on an overall economic performance. I examine nonlinear dynamics in a vector smooth transition autoregressive framework, and illustrate regime-dependent asymmetries in the financial and economic indices using the generalized impulse-response functions. The results reveal more amplified dynamics during the stressed conditions. I further evaluate benefits of nonlinear modeling in an out-of-sample setting. The forecasting exercise brings out the important advantages that nonlinear modeling provides in the identification of the causal effect of financial instability on overall economic performance.

The Rice Outlook and Situation (RO&S) forecasts were compared to the forecasts of a univariate Box-Jenkins (BJ) model. On balance, the RO&S forecasts had lower mean square forecast errors and lower mean absolute forecast errors than the BJ model forecasts. The differences in the squared and absolute forecast errors were not significant, however. Based on the concept of conditional efficiency as set forth by Granger and Newbold, it was found that the BJ forecasts did not add any information that might improve forecast accuracy beyond what was already incorporated in the RO&S forecasts.

This research presents a systematic and unified approach to evaluating forecast rationality that considers the potential of nonstationarity in forecasts and realized values. The approach is applied to one-quarter ahead U.S. Department of Agriculture livestock price forecasts from 1982 through 2004. Results show that forecasts and realized prices are integrated of the same order, and those that are nonstationary are cointegrated. However, the stationary price forecasts for hogs, turkeys, eggs, and milk are biased and improperly scaled, and forecast errors tend to be repeated. Similarly, nonstationary forecasts for cattle and broilers are also biased and irrational in the long run, but short-run dynamics are rational.

Forecasts of economic time series are often evaluated according to their accuracy as measured by either quantitative precision or qualitative reliability. We argue that consumers purchase forecasts for the potential utility gains from utilizing them, not for their accuracy. Using Monte Carlo techniques to incorporate the temporal heteroskedasticity inherent in asset returns, the expected utility of a set of qualitative forecasts is simulated for corn and soybean futures prices. Monetary values for forecasts of various reliability levels are derived. The method goes beyond statistical forecast evaluation, allowing individuals to incorporate their own utility function and trading system into valuing a set of asset price forecasts.