Academic journal article International Advances in Economic Research

A Sequential Rationality Test of USDA Preliminary Price Estimates for Selected Program Crops: Rice, Soybeans, and Wheat

Academic journal article International Advances in Economic Research

A Sequential Rationality Test of USDA Preliminary Price Estimates for Selected Program Crops: Rice, Soybeans, and Wheat

Article excerpt

Abstract Despite recurrent evaluations on USDA price forecasts, the performance of USDA price estimates has not previously been examined in publication. To fill the void in research to this important public information, a sequential forecast evaluation procedure is applied to selected USDA price estimates: Rice, soybeans, and wheat. The evaluation procedure reveals that the USDA price estimates are short-run unbiased; however, they are not long-run rational. In addition, short-run optimality and efficiency tests suggest that USDA price estimates need to be properly scaled and fully reflect information embodied in past prices and their estimates--a possible venue to improve the predictability of USDA price estimates for the crops.

Keywords A sequential rationality test * USDA price estimates * Time-series model

JEL D40 * Q11 * C32 * C53 * C10 * Q00


Commodity price forecasts generated by the USDA are intended to provide important public information to both policy makers in Government and market participants in making informed production, marketing, processing, and retail decisions. Price forecasts are useful to policy makers and decision markers only if the price forecasts help them make better decisions. As indicated by Stein (1981), accurate commodity price forecasts derive rational decision making, assure efficient allocation of resources, and thus maximize social welfare.

Numerous researchers have scrutinized USDA price forecasts in terms of absolute accuracy (Elam and Holder 1985; Kastens et al. 1998), bias and efficiency (Sanders and Manfredo 1998), or directional accuracy (No 2007). Although the researchers above have used different forecast evaluation methodologies, they have implemented a simple strategy that N step ahead of the USDA forecast, E([P.sub.[t+n]]|[[OMEGA].sub.t]) is compared with their own forecast, E([P.sub.[t+n]]|[X.sub.t]), where [[OMEGA].sub.t] is a quantitative and qualitative information set available at the USDA and at time t (Vogel and Bange 1999, p.10) and [X.sub.t] is a vector of price related time series as well as actual price available for researchers at time t. Based on various accuracy criteria, their reports on the performance of USDA price forecasts are diverse.

For instance, Elam and Holder (1985) found that USDA rice forecasts have lower mean square forecast errors than random walk model forecasts. Sanders and Manfredo (1998) found that USDA forecasts are statistically more accurate than competing times series forecasts for fluid milk. However, not all investigators were in favor of USDA forecasts; Kastens et al. (1998) reported that extension forecasts are more accurate than USDA forecasts for livestock. No (2007) found that USDA hog price forecasts have lower accurate forecast ratio and higher worst forecast ratio than the forecasts of time-series model, suggesting weaker directional accuracy in the USDA model.

Recently, Sanders and Manfredo (2007) reported that USDA price forecasts for hogs, turkeys, eggs and milk are biased and improperly scaled and that forecast errors tend to be repeated. It is certain that despite numerous and recurrent evaluations on the performance of USDA forecasts, researchers will continue to micronize one-period or multiple-periods ahead of USDA forecasts not because of their publication merits but because of the fact that rational forecasts reduce price uncertainty prevailing in the markets.

Besides its price forecasts, the USDA releases preliminary estimates for various crops in its monthly outlook reports, issued at midmonth, to assist market participants for informed decision making. For instance, the USDA's monthly Rice Outlook on July 13, 2007, released an estimated rice price for the previous month. The estimate was later revised after complete quantitative and qualitative information available, such as demand and supply of rice, seasonal component, and expert opinions. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.