Alternative Forms of Price Expectations in Supply Analysis for U.S

The consequences of frequently used price expectation models are analyzed by comparing the responsiveness of U.S. corn and soybean acreages to six alternative formulations. The trade-off between bias and variance associated with these forecasts is investigated. The results of this study have important implications for future research on supply analysis. Key words: supply response, price expectations, government programs. Identifying the appropriate formation of price expectations is a major challenge in supply re-sponse analysis for two reasons. First, uncer-tainty about how farmers formulate their price expectations has led researchers to propose dif-ferent assumptions and hypotheses regarding the extent to which farmers use available out-look information in their planting decisions. As a result, different schemes have been used as proxies for price expectations in previous studies on supply analysis. The most common ones are adaptive expectations (Nerlove), na-ive expectations (Houck and Gallagher; Shum-way and Chang), futures prices (Gardner 1976; Chavas, Pope, and Kao), effective support prices (Houck and Ryan; Houck et al.; Ryan and Abel), rational expectations formulations (Shonkwiler and Maddala), and other mixed formulas of market and support prices (Gal-lagher; Lidman and Bawden). Muth proposes an expectations model based on "the rational expectations hypothesis " which provides a consistent way of incorporating expectations into economic models. However, the rational expectations hypothesis generally is not used in supply response analysis because of the dif-ficulties encountered in empirical applications (Nerlove, Grether, and Carvalho, p. 301). Second, price expectations as proxies for "true " unobservable prices are subject to the The authors are a postdoctoral research associate and a professor