This dissertation consists of three manuscripts. The goal of the first manuscript is to empirically examine how U.S. domestic wheat markets respond to wheat gluten imports from the EU, Australia, and other exporter nations. Two demand systems, factor demand and inverse demand, are specified and derived. A non-nested generalized likelihood test is applied for model selection to determine whether prices are adjusting to quantities or in reverse in U.S. wheat markets. The key results of this study suggest that gluten imports generally have significant influences on U.S. domestic wheat prices. However, specific influence depends on wheat class and the origin of the gluten imports.

The objective of the second manuscript is to assess the impacts of the new advertising strategy on the effectiveness of promotional efforts and to examine the significance and magnitude of the effect that promotional efforts have had on demand for D’Anjou pears in major U.S. retail markets. Nonparametric regression procedure is used to estimate domestic demand equations for D’Anjou pears across four regions. Results show that: (1) under the new advertising management system, promotional activities achieved significantly higher effectiveness than did the previous system; (2) shipment demand for D’Anjou pears was significantly impacted by own price, patterns of seasonal availability of other pears, patterns of habit formation, and pear imports to the U.S.; and (3) the Ad Buy and Demo promotional efforts, as expected, impacted the demand for D’Anjou pears significantly and positively, with noticeable differences across regions.

In the third manuscript, we apply a recently proposed Bayesian approach to infer the incidence of inefficiency in U.S. commercial banks from 1990 to 2000. To overcome misspecification problem of the estimated DEA efficiency scores, uniform ignorance Bayesian prior is used to infer an appropriate posterior distribution for the latent incidence of inefficiency. Results imply that the inferred latent incidence of banking inefficiency from the Bayesian method could be more accurate than the DEA method when the sample size used in the study is limited.