Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)

First Advisor

Harlon D. Traylor


This dissertation analyzes the effects of quality factors on prices paid producers for long and medium grain rough rice in Louisiana. Rough rice prices, and other information surrounding quality, were collected for the study for the 1986/87 and 1987/88 marketing years from the Louisiana Farm Bureau Marketing Association in Crowley, Louisiana. The relationship between the price of rough rice, and its quality attributes or characteristics, was analyzed in a hedonic price framework. A conceptual model for the Louisiana rough rice market was constructed, and estimated premiums and discounts reported for a set of quality factors believed to influence producer prices. Premiums and discounts were calculated for long and medium grain markets for the 1986/87 and 1987/88 marketing years and for marketing seasons within marketing years. The hedonic model was tested for structural differences across marketing years, marketing seasons, and classes of rough rice. Structural differences were found in all cases. A linear specification was chosen for the base model. However, the Box-Cox transformation indicated a semi-logarithmic specification for the 1987/88 marketing year. In the set of quality factors studied head rice, red rice, and heat damage were the most important monetarily. The monetary value of the quality factors were calculated for an average producer and compared to the cost of controlling quality where applicable. Significant error problems were identified and rigorously analyzed in the hedonic models. In addition to least-squares point estimates, the hedonic model was estimated using an error-in-variables model (EVM). From this model a set of consistent hedonic prices were calculated. Prior information was used to re-specify and estimate the original hedonic rough rice model. The regressions were adjusted to account for the measurement error and statistics were calculated to measure the degree of error. The EVM model provided a likely range (upper and lower bound) for the premiums and discounts, and thus directly assessed the uncertainty about premiums and discounts relative to least squares estimates. The monetary value of premiums and discounts was reassessed based on these bounds.