Estimating U.S. Household Seafood Demand Based on Longitudinal and Cross-Sectional Data

• Main Author: Wang, Huabo
• Other Authors: Keithly, Walter R
• Format: Others
• Language: en
• Published: LSU 2014
• Subjects: Agricultural Economics & Agribusiness
• Online Access: http://etd.lsu.edu/docs/available/etd-06112014-154108/

This overall goal of this research is to examine and update U.S. at-home household demand estimates for seafood (including species and generic products) that can be used in policy formulation (by both the private and public sector). As such, this study estimated quantity-based household demand functions for seafood in aggregate, by generic product form (fresh, frozen, and prepared), and primary species. Given that the role of quality has been shown to significantly influence expenditures and seafood demand, this paper also seeks to incorporate the household quality choice into the demand model. In order to help to tailor the market strategy, this study also estimates household seafood demand in a complete demand system framework to track the substitution and complementation between seafood product forms and other protein sources. Emphasis is also given to the influence of socioeconomic factors on the demand for quantity. This study uses 2005-2006 NOAA Fisheries Seafood Consumption Survey data which consisted of 10798 completed interviews. In the complete demand system analyses, comparison of model results with a quality proxy, without a quality proxy, and with a quality adjusted price were examined to determine a preferable means for incorporating household quality choice. Results suggest that quality does play an important role in seafood demand. As for the quantity-based demand equations, a bivariate model was applied to simultaneously investigate the quality variation and consumer preference. As an outcome of this model, and based on the hypothesis that demand for quality is proportional to the level of aggregation, the study examines whether the demand for quality diminishes in relation to the level of disaggregation. The bivariate model utilized the maximum likelihood method to successfully deal with a truncation problem as well as difficulties of unobserved unit price values.