A comparative study of household demand for meats by U.S. Hispanics

The general objective of this research project was to study the consumption behavior of U.S. Hispanics for meat products. A comprehensive review of literature on the studies of U.S. Hispanic consumers was performed by means of a systematic survey; different sources of information useful in studying Hispanic consumers were documented. The simultaneous increase in population and income of Hispanics in the United States is attractive to food companies, and makes more relevant the study of their consumer behavior patterns for discovering future market opportunities. Censored incomplete demand systems of the LinQuad form were used for recognizing the consumption patterns by Hispanics with those of Whites, African Americans, and households of other minorities. The analyzed dataset was extracted from the Consumer Expenditures Survey released by the U.S. Bureau of Labor Statistics. The estimation of elasticities required two steps; in the first step, the decision to purchase was modeled and in the second step demand equations using instrumental variables were estimated to eliminate selectivity bias. Both steps used a concatenated forward stepwise modeling approach. It was found that Full Information Maximum Likelihood (FIML) under a high level of censoring produces inconsistent own price elasticities and bigger standard errors, due to violations of multivariate normality of the error terms and high level of censoring. The discussed results come from Iterative Seemingly Unrelated Regression since it was able to produce a greater number of significant parameters and more consistent elasticities compared to FIML. The role of ethnicity in the consumption of meat products in the U.S. marketplace was evaluated by means of a demand system that had dummy variables for comparing patterns of Hispanics with the rest of the ethnic groups. Hispanics on average allocated more for total food expenditures, consumed more at home, and spent 21.5%, 8.1%, 5.4% more on meat products than Whites, African Americans, and households of other minorities, respectively. Three sets of demand systems are presented. The first set includes elasticities from demand systems that only included prices and income; the second set is augmented with household size in Amsterdam scale; the third set uses the complete proposed set of demographics