Testing for Preference Change in Consumer Demand: An Indirectly Separable, Semiparametric Model

This article considers the hypothesis that preference changes may partly explain observed consumption patterns. A model is developed based on indirect weak separability. The kernel estimator of the regression function models the unknown shape of demand functions, and a parametric structure models seasonality, dynamics, and preference change. This semiparametric specification is applied to U.S. meat demand. The results support the notion that changes in consumer preferences, due to consumers' awareness of the health hazards of cholesterol and saturated-fat intake, may explain an increased consumption of white meat and a decreased consumption of red meat.