Estimating a continuous hedonic-choice model with an application to demand for soft drinks

Using micro-level scanner data, I study empirically the consumer demand for soft drinks, which is characterized by multiple-product, multiple-unit purchasing behavior. I develop a continuous hedonic-choice model to investigate how consumers choose the best basket of products to satisfy various needs. My model's embedded-characteristics approach both helps to reduce the dimensionality problem in model estimation and generates flexible substitution patterns. Hence, the model is useful in application to data with many product choices that are correlated with each other at the individual level. The estimation results show that interesting substitutability and even a form of complementarity exist among soft drinks.