The Supply of Health in Food: The Economics of Innovation Behaviour, Brand Competition and Changing Consumer Preferences
Final Report Abstract
Recent controversial policy proposals have aimed at creating a healthier food supply by means of taxation, minimum quality standards or nutritional labeling. Yet the outcomes of such policies strongly depend on the competitive structures and thus substitution processes of individual products within categories, which are not well understood. The objective of this project was to quantify the source and impact of differentiation in ingredient formulation and especially product health attributes on the competitive positioning of brands under heterogeneous consumer preferences. I employed Berry, Levinsohn and Pakes’ (1995) random-coefficient logit framework to estimate product-level demand for highly differentiated Potato and Tortilla chips in the U.S. retail market. I was specifically interested in the extent to which heterogeneous consumers respond to changes in product formulation, pricing and brand attributes. My results supported the unhealthy-tasty intuition hypothesis with consumers’ utility increasing in sodium and saturated fat levels but decreasing in energy and total fat content. Results further suggested strong impacts of price, brand, and flavor effects on band-level market shares. The analysis underlines the trade-offs involved in food manufacturers’ decisions to reformulate products in order to comply with policy and public demands for healthier products options that do not sacrifice taste. Future work should be directed to a deeper and probably interdisciplinary study of the nutritional and sensory attributes of consumer products through the integration of formal econometric modelling and complementary experimental and/or survey approaches. Ideally, such information would also allow the econometrician to observe changes in product formulation over time, which could play a critical role in the context of the objective in this paper. Hence, data on consumer product choices that include attitudes and actual eating behavior, including stated taste preferences, individual health attitudes or health status, would make for better determinants of product choice compared to the basic U.S. Census variables available to us in this study. Alternatively additional information could be obtained via nutritional panel surveys such as U.S. NHANES or household panel data from major providers of market research (e.g. Nielsen, GfK).
