We apply the BLP random coefficient logit model demand model to fluid milk sales data from two north-south Italian cities: Turin and Naples. By virtue of their location and socioeconomic differences, these cities provide a natural experiment for contrasting consumer choices and retail market power related to milk physical and marketing characteristics. Results reveal that, regardless of location, consumers negatively value price increases, fat content and ultra-high temperature (UHT) treatment. However, location matters with respect to brand and type of milk purchased. While in Turin (the higher-income region) demand for the leading manufacturers’ brands is the most price inelastic, in Naples consumers have the lowest price elasticities in case of cheaper milk, often small manufacturer or private label brands. Unlike previous studies, we do not find price elasticities for private labels to be consistently lower (or markups to be higher) compared to manufacturer brands, indicating that private labels have reached maturity in these markets. Further, while demand for fresh milk is more price inelastic in Turin, it is more inelastic for UHT milk in Naples. Likewise, markups and Lerner indexes are higher for fresh milk in Turin and for UHT in Naples corresponding to the more inelastic demands under Bertrand price competition.     «

We apply the BLP random coefficient logit model demand model to fluid milk sales data from two north-south Italian cities: Turin and Naples. By virtue of their location and socioeconomic differences, these cities provide a natural experiment for contrasting consumer choices and retail market power related to milk physical and marketing characteristics. Results reveal that, regardless of location, consumers negatively value price increases, fat content and ultra-high temperature (UHT) treatment....     »