Measuring the impact of retailers’ strategies on food price inflation using scanner data

We implement a three-way panel data model to test the effect of retailers’ marketing mix on food price inflation. Using high frequency scanner data for different dairy products in Italy we compute a weekly drift-free price index, specific for product category, chain and type of store. Exploiting the panel data structure to control for unobservable marketing mix by chain, type of store and time, we test if unobservables are statistically significant in influencing the food inflation rate on each of the products covered by our analysis. In general chain and type-of-store specific unobservables have a significant role in controlling the rise of prices. Moreover, we identify the role of some observable marketing mix on controlling or on facilitating price inflation rates. Results show that while higher PL shares help on limiting an upward inflation rate, reversely a higher PL line extension tends to accelerate it. Sales, as expected, alleviate the burden of a general increase in prices. However, PL sales have an effect on reducing the price inflation rate which is proportionally smaller than the overall average, indicating that proportionally sales on PL contribute less intensively to reduce a generalized upward price trend. Finally, assortment has a mixed effect depending on the competition environment of the market we refer to.