Loss leaders and cross-category retailer pass-through: A Bayesian multilevel analysis

Retail pass-through has been extensively analyzed analytically and empirically, and recent empirical work has stressed the importance of appropriate methodology and data for inferring correct retail pass-through. However the literature on retail pass-through has interpreted ‘pass-through’ as being confined to a specific product category, and only to brands within that category. This category restriction has been derived from a tradition of modeling retailers as ‘category profit maximizers’. Yet it is widely accepted that retailers strive to maximize profits across categories, with several categories specifically functioning as ‘loss leaders’. In this paper we argue that this pragmatic view of retailers makes it necessary to reevaluate retailer pass-through from being merely a ‘within category’ phenomenon to also a ‘cross category phenomenon’. Using a unique dataset we construct category price indices and empirically evaluate cross category pass-throughs with a variety of categories – selected on the basis of profitability. We find that by and large cross category pass-throughs tend to be negative (about eighteen percent as compared to nine percent positive), that is, price cuts in a focal category being accompanied by price increases in other categories. Category characteristics such as price elasticity and proportion of loss leaders increase the probability of negative cross-category pass-throughs. We conclude that future work on retailer pass-throughs needs to incorporate cross category analysis in order to capture the ‘true’ strategic behavior of the retailer.