Packaging as Commodity Bundling and Inter-type Competition in Retailing by

* Earlier versions of this paper were presented at The Economics of Retail Systems seminar at the University of Maryland. I thank the participants for their comments while claiming responsibility for any errors and omissions. 1 This work also provides references to the earlier literature. 2 Introduction There is a substantial literature on commodity bundling. A seminal point of departure for the modern literature is the work of Adams and Yellen (1976). 1 It eliminates what one may consider two fundamental explanations for bundling: preferences (complementarity of demand) and technology (economies of scope). Instead, it focuses on commodity bundling as a mechanism for price discrimination between heterogeneous consumers. It defines three possible strategies for a monopolist selling two package sizes of the same items. Offering just the large size, which is called pure commodity bundling; offering both sizes with the large one priced at a discount, which is called mixed commodity bundling; and, implicitly, offering the small size which would be simple monopoly pricing or the pure components strategy in their terminology. Subsequent work proceeds by relaxing the assumptions made by Adams and Yellen (1976) but, by and large, preserves the emphasis on commodity bundling as a form of price discrimination. Lewbel (1985), for example, extends their results by showing that there can be a profit incentive to bundle in the presence of substitutes and a profit incentive not to bundle in the presence of complements. McAffee, McMillan and Whinston (1989) characterize mixed commodity bundling as an optimal strategy in very general terms, by deriving the results in terms of joint probability distributions for the reservation values of consumers rather than in terms of discrete consumers and by allowing for premia as well as discounts, and they extend the result to a duopoly. More recently, the literature stressing the price discrimination aspect of commodity bundling 3 has moved to the incorporation of strategic interaction features, for example Carbajo, de Meza and Seidman (1990) and Matutes and Regibeau (1992). Of special relevance to our present purposes are three other contributions. Two of them stress the so called fundamental reasons for commodity bundling. Chae (1992) assumes an extreme form of economies of scope in the distribution technology and demonstrates that pure commodity bundling can increase profits and consumer surplus relative to a pure components strategy. This analysis and its application to the subscription television market is described by the author as a …