اندازه بسته و تبعیض قیمت در بازار حوله کاغذی

Estimates from a structural model of consumer behavior and firm conduct are used to decompose the extent to which quantity discounts for paper towels are consistent with second degree price discrimination as opposed to cost differences across sizes. Counterfactual exercises assuming that firms offer only one package size or charge uniform prices across sizes indicate that competition in the multi-roll package size segment results in increased consumer surplus and lower prices for all consumers.

Several consumer goods markets are characterized by brands that simultaneously offer different sized packages of the same product. These otherwise identical goods are often sold at different unit-prices — typically at volume discounts. While it is easy to find examples of products with prices that are nonlinear, attributing nonlinearities in observed prices to the presence of second-degree price discrimination is much more difficult. Invariably, cost-side explanations could also account for the nonlinearity in prices — for example, one may argue that larger package sizes can be produced at lower unit cost, or that inventory, transaction, or restocking costs are lower for larger sized packages. Much attention, therefore, has been given to econometrically identifying price discrimination when observed prices are nonlinear. The flavor of this work has been to determine whether observed nonlinearities in prices can be explained by variables that capture the ability of firms to price discriminate — e.g., the number of firms, the availability of competing products, and demographic patterns. Costs are controlled for either by including cost-related variables in the regression or by providing an explanation of why costs (in theory or in the data) cannot explain the observed price nonlinearities. I examine the effects of multiple package sizes in the paper towel industry by estimating a structural model of consumer demand and firm pricing behavior using readily available aggregate level data. I am able to use the model to decompose the unit-price of each product into a markup term (that reflects demand and competitive conditions) and a cost term. The extent to which price discrimination is driving the differences in unit-prices of the same underlying good is captured by differences in the unit-markups of small and large sized products of the same brand. These estimates are then used to determine the proportion of the observed within-brand unit-price variation across different package sizes that is cost-driven, and how much is consistent with price discrimination. I find that between 35 and 45% of the observed unit-price variation is consistent with price discrimination. Estimating a structural model has the additional advantage of permitting the researcher to predict market performance under counterfactual behavioral assumptions. I explore two such assumptions: (1) each brand may only offer its small size; and, (2) each brand must charge a uniform unit-price for all its sizes. In the counterfactual in which only the small size is offered, the model predicts modest price increases and a decrease in consumer surplus on average of about $100,000 to $140,000 per million consumers per quarter (due mainly to consumers forced to substitute away from the large size). In the uniform pricing counterfactual, the model predicts modest price decreases for small packages and a significant price increase for large packages with a negligible decrease in consumer surplus. In Section 2, I discuss the data used in this paper and provide some background on the paper towel industry. In Section 3, I describe the role of package size in consumer products and discuss how offering multiple package sizes enables firms to sort between different types of consumers. Section 4 reviews methods that have been used to analyze price discrimination in settings with multiple firms. Section 5 presents the model of consumer behavior, and Section 6 presents the model of firm pricing behavior used to identify costs. Section 7 describes the estimation procedure, Section 8 provides results, and Section 9 concludes.

This paper is an attempt to use recent advances in the structural estimation of differentiated product models of demand and market equilibrium to analyze second degree price discrimination in an oligopoly market. Using readily available aggregate level data, I have developed one measure of the extent to which the large observed unit-price differentials across different sizes of the same brand are the result of price discrimination. The data show significant discounts in unit-prices for multi-rolls relative to single rolls. My measure indicates that between 34 and 46% of these discounts are consistent with price discrimination, with the rest being cost-driven. In addition, I have used the structural model to assess the affects of multiple package sizes sold at different unit-prices on market performance. These counterfactual exercises suggest that competition among multi-roll packages is an important feature of this market. Consumers who purchase multi-roll packages benefit from more options and lower prices while consumers who purchase one-roll packages also enjoy lower prices as a result of competition from the multi-roll segment.