Essays on Pricing and Consumer Demand in the Retail Sector

Abstract

This dissertation consists of two independent chapters on pricing and consumer demand in the retail sector. In chapter 1 develop an empirical model of Consumer Supermarket Choice that enables identification of heterogeneous consumer travel costs and is suitable for a wide range of policy experiments and the study of local competition. Chapter 2 is a theoretical investigation on pricing patterns in multi-product retail markets, when boundedly rational consumers' choice of a store is based on the price and valuation of a subset of goods. Estimation of demand systems in spatially differentiated retail markets is fundamental for understanding local competition and the impact of policy changes. It is also challenging, because shopping decisions consist of multiple dimensions: when to shop, where to shop and what to buy. In chapter 1 I develop an empirically tractable model of store choice in the supermarket industry and provide a way to identify consumers' heterogeneous travel costs without imposing restrictions on bundle choice. Using micro level data on a small market in New England, I estimate demand for stores using both a moment inequality approach and standard discrete choice techniques. I specify utility as a function of both store and bundle characteristics, and control for the endogeneity of expenditure on the bundle. I use the estimates of the discrete choice model to evaluate the welfare impact of 1) the closing of each individual store in the market and 2) the relocation of one of the stores. I find that travel costs are heterogeneous and marginally decreasing; that people like to shop at stores that are close, but also like to shop at multiple stores. Furthermore, people value stores differently (across consumers and shopping occasion) and trade off additional travel time for better store characteristics; utility differentials in preference for stores correspond to a distance ranging between zero and up to 3.3 miles. Variation in demand and substitution patterns across stores are explained by differences in store characteristics and by the shopping habits and geographic distribution of heterogenous consumers. Changes in market structure, like store entry and exit can have significant impact on consumer welfare. For example, removal on one of the stores results in a loss in CS that ranges between 8% and 44%. The assumption of rationality in retail shopping decisions appears very problematic when stores sell thousands of products and frequently vary their assortments and prices. Consumers are typically uncertain about prices at different stores and for a consumer to consider the entire distribution of bundles and prices might be a far too complex decision process. Furthermore, models with rational consumers are incapable of fully explaining important features of retail markets such as price dispersion, advertising and leader pric- ing. In chapter 2 I attempt to characterize optimal pricing by multi-product retailers when imperfectly informed consumers buy more than one product. The distinctive feature of the model is that there are two relevant moments to all purchase decisions. First, the choice of a store to visit, and second, the choice of the items to purchase. While consumers might rationally choose a store to best meet their specific needs and desires, the choice of the items to purchase is made only once in a store. Whether guided by impulse, contingent and unforeseen needs or in-store learning about a product, consumers often end up buying additional products which can generate higher profits for the stores. To examine the implications on retail pricing of this kind of behavior, I depart from a standard rational setup and introduce the concept of attractor goods. Using an an approach similar to that found in Osborne and Rubinstein (1998) and Spiegler (2006) I consider boundedly rational con- sumers whose choice between stores is based solely and entirely on the price and valuation of a subset of goods, the attractors. I show that retailer's pricing decisions have to take into account not only the direct effect of prices on a product's demand but also the effect on the demand for the other products sold in the store. The optimal pricing schedule will be a decreasing function of the goods' attractiveness, and pricing below marginal cost might be optimal for some goods. The model provides a rationale for the strategy of loss leader pricing and offers an intuitive explanation to countercyclical markups.