Determinants of retail price margins and quality adjustment of price indices in retail food stores

Properly measuring a change in retail prices over time is important in constructing an accurate price index for the retail sector. Since a retailer implicitly bundles the physical good it sells with some distribution services, it is necessary to measure and control for the services provided by the retailer. Using five categories of retail distribution services (location accessibility, product assortment (breadth and depth), assurance of product delivery (time and form), information (prices and characteristics of the good), and ambiance) outlined in Betancourt and Gautschi (1988), the first half of my research investigates the determinants of retail price margins using a unique data set of retail food store price margins and store characteristics for a nationally representative sample of food stores. I find that all five services, when approximated by specific store characteristics, play a significant role in determining variation in price margins across stores with the magnitude of impact dependent upon the format of the store. The price margins also vary significantly across departments within each store and different distribution services play larger or smaller roles determining price margins depending on the specific department one looks at.; Based on these findings, the second half of my research presents a method for accounting for differences in distribution services across stores in calculating a producer price index. The marginal value of each of the distribution services is used to explicitly adjust the observed price changes over time. By implementing a quality adjustment methodology that controls for changes in observed characteristics of a store over time, I am able to compare a quality adjusted index to the unadjusted index and find there to be an upward bias (1% to 2%) in the producer price index for food stores. This bias is due to the fact that an observed price change is often associated with a change in distribution services provided by the retailer—a portion of the price change is due to this quality change to the good. The magnitude of the bias is sensitive to the specific type of change that occurs as well as the number of products that experience a characteristics change in a given month. This bias may exist in other retail industries and a similar adjustment to observed price changes would provide a more accurate measure of pure price change in the retail sector.