A Model of Retail Formats Based on Consumers' Economizing on Shopping Time

Paul R. Messinger and Chakravarthi Narasimhan

Retailers in developed countries increasingly provide one-stop shopping. Why? Are they responding to growing demand for time-saving convenience? Or are they responding to economies of scale made possible by new retail technology? And what role, if any, is played by improvements in automotive, refrigeration, and other consumer technology?

To examine these questions, we develop a model that can help explain the growth of one-stop shopping, estimate the model with aggregate grocery retail data, and compare the implications of our model with competing explanations.

Our model includes several desirable features:

(1) The extent of one-stop shopping is endogenously determined.
(2) Consumer store choice is determined in an explicit household-production, utility-maximizing context. This choice is based on a tradeoff between timesaving shopping convenience and price.
(3) The model is readily amenable to empirical testing.

The key endogenous parameter of our model describes the number of categories carried by a store. In the context of grocery retailing, this parameter can distinguish between a single-category specialty store, a simple grocer, a supermarket, and a large superstore. This parameter is determined by the interplay of consumers economizing on shopping time and retail competition in the presence of technological constraints which relate operating cost with assortment size.

Our approach may also be viewed as an example of how the boundary between retail service production and household service production can be modeled as a market outcome (in the spirit of Betancourt and Gautchi 1990 and Wernerfelt 1944). The nature of extant retail formats and other channel-related institutions circumscribes the boundary between retailers and households. With the notable early exception of Baumol and Ide (1956), few explicit models describe retail format determination as an equilibrating market mechanism. Our model describes the determination of one aspect of retail formats (namely, the extent of one-stop shopping). Our hope is that similar approaches can be used to describe other aspects and types of emergent retail formats.

We estimate our model using U.S. aggregate annual data for a 26-year period, and later corroborate parts of the analysis using cross-sectional data for the 48 contiguous states of the US The equilibrium conditions of our model translate into a system of two equations with two endogenous variables. Estimates of the reduced from parameters provide our central empirical finding--- that per capita disposable income has had a significant positive effect on both supermarket assortment (as expected from our model) and store operating costs (perhaps somewhat less expected). This suggests that greater prevalence of one-stop shopping has been a response to growing demand for timesaving convenience. The estimates of the underlying structural parameters also provide a measure of the imputed net savings to consumers from supermarket shopping, taking into account reductions in shopping time made possible by supermarkets (savings equivalent to approximately 2.2% of expenditures on grocery products over the 1961-1986 period).

We lastly consider competing hypotheses. While a number of explanations can account for increased store assortment, these explanations have divergent implications for retail margins, operating costs, and profits. These implications suggest that retail scale economies were not the drivers of one-stop shopping in the years of our data. Nor was the impetus to increase store size primarily a desire to bring in new higher-margin items. However, we acknowledge that transportation and inventory-holding technologies are prerequisites to one-stop shopping and that improvements in these technologies were particularly important in the early growth period of supermarkets.

(Store Formats; Retailing; Channels of Distribution)