In many competitive settings, some consumers buy multiple types of goods from a single location to save on shopping costs, while others use different sellers for different goods. We call these one-stop and multi-stop shoppers, respectively.
One-stop shopping can intensify competition by turning co-located goods into pricing complements. This is because when you cut the price of one product and gain a one-stop shopper, you gain the shopper’s demand for all goods. But clearly the effect is less strong when there are multi-stop shoppers – such shoppers need not transfer their demand for all goods.
If the seller internalises this complementary effect, she has a greater incentive to cut prices, with a potentially dramatic reduction in profit margins (Nalebuff 2000).
Whether these cross-effects are internalised by price setters depends on how the market is organised (Beggs 1992, Smith and Hay 2005). They are internalised in a supermarket, but not in traditional alternatives – for example, streets, malls, and public market places – where products are sold by separate single-category sellers (e.g. butchers, liquor stores, and bakers).
We consider two questions. First, how much does internalisation reduce market power – that is, how much more competitive is supermarket organisation compared to streets, malls, and market places? Second, do one-stop shoppers or multi-stop shoppers have the greater pro-competitive impact?
These questions are relevant for understanding competition in the grocery retail market – a market that receives much antitrust attention (see, for example, the UK’s Competition Commission 2000, 2008, and the US Federal Trade Commission’s analysis of the Whole Foods/Wild Oats merger).1,2
We address these questions in light of results obtained in Smith and Thomassen (2012) and Thomassen et al. (2017), using a model of shopping choice estimated with household shopping data.3
How much does internalisation by supermarkets reduce market power?
We first look at the data to see how grocery consumers shop. Table 1 shows that consumers tend to buy many goods per trip (about 30) but visit only a few stores (between one and two per week). This pattern – many products, few stores – is consistent with the presence of shopping costs. But shopping costs are not so high that all households are one-stop shoppers.
Table 1 Many goods, few stores: An indication of shopping costs
Notes: Definition of goods include, for example, shampoo, breakfast cereal, yoghurt. SD denotes standard deviation.
Source: Smith and Thomassen (2012).
To look at the incentives of supermarkets relative to specialist sellers, the model aggregates the products in supermarkets into eight demand categories that correspond to the items sold in the specialist vendors of a street or market place: bread products (as sold in bakery stores), drink products (liquor stores), meat products (butchers), fruit and vegetables (greengrocers), etc. Each consumer decides whether to buy the categories from a single supermarket or from more than one – that is, he can choose between one-stop and multi-stop shopping depending on his shopping costs.
Table 2 (Intra-firm) cross-category elasticities for ASDA and Tesco
Note: Elasticity of column demand with respect to row price. All elasticities are intra-firm. Not all categories shown.
Source: Thomassen et al. (2017).
Table 2 reports the cross-elasticities between these categories for the two largest supermarket firms in the UK – ASDA and Tesco. Note that all the within-firm cross-category elasticities are negative. Thus, when one of these firms increases the price for goods in one category, it loses demand for the other categories it sells. Thus, categories at each supermarket are pricing complements; this is not intrinsic to the products, but a consequence of shopping costs.
How much does internalisation of these cross-category effects affect market power? We assess this question using the concept of a cross-category externality. This is the profit gain to the supermarket on all other categories when it cuts the price of one category by enough to increase its demand by an extra unit. A supermarket internalises this externality, but a single-category seller does not. We express the externality as a fraction of the product’s price. As Table 3 shows, it is on average 47% of prices.
Table 3 also includes marginal costs and marginal revenues for the category. A supermarket seller sets a category’s marginal revenue, plus its marginal externality, equal to its marginal cost. We can see that the marginal externality is large in comparison to the marginal costs and revenues.
Table 3 Marginal revenues, marginal costs, and (intra-firm) cross-category externalities
Note: Marginal externality plus category-specific marginal revenue equal marginal cost (by profit maximisation).
Source: Thomassen et al. (2017).
How do we interpret the externality and its magnitude? Suppose the supermarket owner were to delegate pricing of each category to a category manager who considers only his category’s profit. Then the externality is the (Pigouvian) marginal subsidy needed to ensure that prices do not increase. Our estimates indicate that this marginal subsidy is around half (47%) of supermarket prices. This is a substantial pro-competitive effect of supermarket organisation.4 Its size is consistent with theory (see numerical examples in Nalebuff 2000).
Who constrains prices the most: One-stop shoppers or multi-stop shoppers?
We now turn to our second empirical question: which type of shopper has the greater pro-competitive impact. Theory does not give an unambiguous guide – the two-stop shopper can switch demand between stores with relative ease when a price changes, but the one-stop shopper generates large complementary price effects by switching demand for all goods.
To analyse this empirical question, we divide each firm’s shoppers into two groups: those that lean towards one-stop shopping, and those towards multi-stop shopping. (Which group a shopper belongs to is determined by their exogenous preferences.) Starting at optimal prices, we compute the marginal effect on profits of unilaterally increasing the price of each category at each firm.
The effect of any price change on a firm’s total profit (derived from all shoppers) is of course zero – the prices are optimal. But it would be a fluke if the effect were also zero on the profits derived from any sub-group of shoppers. The group for which profit falls constrains prices more than the group for which profit rises.
Table 4. One-stop and two-stop shoppers: Which group constrains prices the most?
Note: The results are based on marginal price changes at each category and firm.
Source: Thomassen et al. (2017).
Our estimates give a clear result (see Table 4). For the overwhelming number (89%) of marginal price hikes (across firms and categories), the profit from one-stop shoppers falls, and the profit from two-stop shoppers rises. The results are statistically significant – we reject the null hypotheses that the median effect (across firms and categories) is positive for one-stop types and that it is negative for two-stop types (both at the 2.5% significance level). Thus, the one-stop shoppers have the greater pro-competitive effect, because of their greater cross-category effects.
In many settings, buyers have shopping costs and concentrate their shopping in a few stores. For the UK supermarket industry, we find that this generates large positive externalities between products, which in turn means that supermarket organisation (which internalises these effects) greatly improves the competitiveness of grocery supply.
Adam Smith observed that “it is not from the benevolence of the butcher, the brewer, or the baker that we expect our dinner, but from their regard to their own self-interest.” But as we find, these single-category sellers (when located together) will set more socially efficient prices if they consider their joint interests. Put differently, specialty stores do not take into account the positive externality they impose on nearby stores by reducing prices, while supermarkets do, which increases social efficiency. Our results show that this gain is considerable.
This does not change the fact that having a few supermarket firms usually results in higher margins. But our conclusion indicates that if one wishes to go back to a traditional market organisation with only specialty stores, or if supermarkets were to delegate pricing decisions to category managers, one should expect an efficiency loss.
We also find that it is not the more active kind of shopper – the multi-stop shopper, shopping around at different places – that constrains supermarket prices the most. Rather it is the high shopping-cost consumer (who visits a single store) that has the greater pro-competitive impact on supermarket pricing. A market where consumers shop around is not necessarily a competitive one.
Authors’ note: We are grateful to DEFRA and the Milk Development Council for funding. The views expressed here are not necessarily the views of these organisations.
Armstrong, M and J Vickers (2010), “Competitive non-linear pricing and bundling”, Review of Economic Studies 77(1): 30–60.
Beggs, A (1994), “Mergers and malls”, Journal of Industrial Economics 42(4): 419–28.
Competition Commission (2000), Supermarkets: A report on the supply of groceries from multiple stores in the United Kingdom, London: The Stationery Office.
Competition Commission (2008), The supply of groceries in the UK, London: The Stationery Office.
Farrell J and C Shapiro (2010), “Antitrust evaluation of horizontal mergers: An economic alternative to market definition", The BE Journal of Theoretical Economics 10(1): 1-41.
Matutes, C and P Regibeau (1988), “Mix and match: Product compatibility without network externalities”, RAND Journal of Economics 19: 221–234.
Nalebuff, B (2000), “Competing against bundles", in P Hammond and G Myles (eds) Incentives, Organization, and Public Economics, Oxford: Oxford University Press.
Smith, H (2004), “Supermarket choice and supermarket competition in market equilibrium”, Review of Economic Studies 71(1): 235.
Smith, H and D Hay (2005), “Streets, malls, and supermarkets”, Journal of Economics & Management Strategy 14(1): 29–59.
Smith, H and Ø Thomassen (2012), “Multi-category demand and supermarket pricing", International Journal of Industrial Organization 30(3): 309-314.
Thomassen, Ø (2017), "An empirical model of automobile engine variant pricing", International Journal of the Economics of Business.
Thomassen, Ø, H Smith, S Seiler and P Schiraldi (2017), "Multi-category competition and market power: A model of supermarket pricing", American Economic Review 107(8).
 FTC v. Whole Foods Markets, Inc., 533 F.3d 869 (DC Cir. 29 July 2008).
 The possibility that two-stop shoppers (also known as cross-shoppers) constrained prices more than one-stop (or core) shoppers played a central role in both examples.
 There is a large theoretical literature with multiple stores, multiple goods, and shopping costs (Matutes and Regibeau 1988, Armstrong and Vickers 2010, and references cited therein). The model we estimate is in this class. In the model, consumers select up to two stores and up to eight categories from these stores. Shopping costs are partly a result of transport costs and partly the inconvenience inherent in shopping in two supermarkets. In modelling a continuous choice (of goods) within a discrete choice (of stores) the model draws upon earlier work in Smith (2004) and Thomassen (2017).
 See Farrell and Shapiro (2010) for a discussion of such subsidies in the context of merger control. In Farrell and Shapiro, a merger results in upward pricing pressure because the goods are substitutes. In our setting, a supermarket is analogous to a merger of complementary products and results in downward pricing pressure. The magnitude of the effects we estimate exceed those that usually flag a significant change in competition from a merger in antitrust practice.