In many markets, the buyer negotiates with competing sellers, and gets an individualised price. This is common in intermediate goods markets, where the sellers are familiar with the tastes of the buyer. We consider individualised pricing in this multi-seller, complete-information setting.1
In recent years many antitrust investigations have considered such markets. In recognition that they deserve special treatment, the 2010 revision to the US Horizontal Merger Guidelines (HMG) added the section Bargaining and Auctions. Examples of markets where this section was relevant include consumer-generated ratings platforms, bought by online retailers (Power Reviews/ Bazaarvoice, 2014), marine water treatment products, bought by owners of fleets of ships (Wilhelmsen/ Drew Marine, 2018), and private label breakfast cereals, bought by supermarket retailers (Post Holdings/TreeHouse Foods Inc, 2020). All are business-to-business markets. See Miller (2014) and Sweeting et al. (2020) for a discussion.
We consider two policy-relevant questions. Does individualised pricing abate market power (for any given market structure)?2 And does individualised pricing mitigate the effect of mergers? We focus on situations where the buyer selects a single seller.3
Market power and merger effects: A tale of two surplus advantages
These policy questions are natural. Individualised pricing works in a different way from standard uniform pricing. With uniform pricing, equilibrium markups depend on market-wide price elasticities and product-ownership portfolios (see Nevo 2001). With individualised pricing on the other hand, an individual buyer’s markup depends on whether she has an attractive substitute product.
To explore the nuts and bolts of individualised pricing, consider a discrete-choice situation for a buyer with multiple competing sellers. We define three key concepts – the first-best product, the runner-up product, and the first-best’s surplus advantage. The first-best product generates the greatest surplus from trade. The runner-up product generates the next-best surplus, not counting the products in the first-best seller’s portfolio. And the first-best’s surplus advantage is the difference between the surplus from trade from these two products.
The first-best’s surplus advantage hands bargaining leverage to the seller of the first-best product and is a key determinant of the markup (defined as price minus cost). In the case of take-it-or-leave-it (TIOLI) pricing, in which sellers make the offers, the equilibrium markup is equal to the first-best surplus advantage (see Thisse and Vives 1988, Miller 2014). Alternatively, suppose the buyer or the sellers can make offers. Then, for several reasonable bargaining specifications, the markup is upper-bounded by the first-best surplus advantage (see Binmore 1985, Manea 2018).4
With this pricing mechanism in hand, how do merger effects work? Consider a merger that changes product ownership but not the set of products or the surplus from any product. Such a merger cannot change the buyer’s choice, because it does not change the first-best product. It does, however, change the first-best’s surplus advantage, if the buyer’s first-best and runner-up products are both owned by insiders to the merger. The size of this change depends on another surplus advantage, the runner-up’s surplus advantage – i.e. the difference in the surplus from trade with the pre-merger and post-merger runner-up.
Thus, to coin a phrase, market power and merger effects, for any individual buyer, is a tale of two surplus advantages – the first-best’s surplus advantage is key for market power and the runner-up’s surplus advantage is key for merger effects.
Comparing uniform and individualised pricing: Theory and evidence
We have now seen that the mechanisms that drive individualised and uniform pricing are different. But does this difference matter? That is, does it abate market power or mitigate merger effects?
A large theoretical literature considers the effect of individualised pricing on market power. In some cases the effect is dramatic. In the TIOLI Hotelling set-up of Thisse and Vives (1988), it reduces average markups by 50% and all buyers benefit.
A smaller theoretical literature considers the effect on mergers. Cooper et al. (2005) considers an example with a triangular Hotelling model, and Sweeting et al. (2020) considers a calibrated logit model. Both find that individualisation can mitigate the effects of a merger on markups.
The theoretical results, however, depend critically on the type of product differentiation that is assumed. This points to the value of empirical modelling.
In Beckert et al.(2021), we consider evidence from the UK, using the market for bricks delivered to house-builders. In this market, the sellers are multi-product manufacturers and the buyers are house-builders with projects in different locations. Both sides of the market are concentrated. We use transaction-level data, giving the negotiated price and chosen product.
We estimate a model of multi-seller bargaining (as in Binmore 1985 and Manea 2018), where the buyer selects the first-best product, and pays a markup that depends on the first-best’s surplus advantage. Product differentiation is modelled using the mixed-logit discrete-choice framework, with the buyer’s location playing a role in determining which products are first-best and runner-up.
We use the estimated model to answer the two questions mentioned above.
First, we investigate whether individualised pricing abates market power. The results from the empirical model imply that average markups are indeed lower with individualised pricing. However, although this average fall is substantial, we do not get the extreme all-markups-fall result of Thisse and Vives (1988). Most markups fall but some rise.
Second, in order to compare merger effects with uniform and individualised pricing, we simulate the exact same counterfactual mergers under the two alternative pricing assumptions. We model mergers as rearrangements of product ownership. We saw, in the discussion above, that merger effects with individualised pricing depend on (i) whether the first-best and runner-up products are party to the merger and (ii) the size of the runner-up’s surplus advantage. The empirical model is able to measure each of these effects for any transaction and the distribution of merger effects across transactions. The results from the model imply that average markup effects are lower with individualised pricing than with uniform pricing. In this sense, the merger effects are indeed mitigated. However, we also find that the distribution of effects from merger is unequal across transactions, and the harm can be greater in some than with uniform pricing.
The 2010 US HMG suggest that merger control should account for individualisation of prices. They suggest looking out for the pre-merger frequency with which merging parties jointly occupy first-best and runner-up status, and whether there is a third product that can stand in effectively for the runner-up post-merger. The current EU and UK merger guidelines by contrast give much less explicit guidance.5
The empirical results reported here, although based on a specific market, provide empirical support for the view that merger analysis should indeed account for individualisation, as suggested in the US HMG. They show that individualised pricing matters. It can abate market power and can mitigate the average effect of mergers. However, the results also suggest that distributional impacts should be considered, particularly for buyers in line for the largest markup increases.
Albæk, S, P Møllgaard, and P B Overgaard (1997), “Government-assisted oligopoly coordination? A concrete case”, The Journal of Industrial Economics, 45(4): 429–443.
Beckert W, H Smith, and Y Takahashi (2001), “Competition in a Spatially-Differentiated Product Market with Negotiated Prices”, CEPR Discussion Paper 15379.
Binmore, K (1985), “Bargaining and Coalitions”, ch. 13 in A E Roth (ed.), Game Theoretic Models of Bargaining, 269-304, Cambridge University Press.
Cooper, J C, L Froeb, D P O'Brien, and S Tschantz (2005), “Does price discrimination intensify competition? Implications for antitrust”, Antitrust Law Journal 72(2): 327-373.
Manea, M (2018), “Intermediation and resale in networks”, Journal of Political Economy 126(3): 1250-1301.
Miller, N H (2014), “Modeling the effects of mergers in procurement”, International Journal of Industrial Organization 37: 201-208.
Nevo, A (2001), “Measuring market power in the ready-to-eat cereal industry”, Econometrica 69(2): 307-342.
Nevo, A. (2014), “Mergers that Increase Bargaining Leverage”.
Osborne, M, and A Rubenstein (1990), Bargaining and Markets, Academic Press.
Sweeting, A, D J Balan, N Kreisle, M T Panhans, and D Raval (2020), “Economics at the FTC: Fertilizer, Consumer Complaints, and Private Label Cereal”, Review of Industrial Organization 57(4): 751-781.
Thisse, J-F, and X Vives (1988), “On the strategic choice of spatial price policy”, The American Economic Review, 122-137.
1 We use the term negotiated to include settings with take-it-or-leave-it offers and those where offers alternate between the buyer and the (competing) sellers.
2 A classic example, relevant for spatially differentiated markets, which dates back to FTC vs. Cement Institute 1948, compares uniform and location-specific (mill-) pricing, for products such as steel and cement, where the seller can condition on the buyer’s location (see Thisse and Vives (1988)). Another example of the debate, which we do not discuss, is whether the secrecy that is associated with individualised pricing tends to change market power. See Albæk, S., P. Møllgaard, and P. B. Overgaard (1997).
3 See Nevo (2014) (and references cited there) for a discussion of situations where the buyer selects several sellers.
4 The result in Binmore (1985) is sketched. Chapter 9.3 of Osborne and Rubenstein (1990) derives it formally.
5 However, like the US guidelines, they do mention the need to account for the possibility that specific consumers may be targeted for price rises when price discrimination is possible.