An Empirical Analysis of Countervailing Power in Business-to-Business Bargaining

Snyder (1996) provides an explanation for discounts that are granted to large buyers, in the context of an infinitely repeated game with upstream competition vis-à-vis a single downstream buyer; there, buyer size makes a seller’s deviation—in the form of discounts—from upstream collusion profitable relative to possible punishment. Smith and Thanassoulis (2008) demonstrate how upstream competition can endow large buyers with market power by inducing supplier-level volume uncertainty. See also Inderst and Mazzarotto (2006) for a survey of theoretical buyer power results and policy implications.

Chipty and Snyder (1999) show that the effect of size on the buyer’s bargaining position depends on the curvature of the supplier’s gross surplus function. Increases in firm size—e.g. as a consequence of a merger—enhance (worsen) the buyer’s bargaining position if the gross surplus function is concave (convex).

Ellison and Snyder (2010), in a study of US wholesale prices in pharmaceutical markets, distinguish between different buyer types—health maintenance organisations and chain and independent drug stores—controlling for institutional differences in drug administration. Using cross-section data, as opposed to the transaction panel data that are used in this study, they cannot account for buyer heterogeneity within type.

See also Adelman (1959), Boulding and Staelin (1990), Brooks (1973), Buzzell et al. (1975), Lustgarten (1975), Clevenger and Campbell (1977), McGukin and Chen (1976),and McKie (1959) for early nonstructural works that provide evidence on countervailing power.

The Competition and Markets Authority (CMA)—established as the successor of the first-stage authority, the Office of Fair Trading (OFT), and the second-stage authority, the Competition Commission (CC), in 2014—has adopted the merger guidance that was originally published by the OFT and the CC.

Competition Commission Guidleines (2003) for market investigation references, paragraph 3.37. Additions in square brackets by the author.

Competition Commission (2007, Appendix C, Table 1).

Competition Commission (2007, Figure 4) shows plant locations. Bricks have a high weight-to-value ratio, while the main fuel to operate kilns, natural gas, has a low weight-to-value ratio. So it is economical to transport the fuel to brick plants at the clay deposits, not to transport clay to delivery points of the natural gas pipeline network. Furthermore, the extracted clay needs to be pre-processed before it can be used to form and fire bricks. Pre-processing involves crushing, filtering, and drying. This requires a processing plant to be close to the clay deposit, even before the brick-forming and firing processes.

The manufacturers in the data also produce a small number of concrete or cinder blocks whose characteristics—notably in terms of load-bearing capacity—are similar to engineering bricks. The CC in its investigation concluded that “the relevant product market affected by the merger [of two brick manufacturers in the sample] included all clay facing and engineering bricks. [...] the relevant product market does not include [...] (b) concrete bricks [i.e. blocks].” (CC Report, paragraph 6) Additions in square brackets are by the author. As concrete blocks do not belong to the same product market as facing and engineering bricks, they are not considered in this analysis.

Competition Commission (2007, paragraphs 8 and 5.20).

They can also be sold to factors, that also buy for resale and that are specialist distributors—acting on behalf of architects or builders—and distinct from merchants in that they do not have physical storage facilities. The accounting data do not distinguish factors from merchants, so the analysis treats them as merchants.

This is confirmed by industry associations that represent brick manufacturers, builders, and merchants. One of the four manufacturers in the data, for example, in 2008 had five regional sales offices in England, one in Scotland, and one in Northern Ireland.

Competition Commission (2007, paragraph 4.54).

I have access only to manufacturer data, not merchant data. Not all manufacturer data provide consistent information on haulage. Therefore, the data do not permit a comprehensive investigation of haulage in the UK brick industry.

Framework agreements set out annual matrices of ex-works and/or delivered prices vis-à-vis brick specifications and target volumes. These are not formal contracts, but are merely indicative commercial conditions subject to which transactions can, but need not, be called off.

According to the UK National Association of Estate Agents, the average detached house in 2006 required 5400 bricks. The CC report, paragraph 4.30, states 7000–10,000 bricks per dwelling.

The restoration and modernisation of the station about a decade ago required 60 million bricks; http://​www.​bifm.​org.​uk/​bifm/​filegrab/​StPancrasFactShe​et.​pdf?​type=​documents&​ref=​2965.

Flues are outlets of waste gases, from boilers, heating, and ventilation systems.

Competition Commission (2007, paragraph 4.45).

Prices of more than £4.50 per hand-made brick are not uncommon, according to manufacturers’ online product price lists.

This appears to be the case of a brick delivery in Cornwall that originated at a plant in the Newcastle area.

http://​www.​bbc.​co.​uk/​news/​magazine-33095262.

Competition Commission (2007, paragraph 5.44).

https://​www.​gov.​uk/​government/​collections/​rural-urban-classification.

Competition Commission (2007, Figure 4) shows the locations of manufacturing plants.

Boyer (1996) and Hyde and Perloff (1995) discuss the merits of reduced form versus structural analysis of market power more generally.

See Competition Commission (2007, Appendix C, paragraph 23) for the non-binding nature of framework agreements.

See Competition Commission (2007, paragraph 5).

According to CEBR (2016), in 2006 the cost of housing per brick was £64.59 in London and £52.37 in Cambridge, for example. This compares to an average price of a facing bricks of £0.75 in the contract data.

See Competition Commission (2007, paragraph 4.32 and Appendix C, paragraphs 15 and 18).

The specification of the pricing equation that is linear in the levels of this variable produced a better fit than a specification that is linear in the log of this variable. There are no a priori theoretic reasons for either specification. This is an empirical matter. The semi-log specification in this application means that the elasticity of price with respect to the buyer’s local commercial significance is not constant, but varies with the level of that variable.

An urban versus rural split is motivated by construction restrictions that are due to rigid nature conservation regulations in the UK.

To economise on space, the first-stage coefficient estimates of the remaining coefficients are omitted, but are available on request.

For both builders and merchants, F-tests failed to reject the hypothesis that interactions with a dummy variable for “small” are jointly zero at the 5% level. The result is stronger for merchants due to the larger sample size: 1161 merchants account for 90.8 % of contracts in the data.

Large buyers cannot pretend to be small by spreading the required volume across several manufacturers, however. Bricks are differentiated products, notably with regard to their aesthetic appearance. This is a consequence of the particular clay that is used. This means, in particular, that products by different manufacturers do not look the same. Therefore, for a given project, only bricks from a single manufacturer can be used, and order division is not an option.

Builders have little, if any, incentive to merge. Delivery by the manufacturer is to a location—a construction site—and the number of bricks that are required at that location is determined by the nature of the building. With bricks being a small part of the overall construction cost, one would not redesign to change the number of bricks required. A merged firm or cooperative would have to arrange transport from the location to which the manufacturer delivers to any other location. To the extent that there is a cost advantage in this, merchants do it already.

The econometric evidence on nonlinear volume effects is presented below.

Since the network of large merchant store locations (large merchants have several hundred stores in the UK) is much denser than the network of manufacturing plants (54), virtually all transactions in the data bypass a large merchant—in the sense that the distance between delivery site and the chosen manufacturing plant exceeds the distance to the closest store of one of the five largest merchants. 99.8% (99.7%) of small builder (merchant) contracts (with manufacturers) bypass the closest of the five largest merchants. And 95.2% (95.1%) of small builders (merchants) have contracts only with manufacturers that bypass these merchants. The median bypass distance is 106 km (93.6 km) for builders (merchants). This suggests that the small buyers in the data reveal themselves as being selective with whom they deal. They could have bought from at least one of the five large merchants, but chose to bypass that merchant and to buy from a manufacturer instead.

Merchants carry a wide variety of construction materials other than bricks, including timber, plumbing, electric equipment, paints, etc.

See Table 9 in the Appendix, which replicates specifications IV(1) and IV(2) in Table 6, with the number of large merchants within 50 km of the buyer site included.

As the estimated model is log-linear, these estimates are calculated as \(\exp (\hat{\alpha })-1\).

Competition Commission (2007, paragraphs 5 and 9, and paragraphs 4.33–4.37).

e.g. Paragraphs 4.61–4.64.

Separate data contain some coarse information on cost at the manufacturing plant level, but even this information does not appear to be consistent across all pants. Therefore, it is not utilised in this study.

Examples of other inquiries into bulk industries where buyer power was considered are Office Of Fair Trading OFT (2011a, b); Competition Commission (2012a, b). An example of a business-to-business relationship that involves bilateral bargaining for services is statutory company audits. The OFT refers to countervailing power of some auditees in its referral document to the CC; see OFT (2011c, paragraph 5.25).