Traditionally, conglomerate mergers have raised little antitrust concern since the merging companies’ products were not perceived to compete with each other or to be critical in the merger parties’ value chain. The assessment of these mergers has generally consisted of a check for potential foreclosure strategies by way of tying or the reduction of technological interoperability. More recently, new theories of harm emerged from bargaining theories and dynamic considerations. These theories acknowledge a greater concern about dynamic effects of mergers on innovation, that an increasing number of markets exhibit bargaining power on both sides of a transaction, and that mergers of complements may not be innocuous in markets for increasingly complex products.
This paper, available here, argues that these new theories are not suitable to generate ex ante decision-making rules; instead, their applicability will need to be empirically validated on a case-by-case basis.
Section 2 deals with the traditional treatment of conglomerate mergers.
The most common theories of harm traditionally used in conglomerate mergers concern recourse to tying and bundling as a method to foreclose rivals. The European Commission relied on such theories in the GE/Honeywell merger. In this case, the Commission alleged that the merged entity would have been able to leverage the parties’ respective dominant positions in jet engines and avionics into each other’s markets due to the ability to provide ‘attractive bundled offers’ that would be difficult for competitors to match. In parallel to decreasing the price of the bundle of products, the parties could increase the price of the stand-alone products where they were dominant, making the bundle relatively more attractive. Although the outcome would be efficient for those who buy the bundle at relatively lower prices, there would be a negative effect on those customers that did not wish to do so. The bundling of products by GE and Honeywell was also considered likely to foreclose producers of the now unattractive stand-alone products, leading to long-term harm to customers.
The analysis of foreclosure strategies by tying or bundling, in one version or another, continues to underpin much of the conglomerate analysis in the EU. Invariably, the effectiveness of a foreclosure strategy by way of bundling will depend on the following factors: the existence of strong market power in one of the markets (tying market); demand externalities across the tying and tied products, i.e., the degree of their complementarity; the degree of customer heterogeneity with respect to the preferences for the different products and components; and the existence of barriers to entry, such as fixed costs or economies of scale.
Section 3 looks at bargaining theories of harm.
Bargaining models are a framework to describe markets where sellers and buyers, both with market power, bilaterally negotiate prices. These models rely on the concept of the welfare gain of a sale, defined as the difference between the value of the product to the customer and its cost to the seller, and examine the split of this welfare gain. The magnitude of the price and of the rents that the buyer and seller obtain from the trade are determined by their respective ‘bargaining power’ as well as by the gains that each party expects from the trade, i.e., their incentives to complete the trade. These gains are calculated by comparing profits obtained with the transaction with those obtained absent the transaction, assuming the next best alternative trade (or no trade). Higher gains from the trade imply a higher willingness to concede rent. Mergers that increase bargaining positions are scrutinised for harm to customers by way of their likely higher prices.
By focusing on changes in the value of the agreement to the parties and their customers (or on the cost of not reaching an agreement), bargaining theory presents a new framework for the analysis of mergers. However, it is unclear that it extends the scope for enforcement, since the conditions that lower the cost of a non-agreement are conceptually close to those that enable foreclosure in vertical or horizontal mergers. Further, while bargaining theories may provide grounds for new theories of harm in conglomerate mergers, the relevance of these theories will depend on consumer preferences and on the existence of crucial economic relations across products. Such theories are unsuited to give rise to presumptions of harm, and their actual relevance is yet to be tested.
The increased scope of activity of large firms, which increases the prominence of bargaining theories, also calls for more prominence to be awarded to efficiencies in the analysis of conglomerate mergers. Efficiencies related to production, marketing and distribution could be significant in mergers involving portfolio effects, because common ownership of products exhibiting strong demand-side linkages allows suppliers efficiently to provide a more complete range of products. Conglomerate mergers can also induce economies of scope that reduce the merging producers’ production and distribution costs.
Section 4 looks at mergers that could harm innovation.
Regulators seem increasingly inclined to address the effects of mergers on innovation. Can the growing focus on the effect of mergers on innovation extend to conglomerate mergers? For a theory of harm on innovation to hold in the case of a conglomerate merger some conditions must be fulfilled. These include demand-side linkages, specific margins of different products, barriers to entry and the nature of the innovation threatened by the merger. A merger is more likely to have harmful effects as regards gradual improvements; harmful effects are unlikely in the case of disruptive or significant innovations.
The main virtue of this paper is to describe a number of theories on the possible competition effect of conglomerate mergers. As to its main thesis, it is uncontroversial, I believe: the direction of likely price effects as a result of a conglomerate merger are not certain, and depend on a series of empirical facts regarding market power, demand substitution patterns, demand externalities across products, or the characteristics of innovation. Nonetheless, it would have been nice to have a more thorough summing-up of how these ‘empirical facts’ need to line up in order for a conglomerate merger to have anticompetitive effects, even this already is an informative paper as it stands.