Suppose that firms A, B, C, and D sell differentiated, competing products (p1…p4), for example, cereal, wine, or word processors. Suppose further that p1 and p2 are each other’s closest substitutes, that is, for customers of p1, p2 is the next best choice, and for customers of p2, p1 is the next best choice. Neither A nor B can raise their prices any further. If A were to raise its prices by 10%, it would lose 15% of its customers; the same is true for B. Now suppose that A acquires B. After the merger, AB would have greater pricing flexibility than was available to either A or B in the pre-merger world. For example, the price elasticity of demand for p1 and p2 combined might be significantly lower than that of either p1 or p2. In that case, AB could profitably raise prices for both p1 and p2. Also, and usually more importantly, AB could raise prices significantly for p1 and moderately for p2. In that case, a 10% price increase for p1 would still lead to 15% customer defection, but if most of these customers defect to p2, AB would capture the otherwise lost profits, which would make the post-merger price increase profitable. As a secondary effect, C and D might also be able to raise prices, because some of the pre-merger pricing pressure from p1 and p2 on p3 and p4 would now have been relaxed. That, in a nutshell, is the standard theory of unilateral effects from a merger of firms with differentiated products. Here is a graph:

In evaluating the effects of the merger, the above model does not take into account product repositioning. That is a significant shortcoming, because price competition is in many cases secondary to competition on features, design, service, novelty, and other dimensions of quality. The 1992 Merger Guidelines therefore acknowledge that

[a] merger is not likely to lead to unilateral elevation of prices of differentiated products if, in response to such an effect, rival sellers likely would replace any localized competition lost through the merger by repositioning their product lines. (Guidelines, §2.212).

Similarly, in U.S. v. Oracle, 331 F.Supp 2d 1098 (N.D. Cal. 2004), the court held that:

the plaintiff must demonstrate that the non-merging firms are unlikely to introduce products sufficiently similar to the products controlled by the merging firms to eliminate any significant market power created by the merger.” (Id., at 1118).

While certainly a step in the right direction, product repositioning by the non-merging firms (in our example C and D) is only half of the story. Of course, if p1 and p2 are close substitutes to begin with and no repositioning of p1 and p2 occurs after the merger, then it is important to know whether p3 and p4 can move into p1’s and p2’s space in the event of a price increase. But what about the effects from repositioning the products of the merging parties, that is, p1 and p2? In many instances, the ability for post-merger product repositioning is an important business reason for the acquisition. But what are the competitive effects of AB moving p1 and p2 away from each other to reduce product cannibalization following the merger? A new paper by Gandhi, Froeb, Tschantz, and Werden makes a significant contribution to answering this question. In many ways, the paper formalizes what antitrust lawyers and their clients have been arguing to the government for years with mixed results. In a nutshell, “merger effects in the price-location [= brand positioning] model are significantly more complex than those in the price-only model,” (Id., at 1) as depicted in the chart below:

Taking both dimensions, price and brand position, into account, the effects from the merger are threefold:

1. Moving p1 and p2 away from each other increases product variety, and at the same time reduces substitutability. As a consequence, unilateral post merger price increases of p1 or p2 become less likely. These are both positive effects from the merger.
2. As AB moves p1 and p2 away from each other, there is an incentive for C and D to position p3 and p4 in the space created between p1 and p2. As a consequence, competition between p3 and p4 may well increase post merger. The more similar p3 and p4 become, the greater the significance of price competition gets between the two brands. This, too, is a positive effect.
3. The further p1, p2, p3, and p4 move away from each other, the greater the ability of each differentiated product to increase price, assuming that there are no other competitors. Whether this is a net-positive effect depends on whether the benefits from increased variety outweigh the “softening of price competition” among spread-out products.

Recognition of the potential welfare gains from (1) and (2), which are usually disregarded in merger analysis, leads Gandhi, Froeb, et al. to the following conclusion:

In our simple model, post-merger product repositioning substantially alters the effects of a merger because the merged firm finds it optimal to separate closely competing products combined by the merger. The merged firm’s product repositioning both mitigates the reduction in consumer welfare the merger otherwise would produce and allows the merged firm to capture a much larger portion of the profits the merger generates. […] The Guidelines and the case law anticipate the possibility that the anticompetitive effects of a merger are mitigated by the repositioning of non-merging products, but not the possibility that the anticompetitive effects of a merger are mitigated by the repositioning of merging products. Our analysis finds that the latter is more important, and repositioning of merging products was observed following the Princess-Carnival merger of two cruise lines, as the Cunard Line was repositioned as a premium brand and P&O Cruises was repositioned as brand designed to appeal primarily to British consumers.

The closer p1 and p2 are to each other in the pre-merger world, the greater the effects from (1) and (2), so that the price + brand position model would predict a more favorable net effect from the merger than the price-only model.

A very interesting paper, highly recommended for anyone involved in dealing with differentiated product mergers.

(Note: This entry was posted originally on the Law & Society Weblog in August 2005.)

Comment feed: RSS 2.0. You can leave a response, or trackback from your own site.