Scale economies and synergies in horizontal merger analysis

Antitrust Law Journal

Volumn 68, Issue 3, 2001, Pages 685-709

  Farrell, Joseph ^a   Shapiro, Carl ^a  

^a UNIVERSITY OF CALIFORNIA   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 0346491841     PISSN: 00036056     EISSN: None     Source Type: Journal    
DOI: None     Document Type: Conference Paper

References (50)

1. U.S. Department of Justice and Federal Trade Commission Horizontal Merger Guidelines § 4 (1992, revised 1997), reprinted in 4 Trade Reg. Rep. (CCH) ¶ 13,104 [hereinafter Merger Guidelines]. An important question, but one that we do not treat here, is how the standard weighs consumer surplus against surplus accruing to the parties or other non-consumers. At present there appear to be differences among jurisdictions on this point. A recent case before Canada's Competition Tribunal (Canada (Commissioner of Competition) v. Superior Propane Inc., 2000 Comp. Trib. 15 (Aug. 30, 2000)) affirmed a total-surplus standard, while a recent U.S. case (United States v. Franklin Electric, 00- C-0334-C (W.D. Wise.) (citing FTC v. Elders Grain, 868 F.2d 901, 904 (1989)), affirmed a standard that requires benefits to consumers (or non-parties, such as consumers). The European merger control statute forbids a "concentration which creates or strengthens a dominant position as a result of which effective competition would be significantly impeded . . ." (Article 2, Section 3), and this could be interpreted as hostility to efficiencies achieved by dominant players. Some indeed have referred to an "efficiencies offense" in Europe.
2. U.S. Department of Justice and Federal Trade Commission Horizontal Merger Guidelines § 4 (1992, revised 1997), reprinted in 4 Trade Reg. Rep. (CCH) ¶ 13,104 [hereinafter Merger Guidelines]. An important question, but one that we do not treat here, is how the standard weighs consumer surplus against surplus accruing to the parties or other non-consumers. At present there appear to be differences among jurisdictions on this point. A recent case before Canada's Competition Tribunal (Canada (Commissioner of Competition) v. Superior Propane Inc., 2000 Comp. Trib. 15 (Aug. 30, 2000)) affirmed a total-surplus standard, while a recent U.S. case (United States v. Franklin Electric, 00-C-0334-C (W.D. Wise.) (citing FTC v. Elders Grain, 868 F.2d 901, 904 (1989)), affirmed a standard that requires benefits to consumers (or non-parties, such as consumers). The European merger control statute forbids a "concentration which creates or strengthens a dominant position as a result of which effective competition would be significantly impeded . . ." (Article 2, Section 3), and this could be interpreted as hostility to efficiencies achieved by dominant players. Some indeed have referred to an "efficiencies offense" in Europe.
3. The FTC voted 3-2 in July 2000 to challenge Heinz's acquisition of Beech-Nut. According to the FTC, after the merger, two firms, Gerber and Heinz/Beech-Nut, would control 98 percent of the U.S. baby food market. In October, Judge James Robinson denied the FTC's request for a preliminary injunction to block the merger. FTC v. H.J. Heinz Co., No. 1:00 CV 01688, slip op. (D.D.C. Oct. 18, 2000). While Judge Robinson agreed that the FTC had established its prima facie case based on market definition, concentration, and entry barriers, he found for the defendants based on efficiencies: "My conclusion in this case does not rest on aspirational testimony, but instead credits powerful evidence in the record about the efficiencies realized by the merger and about the enhanced prospects of the merged entity to introduce innovative products to compete with Gerber. That evidence shows, in my view, that the Commission's prima facie case inaccurately predicts the merger's probable effect on future competition. See Baker Hughes, 908 F.2d at 990." Id. at 17-18 (footnote omitted). The particular efficiencies credited by Judge Robinson were cost savings from the consolidation of production at the more efficient Pittsburgh plant, and from using Heinz's more efficient distribution network. The court also accepted the parties' assertion that "those efficiencies will enable Heinz to provide the best of the two companies' recipes under the new Heinz/Beech-Nut (or Beech-Nut/Heinz) label, and to apply its value pricing strategy to the entire combined product volume." Id. at 20. Judge Robinson also agreed with Heinz and Beech-Nut that the merged entity would have enhanced incentives to innovate and introduce new products because more stores would be carrying its combined product line. See id. The FTC appealed, and, on November 8, 2000, the D.C. Circuit Court of Appeals (Judges Ginsburg, Tatel, and Garland) issued an injunction pending appeal, describing the efficiencies argument as "a novel defense, which the Supreme Court has not addressed since the 1960s (and then unfavorably)," and hinting at particular skepticism in highly concentrated markets.
4. We are certainly not the first to address the question of how efficiencies have been treated under the revised Merger Guidelines. For a very informative set of papers on just this topic, see Symposium, The Changing Face of Efficiency, 7 GEO. MASON L. REV. 485 (1999). In particular, see Robert Pitofsky, Efficiencies in Defense of Mergers: Two Years After, 7 GEO. MASON L. REV. 485 (1999); Craig W. Conrath & Nicholas A. Widnell, Efficiency Claims in Merger Analysis: Hostility or Humility?, 7 GEO. MASON L. REV. 685 (1999); Jerry A. Hausman & Gregory K. Leonard, Efficiencies from the Consumer Viewpoint, 7 GEO. MASON L. REV. 707 (1999); and Timothy J. Muris, The Government and Merger Efficiencies: Still Hostile After All These Years, 7 GEO. MASON L. REV. 729 (1999).
5. We are certainly not the first to address the question of how efficiencies have been treated under the revised Merger Guidelines. For a very informative set of papers on just this topic, see Symposium, The Changing Face of Efficiency, 7 GEO. MASON L. REV. 485 (1999). In particular, see Robert Pitofsky, Efficiencies in Defense of Mergers: Two Years After, 7 GEO. MASON L. REV. 485 (1999); Craig W. Conrath & Nicholas A. Widnell, Efficiency Claims in Merger Analysis: Hostility or Humility?, 7 GEO. MASON L. REV. 685 (1999); Jerry A. Hausman & Gregory K. Leonard, Efficiencies from the Consumer Viewpoint, 7 GEO. MASON L. REV. 707 (1999); and Timothy J. Muris, The Government and Merger Efficiencies: Still Hostile After All These Years, 7 GEO. MASON L. REV. 729 (1999).
6. We are certainly not the first to address the question of how efficiencies have been treated under the revised Merger Guidelines. For a very informative set of papers on just this topic, see Symposium, The Changing Face of Efficiency, 7 GEO. MASON L. REV. 485 (1999). In particular, see Robert Pitofsky, Efficiencies in Defense of Mergers: Two Years After, 7 GEO. MASON L. REV. 485 (1999); Craig W. Conrath & Nicholas A. Widnell, Efficiency Claims in Merger Analysis: Hostility or Humility?, 7 GEO. MASON L. REV. 685 (1999); Jerry A. Hausman & Gregory K. Leonard, Efficiencies from the Consumer Viewpoint, 7 GEO. MASON L. REV. 707 (1999); and Timothy J. Muris, The Government and Merger Efficiencies: Still Hostile After All These Years, 7 GEO. MASON L. REV. 729 (1999).
7. We are certainly not the first to address the question of how efficiencies have been treated under the revised Merger Guidelines. For a very informative set of papers on just this topic, see Symposium, The Changing Face of Efficiency, 7 GEO. MASON L. REV. 485 (1999). In particular, see Robert Pitofsky, Efficiencies in Defense of Mergers: Two Years After, 7 GEO. MASON L. REV. 485 (1999); Craig W. Conrath & Nicholas A. Widnell, Efficiency Claims in Merger Analysis: Hostility or Humility?, 7 GEO. MASON L. REV. 685 (1999); Jerry A. Hausman & Gregory K. Leonard, Efficiencies from the Consumer Viewpoint, 7 GEO. MASON L. REV. 707 (1999); and Timothy J. Muris, The Government and Merger Efficiencies: Still Hostile After All These Years, 7 GEO. MASON L. REV. 729 (1999).
8. We are certainly not the first to address the question of how efficiencies have been treated under the revised Merger Guidelines. For a very informative set of papers on just this topic, see Symposium, The Changing Face of Efficiency, 7 GEO. MASON L. REV. 485 (1999). In particular, see Robert Pitofsky, Efficiencies in Defense of Mergers: Two Years After, 7 GEO. MASON L. REV. 485 (1999); Craig W. Conrath & Nicholas A. Widnell, Efficiency Claims in Merger Analysis: Hostility or Humility?, 7 GEO. MASON L. REV. 685 (1999); Jerry A. Hausman & Gregory K. Leonard, Efficiencies from the Consumer Viewpoint, 7 GEO. MASON L. REV. 707 (1999); and Timothy J. Muris, The Government and Merger Efficiencies: Still Hostile After All These Years, 7 GEO. MASON L. REV. 729 (1999).
9. Merger Guidelines, supra note 1, § 0.1.
10. The Guidelines describe a market as "unconcentrated" when post-merger HHI is below 1000. Id. § 1.51(a).
11. FTC v. Cardinal Health, Inc., 12 F. Supp. 2d 34, 43-44 (D.D.C. 1998). One of the authors, Carl Shapiro, served as an expert witness for the FTC in this case.
12. Id. at 61. The merging parties also emphasized that entry into drug wholesaling was easy and would protect customers from any anticompetitive price increases. They were hard pressed, however, to explain why small entrants would be efficient even while the merging firms, already far the largest in the industry, needed to grow yet larger to become efficient. Id. at 54-55. As often occurs, the parties' entry arguments and efficiency claims were at odds with each other. All those who have studied economics are familiar with classical "U-shaped" average cost curves, where costs per unit are high at very small scales, fall to a minimum at some optimum scale, and then rise at even larger scales. Merging parties all too often assert (implausibly) the exact opposite: a hump-shaped average cost curve whereby small entrants are efficient enough to discipline the market, yet the merging firms (with intermediate scale) could significantly lower their costs if they were to grow even larger. (And even if costs took this surprising form, merger would not be the only way to get off the hump.)
13. Id. at 62.
14. Id.
15. Id. at 63.
16. This distinction follows the formal definitions we gave ten years ago in our more theoretical paper, Horizontal Mergers: An Equilibrium Analysis, 80 AM. ECON. REV. 107 (1990).
17. We are focusing here on the dominant consumer-welfare standard for the evaluation of mergers. With a total-welfare standard, i.e., if one includes accounts for profits as well as consumer welfare, the delay in achieving efficient scale would weigh more heavily. For example, with highly inelastic demand (up to a reservation price), even a merger to monopoly, while leading to higher prices, might create little by way of conventional deadweight loss yet permit efficient production sooner than would arise under competition. Of course, even this observation could be reversed if competition provides the pressure necessary for the supplier(s) actually to achieve cost reductions. See generally Gregory J. Werden, An Economic Perspective on the Analysis of Merger Efficiencies, ANTITRUST, Summer 1997, at 12.
18. Merger Guidelines, supra note 1, § 4.
19. This criticism is similar to a concern about attacks on collaborations generally based on the availability of "less restrictive alternatives" that could, in theory, achieve the same legitimate business ends of the parties with less of an adverse impact on competition.
20. Merger Guidelines, supra note 1, § 4 (footnote omitted).
21. This appears to be the position taken by Hausman and Leonard: "Our recommendation is that when a given merger increases consumer welfare [relative to the pre-merger situation], it should be deemed pro-competitive." Hausman & Leonard, supra note 3, at 719. Hausman and Leonard take the view that alternatives to the merger other than the status quo "are typically speculative and not based on a reliable analysis." Id. In contrast, Conrath and Widnell argue that it is the efficiency claims themselves that tend to be speculative, so that the agencies and the courts should treat many of them with skepticism. See Conrath & Widnell, supra note 3, at 705. The massive disruptions of service following the Union Pacific/Southern Pacific railroad merger, approved by the Surface Transportation Board based in part on the STB's expectation of enormous efficiencies, is a recent and highly visible case of faulty efficiency claims. Id. at 697-99. This merger was approved over the objections of the Department of Justice; Shapiro served as Deputy Assistant Attorney General for Economics at the Antitrust Division while the UP/SP merger was reviewed there.
22. Among other problems, to do so would also create undesirable incentives for firms contemplating mergers to postpone ordinary efficiency improvement.
23. 2. Of course, in practice there tends to be some degree of product differentiation in most markets, but (for example) we are hesitant to characterize as "synergies" the combination of one firm's brand name with another firm's production capabilities unless the brand name is commercially quite important and the production facilities are specialized and could not easily be replicated by the company owning the brand name.
24. Farrell & Shapiro, supra note 11.
25. Id. at 110-11.
26. We recognize that the general definition of "synergies" has some inherent lack of precision. If one firm could replicate the other's information technology capabilities simply by hiring a new Chief Information Officer, which we would consider an "asset" available in the market, then improvements in information technology based on learning from the other firm are not "synergies." Alternatively, if the specific skills at one firm are extremely hard, or even impossible, to find outside that firm, such learning would be a "synergy." As a practical matter, we believe that the enforcement agencies and merging parties should be asking whether the assets being combined are "generally available" or "highly specific" to the merging firms. But also note an efficiency based on purchasing assets or inputs that are readily available in the market may well not be merger-specific in the first place, because one firm may well be expected to acquire the necessary assets or inputs without the necessity of entering into an otherwise anticompetitive merger.
27. Merger Guidelines, supra note 1, § 4.
28. Id.
29. In a declining industry, a similar question could arise even if historically both parties have achieved efficient scale but at least one is now in danger of dropping well below MES. As discussed above, it is really absent-merger versus with-merger that matters, not literally pre-merger versus post-merger.
30. The parties argued that not only would the combined firm immediately be able to serve a larger set of locations than either could pre-merger, but that it would also increase the incentives to expand to locations outside either party's territory.
31. Petition to Deny of Sprint Communications Company, LP at 49, In re SBC Communications, Inc., FCC Docket No. CC 98-141. Farrell was a consultant for Sprint in this merger proceeding. As is often the case in telecommunications, the distinction here between scale and scope is not simple.
32. Farrell & Shapiro, supra note 11, at 111-13. We caution, however, that the result is a little more subtle than this: in particular, it does not rule out expansion of output by one of the partners, but only finds that such expansion must be more than undone by output reduction at the other partner's facilities.
33. FTC v. Staples, Inc., 970 F. Supp. 1066 (D.D.C. 1997).
34. Pitofsky, supra note 3, at 488.
35. We repeat that the no-synergies theorem makes a strong statement about consumer welfare; it by no means rules out the possibility that a merger without synergies can raise total welfare, e.g., through the reduction of substantial fixed costs, which leads to sharply higher profits.
36. See Richard G. Parker & David A. Balto, The Merger Wave: Trends in Merger Enforcement and Litigation, 55 BUS. LAW. 351 (1999), for a further discussion of Judge Sporkin's treatment of efficiencies specifically, and of recent merger enforcement more generally.
37. Formally, this is the case of a positive conjectural variation parameter: each firm expects the other to match output expansions or reductions. Such behavior is less competitive than Cournot behavior. As should be clear, the suggestion is that substantial economies of sheer scale may be more plausibly merger-specific the less competitive the pre-merger behavior of industry participants.
38. Even in this example, our familiar tension resurfaces: if there are many such new customers, the prospect of serving them (through a strategy of price discrimination) will make it more attractive for one of the merging firms to expand production without a merger, as this will allow it to achieve the lower marginal cost, attendant to greater scale, needed to make serving the new customers profitable.
39. Of course, in a fully competitive market, mergers may be one of the means by which firms most rapidly can achieve newly available economies of scale, and there is no need for merging parties to mount an efficiencies defense because, even together, they lack market power.
40. Again, the no-synergies theorem does not state that economic welfare must fall with a no-synergies merger. Indeed, our 1990 article showed that total surplus may rise, even as consumer surplus falls, with such a merger. See Farrell & Shapiro, supra note 11.
41. Difficult issues might arise here concerning firms' incentives to experiment and invest in making a nonmerger relationship work, when that is possible but difficult, if failure to do so will facilitate merger clearance.
42. An example might perhaps be the inefficiency associated with roaming arrangements between wireless telephone providers in different territories; by charging each other abovecost per-minute roaming fees, the two providers artificially inflate their marginal costs of provide roaming services to their customers, possibly (though there are other theories) raising the price of such services above what an integrated company would charge. A rather similar potential efficiency appears to have been a source of concern to European authorities in the proposed merger between Telia and Telenor, the Swedish and Norwegian incumbent telephone companies.
43. This example is taken from the recent merger of BP and ARCO; author Carl Shapiro served as a consultant and expert witness for BP and ARCO in this merger.
44. As a general matter, one can argue that something must be changing in the industry because the proposed merger was not previously proposed. It is unclear how far one ought to push that argument, and we merely note it for now.
45. This, and the rest of the discussion, of course presupposes that the lines converge at each end at some type of terminal, so that trains can shift from line to line (even if this was not the practice pre-merger).
46. Here we assume that improvements to both firms' prowess are not passed through to consumers, because they jointly lack market power.
47. Firms without market power may also be less likely to be accused or convicted of collusion if they reach a cooperative agreement that implements synergies.
48. Why do we describe economies of scale as a synergy here, but not above? The point is that here we are considering a situation in which there are hard-to-trade "capital stocks" that affect the production function and that will be intimately combined post-merger, so that the two firms jointly do indeed have a different production function than the sum of the two.
49. See, e.g., Gregory J. Werden, A Robust Test for Consumer Welfare Enhancing Mergers Among Sellers of Differentiated Products, 44 J. INDUS. ECON. 409 (1996); Luke M. Froeb & Gregory J. Werden, A Robust Test for Consumer Welfare Enhancing Mergers Among Sellers of a Homogeneous Product, 58 ECON. LETTERS 367, 367-69 (1998). Hausman and Leonard offer a numerical example in which a merger between a brand with a 50% market share and a brand with a 5% market share, which leads to a 10% reduction in marginal costs for both merging brands, generates net consumer benefits. Hausman & Leonard, supra note 3, at 715-16.
50. See, e.g., Gregory J. Werden, A Robust Test for Consumer Welfare Enhancing Mergers Among Sellers of Differentiated Products, 44 J. INDUS. ECON. 409 (1996); Luke M. Froeb & Gregory J. Werden, A Robust Test for Consumer Welfare Enhancing Mergers Among Sellers of a Homogeneous Product, 58 ECON. LETTERS 367, 367-69 (1998). Hausman and Leonard offer a numerical example in which a merger between a brand with a 50% market share and a brand with a 5% market share, which leads to a 10% reduction in marginal costs for both merging brands, generates net consumer benefits. Hausman & Leonard, supra note 3, at 715-16.