A symposium on the implications of recent merger litigation for market definition.
Key words: market definition, merger analysis, competition policy
This issue of the Antitrust Bulletin focuses on the implications of a number of recent merger litigations for market definition and, therefore, the future of competition policy. In light of the structural presumption, market definition has long played an important and on occasion, controlling, role in merger analysis. (1) Although many economists are willing to dispense with market definition and the 2010 Merger Guidelines introduced the idea of direct evaluation of likely competitive effects, (2) courts are still attached to market analysis, especially for prospective mergers. (3) Of course, as our authors clearly recognize, the analytical procedures for market definition are likely to change, driven by better understanding of the competitive concerns at issue, and these changes may have important implications for predictions of likely competitive effects. Thus, to understand how merger policy is likely to evolve, it is important to study the market definition process used in merger review.
One crucial theme running through many of the articles is the relevance (or lack of relevance) of price competition in driving market analysis. Antitrust policy evolved following the development of an industrial economy in the late 19th century. Products such as steel, oil, and coal lent themselves to modeling as homogeneous goods, and price was the key avenue of competition. Cost often drove pricing with innovation playing a role over the long run. Competitive prices clearly benefited consumers, while high prices, brought on by monopolies, enriched stockholders. As the consumer goods economy developed in the 20th century, price remained a relevant concern in markets as diverse as cigarettes, flexible wrapping materials, and shoes. The 1950 reform of the Clayton Act broadened antitrust law to address asset acquisitions that may substantially lessen competition, in effect establishing the legal foundation for most of modern merger policy. In 1982, the Department of Justice took the next major step, introducing the modern Merger Guidelines with its price-based hypothetical monopolist test and its focus on adverse competitive effects. (4) By the late 20th century, further changes in the economy highlighted the need to address issues other than price competition. For example, in Microsoft, the price of the operating system was never the focus of the innovation-based analysis. (5)
By changing the competitive dynamics of a market, innovation may create the need for new institutions, and those institutions will further change the competitive process. Thus, antitrust analysis must identify both direct and indirect effects of innovation on the competitive process. This situation is particularly clear in health-related markets, as innovation created a range of costly new procedures and treatments and these developments led to new institutions, such as managed care and pharmacy benefit management, designed to influence the competitive process. (6) In other markets, the effects of innovation were more direct, creating new markets, destroying old markets, and lowering costs in existing markets.
Against this dynamic background, merger policy faces pressure to change, although it is far from clear that adopting a policy designed to directly estimate price effects of mergers is a credible choice. Instead, it is likely that we will see theory play a more important role in organizing the merger review, although the choice of theory will be dictated by the factual situation in the specific case. Our studies offer insights into the evolution of merger policy through their discussions of recent merger cases decided in four different industry settings--hospitals, pharmaceuticals, consumer goods, and producer goods. For each industry, a pair of articles is presented, one supportive of relatively aggressive policy and another more attuned to a less aggressive merger regime. For the hospital and pharmaceutical industries, comments are available to provide further perspective.
The first pair of articles addresses mergers involving local hospitals. One transaction, Evanston-Highland Park, was consummated prior to challenge and thus the FTC could evaluate data on market performance, while the others were proposed and thus the analysis was focused on predicting likely competitive effects. To study hospital competition, it necessary to understand the dynamics of the health care sector in the economy. Competition has changed dramatically from the mid-20th century model in which the physician controls everything to the 21st century structure in which physician group practices associate with particular hospitals to create integrated health care systems in which health care could end up institutionalized as a set of treatment protocols. (7)
At the risk of oversimplification, private health care services are purchased by insurance companies, serving as agents for collections of individuals drawn together through their common employment relationships. (8) Institutionally, employer decisions to provide health insurance and the managed care choice of the insurer to contract directly with health care providers are designed to address adverse selection and moral hazard issues that, left uncovered, would significantly undercut the performance of the health care marketplace. (9) The outcome of these institutional constraints involves a set of limitations on consumer choice.
Thus, in studying competition in hospital markets, it is important to understand the key institutional parameters that control competition. As an aggregator of hospital services for their clients, the insurer has an obvious interest in low prices. However, the insurer is also interested in contractual innovations that serve to control the quantity of services provided to patients, while at the same time resolving the medical problems. These innovations could allow the insurer to prefer a higher price, as long as the quantity control arrangements depress the overall costs of medical services.
Our two articles, one by Capps and the other by May and Noether artfully address these and other issues in the context of the recent hospital merger decisions. Capps traces the evolution of hospital merger decisions through their 1990s focus on geographic markets and then explores the impact of the increased relevance of managed care on merger analysis. By offering a more descriptive conceptual frame
work for merger review, as well as new models through which to apply the analysis, economics aided the courts in drawing better geographic markets that facilitated the identification of the competitive concerns. May and Noether agree with this basic story and recognize the impact of the managed care conceptual framework on market definition, but then highlight both new concerns with product market definition and practical limitations associated with the application of the innovative analytical models to hospital mergers. They suggest caution in using models of competitive concern to evaluate the likely effects of hospital mergers. The section concludes with comments by Guerin-Calvert and Wendling and Wilson. Guerin-Calvert's thesis is that competitive effects analysis should reflect the incentives, ability, and effects of insurer responses to an anticompetitive price increase. Wendling and Wilson suggest the analysis focus more deeply on the specific factors that drive consumer substitution and, therefore, the merger's likely competitive effect.
A specific pharmaceutical merger, Lundbeck, is addressed in the next pair of articles. Mergers in the drug industry regularly combine competitive pharmaceutical products as each transaction assembles a portfolio of products for the particular company that is best able to innovate in particular technical areas. As any specific drug is usually ancillary to the merger, acquiring firms are generally willing to divest products in response to a merger challenge. However, in Lundbeck, the acquisitions were limited to a few products, and the ones that were the focus of the FTC investigation were the key to the value of the transaction. Thus, apparently the acquiring firm was unwilling to divest the assets when the deal was challenged.
Historically, pharmaceuticals were heavily differentiated, as firms designed innovative drugs with unique attributes and then "detailed" (marketed) the products to physicians (and hospital boards with the responsibility to purchase drugs used only within the hospital). (10) New products were protected by patents, although successful products eventually faced generic competition. Generic substitution laws, coupled with pharmacy benefit management rules, allowed these generic rivals to undercut the pricing of the branded products and capture the bulk of the business. Generic substitution inspired the growth of competition among therapeutic substitutes. Here, pharmacy benefit managers (or when relevant, hospital pharmaceutical boards) evolved to draw up formularies that listed branded drugs that could be substituted for each other.
Competition caused providers to bid for formulary placement; this competitive process allowed insurers to use rivalry to force price down. However, if the particular drug had no close substitutes for specific medical indications, that drug would retain its historical pricing power until the product went generic. Each formulary decision was made on case-specific data, evaluated by health care professionals, and once made, would generate certain legal liabilities. Thus, formulary substitution would only be specified if the expected savings from competition outweighed the transaction costs associated with the implementation of the formulary entry.
This understanding of the limits of price as a market defining tool, and what can be done in those cases when price--and hence the normal economic tools such as formulary analysis--does not apply, may be crucial in Lundbeck. As discussed in both articles, the FTC brought the challenge under its traditional approach, one that assumed that price competition between the acquired product and those already in hand existed prior to the merger. Sagers and Brunell observe that the available evidence identified significant competition, showing the merger created a monopoly and that even if the firms did not compete on price, they competed on nonprice considerations. In either case, theories of competitive concern require antitrust policy to protect consumers from possible monopolies. However, the court did not consider the products to be competitive; instead, the court found that competition focused on selecting the best product for the particular patient in the neonatal intensive care unit, not on price or marketing. Although the two drugs in question were both usable to treat the same rare condition, their therapeutic profiles differed, and the court found that no reasonable price differential would cause switching from one to the other. Bernard provides a comprehensive discussion of the court's analysis, explains why the court was correct, and then moves beyond the specifics of the case to propose a method to analyze nonprice, patent-centric cases that takes into account the actual drivers in the pharmaceutical industry. Focusing on the comments, Dolin offers the medical perspective that to properly evaluate mergers in the pharmaceutical industry, it is necessary to consider the entire range of factors that go into the physician's choice of drug. Finally, Werden explains that standard merger analysis is derived from economic models built on behavioral assumptions that might not hold for prescription drugs, and therefore, the usual insights are not necessarily relevant.
The third merger, H&R Block, raises issues associated with price competition in consumer goods industries. The merger parties, H&R Block and Tax Act, offer do-it-yourself software to aid in tax preparation. These products may or may not compete closely with storefront tax preparation services, tax professionals, and the taxpayer with pen and paper. In consumer goods, market definition is normally evaluated by observing market behavior and inferring the boundaries of the market. (11) Customer input into the merger analysis is usually limited due to the final consumer status of the purchasers. Hence, for consumer goods, business documents, coupled with any retailer insights tend to control the review process. On occasion, empirical models, some simple (such as critical loss) and others complex (Nash-Bertrand simulation) can be applied to identify markets. Theorists have recently introduced a diversion test which uses estimates of sales lost to close rivals to infer the boundaries of a market (that is, sales not lost to close rivals must be lost to distant rivals and, if lost in sufficient quantity, will negate the proposed market definition). (12) However, such a model would define markets indirectly, using theory to predict the scope of the market. In contrast to the industries discussed above, institutional evidence does not suggest the application of a particular theory. Although it may be reasonable to assume price competition drives a game theoretic equilibrium in some consumer goods industries, other markets involve either simpler or more complex competitive processes. (13) Thus, it would appear necessary to show some type of empirical evidence to justify the application of the theory prior to its use.
In H&R Block, the court found both merger simulation and diversion analysis to be useful, but not dispositive. After a discussion of how courts could identify closeness of competitors in the market analysis, Remer and Warren-Boulton detail the applicability of both merger simulation and diversion analysis for market definition. By using direct evidence to aid the market definition process, the analyst can develop the standard two-stage Merger Guidelines review as a useful "reality check" on the implications of the more direct economic study of the merger's competitive effects. Coate questions the applicability of complex economic models to merger review, suggesting that theory should be used mostly to provide organizational principles for a fact-based analysis. To understand competition in H&R Block, Coate describes a consumer choice model, which if parameterized with data from the record, could have determined whether the market was narrow or broad by identifying the number of marginal consumers. Moreover, empirical analyses, such as the study observed in Staples, are discussed as more relevant than theoretical modeling, as stronger insights are available from "reduced form" models on the effects on competition. (14) Price discrimination is advanced as an alternative approach to test for narrow markets in consumer goods markets.
The final merger analysis focuses on Polypore, an industrial products merger involving battery separators (the insulator installed between the positive and negative plates of the lead-acid battery to force the electrical current to flow in the required direction). For industrial products, customers are knowledgeable about potential substitutes, because the replacement of high cost with low cost inputs is a standard business strategy designed to improve the competitiveness of an end-use product. Thus, the Merger Guidelines' hypothetical monopolist test is almost custom-made to understand the competitive issues facing the firms in producer goods markets. Given clear information from customers, it is not unusual for the market definition process to be perfunctory and thus the number and significance of the rivals affected by the merger will be well defined. As the likely competitive effect of a merger turns on collusion concerns, entry issues, and the potential for efficiencies, the analyst must build a solid understanding of competition within the market. Here, evidence of poor market performance can often justify a merger challenge, as can a credible argument that the merger creates a dominant firm or eliminates an important rival.
Battery separators are used in four different types of batteries: automotive (starter-lighter-ignition, motive, deep cycle, and uninterruptible power supply). All these batteries have different designs and require customized separators. As discussed in the Simpson and Robertson article, price appears controlling in the starter-lighter-ignition niche, and thus a standard merger analysis seems to apply, while in two other niches (motive and deep cycle), superior technology appears to allow one provider to dominate the market. Structural analysis posits somewhat higher prices for both niches once the rival is acquired. Only Polypore competes in the fourth market, and thus it is not affected by the merger. Simpson and Roberson highlight the evidence of anticompetitive effects identified in the investigation as confirming these market-based predictions.
Kahwaty and Tyler take a different approach to the merger review and stress the importance of a consistent analysis. Two mergers are discussed. In a Canadian merger affecting hazardous waste disposal, they stress the importance of matching the analysis used for market definition with the model of competitive effects. Both models are meant to represent the actual competitive process within the market and therefore both must be compatible. In the discussion of Polypore, Kahwaty and Tyler focus more on the need to align the remedy with the theory of concern, itself closely linked to market definition. In particular, if the market is limited to North America and a competitive concern is identified in that area, it makes little sense to make the acquiring firm divest a plant outside the relevant market. To the extent that the plaintiff desires a broader remedy, it should have the burden of linking the success of the remedy with ownership of the foreign plant.
Taken together the articles in our issue are reasonably compatible with a broad interpretation of the concept of direct estimation of competitive effects introduced in the 2010 Merger Guidelines. Instead of thinking of direct estimation as econometric magic to define the likely competitive impact of a merger, the idea uses economic theory to structure the merger review process and produce an integrated study that is able to predict the likely effect of the merger, while at the same time, isolate a relevant market. The analysis can collect price data and estimate a quantitative model or obtain a solid understanding of a competitive process and build a qualitative model of the marketplace. In either case, theory serves to focus both the market and competitive reality on the most relevant facts.
In the hospital business, Capps, together with May and Noether, agree that managed care creates a conceptual framework for organizing the analysis around how the merger affects the ability of managed care firms to create cost-efficient networks for their subscribers, although they differ on the general applicability of specific models. In effect, although managed care considerations tend to drive market definition, the competitive effects analysis may need to look at the totality of the evidence when predicting competitive effects. Diversion-style models may be useful for some transactions and problematic in others.
Moving on to pharmaceuticals, it is possible to see disagreement on the basic approach, with Bernard highlighting how important it is to understand the institutional limits on price competition as well as the actual drivers of investment decisions, while Sagers and Brunell suggest the closest rival must have some constraint (often a price constraint) on the process. In a merger challenge, the judge would try to balance the evidence presented by the parties and resolve the dispute. This evidence can be either qualitative or quantitative in nature, but must offer insights into the competitive process under review.
The same issues are raised in consumer goods markets. Here, Remer and Warren-Boulton detail the application of a specific price model, here one that focuses on competition among close competitors, while Coate argues for retaining the standard Merger Guidelines focus on case-specific evidence of competition. In contrast to the pharmaceutical case in which customer testimony can be collected to determine whether institutional considerations require the analysis to focus on a narrow market, a consumer goods analysis may be based on the assumption of a modeling structure. The court felt the internal documents were sufficiently clear to allow the application of the narrow modeling structure, a structure that then led directly to a competitive concern. Had the analysis focused on a price discrimination model, the plaintiff would have been expected to present a reason for certain consumers to prefer specific tax preparation products.
In the fourth case, Simpson and Robertson argue for the application of standard Merger Guidelines principles to a producer goods merger. Kahwaty and Tyler appear willing to concur on the modeling structure, but note even the Guidelines-based modeling structure must be carefully applied to avoid internal inconsistencies in the merger review process. Thus, organizing principles used to define markets may very well constrain the later analysis of competitive effects, entry, or remedy. To the extent that significant inconsistencies evolve, the merger analyst should identify the problem and clarify the modeling structure to avoid the concern. Sometimes, this may require the use of an integrated approach focused on direct evidence, while other times, this may preclude complex modeling and leave the analyst with the standard two-stage Merger Guidelines process. Every merger is different and requires a review customized to the available facts. Differences in analyst opinion must be expected.
AUTHORS' NOTE: The analyses and conclusions set forth in this issue are those of the authors and do not necessarily represent the views of the Commission, any individual Commissioner, or any Commission Bureau.
MALCOM B. COATE, Economist, Federal Trade Commission.
Kent Bernard, Adjunct Professor, Fordham University of Law.
(1) United States v. Phila. Nat'l Bank, 374 U.S. 321 (1963), concluded that the plaintiff was entitled to a presumption of a concern once it showed the merger affected a concentrated market. United States v. Baker Hughes Inc., 731 F. Supp. 3 (D.D.C. 1990), aff'd, 908 F.2d 981 (D.C. Cir. 1990), clarified the review as a three-stage process in which the plaintiff earns a presumption by establishing a concentrated market affected by the merger, the defendant avoids immediate liability by rebutting the presumption of concern, but the plaintiff can still prevail by meeting the burden of proof.
(2) U.S. Dep't of Justice & Federal Trade Comm'n, Horizontal Mercer Guidelines 9 (2010), available at http://www.ftc.gov/os/2010/08 / 100819hmg.pdf.
(3) For unconsummated transactions, likely competitive effects must be inferred from historical evidence, either directly from natural experiments or indirectly through structural analysis.
(4) U. S. Dep't of Justice, Merger Guidelines, Antitrust Trade Regulation Report, No. 1069 (1982) [hereinafter 1982 Merger Guidelines], Although the courts have not adopted the hypothetical monopolist test as the controlling analytical methodology, it plays an important role in the market definition process.
(5) United States v. Microsoft Corp., 87 F. Supp. 2d 30 (D.D.C. 2000), rev'd, 253 F.3d 34 (D.C. Cir. 2001).
(6) Other countries have substituted government control for market processes to address these concerns, as well as to affect the distribution of income, as many of these programs predate the dramatic boom in health care technology. The United States has a very large government sector, but that portion of the health care industry is rarely considered relevant in antitrust analysis.
(7) In the post-World War II era, physicians seemed to see hospitals as inputs into their management of health problems among their patients. Today, more and more physicians see hospitals as their employer. In light of the nonprofit nature of many hospitals, it remains to be seen whether the physician staff will end up in control of the hospital or whether health care managers will use their near-monopoly on business information to obtain de facto control of the institution. Others might suggest that the hospital and its employees will end up controlled by managed care organizations. And others would posit that the government will rule them all.
(8) Due to a collection of government policies, large employers usually find it efficient to serve as the insurer for their employees, hire a specialist insurance company to manage the network selection and payment process, and diversify away the risk of catastrophic losses on their pool of employees in a reinsurance market. Employer provision of insurance tends to generate universal coverage for the substantial portion of consumers with covered employment. Thus, government policies that undermine employer-sponsored insurance also undermine the policy goal of universal coverage. Institutional reforms are well advised to follow the well-known medical adage of "First, do no harm." For consumers lacking employer coverage, health insurers offer coverage through government-based pools and private offerings. Government payment also plays a large role in health care, but its ability to dictate price minimizes antitrust issues in that segment of the industry.
(9) Two issues that cause particular problems are (1) the correlation of health risk over time, so once an adverse health effect is recognized, the treatment cost in both the current and future periods are expected to be high and (2) the consumers' incentive to demand all possible services once their health is subject to an adverse effect. Employee pooling of risk addresses the first problem and explicit controls on care imposed by the relevant managed care provider limit the second problem. Numerous other problems also affect competition in health care.
(10) Government regulation allows some drugs to be marketed directly to consumers in the over-the-counter segment of the business.
(11) The 1982 Merger Guidelines list various sources for the evidence needed to apply the hypothetical monopolist test. See 1982 Mercer Guidelines, supra note 4.
(12) Joseph Farrell & Carl Shapiro, Improving Critical Loss, Antitrust Source, Feb. 2008, http://www.abanet.org/antitrust/at-source/08/02/Feb08 -Farrell-Shapiro.pdf.
(13) For a detailed criticism of this approach see Malcolm B. Coate & Joseph J. Simons, In Defense of Market Definition, 57 Antitrust Bull. 667 (2012).
(14) A structural model attempts to model the competitive process of interest, and once parameterized with relevant coefficients can predict the effect of changes in the exogenous variables (such as the number of competitors) on the variable of interest (market performance). A reduced form model directly estimates a relationship between changes in exogenous variables and the variable of interest with a generic specification. For a discussion of these issues, see Malcolm B. Coate & Jeffrey H. Fischer, Why Can't We All Just Get Along: Structural Modeling and Natural Experiments in Merger Analysis, 8 Eur. Competition J. 41 (2012).
|Printer friendly Cite/link Email Feedback|
|Author:||Coate, Malcolm B.; Bernard, Kent|
|Date:||Sep 22, 2014|
|Previous Article:||Proactive or reactive? An empirical assessment of IPR policy revisions in the wake of antitrust actions.|
|Next Article:||From Rockford to Joplin and back again: the impact of economics on hospital merger enforcement.|