FEDERAL TRADE COMMISSION
HEARINGS ON GLOBAL AND INNOVATION-BASED COMPETITION
Efficiencies in Dynamic Merger Analysis
Statement of Steven C. Salop
November 2, 1995
Statement of Steven C. Salop
I feel privileged to participate in the Commission's hearings on Global and Innovation-Based Competition. The Commission and its staff can improve our understanding of modern industrial competition as well as the proper role of antitrust in regulating that competition to maximize consumer welfare and economic progress. Based on the list of topics and the mix of academic and business participants, these hearings are sharply defining the agenda for antitrust policy makers and academic scholars.
Today's panel is one of a series on the role of efficiencies analysis in antitrust, particularly merger enforcement. This subject has been an area of my recent research and teaching, so I am delighted to be here. I hope that I can make a contribution to the important questions you have posed for the panel.
My comments today on dynamic efficiency analysis grow out of joint work with Gary Roberts, the Associate Director for Antitrust in the Bureau of Economics. I also am submitting our article as an attachment to this testimony. Before I discuss this work, however, I want to set forth some caveats and acknowledgments.
First the caveats. The opinions in the article are not necessarily those of the Commission, individual Commissioners or Commission staff, other than Gary Roberts, of course. (But, we certainly hope the Commission eventually will come to agree with us.)
As for the acknowledgments, I want to emphasize the large number of important articles on efficiency analysis. We are not the first to study this issue. Three authors stand out. First, the work of Oliver Williamson, arising out of his own sojourn at the Department of Justice, obviously have been seminal to every antitrust scholar. Second, as in many areas, Joseph Brodley has brought law and economics together in a rigorous and subtle way. These articles and others have pointed the way for the rest of us. Third, Chairman (then Professor) Pitofsky's recent article has propelled my own work by framing the policy issues and introducing a number of provocative policy suggestions. In light of my initial strong disagreements with much of his analysis, I have been surprised at how much I agree with him now.
I want to discuss the efforts of Gary Roberts and myself in formulating a dynamic framework for analyzing and evaluating efficiency claims. This dynamic framework attempts to provide a rigorous economic foundation for a potentially greater role for efficiency claims within the context of the antitrust goal of maximizing consumer welfare. It also leads to a number of recommendations for the way in which efficiency claims in merger enforcement should be evaluated and balanced against anticompetitive concerns.
II. The Limits of Static Efficiency Analysis
In Brown Shoe, the Supreme Court viewed the increases in efficiency flowing from the merger as part of the offense. Antitrust has come a long way since then. The efficiency benefit of cost savings and superior products are now recognized as a legitimate justification in antitrust. The rule of reason is designed to balance the tradeoff between the likelihood of significant efficiency benefits and the likelihood of potential anticompetitive harms. Even horizontal price fixing cases like BMI qualify for such rule of reason evaluation if the joint pricing creates efficiency benefits.
It is clear that the motivation and effect of many mergers is to reduce costs and improve products. Mergers involve the real asset integration that is associated with increases in efficiency. The Guidelines present the state of the art in analyzing other issues. Yet, the Guidelines' analysis of efficiencies is nothing more than a brief "placeholder." Surprisingly, even that placeholder is not without controversy.
Why? The fundamental reason why efficiency analysis has lagged behind analysis of market power may be the view that it is impossible to balance cost savings against price increases in a way that is consistent with the goal of antitrust as the maximization of consumer welfare. Until the issue of the antitrust welfare standard can be resolved, the process appears stalemated.
This controversy over the antitrust standard involves the conflict between the aggregate economic welfare (AEW) standard and the pure consumer welfare (PCW) standard. The AEW standard balances the cost savings efficiencies from the merger against the deadweight loss harm resulting from a post-merger price increase. The monopoly overcharge paid by purchasers to stockholders of the firm resulting from this price increase is treated as a transfer from one member of society to another and so is ignored in the balance. Bork claims that the AEW standard reflects the proper consumer welfare standard.
Proponents of the PCW standard disagree. They argue that the monopoly overcharge represents a real harm to consumers, not "just" a transfer from one consumer to another. In effect, they accuse Bork of nominating rich stockholders as "honorary" consumers. Instead, they argue that the proper standard would count only the welfare of purchasers and that standard would be violated by a merger that raises price, regardless of the magnitude of the cost savings. Only if cost savings are passed on to purchasers in the form of price reductions should they constitute cognizable efficiency benefits.
The PCW seems to better reflect the mainstream view of consumer welfare. However, adoption of the static PCW standard unfortunately leaves little room for efficiencies compared to the AEW, at least for low demand elasticities. In this case, only dramatic cost savings would be passed on in the form of price reductions. For example, in the Roberts-Salop simulation model, a merger between two firms in a ten-firm market would only lead to lower prices if variable costs were reduced by 11.1 percent. In a six-firm market, the required variable cost reduction is 20 percent. In contrast, if the static AEW standard were used instead, the critical cost reductions would be 1.1 and 3.4 percent respectively.
Adoption of a dynamic framework for analyzing merger efficiencies has the potential to eliminate this paralysis. The dynamic framework provides a far more realistic account of the manner in which merger efficiencies increase competition. In particular, the dynamic framework recognizes that cost savings achieved by the newly merged entity generally will diffuse at least partially to competing firms over time. As this diffusion occurs, the aggregate cost savings multiply. This diffusion also increases competition and increases the likelihood that firms will pass the cost savings on to consumers by reducing prices and, thus, raise consumer welfare.
As a result of this diffusion and increased price competition, the tension between the static AEW standard of Bork and the static PCW standard can be at least partially resolved by the dynamic framework. One interpretation of the claims made by proponents of the AEW standard is that the cost savings will increase social efficiency and consumer wealth in the long run. The dynamic framework recognizes that the mechanism by which the initial cost savings are passed on to consumers is diffusion of the cost savings to rivals and, consequently, intensified price competition that leads to lower prices and higher real consumer wealth.
III. Dynamic Analysis of Merger Efficiencies
Efficiency improvements are not static, one-time-only events. Rather they occur as part of a rich dynamic process in which efficiency improvements are introduced for private gain but then frequently stimulate competition that creates significant spillover benefits for consumers. Mergers can speed the pace of technical progress and reduce prices by facilitating innovations that initiate technological diffusion and induce competitive innovations.
The spread of process and product innovations through technology diffusion is common. Intellectual property rights and secrecy often are insufficient to prevent substantial leakage of critical information. For example, Henry Ford's assembly line process was imitated by other auto companies. Similarly, GM and Toyota tried to justify their joint venture on the grounds that the innovations they developed would spread. As a result, innovations spread to the competitors of the innovator over time. Mergers also can lead to diffusion of cost savings over time through the broader process of inducing competitive innovation. Competitive pressure may spur rival firms to increase their independent investments to keep up with the newly merged entity. When rivals fall behind, they may redouble their efforts to catch up by reducing their own costs and improving their products.
This process of technological diffusion and competitive innovation reduces the financial returns to the initial innovator. However, it typically does not destroy incentives. The diffusion and response process generally takes time and may be incomplete as well. This also means that prices may rise for a period before they fall.
Mergers can increase the speed and magnitude of cost savings by enhancing the magnitude of financial returns from investment in innovation. First, a merger may combine complementary assets in a way that increases efficient resource use. Second, a merger can allow the entity to spread unit cost savings over a larger output base. Third, a merger may reduce the risk associated with the investment. Fourth, a merger may allow the combined firm to implement efficiency improvements more rapidly than they could independently.
This dynamic analysis of diffusion raises an important potential criticism, however. It is well recognized that if efficiencies can be achieved unilaterally, they can not justify a merger or joint venture. It might be argued that this is inconsistent with diffusion of innovations to rivals. On the one hand, if the merger is necessary to induce these efficiencies, then imitation by rivals seems unlikely. On the other hand, if diffusion is likely, then the innovation must be predictable and so the merger must not be necessary to induce the innovation to begin with.
This criticism generally ignores the potential for competitive innovations by rivals. It also fails to recognize the important role that mergers may play in promoting investments that may be copied if they succeed. Consider the following two examples.
Example 1: A merger predictably can induce innovations solely by increasing the scale of unit cost savings. In particular, merging can increase the expected return from a risky investment by providing a larger scale of output on which to reduce production costs, if the investment turns out to be successful. If the risky investment does succeed in reducing costs (and a certain fraction will) and competitors find out, then those competitors risklessly and unilaterally can imitate this cost savings approach. In contrast, if the merger is not permitted, there will be no innovation to imitate. Thus, the necessity of a merger to facilitate the first innovation is consistent with diffusion to competitors.
Example 2: Suppose that two merging firms combine the following complementary assets, know-how or creativity of the first firm with the technology or facilities of the second firm. The merger may reduce costs by applying the first firm's know-how to the second firm's facilities or technology. The exact mechanism by which the know-how will lead to lower costs may be uncertain, but the likelihood that costs will be reduced may be fully predictable in advance. In this example too, these merger-specific cost savings easily can diffuse to competitors. Once the exact method of cost reduction becomes public, the first firm's know-how is no longer necessary to the process. Any firm with the same technology or facilities in effect can appropriate the know-how and reduce their costs unilaterally.
These two examples demonstrate how diffusion is consistent with cost savings being merger-specific. However, these examples raise an additional possible criticism. If complete diffusion were to occur instantly, for example, then the economic incentive for the merged firm to invest in the innovation would disappear. This is a valid point. However, most diffusion is neither instantaneous nor complete. Incomplete or delayed diffusion is fully consistent with the incentive to innovate.
To carry out a dynamic analysis of merger efficiency benefits, the likely effect of the merger on costs, prices and quantities over time must be predicted and evaluated against the appropriate dynamic welfare standard. The analysis must focus on the likelihood and speed with which some or all of the initial cost reductions will spread to other competitors by technological diffusion or competitive innovation.
Dynamic welfare standards can be formulated by extending the static welfare standards to this dynamic environment. A dynamic version of the PCW standard, for example, would balance any consumer harms flowing from short run price increases with consumer benefits from price decreases in the longer run resulting from diffusion of the merger-induced cost reductions to other competitors. Application of an appropriate discount rate to future time periods ensures that greater weight is given to relatively more certain, short run effects.
In the Roberts-Salop paper, we apply this methodology to balance the potential competitive benefits and harms from mergers. We calculate the critical variable cost reductions necessary to satisfy a number of alternative dynamic welfare standards. In making these calculations, we consider a number of alternative specifications for the speed and degree of diffusion of the cost reductions to rivals.
We also make these calculations for a variety of pre-merger market conditions and market demand elasticities. As in the static welfare analysis, the critical cost reduction is higher for mergers in more concentrated markets and in markets with lower demand elasticities. These facts and the resulting calculations would be used in a rule of reason evaluation. The calculations determine the required cost reduction needed to balance out a likely competitive harm that would occur but for the cost reduction.
Table 1 presents illustrative results for the dynamic PCW standard for three different assumptions regarding the speed and likelihood of diffusion of the initial cost reduction: (a) Complete and instantaneous diffusion to all competitors; (b) Complete diffusion to all competitors over five years; and, (c) Partial diffusion (equal to half the initial cost reduction) over five years.
The figures are based on a unitary market demand elasticity. To take account of the general uncertainty regarding diffusion, we discounted future welfare at the annual rate of 25 percent. Table 1 also reports the corresponding figures for the static PCW and static AEW standards, in which diffusion is not taken into account.
Table 1 presents results for two different pre-merger market structures. One structure is a merger of two of the firms in an initial market of ten equal-sized firms. The other structure is a merger of two of the firms in an initial market of six equal-sized firms.
As a benchmark, we first consider the case of complete and instantaneous diffusion. In this case, the critical variable cost reduction needed to offset the potential anticompetitive harms from the merger is only 1.2 percent for the ten-firm market, rising to 4.0 percent for the six-firm market. These figures are much lower than the critical cost reductions for the static PCW standard in which diffusion was ignored. Those corresponding figures are 11.1 percent and 20.0 percent respectively.
Despite the sole focus on consumer welfare, these critical cost reductions are quite low. They suggest a more important potential role for efficiency analysis in mergers, even in concentrated markets, if diffusion rates are rapid and complete.
These figures also illustrate the dynamic interpretation of the static AEW standard discussed earlier. In fact, when diffusion is instantaneous and complete, the static AEW standard is not a bad proxy for a dynamic PCW standard that takes diffusion into account. That is, the static AEW standard requires variable cost reductions of 1.1 percent and 3.4 percent respectively for the ten-firm and six-firm markets. These are not far from the corresponding results for the dynamic PCW standard (equal to 1.2 and 4.0 percent respectively).
Of course, diffusion is seldom so rapid and complete as this. (And as discussed earlier, if it were, there would be little or no incentive by the merged firm to innovate.) When complete diffusion takes place over a five year period, the critical cost reductions for the dynamic PCW standard rise, to 1.7 percent and 5.4 percent respectively for the two market structures. Moreover, when we assume that only partial diffusion takes place over the five year period, in the sense that half of the competitors achieve the cost reduction achieved by the merged firm, the figures rise further, to 3.0 percent and 8.5 percent respectively. For the more concentrated market, the critical cost reduction becomes quite significant.
There clearly is less scope for efficiencies to be dispositive under the PCW standard when diffusion is both delayed and partial. However, the following four factors mitigate this general result. First, the required cost reductions are reduced substantially for larger market demand elasticities. For example, as reported in Table 2, the critical cost reduction for a market demand elasticity of 2 falls to only 3.9% for the six-firm pre-merger market when diffusion is partial and delayed. For an elasticity of 3, the cost reduction is only 2.5%. In contrast, for an elasticity equal to 0.5, the critical cost reduction for the six-firm market is 21.3%. These examples demonstrate just how important the demand elasticity is for efficiency analysis.
Second, it should be noted that these reported critical cost reductions solely involve variable costs. Fixed costs do not enter the PCW calculation at all. Although it may be difficult to reduce average total costs by 10 percent, eventually reducing variable costs by 10 percent may be more likely.
Third, the Roberts-Salop paper also analyzes an alternative welfare standard, that we refer to as a weighted social welfare (WSW) standard. The WSW standard is a more general standard which can be used to create a compromise standard between the PCW and the AEW. Such a WSW standard would be motivated only by consumer welfare but could recognize that even middle class consumers earn some of their income from stock ownership, either individually or through pension plans. For example, in 1989, taxpayers with adjusted gross incomes of $50,000 or less accounted for 53 percent of all income earned, 34 percent of dividend income and 19 percent of capital gains.
Table 3 reports the critical cost reductions for a dynamic WSW standard geared towards the interests of taxpayers earning $50,000 or less. The critical cost reductions are below those of the dynamic PCW standard. This is because the WSW takes into account the fact that the additional income flowing to this consumer group from price increases partially offsets the harm they suffer from those price increases. For example, for the six-firm market structure, with a unitary demand elasticity, the figures are 2.4%, 3.2% and 4.9% respectively for the cases of (a) complete and instantaneous diffusion, (b) complete and delayed diffusion, and (c) partial and delayed diffusion. These figures fall far below the corresponding figures for the dynamic PCW standard.
Fourth, this dynamic approach also suggests a new set of remedies short of divestiture to increase the diffusion rate. In certain matters, the government or court could require the merging firms to increase the speed and completeness of diffusion by licensing technology at below market rates or even royalty-free. Assuming that such a remedy would not render the merger unprofitable or destroy innovation incentives, it may permit the companies to merge even while having consumers share in the efficiency benefits.
IV. Implementing Dynamic Efficiencies Analysis
This dynamic framework thus represents a method for evaluating efficiencies that is consistent with the consumer orientation of the antitrust laws but nonetheless could make efficiencies dispositive in a larger number of cases. This approach would balance the benefits of cost reductions against the potential harms from mergers, taking into account the potential for diffusion of merger-specific cost reductions over time to competitors.
Our approach would imply a sliding scale standard for merger efficiencies, of course. Markets in which concentration and market shares are higher, market demand elasticity is lower, entry is more difficult and anticompetitive effects are more likely would require greater efficiency benefits to offset the higher likelihood of anticompetitive harm. Any short run price increases would be balanced against longer run consumer benefits from lower prices and enhanced competition.
To implement this dynamic efficiencies analysis, a number of additional issues must be examined. First, implementation may involve a requirement that efficiencies be merger-specific. Second, implementation involves a number of measurement issues. Third, implementation involves determination of the cognizability of different classes of cost savings.
A. Merger-Specific Efficiency Requirement
It is a general tenet of antitrust that efficiency claims are only cognizable to the extent that the allegedly anticompetitive conduct is reasonably necessary to achieve the claimed efficiency benefits. There is no reason why that same constraint should not be placed on merger efficiency claims. First, if the claimed efficiencies can be achieved unilaterally, then the merger is not reasonably necessary. We have already discussed this issue with respect to the criticism that diffusion might appear inconsistent with merger-specific efficiencies.
Second, the reasonable necessity standard also applies to cooperative conduct short of merger. This standard makes sense in general, but it is important not to take the test to extremes. Limited contracts may not provide the same degree of cooperative incentives and monitoring ability as full integration and may create opportunistic possibilities. Moreover, at some point, contracts can become so complex that they are either unmanageable or limit independence as much as a merger would. Thus, this standard should be interpreted sensibly.
B. Measurement Issues
One concern with case-by-case analysis of efficiency claims is the measurement of the magnitude of the cost reductions. These concerns are magnified in a dynamic efficiencies analysis in which the speed and likelihood of diffusion also are relevant. It might be argued that these economic factors are too difficult to measure reliably and to take into account in the typical HSR merger investigation.
I doubt that the factual determination of the cost savings is inherently more difficult than the determination of relevant market or ease of entry. Initial cost savings of the merging firms ought to be easier to evaluate because they do not implicate market dynamics, only the conduct of a single business entity. It is true that the merging parties control the initial information. For that reason, it would be appropriate to place the burden of persuasion on them. In addition, the government can use discovery from the parties and other firms to evaluate the validity of the efficiency claims. Finally, once efficiency claims begin to be introduced as a matter of course, the agencies and the bar will gain experience and expertise.
It seems a bit peculiar for the agencies to want to disregard evidence generated after the merger offer is made. For one thing, such a preference soon will lead the parties merely to delay their offers until the antitrust lawyers have evaluated the efficiencies studies. Moreover, antitrust analysis properly is focused on effect, not intent. A pre-merger efficiencies study may be indicative of the party's motivation for the merger, and motivation may indicate likely effect, but subsequent studies also are relevant to evaluating likely effect. They do not inherently lack credibility.
As discussed earlier, the process of diffusion of innovations is well established. Diffusion of innovations is not a matter of speculation in general. Moreover, dynamic analysis accounts for uncertainty over the likelihood of diffusion by discounting future effects at an appropriate discount rate.
However, I understand that it may be difficult to accurately predict the likelihood and speed of diffusion. Although this evaluation may be straightforward in some cases, it is certain to be quite problematic in others. A possible solution would be to create rebuttable presumptions of these factors. Then, in those industries or cases where credible case-specific evaluation of diffusion is possible, the merging parties would be permitted to offer additional proof.
I also am sympathetic to the practical difficulties involved in using a dynamic welfare standard. Estimating the present discounted values of consumer and producer surplus on a case-by-case basis could be a formidable task. At the same time, however, the real difficulties should not be overstated. Antitrust fact finders already must evaluate complex dynamic economic phenomena under the rule of reason, and the use of present discounted values is common in damages and valuation proceedings.
However, the merger evaluation and balancing process also can be simplified. Critical cost reduction standards can be formulated along the lines of those presented here and published. Analysts and courts can use these published standards directly instead of recalculating the underlying values themselves. After all, that is essentially the methodology of the Merger Guidelines. Fact finders do not explicitly calculate the probability that a merger will lead to the exercise of market power. Instead, they rely on the specific market definition, concentration and entry tests and then combine those results with a qualitative analysis of competitive effects.
In addition, I also think (but have not yet proved) that one can formulate static WSW standards that can serve as proxies for the underlying dynamic welfare standards, once the speed and completeness of diffusion is specified. For example, I discussed earlier the convergence of the critical cost reductions under the dynamic PCW standard with those under the static AEW standard in the case of instantaneous and complete diffusion. Thus, policy makers who want to think about their task in terms of a particular static balancing standard can use the proxy static standard instead of the more complicated underlying dynamic standard.
C. Cognizable Efficiency Benefits
Part of the controversy over efficiency analysis is the identification of the type of cost reductions that are cognizable. Cognizability of specific types of cost savings varies with the welfare standard adopted. I do not have the time to go into all the details here. To summarize, however, the proper treatment of certain cost savings, such as fixed cost reductions, input price decreases and tax savings, depends on the choice of welfare standard. In addition, the proper treatment sometimes may be counterintuitive or complex. When the Guidelines are revised to take detailed account of efficiencies, this part of the analysis will deserve close attention.
For example, tax savings are not real resource savings, but nonetheless would be cognizable under the PCW standard if they lead to lower prices. Fixed cost reductions may or may not be passed on to consumers, depending on the market growth and other factors. Merger-induced reductions in input prices sometimes do and sometimes do not involve real resource savings (rather than always being just transfers) that are cognizable under the static AEW. In fact, under certain circumstances, the real resource savings can exceed the cost savings achieved by the firm.
kV. Revitalizing Distressed Industries
Chairman Pitofsky proposes loosening merger enforcement in distressed industries to permit firms to rationalize excess capacity and reduce costs as a way to revitalize their competitiveness. Free markets do not always lead to efficient exit dynamics, so some exceptions may be justified. However, it is clear that antitrust should not return to Appalachian Coals. The question is how to facilitate revitalization rather than higher prices.
Chairman Pitofsky's proposals are carefully circumscribed. Some might say overly so. He apparently would limit the defense to "moderately concentrated" industries with excess capacity. He suggests that prices would not rise because of easy entry and substantial import competition. But, under these market conditions, the firms would not need to raise efficiency claims at all, let alone a special distressed industry defense.
The question is how to deal with the harder cases. Facilitating industry revitalization is a legitimate goal and the political alternative of protection from foreign competition may be worse than permitting questionable mergers. However, antitrust enforcers nevertheless must remain cautious and highly skeptical of lawyers who will try to turn every temporary recession into a golden opportunity for industry cartelization by merger.
As this testimony makes clear, I strongly favor adding case-by-case evaluation of efficiency benefits to the HSR merger evaluation process. Recognition of efficiency creation and diffusion as a dynamic process permits efficiency analysis to be placed on a strong foundation of consumer welfare. It also allows greater perspective on the so-called "killer qualification" -- proof that efficiencies will be rapidly passed on to consumers in the form of lower prices.
However, it will take some hard work to transform these ideas into practical guidelines. The best way of dealing with uncertainty over diffusion rates must be determined and the complexities of cognizability must be hammered into usable form. However, these are not insurmountable hurdles. The only real question is whether the we are prepared to make a commitment to improving the accuracy and efficiency of merger enforcement.
Summary of Critical Cost Reductions for Dynamic PCW Standard
Unitary Market Demand Elasticity
TABLE DID NOT SURVIVE HTML CONVERSION - SEE PAPER IMAGE
Summary of Critical Cost Reductions for Dynamic PCW Standard
Partial and Delayed Diffusion
TABLE DID NOT SURVIVE HTML CONVERSION - SEE PAPER IMAGE
Summary of Critical Cost Reductions for Dynamic WSW Standard
TABLE DID NOT SURVIVE HTML CONVERSION - SEE PAPER IMAGE
 Professor of Economics and Law, Georgetown University Law Center, and Special Consultant, Charles River Associates. I would like to thank Gary Roberts, Jonathan Baker, Tom Overstreet and John Woodbury for helpful comments.
 Broadcast Music, Inc. v. Columbia Broadcasting System, Inc., 441 U.S. 1 (1979). Similarly, in per se tying cases, efficiency benefits are taken into account through the single product defense. Jefferson Parish Hospital District No. 2 v. Hyde, 466 U.S. 104 (1984). They also would arise in the emerging "business justification" rationale for rule of reason treatment. Mozart Company v. Mercedes-Benz of North America, Inc., 833 F. 2d 1342 (1987).
 Chairman Pitofsky also adds two additional reasons, measurement difficulties and the potential for achieving efficiencies by conduct short of merger. Pitofsky, supra note 5 at 209-210. I think that if the issue of standard can be resolved, these other two will seem less formidable.
 Bork, THE ANTITRUST PARADOX (1978). See also Muris, The Efficiency Defense Under Section 7 of the Clayton Act, 30 CASE WESTERN L. R. 381 (1980). However, the AEW standard is inconsistent with indifference to harm to competitors often expressed by courts and commentators including Bork. Under the AEW standard, the profits of (say) excluded competitors would count equally with deadweight losses or cost saving benefits in evaluating welfare.
 These results are based on the assumption of a unitary market demand elasticity. As discussed in more detail below, the critical cost savings fall significantly for larger demand elasticities. Thus, along with pre-merger market structure, the market demand elasticity is a key factor determining the magnitude of the critical cost reduction. Note that these figures are intended to illustrate the economic concepts rather than to serve as definitive estimates.
 For example, Mansfield found that information about new innovations typically is known to some rivals within twelve to eighteen months of the date of the initial decision by the innovator to develop the new product or process. Mansfield, How Rapidly Does New Technology Leak Out, 34 J. INDUSTRIAL ECON. 217 (1985). See also, Levin et. al., Appropriating the Returns from Industrial Research and Development, BROOKINGS PAPERS ON ECONOMIC ACTIVITY 783 (1987). Porter explains how close geographic proximity can increase diffusion. See Porter, THE COMPETITIVE ADVANTAGE OF NATIONS (1990)
 This raises the question of why one firm can not increase its scale unilaterally by expanding its own production. However, internal expansion may not be a good substitute when time and additional sunk cost investments would be needed to expand, perhaps because products or distribution networks are differentiated. Internal expansion also may be deterred or reduced by the expectation of market price decreases that would reduce profitability.
 Dynamic analysis also should account for the potential for rent-seeking. Any cost increases that arise from rent-seeking would raise the critical cost reductions required to justify the merger in light of the market structure.
 The rule of reason sometimes is formulated as determining the likelihood of competitive harm that would be necessary to exceed a given level of likely efficiency benefits. Our formulation flips the question around. We determine the level of efficiency benefits necessary to exceed a given likelihood of competitive harm.
 For this, we assume one fifth of the competitors achieve the cost reductions each year over the five year period, beginning at the same time that the merging parties begin to reduce their costs. We use a discount rate of 25% and assume a unitary demand elasticity.
 Under the Merger Guidelines with a "ssnip" test of 5%, a relevant market may have any market demand elasticity of less than 20, depending on the pre-merger price-cost margin. See Harris and Simons, Focusing Market Definition: How Much Substitution is Necessary?, 12 RESEARCH IN LAW AND ECONOMICS 207 (1989).
 Internal Revenue Service, Statistics of Income - 1989: Individual Income Tax Returns, Table 14. These figures apparently ignore employer-based pension plans that do not show up on the employees' tax returns. Because such plans affect taxpayers of all incomes, we do not know how including them would affect the percentages we report.
 Similarly, reasonable necessity is not gauged relative to other potential mergers (for example, with a smaller competitor). Efficiencies analysis is not like the failing firm doctrine. The Merger Guidelines gauge the likelihood of market power relative to the pre-merger situation, not relative to some other hypothetical acquirer of the acquired firm. The same type of benchmark should apply to efficiencies analysis.
 There currently seems to be a chicken-and-egg problem at work here. The agencies view evaluation of efficiency claims as daunting because they lack experience. But, merging parties seldom offer such evidence because the agencies imply that they will not consider it seriously.
 I am sympathetic to proposals by Chairman Pitofsky and others for a "trial period" that would induce firms to make only credible claims by penalizing those whose claimed efficiencies evaporate in the aftermath of reality. See Pitofsky, supra note 5; Brodley, supra note 4; Scherer, Comment in BROOKINGS PAPERS ON ECON. ACTIVITY (Baily and Winston eds.,1991). The penalty might be divestiture, a requirement that prices be reduced below pre-merger levels or a fine equal to the shortfall in claimed cost reductions. In contrast, if the penalty were just a modest fine, that fine simply would permit the merging entity to purchase market power from the government.
 The fact that diffusion may occur some time after the merger does not render the likelihood of diffusion speculative. For example, the certainty of one's death is not speculative, despite uncertainty over the date. Market definition and merger analysis do not focus on a 1-2 year period because a longer period would be speculative. Instead, this period is the focus because short run substitution is said to be necessary to deter anticompetitive mergers and price increases. Of course, analysis of innovation markets also has lengthened the period of focus. Finally, the estimated market power effects of many mergers also often involve a likelihood, not a 100% certainty.
 Alternatively, certain markets or sectors might be classified as high or low diffusion on the basis of economic studies. Antitrust enforcers and courts then could rely on that classification in specific cases.
 Chairman Pitofsky's reinvestment commitment over a trial period clearly is a good idea. Moreover, it obviously makes more sense to have antitrust enforcers overseeing the revitalization instead of just the Commerce Department.