on Merger Enforcement
Panel on Efficiencies/Dynamic Analysis/
Wilmer Cutler Pickering LLP
February 19, 2004
FTC/DOJ Joint Workshop on Merger Enforcement
Panel on Efficiencies/Dynamic Analysis/Integrated Analysis
February 19, 2004
It has been seven years since the efficiencies section of the Horizontal Merger Guidelines was last revised in 1997. In these seven years, agencies have conducted full-scale Second Request investigations of several hundred mergers. Efficiencies have become an increasingly important focus of these investigations. Now is a good time to review what we have learned over these seven years and to discuss whether further revision of the efficiencies section of the Merger Guidelines is now in order.
I believe that, by and large, the analytical framework set out in 1997 revisions is sound. The problem is not with the framework, but that too many practitioners have not made a serious effort to understand and apply that framework. As a result, the agencies see too few well-presented and well-substantiated efficiency claims.
I want to focus today on three important issues: (1) What is a cognizable efficiency? (2) To what extent should "fixed cost" savings that do not reduce marginal costs be taken into account? (3) Under what circumstances should efficiencies in other markets be found to justify a merger that reduces competition in a particular market?
In discussing these three issues, I do not intend to offer fully formed views, much less recommendations for action. Rather, I hope to raise some questions designed to stimulate further discussion and analysis.
What is a cognizable efficiency?
Lawyers and businessmen too often confuse efficiencies with cost savings. While in some sense every efficiency can be translated into cost savings, in practice many of the most important efficiencies are the result of synergies from the combination of complementary assets rather than cost savings as such.
Take, for example, a merger of two software companies, both of which produce spreadsheet and word processing programs. Assume that Company A has a superior spreadsheet program and Company B a superior word processing program. The two companies compete also in the market for office suites that combine word processing and spreadsheets. By merging, the two companies expect to be able to offer an Office Suite program with the best available spreadsheet and word processing program, and to be able to use their complementary knowledge to create even better spreadsheet and word processing software in the future.
Now suppose further there is only one other company, C, that produces both word processing and spreadsheet programs and offers an office suite. At present that company's combined office suite program is better than either A's or B's and it, therefore, has a larger market share. After the merger, the combined AB's program will be markedly superior and it is anticipated that the combined firm will take over market leadership from C.
Should we prohibit this merger? A conventional answer might well be yes. The merger I've posited is a 3-to-2 merger, and one can imagine the staff telling the usual coordinated effects story or even a unilateral effects story. But is that story credible in these circumstances? Given that competition in this market is driven less by price and more by innovation, especially since without innovation, users would have no incentive to replace their existing office suite software with a newer version, how likely is it that this merger will, in fact, reduce competition. Do we really believe that C will sit still and accept losing its market leading position. Won't the merger stimulate C to improve its own product in order to recapture the lead?
Ah, but, you say, A and B could combine their superior products by contract, rather than by merger. This argument raises a series of further questions. Why would that necessarily leave the market any more competitive? And how does each firm assure that its contribution to the success of the joint product is fairly valued. What if firm A makes a breakthrough in spreadsheet technology that causes the joint product's share to increase dramatically? Will B accept a reallocation of relative shares of profits from the joint product? These are all issues of transactions cost economics that antitrust lawyers too often overlook.
Now change the facts. Suppose A and B are already the two leading firms and C is in third place. Would you now reach a different result and, if so, why? The merger will still leave users better off, at least in the short run. But now the fear would be that C will drop even further behind and will no longer be able to constrain the conduct of the combined AB, which will now be able to behave as a monopolist. Well, yes, but so what? Won't AB have to continue to strive to improve its product in order to give users an incentive to buy new versions? And won't C have an incentive to do whatever it needs to in order to get back in the ball game, perhaps by merging with the company with the best email/calendar program.
I'm not suggesting that on these facts the agencies should necessarily clear a merger to monopoly or near monopoly, but only to suggest that the answer is not as self-evident as we might have thought in the past. In these circumstances, the agencies need to look carefully at whether and, if so, by how much a hypothetical monopolist would be able to raise quality-adjusted prices and whether, even with that price increase, the savings to society from the enhanced quality of the products resulting from the merger will outweigh the harmful effects of that price increase.
The other point I want to leave you with is that the story I've just told has nothing to do with cost savings as such, except in a very attenuated way. It is true, I suppose, that cost plays a role, because at some point A and B could invest enough to develop superior word processing and spreadsheet programs on their own. And you need to know what that cost is in order to determine whether they are likely to be willing to invest as much as would be necessary. But note that I'm now talking about R&D and investment costs, not variable costs, which leads us naturally into our next issue.
How much weight should be given to fixed cost savings?
The 1997 revisions emphasize that in determining whether cognizable efficiencies are of a character and magnitude such that the merger is not likely to be anticompetitive in any relevant market, the agencies will consider whether the cognizable efficiencies "likely would be sufficient to reverse the merger's potential to harm consumers in the relevant market, e.g., by preventing price increases in that market." The guidelines add that, for this reason, efficiencies that "enable the merging firms to reduce the marginal cost of production," are likely to receive the greatest weight. The guidelines acknowledge, however, in footnote 37, that the agencies "will also consider the effects of cognizable efficiencies with no short-term, direct effect on prices in the relevant market."
If there is anything in the 1997 revisions requiring further clarification, it is probably this series of statements. The paradigm on which these statements are based is plainly a smokestack industry in which price is driven primarily by marginal cost. While a perfectly sound static economic model, that model has little to do with competition in many sectors of our economy where price and competitive behavior is driven far more by innovation and by recurring R&D costs than by production cost. In these markets a merger that enables a firm, in the very short-term, to raise prices relative to marginal cost may actually enhance competition by enhancing its ability to fund the R&D necessary to innovation.
The most instructive treatment of this subject I've seen to date is in William Baumol's book "The Free Market Innovation Machine." In his book, Baumol extends the argument first posited by Joseph Schumpeter showing that in many markets, especially those driven by innovation, there is a tension between static allocative efficiency and productive efficiency. I've developed this point at some length in an article that is forthcoming in Antitrust Bulletin, entitled What is Competition? A Comparison of U.S. and European Perspectives, a copy of which is available outside.
As I explain in that article, in markets driven by innovation there is no necessary correlation between competition and concentration or even between price-cost margins and competition. The most intensely competitive markets are often ones with only one or two competitors in which price-cost margins are high because of the need to support recurring R&D expenditures. Without high price-cost margins, there would be no way to recoup risky investments in R&D and no incentive to innovate. Studies show, not surprisingly, that price-cost margins are highest in R&D-intensive industries. They also show a U-shaped relationship between price and concentration, suggesting that increases in concentration are often associated, up to a point at least, with increases in productive efficiency that are passed on to consumers in the form of lower prices.
As Baumol explains, competition in markets driven by innovation resembles the "Red Queen Game" in Alice in Wonderland, in which a firm needs to run as fast as it can just to stand still. According to Baumol, the competitive model that should be used to analyze competition in these markets is the contestability model that Bobby Willig and he developed in the late 1970's and early 1980s. Few markets are perfectly contestable, just as few markets are perfectly competitive. Nevertheless, the contestability model helps us better understand how these markets perform. In markets driven by innovation, firms must incur substantial and recurring sunk costs in the form of R&D expenditures in order to compete. In these markets, prices will be substantially above marginal cost, but that is not a sign of market power because taking into account the riskiness of the investments in R&D, firms in these markets often earn only a normal return on capital. Moreover, these firms are price takers, in that they can set prices no higher than the level the would induce entry by others, and therefore do not have "market power" in the antitrust sense.
In these markets, as in other contestable markets, it is the threat of entry that constrains the incumbent's pricing. Entry will occur if the incumbent sets prices above the level necessary to recover its sunk or common costs. Thus, potential entry will drive prices to the levels that just permit a competitive return on those sunk costs, but will not depress them to marginal cost. These markets also tend to be characterized by discriminatory pricing because price discrimination is the optimal strategy for allocating common costs among customers depending on their price sensitivity. In these circumstances, price discrimination is welfare enhancing because it helps provide the return needed to support the necessary investment.
In these markets, in particular, efficiency analysis needs to examine, not what the impact of the merger is on marginal production costs, but rather on its impact on the merged firm's ability and incentive to make the recurring R&D investments necessary for effective competition through innovation. Efficiencies that serve to reduce recurring common costs, even though those costs would be viewed as "fixed" in the static price theory model because they do not vary with output, may over the onger term benefit consumers through lower quality-adjusted prices because they will reduce the prices a firm will need to charge to recoup those sunk costs.
To what extent can a lessening of competition in one market be justified by efficiencies in other markets?
One of the most striking shifts in merger enforcement over the seven years since the 1997 revisions were adopted has been the increasingly rigorous application of the "hypothetical monopolist" or "SSNIP" test for market definition, especially in markets characterized by price discrimination, as most markets now are. As a result, antitrust markets are being defined more narrowly than ever. A good recent example is the First Data/Concord merger where the Antitrust Division alleged a market for "processing PIN-debit cards at point-of-sale." Indeed, in markets characterized by price discrimination, it is not uncommon for the agencies, in evaluating a merger, to view each customer as a separate market, even if that is not how they define the market when they file a complaint.
As markets are defined more and more narrowly, it becomes increasingly important for the agencies to take into account efficiencies in other, closely related markets that would be lost if a merger is blocked or a divestiture ordered.
The paradigmatic example is a merger that Andrew Dick and I describe in our Antitrust Law Journal article on the role of efficiencies in merger review. That merger involved the combination of two natural gas gathering and processing systems in West Texas, near President Bush's hometown of Midland. These systems were located in an area with very mature natural gas fields that were experienced sharply declining production. As a result, both systems were operating well below capacity and were experienced sharply rising per unit operating costs, with no sign of any turn around. There were only a small number of wells that were close enough to both systems to be served by either, but as to these this was a merger-to-monopoly because no other systems were close enough to service those wells. At the same time, both those wells, and the much larger number of wells that were captive to one system or the other, would all benefit from the substantially reduced processing costs resulting from the merger. And there was no divestiture that could have fixed the competitive problem without sacrificing those efficiencies. In these circumstances, the FTC wisely granted early termination of the Second Request waiting period once the staff understood that the benefits from the efficiencies overall outweighed the trivial loss from reduced competition for a few wells.
Another more controversial example was the Union Pacific/Southern Pacific merger, where the Surface Transportation Board recognized that the merger would eliminate competition in a large number of 2-to-1 routes out of the Gulf Coast region, but concluded the merger would also generate substantial efficiencies across the entire networks of the two merging railroads. In this case, the STB sought to preserve the efficiencies, while addressing the competitive concerns, by using trackage rights, rather than divestiture, to remedy the anticompetitive effects. This is a good illustration also of how efficiencies can and should be taken into account in structuring relief in a merger case.
This is an area in which there is a marked difference in approach between the United States and the EU. The US guidelines provide that the agencies will, as a matter of prosecutorial discretion, take into account out-of-market efficiencies that are "inextricably intertwined" with the anticompetitive effects in deciding whether to challenge a merger. The new EU Notice on Horizontal Mergers provides the European Commission no similar flexibility. I would be very interested in having Vincent Verouden comment on how the Commission would address this issue in cases like the two I've just described.