Statement of Chairman
Robert Pitofsky and
File No. 971-0033
The consent agreement negotiated in this matter, which the Commission has today accepted and placed on the public record for comment, eases competitive concerns raised by Cadence Design Systems, Inc.'s ("Cadence") acquisition of Cooper & Chyan Technology, Inc. ("CCT").
The Commission's complaint alleges that Cadence is the dominant supplier of complete software "layout environments" for the physical design of integrated circuits, or "chips," the postage-stamp sized electronic components used in devices as diverse as personal computers and kitchen appliances. CCT sells a software tool, called a "router," that works within a layout environment and allows users to plot the connections among the millions of components within an integrated circuit. The proposed complaint alleges that CCT is the only firm to have developed a "constraint-driven, shape-based" router, state-of-the-art technology that is expected to solve the next generation of problems that will face integrated circuit producers designing ever more powerful chips.
The Commission's proposed complaint alleges a well-established vertical theory of competitive harm, laid out in the 1984 Merger Guidelines.(1) The Guidelines explain that a vertical merger can produce horizontal anticompetitive effects by making competitive entry less
likely if (1) as a result of the merger, there is a need for simultaneous entry into two or more markets and (2) such simultaneous entry would make entry into the single market less likely to occur.(2) While the dissenting Commissioners may take issue with the "dual-level entry" theory of vertical mergers that the 1984 Guidelines articulate, the available evidence suggests that the Cadence/CCT merger, which combines Cadence's dominant position in integrated circuit layout environments with CCT's current monopolistic position in constraint-driven, shape-based integrated circuit routers, presents a straightforward case of anticompetitive effects caused by vertical integration. We believe that this type of competitive harm merits our attention.(3)
When considering the effects of mergers in dynamic, innovative high-tech markets, such as those present here, it is particularly important to investigate whether such mergers will create barriers to entry. New entrants often bring innovation to the market, and the threat of entry leads incumbents to innovate. Therefore, we must be vigilant to preserve opportunities for entry.
As the Analysis to Aid Public Comment explains, unless a would-be supplier of routing tools had the ability to develop an interface to the Cadence integrated circuit layout environment, it would not be able to market its routing product effectively to the vast majority of potential customers which use the Cadence layout environment.(4) Without an expectation that it could design software compatible with Cadence's installed base, a would-be entrant might well decide not to compete.(5)
After the proposed Cadence/CCT merger, Cadence would have an incentive to impede attempts by companies developing routing technology competitive with CCT's constraint-driven, shape-based router technology, IC Craftsman, to gain access to the Cadence integrated circuit layout environment. Following the proposed merger, successful entry into the routing tool market is more likely to require simultaneous entry into the market for integrated circuit layout environments. Without a consent that mandates access to Cadence's layout environment, and thus lowers the barriers to entry in the market, a combined Cadence/CCT will face less competitive pressure to innovate or to price aggressively. Thus, competition would likely be reduced as a result of the proposed acquisition.
The proposed remedy in this matter preserves opportunities for new entrants with integrated circuit routers competitive with IC Craftsman by allowing them to interface with Cadence's layout environments on the same terms as developers of complementary design tools.(6) Specifically, the proposed order would require Cadence to allow independent commercial router developers to build interfaces between their design tools and the Cadence layout environment through Cadence's "Connections Program." The Connections Program is in place now and has more than one hundred participants who have all entered a standard form contract with Cadence.
The separate statements by Commissioners Azcuenaga and Starek question this enforcement action. We respectfully disagree.
First, Commissioner Azcuenaga argues that the Commission should have brought an action based upon a horizontal theory of competitive harm. We certainly agree that horizontal competitive concerns deserve our close attention and recognize that horizontal remedies often cure vertical problems. If we had credible support for the theory that the proposed merger would combine actual or potential horizontal competitors and would substantially lessen competition in an integrated circuit routing market or an innovation market for integrated circuit routers, we would not hesitate to advance that case. But after a thorough investigation by Commission staff, we have not found sufficient evidence to conclude that, absent the acquisition, Cadence would have been able to enter the market for constraint-driven, shape-based integrated circuit routers successfully in the foreseeable future.
The dissenting statements fail to give full weight to all the incentives at work in the vertical case. It is true that Cadence would be motivated by the entry of new, promising routing technology to allow an interface to its layout environment to sell more of its complementary products. And absent the merger, that would be its only incentive. But with the merger, Cadence clearly also has an incentive to prevent loss of sales in its competing products. And while these two incentives may compete as a theoretical matter, the evidence in this case indicates that Cadence has acted historically according to the latter incentive. There is some reason to believe that Cadence in the past has thwarted attempts by firms offering potentially competitive technology to develop interfaces to its layout environment (including at one point, CCT). Now that it has a satisfactory router to offer its customers, there is no reason to think that absent the consent, Cadence would treat developers of routers that would compete with IC Craftsman any differently than it once treated CCT.
Commissioner Azcuenaga also suggests that the consent order is unnecessary because a company developing a router to compete with IC Craftsman could proceed, as CCT did, without an interface to Cadence's design layout environment. The evidence shows, however, that CCT's management thought that ensuring compatibility with Cadence's layout environment was critical and that marketing without that compatibility, which it had done, was not sufficient.(7) It took the extreme measure of inducing a third party to write software for CCT to interface IC Craftsman with the Cadence layout environment without Cadence's knowledge. Moreover, despite CCT's success in developing a routing program, its sales were modest before the merger announcement.(8)
Commissioner Azcuenaga is further concerned that mandating access to the Connections Program for developers of routing software on terms as favorable as for other Connections participants might have unintended consequences. In particular, she is concerned that the order may prompt Cadence to charge higher prices to all Connections partners. But the Connections Program is an existing program with over one hundred members, and Cadence would have significant logistical difficulties, and would risk injuring its reputation, if it suddenly altered the terms of the program. Also, Cadence has good reasons for having so many Connections partners--they offer Cadence customers valuable tools, most of which do not compete with Cadence products. It seems unlikely that Cadence would be motivated to make the Connections Program less appealing to those partners.
Both Commissioners Azcuenaga and Starek suggest that the proposed remedy may be difficult to enforce. Any time this Commission enters an order, it takes upon itself the burden of enforcing the order, which requires use of our scarce resources. However, we think the proposed order, which simply requires Cadence to allow competitors and potential competitors developing routing technology to participate in independent software interface programs on terms no less favorable than the terms applicable to any other participants in such programs, is a workable approach.(9) Connections partners all sign the same standard-form contract and there has been a consistent pattern of conduct with respect to the program to use as a baseline for future comparisons. Moreover, the Commission has had experience with such non-discrimination provisions, and can rely on respondent's compliance reports required under the order as well as complaints from independent software developers to ensure compliance with the consent. We think the dissenting Commissioners' scenarios about intractable compliance issues are unfounded.
In sum, we believe that the consent order will preserve competition in the market for cutting-edge router technology by reducing barriers to entry.
1. See U.S. Department of Justice Merger Guidelines, 4 Trade Reg. Rep. (CCH) ¶ 13,103 (June 14, 1984) (hereinafter "1984 Merger Guidelines"). When the agencies issued the 1992 Horizontal Merger Guidelines, U.S. Department of Justice and Federal Trade Commission Horizontal Merger Guidelines, 4 Trade Reg. Rep. (CCH) ¶ 13,104 (April 7, 1992), they explained that "[s]pecific guidance on non-horizontal mergers is provided in . . . [the] 1984 Merger Guidelines." U.S. Department of Justice and Federal Trade Commission Statement Accompanying Release of Revised Merger Guidelines, 4 Trade Reg. Rep. (CCH) ¶ 13,104 (April 2, 1992). See generally Herbert Hovenkamp, Federal Antitrust Policy §§ 9.4, 9.5 (1994) (suggesting that vertical mergers may create barriers to entry when one of the parties is a monopolist or near-monopolist).
2. See 1984 Merger Guidelines § 4.21.
3. Contrary to Commissioner Starek's assertions that enforcement action here, in the context of a merger, leads logically to enforcement action against internal vertical expansion, see Dissenting Statement of Commissioner Roscoe B. Starek III at n.8 & accompanying text, such unilateral action has been known to present a completely different set of questions under the antitrust laws for more than one hundred years.
4. Not only is Cadence the dominant layout environment, but its competitors are in a state disarray. For example, Cadence's most significant competitor, Avant! Corporation, and several of its top executives have recently been charged with theft of trade secrets from Cadence.
5. CCT decided that it was so important to gain access to Cadence's layout environment that when Cadence refused to allow the IC Craftsman product (CCT's constraint-driven, shape-based router technology) to interface with the Cadence layout program through the "Connections" Program, CCT induced a third party that was a Connections partner to write an interface to the Connections Program for IC Craftsman without Cadence's knowledge. Cadence thereafter sought to impede CCT's attempts to gain access to the Cadence integrated circuit layout environment by suing CCT.
6. At the same time, the proposed order preserves any efficiencies of vertical integration resulting from the proposed merger, which may benefit customers.
7. Interfacing with another firm's design layout environment is also not a feasible alternative because of Cadence's dominant position in the market. Without hope of marketing to the vast majority of customers, developers of an alternative router have minimal incentives to compete. In addition, the competitive significance of Cadence's few competitors is questionable.
8. Products offering incremental innovation rather than the revolutionary breakthrough of IC Craftsman would have an even more difficult time entering.
9. The language of the consent is clear in requiring that terms for routing companies be no less favorable than for any other participant in the Connections Program. Thus, we do not understand Commissioner Starek's conclusion that the consent could be interpreted to require routing companies to pay a "fee no higher than the highest fee." And as his own dissent acknowledges, if the order could be interpreted to allow Cadence to terminate router developers from the Connections Program after thirty days, the proposed order would be meaningless.