JOEL I. KLEIN
San Antonio Marriott Rivercenter
May 2, 1997
I am happy to be here today -- or, should I say, as happy to be here as any head of the Antitrust Division could ever be facing down a horde of intellectual property lawyers at the Alamo. Seriously, I'm flattered to have been invited, and grateful for the chance to speak with you at the close of your Spring Meeting. I view it as a sign of how far we at the Antitrust Division have come in the last few years that I can come before you with the sense that we share a lot of common ground. I especially want to mention how gratified I am by the cooperative and productive relationships we have developed with our colleagues at the Patent & Trademark Office and the Copyright Office, and want to thank Commissioner Lehman, Register Peters and Solicitor Linck. The open lines of communication among our offices have helped us understand and take into account each other's concerns, and I know, for our part, that our own work as competition advocates has benefited greatly as a result.
I. Antitrust Enforcement and Intellectual Property
Let me make clear at the outset, the intersection of antitrust law and intellectual property has become a major agenda item for the Antitrust Division. It figures prominently in the work of three of the Division's sections. Moreover, it raises such important and difficult questions, frequently involving consultations with other parts of the Executive Branch, that my predecessor Anne Bingaman appointed a senior counsel specifically for intellectual property and antitrust. This high priority makes sense to me, too. In a world where so many important products and services are driven by technology, it is becoming increasingly apparent that competition among technologies, and in the development of those technologies, is often as important to consumer welfare as the price competition that has been the traditional focus of antitrust attention. This point was repeatedly emphasized in the fine hearings that the Federal Trade Commission recently held concerning the effects of globalization on the U.S. economy.
And, I should also add, the competitive significance of intellectual property rights is not limited to patents: on the contrary, our most recent public activity has involved material protected by copyright. Most recently, in an extensive investigation of competitive conditions in the market for legal research materials, we saw the competitive barrier that West Publishing Company was imposing through its assertion that other case law publishers' use of star pagination, showing the fact of where page breaks appeared in West's own case reports, infringed its copyright. Consequently, we filed amicus briefs in two declaratory-judgment actions brought by other publishers, contending that star pagination does not infringe West's copyright.(1) And, in another matter, only last August, we sued General Electric over restraints it imposed on over 500 hospitals and clinics that had licensed software for use in maintaining GE medical imaging equipment -- restraints that, we contend, amount to naked agreements keeping the licensees from competing with it in servicing and maintaining medical equipment of any kind. The effect of the agreements is to limit competition in providing service for medical equipment, raising the cost of that service for hospitals and clinics, and ultimately harming competition and raising prices in medical equipment markets too -- all leading to higher health-care costs. We are also grappling with a case we filed a little earlier -- in 1941, to be exact. Our antitrust consent decrees governing the licensing and royalty-distribution functions of both ASCAP and BMI are getting serious study and analysis, as we consider whether the licensing model the decrees employ is the right one in a world that could barely have been imagined when we entered into the BMI decree in 1966, let alone in 1941, when the original ASCAP decree took effect.
More generally, a great deal of attention given to our work involving intellectual property has focused on the drafting, issuance and application of the Antitrust Guidelines for the Licensing of Intellectual Property, which the Antitrust Division and the Federal Trade Commission released a little over two years ago. As the Guidelines indicate, our public focus has been primarily on fairly straightforward licenses, in which protected technology or creations move in one direction -- from the licensor to the licensee.(2) We've said little about our analysis of licensing transactions that combine the intellectual property of different owners -- i.e., cross-licensing -- or patent pools, if you will.(3) And we've said even less about the subset of licensing arrangements that bundle intellectual property rights together and present them to licensees in a mandatory package. But, as some of you may know, we've spent a considerable amount of time looking at both of these complex subjects. And, while I can't pretend to tell you that we've got all the answers now, a few things have occurred to us along the way. In the future, I'll have something to say about the bundling issue. Today, I want to focus principally on the competitive issues posed by cross-licensing and pooling agreements. They concern both substantive issues that are likely to get our attention when we review combinations of intellectual property rights and some procedural ideas that should help us do a better job of identifying those combinations that are likely to be anticompetitive, while not standing in the way of the rest of them.
To put matters in perspective, I should make clear at the outset that I would expect that by far most cross-licenses and pools are, on balance, procompetitive. That means that, at bottom, they help sellers provide consumers with better products and services at lower prices because of benefits ranging from cost savings -- due to more efficient production technologies -- to improved product quality -- resulting from combining complementary inventions.(4) At the same time, cross-licenses can have anticompetitive effects, too, including increased prices, cutbacks in production, and reduced innovation. This is particularly true when the cross-license is between firms that are competitors, whether in producing goods or services, licensing intellectual property, or in R&D. In that case, our antitrust antennae go out: we have to be alert to the possibility that the cross-license can serve the interests of the parties, at the expense of competition and consumers. Based on our experience, we think this possibility is probably greatest in the context of settling infringement litigation. The stakes are high, particularly if the dispute involves a market with a small number of competitors to begin with or a particularly broad or fundamental intellectual property claim. The defendant may be facing the possibility of continuing in business only at the sufferance of the plaintiff; for the plaintiff, the litigation may determine its ability to decide who it will or won't allow to compete. Consequently, settlements are often based on considerations that lead parties to give up rights that they might well vindicate if they went to the mat. And when intellectual property rights are at stake, the consequences of those compromises can align the settlers' interests against the interests of consumers.
Essentially, we analyze these cross-licensing or pooling issues in the same way as we look at all licenses under Section 1 of the Sherman Antitrust Act. We ask, first, does something about the license hurt competition that either already existed or likely would have come into being without it? And if the license does have such an anticompetitive effect, we then ask, is that harm reasonably necessary in order to bring about some even greater procompetitive benefit? These two basic questions -- what are the anticompetitive effects, what are the procompetitive benefits -- constitute what's known in antitrust law as the Rule of Reason.(5) The application of such a rule-or-reason analysis to cross-licenses is pretty straightforward. Obviously, our principal concern is whether the patents or their owners are using the arrangement to blunt competition that would otherwise take place -- a rather-switch-than-fight strategy, if you will. And so when we look at one of these arrangements, we generally analyze the following particulars, which have been common to our competitive analysis from the very beginning. They include the relationship of the intellectual property rights being combined; the nature of the markets in which those rights, and the goods or services in which they're used, compete; the extent to which the pool controls access to those rights; the openness of the pool to outsiders; and the extent to which the cross-license controls the terms on which future innovations in the field will reach the market. Fortunately, I can give some color to this rather arid description of what we do by taking an historical example -- involving airplanes and World War I -- that is remarkable for its endurance. If nothing else, when one relies on some venerable precedent to highlight contemporary concerns, it always has the sound of erudition.
II. The Manufacturers Aircraft Association
By the time of America's entry into World War I in April 1917, our Government had become an eager consumer of airplanes. Airplanes were still so new that the Wright brothers' basic patent, then in the hands of the Wright-Martin Aircraft Corp., still blocked would-be manufacturers, at least for all practical purposes.(6) So too, apparently, did "numerous important patents" in the hands of the Curtiss Aeroplane & Motor Corp.(7) The two firms were demanding royalties from other aircraft manufacturers, although the Attorney General singled out Wright-Martin for seeking "high" royalties.(8) The upshot for the Government was not only "excessive" airplane costs but an inability even to obtain enough planes, since the cost of licensing constrained industry capacity. In the face of this problem, an obvious and time-honored Washington solution presented itself: an advisory committee. This one, though, had a better-than-average pedigree, having been convened by Assistant Secretary of the Navy Franklin D. Roosevelt.(9) After a series of consultations with the truculent patentees,(10) the committee recommended the formation of a patent pool, which was in place by the end of July 1917. The aircraft pool, which encompassed "practically all" airplane manufacturers,(11) resolved all pending infringement claims and bound the members to give each other nonexclusive licenses to "all airplane patents of the United States (with unimportant exceptions) now or hereafter owned or controlled by them."(12) Membership was open to three types of entities: (1) any "responsible" present or potential airplane manufacturer; (2) any manufacturer to which the federal government had awarded a contract for ten or more planes;(13) and (3) any owner of U.S. aircraft patents.(14) The pool's charter contemplated only 100 members, but that was apparently more than adequate to accommodate the potential membership.(15) Members promised not to put relevant patents out of the pool's reach, as, for instance, by taking an exclusive but non-sublicensable license.(16) In consideration for access to the basic patents, manufacturing members would pay a flat $200-per-aircraft royalty, from which 67.5% went to Wright-Martin, 20% to Curtiss, and the remainder to the pool entity for administrative costs.(17) Wright-Martin and Curtiss were entitled to these shares, up to a maximum of $2 million each, until their patents expired.(18) An arbitration panel would decide what other pooled patents merited royalties and would set those royalties.(19) Unlike the flat $200 royalty,(20) the fee for other royalty-bearing patents was predicated on actual use; the rest of the patents were available to the pool members on a royalty-free basis.(21)
In September 1917, the Secretary of War passed along to the Attorney General a request for an antitrust advisory opinion.(22) The Attorney General responded in less than three weeks, concluding that there were no anticompetitive effects that came close to outweighing the very real procompetitive benefits resulting from assembling these patents into an affordable package available to all comers.(23) In reaching that conclusion, he touched on one after another of the issues that we would raise today as well. Let me go through them with you. An unspoken but important premise of the Attorney General's analysis is that the patents were valid. Of course, restraints imposed in conjunction with licenses of invalid or illusory intellectual property rights would be highly anticompetitive.(24) Moreover, as the possessor of its own blocking, and presumably valid patents, Curtiss had more to gain than anyone else from challenging the validity of Wright's patent. If there were any significant doubt about validity, the cross-license between the firms could entail a real loss to the market, since Curtiss would no longer have that impetus to eliminate a principal entry barrier.(25) The Attorney General's stated analysis began with the patents' relationship to the airplane industry and each other. In particular, he observed that the pool solved what we now call the bilateral monopoly problem. Both Wright-Martin and Curtiss separately had the industry over a barrel. Anyone that wanted a license from Wright-Martin had to pay the "excessive" royalty the Attorney General mentioned, only to find that Curtiss wanted something, as well. Combining the Wright and Curtiss patents in the pool resolves this dilemma: the Attorney General noted that the pool royalty is "materially lower" than what Wright-Martin alone had demanded, leading to "a substantial saving to the Government."(26) Implicit in this analysis is that Wright-Martin and Curtiss had each other over a barrel as well.(27) If each had not needed the other's intellectual property to make a viable aircraft, their technologies would have been competitors, not complements, and the combination of the patents would have been a very different story. In that case, instead of reducing a series of monopoly rent-seekers to one, it would have made a monopolist out of the only two competitors in the field. Then, manufacturers probably would have faced a higher royalty than the two competitors quoted separately.(28) As it was, though, the complementary relationship of the patents made their combination in the pool procompetitive. Just how procompetitive it was depended in part on whether the benefit that Wright-Martin and Curtiss were bestowing on each other would be freely available to their manufacturing competitors. Because the pool was open to all, the Attorney General found, the cross-license "render[ed] competition freer by giving every responsible manufacturer of aircraft access to all the inventions in the field."(29)
On the other hand, an agreement by the two firms not to license their patents to others would have drastically limited the procompetitive benefits that the Attorney General saw in the pool. If the two firms' patents were competitors, such an agreement would be little more than a horizontal boycott by competitors, which is a clear Sherman Act violation. To be sure, when, as here, the patents compete neither with each other nor with any others, there is no current competition to be harmed. But if there were or might soon have been alternatives to the pooled patents, an agreement not to license to others might then have prevented Curtiss, for example, from combining its patents with some technology that substituted for Wright-Martin's. This could be a potent strategy for stifling new rivalry. On the other hand, in a robustly competitive market, mutually exclusive cross-licenses might not be at all troublesome.
What if Wright-Martin and Curtiss agreed that they would make their pooled patents freely available, but only as a package? Again, as long as the firms' patents truly blocked the industry, and each other, it is hard to see how such an agreement would harm competition. But once plausible alternatives appeared for either Wright's or Curtiss' inventions, the mandatory package would begin to look more like a tying arrangement. But even that's not the end of the issue. Under our Intellectual Property Guidelines, tying arrangements are generally subject to an analysis that balances procompetitive benefits against anticompetitive effects.(30) And, as I indicated at the outset, we will be giving some additional guidance in the future about the tying and bundling together of intellectual property. For now, suffice it to say that the general question is whether the tie really helps to increase output, lower costs, or create an otherwise unavailable product -- or whether it simply serves to raise the price of the tying product. The fact that the Wright and Curtiss patents blocked the industry also made the flat per-airplane royalty acceptable. This provision, the Attorney General observed, "at first sight seems objectionable as possibly designed to extend the patent rights of these corporations to objects not covered by their patents."(31) While this is phrased in the parlance of misuse law, the concern is much the same as the one raised by Microsoft's per-processor license for DOS, which was at issue in the case we brought against Microsoft in 1994.(32) Since the flat royalty must be paid whether or not you use the licensed technology, it discourages you from using competing technologies. But in the 1917 aircraft industry, there was no OS/2, no operating system from Apple, and apparently, no new entrants on the horizon. In short, since there was essentially no technological competition out there to be squeezed, the convenience that the per-airplane royalty offered could not be outweighed by any anticompetitive effects. The true antitrust cognoscenti among you will probably be amazed to learn that nowhere in the opinion letter did Attorney General Gregory discuss the concept of innovation markets. I can attribute this lapse only to the press of wartime, and perhaps the fact that our Intellectual Property Guidelines were not issued until 78 years later.(33) But while he didn't explicitly enunciate the concept of a specific "market" in which innovation takes place, competition among the pool's members in aircraft research and development was clearly an important factor in the Attorney General's analysis. He noted that the members' obligation to contribute future inventions to the pool, for compensation to be determined by arbitrators, "might possibly be used to secure valuable inventions at unreasonable compensation."(34) But taking into account that the obligation helped ensure that each subscriber's patents would be open to all, the Attorney General concluded that the obligation's "possible abuse . . . scarcely justifies its condemnation in the absence of such abuse."(35) Today, I believe that the impact on the members' incentives to innovate would receive at least a bit more attention. The obligation to contribute future patents was not limited to improvements on the fundamental patents; thus, it didn't serve simply as insurance against the possibility that a licensee might expropriate the entire value of a patent in the field. Instead, it guaranteed that any innovation, no matter how revolutionary, would be thrown into the pot, probably for nothing and otherwise for an amount to be determined by arbitration. Now, it may be possible that the arbitration mechanism was so fair that other manufacturers could have felt confident that their research and development would be competitively rewarded. But in ordinary circumstances, I wonder whether "the purpose of keeping [all] the patents of each of the subscribers open to all" ought to prevail over the drag such a broad obligation might exert on innovation.
Of course, these weren't ordinary circumstances: there was a war on and the U.S. needed airplanes quickly. The government was willing to give up long-term innovation -- and accept what we now have learned to call "dynamic inefficiencies" -- in exchange for the current technology at an attractive price. And in this situation, that "policy choice" -- even with the benefit of hindsight -- would still appear to make sense. Unless World War I had developed into another Thirty Years' War, the promise of long-run innovation competition would have been pretty cold comfort to our troops.
Still, as long as there isn't a war on the next time you form a patent pool, give some thought to members' future-patent obligations. Do they have to contribute only improvements? Or anything they invent in the field? If the latter, what is the justification? Are members free to license their future patents independently of the pool? Or is the pool going to be the only route the technology can travel to the market? Again, if the latter, what's the justification?(36) If the answers to these questions don't tell a very persuasive story about how the future patent obligations contribute to the pool's procompetitive purpose, you may want to rethink what's going on, or else risk Sherman Act exposure. The question of compensation that the Attorney General addressed is worth considering, too. The most obvious threat to the innovation incentive is if pool members have to give royalty-free licenses. But even if the pool agreement provides for a "reasonable" royalty, the actual royalty for new inventions may be artificially low if it set by the rest of pool -- i.e., by the innovator's competitors.(37)
III. The Need for a New Notification Regime
All this thinking about how we look at cross-licenses won't amount to much, of course, if we don't have the ability to investigate them effectively. Based on my activities at the Division over the past couple of years, I am convinced that, unlike with respect to mergers and many other kinds of business practices, cross-licenses have remained largely off our screen. I also believe that there are some cross-licenses in place now that, had we been aware of them at the time of their inception, we might have sought to block under the antitrust laws. Indeed, these two conclusions I have drawn apply not only to cross-licenses, but to license agreements generally. The reason that they have stayed off our screen is because they are not by any means open and notorious, unlike, say a merger, and the two statutes that might have been thought to bring such arrangements to our attention generally do not. Let me elaborate.
Those of you who are familiar with the antitrust laws know that the premerger notification provisions of the Hart-Scott-Rodino Antitrust Improvements Act of 1976 require notification to both the Antitrust Division & the FTC of stock and asset acquisitions that give the acquirer an interest of more than $15 million in the acquired company's stock or assets, if the acquirer and acquiree exceed some size thresholds.(38) You probably also know that the FTC Premerger Notification Office, which is the arbiter of Hart-Scott-Rodino notification rules, views some intellectual property licenses as asset acquisitions.(39) But this has not made Hart-Scott-Rodino notification an effective net for potentially anticompetitive licenses. For one thing, only licenses that on their face are literally exclusive -- retaining no rights in the licensed field, even to the licensor -- qualify for HSR notification.(40) That is understandable, since HSR is meant to deal with asset acquisitions, in which the seller parts with beneficial ownership of whatever the acquirer is getting. I think we could all agree, though, that licenses not literally exclusive on their face can also raise competitive issues. For one thing, facially nonexclusive licenses can be de facto exclusive, due to devices such as royalty structures that penalize additional licensing, or, as in our suit against S.C. Johnson & Sons and Bayer AG, a side agreement.(41) For another, a nonexclusive license between two competitors could lead one of them to give up on its own technology, making its own product more like its competitor's and giving consumers less choice. In other situations, as where the royalty on a an input accounts for a large proportion of the finished good's sales price, a nonexclusive license could enable the parties to coordinate their prices with each other. These situations may not involve asset acquisitions within the purview of Hart-Scott-Rodino, but we as antitrust enforcers would still like to know about them. Even among exclusive licenses, only those that pass Sec. 7A's size-of-person and size-of-transaction tests require notification. Essentially, anything acquired from a non-manufacturing firm that has assets of less than $10 million and sales of less than $100 million will escape notification in any event. When it is an intellectual property right that is being acquired, it does not strain the imagination that a large, even dominant firm, could acquire a competitively important technology from a research and development outfit with assets and sales below the threshold levels. Such a transaction flies under the Hart-Scott-Rodino radar. So too will the licenses that fall short of the $15 million size-of-transaction test, a test difficult to apply objectively to a stream of royalty payments keyed to uncertain factors such as usage, production, or sales.(42)
Given all this, you won't be surprised to learn that licenses account for no more than a tiny fraction of the approximately 3000 transactions that are reported annually pursuant to Hart-Scott-Rodino filings. The problem is not with how Hart-Scott-Rodino is construed. The real problem is that the statute simply wasn't designed to apply to licenses in the first place.
Some of you might think that filings of patent interference settlements, pursuant to 35 U.S.C. § 135(c), might help to bring some potentially important matters to our attention. Enacted in 1962 to discourage competitors from using the medium of an interference settlement to collude,(43) the statute requires "any agreement or understanding between parties to an interference, including any collateral agreements referred to therein, made in connection with or in contemplation of the termination of the interference" to be filed with the PTO, from which the Antitrust Division may obtain it for review. The filing requirement not only gives us a look at settlement agreements but also should dissuade parties from entering into the most egregiously anticompetitive agreements.
At least, that is, so long as parties obey their obligation to file their entire agreement. Disobedience carries a facially harsh penalty: it renders the agreement and any related patents unenforceable. But I still can't help but wonder whether the statute is fully effective in ensuring filings. The PTO is not in a position to police the requirement that there be no unfiled side agreements. Nor are third parties, who may find it difficult even to see what has been filed; most settlers designate their agreements as confidential, shielding them from public view except upon "good cause." And only way we're likely to find out about a failure to file is through serendipity in an already-opened investigation. That did happen once, but it's not a happy story. In the course of an investigation, we concluded that the FMC Corporation had failed to honor its obligation to file all the agreements connected with an interference settlement it had entered into. We sued, asking the court to declare FMC's resulting patent unenforceable under Section 135(c). The district court found no violation of the statute;(44) on appeal, the Third Circuit held that the statute did not give the Government standing to have the patent declared unenforceable.(45) The Department did not seek certiorari, and FMC is the sole case on this issue. A bill to add an explicit right of action for the Government was introduced into Congress and died there in 1985. So, with the Antitrust Division apparently unable to enforce the statute, who can? Infringers, who might otherwise raise unenforceability as a defense, are unlikely to learn of a failure to file on their own, and almost certainly won't be able to learn about it from us. The Antitrust Civil Process Act imposes strict confidentiality requirements on our use of documents and information we obtain through Civil Investigative Demands.(46) Consequently, when, as in FMC, we learn through CIDs of a violation of Section 135(c), we cannot place an announcement in the New York Times. What this means, then, is that the only people who are likely to know about a violation of Section 135(c) and to be able to do something about it are the parties to the settlement themselves. Of course, sometimes even cartels break down. But to hinge law enforcement on changes of heart by violators places more hope in redemption of the human spirit than one in my position can afford to have. Even the Third Circuit recognized that denying us standing under Section 135(c) makes the filing requirement toothless. But that, the panel concluded, was a matter for Congress to redress.(47) As I indicated, Congress did take a brief look in 1985 at a bill that would have solved the problem with a simple addition to the end of Section 135(c): "The U.S. may bring an action for equitable or declaratory relief to enforce the provisions of this Section."(48) The bill enjoyed the support of the Commerce Department, but did not emerge from the Judiciary Committee. I think it may be time for another, more serious look at the same simple legislative fix. At the same time, if the opportunity to overturn FMC came along through litigation, I would jump at it. Like William Baxter, the distinguished leader of the Antitrust Division at the time of FMC, I think the case was wrongly decided. But the chance to persuade a court of that may not come again soon. FMC was one of only a very few instances where we came across a failure to file in the course of an antitrust investigation; who knows, given the difficulty of determining whether a required filing has been made, when the next instance will be. But whether we do it by legislation or litigation, vindicating our right of action would give the Section 135(c) filing requirement teeth, and consequently have some measure of confidence that we will learn about interference settlements that could otherwise stifle competition from outsiders and protect potentially invalid patents. But that fix, alone, addresses only one subset of potentially anticompetitive agreements, which may or may not involve the cross-licenses I've been talking about.
IV. A Proposed System of Notification
As I just indicated, the problem I see is not limited to just one particular setting, such as the settlement of interferences, or one particular device, such as a cross-license. The problem is with our inability to learn of a whole range of agreements involving intellectual property rights, agreements which may impede competition while affording no countervailing competitive benefits. One solution, and of course the only way to assure full detection of all licensing arrangements, would be a mandatory filing system. But, frankly, that would swamp the boat as far as I'm concerned. A practical solution hinges on our ability to articulate some special characteristics common to these agreements, so that we could insist on being told about some -- where competitive scrutiny might be worth the candle -- while letting the large majority go. I think that eventually we can come up with those identifying characteristics, based on the essence of the agreements themselves rather than what specific form they take or whether they result from a lawsuit, an interference, a cease-and-desist letter, or a government advisory committee. In the meantime, though, I propose that we start with the setting that concerns us the most: settlement of infringement disputes.
We should identify a category (or categories) of infringement cases in which we get notification at the outset. We could then monitor such cases and, if the parties decided to settle, we could either step into the defendants' shoes if we thought the settlement was anticompetitive or, perhaps, adopt a regime where the court, based on our comments and those of other interested parties -- could reject the settlement on public interest grounds. I should note that Bill Baxter made a similar proposal just last year. Our experience with West's star-pagination infringement claim provides an illustration of how this might work and a demonstration of the limitations of the status quo.
First, let me remind you of what is at stake for competition in the fight over star pagination. To put it simply, to produce a viable case law product, any case law publisher other than West itself must be able to star-paginate to tell the reader where each portion of each case it publishes may be found in West's printed system, so that the reader may employ pinpoint citation to the particular part of the case relevant to his analysis. West, however, contends that star pagination to its system infringes its copyrights by copying the "arrangement" of those volumes. It finds support in West Publishing Co. v. Mead Data Central, Inc.,(49) where the Eighth Circuit held that Lexis' use of star pagination did indeed infringe West's copyright. West ultimately granted Lexis a license but, until it agreed with the Department recently in an antitrust consent decree to license the right to star paginate on request,(50) West had licensed only one other publisher to star paginate -- in a set of Virgin Islands case reports. And competition has not thrived. In the meantime, though, the Supreme Court's Feist decision(51) pulled the rug out from under the Eighth Circuit's analysis, at least as we see it. Thus you shouldn't have to have a license to star paginate, but as long as West sees it the other way, the difficulty of getting a license is a considerable entry barrier into the case law publishing business. Because of our interest in competition in the provision of legal research products and services, we followed the two declaratory-judgment actions brought against West by rival case law publishers with considerable interest.(52) We sought to appear as amicus in both at about the same time, but at very different stages. In the litigation brought by Matthew Bender and HyperLaw against West in the Southern District of New York, we filed our brief during the pendency of the plaintiffs' motions for summary judgment. The brief we filed in Oasis v. West, though, was in the Eighth Circuit; the District Court, following West v. Mead, had already entered summary judgment for West. Ordinarily, we are wary of weighing in at the District Court level, since the record frequently is not as well developed, and the legal issues are not always as clearly delineated, as we might like. In this instance, though, the legal issue -- whether star pagination infringes West's copyright -- was straightforward, and so were the facts, enabling us to contribute as intelligently to the summary-judgment question before Judge Martin in New York as we were to the Eighth Circuit's appellate review. And, as it turned out, the one decision that has come down on this issue where we have been heard as amicus has gone our way, whereas the courts that did not hear us, both the District Court in Oasis and the courts in West v. Mead, went the other way. But even though our amicus filing may have had an impact here, I am not convinced that such a role is enough.
For one thing, although we were able to be heard in these proceedings, we cannot prevent the parties from settling and taking the issue away from us. A settlement in the Oasis matter, for example, would not only still our voice but also leave West v. Mead intact. For another thing, we were lucky: the West cases are the rare instance where we actually are aware of infringement litigation because we had a pre-existing interest in this industry. In the more usual situation, an infringement dispute could go to court and settle without ever coming to our attention. Notice of an infringement action would at least allow us to assess what is at stake for competition in the matter while it is pending, putting us in a position to decide quickly and, I would hope, confidently when confronted with a settlement. If we had that right to be heard, we could ensure that meritorious defenses would not be abandoned, and questionable intellectual property claims would not triumph, without at least an opportunity for us to consider whether broader societal interests in competition warrant putting the claims to their proof, and to bring those considerations to the court's attention. Then, those broader interests would not be held hostage to the defendant's own economic interests, which may be subject to limited resources for litigation and a strong aversion to the consequences of defeat, no matter how remote the chances. Of course, we might well still wish to be heard as amicus on the merits when the dispute does go to litigation, as it has in the current star-pagination cases. But it is easy to imagine a case being at a stage where amicus participation might not yet be prudent, for the reasons I alluded to above, yet a settlement would be very troublesome. There, the additional right to be heard in some manner in advance of any settlement would be essential to ensure that competitive concerns be heard at all.
In sum, I see a lot of merit in a new notification requirement, and I'm giving serious thought to how we can make it work. Of course, such a system could require a significant investment of Antitrust Division resources, both in reviewing proposed settlements and in following through in the occasional case by litigating on behalf of the settling defendant or providing comments to the court. But it's an important enough challenge that we ought to figure out how we can meet it. I think we could devise a notification regimen that would impose only mildly on the settling parties. I recognize that many settlements, for understandable reasons, are confidential. Those of you who have dealt with the Division know that our staff deals daily with the most sensitive confidential business information, and has a well-deserved reputation for guarding it carefully. There is every reason to think that we could accord the same degree of care to confidential settlements.
I recognize, of course, that this may not be much comfort to patentees, who might find themselves put to their proof more often. But I remain convinced that the current situation means that, whenever there is even a more than trivial possibility of infringement, the costs of litigation skew the parties' decisions, steering them away from a serious test of the bounds of the rights of the patentee or copyright holder and towards agreements that too often make teammates out of rivals. Since society picks up the tab for these agreements over the long run, I think it may be worth an investment of our resources up front to head them off where necessary.
Lastly, lest there be any misunderstanding here, I want to make it clear that I don't think it is necessary for the Division to get notice of every last infringement settlement. We should be able to come up with some easily understood and applicable criteria. Whether they ought to be on the basis of the size of the parties, their shares of the relevant markets, the economic significance of the alleged infringement or the product or service in which the allegedly infringed intellectual property is put to use, or some combination of the above, needs to be considered carefully. Whatever criteria we adopt, I want to assure you, will be designed to impose a minimal bar to settlements that don't pose a serious competitive threat, and to limit the call on the Antitrust Division's resources to reviewing economically significant and competitively troubling agreements.
Again, I want to say how grateful I am for the chance to speak with you today. I am very interested in your thoughts as to the ideas I have put forward, and more generally about the Antitrust Division's protection of competition in the creation and exploitation of intellectual property. As the advocates for the competitors in this arena, you can and should be significant contributors to our efforts to develop antitrust doctrine for intellectual property that is at once vigilant, sensitive to the facts of each case, capable of consistent application, predictable, and, of course, to come back to patent law concepts, useful without necessarily being too novel.
1. 1See Matthew Bender & Co. and HyperLaw, Inc. v. West Publishing Co., No. 94 Civ. 0589 (JSM) (S.D.N.Y. November 22, 1996)(bench ruling), appeal withdrawn, No. 96-9711 (2d Cir., March 4, 1997); Oasis Publishing Co. v. West Publishing Co., No. 96-2887 (8th Cir., oral argument held March 10, 1997).
2. See, e.g., United States v. General Electric Co. (D. Mont., complaint filed August 1. 1996); United States v. S.C. Johnson & Son, Inc., 1995-1 Trade Cas. (CCH) ¶ 70,884 (N.D. Ill. 1994)(consent decree); United States v. Pilkington plc, 1994-2 Trade Cas. (CCH) ¶ 70,842 (D. Ariz. 1994)(consent decree); United States v. Microsoft Corp., 1995-2 Trade Cas. (CCH) ¶ 71,096 (D.D.C. 1995)(consent decree). The complaints in General Electric, Pilkington, and Microsoft, as well as numerous pleadings from the many other cases we have brought over the last few years, are available at the Antitrust Division's web site, www.usdoj.gov/atr.
3. By "cross-licensing," I mean the interchange of intellectual property rights between two or more persons. By "patent pool," I mean the aggregation of intellectual property rights which are the subject of cross-licensing, whether they are transferred directly by patentee to licensee or through some medium, such as a joint venture, set up specifically to administer the patent pool. I hope my use of these terms comports reasonably well with your own; I take some comfort in knowing that the Supreme Court has stated that "patent pool" is not a term of art. United States v. Line Materials, 333 U.S. 287, 313 n.24 (1948).
4. Our Intellectual Property Guidelines give several examples of intellectual property licensing's procompetitive benefits. Antitrust Guidelines for the Licensing of Intellectual Property ("IP Guidelines"), § 2.3 (1995).
5. Only a very few licenses are likely to fall prey to Section 1's per se rule, which automatically condemns the most egregiously anticompetitive agreements. IP Guidelines, § 3.4.
6. Manufacturers Aircraft Association -- Antitrust Laws, 31 Op. Atty. Gen. 166 (1917).
7. Id. at 167.
8. Id. The Wright company was demanding $1000 per plane, George Bittlingmayer, Property Rights, Progress, and the Aircraft Patent Agreement, 31 J. L. & Econ. 227, 232 (1988), which then amounted to five percent of airplanes' $20,000 average cost, Harry T. Dykman, Patent Licensing Within the Manufacturer's Aircraft Association (MAA), 46 J. Pat. Off. Soc'y 646, 649 (1964).
9. Dykman at 647.
10. Id. Dykman characterizes the aircraft manufacturers and patentees as classic rugged individualists who were damned if they would readily give their competitors access to their inventions "under anything like a reasonable basis." Id. Of course, this tends to suggest that, if they were each that determined to go it alone, they either were not blocked by, say, the Wright patent, or they were deluding themselves. In any case, it appears that the specter of eminent domain helped bring them all to the table. See Bittlingmayer at 232.
11. 31 Op. Atty. Gen. at 167.
12. Id. at 168. Among them were some very basic patents indeed, such as "Method of Getting a Hydroairplane Off the Water Into the Air," and "Heavier Than Air Flying Machines." Dykman at 648-49.
13. Evidently the award of a government contract was deemed a satisfactory indicator of responsibility.
14. 31 Op. Atty. Gen. at 171.
15. Of course, 100 would seem like an absurdly high number for today's aircraft industry, but that's another speech.
16. Id. at 168-69.
17. Dykman at 650. The following year, the Secretary of the Navy, armed again with the threat of eminent domain, jawboned the per-unit royalty down to $100, although the $2 million ceilings remained in place. Id. at 655; see also Manufacturers Aircraft Association v. United States, 77 Ct. Cl. 481, 490-93 (1933)(action by pool for payment of royalties on airplanes made for U.S. government by non-members of pool).
18. 31 Op. Atty. Gen. at 169.
19. Id. at 169.
20. Id. at 169, 170.
21. Dykman at 650.
22. The Secretary may have been acting under duress, since the pool's formation had provoked an uproar about an "aircraft trust." Bittlingmayer at 235 n.30. Today, if you were to ask us for a business review letter on conduct your client had embarked on five months earlier, you'd be out of luck. See Antitrust Division Business Review Procedure, 28 C.F.R. §50.6, ¶ 2.
23. 31 Op. Atty. Gen. at 170-71.
24. This was the point of our suit against Pilkington PLC, which employed a web of exclusive licenses on its expired patents and commonly known know-how to maintain an international cartel in the manufacture of float glass. See United States v. Pilkington plc, n.2, supra.
25. While the Attorney General had to take validity on faith, soon after, the Bureau of Aircraft Production submitted the question to a panel of patent experts, who confirmed both validity and the bargain the $200 royalty represented. See Manufacturers Aircraft Association v. United States, 77 Ct. Cl. at 489-90.
Because the pool provided that Wright-Martin's and Curtiss' royalties would terminate upon the expiration of their patents, the Attorney General did not have to consider the implications of any mechanism to provide post-expiration income to the patentees. It is not clear from the Attorney General's opinion whether the two firms' role in the pool's governance was to expire with their patents.
26. 31 Op. Atty. Gen. at 167. Absent government arm-twisting, one might not expect every such pool to lead to such dramatically reduced aggregate royalties. But any combination of complementary intellectual property rights brings with it the potential for genuine savings in transaction costs and provides a forum for owners of blocking complementary patents to reach a royalty agreement that will make access to their technologies feasible.
27. All that was judicially determined was that Wright had Curtiss over a barrel. Wright sued Curtiss in 1909, contending that Curtiss' use of wing flaps (suggested by Alexander Graham Bell) infringed the basic Wright patent. Bittlingmayer at 231. Although Wright won, Curtiss altered its invention, reopening the dispute, which was still underway at the time of the pool's formation. Id. At a minimum, the Curtiss invention seems to be a valuable improvement on Wright's.
28. Admittedly, they would save the cost and inconvenience of negotiating a competitive royalty.
29. Id. at 171. This conclusion, which focuses on the pool's cross-license, seems to depend upon the representation made to the Attorney General that the by-laws' limitation of membership to 100 was more than adequate to accommodate the likely interest in membership. Id. However, the Attorney General stated, that limitation "may prove objectionable" if the industry expanded. Id.
30. IP Guidelines, § 5.3.
31. Id. at 170.
32. 32See United States v. Microsoft, n.2, supra.
33. The Attorney General would have found Section 3.2.3 especially helpful.
34. Id. at 170. This was essentially the very same issue that fueled our 1994 inquiry into the intellectual property policy of the European Telecommunications Standards Institute, which raised the specter of a technology buyers' cartel. Because ETSI changed its policy, we chose not to pursue the matter further.
36. The answer to this question could be very important if the new technology were itself a rival of the basic pool technology or could be used in combination with such a rival.
37. That our own analysis today of the pool at its outset would be very much along the lines of the Attorney General's doesn't mean that our stance towards conduct we assess in a business review might not change over time if the facts or circumstances change, or antitrust analysis evolves. In fact, while the aircraft pool got a clean bill of health from us in 1917, it was the subject of several antitrust investigations, and ultimately was ended by a consent decree we obtained after suing in 1972 to break it up, on the ground that it had, in fact, retarded innovation. See United States v. Manufacturers Aircraft Assn., 1976-1 Trade Cas. (CCH) ¶ 60,810 (S.D.N.Y. 1975). This suit has been the subject of thought-provoking commentary. See, e.g., Bittlingmayer at 235-40. The moral of this story is that not even a business review is forever, and that, to be responsible enforcers of the antitrust laws, we have to reserve the right to review the effects of old practices as times change, especially if things pan out differently from how they were originally presented to us.
38. 38See 15 U.S.C. § 18A.
39. See Premerger Notification Practice Manual, 1991 ed., American Bar Assn., Section of Antitrust Law, Interpretation 49 at 45.
41. 41Bayer had undertaken extensive preparations to enter the U.S. market with a patented new insecticide, which was likely to provide potent competition to S.C. Johnson & Sons, the incumbent market leader. At the last minute, though, Bayer instead licensed the product to Johnson. Because we concluded that the license would be the only one Bayer would grant in the U.S., it deprived US consumers of a likely entrant in the insecticides market -- Bayer itself -- without any countervailing benefits. The consent decree we obtained enjoined the license. See United States v. S.C. Johnson & Sons, n.2, supra.
42. See, e.g., id., Interpretation 116 at 95.
43. See United States v. Singer Mfg. Co., 374 U.S. 174 (1963). There, the defendant and a Swiss competitor settled an interference, scuttling evidence that cast doubt on the validity of the Swiss firm's patent, as part of a larger set of understandings that allowed them to focus their energies on excluding Japanese competition. See id., 374 U.S. at 197-99 (White, J., concurring).
44. United States v. FMC Corp., 514 F. Supp. 1166 (E.D. Pa. 1981), reversed, 717 F.2d 775 (3d Cir. 1983).
45. 717 F.2d at 787.
46. 15 U.S.C. § 1313(c)(3).
47. 717 F.2d at 787.
48. S. 1358, 99th Cong., 1st Sess. (1985).
49. 799 F.2d 1219 (8th Cir. 1986), cert. denied, 479 U.S. 1070 (1987).
50. See United States v. The Thomson Corp. & West Publishing Co., Civ. Action No. 96-1415 (PLF) (D.D.C., consent decree entered March 7, 1997) (resolving competitive concerns raised by West's acquisition by The Thomson Corporation).
51. Feist Publications, Inc. v. Rural Telephone Service Co., 499 U.S. 340, 349 (1990).
52. See n.1, supra. Our brief in Oasis, which is largely the same as the one we filed in Bender, is available at our web site. See n.2, supra.