< Return To Hearing
Mark A. Lemley
Patent Reform Legislation
Testimony of Mark A. Lemley, Stanford Law School before the Senate Committee on the Judiciary
June 14, 2005
Introduction and Executive Summary
Reforming the patent system is important. Patents are critical to innovation, and the patent system generally works well in encouraging invention. But the system also has problems, and is in need of an overhaul. In particular, improvements can be made in two main areas: (1) streamlining the law by simplifying unnecessarily complex rules and harmonizing our laws with foreign laws to the extent possible; and (2) preventing abuses of the system by people who use patents not for their intended purpose of supporting innovation, but to hold up legitimate innovators.
Let me be clear at the outset that these are both important problems, and patent reform that addresses those problems will be an important step in encouraging innovation in the United States. It is particularly important that Congress act to prevent abuses of the patent system by so-called "patent trolls," who use the patent system not to develop and make products but to squeeze money out of those who do. While there are no reliable statistics on the extent of the troll problem, there is no question that it is a widespread and extremely serious problem in the semiconductor, computer, and telecommunications industries. Large, innovative companies such as Intel and Cisco never have a week go by without threats of suit from a non-manufacturing patent owner claiming rights in technology that the defendants did not copy from the patent owner - usually they've never even heard of the patent owner - but instead developed independently. While there is a legitimate role for small and individual inventors who patent their technologies and license their ideas to others, increasingly the patent owners are not contributing ideas at all, but popping up years or even decades later and trying to fit an old patent to a different purpose. Trolls do this because the law permits it, and because it gives them a chance to make a lot of money - under current law, far more money than their technology is worth.
Patent reform needs to deal with these abuses of the system without interfering with the normal, legitimate use of the system to protect and encourage innovation. Doing so requires careful balancing of the interests of patent owners, technology companies, and the public.
One fact that complicates patent reform efforts.is that the patent system works very differently in different industries. See Dan L. Burk & Mark A. Lemley, Policy Levers in Patent Law, Va. L. Rev. (2003). While innovators in the semiconductor, computer, Internet and telecommunication industries identify abusive patent litigation as the major problem they face, there is no similar problem in the medical device, biotechnology and pharmaceutical industries. Those industries have very different characteristics - pharmaceutical patents are more likely to cover a whole drug, rather than one of 5,000 different components of a semiconductor chip. So patent owners in the pharmaceutical industries don't have to worry about and endless stream of patent owners asserting rights in their drugs. Further, innovators in the biotechnology and pharmaceutical industries consider patent protection far more important to their R&D efforts than do the information technology industries. The challenge is to craft a unitary patent law that can accommodate the very different needs of each of these important industries.
Because patents are so important to a large group of stakeholders, and those stakeholders have such diverse interests, it may not be possible to get universal agreement on all aspects of a comprehensive reform bill. A workable bill will necessarily involve compromises, and won't leave everyone happy. That is not a reason to abandon the effort. Rather, it suggests the need to take measured steps towards reforming the system.
To date, the only patent reform legislation introduced in this Congress is H.R. 2795 (Rep. Smith). As a result, I will organize my specific thoughts on particular proposed reforms around that bill. H.R. 2795 is an important step improving the patent system. In general, I am in favor of virtually all of the reforms in the bill. I do have a few suggestions for improvement, however. What follows is a section-by-section analysis.
Summary: This is an important change, and this section of H.R. 2975 needs no revision. However, the section works only if the bill continues to include the provisions of section 9 requiring publication of all patent applications and expansion of prior user rights. If those provisions are not included, Congress should oppose the move to first inventor to file.
The move to a first-inventor-to-file system is an important step for several reasons. First, it simplifies the complex of rules for deciding whether a patent applicant is the first inventor. One way a focus on the filing date simplifies things is to eliminate the need to determine when an invention occurred in the vast majority of cases, an inquiry that has proven difficult. But the changes to section 102 also get rid of three other confusing rules that add uncertainty to the patent system: the "secret prior art" rules governing commercial but nonpublic use, and that differ depending on whether the user is the patentee or not; the "experimental use" exemption based on a totality of the circumstances analysis; and the perplexing definition of when an invention is on sale. All these rules have created inconsistent judicial guidance and made it hard to know when an inventor was entitled to a patent.
Second, first inventor to file recognizes the international nature of today's markets. The current statute defines prior art differently depending on whether a sale or a conference occurs in the U.S., Canada or Europe. Eliminating this distinction makes sense in the modern world. Because the rest of the world already uses filing rather than invention date to measure priority, first inventor to file will take an important step towards global harmonization, permitting U.S. inventors to more easily seek patent protection not just in the U.S. but in other countries as well.
In the past, small inventors have expressed concern that a first to file system will disadvantage them because large companies have the resources to file patents more quickly. More recent evidence demonstrates that that is not true. It is large inventors, not small inventors, who most benefit from the complex and expensive interference system that determines who was first to invent. And large inventors challenge the patents of small inventors in an interference proceeding more often than the reverse. Eliminating interferences will help, not hurt, small inventors.
Further, H.R. 2795 contains an important deviation from a pure first-to-file system: it gives inventors who sell, use or publish their invention a year to get a patent application on file. This is a reasonable grace period. A small inventor concerned about losing a race to the patent office can publish the invention on a Web site. Doing so will prevent anyone else from getting a patent, while giving the inventor a year to find a patent attorney and file a patent application. Given the existence of simple provisional applications, that is a reasonable accommodation.
Summary: H.R. 2795 needs no revision.
The National Academy of Sciences recommended the elimination of unnecessary mental states in patent law. This would simplify patent litigation, reducing its uncertainty and hopefully its cost.
The most important change in this section is the elimination of the best mode requirement. That requirement invalidates patents when the inventor has not disclosed her preferred way of implementing the invention, even if she has given enough information to enable scientists in the field to make and use the invention. The best mode requirement does serve a purpose - it prevents inventors from obtaining the benefits of a patent without giving the public the full benefit of disclosure. But on balance, the benefits of the doctrine aren't worth the costs. Because the best mode doctrine is based on the beliefs and intent of the actual inventor, the doctrine serves as a "gotcha" that can invalidate novel and nonobvious patents regardless of the good faith of the company that owns them. Indeed, the doctrine has been responsible for more than 10% of all the patents invalidated in court during the 1990s. John R. Allison & Mark A. Lemley, Empirical Evidence on the Validity of Litigated Patents, 26 AIPLA Q.J. 185 (1998). The enablement and written description requirements, properly applied, can require sufficient disclosure to benefit the public.
Summary: Eliminating inequitable conduct from litigation is a major change that should not be entered into lightly because it will encourage deceit by unscrupulous patent applicants. This section should be included in patent reform legislation only if counterbalanced by significant limitations on abuse of the patent system. Even if it is, some textual changes could improve the operation of H.R. 2795.
The doctrine of inequitable conduct discourages deception by patent applicants during the ex parte patent examination process. It is very important to discourage that deception, because the patent examiner can't independently verify what the applicant tells him. At the same time, the doctrine of inequitable conduct is often abused by litigation defendants, who assert bogus claims of inequitable conduct.
H.R. 2795 as currently drafted essentially abolishes the litigation defense of inequitable conduct, replacing it with an administrative system within the patent office. Courts could not consider inequitable conduct unless and until a patent claim had already been invalidated. This tightens up the inequitable conduct standard, requiring that a deception of the patent office actually have led to a patent that would not otherwise issue. Even then, they could not render the patent unenforceable unless a business person within the company, rather than just an in-house or outside lawyer, participated in the fraud. This tightening of the standard may be appropriate as a means of preventing abuse of the defense in litigation. But as the statute is currently drafted, it raises the prospect that an unscrupulous patent owner could get away with filing fraudulent patent applications in certain circumstances. This is a particular risk if the bill does not contain adequate provisions deterring patent "trolls," who have particular incentives to mislead the PTO.
Limiting abuse of the patent system is an important element of the reform package. Eliminating inequitable conduct could actually make the problem worse, not better. So, while there are reasons to do so, Congress should take such a drastic step only if there are other provisions in the legislation that effectively limit abuse, such as appropriate restrictions on injunctive relief, excessive damages, and continuation applications.
Even if Congress decides to go forward with the effective elimination of the inequitable conduct defense, the specific provisions of H.R. 2795 could be improved in several ways, detailed below.
Problem 1: Patent owners could avoid rulings of inequitable conduct by strategically deciding not to litigate particular claims in the patent. They may assert claims, then drop them when it becomes apparent that their claim will be invalidated or their misconduct discovered. Because H.R. 2795 as currently written requires that the validity of a claim be an issue in the litigation, and actually be adjudged invalid, a patentee who strategically drops claims will avoid ever being called to account for its conduct.
Solution: Section 136(d)(2) in H.R. 2795 should be modified by adding at the end of subsection (A): ". If a claim is asserted by the patentee in the action, the court shall have jurisdiction to determine its validity even if the patentee later withdraws that claim;"
Problem 2: The newly created administrative bureaucracy will work well only if people take it seriously. The administrative sanction is relatively small, and patent applicants with a lot of money at stake may decide to "roll the dice" and lie to the PTO because even if they are caught, they will receive only a relative slap on the wrist.
Solution: The penalties need to be increased. Section 136(e)(6)(B) in H.R. 2795 should be modified by striking "$1,000,000" and replacing it with "$10,000,000" and by striking "$5,000,000" and replacing it with "$50,000,000." Further, section 136(d)(2)(B) should be modified by adding at the end ", or the Director has made a nonappealable adjudication of misconduct under this section;"
Problem 3: The language of H.R. 2795 regarding preemption of other actions is self-contradictory. Section 136(g) provides that other proceedings can be brought based on a final determination of misconduct, but then says "except that nothing in this subsection shall authorize any investigation or determination of misconduct that is otherwise preempted under this section." But section 136(c)(1) preempts all such other proceedings, rendering section 136(g) of no effect.
Solution: The language "except that nothing in this subsection shall authorize any investigation or determination of misconduct that is otherwise preempted under this section." should be deleted from section 136(g) of H.R. 2795.
Problem 4: Finally, one apparently unintended consequence of H.R. 2795 would be to prevent the government from asserting inequitable conduct as a party to litigation, whether when defending itself in a patent suit or engaging in an antitrust investigation. This flows from the language of section 136(c)(1), which forbids not just courts but any "federal department or agency" and any "other Federal or State governmental entity" from investigating claims of inequitable conduct. Read literally, this would prevent the federal government or any state from asserting the defense even when private defendants could do so. It would also prevent the Antitrust Division or the Federal Trade Commission from investigating or charging antitrust violations based on fraud on the patent office (so-called "Walker Process fraud").
Solution: These problems could be solved by changing the language "may investigate or make a determination or an adjudication" in section 136(c)(1) of H.R. 2795 to read "may adjudicate", and to add at the end of that subsection "or in the antitrust laws."
Summary: The change to the entire market value rule in reasonable royalty damages and the limitation of willfulness claims are both extremely important. The reasonable royalty portion of H.R. 2795 does not need any modification. The willfulness provision of that bill improves the current law in certain respects, but could be made better still.
The reasonable royalty provisions in the existing law create significant problems in those industries in which patented inventions relate not to an entire product, but to a small component of a larger product. Because courts have interpreted the reasonable royalty provision to require the award of royalties based on the "entire market value," juries tend to award royalty rates that don't take into account all of the other, unpatented components of the defendant's product. This in turn encourages patent trolls, who can obtain damages or settlements that far exceed the actual contribution of the patent. There are numerous cases of just this problem occurring. I am currently conducting an empirical study to determine the extent of the problem, but there is no doubt that it exists. The bill solves this problem by encouraging the courts to consider the contribution of other elements of the invention.
The doctrine of willfulness is a mess. Over 90% of all patent plaintiffs assert willful infringement, even though most of the defendants in those cases developed their products independently and had never heard of the plaintiff or its patent. They are not "willful" in any ordinary meaning of the term. Rather, the way the courts have interpreted patent law has created a bizarre game. By sending a carefully crafted letter, patent owners can cause companies to have to obtain written opinion letters and waive the attorney-client privilege, and if they don't can declare them willful infringers for continuing to sell products they designed in good faith. It is important to clean up the willfulness doctrine. [While some have proposed eliminating it altogether, I think that goes too far. Enhanced damages for willfulness serve as an important deterrent in those cases where the defendant really does steal the technology from the patent owner.]
H.R. 2795 makes two important changes that reduce the abuse of willfulness. First, it requires a letter that puts the defendant on notice of a patent to be sufficiently specific that a defendant can file a declaratory judgment action asserting its innocence. This should reduce the casual, off-hand sending of such letters. Second, by requiring the pleading and litigation of willfulness only after a defendant has been found to infringe, H.R. 2795 eliminates many of the harms associated with the court's reliance on advice of counsel, because the defendant will not have to decide whether to waive the privilege until after the primary trial has ended. Further, by requiring bifurcation of willfulness, the bill simplifies the patent litigation process by separating out discovery as to willfulness and eliminating the need for that discovery in the cases where the patent is ultimately held invalid or not infringed.
However, H.R. 2795 as currently written leaves intact the opinion letter "game" for many patent lawsuits. Because a defendant's only defense to willfulness under the statute is the existence of "an informed good faith belief" in invalidity or noninfringement, defendants are as a practical matter extremely likely to decide they have to obtain an opinion, rely on the advice of counsel, and therefore waive the attorney-client privilege. This waiver distorts legal advice in difficult ways, making settlement more difficult. See Mark A. Lemley & Ragesh K. Tangri, Ending Patent Law's Willfulness Game, Berkeley Tech. L.J. (2003).
This problem could largely be solved if defendants could rely on strong (though ultimately unsuccessful) arguments to avoid a finding of willfulness. To do this, section 284(b)(3) of H.R. 2795 should be modified by adding after "under paragraph (2)" the following: "if the infringer offered an objectively reasonable defense in court or". This would make either an objectively reasonable argument or a subjectively good faith belief grounds for avoiding willfulness. It makes little sense to conclude that defendants are acting willfully if the case was a close one. Adding an objective reasonableness defense would permit defendants who think they have a strong argument to rely on that argument, rather than having to waive privilege.
Summary: Injunctive relief is an important part of the patent right, but it is subject to abuse by patent trolls. It is important to preserve the right of injunctive relief in the case of legitimate patent claims, while preventing those who abuse the system from using the threat of injunctive relief to extort money from legitimate innovators. H.R. 2795 takes a small step in the right direction by giving courts the power to stay injunctive relief pending appeal where doing so wouldn't harm the patentee. It takes a more significant step by explicitly introducing fairness concerns, but doing so may have unintended consequences. Limiting injunctions based on patents that cover only a small part of a product would be a significant improvement, and need not raise any of the concerns that have been expressed about compulsory licensing.
The goal of any revision to the injunctive relief sections of the patent law should be to ensure that people who actually need injunctive relief to protect their markets or ensure a return on their investment can get it, but that people can't use the threat of an injunction against a complex product based on one infringing piece to hold up the defendant and extract a greater share of the value of that product than their patent warrants.
Some have suggested that any restriction on the right to injunctive relief amounts to compelling patent owners to license their competitors. That's not true. The presumptive right to injunctive relief is an important part of the patent law, and in most cases there will be no question as to the patentee's entitlement to such relief. To begin, an injunction is warranted if the patentee practices the patent. Even if they don't, if the patentee sells a competing product in the marketplace, they should be entitled to an injunction to prevent their own invention (in the hands of an infringer) from competing with themselves. Similarly, if they assign or exclusively license the patent to someone who competes in the marketplace, they should also be entitled to injunctive relief. And even if the patentee hasn't done these things in the past, if they begin to do so in the future they should have a right to injunctive relief. Patentees also ought to be entitled to an injunction in cases of willful infringement, even if they are not participating in the market and have no plans to do so. Infringers shouldn't be able to intentionally take the patented technology knowing they will only have to pay a royalty. Even if none of these things are true, some injunctions won't lead to a risk of holdup, and so even patentees who don't meet any of the criteria listed above should be entitled to an injunction in ordinary circumstances.
That said, an absolute entitlement to injunctive relief can and does permit patent trolls to "hold up" defendants by threatening to enjoin products that are predominantly noninfringing. In numerous cases, the parties settle for an amount of money that significantly exceeds what the plaintiff could have made in damages and ongoing royalties had they won. In these cases it is not the value of the patent, but the costs to the defendant of switching technologies midstream, that are driving the high price being paid. For example, on patent owner charges a 0.75% royalty for patents that don't cover industry standards, and 3.75% for patents that do cover industry standards. The technology isn't any better, but they can demand five times as much money once the industry has made irreversible investments. This is of particular concern when the patent itself covers only a small piece of the product. An Intel microprocessor may include 5,000 different inventions, some made at Intel and some licensed from outside. If Intel unknowingly infringes a patent on one of those inventions, the patent owner can threaten to stop the sale of the entire microprocessor until Intel can retool its entire fab to avoid infringement. Small wonder, then, that patentees regularly settle with companies in the information technology industries for far more money than their inventions are actually worth. The companies are paying holdup money to avoid the threat of infringement. That's not a legitimate part of the value of a patent; it is a windfall to the patent owner that comes at the expense not of unscrupulous copyists but of legitimate companies doing their own R&D.
The question is how to accommodate these competing concerns. H.R. 2795 takes a two-pronged approach. First, it provides that courts have the power to stay injunctions pending appeal if they find that doing so won't irreparably harm the patent owner. This has long been the law, until the Federal Circuit's decision this year in MercExchange v. eBay suggested stays are inappropriate. The old rule permitting stays caused no mischief. Restoring the equitable power of courts to stay injunctions is a good idea. It will give companies time to retool their factories to avoid infringement. At the same time, the irreparable harm limitation ensures that patent owners that actually need injunctive relief, like pharmaceutical companies litigating against generics, will be entitled to get it.
The second change H.R. 2795 would make is to require a court to consider the fairness of injunctive relief to all parties in deciding whether to issue an injnction at all. This is a fairly minimal change to the statute, which already provides that courts should only issue injunctions "consistent with principles of equity" and "on such terms as they deem reasonable." 35 U.S.C. § 284. In short, permitting courts to consider equitable principles in deciding whether to grant injunctions was what Congress intended in writing the 1952 Patent Act; it's just that the Federal Circuit has strayed from the statutory language. The "fairness" provision merely highlights for the courts the duty they already have in the existing statute.
There are reasonable concerns, however, that a general "fairness" exemption could be misused to deny patent owners injunctive relief when they should in fact be entitled to it. While it is possible that this will happen in U.S. courts, the greater risk is that developing countries will seize on this language to try to deny any rights to pharmaceutical patent owners.
Because of this concern, I think it preferable to try to tackle the troll problem more directly while protecting the general right to injunctive relief. Thus, rather than grant general powers to consider fairness, I would specify a limited set of cases in which courts could - not must, but could - deny injunctive relief. A logical way to proceed would be to take the language H.R. 2795 introduces to deal with a similar problem in the context of damages, modifying it slightly to further insure that the normal right to injunctive relief is protected both in cases where the patentee is in the market and where the infringer acts willfully.
Summary: Abuse of continuation practice remains a significant problem in the patent system. H.R. 2795 as currently written does not adequately address that problem.
While there are some legitimate reasons to use patent continuation practice, there are also a number of problematic reasons to do so. I have detailed those problems elsewhere; see Mark A. Lemley & Kimberly A. Moore, Ending Abuse of Patent Continuations, 84 B.U. L. Rev. 63 (2004).
H.R. 2795 as currently drafted gives the Director authority to limit continuations, but does not compel him to act. Because this is such an important issue, I would prefer to see Congress act to limit abuse of patent continuations, either by restricting the total number of continuations that can be filed, or by preventing patentees from broadening their patent claims during a continuation application.
At a bare minimum, however, even if Congress does not act more directly to forbid abuse of continuations, Congress should ensure that the Director acts by changing "The Director may" to "The Director shall". It should also draft the statutory provision in such a way that the PTO actually limits the abuse of continuation applications, rather than merely permitting such use so long as the applicant pays a fee.
Summary: Requiring publication of all patent applications, expanding prior user rights, and creating a post-grant opposition system are all important changes that will improve the patent system. The provisions of H.R. 2795 need no revision.
Section 9 of H.R. 2795 contains three important reforms that are critical to making a first-inventor-to-file system work. If the first inventor to file system is to work, it is absolutely essential that the patent system require prompt publication of all U.S. patent applications. The definition of who is first to file in section 102(a)(2) of H.R. 2795 treats as filings only issued patents and patents published under section 122(b). Because section 122(b) currently permits some patent applications to avoid publication, it would eliminate those applications both as prior art and for priority purposes. A small inventor who filed such an application would not have their filing date count for priority purposes if it were unpublished. Section 9 of H.R. 2795 solves this problem.
Second, the move to first inventor to file makes it important that the bill be amended to provide prior user rights for those who engage in non-public use before the patentee files his application. The bill eliminates the existing categories of non-public prior art. Doing so risks permitting more, not fewer, patents to issue to people who were not truly the first inventor. Granting prior user rights to those who were already using the invention is a reasonable counterweight, because it gives the owners of such secret prior art at least the right to continue using technology they invented. The modifications to section 273 solve this problem by expanding a limited right that has been in the law for six years without creating any problems.
Finally, post-grant oppositions are a valuable addition to the patent system that will help identify and weed out bad patents without the cost and uncertainty of litigation. The post-grant opposition bill is well-written and will significantly improve the patent system.
H.R. 2795 adds new section 323, which permits a post-grant opposition to be filed either within 9 months after a patent issues or within 6 months after the opposer is notified of infringement, whichever comes later. The addition of the second, 6-month window has been controversial in some circumstances, but it is critical to the success of the post-grant opposition procedure. Because of the long timelines associated with many patents, and the fact that patent trolls often wait for years after patents issue before asserting them, limiting opposers to a 9-month window after the patent issued would render post-grant opposition ineffective for the majority of patents. An example is pharmaceutical patents. Because of the long FDA approval process, potential generic manufacturers will likely have no idea at the time a patent issues whether the drug it covers will survive clinical trials and be approved for sale. By the time they know which patents are actually important, it would be too late to oppose them. This problem may extend to other industries as well. Submarine patentees and other trolls often sit on patent rights for many years before asserting them against manufacturers. In order to take advantage of the bill's opposition procedure, those manufacturers would have to guess which of the millions of patents in force might become important a decade from now. Since only 1% of patents are ever litigated, forcing them to make such a guess would make the system worthless to most of the people who would use it.
H.R. 2795 solves this problem appropriately, by including a second window for defendants who were not on notice of the patent when it issued. This gives a short period in which to oppose patents once they are brought to a company's attention, without permitting undue delay.
Finally, there are a few other matters not included in H.R. 2795 that deserve Congressional attention.
Experimental Use Defense: Traditionally, courts have not held defendants liable for using a patented invention in the course of experimentation to try to design a new product that does not infringe the patent. That changed in 2002, when the Federal Circuit essentially eliminated that defense. Major bar groups such as AIPLA have endorsed the creation of a limited experimental use defense, and I agree that doing so would be a good idea. In doing so, however, Congress should take care not to eliminate the incentives for investing in research tools. While experimenting on a commercial product in order to improve on it or design around it is legitimate activity that deserves to be exempt from patent infringement, the use of a patented research tool for its intended purpose in research should not be exempt from patent infringement.
Fee Diversion. Every year, Congress diverts user fees paid by patent applicants away from funding the PTO to the general federal revenue. Title III of the COMPETE Act, S. 1020 (Sen. Coleman) would end that diversion. I support ending fee diversion. In an environment where there are numerous reports of bad patents issuing, we should make sure the PTO has the resources to do its job. I want to emphasize, however, that merely giving the PTO more money is not a substitute for real reform of the patent litigation process. No reasonable amount of money spent at the PTO will end the problem of patent trolls so long as they have incentives to game the system.
Federal Circuit Jurisdiction: The Supreme Court's interpretation of the patent jurisdiction statute, 28 U.S.C. § 1338(a), in the Vornado v. Holmes case has eliminated the uniformity that came with the Federal Circuit hearing all patent appeals. After Vornado, regional circuit courts and even state courts are now hearing some patent cases. The Federal Circuit Bar Association has proposed statutory changes to section 1338(a) that would restore to the Federal Circuit exclusive jurisdiction over patent cases, and that proposal was the subject of a hearing in the House in March 2005. I support the Federal Circuit proposal.