The patent system grants exclusive rights to innovators that disclose novel, useful, and non-obvious inventions. A patent increases the rents that an inventor can appropriate by restricting market competition. The social benefit of the balance between innovation and competition hinges on granting patent protection only to those inventions that legitimately deserve it. Unfortunately, determining whether an applicant’s disclosure warrants patent protection is very challenging and time-consuming and, despite the best efforts of examiners, many patents are granted in error. As such, society benefits from good-faith efforts to police the system and invalidate patents that should not have been granted in the first place.
To that end, the Patent Trial and Appeal Board (PTAB) provides an administrative forum in which private parties may challenge the validity of granted patents, and which avoids the notoriously high costs and long timeframes associated with patent litigation. In light of these efficiencies, it is not surprising that the frequency of PTAB petitions has grown substantially over time since the inauguration of the board in 2012.
However, notwithstanding its many benefits, PTAB may be exploited as a convenient platform for reaching potentially-anticompetitive “reverse payment” settlements. This is an agreement in which a monopolist (the patent owner) pays a potential market entrant (the petitioner) to drop its challenge and stay out of the market. Such agreements typically arise in pharmaceutical markets, with the seller of a patented brand-name drug paying off a generic firm that is attempting to enter the market. In its recent Actavis decision, the Supreme Court held that such agreements may violate the antitrust laws.
Our research paper examines the potential formation of reverse payment settlements in PTAB. Our empirical analysis focuses on PTAB trials in which a generic drug maker challenges one or more patents covering a brand-name drug. The PTAB database allows us to see whether, and at what date, the firms settle. But the terms of settlement are always confidential. This presents a challenge, because one possible settlement format involves the patent holder’s provision of a license to the petitioner, and this kind of agreement would not raise antitrust concerns. However, we can get around this ambiguity by exploiting data published by the Food and Drug Administration (FDA). In particular, we rely on the FDA’s “Orange Book” listings, which provide information on FDA-approved drugs, any applicable patents, and – most importantly – a list of any firms that have obtained approval to sell the drug or a generic equivalent.
Under Paragraph IV of the Hatch-Waxman Act, a firm cannot obtain FDA approval to sell a generic version of a patented drug without certifying that such patent is either invalid or would not be infringed by its proposed generic – or else by obtaining a license from the patent holder. Thus, in the PTAB trials we examine, the generic-petitioner did not have FDA approval – and thus was not listed in the Orange Book – as of the PTAB filing date. But, by looking at Orange Book listings that appear after the parties settle, we can infer what the generic-petitioner received in the PTAB settlement. In particular, if the petitioner is listed in the Orange Book, we conclude that it must have obtained a license in the settlement. If not, we infer a strong likelihood of reverse payment.
As a threshold matter, our empirical analysis shows that many PTAB petitions involving drug-related patents are settled – about 38%. Of those trials that settle, about 75% meet our criterion for inferring reverse payment. Curiously, about half of these settlements occur after a decision to institute the petition. It seems unlikely that a generic firm would settle for nothing right after the judge signals that it has a good chance of prevailing on final judgment, so the inference of reverse payment is particularly strong in these settlements. Another point to note is that some of the settlements occurred relatively recently. But even if we restrict focus to settlements occurring more than a year ago (to allow for possible delays in post-settlement FDA approval), we still find that about 75% of settlements meet our criterion for inferring reverse payment.
We should note that our sample size is small, because PTAB is relatively new, having been inaugurated in 2012. As such, there are only about 50 distinct settlements involving Orange-Book patents. Another important point is that we cannot identify the magnitude of any reverse payments, which is an important element of the antitrust question. Thus, we cannot speculate as to the likely outcome of an antitrust inquiry into any particular settlement.
In addition to the empirical analysis, we offer a number of detailed examples of settled PTAB trials. For instance, in one of the trial (which had recently been instituted), the parties’ joint motion to settle purports to forestall “any potential Hatch-Waxman litigation on this patent between these parties.” But the evidence suggests the petitioner did not obtain a license in the settlement, suggesting that it agreed to stay out of the drug market until the patent expires, notwithstanding that the institution decision suggests a reasonable likelihood of invalidation.
In many cases, the parties enter into parallel settlements simultaneously – one in PTAB and one in district court. A second example shows how the parties may use this strategically. In the district court settlement, the parties obtain a consent decree holding the patent valid and infringed (and enjoining the generic firm). This has a strong claim-preclusive effect that will keep the generic firm out of the market until the patent expires. The consent decree does not provide for a reverse payment, however, which ensures that a federal judge will not challenge it on antitrust grounds. But the parties can still achieve reverse payment by relegating it to the PTAB settlement, where the judge has no authority to enforce the antitrust laws. The result of this bifurcated settlement appears to be a de facto pay for delay agreement that lasts for the full remainder of the patent term.
One reason the PTAB is convenient for reaching reverse payment deals is that there is no direct antitrust oversight, since its judges are administrators with very narrow legal authority. Further, while there is a statute requiring agreements between Hatch-Waxman firms to be submitted to the Federal Trade Commission (FTC) for antitrust review, the parties may attempt to evade it. The statute requires submission of agreements that relate to the manufacture or sale of either firm’s drug. The parties might therefore feel justified (rightly or wrongly) in not submitting a consent decree stating that the patent is valid and would be infringed by the proposed generic, since this does not expressly address manufacture or sales. Alternatively, it could be that the parties submit a district court consent decree (which includes no reverse payment), but not the PTAB settlement (which could include a reverse payment). After all, a PTAB settlement simply says that the parties agree to terminate the IPR – it need not declare the patent valid – and this arguably does not relate to manufacture or sales. A final point is that, if the generic firm brings its challenge before filing an Abbreviated New Drug Application under Paragraph IV, then the FTC reporting statute would not apply.
We offer a few proposals to deal with these settlements. Reverse payment PTAB settlements between competitors should be scrutinized by the FTC, at least in industries like pharmaceuticals, where individual patents may substantially shape the intensity of competition. To accomplish this, the statutory rules requiring FTC antitrust review should be broadened so that a large reverse payment is itself sufficient to trigger the review requirement. As a failsafe, it would be beneficial if PTAB judges were instructed to submit settlements to the FTC if they involve large reverse payments between competing pharmaceutical companies. Further, if a PTAB settlement coincides with settlement of district court infringement litigation between the same parties, then the district court judge should view both settlement proposals as a consolidated agreement between the parties.
Although we focus principally on potentially-anticompetitive settlements, our paper also addresses the so-called “reverse patent troll” phenomenon. This is a non-producing company that uses PTAB purely as a holdup device. That is, a reverse troll uses the threat of invalidation to extract reverse payment settlements. The incentives of reverse trolls differ from those of traditional “patent trolls,” i.e. non-producing entities (NPEs), which are firms that earn profits by enforcing patents that they do not practice. A reverse troll does not directly benefit from a successful invalidation, while a NPE directly benefits from a holding that its patent is infringed. Thus, a NPE can earn money either through settlement or a successful judgment, whereas reverse trolls depend critically on the former. Furthermore, every NPE brings legal claims that no other firm can bring, for it is the sole owner of the patents it asserts. By contrast, when a reverse troll challenges a patent, any other party is free to bring the same challenge, and this may lead others to “free-ride” on its petition. These factors undermine the reliability and profitability of the reverse troll model, which may be the reason why this business model is not widely observed in practice.
Importantly, a reverse troll has no interest in “policing” the patent system. Indeed, it seeks only settlements, which leave no mark on the patent system. As such, it may be beneficial to discourage this kind of conduct without deterring good faith petitions. One way to accomplish this would be to instruct PTAB judges to reject settlement proposals involving reverse payments that are large in relation to the cost of bringing a PTAB petition. This would not preclude reverse pay settlements altogether – and it would never prohibit licensing settlements – nor would it prevent any party (including a non-operating company) from taking a petition to final judgment. Rather, this would simply ensure that firms cannot earn a substantial living by doing nothing more than filing petitions and settling before a decision is reached.