Wednesday, January 6, 2021

Lear v Adkins Should Not Be Followed

Loops LLC v Maxill Inc 2020 ONSC 5438 Lederer J: Swinton, Penny, JJ revg 2020 ONSC 971 Templeton J

            2,577,109 / Toothbrush

In Monday’s post I reviewed Lederer J’s decision in this case, holding that a no-challenge clause in a settlement agreement is enforceable, notwithstanding the decision of the US Supreme Court in Lear Inc v Adkins 395 US 653 (1969) which abolished the doctrine of licensee estoppel in US law. In Tuesday’s post I explained that the doctrine of licensee estoppel is not based on estoppel at all, but is based entirely on contract law. Absent an express agreement to the contrary, the licensee’s obligation to pay royalties will not normally be contingent on the validity of the patent. This is partly because the courts will not normally imply a warranty of validity into the agreement and partly invalidity will not normally constitute complete failure of consideration, but might instead give rise to a counter-claim for damages. In this post I will step back and consider the bigger question: as a matter of policy, should Lear v Adkins be followed in Canada? Recall that while Lederer J was of the view that holding a no-challenge clause in a settlement agreement to be unenforceable was a leap too far [64], but he left open the possibility that it might be adopted in more incremental steps. That is the question I will focus on.

The first question is what it would mean to follow Lear. Most obviously, the courts might simply hold that a licensee is not estopped from disputing the validity of the patent. But such a holding would have no effect. A licensee would be permitted to plead invalidity, and introduce evidence going to validity, and argue invalidity; but if, after all that, the licensee was successful in establishing invalidity, the court would say something to the effect of “That’s all well and good, but it is immaterial to your obligation to pay under the contract.”

The next most conservative step would be to change the default rule, so that absent language to the contrary, the courts would imply a warranty of validity that would give rise to a counter-claim in damages. This is a fairly modest change in the law which is reasonably within the normal capacity of the law to evolve incrementally.

But that only makes such a change feasible; it does not mean it is necessarily desirable. There are some reasonable objections even to such a modest change. The current rule has been established for over a century and changing the rule now might disrupt the expectations of parties who have drafted their agreement in light of it. Secondly, the current rule is not really a default rule as such, but rather a tendency arising from the recognition that the licensee is not necessarily in the best position to assess validity. The most accurate statement of the law is that “Courts are not prepared to readily extend the meaning of doubtful words so as to convert them into an express warranty of the validity of a patent” Rymland (1966), 51 CPR 137 (Man CA), 141. The effect of changing the default rule would be to introduce an opposite tendency, that would have effect only in marginal cases. So, changing the default rule would create uncertainty without effecting much change.

More importantly, if only the default rule were changed, the parties could and likely would contract out of it, simply by using terms that are sufficiently clear that payment is not contingent on validity, or by using a no-challenge clause. There are various reasons, discussed below, why parties would often mutually agree to such terms. And in any event, to the extent that the parties generally prefer to include terms that absolve the licensee of its obligation to pay if the patent is invalid, then there is no problem to solve in the first place. The whole question of whether we should follow Lear, and how far, is only significant in cases in which the parties would not contract for invalidity to be a defence. Otherwise we are only dabbling at the margins where the default rule governs because the parties are not sophisticated enough to think of addressing the point in their contract.

So, to follow Lear it would be necessary to hold that there is an implied warranty of validity, and that the parties are not free to exclude it, so that any term to the contrary, such as a no-challenge clause, would be unenforceable. So, in Lear the US Supreme Court did in effect impose a term that trumps contract law. Adkins argued that the terms of the agreement obliged Lear to pay until such time as the patent had been held invalid, and that Lear was obliged to comply with its contract (673). The US Supreme Court rejected this on the basis that the contract was not controlling and the payment clause was not enforceable, in light of the “the strong federal policy favoring the full and free use of ideas in the public domain” (674).

This is in itself a departure from the normal incremental evolution of the law, at least so far as contract law is concerned. The general rule in Canadian law is that “the Courts will give effect to the intentions of the parties as clearly expressed by contract” (Louiseville Spinners (1972) 7 CPR(2d) 18 (Que CA) 20). The courts are loathe to impose mandatory terms on the parties that cannot be altered by agreement: Bhasin v Hrynew 2014 SCC 71 [74]–[75], imposing a duty of honest performance, is the exception that proves the rule. (Note that this was not directly raised in Loops v Maxill, which concerned a motion for an interlocutory injunction. An interlocutory injunction is inherently discretionary and it is well-established that the public interest may be taken into consideration.)

A second problem is that an implied warranty is not always appropriate on the facts. Suppose a small research based firm develops and patents a new invention and then grants an exclusive licence to a large established firm for development and commercialization. In such a case, the licensee may well be in a better position to know whether the invention is novel and non-obvious, and the burden of that risk would appropriately be placed on the licensee, as the early courts recognized. (Of course, in some cases the risk is better placed on the licensee, and under current law the parties are free to agree to that effect.) Thus, the law would have to impose a non-negotiable warranty into every licence agreement, knowing that it is not always appropriate to the actual relationship, on the view that the overall benefit from encouraging patent challenges more than outweighs the harm from imposing inappropriate terms on a relationship.

A third problem is that even a mandatory warranty allowing the licensee to counter-claim for damages would have relatively little effect in encouraging patent challenges, as damages will typically be zero. Lear was a fairly typical scenario in which the licensee was seeking to establish the invalidity of the patent as a defence to an action for royalties owing. In Lear and similar cases, the patent had not yet been held invalid, and so the licensee will have had the benefit of exclusivity in fact, even if the patent turns out to have been invalid all along. A counter-claim for damages based on a warranty of validity may be a live issue where the patent has been infringed by a third party and so the licensee has suffered a real loss if the patent is invalid; but in that case it will not have been the licensee who challenged the patent, and the point of Lear is that it is the licensee who has the greatest incentive to challenge validity. (This situation is more commonly dealt with by a clause that the licensor agrees to defend the patent.) If the patent has been held invalid as a result of a third party action, then it may seem unfair if the licensee has to continue paying royalties, but so far as the rationale in Lear is concerned, the important point is that the patent has been invalidated.

Now, the licensee may have an incentive to raise invalidity in order to avoid royalty payments going forward. But we don’t need Lear to provide that incentive; it is enough if the contract contains a unilateral termination clause allowing the licensee to terminate at will. Then, if the licensee discovers potentially invalidating prior art, it can terminate the licence, continue to use the invention, and defend the infringement action on the basis of invalidity. The contract at issue in Lear allowed Lear to terminate and stop paying royalties, but only if it also stopped using the invention: Adkins v. Lear, Inc., 67 Cal2d 882 (1967) 904-07. We could have a discussion about whether a unilateral termination clause should be implied into every licence agreement, and if so, whether it should be mandatory, but that’s a different discussion. (Note that in Brulotte v Thys Co 379 US 29 (1964) the US Supreme Court held a contract requiring post-expiry royalties is unenforceable; there is a widespread view among academics that Brulotte was wrongly decided.)

The next step would be to imply a mandatory guarantee of validity, so that if the patent is invalid, the licensee would be able to continue using the invention, without paying royalties, and, presumably, would be entitled to repayment of any royalties that were already paid. (Note that in Lear itself the licensee had started withholding payment early in the term, and the holding was that Adkins was not entitled to damages for breach of contract.) This would certainly provide an incentive to challenge validity. If the licence ran for the term of the patent, say 15 years, and the licensee used the invention and paid royalties for the first 13 years, then discovered invalidating prior art, it would be able to get repayment of all its royalties.

The difficulty is that this runs into the second traditional rationale for the licensee estoppel rule, namely that invalidity does not usually constitute complete failure of consideration. Even if the patent is invalid, the licensee will have had the practical benefit of market exclusivity, as well as the benefit of any transfer of know-how; after all, the fact that the patent is invalid for anticipation, for example, certainly does not mean that the licensee could have come up with the invention at minimal cost. Even apart from know-how, once the licensee has been told the solution to its problem by the licensor, it may be able, with the benefit of hindsight, to find the prior art that discloses the same solution; it is quite another thing for the licensee to identify which of the myriad prior patents is the one that actually provides a solution to its problem. So, in the early English case of Taylor v Hare (1805) 127 ER 461 the licensee had practised the invention for 6 years before discovering prior art said to anticipate, and sought on that basis to recover all the royalties that had been paid to date. The court rejected the action, with Heath J stating at 462 “It might as well be said, that if a man lease land, and the lessee pay rent, and afterwards be evicted, that he shall recover back the rent, though he has taken the fruits of the land.”

This is not farfetched. The story in Lear started in the 1950s as the faster airplanes of the post-war boom required more accurate gyroscopes. The method used by Lear for producing accurate gyroscopes was expensive, requiring precision work by a skilled machinist, and even so there was a high incidence of rejects (67 Cal2d at 894-95). Developing a more economical production method was a “vexing problem” (655) that Lear’s engineers had not been able to solve. Shortly after Adkins was hired he developed a method of construction which improved accuracy at a low cost. “Lear almost immediately incorporated Adkins' improvements into its production process to its substantial advantage” (655). Adkins applied for a patent and at the same time entered into negotiations with Lear. In 1955 the parties entered into a licence agreement, after extensive negotiations in which both sides were represented by counsel (67 Cal2d at 897). Around 1957, when Adkins’ patent application was still being prosecuted, Lear had done a search of the patent prior art and concluded that Adkins invention was not novel. Lear thereupon refused to pay royalties. Lear no doubt believed that it could have or even did come up with an equally good solution on its own (67 Cal2d at 895, 907), but hindsight is 20/20, and the fact is that Lear continued to use the method developed by Adkins and it also infringed the patent that ultimately was granted (67 Cal2d at 915).

So, whether or not the granted patent was valid, it is perfectly clear that in fact Lear gained a “substantial advantage” from the use of Adkins’ method. This was not a case of partial failure of consideration, as for example if the licensee had been relying on the patent for market exclusivity, and the patent had been held invalid with the result that a competitor entered the market. Lear had received a substantial advantage both in terms of technology transfer and market exclusivity (as the patent had not actually been held invalid even at the time of the US Supreme Court decision). Indeed, it had received the full benefit of the contract it had bargained for, as the Supreme Court of California had held, in a holding that was not challenged on appeal.

Now, it might be said that this result is too extreme, and Canadian courts, under some doctrine or other, perhaps quantum meruit, would ensure that the inventor was rewarded for the actual value of its contribution. But that would have the same effect as a warranty of invalidity, allowing the licensee to counter-claim for damages, and we are back to the point that a warranty does not provide a substantial incentive to challenge the patent, along with the various other problem outlined above. Moreover, we must recognize that the result in Lear itself was extreme: notwithstanding that it had received the full benefit that it had bargained for, in the result, Lear was not required to pay any royalties for the period after the patent was granted.

So, following Lear, even incrementally, by abolishing licensee estoppel in the context of licence agreements, not settlements, where there is no express “no-challenge” clause, would require a major change to Canadian law, by effectively imposing a mandatory clause in a contract, which the Canadian courts are loathe to do. Moreover, we know that this kind of mandatory clause would be inappropriate for some parties. We also know that it would work a serious injustice in cases in which the licensee would have received the full benefit of its technology, and the licensor would nonetheless be denied the reward that the licensee had expressly agreed to give in exchange. I would suggest that there would have to be a very powerful countervailing argument to follow Lear, even to that extent.

What, then, is the argument in favour of following Lear? Remarkably, the US Supreme Court itself did not articulate any clear rationale. It spoke generally of “the important public interest in permitting full and free competition in the use of ideas which are in reality a part of the public domain” (670), as well as “the demands of the public interest” (670), “overriding federal policies” and “the aims of federal patent policy” (673). As justification for overriding freedom of contract, the Court referred to “the strong federal policy favoring the full and free use of ideas in the public domain” (674); it is this phrase which is perhaps most often relied on.

With respect, these are handwaving arguments, not much better than “information wants to be free.” The most specific statements as to what the overriding policy actually entails are the following (668):

On the other hand, federal law requires, that all ideas in general circulation be dedicated to the common good unless they are protected by a valid patent.

First, this doesn’t support the result in Lear. The invention at issue may have been in the public domain, but it certainly was not “in general circulation.” The mere fact that an idea can be found in a patent application in some patent office somewhere does not in any way imply that it is generally known and used. As the Court itself noted “By promising to pay Adkins royalties from the very outset of their relationship, Lear gained immediate access to ideas which it may well not have learned until the Patent Office published the details of Adkins' invention in 1960" (672). And realistically, unless Adkins had contracted with Lear, there is no particular reason to believe that Lear would have learned of Adkins ideas even when his specification was published; even if Lear had been scanning the published patents, many patented inventions, including valid ones, are not commercializable, and it is not self-evident that Lear would have recognized that Adkins’ method was actually an effective solution to their problems.

The Court also referred to (658)

the strong federal policy favoring free competition in ideas which do not merit patent protection.

That’s fine as a principle, but how far does it extend? Suppose I have an idea for a process which I am protecting as a trade secret. I have no intention of seeking a patent. I licence it to you, on a confidential basis. You incorporate it into your manufacturing process to your great advantage. With the information I had disclosed to you, you were able to find an Italian patent application from 20 years ago that anticipated my idea; that application had been abandoned and never commercialized, and without the knowledge derived from my transfer of know-how to you, you would never have discovered it. Is our contract enforceable or does some “strong” policy mean that you can use my idea without paying me the agreed price, given that my idea does not merit patent protection? That such a contract is unenforceable is a direct implication of Lear; whether or not the invention was patentable, Lear did in fact receive the full benefit it had bargained for, but the Supreme Court nonetheless held that Adkins was not entitled to the reward specified by the agreement, in light of what the Court viewed as the overriding policy favouring free competition in ideas “which do not merit patent protection.” If that is the policy, it is difficult to see why it should matter whether or not the licensor had actually applied for and been granted a patent, if the idea did not merit a patent, ie if the invention was not in fact patentable. 

The Court in Lear recognized this problem, but avoided it on the basis that it was a matter for state law. The Court recognized that its holding in Lear might “revolutionize” the law of trade secrets, but it declined to define “in even a limited way the extent, if any, to which the States may properly act to enforce the contractual rights of inventors of unpatented secret ideas” (675). The only authorities the court relied on to elucidate the meaning of this policy were Sears, Roebuck & Co v Stiffel Co 376 US 225(1964) and Compco Corp v Day-Brite Lighting, Inc 376 US 234(1964). These cases stand merely for the proposition that the states cannot set up an intellectual property regime that supplants the patent regime, but with a longer term of protection. Whatever the merits of that proposition as a matter of federalism, it does not lead ineluctably to the conclusion that revolutionizing trade secrets law is in the public interest.

Thus, not only would following Lear, even in the most closely analogous case, require a dramatic change to the law of contract, it would also require us to entirely reconceptualize the law of trade secrets, somehow tying the ability of parties to licence confidential information to the patentability of that information.

Thus the Court in Lear does not give us any sustained analysis to support its conclusions. Let me see if I can do any better. Ultimately the Court’s reasoning turns on two points. First, invalid patents are bad. If we’re going to stop short of saying that all contracts relating to IP are unenforceable unless the IP is patentable, this seems to be the main implication of the Court’s insistence on the importance of free flow of ideas. With this point, at least, I entirely agree. Second, the licensee is the party with “the strongest incentive to show that a patent is worthless” (672). I expect that is probably true, though the point is not quite as clear. Even an invalid patent may give market exclusivity until someone risks the time and money to challenge it. This allows the vendor to charge a higher than competitive price. A licence usually splits those rents between the licensor and licensee. If the patent is invalidated, the licensee will not have to pay royalties, but it will lose the rents from market exclusivity. So it is not clear in general that the licensee has a strong incentive to challenge the patent validity. But my argument doesn’t turn on this. For the present purposes I’m willing to assume that the licensee has the strongest incentive to challenge validity. This is the assumption that provides the strongest argument in favour of Lear.

So far as I can tell, these are the premises on which Lear rests: bad patents are bad, and the licensee is the party with the strongest incentive to challenge the validity. Taking these premises, does it follow that the rule in Lear is sound? Remarkably, the Court in Lear never addressed this question. The closest it came is the following conclusory assertion (670-71):

We think it plain that the technical requirements of contract doctrine must give way before the demands of the public interest in the typical situation involving the negotiation of a license after a patent has issued.

Well, it’s not plain to me.

The problem is that there is a benefit to the Lear policy—more challenges to bad patents. But there must be some cost to running roughshod over freedom of contract. The Court’s statement simply asserts that the benefit outweighs cost, without giving us any argument to support that assertion. So far as I can see, in the absence of any analysis by the Court on this point, the answer is not clear. Consequently, there is no compelling rationale for following Lear, given the clear drawbacks of doing so.

To see why, compare Option (1), under which the parties are permitted to include an enforceable no-challenge in the licence, with Option (2) (ie Lear), where a no-challenge clause is prohibited, in the context of a simple model where Patentee and Licensee negotiate a licence agreement when the validity of the patent is uncertain, and the parties know that if L does not challenge the patent, no one will challenge. There is nothing to choose between these options if the parties choose not to include a no-challenge clause—in that case the Lear rule doesn’t make any difference—so the difficulty arises if the parties do include a no-challenge clause under Option (1).

In the case where the parties would include a no-challenge clause under Option (1), then the royalty rate will be lower under Option (1) than under Option (2), where they are prohibited from doing so. This is because ex ante, for any given expectation of validity, the minimum return P is willing to accept is the same regardless of whether the clause is included, and since the return is riskier without a no-challenge clause, P will demand a higher rate as compensation. Some of that benefit in the form of a lower rate will be passed on to the public; even a monopolist will lower its price if its costs are lower, and the same is true under monopolistic competition. The parties’ may or may not choose to include a no-challenge clause, depending on their risk-tolerance and relative assessment of the likelihood of invalidity.

From the public point of view, Option (1) means a lower royalty rate and therefore lower product prices and lower social deadweight loss, at least initially. Option (2) means a higher initial product price and therefore a higher deadweight loss relative to Option (1)—until such time as the patent is successfully challenged and the royalty is eliminated entirely, resulting in a lower product price and lower deadweight loss than under Option (1). Whether Option (2) is better than Option (1) from a societal perspective reduces to the question of whether, on average, the higher deadweight loss prior to a challenge more than outweighs the lower deadweight loss after a challenge. We also need to take into account the fact that the benefit in Option (2) from a lower deadweight loss post-challenge is a future benefit that has to be discounted accordingly–if L had reason to believe it could challenge immediately it would not enter into the licence agreement at all. I cannot see any particular relationship between the two that would allow us to answer that question. The lower the probability of invalidity is low, the less the advantage of Option (1) from lower initial deadweight loss, but also the less the expected advantage from Option (2) of eventual much lower deadweight loss.

Thus, even though the patent will not be challenged under Option (1), this does not mean that the deadweight loss from Option (2) is necessarily lower than Option (1).

While this reasoning may seem complex, it is simply an application of the basic argument in favour of freedom of contract. Restricting freedom of contract is prima facie bad because it reduces gains to trade between the parties, and no matter what type of market the parties are in, some of those gains will be passed on to consumers.

So what about the policy relating to free flow of ideas? This points to the fact that there is an externality to bad patents—much of the cost of unjustifiably restricted ideas is borne by the public, rather than by the parties to the licence. This is why, as discussed above, the licensee might prefer not to challenge a bad patent. Option (1) is clearly better than Option (2) from the parties’ perspective, because under Option (1) they will include a no-challenge clause only if it is mutually beneficial. Because of the externality, we can't infer the same is true from a social perspective; but the fact that we don’t know that Option (1) is socially better, doesn’t mean it is socially worse. This is because Option (2), prohibiting a no-challenge clause, doesn't cause L and P to internalize the externality, either when they bargain initially or when L makes the decision to challenge the patent. If prohibiting a no-challenge clause caused the parties to internalize the externality, then there might be some kind of general justification for Option (2). But it doesn’t so there's not.

Now, there are no doubt specific situations where no-challenge clauses may be undesirable. The most prominent is the reverse payment settlements that have been the subject of scrutiny under Actavis 570 US 136 (2013). But even if no-challenge clauses are undesirable in that specific context, that doesn’t imply that no-challenge clause should never be enforced. It means they sometimes shouldn’t be enforced. This is illustrated by Actavis itself, where the US Supreme Court held that even in that context, reverse payment settlements are sometimes permitted, depending on their particular terms. Even the FTC was arguing that an agreement not to challenge the patent should be presumptively illegal only in this particular context, not as a general matter. That prohibiting no-challenge clauses is against the public interest in some cases reinforces my point that it is not against the public interest in all cases, which is what Lear would have us believe. Actavis illustrates another important point; the desirability of no-challenge clauses turns on a fairly detailed economic analysis of the consequences. This is a matter that is within the purview of competition law, which has trained economists and specialized tribunals to deal with such arguments. Competition law is the place to address concerns related to no-challenge clauses, not contract law.

Put another way, following Lear amounts to a rule that no-challenge clauses are per se illegal as a matter of competition law. Per se illegality under competition law is appropriate only when we know that the rule will essentially always have anticompetitive effects, as for example price-fixing conspiracies. But the US Supreme Court in Actavis has held that no-challenge clauses are not even per se illegal even in the problematic context of pay-for-delay agreement. Now, I don’t want to rely on the US Supreme Court as an authority on economic reasoning, so I will also point out that much of the development of competition law has been a shift away from per se rules, towards a “rule of reason” approach, that recognizes that some types of agreements may be harmful in some cases but beneficial in others. That is exactly the approach I am arguing we should take to no-challenge clauses, and by extension the licensee “estoppel” rule. To the extent that a no-challenge clause is socially harmful, we should address it by means of competition law, which has the institutional competence to undertake the necessary economic analysis, and not by means of a sweeping restriction on freedom of contract, underpinned by nothing more than some hand-waving statements of policy.

In summary, maybe the rule in Lear gives a net benefit to the public, and maybe it doesn’t. Whether it does depends on a fairly intricate chain of economic reasoning. Perhaps sustained economic analysis could give a definitive answer, but I haven’t seen it. The idea that “maybe” there will be some public benefit from Lear is not a sufficiently compelling reason to follow it, in light of the dramatic change to the law that would be required, and the clear drawbacks including foisting contract terms on parties that are not in their private interests and the inequity to the licensor that would certainly be entailed, as in Lear itself.

Lear is not a precedent that we should follow, even in incremental steps. Rather, it is a cautionary tale; it is egregious example of the perils of judicial activism undertaken on the basis of a ill-defined policy considerations in an astonishingly thinly reasoned decision.

Additional reasoning (my thanks to Professor Dmitry Karshtedt for directing me to these articles)

Rochelle Cooper Dreyfuss, Dethroning Lear: Licensee Estoppel and the Incentive to Innovate,72 Va L Rev 677 (1986) (arguing that Lear severely restricts the ability of small, specialized, research firms to deal with risk)

Rochelle Cooper Dreyfuss†and Lawrence S. Pope, Dethroning Lear? Incentives to Innovate After MedImmune, 24 Berk Tech LJ 971 (2009) (discussing MedImmune, which extended Lear by allowing licensees in good standing to bring declaratory judgment actions, and describing MedImmune as “ahistorical, convoluted, and blind to the realities of licensing practice”).

Stephen Yelderman, Do Patent Challenges Increase Competition? 83 U Chi L Rev 1943 (2016) (arguing that the circumstances in which patent challenges increased competition are very limited)

The strongest argument against no-challenge clause I am aware of is Alan D. Miller & Michal S. Gal, Licensee Patent Challenges, 32 Yale J Reg 121 (2015), arguing that any procompetitive justifications for no-challenge clauses are “generally” outweighed by their anticompetitive effects (156). Their substantive analysis of the effect of no-challenge clauses is full of estimates: for example “However, in patent challenge cases the costs of litigation may well be outweighed by benefits to society” (138); “The cost of this uncertainty is likely to be of second order when compared with the large social benefit that arises from the increased level of competition that follows the invalidation" (141). Even if they are ultimately right in these assessments, these assessments of the harmful effect of no-challenge clauses are too speculative to justify the courts in abrogating freedom of contract and overturning long-standing precedent. And indeed, Miller & Gal argue that “the legislative and executive branches should be primarily responsible for setting the optimal balance between these areas of law” (105).

No comments:

Post a Comment