Relevance of Infringer’s Outside Option in Non-infringing Alternative Analysis

Apotex Inc v Merck & Co, Inc / Lovastatin Damages 2015 FCA 171 Dawson JA: Stratas, Boivin JJA aff’g 2013 FC 751 Snider J
            1,161,380 / lovastatin / MEVACOR

In Lovastatin Damages the FCA held that a non-infringing alternative should be taken into account in assessing damages for patent infringement if the alternative “could and would” have been used by the infringer. Yesterday’s post discussed the “could” branch of this test (which the FCA held was dispositive on its own [89]). This post reviews the “would” branch. The FCA held that even if the NIA was available to the infringer in the sense that it would have been technically feasible (the “could” branch), it should not be considered in assessing damages if the NIA would not have been used because there was some other product entirely that would have been more profitable than the NIA. This raises a tension between treating causation purely as a matter of establishing what would most likely have happened but for the infringement, and a purposive approach to the interpretation of s 55(1) of the Act.

As noted in yesterday’s post, I consulted for Apotex in this case at both the FC and FCA level. The views expressed in this post are my own.

The FCA’s discussion of the “would” branch is quite brief and somewhat cryptic, as the court listed a number of factors and then stated that they collectively led to the conclusion that Apotex had not established that it would have sold the non-infringing product in the "but for" world [90]-[94]. The most direct statement was:

Specifically, Apotex did not point to evidence that demonstrated the profits that it would have made through the non-infringing alternative would have been greater than value lost in any of the identified scenarios (for example, the research and development activities foregone by repurposing the Winnipeg facility) [94].

If I understand correctly, the point is that it may be that the NIA was so expensive that Apotex would have left the market entirely rather than compete using the NIA. As the FCA acknowledged, it was agreed that the NIA would have been profitable [94], so more precisely, the point is that even though Apotex had the capacity to make the NIA profitably, it failed to show that there was nothing else that it could have made even more profitably. If there was an outside option that it could have made even more profitably, then rather than competing with the AF-4 process Apotex would have abandoned the lovastatin market entirely and devoted its efforts to the other, more profitable endeavour. Because the burden is on Apotex to establish that it would have used the NIA, it must establish that competing using the NIA would have been at least as profitable as leaving the lovastatin market entirely, and it did not meet that burden [94].*

Leaving aside the facts of this case and the issue of the burden of proof, the fundamental point of importance going forward is that if an outside option would have been more profitable than competing with the NIA, then the infringer would not have used the NIA in the "but for" world and so the NIA is not relevant in assessing damages.**

While this reasoning is logical as a matter of what would in fact have happened but for the infringement, this result runs contrary to the purpose of considering the NIA in the first place. As the FCA explained:

Thus, in the event of infringement, under-compensation of an inventor discourages research and development, and the disclosure of useful inventions. Equally, over-compensation of an inventor chills potential competition to the extent that a potential infringer is uncertain about the scope and validity of a patent [42].

That is, from a public interest perspective, over-compensation is bad because we want the generic to have an incentive to challenge a patent which is probably invalid, so that society at large will benefit from the elimination of an unwarranted monopoly.***

For example, suppose the patentee has a process patent for producing a drug, which costs $1,000/kg. No one else has a process for producing the drug in commercial quantities, and the patentee sells the drug for $10,000/kg. The generic then develops a non-infringing process, which costs $4,000/kg. The generic believes that the patent is probably invalid. The generic has two choices: it can compete using the NIA or it can infringe and challenge the validity of the patent. In either case it will sell for $8,000/kg, but if it infringes it will make $2,000/kg more in profit. Suppose for the moment that either choice would be more profitable than the best outside option. The generic's decision will depend on the damages it will face if the validity attack fails (as well as the profit margin and market share it expects under both scenarios and the likelihood of success in the validity attack). If it infringes and it can raise the NIA in assessing damages, it will be liable for a reasonable royalty on the infringing process, which would be at most $3,000/kg. If it cannot raise the NIA argument, it will be liable for the patentee's lost profits of $9,000/kg. Obviously, it will be much less likely to challenge the validity of the patent if the NIA cannot be considered in assessing damages. Excluding consideration of the NIA chills potential competition as compared with considering the NIA.

Why is it preferable to consider the NIA? It might be said that the higher award is better, as it discourages infringement. The answer, as the FCA explained, is that “only by comparing the patented invention to non-infringing alternatives can a court discern the market value of the patent owner’s exclusive right, and therefore his expected profit or reward” [56]. In our example, the product in question can be produced by either the infringing or the non-infringing process. Even if the patent is valid, the true value of the patented process is only its cost advantage over the non-infringing process. If the generic launched with the NIA, the patentee could exploit its cost advantage by pricing its product at $4,000/kg, which would allow it to capture the entire market from the generic. This is exactly what we want; competition from the NIA forces the patentee to lower its price, the product ends up being made by the cheapest process, and the patentee’s reward is the difference between the cost of the patented process and the next cheapest process. On the other hand, if the generic launched with an infringing product and its attack on the patent fails, if it is required to pay only the difference in value between the infringing and non-infringing processes, then the reward to the patentee will be the same as if the patentee had itself sold at $4,000/kg, which means it will still be rewarded with the difference between the cost of the infringing process and the non-infringing process. And if the patent is invalid, the generic will have done a public service by destroying an unwarranted monopoly, with the result that all entrants will adopt the formerly patented process and prices will drop to something just over $1,000/kg. So, the advantage of considering the NIA in assessing damages is that the infringer is encouraged to challenge a patent that is potentially invalid, and even if the challenge fails, the patentee will receive a reward equal to the true value of the patent. On the other hand, if the infringer is required to pay $9,000/kg if its attack fails, the patentee’s reward will be as if it had a patent on the product itself. And if the prospect of paying those damages deters the generic from challenging the patent in the first place, prices will stay at $4,000/kg, even if the patent is invalid and the price should be $1,000/kg. By deterring a challenge to the patent, excessive damages can result in a patentee reaping a monopoly reward with an invalid patent. This is why, as the FCA explained, taking into account an NIA strikes the right balance of encouraging innovation without chilling socially desirable challenges to patents that are probably invalid.

Now consider how this is affected by the FCA’s requirement that the infringer show that the NIA is more profitable than the alternatives. Suppose we have a situation in which the generic would choose to infringe if and only if the NIA can be considered if its validity attack fails and it is required to pay damages. Under the FCA’s holding, whether the generic decides to challenge the validity of the patent will depend on the profitiability of its best outside option. Suppose the profit margin on the non-infringing process is 10%. The generic will challenge the patent if the profit margin on the outside option is 9%, but not if the margin on the outside option is 11%. Thus the generic's decision as to whether to challenge the validity of the patent will turn on the value of its outside option. The problem is that as a matter policy the damages should reflect the value of the patent owner’s exclusive right, which is the difference in value between the patented invention and the best non-infringing alternative [56]. By the same token, we want the generic’s decision as to whether to infringe to turn on that same difference in value. The value of the generic's outside option is unrelated to the value of the invention, and yet the effect of the FCA’s rule is that the generic’s decision to challenge the patent may turn on the profitability of the generic’s outside option. Whether the value of the generic’s outside option is 9.9% or 10.1% is irrelevant to the value of the invention, but under the FCA’s rule, it might make all the difference as to whether the generic decides to challenge a patent which it believes to be invalid.

This indicates a tension between two basic principles. On the one hand is the view that perfect compensation requires that the patent be put in the position that it would have been in but for the infringement. If in fact the generic would not have entered the market at all, for whatever reason, then this should be reflected in the damages award. On the other hand, the purpose of the Patent Act and s 55(1) in particular is to strike the right balance between encouraging innovation and avoiding the chilling effect of invalid patents [42]. While patent infringement is often likened to a species of tort for the purpose of assessing damages, it is ultimately a matter of statutory interpretation, which requires that s 55(1) be interpreted so as advance the statutory purpose. A strictly factual approach to "but for" causation is not sacrosanct: in Hamilton v Open Window Bakery Ltd 2004 SCC 9, the SCC departed from a strict consideration of what would in fact have happened but for the wrong, in favour of a competing principle which was applicable in the contract context. Similarly, if there is indeed a tension between a tort-like approach to causation and one which advances the purpose of the Act, it seems to me that the latter should be paramount.

*The FCA also noted that Apotex infringed intentionally [91], [93] (this was an inference drawn by the FCA, not a finding at trial); that Apotex believed the 380 patent was invalid [92], [93]; and that Apotex conceded in oral argument that “‘Brazen’ infringement in the real world makes it very difficult to prove that the defendant would have deployed the non-infringing alternative in the ‘but for’ world” [90]. If my understanding of the court’s main point (described above) is correct, then it's not clear to me how these points are relevant. One issue relates to the facts. “Brazen” usually means open and shameless, as when a generic concedes that its product infringes, and launches nonetheless, basing its defence solely on invalidity of the patent. On the facts in this case, however, Apotex’ infringement was not brazen in that sense, as Apotex strenuously denied infringement in the liability proceeding [131]-[466]. Perhaps “brazen” was being used to mean “intentional,” which would be consistent with the FCA’s emphasis on the intentional nature of the infringement. But whether “brazen” means “open” or “intentional,” I don’t see why it should normally be “very difficult” to prove the defendant would have deployed the NIA when infringement is brazen. As I understand the FCA’s reasoning, the question of whether the infringer would have used the NIA turns on the relative profitability of (1) infringement, (2) the NIA, and (3) the outside option. If they rank (1), (2), (3) in order of profitability, the defendant would use the NIA rather than exit the market, and if they rank (1), (3), (2), the defendant would exit the market rather than us the NIA. But so long as (1) is most profitable, the defendant will infringe intentionally, and if (1) is profitable because the defendant believes the patent is invalid, it will likely infringe openly. Intentional infringement is equally compatible with either scenario, so I don’t see how intentional or open infringement can support any inference as to whether the NIA or the outside option is more profitable. Certainly, infringement and a belief in the invalidity of the patent are both consistent with the (1), (3), (2) ordering, and perhaps that is why those factors were emphasized by the FCA. In that case, the defendant would have been thinking “It only makes sense to infringe because the patent is probably invalid, so I won’t be liable for damages.” But the same factors are just as consistent with the possibility that the NIA is more profitable than the outside option. That is particularly true if the defendant believes the patent is invalid. Apart from the possibility of being liable for damages, infringing will always be more profitable than the NIA or the defendant would have used the NIA in the first place. If the profitability ranks (1), (2), (3), the defendant would have been thinking “We have a NIA that we could launch that would be more profitable than anything other product in the pipeline, but infringing is even more profitable than that, so given that the patent is probably invalid, we will go ahead and infringe.” Thus, while the FCA emphasized Apotex’ intent and belief about the validity of the patent in concluding that Apotex would not have produced the NIA in the "but for" world, it seems to me that the conclusion more properly turns solely on the burden of proof. Alternatively, it may be that I have misunderstood the FCA’s main point.

**Strictly, it is better to say that it would have to show that the profitability of using the NIA is higher than its cost of capital. This does not affect the analysis in this post.

***More accurately, we want the generic to attack a patent if the cost of litigation is lower than the expected reduction in deadweight loss if the patent is invalid. That might well be true even if the patent is probably valid. But the point is most intuitive if the patent is probably invalid.