Sunday, February 1, 2015

" 'I Could Have Gone Down Metcalfe Street Instead"

Eli Lilly and Co v Apotex Inc / cefaclor, 2014 FC 1254 Zinn J [Cefaclor Damages]
            1,133,0071,146,5361,133,4681,150,725 [“Lilly Patents”]
            1,095,0261,132,5471,136,1321,144,924 [“Shionogi Patents”]

Zinn J’s Cefaclor Damages decision raises a couple of interesting legal issues related to causation, the first of which is the relevance of a non-infringing alternative to the assessment of damages:

Apotex urges the court to find that if there is a non-infringing alternative [NIA] to the infringing product or process that was available to the infringer in place of the infringing product or process, then, even though the infringer did not employ the NIA in the real world, it must be considered in the but-for world. I shall refer to this as the NIA Defence.[21]

The phrase “NIA defence” is certainly convenient, but it is something of a misnomer, as will be discussed below. Zinn J rejected the NIA defence as a pure matter of law [57]. Zinn J’s conclusion in rejecting the NIA defence was the same as that of Snider J in Lovastatin Damages 2013 FC 751 (blogged here), whom he agreed with [52]. An appeal in Lovastatin Damages was heard in mid-January, and presumably we will be getting the FCA’s view on this issue in due course. As Zinn J’s holding is likely to be overtaken by that of the FCA, I will not try to address every argument on this issue. Instead I will focus on what I see as the intuition behind Zinn J’s rejection of the NIA defence, as I suspect that his intuition is widely shared.

Apotex’s argument was that but for the infringement, it could have would have competed just as effectively with a non-infringing alternative. Lilly’s losses were therefore not caused by the infringement, because it would have suffered exactly the same losses in the but for world in which there was no infringement. Because Zinn J rejected the argument as a matter of law, he did not explicitly address whether Apotex could have established the NIA defence on the facts (though there was some evidence to that effect, as will be discussed my next post.) But by the same token, for the purposes of this analysis we can assume that in fact Apotex would have been able to compete just as effectively using a non-infringing alternative, for at least some of the infringing production.

The essence of Zinn J’s objection to the NIA defence seem to me to be captured in these paragraphs:

[32] I concur with Mr. Creber who said: "I am not aware of any Canadian case that has allowed a tortfeasor, the person who committed the tort, to pretend they could have acted differently whether that be personal injury, whether it be negligence, whether it be patent infringement. ... If I drove my car down Elgin Street and I hit somebody, it would be not open to me to argue, 'I could have gone down Metcalfe Street instead and would have avoided hitting the person'."

[35] In short, the causal connection must be examined in the real world. Damages arise if Lilly proves in the real world that, but for Apotex selling infringing product, it would have made some or all of those sales. The causal connection is not examined in the hypothetical world where the infringer engages in different conduct than that in which it actually engaged. Such an approach permits the wrong-doer to escape all responsibility for its conduct.

That is, the NIA defence would allow a wrongdoer to say “But for the wrong, I would have acted so as to avoid the wrong, and so I should escape all responsibility.”

While this argument, and particularly Mr Creber’s analogy, is intuitively appealing, it turns on a failure to distinguish between the liability and damages determinations. The argument “I could have gone down Metcalfe Street instead and would have avoided hitting the person,” is not open to the defendant in the liability phase. In the damages phase, arguments of this type are not only allowed, they are routine. As the SCC pointed out in Athey v Leonati [1996] 3 SCR 458, there is a “fundamental distinction between the way in which courts deal with alleged past events and the way in which courts deal with potential future or hypothetical events” [26]. “Hypothetical events (such as how the plaintiff's life would have proceeded without the tortious injury) or future events need not be proven on a balance of probabilities. . . By contrast, past events must be proven, and once proven they are treated as certainties. In a negligence action, the court must declare whether the defendant was negligent, and that conclusion cannot be couched in terms of probabilities.” [27-28]

So, in Mr Creber’s hypothetical tort case, in determining liability the defendant cannot argue that he might have gone down Metcalfe Street. Past events are treated as certainties; what actually happened, happened. The defendant is therefore held liable.

In the damages phase, the defendant is permitted to argue that but for the tort he would have gone down Metcalfe Street. But where does that get him? If the defendant makes that argument and establishes it on the facts, the court would find that but for the tort he would have gone down Metcalfe Street and consequently would not have hit the pedestrian. That means that in the but for world the pedestrian would have been perfectly healthy. Damages are the difference between the plaintiff’s actual position and the position she would have been in had the defendant acted properly: Athey [32]. In this hypothetical, then, damages would be assessed as the difference between the actual world, in which the pedestrian is injured, and the but for world in which the tortfeasor went down Metcalfe Street and the pedestrian consequently escaped injury. Damages would consequently be equal to the full extent of her actual injuries, as seems intuitively correct. The NIA “defence” would only allow a tortfeasor to “escape all responsibility” if it were permitted in the liability phase, which it is not. That is why it is not a “defence.”

For this reason, in the damages phase the “you could have driven down Metcalfe Street” argument is routinely invoked by the plaintiff, not by the defendant; the standard tort position is that but for the wrong, the defendant would have acted properly. Normally acting properly is taken to mean driving down Elgin Street while taking due care, but if on the facts, it would have meant driving down Metcalfe Street – if, say, Elgin Street was closed for construction, and the defendant’s wrongful conduct was to drive down a closed street – then that will constitute the but for world, and damages will be assessed accordingly.

In the damages assessment, this type of argument only allows the wrongdoer to escape liability if he can prove that the same harm would have occurred had he not acted wrongfully. This is why the driver in Mr Creber’s example would not escape. As we have seen, he is welcome to argue that but for his negligence he would have gone down Metcalfe Street, or driven more slowly or taken better care. The point of the damages assessment is exactly that in any of those but for scenarios, the defendant would not have suffered the harm he suffered in the real world. But when the wrongdoer can prove that the plaintiff would have suffered the same harm – if he can say “I could have gone down Metcalfe Street instead and would have avoided hitting the person, and the person would have suffered exactly the same harm because she had a pre-existing condition,” then the damages will be reduced or avoided entirely.

Indeed, as a technical point, all reasonable royalty damages cases in patent law are of the “I could have driven down Metcalfe Street” type. The theory of reasonable royalty damages is that had it not infringed, the defendant would have licenced, and consequently would not have committed the tort: as Lord Moulton said in Meters Ltd (1911), 28 RPC 157, 164 (CA) “Every one of the infringing articles might then have been rendered a non-infringing article by apply for and getting [a licence]”; and see similarly Lord Shaw in Watson Laidlaw (1914), 31 RPC 104, 120 (HL) stating that “The appellants did this Java trade improperly. Had they done it properly they would have done it under royalty. That royalty the respondents would have obtained.” Licensing is, in effect, a non-infringing alternative. It is true of course, that in that case some money ends up in the hands of the patentee, but the reason is just that we have to find some non-infringing course of action, and in most cases the most likely non-infringing course of action would be for the defendant to have licensed. But the principle is no different if the most likely non-infringing course of action would be for the defendant to have used some entirely non-infringing alternative. In the reasonable royalty context the patentee will not end up with as much money as if it had made the sale itself, and that is because the reasonable royalty is awarded when on the facts, the patentee would not have made the sale itself in the but for world. Whether the patentee would have made a royalty because the defendant would have licenced, or whether it would have made nothing because the defendant would have used an entirely non-infringing alternative, is not a difference of principle.

Zinn J was of the view that, in the contrary, there is a difference principle in the example of a reasonable royalty (and others):

[34] Each example is based on a fact relating to the conduct of the patentee or a third party not on a hypothetical and most certainly not on hypothetical behaviour of the infringer. It is the fact of the patentee's usual conduct in licensing its patents, the fact of the markets into which it sells, the fact of the patentee's distribution and marketing systems, the fact of the willingness of customers to purchase from the patentee, and the fact of the competitors in the real market that prevented the patentee from obtaining full recovery in the above examples.

That may be a distinction, but I do not see the principle. Perhaps the closest analogy to the non-infringing alternative is the crumbling skull rule, which says that “if there is a measurable risk that the pre-existing condition would have detrimentally affected the plaintiff in the future, regardless of the defendant's negligence, then this can be taken into account in reducing the overall award” (Athey [1996] 3 SCR 458 [35]). A reduction in damages as a result of the crumbling skull rule is rare, but that is only because it is rare on the facts, not because it is unsound in principle. The same is true of the NIA defence. If a different generic competitor had developed a non-infringing alternative and entered the market in competition with the infringing defendant and the patentee, and it could be proven that but for the defendant’s infringement, the competing generic would have undercut the patentee and captured all the sales made by the defendant, it is uncontroversial that the patentee would not be able to recover lost profit damages for the sales that would have gone to the generic competitor. What difference does it make if it would have been the defendant, rather than some other generic, that would have made those non-infringing sale?

From a policy perspective, the answer is none. The purpose of the Patent Act is to promote innovation: Free World Trust 2000 SCC 66, [42]; Harvard Mouse 2002 SCC 76, [40], [185]. It does so by providing a reward commensurate with the value of the invention. An invention that is much better than the closest alternative deserves a large reward, while an invention that is only a marginal improvement deserves less. This is the essential economic rationale for considering the non-infringing alternative in assessing damages. In the US context the non-infringing alternative is most commonly considered in the context of reasonable royalty damages, as setting the maximum price that an infringer would be willing to pay to licence. The idea is that any party, including the infringer, would not pay so much for a licence that using the invention would be less profitable than using a non-infringing alternative. In this way, the reward is made commensurate with the value of the invention. Consequently, it should be irrelevant whether the alternative would have been developed by the defendant or by some other party. The essential point is that the reward to the patentee must be commensurate with the benefit to society delivered by the invention, and an invention that is hardly better than the alternative deserves a smaller reward. That, essentially, is the policy argument for considering the non-infringing alternative.

Zinn J rejected this:

[56] If the court were in a position to ignore relevant statutory provisions enacted by Parliament and apply economic logic, then perhaps he and not I would be the decision-maker in this case. However, this is a court of law - not of economics. From the standpoint of economics it may make sense to consider that "but for the infringement, the defendant would have resorted to a noninfringing alternative;" however, it is not appropriate when assessing damages under section 55 of the Patent Act for all of the reasons previously expressed. The position Professor Cotter favours would require an amendment to the Patent Act by the Parliament of Canada. It is not within this court's jurisdiction to sacrifice laws Parliament has enacted written on the alter of economic logic.

With respect, this is contrary to the fundamental principle that a statute must be interpreted in light of its purpose: Rizzo Shoes [1998] 1 SCR 27 [21]; Interpretation Act s 12. It is very well accepted that the purpose of the Patent Act is to promote innovation, and the damages provision must be interpreted in a manner that is consistent with that purpose. The “economic logic” disparaged by Zinn J does exactly that.

Finally, on a somewhat technical point, Zinn J relied in part on Clements, 2012 SCC 32. In that case, Mr Clements was driving a motorcycle with his wife behind him as his passenger. Mr Clements was negligent in that the bike was 100 pounds overloaded and was being driven at 120 km/hr in a 100 km/hr zone. Unknown to him, there was a nail in the bike's rear tire. The nail dislodged, the rear tire deflated, and the bike crashed, leaving Mrs Clements with traumatic brain injury. While not disputing his negligence, Mr Clements argued that the accident would have happened anyway because the nail would have come out and caused a crash even if he had not been driving negligently. Relying on Clements, Zinn J stated [my emphasis]:

[31] The fallacy in Apotex' s submission is that casual connection must be determined based on an examination of the facts as they existed at the relevant time - not on those that could have existed. Just as was no defence for Mr. Clements to say that the nail could have fallen out of the tire before he increased speed and when he was riding a properly loaded bike, and therefore there is no causal connection between his negligence and his wife's injury, it is no defence for Apotex to say that it could have manufactured and sold non-infringing cefaclor and therefore there is no causal connection between its sale of the infringing product and Lilly's Jost sales.

With respect, this misreads Clements. It is perfectly clear that it was a defence for Mr. Clements to say that the nail could have fallen out of the tire before he increased speed and when he was riding a properly loaded bike: “it is open to the defendant to argue or call evidence that the accident would have happened without the defendant's negligence, i.e. that the negligence was not a necessary cause of the injury, which was, in any event, inevitable” [11]. Such evidence was in fact called [51], and so far as is apparent from the SCC decision, that was Clements’ only defence. The trial judge had found Clements liable using a material contribution test for causation. The SCC held that the but for test for causation should have been used, and sent it back for reconsideration of the facts on that basis [53]. If, as a matter of law, it was no defence for Mr Clements to say the nail would have fallen out even had he not been negligent, there would have been no need to send the case back for redetermination on the facts.


  1. Interesting post, as usual.

    It would seem that the courts do make a big distinction between whether the NIA is coming from a third party vs. the defendant; presumably, for the "public policy" reasons listed by Justice Snider in lovastatin. On a simple level, reducing damages to the defendant's incremental profit from infringement creates a sort of "catch me if you can" scenario; ignoring the NIA provides a deterrent.

    Love the blog, I hope you will be posting on the prejudgment interest aspect of this case, which (while possibly misguided in its conclusion) should hopefully open the door a little wider for a rational discussion of the issue in Canada.

  2. Zinn J did raise a version of the "catch me if you can" argument, especially at [35], and I agree it seems to have played an important role in his thinking. I may come back to it in a future post - after I blog on the issue of prejudgment interest.

    Thanks for the compliment!