Thursday, May 31, 2012

Constructing the “But For” World in Section 8 Damages

Teva Canada Ltd v Sanofi-Aventis Canada Inc / ramipril (s 8) 2012 FC 552 Snider J
Apotex Inc v Sanofi-Aventis Canada Inc / ramipril (s 8) 2012 FC 553 Snider J

Apart from the question of the compensable period, Teva / ramipril (s 8) and Apotex / ramipril (s 8) (discussed here and here) raised similar legal issues, and indeed similar factual issues, regarding the assessment of damages, as both cases turned on the construction of the hypothetical “but for” market for ramipril that would have obtained had the prohibition order not been granted. Common legal questions regarding the burden of proof, whether the generics could recover for lost off-label sales, and whether lost business value is recoverable, were also raised. Snider J’s analysis of the legal questions is the same in both cases, and the relevant discussion is often identical save to references to Teva or Apotex. For convenience, in this post I will cite paragraphs to the Teva decision, and refer Teva as the generic, except where there is a relevant difference, but it should be recognized that the same analysis is also provided in the Apotex decision.

Snider J assessed s 8 damages on the basis of the generally applicable principle of “but for” causation, which requires a comparison between the actual state of affairs and the court’s best assessment of what would have happened in the hypothetical world had the statutory stay not been triggered [5]. Generally, Snider J constructed the hypothetical world almost entirely as a matter of determining what would in fact have happened, without regard to various arguments that particular consequences should be ignored for policy purposes.

Multiple Entrants

An initial question in constructing the “but for” world was whether it should be assumed that Teva would be the sole generic, or whether the possibility of multiple entrants should be considered. Teva argued that Sanofi’s liability would be “capped inappropriately” if multiple entrants split the market in the “but for” world [119]. For example, a generic might have entered and shared the market had the order not been granted, yet it might not be liable for compensation if it had not filed a notice of allegation and had not been subject to a order of prohibition. On the other hand, if damages were calculated on the basis that Teva was the sole generic, Sanofi might be liable for more loss than actually caused, if for example, several generics had all been the subject of an order of prohibition [124].

Snider J held that the possibility of multiple entrants should be considered, as this would be consistent with the general principle that “damages are compensatory in nature” [117], and “seek to place a successful plaintiff in the position he or she would have occupied but for the defendant's wrong” [116]. That is, the “but for” world should be that world that would, in fact, most probably would have occurred but for the prohibition order. Snider J also pointed out that a generic is not entitled to a disgorgement of the innovator’s profits, and “[t]here is no requirement that Sanofi's total liability resemble the profits it earned on the sale of ALTACE during the period of the stay” [120]. In other words, while deterring inappropriate listing may be one aim of the NOC system, this cannot be understood as the sole or even the primary goal of s 8.

In my view, Snider J's conclusion is sound. It is not clear that the intent of s 8 is to deter the innovator from inappropriate listing, as opposed to compensating the generic for its loss. Further, it is not evident how an optimal deterrence system would be structured. For example, the NOC system provides more deterrence than the Hatch-Waxman system, in which there is no liability corresponding to s 8, but Hatch-Waxman does provide a 180-exclusivity to the first generic, which further shifts the incentive structure. The deterrent effect of the NOC system is a complex question that would make an interesting academic article, but in light of these uncertainties, it is best to adhere to the well-established "but for" principle of damages law, unless the text of the Regulations specifies otherwise.


If multiple entrants are to be considered, who bears the burden of establishing what the market would have been? Snider J as follows:

[141] Taking all of this into consideration, in my view, the proper approach is the following: Once Teva has led prima facie evidence of its losses, the evidential burden shifts to Sanofi to adduce evidence in response. Sanofi cannot simply allege that other generics would have entered the market without leading evidence in support of such assertions.

[142] In this case, Teva does not, at least initially, bear the burden of disproving the hypothetical sales of third party generics. However, Teva, at all times, bears the legal burden of proving its losses, and the evidence adduced by Teva must ultimately be weighed against any evidence adduced by Sanofi establishing sales by third party generics. To the extent that Sanofi succeeds in discharging its evidential burden by proving third party sales, Teva must address that evidence in order to discharge its legal burden.

On the facts, Snider J, held that Apotex would also have entered the market [163].

Single World?

Sanofi argued that not only should multiple entrants be considered, but also that a single “but for” world should apply to all s 8 claims for ramipril [124]. Sanofi argued that otherwise “damages would have ‘no relationship to the multi-generic market that actually exists’, and ‘would be proportional to the number of players in the market,’” and that "[i]f there [were] 10 generic entrants into the marketplace you could have 10 parties asserting a claim they would have been the sole generic" [124]. These objections are entirely met by Snider J’s holding that the “but for” world must consider multiple entrants. The remaining appeal of a single “but for” world is that of logical consistency; unless there is a single “but for” world, then it is possible that damages in different cases will be calculated on the basis of different hypotheticals. Indeed, this was demonstrated in these cases: in both cases Snider J held that the market would have been Apotex, an authorized generic, and Teva, but in Teva she held all would have entered the market on December 13,2005 [325], while in Apotex, Snider J that the same players would have entered in staggered manner, on April 26, 2004, July 26,2004 and August 1, 2006, respectively [Apotex 298]. Clearly, these scenarios cannot both accurately reflect what would actually have happened had no prohibition order been granted.

However, as Snider J pointed out, quoting Lord Shaw’s “broad axe” dictum, “s. 8 damages are hypothetical. It follows that estimates must be made and a market constructed that will not be perfect” [128]. That is, if only a single “but for” world were constructed, it might be logically possible to pretend that it was perfectly accurate, but that would be no more than an illusion. The only difference with multiple “but for” worlds is that the illusion of perfect accuracy is impossible to maintain. Further, as Snider J noted, the result on each case must turn on its own evidence [126]. It is difficult to see how, procedurally, a single “but for” world could be constructed without joining the cases. Given the illusory accuracy of a single world, there is no compelling reason to go to any great lengths to identify a single “but for” world.

Authorized Generic

Another important issue was whether entry by an authorized generic (AG) should also be considered. Apotex’s Dr Sherman raised policy concerns related to considering AGs as part of the "but for" world [178], again largely related to the deterrent effect. If entry by an AG is considered, the mere possibility of AG entry serves to deter generic entry by reducing the innovator’s s 8 liability. While Snider J shared these concerns, [178], [184], she nonetheless held that entry by an AG should be considered, again on the principle that damage should be assessed on what would have been likely to have happened in fact [178], [184]. This is particularly so because the Regulations as a whole contemplate the existence of AGs, and the 2006 RIAS specifically mentioned the problem, yet the amendments did not exclude AGs from s 8 [181]-[182]. If a departure from a strict fact-based “but for” approach is warranted in order to optimize the deterrent effect of s 8, that is a matter of the Governor in Council [184]. I share Snider J’s view; as noted above, the deterrence implications of departing from the general legal principle are too uncertain to be resolved through the case law. On the facts, Snider J held that Sanofi would probably have launched an AG.

Off-Label Sales

Snider J’s factual approach to constructing the “but for” world did not always work against Teva. Sanofi had separate patents covering the so-called“HOPE” indications. Teva did not address the HOPE patents, as it chose instead to withdraw those indications from its product monograph [311]. Snider J held that Teva was nonetheless entitled to recover for lost sales attributable to off-label indications, on the basis that in fact, such sales would likely have taken place [319], and the off-label prescribing by physicians is not itself illegal [314]. (This implies that it is at least an open question as to whether the losses would be excluded if the sales in question could be shown to be illegal, even if they were likely in fact to have happened.)

Lost Business Value

The qualification to Snider J's strictly factual approach to constructing the “but for” scenario concerned Teva’s claim for "lost business value" and "duplicate ramp-up period." Snider J held that while these losses were caused during the compensable period, they were incurred after that period. The claim for lost business value, for example, was simply a present value capitalization of the lost future profits [244]. Consequently, under the rule established in Merck Frosst Canada Ltd v Apotex Inc / alendronate (NOC) 2009 FCA 187, those losses were not recoverable as a matter of law [254].


At the end of the day, Snider J was not able to actually finalize a quantum of damages [324], because prior to resolution of the various legal questions there were too many possible "but for" worlds for the detailed calculations to be carried out in respect of each. However, it does appear that Snider J's decision would allow the parties to agree on the final quantum.

No comments:

Post a Comment