This second post on
the Court of Appeals for England and Wales’ recent decision in Lufthansa Technik AG v. Astronics
Advanced Electronic Systems,
[2026] EWCA Civ 964,
will focus on Lord Justice Arnold’s analysis of the lower court’s award of a
portion of the profits earned by defendants Astronics, Panasonic, and
Safran. For the facts of the case, see
my first post in this series, published last Thursday.
The principal issue
under consideration, as the court states, is how to “distinguish between ‘the
profits derived . . . from the infringement’ and profits which are not ‘derived
. . . from the infringement’” (para.
78). To answer this question, the
opinion sets out to analyze the relationship between a “differential profits”
approach and apportionment. Paragraphs
80 through 82 state the parties’ competing positions:
80. Lufthansa contends that, in most cases, the
correct way in which to distinguish between “profits derived … from the
infringement” and profits which are not “derived … from the infringement” is by
ascertaining the difference between the profits which the defendant made from
its infringing activities and the profits which the defendant would have made
from the nearest available NIA [noninfringing alternative]. This type of
analysis is most commonly referred to in the case law and the academic
literature as a “differential profits” analysis, although other terms (such as
“incremental profits”) are sometimes used. Once a differential profits analysis
has been undertaken, Lufthansa contends that there is no legal or economic
justification for applying an apportionment of the profits: these are alternative
approaches to the isolation of profits caused by the infringement. Lufthansa accepts
that there may be cases in which apportionment is appropriate, but it argues
that a differential profits analysis is usually preferable. What is not
legitimate is to apply both. Lufthansa also accepts that, having carried out a
differential profits analysis, it remains necessary to ask whether any of the
resulting profits were too remote (in the broad sense) from (i.e. not legally
caused by) the defendant’s acts of infringement. Lufthansa argues that this
addresses the problem of long chains of causation, and does not justify
apportionment.
81. Lufthansa contends that, since the judge found
that the Defendants had failed to establish any NIA, the result of the
differential profits analysis is that all of the profits in question “derived …
from the infringement”. The Defendants’ failure to establish an NIA
demonstrates the Patent was a “gateway” patent which controlled access to the relevant
market. Accordingly, Lufthansa says, it is just that Lufthansa should recover all
of the profits generated as a result.
82. Astronics and Panasonic contend that apportionment
is a well-established approach in English law to determining what profits are
“derived … from the infringement”, and that the use of apportionment is not
precluded by a finding that all of the profits in issue were factually caused
by the infringing acts. . . . It would be unjust for Lufthansa to recover all
of the profits in issue, because there were multiple causes for the generation
of those profits. Although infringement of the Patent was necessary for those
profits to be generated, it was not sufficient.
The opinion then
spends several pages (paras. 84-132) discussing the relevant common-law
precedents, including (among many others) Celanese International Corp v BP
Chemicals Ltd [1999] RPC 203, a case I discuss and critique at some length
in my book Comparative Patent Remedies, and Nova Chemical Corp v Dow
Chemical Co [2022] SCC 43 (Can.), a decision that Norman Siebrasse
discusses and critiques in two papers, Nova v. Dow: Intuition or Principle
in the Accounting of Profits Remedy, Part I, 35 I.P.J, 249 (2023); Norman
V. Siebrasse, Nova v. Dow: Intuition or Principle in the Accounting of
Profits Remedy, Part II, 36 I.P.J. 81 (2023). The opinion also discusses and critiques some
academic works that Siebrasse and I, among others, have either authored or coauthored
(paras. 133-42). The opinion’s principal
critique of some of this work is that it doesn’t answer the question of how a
court should proceed when the evidence does not indicate what the NIA would
have been. (Later in the opinion, as
noted below, the opinion rejects the argument that courts should resolve this
issue simply by allocating the burden of production or burden of proof to one
party or the other.) I commend all of
this to readers’ attention, but in the
interest of conciseness I will cut to Lord Justice Arnold’s resolution of the
issue, which begins at paragraph 143:
143. Analysis of the law. As is common ground,
factual causation is not enough on an account of profits any more than it is on
an inquiry as to damages. The reason is simple: factual causation proves too
much. Applying the usual “but for” test of factual causation, the infringer may
have made profits which it would not have made but for the infringement, but
which are not in truth derived from the infringement as opposed to other
factors. . . . Accordingly, as is also common ground, legal causation is
required as well as factual causation. . . .
145. Lufthansa argues that a correct application of
the “but for” test of factual causation involves the identification of a
counterfactual world in which the infringements did not take place. That
requires identification of the nearest NIA available to the defendant. The
difference between the profit the infringer in fact made and the profit the
defendant would have made had it adopted that NIA (i.e. the differential
profit) represents the profit derived from the infringement. On this argument,
a differential profit analysis identifies with precision the profits
attributable to infringement as opposed to other factors. Thus it accounts for
the causal potency of the different factors. Accordingly, the only role for
legal causation is to police the length of the causation chain, and to exclude
profits which are too remote from the infringing acts (e.g. profits made by reinvestment
of the profits from the infringing acts into a cryptoasset which performs very
well). . . .
Lord Justice Arnold,
however, expresses agreement “with Astronics and Panasonic that the ‘but for’
test of factual causation does not necessarily require the identification of a
specific NIA,” and “that the role of legal causation in this context is not
limited in the manner contended for by Lufthansa” (para. 148). In my view, the next several paragraphs are
the most important portion of the opinion, so I’m going to quote them with only
minimal editing for concision.
149. As Astronics and Panasonic accept, there may be
cases in which a differential profit analysis is useful for this purpose. As
they submit, however, differential profit analysis is fraught with difficulty. .
. .
150. The reason why differential profit analysis is
fraught with difficulty is that, in order to identify with precision the
profits attributable to the infringement, as opposed to other factors, it is
necessary to identify a counterfactual in which all other factors are held constant
and the minimum change is made to ensure that the patent is no longer infringed,
so that the economic impact of the infringement can be isolated from the economic
impact of other factors. In theory, this should present no difficulty. In the
real world, the opposite is true.
151. The first question is who bears the burden of
identifying and proving the NIA. The Canadian courts have held the burden rests
on the defendant, but it is not clear to me why this should be so. One could
argue that the burden should lie on the claimant, since the claimant is
claiming the profits derived from the infringement, and if it relies upon differential
profits analysis to quantify those profits, then the claimant must prove the NIA
which should be used for that purpose. A potential difficulty with that
approach is that it would enable the claimant to skew the differential profits
analysis by selecting a very unprofitable NIA. As I understand the
jurisprudence of the US courts on this question, they apply a shifting burden
of proof under which the claimant must first identify an NIA, and then the onus
is upon the defendant if it wishes to rely upon a different NIA as being a
better one. That is a principled approach, but in many cases it will lead to an
evidential burden on both parties.
152. If the defendant bears the burden of proof either
in full or in part, the next question is what happens if the defendant proposes
an NIA which the claimant contends is not an NIA because it also infringes the
patent in suit? This is what happened in the present case. The first problem
with this is that it required the judge to undertake a patent infringement
trial as part of the account of profits, with all the attendant complexity and expense.
153. The next problem is what happens if the claimant
turns out to be correct, and the proposed NIA actually infringes. Lufthansa
argues that, because the Defendants failed to prove their chosen NIA, Lufthansa
can claim all the profits. But all this shows is that the supposed NIA is
inapposite for a differential profit analysis because it is not actually an
NIA. It should not mean that the court is relieved from the burden of
identifying an NIA, because differential profit analysis requires an NIA. One
answer to this would be for the defendant to plead and prove a series of
alternative potential NIAs, each further away from the claimed invention than
the last, but that would simply compound the first problem.
154. The next difficulty is that a question may arise
as to whether the defendant could have undertaken the NIA. Suppose that the NIA
requires access to a particular raw material or part, but the defendant did not
have access to that material or part at the relevant time. As I understand
Lufthansa’s argument, this means that the defendant cannot rely upon the NIA,
but I question why not. The availability of that material or part does not alter
the inventive contribution of the patent. The object of a differential profit
analysis is to identify what profits are caused by the use of invention, not
what profits are caused by adventitious commercial factors. A similar problem
arises if the defendant is prevented from using the NIA by regulatory factors
unrelated to the invention. . . .
155. Furthermore, a question may arise as to whether,
even if the defendant could have undertaken the NIA, it would probably have
done so. This seems to me to even less relevant, since ex hypothesi we are
considering a counterfactual. A counterfactual is a thought experiment whose
purpose is objectively to identify the consequences of what the defendant
actually did. It does not depend on the probability of the defendant doing the
alternative in the counterfactual world. . . .
156. The next difficulty is the one I mentioned when
discussing the academic literature. In adversarial litigation courts depend on
the parties to adduce evidence. What happens if the evidence does not enable
the court to identify a suitable NIA? Counsel for Lufthansa argued that it is
always possible to postulate an NIA, even if it is simply not producing the
product in question at all. The problem with this argument is that an NIA only
serves the purpose of the differential profit analysis if it enables the court
to distinguish between the profits derived from the infringement from the
profits derived from other factors. As explained above, this requires the
identification of an NIA in which the minimum change necessary to avoid
infringement is made, but all other factors are held constant. As the present
case illustrates, the evidence may not permit this satisfactorily to be done.
157. The final problem I will mention is the one
touched on by Laddie J in Celanese v BP at [43] (paragraph 102 above).
In the real world, it is often the case that a complex product or process is
covered (or arguably covered) by multiple patents relating to different aspects
of the product or stages of the process. How does differential profits analysis
work on a claim for an account of profits for infringement of just one of those
patents? It cannot be correct to treat all of the profits generated by the
manufacture and sale of the complex product or process as attributable to that
infringement and none as attributable to the use of the other inventions. It
could be argued that this depends on whether the other patents are (a) valid
and (b) infringed, but that raises the spectre of determining the validity and
infringement of each of those patents, without the participation of the owners
of those patents, for the purposes of an account of profits. . . .
158. In short, while differential profit analysis has
much to be said for it in terms of legal and economic theory, applying it in
real world litigation is at best difficult, costly and uncertain.
159. The conclusions which I draw from this discussion
are as follows. First, as Astronics and Panasonic accept, there are some cases
in which it is possible to say that all of the profits in issue are derived
from the infringement. As Laddie J explained in Celanese v BP at [47]
(paragraph 104 above) and Lewison LJ noted in Abbott v Design & Display
at [28] (paragraph 108 above), these are cases where, without the infringement,
the infringer’s product or process would not have existed at all or where the
invention was the essential ingredient in the creation of the infringer’s whole
product or process. It may be possible, as discussed above, to reconcile such
cases with differential profit analysis on the basis that the NIA is not
manufacturing products at all, but that does not seem satisfactory. In any
event, that is not how they have been analysed in the English or Australian
case law. I shall return to this question in the context of ground 2.
160. Secondly, as Astronics and Panasonic also accept,
there may be some cases in which differential profit analysis is a useful tool
to identify the profits derived from the infringement, rather than from other
factors. These will be cases where there is a well-defined and uncontested NIA
which only changes the defendant’s product or process to the minimum extent
necessary to avoid infringement and holds all other factors constant. For the
reasons given in paragraphs 149-157 above, I am sceptical as to whether there
are likely to be many such cases.
161. Thirdly, the English and Australian case law
demonstrates that there is a well-established alternative to differential
profit analysis, which is for the court to make a fair apportionment of the
profits in issue. This approach is to be adopted when the case does not fall
into either of the two categories discussed in paragraphs 159-160 above, that
is to say, it is not a case where all of the profits are derived from the
infringement or where a differential profit analysis can readily be undertaken.
I will discuss how apportionment is to be carried out when I come to ground 4.
I’ll stop there for
now, and comment a bit on the preceding paragraphs.
First, I agree with
Lord Justice Arnold that, in the real world, and particularly in cases
involving complex products, it often may be unduly difficult or impossible to calculate
the profit attributable to the infringement by means of the differential
profits approach; and that, in recognition of these difficulties, what courts
often tend to do is to apportion, as best they can, the profit attributable to
the invention in comparison with the other features of the accused product. This is a point I do make when I teach my IP
remedies course, but I probably have not made it sufficiently clear in my
scholarship, which has tended to emphasize why I believe that, in general, the differential
profits approach is correct in theory and should be employed
when it is feasible to do so. A related
point that I have tried to emphasize, however, in papers such as Patent
Damages Heuristics, is that there often is a tradeoff between (theoretical)
accuracy and administrability; and that sometimes the net benefits of a more
easily administered rule outweigh the net benefits of (what might seem to be) a
more precise, but also more costly to implement, approach.
Second, I agree with much,
though not all, of Lord Justice Arnold’s analysis of the problems that can
arise when trying to carry out a differential profits analysis.
It is, of course, often difficult to determine precisely what the NIA
was (or to disentangle how its use might have affected other aspects of the
accused product). There is also the
difficulty of determining how to proceed if a proposed NIA itself turns out to
be patented, which is a matter my coauthors and I briefly noted at pages 20-22, 62 of Patent Remedies and Complex Products, but
didn’t attempt to resolve. Where I might
respectfully disagree with Lord Justice Arnold is in his discussion at
paragraphs 154-55 of whether a differential profits analysis should take into
account “adventitious commercial factors” or what the defendant would
have done but-for the infringement as opposed to what it could have done. More generally, I tend to agree with Norman
Siebrasse, in his critique of the Canadian Supreme Court decision in Nova v.
Dow (a case discussed in the present decision at paras. 127-32), that the
correct approach when applying the differential profits analysis is to identify
what the defendant would have done but-for the infringement, even if that noninfringing
option consists of deploying its resources to make an entirely different
product. The goal of the disgorgement remedy should be to determine
how much the defendant benefited from the infringement, and that means taking
into consideration what action the defendant would have taken had it not
infringed and estimating what benefits, adventitious or not, it would have
derived from doing so. For further discussion of this issue, see Siebrasse's article on Nova v. Dow, Part 1, particularly pp. 299-301 (arguing, inter alia, that "[a]n effort to determine the true value of the invention, apart from such happenstance, is akin to an effort to determine the true harm from negligent driving by awarding damages according to some ideal or average harm that would be caused by a negligent driving accident, rather than the accident that actually happened").
Third, and related to
the preceding points, the differential profit approach can run into problems if
there are two or more patents that are essential to the production of the
product in question. Imagine, for example, that two patentees each own an essential patent that is infringed by the maker of the accused product, and that each patentee files its own independent infringement action. Each might claim that, absent the use of the
patented technology in suit, the defendant would have earned zero profits; but surely
it would make no sense to award each patentee the entire profit earned from
sales of the infringing product. Some sort of apportionment would therefore seem to be necessary instead. (Perhaps the correct theoretical approach solution in such a case would
be to apply some version of Shapley Pricing to isolate the inventive
contribution of each essential patent, as Siebrasse and I proposed in The
Value of the Standard--though we cautioned there that our analysis was not “intended
to describe how we think a real world royalty setting process should work,” but
rather as presenting “conceptual benchmarks for assessing a FRAND royalty”
(p.1199). I would also call interested
readers’ attention to a somewhat analogous issue discussed by Jason Reinecke in
his article Lost Profits Damages for Multicomponent Products: Clarifying the Debate, 71 Stan. L. Rev. 1621 (2019), in the
context of lost profits and multiple essential patents.)
Returning to the Lufthansa
decision, the next portion of Lord Justice Arnold’s opinion concludes, in
brief, that the trial court was correct to apportion profits, despite some
language in the lower court opinion suggesting that it was doing so despite
having found that those profits were not legally (proximately) caused by the
infringement (see paras. 162-76). In
fact, there was sufficient evidence that an allocable portion of the profits
were legally caused by the infringing use of the patent (paras. 177-90).
That leaves for
discussion Lord Justice Arnold’s analysis of the method of apportionment, the
double recovery issue, and interest, as well as the two short concurring
opinions by Lord Justices Nugee and Lewison.
I will return to these in a subsequent post or posts.
