Wednesday, May 15, 2013

Interesting Theoretical Point Raised in Recent U.K. Decision on Lost Profits

Judge Birss’s opinion in Xena Systems Ltd. v. Cantideck, [2013] EWPCC 1 (Pat. Cty. Ct.2013), came out too late for inclusion in my book, but it’s worth a look for its detailed application of U.K. law in awarding a prevailing patent owner its lost profits.  EPLAW and PatLit have already blogged on various aspects of the decision and I need not repeat their analysis here, but I would like to note two aspects of the opinion that I found particularly interesting.  One is that the court awarded the patentee damages for higher borrowing costs it incurred as a result of the infringement; the court viewed these costs as foreseeable consequences of the infringement and recoverable under the House of Lords’ decision in Sempra Metals v. Inland Revenue Commissioners [2008] 1 AC 561.  The other is a theoretical point on the recoverability of lost profits on so-called convoyed sales, that is, on sales of nonpatented complementary goods or services that the patentee would have made along with the sales of patented articles that it would made but for the infringement.  Lost profits damages for such goods are recoverable under English law, as long as they are foreseeable consequences of the infringement and not too remote (see my book, p. 187).  Defense counsel, however, raised an interesting theoretical point which the court addressed as follows:

44  In relation to the claim for losses of unpatented fixed platform hiring, Cantideck accepted that losses for so called "convoyed goods" were recoverable in principle in a patent case (Gerber). Mr Abrahams pointed out that in Gerber itself the convoyed sales were goods sold together with the patented products and he argued that Xena were actually seeking to claim for the loss of a chance of a chance, which is not allowed. He argued that in Gerber the court had estimated the lost chance of a sale at 60% and awarded 60% of the profit of the patented goods and the convoyed goods. It did not assess the chance of the convoyed goods separately because that head was only awarded on the basis that it goes with the patented goods. So, he submitted:

"For example, if the chance of the patentee selling the patented item was 50%, and the chance of selling the convoyed goods was 50% in any case where patented goods were sold, the right award is 50% of the profit on the patented goods and not a further award of 25% of the convoyed goods. 

This is because where the sale of the patented item gives the seller a mere chance to sell other goods, that is too remote […]. Otherwise a supermarket that sold an infringing item would have to pay damages in respect of every lawful item in the shop."

45  Mr Cuddigan called this a novel submission and did not agree with it. In my judgment Mr Abrahams' argument, when it is put as a matter of principle, is not right. By approaching it as a "chance of a chance" the argument sets too much store by one mathematical approach to assessing damages. It may be that there is something about the defendant's sales which means none of them, had they been made by the patentee, would have included the convoyed article or maybe all of them would have but these are all matters of fact open to be proved if necessary or proportionate by one side or the other. If, on the evidence available, there is nothing to distinguish sales which included the convoyed goods from sales which did not then one can say that the overall fraction of the patentee's sales which included convoyed goods reflects a probability for each sale that it would have included convoyed goods. Assuming that chance is a substantial one, then I can see no objection to taking it into account as long as the requirements of foreseeability and remoteness are satisfied. One could equally well divide the patentee's lost chance into two separate lost chances, a lost chance of high value sale of a patented product plus convoyed article (25% on Mr Abrahams' figures) and a lost chance of a lower value sale of a patented product alone (also 25%).

46  However, although I do not accept Mr Abrahams' argument when it is put as a matter of principle, it seems to me that care needs to be taken with the facts in this sort of case. When a patentee sells two products side by side, one patented and the other not, it does not follow that just because, averaged over a period, a fraction of the customers who bought the patented article also bought the non-patented one, the patentee can claim for sales of the non-patented article. Gerber is clear authority that the scope of recovery is not restricted to activities which themselves constituted infringements but it is still limited by causation and remoteness. As Staughton LJ said (at 456 ln5-15):
"Beyond that the assessment of damages for infringement of a patent is in my judgment a question of fact. There is no dispute as to causation or remoteness in the present case; nor can I see any ground of policy for restricting the patentees' right to recover. It does not follow that, if customers were in the habit of purchasing a patented article at the patentee's supermarket, for example, he could claim against an infringer in respect of loss of profits on all the other items which the customers would buy in the supermarket but no longer bought. The limit there would be one of causation, or remoteness, or both. But the present appeal, in so far as it seeks to restrict the scope of recovery, should be dismissed."

The relevant language from Gerber Garment Technology Inc. v. Lectra Systems Ltd., [1997] R.P.C. 443, reads as follows:

As I have already said, the judge awarded damages on the basis that the patentees would have achieved 15 sales if the infringers had not made their 25 wrongful sales. There was extensive evidence of the circumstances of those 25; but the judge did not identify 15 which would in fact have been achieved by the patentees. Nor, in consequence, did he identify the profit lost on each particular machine featuring in his calculation. He simply awarded 60 per cent of the total sum claimed as loss of profit on the 25 machines. One can infer that he must have found the figures for loss of profit for each machine proved.

The judge adopted a similar method of calculation for spare parts, servicing and CAD systems. Mr. Hobbs Q.C. for the infringers says that the judge was wrong in law to adopt the method which he did; and he should have made findings relating to the individual machines. His clients, as he put it, were entitled to a speaking award. The judge's calculation of price depression and springboard damages was subject to similar criticism. . . .

In my judgment the issue as to the amount of the patentees' loss in the present case was a question in the second class; it depended on the hypothetical actions of third parties, that is to say the buyers of the infringing machines (or spare parts, servicing and CAD systems). The judge was entitled to conclude that the patentees had lost a chance of making sales to those buyers - no doubt a chance of differing probability in each case. He was entitled to evaluate the chances as a whole, rather than separately, if he chose to do so. The contrary view, that if the judge found 25 chances of a sale, each of 49 per cent probability, he should award nothing is absurd.

I’m inclined to think that in both cases the court got it right—that the plaintiff should recover lost profits on convoyed sales that it reasonably would have made, but for the infringement (note that U.S. law is a bit different on this score; see my book at pp. 117-18); that the application of this principle should turn on the specific facts of the case; and that courts should be wary of undercompensating patent owners and underdeterring infringement by applying overly rigid probability cut-offs. 

Of course, in other contexts a successful tort plaintiff who can show, say, only a 40% chance that the defendant’s conduct caused it to suffer a loss of $1 million generally doesn’t recover $400,000, but rather $0, a result that could lead to underdeterrence in some contexts.  On the other hand, a plaintiff who can show a 60% chance recovers the full $1 million, which can result in overdeterrence.  The problem is well known in the law and economics literature.  See, e.g., Ariel Porat & Eric Posner, Aggregation and Law, 122 Yale L. J. 2 (2012).  Off the top of my head, however, I can’t think of other instances in which the problem has been considered in the context of patent damages in particular.  Perhaps a good topic for a future article . . .    


  1. I don’t think there is any inconsistency between Gerber and the rule that causation must be established on the balance of probabilities. In Canadian law, at least, the rule (Athey v Leonati [1996] 3 SCR 458 ¶¶ 27-28) is that:

    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. Instead, they are simply given weight according to their relative likelihood.
    By contrast, past events must be proven, and once proven they are treated as certainties.

    Hypothetical events include the events of the but for world, such as what sales the patentee would have made but for the infringement. In effect, balance of probabilities is used in establishing the substantive tort, or infringement and validity in patent law, but a probabilistic approach is used in assessing damages.

  2. Interesting point. I think that approach makes sense, and thanks for the Canadian cite.