Grain Processing Corp. v. American Maize-Products Co.

979 F.Supp. 1233 (1997)

GRAIN PROCESSING CORPORATION, Plaintiff,
v.
AMERICAN MAIZE-PRODUCTS COMPANY, Defendant.

No. H81-237.

United States District Court, N.D. Indiana, Hammond Division.

September 26, 1997.

Geoffrey G. Gilbert, McBride Baker and Coles, Chicago, IL, for Plaintiff.

Robert L. Baechtold, Fitzpatrick, Cella, Harper and Scinto, New York, NY, for Defendant.

Opinion

EASTERBROOK, Circuit Judge.^[*]

American Maize Products Co. (AMP) infringed Claim 12 of U.S. Patent No. 3,849,194. A reasonable royalty for the patented product would have been 3% of sales, yielding an award of approximately $2.5 million to plaintiff Grain Products Corporation (GPC). What is now before the court, following remand from the Federal Circuit, is whether a calculation from a lost-profits perspective would produce a higher award. To avoid repetition, I assume that the reader is familiar with the two published opinions in this case: 840 F.2d 902 (Fed.Cir.1988), and 893 F.Supp. 1386 (N.D.Ind.1995).

Claim 12 of the patent (from now on, any reference to the patent is to Claim 12) covers a waxy starch hydrolysate with a "dextrose equivalent" (D.E.) between 5 and 25, and a "descriptive ratio" (D.R.) greater than "about 2". This claims one variant of a substance known as low-dextrose malto-dextrin. GPC and AMP, along with several other firms, sell such products for use in the food industry. District Judge Parsons and the Federal Circuit issued a series of opinions collectively establishing that Lo-Dex 10, one of AMP'S line of low-dextrose malto-dextrins with a D.E. of 10, infringed the patent, but that AMP's products with higher and lower D.E. values did not infringe. Judge Parsons retired before damages could be determined. After a damages trial to the bench in 1995, I concluded that AMP could have produced a non-infringing product (as it ultimately did) with a D.R. reliably less than 1.9, and that it infringed the patent only because it misunderstood which test would be used to measure the D.E. value, the denominator in the fraction that defines the D.R. value. After analyzing the production methods available to AMP during the period of infringement, I wrote: "My conclusion that AMP has carried its burden of establishing that it could have produced a non-infringing product no later than October 1979 scotches GPC's request for lost-profits damages." 893 F.Supp. at 1392. The court of appeals responded, in an unpublished order:

We agree with GPC that the [district] court erred in refusing to award lost profits based on a noninfringing substitute that was not available at the time of infringement. The issue of noninfringing substitutes in a lost profits analysis comes from the test established in Panduit Corp. v. Stahlin Brothers Fibre Works, Inc., 575 F.2d 1152, 197 USPQ 726 (6th Cir.1978). In that case, the court required the patent holder, in order to obtain damages in the form of lost profits, to establish (1) demand for the patented product, (2) the absence of acceptable noninfringing substitutes, (3) manufacturing and marketing capability to exploit the demand, and (4) the amount of profit it would have made. Panduit, 575 F.2d at 1156, 197 USPQ at 730.

The court in Panduit established the proper time for considering the availability of noninfringing substitutes as the period of infringement. See id., 575 F.2d at 1160-62, 197 USPQ at 732-35. On this issue, the Panduit court went further and said that switching to a noninfringing product years after the period of infringement did not establish the presence of a noninfringing substitute during the period of infringement. Id. at 1162, 197 USPQ at 735.

Likewise, in State Industries, Inc. v. Mor-Flo Industries, Inc., 883 F.2d 1573, 1579, 12 USPQ2d 1026, 1030, (Fed.Cir.1989), this court rejected an attempt to rely on a noninfringing substitute that was not available during the period of infringement. Thus, the law is clear — to be an acceptable noninfringing substitute, the product or process must have been available or on the market at the time of infringement. The district court's holding to the contrary is erroneous.

Nos. 95-1506 & 95-1507 (Fed.Cir. Feb. 20, 1997), slip op. 3-4 (table entry 108 F.3d 1392).

I do not wish to be presumptuous, but it seems to me that my opinion did what the court of appeals believes ought to have been done. That is, I found as a matter of fact that a "noninfringing product" was available "no later than October 1979". The Federal Circuit did not conclude that this finding was clearly erroneous; instead it believed that there had been a legal blunder. Yet, as the court of appeals wrote, "to be an acceptable noninfringing substitute, the product or process must have been available or on the market at the time of infringement" (emphasis added). I found and reiterate (a) that noninfringing substitutes for the patented product were on the market at the time of infringement, and (b) that AMP had "available" at the critical time a process that would have ensured that its own product did not infringe the patent.

Proposition (a) is common ground between the parties and therefore was not expressly identified as a "finding": GPC itself makes a variety of low-dextrose malto-dextrins that fall outside the patent; so does AMP (this was established at an earlier stage of the case); firms other than GPC and AMP likewise do so. It is easy to make such a product: just start with a non-waxy feedstock. Customers do not care whether the feedstock contains a waxy starch or whether the D.R. of the finished product is greater than "about 2"; they care about its D.E. value. Moreover, GPC faced competition from non-infringing low-dextrose malto-dextrins that AMP made, even though its feedstock is a waxy starch. Lo-Dex 5 and Lo-Dex 15, which for many end products are substitutes for the infringing Lo-Dex 10, did not infringe. Other firms in the industry also made low-dextrose malto-dextrins that did not infringe the patent. GPC believes that A.E. Staley, one of these rivals, did infringe, though this was never established in litigation; at all events, GPC does not contend that all competing low-dextrose malto-dextrins infringed. So noninfringing substitutes were on sale. GPC did not argue otherwise at the damages trial. What it did argue is that AMP could not have made a non-infringing D.E. 10 product, that D.E. 10 malto-dextrin occupies a market apart from the D.E. 5 and D.E. 15 products, and that if AMP had dropped out of the D.E. 10 line all of its sales would have gone to GPC rather than to the other producers.

It is obvious in retrospect that my discussion of proposition (b) was not as clear as it needed to be. I set out to articulate a finding that although, until 1991, AMP did not sell a non-infringing D.E. 10 malto-dextrin, such a product was "available" in the way the Federal Circuit uses that word. AMP did not have to "invent around" the patent; all it had to do was use a glucoamylase enzyme in its production process. Until 1991 AMP had not used this enzyme, but the sole reason was economic: glucoamylase is more expensive than the alpha amylase enzyme that AMP had been using. The chemical effects of glucoamylase have been known for a long time; my opinion cites a 1967 textbook and the pre-1974 advertising of enzyme manufacturers, which touted that glucoamylase could be used to achieve the effect (breaking starch molecules in hydrolysis at the end of the chain rather than in the middle) that would reduce DP1-6 the numerator in the fraction that yields the D.R. value. See 893 F.Supp. at 1391. By reducing the numerator while holding the denominator (the D.E. value) constant, AMP produced Lo-Dex 10 with a D.R. value reliably less than 1.9. It could have done this at any time during the period for which GPC is entitled to damages, I concluded, and did not do so only because of a legal error. AMP believed, after using the Lane-Eynon test to determine the D.E. value, that Lo-Dex 10 made by Process II and Process III did not infringe the patent; only when the Federal Circuit held in 1991 that the Schoorl test was the right benchmark did AMP recognize the need to change production processes. It was able to do so within two weeks of the Federal Circuit's decision, which for large-scale production is practically instantaneous. I did not, and do not, say that the simple fact of switching established the availability of a non-infringing substitute; my finding of availability was based on a number of subsidiary findings about the technology of enzyme-assisted starch hydrolysis that the court of appeals did not mention. To the extent the federal circuit believed that I made such a leap from the switch to earlier availability, I trust that this clarification will set things straight.^[†]

I am confident that the court of appeals did not mean to say that a particular product must be sold contemporaneously with the infringement to count as an "available" non-infringing substitute. Lost-profits damages are designed to give the patent holder the economic benefits it would have enjoyed had its intellectual property been respected. Aro Manufacturing Co. v. Convertible Top Replacement Co., 377 U.S. 476, 507, 84 S.Ct. 1526, 1543, 12 L.Ed.2d 457 (1964) (plurality opinion). This rule calls for a reconstruction of the way the market would have developed in the absence of infringement. Reconstruction takes account not only of substitutes actually produced but also what would have been produced, had it been economically advantageous to do so. This is the point of the Federal Circuit's statement that "to be an acceptable noninfringing substitute, the product or process must have been available or on the market at the time of infringement" (emphasis added). A product that is within a firm's existing production abilities but not on the market — in this case, Lo-Dex 10 made by Process IV (see 893 F.Supp. at 1389-90) — effectively constrains the patent holder's profits. Potential competition can be as powerful as actual competition in constraining price. William J. Baumol, John C. Panzar & Robert D. Willig, Contestable Markets and the Theory of Industry Structure (1982). See also George J. Stigler, The Theory of Price 203-10 (4th ed.1987). To see this, consider two examples.

The dominant storage medium for digital data today is the Winchester drive mechanism, the basic technology behind modern hard disks. Bits are encoded as magnetic domains on a metal oxide and read by a head on a mechanical arm as a platter spins beneath it. One rival on the horizon is holograms stored in crystal lattices, read by lasers. See Demetri Psaltis & Fai Mok, Holographic Memories, 273 Scientific American 70 (Nov.1995). Suppose this technology is patented and brought to market. It is likely to be more reliable (it has fewer moving parts), faster (beams of light can be deflected faster than arms can move or disks turn), and cheaper (light can encode more data per unit of area, and the crystal can store in three dimensions while the magnetic disk stores in only two). Suppose that hard disks can be manufactured for 10¢ per megabyte, while holographic memory costs 6¢; consumers accordingly turn to crystals, and hard disks vanish from the market. Next suppose that someone infringes the patent on holographic crystal memory. What will be the patent holder's lost profits? Although they could be less than 4¢ per megabyte of memory sold, they cannot be more-for if the patent holder put its devices on the market at 11¢ per megabyte, it would pay to bring Winchester drives back into production. (The patent holder's implicit threat to engage in limit pricing in response to entry could be defeated by long-term contracts, a strategy Baumol and colleagues discuss.) The profit could well be less than 4¢ per megabyte: improvements in hard disk technology might make reentry at a price below 10¢ attractive, and at all events even a monopolist might find that the profit-maximizing price (where marginal cost equals marginal revenue) is less than 10¢ per megabyte. All that matters for current purposes, however, is that a product missing from the market can strongly affect, if not determine, the price a patent holder can obtain, and therefore the profit lost by infringement.

The computer-memory example assumed that the product was patented. Suppose instead that a process patent covers the most economical way of producing an unpatented product. The gasoline cracking patent case, Standard Oil Co. (Indiana) v. United States, 283 U.S. 163, 51 S.Ct. 421, 75 L.Ed. 926 (1931); see Ward S. Bowman, Jr., Patent and Antitrust Law 203-18 (1973), provides the inspiration for this second example. Gasoline used to be made by simple distillation, which converts only a small portion of crude oil into the lighter fractions such as gasoline and kerosene that consumers value highly. Catalytic cracking of the residue left after simple distillation produces a much higher yield per barrel of oil. An antitrust problem arose when holders of different patented cracking technologies pooled their patents; the effects of this merger depended on whether the patents were competitive or blocking. That dispute does not matter for current purposes. Suppose there was only one patented cracking technology, and that the unpatented distillation technologies remained available. At what price could the owner of the cracking patent sell a gallon of gasoline? Not for a penny more than gasoline made by the old technology, because gasoline is fungible. Thus the maximum lost profit, in the event of infringement of the cracking patent, would have been the difference between the actual price of gasoline and the price a manufacturer using the old distillation technology would have charged.

GPC was in the same economic position as the holder of the cracking patent in this second example. The economically-significant product — gasoline or low-dextrose malto-dextrin — is unpatented. True enough, GPC's patent covers a product rather than a process, but as I have emphasized before and will again, the patent claims a particular attribute of waxy starch hydrolysates (a D.R. value greater than "about 2") that consumers do not value. What they want is low-dextrose malto-dextrin, which like gasoline does not infringe a patent. So GPC's patent, although in terms a product patent, has the same economic consequences as a process patent, like the gasoline cracking patent. Low-dextrose malto-dextrins made by some processes will have D.R. values greater than "about 2" and will infringe if they satisfy its other elements; low-dextrose malto-dextrins made by other methods (starting with non-waxy feedstocks, or hydrolyzed using glucoamylase enzymes) will not have infringing D.R. values. The patent effectively forecloses the least-cost method of turning waxy starch into low-dextrose malto-dextrin, and therefore the profit lost from infringement is the cost and market price difference attributable to using glucoamylase.

Now I leave hypothetical reconstructions behind and approach the lost-profits question from a different angle. Suppose that for formal reasons the only proper way to look at the patent is strictly as a product claim. Panduit and, more pertinently, Rite-Hite Corp. v. Kelley Co., 56 F.3d 1538, 1545-46 (Fed.Cir.1995) (en banc), identify demand for the patented product as an essential element of the patent holder's lost-profits claim. (Rite-Hite is more pertinent than Panduit because GPC did not practice its patent. Lost-profits damages therefore depend on the approach Rite-Hite articulated for such cases.) Is there a demand for the product covered by the patent?

To answer this question, it is essential to be precise about what the patented product is. GPC emphasizes that there is a substantial demand for D.E. 10 malto-dextrins, which no one doubts. It also contends that D.E. 5 and 15 malto-dextrins are not good substitutes for D.E. 10 products, which is a hotly contested issue. Assume that D.E. 10 malto-dextrin is an economically significant product, at least in the short run. (Over a longer run, food producers that now use D.E. 10 products could reformulate their products so that other binders and fillers would serve their needs. GPC has never attempted to estimate the long-run elasticity of substitution among different dextrose equivalents. Given the conclusion that follows, the elasticity does not matter.) Still, GPC does not have a patent on D.E. 10 malto-dextrin. Its patent covers:

A waxy starch hydrolysate having

a dextrose equivalent value between about 5 and about 25,

a descriptive ratio greater than about 2, said descriptive ratio being the quotient obtained by dividing the sum of the percentage of saccharides, dry basis, having a degree of polymerization of 1 to 6, by the dextrose equivalent value,

a monosaccharide content in the range of from about 0.1 percent by weight, to about 2.4 percent by weight, dry basis,

a disaccharide content in the range of from about 1.3 percent to about 9.7 percent, by weight, dry basis, and

being further characterized as capable of producing an aqueous solution of exceptional clarity and substantially complete lack of opaqueness when said hydrolysate is added to water.

The question is whether there is an economically significant demand for a product having all of these attributes. To this the answer is no. Two of the essential elements of this claim — that the starch be "waxy" and that the "descriptive ratio [be] greater than about 2" — are irrelevant to consumers. There is no demand, none at all, for the patented product; the demand is for low-dextrose malto-dextrins of which the patented product is just one exemplar, functionally identical to its unpatented rivals in the marketplace.

This is why the patent's value rests on its effect on manufacturing processes. Suppose that in lieu of the descriptive ratio clause the patent read "glows when subjected to ultraviolet radiation". That change would be irrelevant to the food industry. Suppose further that the cheapest method of hydrolysis used to make a D.E. 10 malto-dextrin just happens to produce fluorescence under ultraviolet radiation. Buyers of malto-dextrins would not care, but firms such as AMP would have to eliminate this attribute to avoid infringement. The higher manufacturing cost of doing so would enable the patent holder to charge a royalty for a license. But the lack of any demand for an ultraviolet-sensitive malto-dextrin would show that the patent holder did not suffer lost profits from the diversion of sales. And so it is with GPC's patent. There is no demand for low-dextrose malto-dextrins with a D.R. value greater than about 2, because the D.R. value is irrelevant to purchasers of the product. GPC does not have a patent on D.E. 10 malto-dextrins, the economically significant product, and therefore cannot recover lost-profits damages on account of AMP's competition.

I anticipate the response that defining the product as the precise combination of attributes claimed in the patent would have knocked out an award of damages in Rite-Hite itself, but this is not so. Rite-Hite and Kelley sold competing devices for securing a truck to a loading dock, in order to prevent the vehicle from moving while loading or unloading was in progress. Kelley's device infringed Rite-Hite's patent — but there were other ways of securing trucks to docks, and another of Rite-Hite's models, not covered by the patent in suit, was foremost among them. The court of appeals held that there could be an economically significant demand for the patented product even though there were other ways to achieve consumers' objectives. It was important to the finding of lost profits that the principal "other way", Rite-Hite's own device, also was patented. Rite-Hite's patents collectively entitled it to sell the whole market's demand, and therefore to collect profits lost by infringement that siphons sales. In other words, Rite-Hite treats the "patented product" for purposes of lost-profits analysis as including the attributes of all of the plaintiff's multiple patents, and not just the attributes of the patent in suit. That approach does not assist GPC, because products that are perfect substitutes for waxy low-dextrose malto-dextrins with a D.R. greater than about 2 are not covered by any patent, let alone by a patent GPC owns. Unlike Rite-Hite, GPC did not have a legal right to prevent competition from other products that consumers deemed identical to the patented product.

GPC's lost-profits claim is flawed in several other respects. Its failure to offer any evidence concerning the elasticity of substitution among D.E. 5, D.E. 10, and D.E. 15 malto-dextrins makes it impossible to know how far (and how fast) consumers would shift to these other products if the price of D.E. 10 malto-dextrin rose. Its lost-profits study is unrealistic: it treats the malto-dextrin business as a high-margin endeavor but is not based on an economically sound means of determining which of its costs are fixed over the relevant time dimension. GPC's offer to license a portfolio of patents, including the one in suit, for less than a 5% royalty means that its own business managers believed that its long-run margins are low — for otherwise the license would be disastrous. (It would be folly to exchange a 40% profit from the sale of a pound of malto-dextrin for a 5% royalty on the sale of that pound by a competitor.) But because I have concluded that there is no economic demand for the patented product, I need not elaborate on these additional obstacles to GPC's lost-profits claim.

One task remains: ascertaining the date at which pre-judgment interest (on the reasonable-royalty award) stops and post-judgment interest begins. Under 28 U.S.C. § 1961 the date of judgment marks the transition to the statutory post-judgment rate. Kaiser Aluminum & Chemical Corp. v. Bonjorno, 494 U.S. 827, 835-36, 110 S.Ct. 1570, 1575-76, 108 L.Ed.2d 842 (1990), holds that a legally defective judgment should be disregarded for the purpose of fixing this date. The court of appeals vacated my original judgment but contemplated the possibility that the same judgment would be reentered on remand. When the same monetary award is entered twice, which of the two judgment dates marks the dividing line? The parties have not addressed this question, and I therefore invite them to file short memoranda within 10 days setting out their views. It would make little economic sense to use the pre-judgment rate, which includes compensation for the risk of non-payment, for a period during which GPC was assured of payment by AMP's letter of credit. Still, the law appears to be murky, see Adkins v. Asbestos Corp., 18 F.3d 1349, 1351 (6th Cir.1994); Charles Alan Wright, Arthur R. Miller & Edward H. Cooper, 16A Federal Practice and Procedure § 3983 at 586-88 (1996), making it advisable to solicit the parties' views. After receiving these submissions, I will enter a new final judgment.

NOTES

[*] Of the Seventh Circuit, sitting by designation.

[†] I do not think that this conclusion is incompatible with Mor-Flo Industries, which the court of appeals read to "reject[] an attempt to rely on a noninfringing substitute that was not available during the period of infringement." Mor-Flo does not address the question whether, or under what circumstances, a readily made product that is not actually on the market is "available".