W HEN ABRAHAM LINCOLN WAS ATTACKED in print covers 129 pages, - - PDF document

SLIDE 1

S U M M E R 2 3 · 8 5

age] chance that the patent would be held invalid and the market would become competitive . . . [T]hose payments are inherently anticompetitive. On expectation the patentee is payment for an advantage that it could not get if it went to trial”).

7 See, e.g., id. at 2, 46–49 (arguing that “exclusion payments that exceed

litigation costs should be deemed illegal per se). http://ssrn.com/abstract=380841, at 36 (“Insofar as antitrust is con- cerned, among the most problematic settlement agreements are those in which the infringement plaintiff pays the infringement defendant for that lat- ter’s abandonment of the market,” because such an “exclusion payment,” “keeps the rival out of the market and induces it to drop its suit in exchange for a payment”). Id. at 46–47 (“[t]he reason the patentee is willing to make this payment is precisely because the settlement will permit it to exclude competition from the market, whereas if it went to trial there is a [percent-

With Malice Toward None: A Brief Rejoinder to Leffler and Shapiro

B Y K E V I N M c D O N A L D

W

HEN ABRAHAM LINCOLN WAS ATTACKED in print by the eminent Horace Greeley, he began his response this way:

I have just read yours of the 19th. addressed to myself through the New-York Tribune. If there be in it any statements, or assumptions

• f fact, which I may know to be erroneous, I do not, now and here,

controvert them. If there be in it any inferences which I may believe to be falsely drawn, I do not, now and here, argue against them. If there be perceptible in it an impatient and dictatorial tone, I waive it in deference to an old friend, whose heart I have always supposed to be right. As to the policy I “seem to be pursuing” as you say, I have not meant to leave any one in doubt. I would save the Union.1

Lincoln, I think, understood that responding to the specifics of the attack, pointing out its liberties from the actual record of what he had said and done, was less important to the public than a firm understanding of his position. While I make no pretense to Lincoln’s higher virtues, nor to his matchless prose, I will also spare those of you who have read this far a detailed list of grievances as to how my ideas have been mangled. What I propose instead is a brief listing of four fundamental principles on which I believe my critics and I do disagree—a subject about which the readers of the three pieces addressing my article have a right to be confused.2 This approach, I hope, will give those readers a chance to form their own views—and, ide- ally, to formulate even better questions than we are asking about the ultimate point: whether the “probability” that a patent may be invalidated during litigation, however small, can be used to jus-

Kevin McDonald is a Partner at Jones Day, Washington, D.C.

tify a legal conclusion that a settlement with payments from the patentee is per se unlawful (according to the Lefflers), or pre- sumptively anticompetitive (according to Professor Shapiro). If nothing else, this approach may help us to understand why the attack on “reverse” payments continues to flounder in the

• courts. In addition to the Schering-Plough decision from the FTC’s

Administrative Law Judge discussed in my prior article, two fed- eral district courts have recently spoken—one dismissing a com- plaint based on a Hatch-Waxman settlement altogether;3 the

• ther refusing to find a settlement per se illegal, and noting

that “so-called reverse payments are a natural by-product of the Hatch-Waxman process.”4 Although one of these opinions covers 129 pages, neither embraces, or even mentions, the “probabilistic” property theory.5 The principles that follow may explain why. Antitrust Laws Impose Only Negative Duties. One common mistake is to analyze competitive conduct

as if the function of antitrust law were to compel firms to maximize competition . . . rather than to prevent them from restricting it. [On the contrary,] there is a difference between positive and negative duties, and the antitrust laws . . . have generally been understood to impose only the latter.6

Accordingly, a practice may not be condemned under Section 1

• f the Sherman Act unless it reduces competition in the first
• place. It is not enough to hypothesize an alternative that a plain-

tiff regards as more competitive (e.g., a license). “It is the choice

• f an unreasonable alternative, not the failure to choose the least

restrictive alternative, that leads to liability.”7 To avoid this fallacy, one cannot define the consumer harm allegedly flowing from patent settlements as the Lefflers do: “It is clear that licensing agreement settlements which allow entry of the alleged infringer are preferable on a static welfare basis to lump sum patent settlements which sustain the exist- ing monopoly supply situation.” Leffler III at 78.8 Thus defined, the harm that they purport to weigh against the conceded bene- fits of a settlement consists of the difference between the set- tlement chosen by the parties (payments) and the settlement they might have chosen (a license). Under the law, however, this is not an anticompetitive effect at all. In the words of the Second Circuit: “The [agreement] is not even amenable to scrutiny under Section 1 unless it is a restraint of trade. The fact that it may be in some sense ‘unnecessary’ does not make it a restraint.”9 Professor Shapiro imposes positive duties with assertions like this: “[L]icenses are generally regarded as anticompetitive if

SLIDE 2

A R T I C L E S A N D F E A T U R E S

8 6 · A N T I T R U S T

the market.17 Two companies may compete vigorously other- wise, but Professor Hovenkamp has explained that, when one has a patent, they do not “compete” for the patented invention in any sense relevant to antitrust liability (he used GE and Westinghouse as an example).18 It thus remains suspect, in my view, to declare competition proved to be infringing under the patent laws “entirely legitimate” while acknowledging, as Professor Shapiro must, that both the sellers of the infringing product and every consumer who buys or even “uses” the product, are breaking the law. He protests that he never sought clemency for the infringing seller (Shapiro at 74) (which was my point exactly; he should have done so, if the com- petition is “entirely legitimate”). When he finally addresses the point that consumers are also acting illegally, we are told that the infringer and the consumers will carefully weigh the “strength” of the patent in order to determine the optimal level of infringement. Shapiro at 74. If the competition produced by these “market forces” proves to be illegal, the patentee “will be compensated as allowed under the patent system.” Id. So, all is well. Except for a few things. First, this is a complete justification for protecting black markets. We will therefore have to rethink the maxim that “[w]e do not want an efficient market in stolen goods.”19 The robbery victim, after all, can be compensated “as allowed for” by the legal system as well. Why not let these same “market forces” determine just how strong the owner’s right to his car really is? Second, I for one will need some empirical evi- dence before finding it plausible that consumers will be evaluat- ing the validity of the patent, rather than evaluating—as would the car thief—whether they’ll be caught. (“Fred, make sure you buy that new generic drug this time; that branded drug makes me suspicious of double patenting.”) I could go on, but the decisive point is whether the antitrust laws care about competition that is proven to be illegal. As I have shown previously, they do not.20 The “Possibility” of Harm to Competition Cannot Support an Inference of Actual Harm. I note the telling absence in the three comments of any challenge to the legal rule that the mere pos- sibility of harm to competition cannot substitute for actual harm. That “[i]n our legal system, . . . a 10 percent chance of rain is not rain.” McDonald at 72. As I noted, this rule has been applied with particular force when a claim of injury is based on the outcome

• f a contest, whether horseracing or litigation: “[N]o matter how

meritorious one’s claim may appear at the outset, the course of litigation is rarely predictable.”21 In light of this law, I suggested that there is a fundamental choice to be made when considering, after the fact, whether a settlement within the scope of a pre- sumptively valid patent harms competition. We may begin by ask- ing (1) whether the patent is invalid, in the same sense and under the same evidence as in actual patent cases, or (2) whether (as the proponents of the “probabilistic” property theory sug- gest) the specific litigation that was settled would have resulted in a defeat for the patent. Shapiro at 70. The first inquiry is difficult, sometimes grueling, but under- taken routinely in thousands of cases; the second, I tried to show, is both misguided and infeasible. By rejecting the second inquiry, I avoid a thicket of analytical problems that my critics they restrict competition in comparison with what would have

• ccurred in the absence of the license.” Shapiro at 70 (empha-

sis added). The Tamoxifen plaintiffs used such reasoning to argue that a settlement was anticompetitive because it pre- vented the parties from agreeing to a “competitive license.” The court disagreed.10 Because positive duties are not required, we cannot avoid the question I raised in the prior article: “No mat- ter who paid what to whom, what lawful competition has been reduced by the settlement?” McDonald at 68. If the patent is valid, and (this is crucial, although ignored by my critics) the set- tlement does not extend the preclusive effect of the patent, then the answer remains “None.”11 “Competition” Under the Antitrust Laws Is Not Defined by Rivalry Alone. The notion that more rivals means more compe- tition is superficially appealing and sometimes true, but has been proven unreliable.12 It nonetheless drives, for example, the Lefflers’ conclusion that a settlement with a license has more “static efficiency” than one with payments. E.g., Leffler III at 78. Over the last three decades, however, as Judge Posner has noted, “the emphasis of antitrust policy [has] shifted from the protection of competition as a process of rivalry to the protection

• f competition as a means of promoting economic efficiency.”13

A corollary of this rule is that the protection of valid patents and the innovation that results are procompetitive benefits. Especially in a market like pharmaceuticals, new inventions can provide dramatic and genuine improvements in product quality, which for an economist is the same as lowering prices. If that requires less rivalry, and even monopoly, nothing in the antitrust laws is offended. This is why Dennis Carlton notes that “[e]ven the introduction of a product subject to monopoly power can represent a gain to society,”14 and why Richard Posner concludes that “[a]ppearances to the contrary notwithstanding, the antitrust laws are not much concerned with monopoly as such.”15 Both are referring expressly to the procompetitive aspects of intellectual property protection. This truth cannot be evaded by accusing me of “defending set- tlements that reduce consumer surplus on the grounds that they promote innovation.” Shapiro at 76. The devil, of course, is in the

• definitions. Since settlements that “promote innovation” are

procompetitive to that extent, they do not necessarily “reduce consumer surplus”; they can (and by definition must, when the settlement does not exceed the scope of a valid patent) increase consumer surplus, properly defined. Thus, I do not trade off inno- vation for consumer surplus; I challenge any definition of con- sumer surplus that fails to account for the procompetitive ben- efits of innovation. Professor Shapiro’s characterization of my argument is an attempt to impose another asymmetrical “pre- sumption” here—that consumer surplus can only “grow” when a patent is defeated. Unless rivalry is deemed an end in itself, how- ever, that presumption will not work. The Antitrust Laws Do Not Protect Infringing Competition. The proposition that “the public [is] not entitled to profit by com- petition among infringers” is older than the Sherman Act itself,16 and it remains unchallenged here. But it is ignored every time my critics assert that the patentee is paying a “competitor” to exit

SLIDE 3

S U M M E R 2 3 · 8 7

either ignore or assume away (e.g., the parties’ asymmetric views of the “chance” of winning the case). The unavoidable need to consider “false positives” when evaluating the competitive result of a specific lawsuit is one factor that renders the con- clusion of per se illegality, or even presumed competitive injury, untenable in my view, as I also tried to show.22 What I plainly did not suggest is a third and different inquiry, which my critics dub “objective validity”: “McDonald (p. 69) nev- ertheless asserts that there is something akin to ‘the patent’s

• bjective validity’—regardless of what a federal court may say.”

Leffler III at 78. The idea is that I believe in some platonic ideal

• f patent validity that could be used to contradict even a court

decision striking the patent down. Of course, none of this can be found in my paper. I do believe that an inquiry into the merit of the patent itself is more “objec- tive” than an inquiry into the result of a specific lawsuit (which the Lefflers suggest may turn on the judge’s bias or the death of a key witness).23 But that does not mean that a prior ruling against the patent—which I pointed out would have eternal res judicata effect against the patentee (McDonald at 74 & n.66) — would ever be rejected in a subsequent antitrust analysis. The Lefflers nonetheless build on this false premise until they exclaim, “He suggests (at 69 and 75) that the true validity of the patent, which a federal court adjudicating the patent litigation failed to ascertain, is to be determined by . . . a federal court!” Leffler III at 79. As exciting as this point may seem to the Lefflers, it entirely abandons the scenario being analyzed. Recall that the original patent litigation was settled. No one will ever know for sure how it would have come out. There is no scenario— and certainly none that I suggested—under which a second federal court will ever second-guess what the first court decided

• n the patent.24

Only by making the possibility of harm into a “property right” can Professor Shapiro indulge his presumption that payments from the patentee are anticompetitive. Let’s look at his expla- nation of why reverse payments are evidence that the patent is “weak,” and recall that he claims that my critique of those who declare patent validity “irrelevant” does not apply to him. Shapiro at 72–73. To “prove” that reverse payments are nefarious, he

• ffers an example of a settlement in which the patentee, believ-

ing it has a “75 percent chance of winning,” agrees to let the challenger enter in the last year of the patent’s remaining four- year life. After a long series of assumptions (that would justify another article), Shapiro states that “both settlement and liti- gation would thus give [the patentee] monopoly profits [of $100 million per year] 75 percent of the time and duopoly profits [of $60 million per year] 25 percent of the time.” Both scenarios would give the patentee $360 million. If the patentee then adds a reverse payment of $16 million, Shapiro concludes that the patentee is settling for less profit than it would expect from the 75 percent chance of winning (360 – 16 = 344), demonstrating that the real strength of the patent is lower—here 65 percent. Shapiro at 71–72. But try this example. The patentee believes his chance is 90

• percent. Using Shapiro’s own numbers and methodology, the

expected value of litigation is $384 million. This settlement, however, allows no entry for the life of the patent (the challenger simply agrees not to infringe) and the same reverse payment of $16 million. This yields profits to the patentee of . . . $384 mil- lion.25 In other words, the exact same reverse payment can occur in a case where the patent’s “strength” is (under Shapiro’s view) accurately perceived as 90 percent, and the patentee receives exactly the same profit as Shapiro would “expect” from litigation. Here is the dilemma. If Professor Shapiro would condemn both settlements (as I suspect he would), then he has developed a model that cannot distinguish between a patent with 65 per- cent “strength” and a patent with 90 percent “strength.” This result can be characterized in various ways, but I think “to ren- der patent validity irrelevant” is more than fair. McDonald at 69. If, on the other hand, he would not condemn both settlements, then he must concede that the same “reverse” payment (1) can be made for a vastly stronger patent than in his example, and (2) can precisely reflect the expected profits of the patent holder. The “presumption” against reverse payments thus collapses.26 No matter which result he chooses, moreover, my example shows that his statement that “the larger the cash payment, the weak- er the patent holder must have regarded its own litigation prospects” (Shapiro at 72) is simply wrong. Let me close with a word about probability and expected

• value. Contrary to Professor Shapiro’s suggestion that I am “con-

fused” about probability, his mistake (substituting the expected value of a binary outcome for the actual outcome itself) is com- monplace in antitrust litigation. In a class action, a plaintiff may define a class improperly to include members entirely uninjured. No problem, says the plaintiff, I will simply compute an “average” class-wide injury. This is when I have heard economists say, unanswerably in my view: “I can average ‘injury’ and ‘no injury,’ but I can’t make the answer meaningful.” So, here, we can (in theory) compute the expected value of the result of a specific lawsuit by “averaging” the probability of two mutually exclusive results, but we cannot claim, as does Shapiro, that it represents “the level of competition that would have prevailed, on average, had the two parties litigated the patent dispute to a resolution in the courts.” Shapiro at 70.27 It is simply a prediction of an uncertain outcome, candidly unknow-

• able. If Professor Shapiro wishes to award consumers a “prop-

erty right” (to use his term) in a portion of that uncertainty, he will need a new statute and, in my view, a justification less dependent on the use of the word “presumption.” It is one thing to label property “probabilistic”; quite another to label possibil- ities property.

1 Abraham Lincoln to Horace Greeley, August 22, 1862, reprinted in A.

Delbanco (ed.), THE PORTABLE ABRAHAM LINCOLN 239–40 (Viking 1992). I know what you’re thinking: I’m no Lincoln. That is a proposition that I do not, now and here, controvert.

2 The responses in the current issue include: Merril Hirsh, Are False Positives

Really So Negative? A Response to Kevin D. McDonald, ANTITRUST, Summer 2003, at 83; (Hirsh); Keith Leffler & Cristofer Leffler, The Probabilistic Nature

• f Patent Rights: In Response to Kevin McDonald, ANTITRUST, Summer 2003,

SLIDE 4

A R T I C L E S A N D F E A T U R E S

8 8 · A N T I T R U S T

at 77; (Leffler III); Carl Shapiro, Antitrust Analysis of Patent Settlements Between Rivals, ANTITRUST, Summer 2003, at 70 (Shapiro).

3 In re Tamoxifen Citrate Antitrust Litig., MDL No. 1408 (ILG), Memorandum

and Order, slip op. (E.D.N.Y . May 13, 2003) (granting motion to dismiss under Rule 12(b)(6)).

4 In re Ciprofloxacin Hydrochloride Antitrust Litig., 1:00-MDL-138 (DGT),

Memorandum & Order, slip op. at 117 (E.D.N.Y . May 20, 2003) (plaintiffs’ motion for partial summary judgment denied; defendants’ motion to dis- miss denied in part and granted in part). I am one of the attorneys for the defendants.

5 Reacting to this Rejoinder, the Lefflers respond that these opinions provide

“little or no support for McDonald’s position.” Leffler III at 82 n.21. By a “little” support, they may have in mind the fact that both the ALJ in Schering- Plough and the district judge in Ciprofloxacin flatly rejected the argument that “reverse” payments are per se illegal. Their most confounding statement is the claim that the Ciprofloxacin court “did ‘embrace’ a probabilistic analy- sis (see pp. 188–92).” Leffler III at 82 n.21. If you look at the pages of the

• pinion they cite, however, you will not find the words “probabilistic” or “prop-

erty,” together or apart. What you will find is the court citing the same author- ity I cited in my prior articles to explain that the procedures of Hatch- Waxman have shifted risk to the patentee in such a way that “[t]o compromise, consideration flows from the patent owner . . . to the challenger . . . and not vice versa as in a traditional context.” In re Ciprofloxacin Hydrochloride Antitrust Litig., supra note 4, at 191. This is exactly the same conclusion that the Lefflers, citing my first article on this subject, brand a “non-sequitur.” Leffler III at 82–83 n.31. According to the Lefflers, then, a court that has (a) adopted the same view I have set forth repeatedly (a non- sequitur, no less) and (b) expressly rejected their view that “reverse” pay- ments are per se illegal, has nonetheless, and by virtue of the same words, “embraced” their probabilistic analysis.

6 USM Corp. v. SPS Technologies, Inc., 694 F.2d 505, 513 (7th Cir. 1982). 7 In re IBM Peripheral EDP Devices Antitrust Litig., 481 F. Supp. 965, 1022

(N.D. Cal. 1979). See, e.g., 11 HERBERT HOVENKAMP , ANTITRUST LAW ¶ 1913c at 306 (1998) (less restrictive alternative “[i]nquiry [is] relevant only after threshold showing” of competitive harm).

8 The assumption that a license necessarily bestows greater benefits to con-

sumers is far from established, and was debunked by Judge Posner in his Brunswick decision. Brunswick Corp. v. Reigel Textile Corp., 752 F.2d 261, 267 (7th Cir. 1984).

9 Buffalo Broadcasting v. ASCAP

, 774 F.2d 917, 933 (2d Cir. 1984). The Lefflers underline the error when they conclude that my “approach simply avoids the efficiency of alternative settlements.” Leffler III at 79. In response to this point, the Lefflers have added a long paragraph asserting that they do not measure harm by comparing the settlement to (note the careful wording) “the most efficient possible outcome.” Id. (emphasis added). By the end of the paragraph, however, they must concede that they are measuring payments against “other profit-maximizing alternatives”—i.e., the “prefer- able” license they rely on (id. at 78). This is precisely the fallacy exposed in Buffalo Broadcasting.

10 Tamoxifen, supra note 3, at 24. 11 That, in any event, is the view of the Federal Circuit and others. Mallincrodt,

• Inc. v. Medipart, Inc., 976 F.2d 700, 708 (Fed. Cir. 1992) (“Should the

restriction be found to be reasonably within the patent grant, i.e., that it relates to subject matter within the scope of the patent claims, that ends the [antitrust] inquiry.”); United States v. Studiengesellschaft Kohle, 670 F.2d 1122, 1129 (D.C. Cir. 1981) (“Defendant has thus sought nothing beyond what the patent itself gave it.”).

12 “The consumer does not care how many sellers of a particular good or ser-

vice there are; he cares only that there be enough to assure him a com- petitive price and quality.” Products Liability Ins. Agency, Inc. v. Crum & Forster Ins. Cos., 682 F.2d 660, 664 (7th Cir. 1982).

13 Olympia Equip. Leasing Co. v. Western Union Tel. Co., 797 F.2d 370, 375

(7th Cir. 1986) (Posner, J.). Indeed, the prior undue emphasis of the courts

• n protecting rivals led to rules of per se liability later rejected as econom-

ically indefensible, as shown by the overruling of Schwinn in GTE Sylvania and the overruling of Albrecht v. Herald in State Oil v. Kahn. United States v. Arnold, Schwinn & Co., 388 U.S. 365 (1967), overruled, Continental T.V., Inc.

• v. GTE Sylvania Inc., 433 U.S. 36 (1977); Albrecht v. Herald Co., 390 U.S.

145 (1968), overruled, State Oil Co. v. Kahn, 522 U.S. 3 (1997). For more detail on this point, see Kevin D. McDonald, Patent Settlements and Payments that Flow the ‘Wrong’Way: The Early History of a Bad Idea, 15 ABA ANTITRUST HEALTH CARE CHRONICLE No. 4, at 2 (Winter 2002).

14 Dennis W. Carlton, A General Analysis of Exclusionary Conduct and Refusal

to Deal—Why Aspen and Kodak Are Misguided, 68 ANTITRUST L.J. 659, 674 (2001).

15 Richard A. Posner, Antitrust in the New Economy, 68 ANTITRUST L.J. 925, 930

(2001).

16 Rubber Tire Wheel Co. v. Milwaukee Rubber Works Co., 154 F. 358, 364 (7th

• Cir. 1907). See, e.g., In re Schering Plough, FTC Docket No. 9297, Initial

Decision 103 (June 27, 2002) (citing Doddridge v. Thompson, 22 U.S. 469, 483 (1824)).

17 Keith Leffler & Cristofer Leffler, Want to Pay a Competitor to Exit the Market?

Settle a Patent Infringement Case, ABA ANTITRUST SECTION ECONOMICS COMM. NEWSLETTER, Spring 2002, at 26 (Leffler I).

18 HERBERT HOVENKAMP

, ANTITRUST LAW ¶ 2040b, at 199 (2d ed. 1999).

19 RICHARD A. POSNER, ECONOMIC ANALYSIS OF LAW 91 (5th ed. 1998). 20 See, e.g., Access Telecom, Inc. v. MCI Telecomms. Corp., 197 F.3d 694, 712

(5th Cir. 1999) (conduct reducing competition in U.S. export market to Mexico for certain telephone service not redressable under antitrust laws “[i]f there is no legal U.S. export market to Mexico”); accord Action Towing and Rental v. U-Haul Int’l, 507 F. Supp. 987, 992 (E.D. La. 1981), aff’d, 683 F.2d 415 (5th Cir. 1982) (where plaintiff had “no legal or contractual right to compete,” there was “no legally protected . . . competition . . . that could be lessened”).

21 Christiansburg Garment Co. v. EEOC, 434 U.S. 412 (1978). 22 Merril Hirsh now believes that my principal argument is that all incentives

to innovate will disappear unless we find a way to purge patent litigation of false positives. Hirsh at 83–84. To the contrary, my analysis would eliminate the need to think about false positives at all: “And these problems are gra-

• tuitous. For in a proper antitrust analysis of Hatch-Waxman settlements, we

should not attempt to measure the possible outcomes of a single lawsuit (which cannot rationally exclude false positives).” McDonald at 75.

23 Leffler I, supra note 17, at 30. 24 Professor Shapiro also tries to change the question presented when he says

that I have “neglect[ed]” key issues: “A settlement that eliminates non-infring- ing competition, or competition based on the inventing around a narrow patent, could well be anticompetitive. McDonald ignores these important cases, all of which are built into the notion of patent strength employed in my analysis.” Shapiro at 74. But it is he who ignores the very explicit limi- tations I placed on the hypothetical chosen “to test the logic of those who argue the irrelevance of patent validity.” McDonald at 75 n.17. Obviously, the validity of the patent alone will not justify an agreement not to “invent around” it, or not to market non-infringing drugs (though other things might). But where, as I put it, “the patent is broad enough to preclude all generic competition and . . . the settlement’s preclusive effect is no broader than the patent itself,” the question that cannot be avoided is: what lawful com- petition has been excluded? If Professor Shapiro wishes to acknowledge that his presumption of competitive harm for “reverse” payments does not apply in this scenario, I would be delighted.

25 Under Shapiro’s methodology, expected value of litigation: (90% x $100mm

x 4 years) + (10% x $60mm x 4 years), or 360 + 24 = $384 million. Actual value of settlement: ($100mm x 4 years) less ($16 million), or 400 – 16 = $384 million.

26 Only by insisting that consumers must be given the benefit of any possibil-

ity of loss can the settlement in my example be deemed to harm competi-

• tion. The reverse payments tell us nothing about patent “strength”—only that

the consideration from the patentee is flowing in one form (payments) rather than another (a license). By declaring one of those forms of consideration suspect, Shapiro’s rule is an attempt to impose the kind of compulsory licensing that the courts have always rejected. See, e.g., King Instrs. Corp.

• v. Perego, 65 F.3d 941 (Fed. Cir. 1995).

27 See also Carl Shapiro, Antitrust Limits to Patent Settlements, 34 RAND J.

• ECON. 391, 396 (2003), available at http://faculty.haas.berkeley.edu/

shapiro/settle.pdf.