Solving the patent settlement puzzle

Texas Law Review

Volumn 91, Issue 2, 2012, Pages 283-330

  Elhauge, Einer ^a   Krueger, Alex ^b  

^a HARVARD LAW SCHOOL   (United States)

^b NONE

Author keywords

[No Author keywords available]

Indexed keywords

EID: 84871726945     PISSN: 00404411     EISSN: None     Source Type: Journal    
DOI: None     Document Type: Article

References (85)

1. These are sometimes called "pay-for-delay" settlements, but we avoid that terminology because it presupposes that the settlement entry date does "delay" entry compared to the expected entry date, which is generally the disputed issue.
2. The literature often instead compares the settlement entry date to the "expected entry date" from litigation, but the latter term is imprecise in cases where, absent settlement, the entrant would have entered at risk and thus could have been excluded well after entry if it had later lost the patent litigation. So we use the more precise term "expected litigation exclusion period." For example, suppose the entrant would enter at risk, the remaining patent term is 10 years, the patent litigation will last two years, and the entrant has a 20% chance of losing and being excluded. The expected entry date is immediate and thus would be exceeded by any settlement that excludes entry for any period at all. But this does not accurately determine whether the settlement harms consumer welfare because the expected litigation exclusion period is 20% of 8 years or 1.6 years. Thus, a settlement that excludes entry for less than 1.6 years does not harm consumer welfare.
3. Delaying entry through a reverse patent settlement is profit maximizing whenever entry reduces the joint profits of the patent holder and entrant, which is true whenever the patent holder has market power that the entrant would constrain to some degree. Further, we show below that if the reverse payment amount exceeds the patent holder's anticipated litigation costs, then the patent holder must have believed it had market power that the settling entrant would uniquely constrain.
4. In re Cardizem CD Antitrust Litig., 332 F.3d 896, 907-08 (6th Cir. 2003). In this case, Professor Elhauge filed an expert declaration for a generic defendant in which he opposed application of the per se rule.
5. Schering-Plough Corp. v. FTC, 402 F.3d 1056, 1065-66, 1075-76 (11th Cir. 2005); id. at 1071 (finding the evidence unrebutted that the settlement "entry date reasonably reflected the strength of [the patent holder's] case"); Valley Drug Co. v. Geneva Pharm., Inc., 344 F.3d 1294, 1312 (11th Cir. 2003) (requiring inquiry into whether the settlement terms exceeded patent protections, "considered in light of the likelihood of [the patent holder] obtaining such protections").
6. Valley Drug, 344 F.3d at 1306-07 (holding that such a settlement could thus be proper even if the patent were later held invalid).
7. In re Tamoxifen Citrate Antitrust Litig., 466 F.3d 187, 203-04 (2d Cir. 2006).
8. Id. at 212-13.
9. id. at 213-16 (holding that the settlement did not exceed the scope of the patent because it did not preclude "non-infringing products," did not delay other potential entrants, and allowed entry before the patent term expired); Ark. Carpenters Health & Welfare Fund v. Bayer AG, 604 F.3d 98, 106 (2d Cir. 2010) (stressing that a "settlement agreement did not exceed the scope of the patent where (1) there was no restriction on marketing non-infringing products; (2) a generic version of the branded drug would necessarily infringe the branded firm's patent; and (3) the agreement did not bar other generic manufacturers from challenging the patent"). 9. In re Ciprofloxacin Hydrochloride Antitrust Litig., 544 F.3d 1323, 1336-37 (Fed. Cir. 2008); id. at 1335 (indicating that if a reverse payment settlement created a bottleneck effect "delaying the entry of other generic manufacturers" or covered "non-infringing" products, then it would clearly lie "outside the exclusion zone of the patent"); FTC v. Watson Pharm., Inc., 677 F.3d 1298, 1312 (11th Cir. 2012). The Eleventh Circuit panel in Watson claimed its conclusion was consistent with prior Eleventh Circuit panels, but it never addressed the language in Valley Drug saying that the exclusionary scope of a patent turned on "the likelihood of [the patent holder] obtaining [patent] protections." Valley Drug, 344 F.3d at 1312. The Watson panel's effort to reconcile Schering-Plough relied on the dubious assertion that Schering-Plough's references to evaluating the "strength of the patent" merely meant the temporal length of the patent. Watson, 677 F.3d at 1311 n.8. But the Watson panel was reasonably concerned that the FTC's proposed standard in that case (whether patent victory was unlikely) failed to reliably identify whether the settlement was anticompetitive and required difficult inquiries into the probability that the patent holder would have won.
10. Id. at 1312-15.
11. In re K-Dur Antitrust Litig., 686 F.3d 197, 218 (3d Cir. 2012).
12. Brief for the United States in Response to the Court's Invitation at 10, 22, 28-32, Ark. Carpenters Health & Welfare Fund, 604 F.3d 98 (Nos. 05-2851-cv(L), 05-2852-cv (CON), 05- 2863-cv (CON)), 2009 WL 8385027, at *10, *22, *28-32. The Antitrust Division used to favor a case-by-case inquiry into whether the patent holder actually would likely have won, but now rejects that sort of objective probabilistic approach in favor of the perceived probabilistic approach described in the text above. Id. at 24-27, 26 n.9.
13. Schering-Plough Corp., 136 F.T.C. 956, 987-91, 1000-03 (2003), vacated, 402 F.3d 1056 (11th Cir. 2005).
14. Id. at 992-98.
15. See FED. TRADE COMM'N, GENERIC DRUG ENTRY PRIOR TO PATENT EXPIRATION: AN FTC STUDY vi (2002) ("Generic applicants have prevailed in 73 percent of the cases in which a court has resolved the patent dispute."); ADAM GREENE & D. DEWEY STEADMAN, RBC CAPITAL MKTS., PHARMACEUTICALS: ANALYZING LITIGATION SUCCESS RATES 1 (2010) (noting that patent holders lose 48% of the cases with generic entrants); Paul M. Janicke & LiLan Ren, Who Wins Patent Infringement Cases?, 34 AIPLA Q.J. 1, 20 (2006) (providing data that demonstrates that patent holders lose approximately 70% of the time).
16. This difficulty might be particularly acute because the Federal Circuit has exclusive jurisdiction over patent cases, while antitrust cases can go to any appellate panel. Thus, the appellate panel hearing the antitrust case might worry it lacks the expertise to predict how the Federal Circuit would decide any patent issues. See FTC v. Watson Pharm., Inc., 677 F.3d 1298, 1314-15 (11th Cir. 2012) (expressing concern that "[t]his Court and the other non-specialized circuit courts have no expertise or experience in the area" of patent law and thus "are ill-equipped to make a judgment about the merits of a patent infringement claim").
17. However, our proof below shows this possibility exists only when the reverse payment amount is lower than the patent holder's anticipated litigation costs, absent a judgment-proof entrant or a separate procompetitive justification.
18. In re Tamoxifen Citrate Antitrust Litig., 466 F.3d 187, 211 (2d Cir. 2006).
19. Ark. Carpenters Health & Welfare Fund v. Bayer AG, 604 F.3d 98, 108-10 (2d Cir. 2010).
20. Schering-Plough Corp., 136 F.T.C. 956, 987 (2003), vacated, 402 F.3d 1056 (11th Cir. 2005); Brief for the United States, supra note 13, at 21-22.
21. See John H. Barton, Patents and Antitrust: A Rethinking in Light of Patent Breadth and Sequential Innovation, 65 ANTITRUST L.J. 449, 450 (1997) ("The patent-antitrust analysis has always had to take into account and balance benefit to consumers by maintaining the competitive structure of existing markets against benefit to consumers by permitting the intellectual property rights system to provide an incentive for research toward new and improved products.").
22. See SUZANNE SCOTCHMER, INNOVATION AND INCENTIVES 100-03 (2004); Partha Dasgupta & Joseph Stiglitz, Uncertainty, Industrial Structure, and the Speed of R&D, 11 BELL J. ECON. 1, 18 (1980); Pankaj Tandon, Rivalry and the Excessive Allocation of Resources to Research, 14 BELL J. ECON. 152, 152, 156-57 (1983). Such a system will also maximize overall total welfare because competing innovators will keep spending on ex ante investments until their investment costs equal their expected ex post profits, so that the profits to patent holders wash out ex ante.
23. See supra note 24.
24. For a theoretical model proving that this is possible, see generally Michele Boldrin & David K. Levine, A Model of Discovery, 99 AM. ECON. REV. 337 (2009).
25. For empirical work showing that expanding patent protections have had net negative effects on patent filings and suppressed later innovations, see generally Josh Lerner, The Empirical Impact of Intellectual Property Rights on Innovation: Puzzles and Clues, 99 AM. ECON. REV. 343 (2009); Fiona Murray et al., Of Mice and Academics: Examining the Effect of Openness on Innovation (Nat'l Bureau of Econ. Research, Working Paper No. 14819, 2009); Heidi L. Williams, Intellectual Property Rights and Innovation: Evidence from the Human Genome (Nat'l Bureau of Econ. Research, Working Paper No. 16213, 2010).
26. In re Tamoxifen Citrate Antitrust Litig., 466 F.3d 187, 211 (2d Cir. 2006).
27. See, e.g., Ian Ayres & Gideon Parchomovsky, Tradable Patent Rights, 60 STAN. L. REV. 863, 864 (2007) (arguing that issuance of too many patents chills innovation).
28. See, e.g., Joshua L. Sohn, Can't the PTO Get a Little Respect?, 26 BERKLEY TECH. L.J. 1603, 1605 (2011) (arguing that federal courts do not give PTO decisions enough deference during invalidity contests).
29. For the same reasons, we think antitrust law should assume the optimality of the Hatch-Waxman Act, which gives pharmaceutical patent holders the additional exclusion right of an automatic 30-month stay on generic entry, which helps incentivize patent holders to incur the costs of new drug applications that can secure FDA approval. 21 U.S.C. § 355(j)(5)(D)(i)-(ii) (2006). Thus, in a Hatch-Waxman case, we would treat the residual patent period as starting once that 30- month stay expires because monopoly profits before the stay expires are part of the special intellectual property reward for investing in new drug applications, which are valuable even if the patent proves invalid. Although the Hatch-Waxman Act provides this limited exclusion right, it nowhere approves anticompetitive settlements that extend beyond that 30-month exclusion right.
30. Altering the model to include discount rates, make profitability differ over time, or both would not change any of the conclusions in the proof but would significantly complicate the mathematical formulas. In fact, adding either of these complications would only strengthen our proof. Adding discount rates would reduce the net present value of the patent holder's anticipated litigation costs but would not reduce the net present value of any reverse payment made at time 0. Discounting future profit streams would only increase the extent to which an entrant would be willing to delay entry in exchange for an upfront settlement payment and reduce the extent to which a patent holder is willing to speed up entry.
31. See, e.g., Timothy F. Bresnahan & Peter C. Reiss, Entry and Competition in Concentrated Markets, 99 J. POL. ECON. 977, 984 (1991); Richard G. Frank & David S. Salkever, Generic Entry and the Pricing of Pharmaceuticals, 6 J. ECON. & MGMT. STRATEGY 75, 84 (1997); David Reiffen & Michael R. Ward, Generic Drug Industry Dynamics, 87 REV. ECON. & STAT. 37, 43 (2005).
32. For example, if the remaining patent term is 100 months, and the parties expect the patent litigation to last 10 months, then L = 0.1. We assume both parties share the same expected litigation duration L because it makes the mathematical model easier to understand but does not change any of the relevant conclusions. If we instead assumed that the entrant was relatively pessimistic about litigation length (so that LE > LP), that would only widen the range of possible settlement entry dates that (without any reverse payment) can provide settlement payoffs to both the entrant and patent holder that exceed their litigation payoffs. If we instead assumed the entrant was relatively optimistic about litigation length (so that LE < LP), that would only increase the extent to which a settlement exclusion period that exceeds the patent holder's estimate of the expected litigation exclusion period will exceed the entrant's estimate of that expected litigation exclusion period.
33. This formula assumes that the entrant has sufficient assets to pay damages, i.e., that it is not judgment proof. We discuss below the case of a judgment-proof entrant. See infra section II(B)(2).
34. Our model assumes firms are risk neutral, but as we show below, our conclusions do not depend on this assumption.
35. The formula in the above text assumes the entrant has sufficient assets to pay any patent damages. If the entrant does not, then it is judgment proof to some extent, which does provide a possible ground for rebuttal that we discuss below in section II(B)(2). The formula in the text also assumes that damages are not trebled for willful infringement. Because we are talking here about a weak patent, where by definition the odds are relatively low that a court would sustain the patent claims, it is very unlikely willful infringement would ever be found, especially because willful infringement is found in only 2.1% of all patent disputes. Kimberly A. Moore, Empirical Statistics on Willful Patent Infringement, 14 FED. CIR. B.J. 227, 234 (2005). In any event, the prospect that damages might be trebled would either: (1) raise damages high enough to deter entry, in which case the strong patent proof would apply; or (2) raise the patent-holder returns from litigation if the patent remained too weak to deter entry, which would make the patent holder demand an even larger settlement exclusion period, worsening all the effects predicted by the model.
36. See, e.g., Robert D. Willig & John P. Bigelow, Antitrust Policy Toward Agreements That Settle Patent Litigation, 49 ANTITRUST BULL. 655, 659-62, 667-77 (2004).
37. 21 U.S.C. § 355(j) (2006).
38. See, e.g., Michael A. Carrier, Unsettling Drug Patent Settlements: A Framework for Presumptive Illegality, 108 MICH. L. REV. 37, 71 (2009).
39. See AIPLA, REPORT OF THE ECONOMIC SURVEY 2011, at 35 (2011) [hereinafter 2011 AIPLA REPORT] (stating that, as of 2011, median litigation costs for a patent infringement suit with more than $25 million at stake were $3 million through the end of discovery and $5 million in total); Meredith Addy, Appellate Strategy Before the U.S. Court of Appeals for the Federal Circuit, in PATENT LITIGATION, NEGOTIATION, AND SETTLEMENT: LEADING LAWYERS ON STRATEGIES FOR EFFECTIVELY RESOLVING PATENT DISPUTES 7, 8 (2006), available at http://www.brinkshofer.com/files/201.pdf ("Generally, once a patent case has gone through a district court trial, it has already cost, on average, $3 to $5 million, or more. Comparatively, the cost of appeal is far less: perhaps a few hundred thousand dollars for an easy case, a few million for a complicated one, but almost always exponentially less than the initial litigation.").
40. See Susan Decker, Apple Patent Battles Create Lawyer Boon at $1,200 an Hour, BLOOMBERG (Aug. 23, 2012, 3:00 PM), http://www.bloomberg.com/news/2012-08-23/apple-patent-battles-create-lawyer-boon-at-1-200-an-hour.html.
41. 2011 AIPLA REPORT, supra note 40, at 35-36, I-154.
42. Id. at I-154.
43. The reports from previous years have similar figures, with the highest 75th percentile reported in any year being $11.5 million for cases in the Los Angeles region in 2009. See AIPLA, REPORT OF THE ECONOMIC SURVEY 2009, at I-129 (2009) [hereinafter 2009 AIPLA REPORT].
44. In re Tamoxifen Citrate Antitrust Litig., 466 F.3d 187, 190, 194 (2d Cir. 2006) (noting the $66.4 million total given to two generics).
45. The patent holder agreed to pay one generic $6 million every three months and the other generic $4.5 million per month for the period from March 31, 1998 to March 1, 2000. Valley Drug Co. v. Geneva Pharm., Inc., 344 F.3d 1294, 1300-01 (11th Cir. 2003). Because of an FTC investigation, the parties terminated this settlement agreement early on August 13, 1999, so $39 million of this was not actually paid out, but that was not part of the original settlement agreement and thus not relevant to the inference at issue here. Id.
46. The patent holder agreed to pay $60 million to one generic and $19-$30 million annually to another generic for the 10.75 years from January 2006 to September 2015. FTC v. Watson Pharm., Inc., 677 F.3d 1298, 1304-05 (11th Cir. 2012). The patent holder also agreed to pay one generic $12 million for backup manufacturing assistance, but if we assume that this $12 million constituted fair consideration for that assistance, then it should be excluded from the calculation of the reverse payment amount. Id.
47. Ark. Carpenters Health & Welfare v. Bayer AG, 604 F.3d 98, 102 n.8 (2d Cir. 2010).
48. See Richard D. Margiano, Cost and Duration of Patent Litigation, MANAGING INTELL. PROP., Feb. 2009, at 150, 150.
49. See Decker, supra note 41.
50. The Third Circuit has recognized the need to allow this sort of rebuttal for other procompetitive justifications. In re K-Dur Antitrust Litig., 686 F.3d 197, 218 (3d Cir. 2012).
51. 21 U.S.C. § 355(j)(5)(D)(i)-(ii) (2006).
52. Aiding this sort of rebuttal evidence was other evidence in that case that minimized the possible anticompetitive effects that needed to be rebutted. The Cardizem settlement differed from the sort we model in this Article because it did not end the patent litigation and set a fixed settlement entry date. Rather, it was an interim settlement that required the parties to continue the patent litigation, precluded entry only during the litigation and only if the litigation did not last too long, and allowed the generic to keep the reverse payment only if it won the litigation. Further, in that case anticompetitive effects were undermined by strong evidence that: (1) the entrant would not have entered at risk anyway, so that such a purely interim settlement did not preclude any entry by the settling generic; and (2) no other generic entry was delayed because (a) under the rules that then prevailed, the settling generic had to win the patent litigation to preclude other generics and (b) no other generic received FDA approval in time to enter any earlier anyway. Given this evidence, the FTC concluded that the settlement had not actually delayed any generic entry. FED. TRADE COMM'N, DOCKET NO. 9293, ANALYSIS TO AID PUBLIC COMMENT 4 (2001), available at http://www.ftc.gov/os/2001/04/hoechstanalysis.pdf.
53. See FTC v. Ind. Fed'n of Dentists, 476 U.S. 447, 460-61 (1986) ("'[P]roof of actual detrimental effects, such as a reduction of output,' can obviate the need for an inquiry into market power, which is but a 'surrogate for detrimental effects.'"); Toys "R" Us, Inc. v. FTC, 221 F.3d 928, 937 (7th Cir. 2000) (Wood, J.) ("[T]he share a firm has in a properly defined relevant market is only a way of estimating market power, which is the ultimate consideration. The Supreme Court has made it clear that there are two ways of proving market power. One is through direct evidence of anticompetitive effects." (citations omitted)); Allen-Myland, Inc. v. Int'l Bus. Machs. Corp., 33 F.3d 194, 209 (3d Cir. 1994) (Easterbrook, J.) ("Market share is just a way of estimating market power, which is the ultimate consideration. When there are better ways to estimate market power, the court should use them." (quoting Ball Mem'l Hosp., Inc. v. Mut. Hosp. Ins., Inc., 784 F.2d 1325, 1336 (7th Cir. 1986))); United States v. Baker Hughes Inc., 908 F.2d 981, 992 (D.C. Cir. 1990) (Thomas, J., joined by Ruth Bader Ginsburg, J.) (same); see also IIB PHILLIP E. AREEDA, HERBERT HOVENKAMP & JOHN L. SOLOW, ANTITRUST LAW: AN ANALYSIS OF ANTITRUST PRINCIPLES AND THEIR APPLICATION 108 (3d ed. 2007) ("'[D]irect' indicators of market power. can be independent of market definition and are sometimes superior to it. [M]arket definition may not be necessary to prove market power.").
54. See FTC v. Watson Pharm., Inc., 677 F.3d 1298, 1315 (11th Cir. 2012) (making such a claim).
55. See supra note 22; see infra Part IV.
56. This data is drawn from Professor Scott Hemphill's invaluable survey of 143 patent settlements from 1984 to 2008. See generally C. Scott Hemphill, An Aggregate Approach to Antitrust: Using New Data and Rulemaking to Preserve Drug Competition, 109 COLUM. L. REV. 629 (2009); C. Scott Hemphill, Drug Patent Settlements Between Rivals: A Survey (March 12, 2007) (unpublished manuscript), available at http://academiccommons.columbia.edu/catalog/ac%3A129331. We use the date of the settlement agreement, rather than the expiration of the 30-month Hatch-Waxman stay, as the best indicator of when generic entry was first possible because sometimes the Hatch-Waxman stay gets extended for other reasons, like pediatric exclusivity. If one instead uses the expiration of the 30-month stay, the residual patent period would be 93.2 months. Because our focus is on the prospective issue of how likely it is that settlements that exclude entry would be anticompetitive if no reverse payment were allowed, we combine results from settlements that did and did not have a reverse payment. If one instead wanted to ask about the likelihood that past settlements without any reverse payment were anticompetitive, then the residual patent period for only those settlements would be 75.4 months and the average monthly sales figure would be $42.4 million. This would not alter our qualitative conclusions. See infra section III(A)(3) (showing that cutting the residual patent period and annual profit level in half would actually make it somewhat more likely that settlements without reverse payments are anticompetitive).
57. See GREENE & STEADMAN, supra note 17, app. C.
58. Frank & Salkever, supra note 32, at 84 fig.3 (reporting 70%); Reiffen & Ward, supra note 32, at 43-44 (reporting 88%).
59. In fact, incumbents increase their drug prices slightly in response to generic entry, but because the price increase is only 0.7% with one generic entrant, we treat it as unchanged. Frank & Salkever, supra note 32, at 87. Apparently, the incumbent makes more money by keeping its price high and selling only to price-insensitive customers than the incumbent would make if it lowered its price to compete with the generic for price-sensitive customers.
60. Reiffen & Ward, supra note 32, at 43.
61. Frank & Salkever, supra note 32, at 89.
62. See Gautier Duflos & Frank R. Lichtenberg, Does Competition Stimulate Drug Utilization? The Impact of Changes in Market Structure on US Drug Prices, Marketing and Utilization, 32 INT'L REV. L. & ECON. 95, 106-07 (2012) (concluding that net volume is unchanged by entry into drug markets because entry leads to a decline in both prices and marketing expenditures, which "approximately offset" each other's effects on output).
63. Because the patent holder profit per sale is unchanged, PY = 55% of PN. The generic who is a single entrant has a price that is 79% of the patent holder's with the same marginal cost of 20%, and thus earns 59% of the patent holder gross sales for 45% of volume, which means average monthly profits of 59% of 45% of $72.46 million = $19.2 million. Thus, if it could obtain those profits for the entire residual patent period, it would get $1.732 billion.
64. Because the suggestion is to allow only buyer patent actions to invalidate the patent prospectively, one need not worry that a risk of paying damages in such a buyer class action would deter the patent holder from ever entering into a settlement with the rival.
65. See Michael A. Carrier, Unsettling Drug Patent Settlements: A Framework for Presumptive Illegality, 108 MICH. L. REV. 37, 75-76 (2009).
66. Herbert Hovenkamp, Mark Janis & Mark A. Lemley, Anticompetitive Settlement of Intellectual Property Disputes, 87 MINN. L. REV. 1719, 1720, 1759 (2003).
67. Carl Shapiro, Antitrust Limits to Patent Settlements, 34 RAND J. ECON. 391, 408 (2003).
68. Professor Shapiro does present proofs on other issues but not on this conjecture.
69. Hovenkamp, Janis & Lemley, supra note 65, at 1734-35.
70. Shapiro, supra note 65, at 408.
71. Carrier, supra note 65, at 77.
72. Professor Carrier would also allow rebuttal if a cash-strapped generic needs cash quickly. Id. To the extent he means to rely on varying risk aversion, we would not allow rebuttal. To the extent he means that the generic might be judgment-proof, we agree with that possible ground for rebuttal, as limited by the conditions we prove are necessary to establish it.
73. Shapiro, supra note 65, at 396, 407-08.
74. Shapiro mistakenly conflates the expected litigation exclusion period with the optimal patent exclusion period, id. at 396, but as we show above, the former can be less than the latter when at-risk entry would have occurred without settlement.
75. Id. at 397.
76. Hovenkamp, Janis & Lemley, supra note 65, at 1762-63. They do suggest another possible rebuttal consisting of evidence that a reverse payment was actually made, but such evidence would mean that this presumption does not apply in the first place.
77. Carl Shapiro, Antitrust Analysis of Patent Settlements Between Rivals, ANTITRUST, Summer 2003, at 70, 72.
78. Carrier, supra note 65, at 76-77.
79. Daniel A. Crane, Exit Payments in Settlement of Patent Infringement Lawsuits: Antitrust Rules and Economic Implications, 54 FLA. L. REV. 747, 779-96 (2002).
80. Thomas F. Cotter, Refining the "Presumptive Illegality" Approach to Settlements of Patent Disputes Involving Reverse Payments: A Commentary on Hovenkamp, Janis & Lemley, 87 MINN. L. REV. 1789, 1807, 1812 & n.92 (2003).
81. Roger D. Blair & Thomas F. Cotter, Are Settlements of Patent Disputes Illegal Per Se?, 47 ANTITRUST BULL. 491, 533-34 (2002) (reporting the views of just Professor Blair).
82. See Willig & Bigelow, supra note 37, at 659-62, 667-77.
83. Id. at 671.
84. Much of their analysis actually addresses a different question: whether a reverse payment might be necessary for patent settlement without showing that settlement would actually be desirable. Our proof shows that although this is true, a reverse payment that exceeds litigation costs is necessary only for undesirable settlements.
85. Id. at 662, 677.