In 2018 Assistant Attorney General Makan Delrahim proclaimed the New Madison Approach (“NMA”) to the interface between antitrust and intellectual property. He specifically addressed disputes involving standard essential patents subject to commitments to license on fair, reasonable, and non-discriminatory terms. After the Ninth Circuit’s Qualcomm decision, which implicitly embraced the NMA, controversy continues to rage over the role of antitrust in these disputes. This paper argues that the NMA and Qualcomm decision are correct and that these disputes do not present a competition issue. Antitrust litigation has been a distraction from the root cause of the disputes: There is neither a consensus nor a controlling edict on the patent owners’ fair share of the bounty from standards-based technology. Opponents of the NMA believe that the patent owners’ fair share is tiny, and proponents of the NMA doubt that.

By Gregory J. Werden[1]


On March 16, 2018 Assistant Attorney General Makan Delrahim proclaimed the New Madison Approach (“NMA”) to the interface between antitrust and intellectual property.[2] The pattern for the name was the New Brandeis Movement, which advocates broad attacks on big business.[3] James Madison’s name was taken because he was most responsible for the Constitution’s Intellectual Property Clause[4] and his advocacy of patent protection counterpoints views of Thomas Jefferson quoted in leading patent decisions.[5]

AAG Delrahim addressed disputes involving standard essential patents (“SEPs”) on which standard setting organizations (“SSOs”) extracted commitments to fair, reasonable, and non-discriminatory (“FRAND”) licensing.[6] He sought to distance the Trump Administration from the Obama Administration’s sympathies to contentions that SEP holders were engaging in “patent hold-up” by waiting until standard implementers made substantial sunk investments then demanding high royalties.

As articulated by AAG Delrahim, the NMA holds that:

(1) patent hold-up is fundamentally not an antitrust problem, and therefore antitrust law should not be used as a tool to police contractual commitments patent holders make to SSOs; (2) SSOs should not become vehicles for implementers to skew conditions in their favor when incorporating a patented technology; (3) SSOs and courts should have a very high burden before adopting rules that severely restrict the right of patent holders to exclude or—even worse—adopting rules that amount to a de facto compulsory licensing scheme; (4) a unilateral and unconditional refusal to license a patent should be considered per se legal from the perspective of the antitrust laws.[7]

A few months after becoming AAG in charge of the Antitrust Division, Makan Delrahim gave a speech asserting that concerted action by standard setting organizations can raise antitrust concerns, but antitrust has little or no role to play in FRAND royalty disputes.[8] He also argued that “the hold-out problem is a more serious threat to innovation than the hold-up problem.”[9] SEP licensees engage in “patent hold-out” by refusing to pay a reasonable royalty. AAG Delrahim explained that royalties are negotiated after inventors incur all unrecoverable development costs but before implementers incur some unrecoverable implementation costs, and this asymmetry makes hold-out more serious than hold-up.

Following the speech, an industry group wrote to AAG Delrahim, arguing that his “novel approaches” “threaten US industry and consumer interests, harm US innovation, and interfere with parties’ right to contract.”[10] Unpersuaded, AAG Delrahim went on to announce the NMA. The following day, 77 academics and former government officials expressed “concern.”[11] They asserted eight propositions, some of which could have benefited from additional exposition. For example, they asserted that “patentees that obtain or maintain monopoly power as a result of breaching a FRAND commitment present a standard monopolization case,” but they did not address the D.C. Circuit’s Rambus decision (discussed below), which is rather to the contrary.[12]

The bedrock issue, which has not gotten the attention it deserves, is the appropriate reward to owners of SEPs. AAG Delrahim doubted that they were getting their fair share and sought “to put implementers and patent holders on equal footing.”[13] Others believe that SEP owners should get just a tiny share of the revenue generated by the sale of smartphones and countless other technological devices that enormously benefit consumers and businesses. Spurred by this belief, many commentators urge the use of antitrust law to address what they believe to be a significant hold-up problem.[14]

Because higher royalties tend to reduce the rewards to device makers and increase device prices (if only by a few dollars on an expensive device), commentators perceive threats to competition and consumers in the efforts of SEP owners to get a larger share. The NMA keeps antitrust in its lane,[15] while its opponents steer antitrust to where they conjecture that it can do some good. Herbert Hovenkamp argues that “when one party is undermining another party’s investments in a way that is conducive to reduced output and higher prices, antitrust intervention is appropriate.”[16] And Carl Shapiro & Mark Lemley contend that “antitrust still has an important role to play when contract law and anti-fraud laws fail to fully address the patent holdup problem.”[17]

Opponents of the NMA reject patent hold-out. Shapiro & Lemley assert that “[p]atent holdout is incoherent” because the vast majority of infringement defendants are unaware of the patent they allegedly infringe.[18] Standards implementers, however, generally know they are using technologies claimed by SEPs, since SSOs require disclosure. Shapiro & Lemley assert that: “Those who express concerns about patent holdout seem to want to increase the returns to patent holders whose inventions add little or no incremental value.”[19] But they ignore the enormous value of some inventions. Shapiro & Lemley argue that the hold-up problem exists because a patent holder “has the law on its side.”[20] Yet the law has systematically denied SEP holders’ requests for preliminary injunctions since the 2006 MercExchange decision.[21]

In rejecting patent-hold out, Hovenkamp focuses on scenarios not normally included within that rubric — concerted conduct by SSOs to exclude particular technologies or extract below-market royalties.[22] Hovenkamp devotes a single paragraph to the hold-out problem in royalty negotiations, and he argues that: “Refusing to accept and pay for a license on an untested patent is not abuse of the system.”[23] He is correct to the extent that he argues that legitimate reasons exist for not concluding licensing agreements sooner, and to the extent that he argues that SEPs could be invalid. But these circumstances set the stage for the opportunism by licensees.

Those opposed to the NMA contend that SEP holders violate antitrust law in royalty disputes, but only one theory of an antitrust violation has been blessed by an appeals court, and that theory involves harm to the competition among technologies in setting a standard. The Third Circuit’s Broadcom decision held that “a patent holder’s intentionally false promise to license essential proprietary technology on FRAND terms,” the SSO’s “reliance on that promise when including the technology in a standard,” and “the patent holder’s subsequent breach of that promise, is actionable anticompetitive conduct.”[24]

Complaints asserting a Broadcom theory can survive dismissal motions, but the theory is unlikely to be sustainable at trial. The underlying fraud on which the theory relies could be committed only by a company well aware that its technology offers little incremental value over the best alternative technology. A company believing instead that its technology offers substantial value could truthfully promise to license on FRAND terms while intending to demand a large royalty. When the Third Circuit decided Broadcom, there was no basis for believing that a large royalty could not also be fair and reasonable.

Another limitation of the Broadcom theory was made clear by the D.C. Circuit’s Rambus decision. The court set aside the FTC’s administrative determination that Rambus had violated Section 2 of the Sherman Act by deceiving an SSO to avoid a FRAND commitment.[25] Critically, the FTC had no idea whether the SSO would have selected a different technology in the absence of deception by Rambus. The D.C. Circuit rejected the FTC’s contention that, “because the ability to profitably restrict output and set supracompetitive prices is the sine qua non of monopoly power, any conduct that permits a monopolist to avoid constraints on the exercise of that power must be anticompetitive.”[26]

The FTC’s suit against Qualcomm provided an opportunity to address hold-up and hold-out without any suggestion of misbehavior in the standard setting process. In ruling for the FTC, Judge Koh implicitly rejected the NMA.[27] In reversing, the Ninth Circuit implicitly embraced it.[28] The Ninth Circuit’s decision was controversial and its timing was unfortunate: The window during which the FTC could have sought Supreme Court review closed before Biden appointees were in place at the FTC and Department of Justice.

Qualcomm pioneered cellular technology and contributed to each subsequent generation. Qualcomm now derives much of its revenue from licensing its patents, which include SEPs subject to FRAND commitments. Qualcomm also earns substantial revenue from the sale of “modem chips,” which cellular phones and similar devices use to transmit and receive data. The FTC’s case presumed that licensing rival suppliers of modem chips would exhaust Qualcomm’s patent rights, but a careful parsing of Qualcomm’s patent claims might indicate that a device incorporating a licensed modem chip still needs a Qualcomm license.

Qualcomm’s business model was to license device makers and to pressure them to take licenses with a “no license, no chips” policy, under which Qualcomm sold chips only to licensed device makers. To comply with FRAND obligations, Qualcomm pledged not to assert its patents against rival suppliers of modem chips. If device makers needed Qualcomm licenses even though they used licensed chips, the pledge was equivalent to royalty-free licenses to rival suppliers of modem chips. If device makers did not need Qualcomm licenses when they used licensed chips, the pledge avoided patent exhaustion. The details of Qualcomm’s patent claims, therefore, perhaps should have been at issue in the case, but they were not.

The FTC convinced the district court that Qualcomm’s “no license, no chips” policy violated Section 2 of the Sherman Act, but the Ninth Circuit reversed in a decision that some commentators find deeply troubling.[29] Whatever one thinks about the propriety of Qualcomm’s conduct, the FTC did not advance a sound basis under Section 2 of the Sherman Act for enjoining the “no license, no chips” policy,[30] so the Ninth Circuit was justified in vacating the injunction and in ruling that the FTC articulated no “cogent theory of anticompetitive harm.”[31]

The FTC argued that device makers were paying Qualcomm both a legitimate “royalty” and an illegitimate “surcharge,” which they were forced to pay by the “no license, no chips” policy. And the FTC argued that the “surcharge” caused rival suppliers of modem chips to get less for their chips. But proving these things did not establish a Section 2 violation under existing law.[32] Antitrust law distinguishes harm to competitors from harm to competition, and the FTC did not effectively explain why the lower chip prices implied harm to competition.

If the FTC was right about how modem chip prices were determined, any Qualcomm royalty, no matter how low or who paid it, reduced rival suppliers’ modem chip prices, but the FTC contended that Sherman Act liability arose only from charging more than an unknowable legitimate “royalty.” The Ninth Circuit reasonably declined “to adopt a theory of antitrust liability that would presume anticompetitive conduct any time a company could not prove that the ‘fair value’ of its SEP portfolios corresponds to the prices the market appears willing to pay for those SEPs in the form of licensing royalty rates.”[33]

Shapiro & Lemley argue that the Ninth Circuit erred as a matter of economics by declaring Qualcomm’s licensing “chip-supplier neutral” because “the royalty surcharge reduces the gains from trade between an OEM and a rival modem-chip supplier but does not reduce the gains from trade between the OEM and Qualcomm.”[34] But the court rightly focused on whether Qualcomm’s licensing distorted supplier choice, and it did not. The FTC relied on precedent holding that Section 2 had been violated by a pricing scheme distorting supplier choice.

Shapiro & Lemley argue that the Ninth Circuit erred as a matter of law by rejecting “the well-established principle that harming customers can be a way of harming competition,”[35] but they neither elaborate this “principle” nor cite any authority for it. Harming competition normally harms consumers, but causation running the opposite direction requires explanation. Shapiro & Lemley charge that the Ninth Circuit assumed away the “surcharge,” but the court held that it was “an issue that sounds in patent law, not antitrust law.”[36] Shapiro & Lemley argue that antitrust law should have been applied because that would have benefited consumers, but the Ninth Circuit rightly held that there must be a limiting principle.

Hovenkamp laments what he clearly views as a wrongheaded decision by the Ninth Circuit.[37] Like Shapiro & Lemley, he pillories the court for ignoring obvious consumer harm,[38] but he does not explain how the consumer harm was associated with harm to competition. Hovenkamp asserts that the “no license, no chips” policy was “classical tying,”[39] but the FTC did not pursue a tying theory. He asserts that the Ninth Circuit misread Aspen Skiing,[40] but the FTC disclaimed reliance on it. Unlike Shapiro & Lemley, Hovenkamp does not defend the legal theory the FTC actually relied upon.

Hovenkamp concludes, in part:

Oversight of FRAND obligations is one area where it is critical for the courts to keep an eye on longer run concerns for innovation. FRAND has evolved into a highly successful but nevertheless vulnerable mechanism for facilitating joint innovation and product development. . . . That system will be undermined, however, if one firm is able to renege on its voluntarily entered obligations, because others will then do the same. The regime of collaborative innovation that FRAND contemplates would very likely fall apart, and at great harm to competition and economic welfare. The Ninth Circuit’s 2020 Qualcomm decision indicates that this fear is more than fanciful.

The Qualcomm decision does indicate that there is a problem, but it has nothing to do with competition or antitrust. No antitrust plaintiff has ever sustained an FRAND theory both on the law and on the facts. What threatens to undermine innovation and SSO collaboration is the failure to address the core question: What is the SEP owners’ fair and reasonable share of the bounty from devices that standards-based technologies enable? Patent hold-up and patent hold-out problems would be trivial if SSOs, or society at large, would just make the call so licensors and licensees could anticipate the outcome of FRAND litigation. Antitrust litigation over FRAND royalty disputes is failed attempt to avoid getting at the root of the problem.

1 The author retired in 2019 from his position as Senior Economic Counsel, Antitrust Division, U.S. Department of Justice. Before retirement, the author had a hand in articulating the New Madison Approach.

2 The speech was published as: Makan Delrahim, The “New Madison” Approach to Antitrust and Intellectual Property Law, 1 J.L. & Innovation 1 (2019).

3 See Lina A. Khan, The New Brandeis Movement: America’s Antimonopoly Debate, 9 J. Eur. Competition L. & Practice 131 (2018).

4 On August 18, 1787, Madison submitted to the Constitutional Convention a long list powers to be granted the legislature, including “To grant patents for useful inventions” and “To secure to Authors exclusive rights for a certain time.” On September 5, the Committee on Detail, of which Madison was a member, proposed the language that was adopted in the Constitution. Madison also campaigned in favor of the Intellectual Property Clause in The Federalist No. 43.

5 The Supreme Court quoted Jefferson’s views on what is essentially patentability. See Bonito Boats, Inc. v. Thunder Craft Boats, Inc., 489 U.S. 141, 147 (1989); Diamond v. Chakrabarty, 447 U.S. 303, 308–09 (1980); Graham v. John Deere Co. of Kansas City, 383 U.S. 1, 7–11 (1966).

6 The author’s views on the subject were set out in Gregory J. Werden & Luke M. Froeb, Why Patent Hold-Up Does Not Violate Antitrust Law, 27 Tex. Intell. Prop. L.J. 1 (2019).

7 Delrahim, supra note 2, at 1–2.

8 Makan Delrahim, Take It to the Limit: Respecting Innovation Incentives in the Application of Antitrust Law, remarks at USC Gould School of Law (Nov. 10, 2017),

9 AAG Delrahim relied on: Luke Froeb h& Mikhael Shor, Innovators, Implementers, and Two-Sided Hold-Up, Antitrust Source, August 2015; Bernhard Ganglmair, Luke M. Froeb & Gregory J. Werden, Patent Hold-Up and Antitrust: How a Well-Intentioned Rule Could Retard Innovation, 60 J. Indus. Econ. 249 (2012); Richard A. Epstein & Kayvan B. Noroozi, Why Incentives for “Patent Holdout” Threaten to Dismantle FRAND, and Why it Matters, 32 Berkeley Tech. L.J. 1381 (2018).

10 Industry Letter to DOJ Regarding Standards, Innovation and Licensing (Jan. 24, 2018),

11 Letter from Michael A. Carrier, Timothy J. Muris, et al. to Makan Delrahim (May 17, 2018), The letter must have been in preparation for some time.

12 Rambus Inc. v. FTC, 522 F.3d 456 (D.C. Cir. 2008).

13 An Interview with Makan Delrahim, former Assistant Attorney General for the Department of Justice Antitrust Division, The Federalist Society (Mar. 22, 2021), See, e.g. Letter of Makan Delrahim to Sophia A. Muirhead, IEEE (Sept. 10, 2020), (on the Antitrust Division website under the heading “comments to states and other organizations”); Brief for the United States, HTC Corp. v. Telefonaktiesbolaget LM Ericsson, No. 19-40566 (5th Cir., Oct. 30, 2019),

14 Shapiro & Lemley purport to present “empirical evidence” demonstrating that patent hold-up is significant, but they show only that substantial sunk investments are associated with implementing technology standards and that industry participants take steps to guard against opportunistic behavior. Carl Shapiro & Mark A. Lemley, The Role of Antitrust in Preventing Patent Holdup, 168 U. Pa. L. Rev. 2019, 2035–36 (2020).

15 Under AAG Delrahim, the Justice Department advocated the NMA in statements of interest filed in district courts and amicus briefs filed in appeals courts. E.g. Statement of Interest of the United States, Cont’l Auto. Sys., Inc. v. Avanci, LLC, 485 F. Supp. 3d 712 (N.D. Tex. 2020),

16 Herbert Hovenkamp, FRAND and Antitrust, 105 Cornell L. Rev. 1683, 1720 (2020). See id. at 1743 (“Among the various legal tools for policing the FRAND process antitrust is only one, but it is an important one and has its own unique requirements and tools for analysis.”).

17 Shapiro & Lemley, supra note 14, at 2058.

18 Id. at 2049.

19 Id.

20 Id. at 2050.

21 eBay Inc. v. MercExchange, L.L.C., 547 U.S. 388 (2006) (holding the usual principles of equity govern injunctions for patent infringement). Patent holders have even met with difficulty in obtaining permanent injunctions following judgments of infringement. See Apple Inc. v. Samsung Elects. Co., 809 F.3d 633 (Fed. Cir. 2015) (reversing denial of permanent injunction).

22 Hovenkamp, supra note 16, at 1729–32.

23 Id. at 1733.

24 Broadcom Corp. v. Qualcomm Inc., 501 F.3d 297, 314 (3d Cir. 2007).

25 Rambus Inc. v. FTC, 522 F.3d 456 (D.C. Cir. 2008).

26 Id. at 466.

27 FTC v. Qualcomm Inc., 411 F. Supp. 3d 658 (N.D. Cal. 2019).

28 FTC v. Qualcomm Inc., 969 F.3d 974 (9th Cir. 2020).

29 See e.g. Hovenkamp, supra note 16, at 1685–86, 1704, 1743–44; Shapiro & Lemley, supra note 14, at 2058–60.

30 The case also involved exclusive dealing allegations, which might have had merit. But the alleged exclusive dealing was not ongoing conduct, and the FTC did not stress these allegations on appeal. The Ninth Circuit seemed well justified in holding that the district court’s injunction could not be supported by the exclusive dealing alone.

31 Qualcomm, 969 F.3d at 998.

32 The theory resembles a margin squeeze theory, which the Supreme Court rejected in Pac. Bell Tel. Co. v. linkLine Commc’ns, Inc., 555 U.S. 438 (2009).

33 Qualcomm, 969 F.3d at 999.

34 Shapiro & Lemley, supra note 14, at 2059.

35 Id.

36 Qualcomm, 969 F.3d at 1002.

37 See Hovenkamp, supra note 16, at 1686–87, 1692, 1696, 1704, 1713, 1725, 1741, 1743.

38 Id. at 1686–87.

39 Id. at 1704. The “classical tying” label does not fit because Qualcomm was using its economic leverage to avoid litigation over whether device makers needed to take licenses and over the FRAND royalty.

40 Id. at 1713, 1725, citing Aspen Skiing Co. v. Aspen Highlands Skiing Corp., 472 U.S. 585 (1985).