Introduction

The admissibility and probative value of settlement agreements to determine reasonable royalties for patent infringement is a topic that has been addressed by Courts many times in the last several years. Guidance from Courts has been meaningful but incomplete. The central questions that remain for damages experts are how, and to what extent, a settlement agreement can be evaluated for comparability to the hypothetical license and how to account for known differences.

Guidance from the Court

In Georgia-Pacific, the Court identified multiple factors that may be considered when analyzing any license agreement, including settlements, for purposes of determining a reasonable royalty.[1]  These factors include the technical comparability of the patents-in-suit; the products accused of infringement; the periods of time covered by the agreements; the parties to the settlement and the hypothetical license; and the dates of the agreements, among other variables.[2]  Under the Georgia-Pacific guidelines, these factors should be considered when comparing both settlements and non-litigation agreements to a hypothetical license in the context of a reasonable royalty analysis.

With respect to settlement agreements, some courts have enumerated additional variables that practitioners should consider in addition to the Georgia-Pacific factors.  For example, the District of Delaware listed eight “litigation-related factors” that may be considered when analyzing licenses that also settled pending litigation.[3]  These factors are seemingly intended to reconcile the differences in real-world settlements and non-litigation licenses from hypothetical licenses negotiated under the theoretical Georgia-Pacific framework.

Further, the Court in Prism effectively argued that the probative value of settlement agreements is a function of both avoidable and sunk costs and how well the parties understood the probabilities of success if the case had been tried to verdict. According to the Court, settlements executed shortly after litigation ensued may be “less probative” because the “expected costs of litigation and uncertainty over the outcome are highest.”[4] The inverse of this conclusion is that settlements executed after many issues had been resolved by the Court are more probative because more costs had been expended, and there was less uncertainty regarding the outcome if the case was tried to verdict.

The presumption seemingly underlying the Court’s guidance regarding settlements is that because the progression of a lawsuit informs the parties’ forecasts about the ultimate outcome, these agreements are distinct from licenses executed before litigation ensued.

Differences Between Settlements, Non-Litigation Licenses and Hypothetical Licenses

In light of the Court’s guidance, it is important to understand the primary economic differences between settlement agreements and hypothetical licenses under Georgia-Pacific and adjust for any disparities.  Several of these differences are discussed below.

Type of Agreement

As settlements are often described by Courts as entirely distinct from licenses that did not settle an ongoing dispute, it is necessary to distinguish between settlements and licenses executed without pending litigation.

The fact that litigation was pending is an obvious difference between settlements and licenses executed before litigation ensued.   From an economic perspective, there is no cause of action for the patent holder unless the defendant had infringed before the lawsuit was filed.  Filing a patent infringement action signals to the potential licensee that the patent holder has the intent, the wherewithal, and the capital (i.e., cash on hand or litigation funding) to carry out litigation against the defendant.  The perspective for the defendant then shifts from the threat of litigation, which always exists, to actual litigation.

A non-litigation license, on the other hand, may be executed before (ex ante) or after (ex post) alleged infringement.  An ex ante non-litigation license is similar to a hypothetical license because both contemplate an agreement executed before any alleged infringement ensued.  The fact that litigation was not pending when the license was executed does not indicate an ex ante agreement, however, because the patent holder could have alleged infringement and threatened litigation preceding the license without having filed suit.  For these ex post non-litigation licenses, there are few economic differences between a non-litigation license and a settlement.

From the practitioner’s perspective, it is typically not possible to determine whether non-litigation agreements are ex ante or ex post, as there is rarely available information about whether infringement had been alleged by the patent holder.  With this understanding in mind, this discussion shifts to a comparison between settlements and hypothetical licenses because this is the framework most often prescribed by Courts if a settlement is to be used as a basis for quantifying reasonable royalties.

Timing of the Agreements

In the context of the hypothetical license, negotiations take place ex ante and provide the potential licensee with an opportunity to consider their best alternatives to an agreement, also described in patent damages parlance as an available, acceptable non-infringing alternative.

In contrast, settlements take place ex post, after the alleged infringer has made use of the patent-in-suit.  Allegations of past infringement made by plaintiffs against companies that have already made investments in the accused product(s) create a liability for the defendant that does not exist in the Georgia-Pacific framework.  This liability for alleged past infringement eliminates the defendant’s opportunity to consider avoiding infringement in the first place, and the switching costs tend to be higher once a product is launched.

Assumptions of Validity and Infringement

The hypothetical negotiation rubric presumes that the parties agree the patents-in-suit are both valid and infringed.  In other words, if the hypothetical licensor sued the hypothetical licensee for patent infringement, there is a 100% probability that it would win on the merits.  There is no presumed discount for the probability that the patent-in-suit is invalid or not infringed.

In contrast, the presumption commonly applied to both settlements and non-litigation license agreements is that the parties disagree whether the patents-in-suit are valid or have been infringed by the licensee.  Given the nature of patent infringement disputes, the parties may have opposing views regarding both validity and infringement, and neither side is certain of the outcome if the dispute is decided by the Court.  Consequently, the probability of plaintiff’s success on the merits is ostensibly less than 100% and the settlement payment, or patent value, is discounted to reflect this lower probability.

Nature of the Agreements

The hypothetical license is framed as an agreement between a willing licensor and willing licensee freely entering into a license for the patent-in-suit.  This framework does not contemplate an adversarial relationship or pending litigation between the parties, and licenses negotiated in this context are not “tainted by the coercive environment of patent litigation.”[5]  Under the Georgia-Pacific framework, the hypothetical license is more similar to an ex ante agreement executed before alleged infringement and before litigation was even threatened.

In contrast to these hypothetical licenses, settlement agreements are negotiated while patent infringement litigation is pending.  As noted by the Court, the “coercive environment of patent litigation” is contrary to the Georgia-Pacific framework, “which assumes a voluntary agreement will be reached between a willing licensor and a willing licensee”[6]

Avoidable and Sunk Costs

Because the hypothetical negotiation framework does not contemplate pending litigation, any corresponding costs are not considered in the reasonable royalty framework.  There may, of course, be other costs associated with negotiating a license, but these costs are presumably smaller and also not customarily considered by the Courts when determining a reasonable royalty for patent infringement.

This is in contrast to pending patent litigation in which both sides understand the enormous costs of trying a case to verdict regardless of which party prevails.

The Litigation-Related Factors

With this understanding of the primary differences, it is apparent that settlement agreements may largely be compared and adjusted by applying the same analytical rubric set forth in Georgia-Pacific.  We turn next to the litigation-related factors suggested by some Courts for further analyzing settlements as distinguished from licenses executed before litigation ensued.

Relevance of Litigation-Related Factors to Non-Litigation Licenses

As an initial matter, all the variables relevant to an analysis of a settlement may also be considered by parties involved in non-litigation license negotiations before the patent holder files suit.[7]  As non-litigation licenses may be executed either ex ante or ex post, there are few differences that distinguish otherwise comparable settlement agreements from comparable licenses negotiated before litigation ensued.  In a license negotiation conducted before a lawsuit has been filed, the threat of litigation and the licensors intent and wherewithal may still be apparent to both parties.  For this reason, “virtually every patent licensing and cross-licensing agreement can be seen as the settlement of a patent dispute,” regardless of whether a lawsuit has been filed.[8]  The rights and remedies available to patent holders do not change because a lawsuit was filed.[9]

Damages Sought in the Litigation

Most of the litigation-related factors influence the amount of damages a patent holder may seek or be awarded, and a defendant may be required to pay, if the case is tried to verdict.  Specifically, factors (ii) through (iv)[10] ultimately impact factor (i), the amount of damages sought in the litigation.[11]  Factors (vii) and (viii)[12] impact the defendants’ views regarding the probability of the plaintiff achieving the damages at trial (i.e., the expected value).  Although the Court may be correct on this point, factoring how each party viewed and measured the odds of success on the merits is impractical even if the information existed.

With respect to factor (i), if plaintiff’s damages demands are known, a practitioner may consider how much the plaintiff was seeking in the settled litigation.  However, this is a complicated endeavor because expert reports are not typically produced, so there is no way to evaluate the veracity of plaintiff’s damages claim or corresponding methodology.  If plaintiff’s damages claim was inflated relative to the actual value of the patent-in-suit, the corresponding settlement may also overvalue the licensed patent(s).  In addition, damages claims asserted by plaintiffs before rulings on motions to exclude under Daubert may never be presented to the Court if the damages expert relied upon disfavored approaches (i.e., the 25% rule).  Finally, without a trial, and corresponding appeal, it is still unknown how the Court would have ultimately ruled on damages sought by the plaintiff.

In considering possible “enhancements” to damages and other penalties, factors (ii) to (iv) (i.e., the issue of willfulness, the possibility of treble damages, and the imposition of sanctions) would serve to increase the amount of damages available to and the negotiating leverage of the plaintiff, if each was available or pending preceding the settlement.  In comparison to a hypothetical license, all three have the potential to increase damages, so all three would increase the settlement payments above the indicated value of the patent-in-suit absent these enhancements.  For these reasons, the presence of enhancements suggests a settlement overvalues the licensed patents.

Despite the wisdom of evaluating litigation-related factors as they pertain to the damages sought, the information to do so is rarely, if ever, available to damages experts.  In the unlikely event that both sides assigned probabilities to the case that settled, weighing the odds of a legal outcome is both beyond the scope of most damages opinions and likely subject to attorney-client privilege.

Avoidable and Sunk Costs

Studies have found that “the median cost to litigate a single-patent infringement claim through trial was about five million dollars for cases with more than twenty-five million dollars at risk.”[13]  Furthermore, the cost to defendants to defend infringement claims in the U.S. “regardless of the amount at risk, hovers around three million dollars.”[14]  Both litigants typically bear these costs regardless of who prevails at trial.[15]

Although avoidable and sunk costs may be estimated, there is no practical way to normalize the amount of a settlement payment for purposes of determining a reasonable royalty.  The hypothetical negotiation does not contemplate the sunk and avoidable costs of litigation, which suggests a settlement agreement is more likely to overvalue a patent compared to the hypothetical license negotiated without these costs pending.  In addition, a patent holder that chooses litigation may be more motivated to recover sunk costs associated with litigation, adding upward pressure to the amount of damages they may seek in the litigation.  Defendants are exposed to both damages (and enhancements) and the costs of defending infringement claims.[16]  These sunk costs incurred by defendants as the litigation proceeds may also apply countervailing downward pressure on the settlement amount, as defendants have already expended resources to defend against the plaintiff’s allegations and may be less willing to spend more as sunk costs increase over the course of a litigation.  Despite the fact that costs likely factor into the parties’ calculus to settle, there is rarely information available to mathematically quantify how much these costs may have increased, or decreased, the settlement payment in comparison to a hypothetical license valued without consideration of litigation costs.

Practical Limitations of Litigation-Related Factors

There are also practical limitations of addressing these factors in the patent litigation context.  These factors, especially with respect to (ii), (iii), and (vi) through (viii)[17], require legal analysis and corresponding odds-making on the merits that are generally outside of a damages expert’s subject matter.  Many of the factors also relate to information that is customarily filed under seal or subject to attorney-client privilege.  For these reasons, the relevant information is not typically disclosed to the public or produced in litigation.

Greater Insight Into Settlement Agreements

In addition to these differences between real-world settlements and hypothetical licenses under Georgia-Pacific, there may be other available information that may be readily considered with respect to settlement agreements.  In particular, the amount of information about settlement agreements is oftentimes larger than non-litigation licenses because of the public nature of patent litigation.  Plaintiffs may describe the allegedly infringing product(s), when the defendant was first alleged to infringe, and which of the licensed patents had been asserted.  For example, the combination of the start of alleged infringement and the expiration date(s) of the licensed patent(s) may be used to establish a comparable timeline and to calculate the effective duration of the settlement agreement.  The same information is often not available, or as well defined, for licenses executed before litigation ensued.  For this reason, the circumstances surrounding a comparable settlement may be better defined than a similarly comparable non-litigation license.

Key Takeaways

The guidance from Courts regarding how to compare and adjust settlement agreements for use in reasonable royalty analyses is largely focused on litigation variables with the potential to change the expected value of a case if tried to verdict.  The CAFC has suggested that the probative value of settlement agreements is a function of both avoidable and sunk costs and how well the parties understood the probabilities of success if the settled case had been tried to verdict.  Some of the guidance suggests that settlement agreements are markedly different from non-litigation licenses and argues that the analysis should focus on the litigation-specific variables like the amount of damages sought by the plaintiff.

However, settlements are not markedly different from ex post non-litigation licenses.  Both involve negotiations in which the licensor/plaintiff alleges that the licensee/defendant infringed the patent(s) before the license was consummated.  The analytical framework and corresponding adjustments that are appropriate for settlements are largely the same as those for non-litigation licenses.

The “litigation-related factors” describe additional variables the Court suggests damages experts should consider when analyzing a settlement agreement for purposes of quantifying a reasonable royalty.  These factors describe some of the variables that could impact the amount paid to settle pending litigation and are ostensibly intended to reconcile real world settlement agreements and non-litigation licenses and hypothetical licenses negotiated under the theoretical Georgia-Pacific rubric.  Despite the potential impact of litigation-related factors on settlements, sufficient information is rarely available to quantify that impact, and many of the factors require expert legal opinion that is outside the scope of damages experts.

Despite the practical and discovery limitations of settlements, settlements may offer more insight into the calculus of the parties than non-litigation licenses.  Because some information about patent infringement lawsuits is published by courts, there is often more available information to consider about a settlement agreement than a non-litigation license.  This additional information may be used to analyze a settlement in the context of both Georgia-Pacific and other litigation-related factors suggested by other courts.

If you have any questions or would like to discuss how we can help, reach out to Christopher Anderson.

To receive Insights, sign up for our newsletter.

Disclaimer: The views expressed in this article are those of the author and do not necessarily reflect the views of StoneTurn Group, LLP, Province, LLC, or their affiliates. This article is provided for informational purposes only and does not constitute legal, financial,  or other professional advice.

[1] Ga.-Pacific Corp. v. United States Plywood Corp., 318 F. Supp. 1116, 1120-21 (S.D. N.Y. 1970), modified and aff’d, 446 F.2d 295 (2d Cir.), cert. denied, 404 U.S. 870 (1971) (“Georgia-Pacific”).
[2] Id.
[3] Memorandum Order (ECF No. 697, at 22-23), EIS, Inc. v. Intihealth Ger GmbH et al., No. 1:19-CV-1227-GBW (D. Del.); Prism Techs. LLC, v. Sprint Spectrum L.P., 849 F.3d 1360, 1368-70 (Fed. Cir. 2017).  The litigation-related factors include: (i) “the amount of damages ultimately sought in the litigation”; (ii) “whether ‘the issue of willfulness’ would ‘have been tried’”; (iii) “whether there was ‘the possibility of treble damages’”; (iv) “whether sanctions had been imposed”; (v) “the nature of the ‘underlying negotiations’”; (vi) “‘the context in which the settlement agreement was reached’”; (vii) “‘the defenses available to [the defendant] at trial’”, and; (viii) “‘the strength of those defenses.’”
[4] Order Regarding Motions to Exclude Damages Experts’ Opinions, Plexxikon Inc. v. Novartis Pharmaceuticals Corporation, No. 4:17-cv-04405-HSG (N. Cal.) (reported at 2021 U.S. Dist. LEXIS 5596, at 10); Prism Techs. LLC, v. Sprint Spectrum L.P., 849 F.3d 1360, 1368-70 (Fed. Cir. 2017).
[5] LaserDynamics, Inc. v. Quanta Computer, Inc., 694 F.3d 51, 77 (Fed. Cir. 2012).
[6] LaserDynamics, Inc. v. Quanta Computer, Inc., 694 F.3d 51, 77 (Fed. Cir. 2012).
[7] Lemley, Mark A., and Carl Shapiro. “Probabilistic Patents.” Journal of Economic Perspectives, vol. 19, no. 2, spring 2005, pp. 76; Noting that “virtually every patent licensing and cross-licensing agreement can be seen as the settlement of a patent dispute.” Also see, Prism Techs. LLC, v. Sprint Spectrum L.P., 849 F.3d 1360, 1368-70 (Fed. Cir. 2017) (Noting the “inherent connection between patent licenses and at least the potential for litigation.” However, the Court does not address any of these litigation-related factors as they pertain to licenses executed before a suit was filed).
[8] Prism Techs. LLC, v. Sprint Spectrum L.P., 849 F.3d 1360, 1368-70 (Fed. Cir. 2017) (Noting “the potential for litigation therefore must loom over patent licenses generally, including those signed without any suit ever being filed.”). Lemley, Mark A., and Carl Shapiro. “Probabilistic Patents.” Journal of Economic Perspectives, vol. 19, no. 2, spring 2005, pp. 76.
[9] Prism Techs. LLC, v. Sprint Spectrum L.P., 849 F.3d 1360, 1368-70 (Fed. Cir. 2017) (Noting that “a (non-exclusive) license to practice a patent is in substance nothing but a covenant not to sue: what such a license is, at its core, is an elimination of the potential for litigation.”).
[10] For reference, (ii) “whether ‘the issue of willfulness’ would ‘have been tried’”; (iii) “whether there was ‘the possibility of treble damages’”; and (iv) “whether sanctions had been imposed.”
[11] Prism Techs. LLC, v. Sprint Spectrum L.P., 849 F.3d 1360, 1369-70 (Fed. Cir. 2017) (Noting that the risk of enhanced damages “would push settlement value above the value of the technology.”).
[12] For reference, (vii) “‘the defenses available to [the defendant] at trial’” and (viii) “‘the strength of those defenses.’”
[13] Day, Gregory, and Steven Udick. “Patent Law and the Emigration of Innovation.” Washington Law Review, vol. 94, no. 1, 1 Mar. 2019, pp. 141.
[14] Day, Gregory, and Steven Udick. “Patent Law and the Emigration of Innovation.” Washington Law Review, vol. 94, no. 1, 1 Mar. 2019, pp. 141.
[15] Day, Gregory, and Steven Udick. “Patent Law and the Emigration of Innovation.” Washington Law Review, vol. 94, no. 1, 1 Mar. 2019, pp. 141.
[16] “License fees negotiated in the face of a threat of high litigation costs ‘may be strongly influenced by a desire to avoid full litigation.’” See Panduit Corp. v. Stahlin Bros. Fiber Works, 575 F.2d 1152, 1164 (6th Cir. 1978) (citing Rude v. Westcott, 130 U.S. 152, 164, 9 S.Ct. 463, 32 L.Ed. 888 (1888)).
[17] For reference, (ii) “whether ‘the issue of willfulness’ would ‘have been tried’”; (iii) “whether there was ‘the possibility of treble damages’”; (vi) “‘the context in which the settlement agreement was reached’”; (vii) “‘the defenses available to [the defendant] at trial’”, and; (viii) “‘the strength of those defenses.’”