No. 23-55299

Appeal from the United States District Court for the Central District of California John F. Walter, District Judge, Presiding

Argued and Submitted En Banc June 24, 2025 Seattle, Washington

Filed February 20, 2026

Before: Mary H. Murguia, Chief Judge, and Susan P. Graber, Ronald M. Gould, Consuelo M. Callahan, Milan D. Smith, Jr., Bridget S. Bade, Daniel A. Bress, Jennifer Sung, Salvador Mendoza, Jr., Anthony D. Johnstone and Ana de Alba, Circuit Judges.

Page 2Opinion by Judge Graber; Dissent by Judge Callahan

SUMMARY *

Reversing the district court's order denying plaintiffs' motion under Federal Rule of Civil Procedure 60(b) for relief from summary judgment in favor of defendant Rubicon Resources, LLC, and remanding for further proceedings in a civil action under 18 U.S.C. § 1595(a), the en banc court held that an amendment to the statute, clarifying that defendants are civilly liable when they attempt to benefit, but do not succeed in benefitting, from human trafficking, has retroactive effect.

Plaintiffs were villagers from rural Cambodia who allegedly were forced to work at seafood factories in Thailand. They alleged that Rubicon marketed in the United States seafood products from those factories, thereby participating in a venture that benefited from human trafficking. The district court entered summary judgment in favor of Rubicon, holding that Rubicon only attempted to benefit from plaintiffs’ forced labor. In “Ratha I,” a threejudge panel of this court agreed, holding that the civil remedy provision of § 1595(a) did not encompass “attempt” liability. Congress subsequently passed the Abolish Trafficking Reauthorization Act of 2022 (“ATRA”), a bill Page 3clarifying that defendants are civilly liable when they attempt to benefit, but do not succeed in benefitting, from human trafficking. Plaintiffs moved for relief under Rule 60(b)(6). The district court denied the motion, holding that Congress did not intend for attempt liability to apply retroactively.

Applying the three-step framework of Landgraf v. USI Film Products , 511 U.S. 244 (1994), the en banc court held that Congress intended its clarifying amendment to have retroactive effect. Under Landgraf , the court first reads the statutory text to see whether Congress explicitly included a retroactive effective date. If not, at step two, the court determines whether application of the statute would have retroactive effect. If so, the court applies a presumption against retroactivity. At step three, the court decides whether the presumption can be overcome by considering all indicia of intent to determine whether Congress made its intention clear that the statute applies retroactively. If so, the statute applies retroactively. To the extent that this court has held in other cases that clarifying amendments to civil statutes fall outside the Landgraf framework, the en banc court overruled those cases as inconsistent with Landgraf .

At step one, the en banc court observed that the ATRA contained no explicit retroactive effective date. At step two, the en banc court determined that the ATRA would have retroactive effect and that the presumption against retroactivity applied. At step three, the panel concluded that Congress nevertheless clearly meant for the ATRA to apply retroactively. The panel concluded that four factors, considered together, clearly established the amendment's retroactivity: (a) Congress expressly stated that its enactment was a technical and “clarifying” update; (b) § 1595(a) was ambiguous before the amendment;

The en banc court held that § 1595’s retroactivity fatally undermined one of the three grounds on which the district court had premised its grant of summary judgment. The en banc court concluded that the two remaining grounds were not sufficient to justify denial of plaintiffs' Rule 60(b) motion. These grounds were: (1) plaintiffs did not demonstrate that Rubicon knowingly participated in a human trafficking venture; and (2) plaintiffs did not demonstrate that Rubicon knew, or should have known, that human trafficking was occurring at the factory. The en banc court held that the district court erred as a matter of law on the participation element because one can participate in a venture without operating or managing it, and the district court did not view all facts and inferences in favor of plaintiffs. The district court also erred as a matter of law on the knowledge element, and a reasonable jury could have found that Rubicon knew or should have known about the alleged human trafficking. Weighing the relevant factors for relief under Rule 60(b)(6), the en banc court held that plaintiffs were entitled to relief from the district court’s final order.

Judge Callahan dissented, joined by Judge M. Smith except as to Part I.A., by Judge Bade in full, by Judge Bress as to Part I only, and by Judge Johnstone in full. In Part I, Judge Callahan wrote that she agreed with the majority that Landgraf governs the retroactivity analysis, that prior cases that created an exception for clarifying amendments must be overruled, and that ATRA would have retroactive effect and thus the presumption against retroactivity applied. Judge Page 5Callahan’s agreement ended there. In Part I.A., she wrote that the four factors cited by the majority did not demonstrate clear congressional intent that the ATRA is retroactive. In Part I.B., Judge Callahan wrote that under a faithful application of Landgraf , the en banc court should decline to give the ATRA retroactive effect because there is no clear indication that Congress intended it to be retroactive. In Part II, Judge Callahan wrote that the majority reached beyond the bounds of this appeal by addressing all of the district court’s reasons for granting summary judgment.

Agnieszka M. Fryszman (argued), Nicholas J. Jacques, Emily Ray, and Madeleine Gates, Cohen Milstein Sellers & Toll PLLC, Washington, D.C.; Paul L. Hoffman and John C. Washington, Schonbrun Seplow Harris Hoffman & Zeldes LLP, Hermosa Beach, California; Catherine Sweetser, UCLA Law Clinics, Los Angeles, California; Dan Stormer, Hadsell Stormer & Renick LLP, Pasadena, California; for Plaintiffs-Appellants.

Barbara E. Taylor (argued), Bryan D. Daly, Melissa K. Eaves, and Charles L. Kreindler, Sheppard Mullin Richter & Hampton LLP, Los Angeles, California; for DefendantAppellee.

Aaron M. Halegua, Aaron Halegua PLLC, New York, New York; Tyler R. Giannini, Human Rights Entrepreneurs Clinic, Harvard Law School, Cambridge, Massachusetts; Margaret Lee, Human Trafficking Legal Center, Washington D.C.; for Amici Curiae Members of Congress, Representative Nadler, et al..

Page 6John Burton, John Burton Law, Pasadena, California; Margaret Lee, Human Trafficking Legal Center, Washington D.C.; for Amicus Curiae Human Trafficking Legal Center.

John Burton, John Burton Law, Pasadena, California; Allison Gill and Johanna Lee, Global Labor JusticeInternational Labor Rights Forum, Washington, D.C.; Avery Kelly and Alicia Brudney, Corporate Accountability Lab, Chicago, Illinois; for Amici Curiae Human and Workers' Rights Organizations and United States Shrimp Producers.

Anne M. Voigts, King & Spalding LLP, Palo Alto, California; Julia Romano, King & Spalding LLP, Los Angeles, California; Zoe M. Beiner, King & Spalding LLP, Washington, D.C.; for Amicus Curiae Professor David Abramowitz.

Joshua M. Wesneski and Sydney Hargrove, Weil Gotshal & Manges LLP, Washington, D.C.; Cory L. Andrews, Washington Legal Foundation, Washington, D.C.; for Amicus Curiae Washington Legal Foundation.

Andrew J. Clopton, Jones Day, Detroit, Michigan; Julia E. Fine, Jones Day, Washington, D.C.; for Amicus Curiae International Franchise Association.

Page 7GRABER, Circuit Judge:

Plaintiffs are villagers from rural Cambodia who allegedly were forced to work at seafood factories in Thailand. They brought a civil action under 18 U.S.C. § 1595(a) against four defendants, including Defendant Rubicon Resources, LLC (“Rubicon”). Plaintiffs allege that Rubicon marketed in the United States seafood products from those factories, thereby participating in a venture that benefited from human trafficking. The district court entered summary judgment in favor of Rubicon, holding that Rubicon only attempted to benefit from Plaintiffs' forced labor. Critically, in the court's view, the civil remedy provision of § 1595(a) did not encompass “attempt” liability. A three-judge panel of this court agreed, Ratha v. Phatthana Seafood Co. (Ratha I), 35 F.4th 1159 (9th Cir. 2022), creating a rift with two other circuits’ interpretation of the statute. Congress disagreed with our interpretation of the statute. Just a few months after this court issued Ratha I, and mere days after the Supreme Court denied certiorari, Congress passed a bill clarifying that defendants are civilly liable when they attempt to benefit, but do not succeed in benefitting, from human trafficking.

Plaintiffs quickly moved, under Federal Rule of Civil Procedure 60(b), for relief from the summary judgment entered in Rubicon's favor. The district court denied the motion, holding that Congress did not intend for attempt liability to apply retroactively. We hold that Congress intended its clarifying amendment to have retroactive effect and that the district court legally erred in its alternativePage 8 grounds for denying Rule 60(b) relief. Accordingly, we reverse and remand for further proceedings.

Because the district court incorporated into its denial of the Rule 60(b) motion its earlier decision on a summary judgment, we set out the facts in the light most favorable to Plaintiffs. See Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 255 (1986) (“The evidence of the non-movant is to be believed, and all justifiable inferences are to be drawn in [their] favor.”); see also Ratha I, 35 F. 4th at 1180 (applying this standard).

Phatthana Seafood Co., Ltd. ('Phatthana'), a Thai corporation, owned a seafood processing factory in southern Thailand. Phatthana’s owner founded Rubicon—a United States company-so that Phatthana and other seafood producers could sell their goods more easily in the United States. Rubicon neither owned factories nor hired Phatthana’s employees, but Rubicon’s employees “coordinated sales and marketing, visited and conducted preaudits of Phatthana’s factories, and arranged for import and shipping of Phatthana’s product.” Ratha I, 35 F.4th at 1166. And in its own marketing materials, Rubicon claimed 'to control every aspect of production.”

Plaintiffs Keo Ratha, Sem Kosal, Sophea Bun, Yem Ban, Nol Nakry, Phan Sophea, and Sok Sang are villagers from rural Cambodia who were recruited from their villages to work at seafood factories, including Phatthana's. Id. at 1165–66. Though Plaintiffs were promised competitive pay and free accommodations, they instead “became victims of peonage, forced labor, and involuntary servitude.' Id. at 1165. Specifically, from 2010 through late 2012, “Plaintiffs were paid less than promised, charged for accommodations, Page 9charged for other unexpected expenses, . . . and subjected to harsh conditions.' Id. Workers who complained faced threats and retaliation; some were forced to crawl on their hands and knees, without clothing, until bloody. Plaintiffs’ passports were confiscated when they arrived at the factories. Plaintiffs could not leave because police officers would arrest them if they left the factories without their passports, 'which [Plaintiffs] were told would not be returned until ‘recruitment fee[s]’ and other amounts were paid.” Id. Some workers tried to leave anyway and were arrested.

The conditions at Phatthana’s factory did not go unnoticed. In October 2011, Rubicon tried to sell fourteen containers of shrimp to Walmart, which rejected the shipment because it had concerns about the factory’s working conditions. Id. at 1166. In January 2012, Plaintiff Ratha told the Phnom Penh Post that he was subject to forced labor and that hundreds of Cambodians at Phatthana's factory 'wanted to leave but were unable to get their passports back.' It is undisputed that in February 2012, Rubicon was aware of the Phnom Penh Post's article reporting Ratha's allegations of human trafficking and describing the factory’s deplorable working conditions. See id. at 1177 ('Rubicon was undisputedly aware of Ratha's whistleblower report.”). More articles came out about the workers' allegations of forced labor at the factories, and Rubicon’s managers circulated them to one another. Rubicon briefly paused its attempts to sell shrimp to Walmart but, beginning in July 2012, it resumed those efforts. Ultimately, Rubicon failed to sell Phatthana’s shrimp in the United States during the time period relevant to this case. Id. at 1166.

Page 10In June 2016, Plaintiffs sued Rubicon under the civil remedy provision of the William Wilberforce Trafficking Victims Protection Reauthorization Act of 2008 (“TVPRA”), 18 U.S.C. § 1595(a) (effective December 23, 2008, through January 4, 2023).¹ Plaintiffs alleged that they were victims of peonage, forced labor, involuntary servitude, and human trafficking in violation of 18 U.S.C. §§ 1581, 1584, 1589, 1590, 1592, and 1593A. Plaintiffs alleged that Rubicon 'knowingly benefited from participation in a venture which [Rubicon] knew or should have known was engaged in peonage, forced labor, involuntary servitude, unlawful conduct with respect to documents, and human trafficking.” They further alleged that Rubicon “provided a market and worked to expand that market,” when it knew that Phatthana was engaged in human trafficking and “intend[ed] to benefit from it.”

Rubicon moved for summary judgment, which the district court entered in December 2017. The district court ruled that “(1) Plaintiffs failed to demonstrate that Rubicon knowingly participated in a human trafficking venture; (2) Plaintiffs failed to demonstrate that Rubicon knew or should have known that human trafficking existed at Phatthana’s Songkhla factory; and (3) there was no evidence that Rubicon benefitted from Phatthana's alleged human trafficking.” Ratha v. Phatthana Seafood Co., Ltd., No. CV 16-4271-JFW (ASx), 2023 WL 2762044, at *1 (C.D. Cal. Mar. 3, 2023).

Page 11Plaintiffs appealed, and a panel of our court affirmed the district court’s judgment. Ratha I, 35 F.4th at 1181. We held that (1) Plaintiffs did not show that Rubicon received a benefit and (2) § 1595(a) did not extend civil liability to attempts to benefit. Id. at 1175–76. We filed that decision on February 25, 2022, and amended that decision and denied rehearing en banc on May 31, 2022; the Supreme Court denied Plaintiffs’ petition for a writ of certiorari on December 5, 2022. 143 S. Ct. 491 (2022).

Fifteen days later, on December 20, 2022, Senators Martin Heinrich, John Cornyn, and Amy Klobuchar proposed an amendment to a pending bill entitled the Abolish Trafficking Reauthorization Act of 2022 (“ATRA”). The proposed amendment added wording to 18 U.S.C. § 1595(a), removing any doubt that victims may bring private actions against parties for attempting to benefit from human trafficking. The Senate passed the amended bill the same day, and the President signed it into law on January 5, 2023. Pub. L. No. 117-347, 136 Stat. 6199 (2023).

Plaintiffs then filed a motion under Federal Rule of Civil Procedure 60(b)(6), asking the district court to vacate its summary judgment order as to Rubicon. Plaintiffs argued that the ATRA clarified that § 1595(a) included attempt liability and, because the ATRA served only as a clarification, it applied retroactively. The district court disagreed and denied Plaintiffs’ motion, holding that (1) the ATRA was not retroactive; and (2) even if it were, the district court premised its summary judgment as to Rubicon on two additional grounds, which were independent of the changesPage 12 made by the ATRA and which the court incorporated by reference.²

Plaintiffs timely appeal. A majority of a three-judge panel affirmed the district court’s decision. Ratha v. Rubicon Res., LLC, 111 F.4th 946, 969 (9th Cir. 2024). Judge Graber dissented, stating that she would have held that the ATRA was retroactive and would have reversed and remanded for further proceedings. Id. at 975–76 (Graber, J., dissenting). A majority of active judges voted to rehear the case en banc, Ratha v. Rubicon Res., LLC, 129 F.4th 1212 (9th Cir. 2025) (order), and the en banc court heard oral argument on June 24, 2025.

We review for abuse of discretion the denial of a Rule 60(b) motion. Phelps v. Alameida, 569 F.3d 1120, 1131 (9th Cir. 2009). A district court abuses its discretion if its denial rests on “an erroneous view of the law.” Id. (citation and internal quotation marks omitted). We review de novo questions of law. Bynoe v. Baca, 966 F.3d 972, 979 (9th Cir. 2020).

Landgraf v. USI Film Products, 511 U.S. 244 (1994), controls the analysis of retroactivity for federal civil Page 13legislation. Landgraf provides a three-step framework for deciding whether an enactment, such as an amendment to a civil statute, applies retroactively. We first read the statutory text to see whether Congress explicitly included a retroactive effective date. Id. at 280. If so, we honor that effective date. Id. If not, we determine, at step two, whether application of the statute “would have retroactive effect.” Id. If the statute would not have retroactive effect—if, for example, it affects only 'the propriety of prospective relief,' or it changes a procedural rule—then it ordinarily applies retroactively. Id. at 273–75. But “[i]f the statute would operate retroactively,” courts apply a presumption against statutory retroactivity. Id. at 280. That presumption is driven by the concern that it may be unfair to “impose[] new burdens on persons after the fact.' Id. at 270. But presumptions can be overcome. Accordingly, the final step of the analysis requires us to consider all indicia of intent to determine whether Congress has made its intention clear that the statute applies retroactively. Id. at 280; see also Rivers v. Roadway Express, Inc., 511 U.S. 298, 304–09 (1994) (elaborating, in Landgraf’s companion case, on the analysis of congressional intent). If so, the statute applies retroactively.

Our own cases generally have comported with Landgraf’s analysis. See, e.g., Bahr v. Regan, 6 F.4th 1059 (9th Cir. 2021). But we pause to clarify two aspects of the retroactivity analysis.

First, we have at times held that “clarifying” amendments are analyzed differently. We have held (a) that such amendments present an exception to Landgraf’s requirements, see, e.g., Beaver v. Tarsadia Hotels, 816 F.3d 1170, 1186 (9th Cir. 2016) (“However, no Landgraf analysis is required if an amendment merely serves to clarify rather than change the substance of existing law.'); and (b) that Page 14there is no presumption against retroactivity for “clarifying” amendments, see, e.g., ABKCO Music, Inc. v. LaVere, 217 F.3d 684, 689 (9th Cir. 2000) (“[C]larifying legislation is not subject to any presumption against retroactivity . . . .”). To the extent that we have held in other cases that clarifying amendments to civil statutes fall outside the Landgraf framework, we now overrule those cases as inconsistent with Landgraf. Nothing in Landgraf, or in any other Supreme Court precedent, suggests that Landgraf’s general framework applies only in some situations. We must analyze all questions concerning the retroactive effect of amendments to civil statutes under Landgraf's three-step framework.

But we overrule those cases only insofar as they bypass a consideration of congressional intent or establish a contrary presumption regarding retroactivity. To the extent that those cases consider various indicators of congressional intent, they remain valid.³

Second, as described above, Landgraf clearly contemplates a sequence of three questions. 511 U.S. at 280; see also Fernandez-Vargas v. Gonzales, 548 U.S. 30, 37–38, 40 (2006) (describing the inquiry as containing three steps). Our cases generally adhere to that understanding. See, e.g., TwoRivers v. Lewis, 174 F.3d 987, 993 (9th Cir. 1999) (describing a “three stage analysis”). But occasionally we have described the inquiry as containing only two steps, with Page 15the second step encompassing both of the final two stages of the Landgraf analysis. See, e.g., Ditullio v. Boehm, 662 F.3d 1091, 1099 (9th Cir. 2011) (describing a “two-step analysis”); Garcia-Ramirez v. Gonzales, 423 F.3d 935, 939 (9th Cir. 2005) (per curiam) (describing a “two-step approach'). For consistency with the Supreme Court's decisions, we adhere to the Court's description of three distinct inquiries.

We turn, then, to applying Landgraf’s framework to the ATRA.

Utilizing that framework here, we conclude that the ATRA’s changes to 18 U.S.C. § 1595(a) apply retroactively. The legislation contained no explicit retroactive effective date, so we cannot stop at step one of the Landgraf framework.

At step two, we determine 'whether the new statute would have retroactive effect, i.e. , whether it would impair rights a party possessed when he acted, increase a party's liability for past conduct, or impose new duties with respect to transactions already completed.” Landgraf, 511 U.S. at 280. The TVPRA, as interpreted by Ratha I, did not provide civil liability for an attempt to benefit from forced labor, and the ATRA unambiguously does so. The ATRA thereby increases Rubicon's liability for its past conduct, so the Page 16ATRA would have retroactive effect.⁴ Accordingly, the presumption against retroactivity applies.

We therefore ask, at the final step, whether Congress nevertheless clearly meant for the ATRA to apply retroactively. The answer here is “yes.” Four factors in this case, considered together, clearly establish the amendment’s retroactivity: (a) Congress expressly stated that its enactment was a technical and “clarifying” update; (b) § 1595(a) was ambiguous; (c) Ratha I created a circuit split concerning the interpretation of § 1595(a); and (d) Congress acted swiftly, and with immediate effect, following our decision in Ratha I. We address those factors in turn.⁵

The full text of the relevant amendment states:

SEC. 102. TECHNICAL AND CLARIFYING UPDATE TO CIVIL REMEDY.

Section 1595(a) of title 18, United States Code, is amended by inserting “or attempts or conspires to benefit,” after “whoever knowingly benefits.”

Pub. L. No. 117-347, 136 Stat. 6199, 6200 (2023).

The section’s title—“TECHNICAL AND CLARIFYING UPDATE TO CIVIL REMEDY”—strongly suggests that Congress intended for it to apply retroactively. Although “clarifying” amendments are not exempted from the Landgraf analysis, an amendment’s “clarifying” nature is evidence that Congress intended the amendment to be retroactive. See Rivers, 511 U.S. at 311 (stating that an enactment's 'restorative purpose' can be evidence that Congress intended the enactment to apply retroactively). 6Page 18 Here, section 102’s title suggests that, as to civil liability for attempts to knowingly benefit from forced labor under § 1589, Congress intended for the ATRA to “confirm what the law . . . always meant.” Beverly Cmty. Hosp. Ass’n v. Belshe, 132 F.3d 1259, 1265 (9th Cir. 1997); see also Almendarez-Torres v. United States, 523 U.S. 224, 234 (1998) (“[T]he title of a statute and the heading of a section are tools available for the resolution of a doubt about the meaning of a statute.” (citations and internal quotation marks omitted)).

We must give great weight to the clarifying nature of Congress’s amendment. The Supreme Court has stated that rule in cases decided both before and after Landgraf. For example, in Red Lion Broadcasting Co. v. FCC, 395 U.S. 367 (1969), the Court stated: 'Subsequent legislation declaring the intent of an earlier statute is entitled to great weight in statutory construction.” Id. at 380–81; see also id. at 381 n.8 (citing earlier cases). The Court repeated the same principle two years after deciding Landgraf. Loving v. United States, 517 U.S. 748, 770 (1996). In Belshe, we cited those cases and observed: “It has been established law since nearly the beginning of the republic . . . that congressional legislation that thus expresses the intent of an earlier statute Page 19must be accorded great weight.” 132 F.3d at 1265 (citations omitted); see also id. at 1266 ('We therefore honor Congress’[s] ‘clarification’ label . . . .”). Other circuits, too, give weight to Congress's clarification label. See, e.g., Brown v. Thompson, 374 F.3d 253, 259 (4th Cir. 2004) (giving effect to Congress's label of an amendment as “technical and clarifying”); Piamba Cortes v. Am. Airlines, Inc., 177 F.3d 1272, 1283–84 (11th Cir. 1999) (giving effect to a legislative declaration of an intent to clarify in the face of a prior ambiguity).

Relatedly, we must assume that Congress knew that we long had considered titles with a “clarifying” label to be a strong indicator of retroactive intent. See United States v. LeCoe, 936 F.2d 398, 403 (9th Cir. 1991) (“Congress is, of course, presumed to know existing law pertinent to any new legislation it enacts.”). By using the term “clarifying” in the title of its amendment, Congress spoke to us in our own language, directing us to apply the amendment retroactively.

It also is worth noting that Congress consciously chose to label the 2023 amendment 'clarifying' when it had labeled the previous changes 'enhancements.' Compare Pub. L. No. 117-347, 136 Stat. 6199, 6200 (2023), with Pub. L. 108-193, 117 Stat. 2875, 2877 (2003), and Pub. L. 110457, 122 Stat. 5044, 5067 (2008). See Rivers, 511 U.S. at 307–08 (considering, as evidence of Congress’s retroactive intent, the difference between Congress’s stated purpose of 'expanding' a law's scope and a prior bill's purpose of “restoring” such scope). “Enhancement” connotes a substantive improvement. See Enhance, Merriam–Webster, https://www.merriam-webster.com/dictionary/enhancement (last visited Nov. 4, 2025) ('[T]o increase or improve in value, quality, desirability, or attractiveness[.]”). By contrast, the 'clarifying' label has a settled meaning that Page 20does not suggest a substantive change. See Clarifying, Merriam–Webster, https://www.merriamwebster.com/dictionary/clarifying (last visited Nov. 4, 2025) (“[M]aking something less confusing or easier to understand[.]”).

The ATRA was not simply labeled 'clarifying'; it actually clarified an ambiguous statute. Before the ATRA, the TVPRA was subject to more than one reasonable interpretation. Ratha I memorializes one such interpretation. 35 F.4th at 1176 (holding that the TVPRA does not create civil attempt liability); see also Dissent at 44-46 (setting forth a similar interpretation of the TVPRA). But we find a different reading—one in which the TVPRA did authorize such liability—to be at least as reasonable. Letting Ratha I speak for itself, we now explain a contrary reading of the TVPRA.

We begin with the text of the statute, which we must read in context and with an understanding of the law’s purpose. Asarco LLC v. Atl. Richfield Co., 866 F.3d 1108, 1118–21 (9th Cir. 2017).

The Victims of Trafficking and Violence Protection Act of 2000 ('VTVPA'), Pub. L. 106-386, 114 Stat. 1464 (2000), made it illegal for a person to knowingly provide or obtain forced labor. Id. at § 112(a)(2), 114 Stat. 1464, 1486– 87. The VTVPA also included § 1594 (General provisions), id. at § 112(a)(2), 114 Stat. 1464, 1489, which functions like a 'definitions' section. For example, as relevant here, § 1594(a) provided that '[w]hoever attempts to violate section . . . 1589 shall be punishable in the same manner as a completed violation of that section.” Id. (emphases added). The VTVPA thus defined an attempt to violate § 1589 as a Page 21violation of § 1589—without extending attempt liability to every crime listed in the chapter.

Three years later, Congress passed the Trafficking Victims Protection Reauthorization Act of 2003. Pub. L. 108-193, 117 Stat. 2875 (2003). That act introduced § 1595—“Civil remedy”—allowing “[a]n individual who is a victim of a violation of section 1589, 1590, or 1591 . . . [to] bring a civil action against the perpetrator.” Id. at § 4(a)(4)(A), 117 Stat. 2875, 2878.

Finally, in 2008, Congress passed the TVPRA. Pub. L. 110-457, 122 Stat. 5044 (2008). The TVPRA expanded civil liability to perpetrators of any violation under the chapter, not just perpetrators of the crimes listed in §§ 1589, 1590, and 1591:

An individual who is a victim of a violation of this chapter may bring a civil action against the perpetrator (or whoever knowingly benefits, financially or by receiving anything of value from participation in a venture which that person knew or should have known has engaged in an act in violation of this chapter) in an appropriate district court of the United States and may recover damages and reasonable attorneys fees.

18 U.S.C. § 1595 (2008) (emphases added); see Pub. L. 110457, § 221, 122 Stat. 5044, 5067.

Considering §§ 1589, 1594, and 1595 together and in context, three things become clear.

First, an attempt to violate § 1589 was a punishable violation of § 1589. As the Supreme Court explained in Carcieri v. Salazar, 555 U.S. 379, 393–95 (2009), a defined term retains its meaning throughout a particular statute. See Cochise Consultancy, Inc. v. United States ex rel. Hunt, 587 U.S. 262, 268 (2019) ('We . . . avoid interpretations that would 'attribute different meanings to the same phrase.'' (citation omitted)); Ratzlaf v. United States, 510 U.S. 135, 143 (1994) ('A term appearing in several places in a statutory text is generally read the same way each time it appears.”). Through § 1594 (General provisions), Congress added 'attempt' into the definition of the conduct that violated § 1589 (Forced labor). Pub. L. 110-457, § 222(b)(3), 122 Stat. 5044, 5068 (2008). In other words, reading the two sections together, an attempt to knowingly benefit from forced labor is and was a violation of the TVPRA.

Second, Congress made clear in the 2008 version of § 1595(a) that any violation of the human-trafficking chapter would give rise to civil liability. The 2003 version of § 1595 (Civil remedy) had limited civil liability to only three of several substantive crimes. Pub. L. 108-193, § 4(a)(4)(A), 117 Stat. 2875, 2878. As of 2008, though, the civil remedy applied (if the other requirements were met) to every “act in violation of [the human-trafficking] chapter.” Pub. L. 110457, § 221, 122 Stat. 5044, 5067. That is, Congress intended every criminal act to have a civil-remedy counterpart. The intended parallelism is further demonstrated by § 1595(b), Page 23which provides that courts must stay any civil action during the pendency of a criminal action involving the same conduct and the same victim. Pub. L. 108-193, § 4(a)(4)(A), 117 Stat. 2875, 2878. Given the purpose of the statute, we can discern no reason why any criminal violation, including an attempt, would fail to give rise to corresponding civil liability.

Third, Congress made a simple drafting choice, apparently for the sake of clarity and efficiency. Congress chose to define, in § 1594 (General provisions), certain violations to include attempts. Pub. L. 106-386, § 112(a)(2), 114 Stat. 1464, 1488 (2000). Section 1594 (General provisions) made “attempt” a part of the substantive crime only as to some human-trafficking violations, but not as to all of them. Id. Had Congress included 'attempt' in § 1595(a) (Civil remedy), that wording would have broadened the parallel civil liability beyond what Congress intended in 2008. Thus, the drafting choice improved clarity.

Of course, Congress could have chosen a less efficient course: to revise each substantive section to add attempt, conspiracy, neither, or both. If Congress had selected that path, there could be no doubt that an attempt to benefit from forced labor was a violation of § 1589. But instead, Congress’s chosen path gave rise to ambiguity.

Both the above interpretation and the reading adopted in Ratha I are plausible. Thus, Congress was clarifying an ambiguity when it passed the ATRA.

The overall circumstances here lend even more weight to Congress’s “clarifying” label. The overarching concern Page 24in Landgraf is the unfairness of imposing new burdens on private actors, punishing behavior that once was legal. Landgraf, 511 U.S. at 280. That concern is greatly minimized in the precise circumstances here. By passing the ATRA, Congress acted to clarify what the law meant all along—consistent with how that law had been interpreted in other circuits. See infra Part A.2.c. Moreover, and specific to the ATRA, no one disputes that pre-ATRA attempts to benefit from forced labor imposed criminal liability; the only pre-ATRA dispute concerned whether victims of that crime could recover civilly. Imposing civil liability retroactively on criminal conduct does not raise the same fairness concerns that are present in other contexts.

The ambiguity of the statute is also evidenced by the circuit split that existed before Congress passed the ATRA. Before we decided Ratha I, two other circuit courts had implicitly concluded that civil liability attaches for attempting to benefit financially, or to receive anything of value, from participation in a venture involving human trafficking.

In Ricchio v. McLean, 853 F.3d 553 (1st Cir. 2017), the First Circuit reversed a district court's dismissal of a plaintiff's civil claims under the TVPRA. Among the plaintiff's claims was Claim 6, which alleged that 'the defendants at the least attempted to violate §§ 1589, 1590, and 1591.” Id. at 557 (emphases added). By reversing the dismissal of that claim, the First Circuit necessarily concluded, at least implicitly, that via § 1594(a) the TVPRA gave rise to civil liability for 'attempt and conspiracy to violate §§ 1589, 1590, or 1591.' Id. at 558 (emphases added). Furthermore, in Claim 1 the plaintiff alleged that the Page 25defendants knowingly benefitted from forced labor, in violation of § 1589, even though the defendants did not extract forced labor from the plaintiff directly. Id. at 556. Taken together, the First Circuit’s analysis of Claims 1 and 6 reinforces that an attempt to benefit from forced labor states a claim under the TVPRA. As the First Circuit concluded, “the objective of forced labor . . . need not be satisfied for liability to attach.” Id. at 558.

The dissent attempts to harmonize Ricchio and Ratha I by drawing a hard line between theories of perpetrator liability and beneficiary liability. Dissent at 48-51. But Ricchio does not support that purported distinction. Rather, Ricchio recognized that benefitting from forced labor is a violation of § 1589—that is, a defendant who benefits from forced labor is a “perpetrator” for purposes of civil liability under § 1595(a). Moreover, Ricchio's broadly worded holding encompasses the concept that, in the dissent’s own words, an 'attempt to violate [§] 1589' states a claim for civil liability. Dissent at 50. That holding conflicts with Ratha I, which concluded that Congress had not “intended to create civil liability under § 1595 for attempts to benefit.” 35 F.4th at 1176. Of equal importance, Plaintiffs here contend that Defendant perpetrated a violation of § 1589. See Compl., D. Ct. Dkt. No. 1 ¶¶ 98–99, 123–26, 128–30.

The dissent's only real response to Ricchio is that Plaintiffs failed to plead adequately an attempted violation of § 1589. Dissent at 47–48, 51 n.5. But, at the summary judgment stage, the parties vigorously litigated the issue of whether an attempt to violate § 1589 states a claim. Indeed, the whole question in this appeal is whether the district court’s ruling after considering the parties’ arguments on that point-that Congress did not intend for an attempted violation of § 1589 to operate retroactively in favor of Page 26Plaintiffs—is legally correct. When the parties litigate an issue and the district court rules on it, the issue is part of the case as if pleaded and, therefore, the issue is preserved for our review. See Fed. R. Civ. P. 15(b)(2) (“When an issue not raised by the pleadings is tried by the parties' express or implied consent, it must be treated in all respects as if raised in the pleadings.'); see also Desertrain v. City of Los Angeles, 754 F.3d 1147, 1154 (9th Cir. 2014) (holding that even if plaintiffs fail to raise a claim properly in the pleadings, the claim may be raised at summary judgment under Rule 15(b)); Madeja v. Olympic Packers, LLC, 310 F.3d 628, 636 (9th Cir. 2002) (holding that a motion to conform a complaint to the evidence is 'irrelevant' and unnecessary when the district court considered an unpleaded claim on the merits). Probably for that reason, the district court and Ratha I did not rely on the complaint’s lack of an express allegation of attempt liability in granting and affirming summary judgment for Defendant. See Ratha v. Phatthana Seafood Co., No. CV 16-4271-JFW (ASx), 2017 WL 8293174, at *6 (C.D. Cal. Dec. 21, 2017) (order) (considering Plaintiffs' theory of attempt liability on the merits at summary judgment, albeit briefly, and finding no “legal support” for it).

Also presumably for the same reason, Defendant's response to Plaintiffs' Rule 60(b)(6) motion did not argue that Plaintiffs failed to plead attempt liability. Defendant argued only that the district court’s judgment is “adequately supported by independent grounds . . . unrelated to Plaintiffs’ attempt to benefit theory,” that the ATRA does not apply retroactively to Defendant’s conduct, and that “there are no ‘extraordinary circumstances’ justifying vacatur under Rule 60(b)(6).' (capitalization altered). Defendant mentioned that it had argued at summary judgment thatPage 27 “Plaintiffs’ attempt to benefit theory had not been pled,” but Defendant did not reassert that argument and thereby failed to preserve it for our review. See BankAmerica Pension Plan v. McMath, 206 F.3d 821, 826 (9th Cir. 2000) (refusing to reach an argument that the appellants initially raised, but later abandoned, before the district court). And, regardless of these principles, any pleading deficiency is irrelevant to the question whether the TVPRA provided for civil liability for an attempt to benefit—the point on which Ricchio and Ratha I disagree.

In Roe v. Howard, 917 F.3d 229 (4th Cir. 2019), the Fourth Circuit held that 'the text of § 1595 shows that it applies coextensively with its [criminal] predicate offenses.” Id. at 243 (emphasis added). The Fourth Circuit thus held that Congress intended to authorize civil actions for all the same conduct that gives rise to criminal liability. It is undisputed that, beginning in 2008, criminal liability attached to any attempt to knowingly benefit from forced labor. Pub. L. 110-457, § 221, 122 Stat. 5044, 5067 (2008).⁷

After the First Circuit’s decision in 2017, and again after the Fourth Circuit’s decision in 2019, Congress saw no need to act. But we reached the opposite statutory interpretation in Ratha I. We held that the lack of the words “or attempts” in § 1595 itself was fatal to Plaintiffs’ claims. Ratha I, 35 F.4th at 1176. In so doing, we rejected the interpretations of the First and Fourth Circuits, namely, that under the TVPRA criminal liability gives rise to coextensive civil liability, including for attempted violations of § 1589. Accordingly, Page 28when Congress passed the ATRA, it ended the circuit split created by Ratha I.

The dissent is simply wrong when it asserts that the existence of a circuit split is irrelevant to our assessment of retroactivity. Dissent at 48. See, e.g., ABKCO Music, 217 F.3d at 690-91 (considering, when deciding whether an amendment was retroactive, whether an ambiguity in the earlier statute created a circuit split before Congress chose one of the competing interpretations); Callejas v. McMahon, 750 F.2d 729, 731 (9th Cir. 1984) (noting that “a ‘dispute or ambiguity, such as a split in the circuits, [is] an indication that a subsequent amendment is intended to clarify, rather than change, the existing’” meaning of a statute (alteration in original) (citation omitted)); see also Adhikari v. Kellogg Brown & Root, Inc., 845 F.3d 184, 201–02 (5th Cir. 2017) (noting that the existence of a conflict or ambiguity before Congress passes an amendment is a factor that informs whether the amendment merely makes what Congress intended all along even more unmistakably clear, and thus is retroactive, and holding that an amendment to § 1596 of the TVPRA was not retroactive in part because “no circuit split or conflict . . . ‘provoked’ Congress to ‘enact an amendment to clarify rather than [to] change the law’” (citation omitted)).

Congress also indicated its retroactive intent by acting quickly. Congress’s overturning of a court’s interpretation of a statute does not-'by itself'-signal retroactivity. Rivers, 511 U.S. at 304–05. But “[i]t is the duty of a court in construing a statute to consider time and circumstances surrounding the enactment as well as the object to be accomplished by it.” Callejas v. McMahon, 750 F.2d 729,Page 29 731 (9th Cir. 1984). As we and other circuits have held, a fast-acting legislative body that amends a statute in the face of an ambiguity or a dispute among courts as to the meaning of the statute suggests that the change is a mere clarification, rather than a substantive change in the law. See, e.g., id. (“[A] ‘dispute or ambiguity, such as a split in the circuits, [is] an indication that a subsequent amendment is intended to clarify, rather than change, the existing law.'' (second brackets in original) (citations omitted)); McCoy v. Chase Manhattan Bank, USA, Nat’l Ass’n, 654 F.3d 971, 974 (9th Cir. 2011) ('If the amendment was enacted soon after controversies arose about interpretation of the original act, it is logical to regard the amendment as a legislative interpretation of the original act . . . .” (quoting 1A Norman J. Singer & J.D. Shambie Singer, Sutherland Statutes and Statutory Construction, § 22.31 (7th ed. 2011))).

We first published Ratha I on February 25, 2022. 26 F.4th 1029, as amended and superseded on denial of reh’g en banc, 35 F.4th 1159 (9th Cir. May 31, 2022). Before Ratha I, no other circuit court had interpreted the TVPRA as we had on the issue of civil attempt liability; thus, up to that point, Congress had no reason to act. But, less than seven months after our amended opinion issued in Ratha I, and a few days after the Supreme Court denied certiorari, 8 Congress amended the ATRA, adding the precise words that we had held were missing from the statute. See All Actions:

Page 30S.3946 — 117th Congress (2021–2022),

https://www.congress.gov/bill/117th-congress/senatebill/3946/all-actions (last visited Nov. 4, 2025) (noting that the Senate and the House considered and passed the bill on December 20, 2022, and December 22, 2022, respectively).

That timeline strongly suggests that Congress acted to resolve the disagreement between circuit courts and to correct Ratha I’s error. Congress often moves slowly, and we have considered timeframes longer than ten months to be quick enough to weigh in favor of retroactivity. See, e.g., McCoy, 654 F.3d at 974-75 (finding legislation passed within fifteen months of an erroneous judicial interpretation to be clarifying). Additionally, lawmakers introduced the relevant amendment on the same day that the Senate passed it. See All Actions: S.3946 — 117th Congress (2021–2022). Had the ATRA created an entirely new form of liability, one might expect a last-minute amendment to receive at least some discussion on the legislative floor. Congress’s speed and its minimal discussion strongly suggest that Congress intended to clarify retroactively what the law already meant, not to make a substantive change in the law.

It also is worth noting that the ATRA took effect immediately when the President signed it in January 2023. There was no notation that the ATRA would take effect on a later date, so the general rule that acts take effect immediately, once enacted, applies. See Gozlon-Peretz v. United States, 498 U.S. 395, 404 (1991) (stating rule). A later effective date would have suggested that Congress did not intend for the ATRA to be applied retroactively. See Fitzgerald v. Century Park, Inc., 642 F.2d 356, 359 (9th Cir. 1981) (interpreting an amendment as not intended to be retroactive when it took effect six months after its enactment, noting that it was “unlikely that Congress would delay the Page 31effective date of amendments which are to be applied retroactively”).

In sum, Congress intended for the ATRA to have retroactive effect. Congress made its intent clear by clarifying the meaning of an ambiguous statute, by labeling the amendment “technical and clarifying,” and by passing its clarification soon after we misinterpreted the statute’s meaning. These indicia are similar to those that we have found sufficient in the past to warrant retroactive application of a law. See, e.g., McCoy, 654 F.3d at 974 (holding a law to be retroactive because it was enacted soon after courts disagreed as to its original meaning); Belshe, 132 F.3d at 1266 (holding a law to be retroactive because it clarified an ambiguity that had led to a “split of authority”).

The dissent faults us for considering indicia of congressional intent that are not identified in Landgraf. Dissent at 40. But Landgraf does not spell out the factors to be considered when determining whether Congress intended to act retroactively. See 511 U.S. at 262-63 (considering legislative history and statutory framework). As a general matter, we consider 'all available evidence' to determine Congress’s intent. See Carnero v. Bos. Sci. Corp., 433 F.3d 1, 8 (1st Cir. 2006) (quoting Sale v. Haitian Ctrs. Council, Inc., 509 U.S. 155, 177 (1993)). Indeed, in Rivers, a companion case to Landgraf, the Court looked to some of the factors that we consider here: the section’s title and whether the statute had a “restorative purpose,” meaning the purpose to override a judicial interpretation of the statute at issue. Rivers, 511 U.S. at 307-08, 311. And contrary to the dissent's contention, cases decided before Landgraf are relevant to the present inquiry. See, e.g., Fitzgerald, 642 F.2d Page 32at 358-59 (analyzing congressional intent to determine whether an amendment applied retroactively). The same is true of our later cases that we now overrule insofar as they did not apply Landgraf. See, e.g., Belshe, 132 F.3d at 1265– 66. We need not throw out the baby with the bathwater: those decisions remain valid to the extent that they considered indicia of congressional intent.

Finally, the dissent errs by looking at each indicator of congressional intent in isolation. Cf. United States v. Arvizu, 534 U.S. 266, 274 (2002) (rejecting a “divide-and-conquer analysis'). We do not suggest that any factor alone is sufficient. Contra Dissent at 39-40. For example, the 'clarifying' label does not, by itself, indicate Congress's intent. See Rivers, 511 U.S. at 311 (discussing the weight accorded to a statute’s stated purpose). But considering all of the available evidence together, we conclude that Congress made its intent clear.

Section 1595's retroactivity fatally undermines one of the three grounds on which the district court had premised its grant of summary judgment. Rubicon argues that, even if § 1595(a) is retroactive, the two remaining grounds are sufficient to justify denial of Plaintiffs’ Rule 60(b) motion: (1) Plaintiffs did not demonstrate that Rubicon knowingly participated in a human trafficking venture; and (2) Plaintiffs did not demonstrate that Rubicon knew, or should have known, that human trafficking was occurring at Phatthana’s factory. We must now determine whether either of those grounds is legally sound. We may review the district court’s reasons for granting summary judgment because the court incorporated them into the decision under review by relying Page 33on them to deny Plaintiffs’ Rule 60(b) motion. As the district court stated: '[T]he Court agrees with Rubicon that the Court’s decision [on summary judgment] is fully supported by the first and second grounds and the change to Section 1595(a) does not apply to or in any way affect the first and second grounds of the Court's decision. As a result, the Court concludes that because the change to Section 1595(a) would not alter the Court’s decision in favor of Rubicon,” the Phelps factors cannot avail Plaintiffs. Ratha, 2023 WL 2762044, at *4 (emphases added); see Mitchell v. United States, 958 F.3d 775, 784 (9th Cir. 2020) (noting that we review de novo 'questions of law underlying the district court’s” denial of a Rule 60(b) motion (emphasis added)).⁹ Additionally, we are not bound by any earlier decision of our court with respect to those two issues, as Ratha I upheld summary judgment for Rubicon solely on the grounds that Rubicon did not actually benefit and that the statute did not provide a civil remedy for attempts to benefit. 35 F.4th at 1175–76.

The district court made two related errors of law when it held that Plaintiffs did not present evidence that Rubicon participated in a human trafficking venture.

First, the district court applied a legally erroneous definition of “participate,” requiring Plaintiffs to show that Rubicon 'took some action to operate or manage the venture.' Ratha, 2017 WL 8293174, at *4-5 (citation omitted). The district court's definition is contrary to the ordinary meaning of “participate,” which is “to take part.” Participate, Merriam–Webster, https://www.merriamwebster.com/dictionary/participate (last visited Nov. 4, 2025); see Doe #1 v. Red Roof Inns, Inc., 21 F.4th 714, 724– 25 (11th Cir. 2021) (defining “participate” as “to take part in or share with others in common in or in association” (citing Black’s Law Dictionary (11th ed. 2019) (defining “participation” as “The act of taking part in something.”) and Oxford English Dictionary 268 (2d ed. 1989) (defining “participate” as “[a] taking part, association, or sharing (with others) in some action or matter.”))). The ordinary meaning applies because the statute contains no special definition. See Chacon v. Wilkinson, 988 F.3d 1131, 1134 (9th Cir. 2021) (holding that we interpret a term according to its ordinary meaning when Congress does not define it). The district court's erroneous definition of 'participate' held Plaintiffs to too high a bar; one can participate in a venture without operating or managing it.

Second, the district court did not view all facts and inferences in favor of Plaintiffs. See Anderson, 477 U.S. at 255 (“The evidence of the non-movant is to be believed, and all justifiable inferences are to be drawn in [their] favor.”).

Page 35The district court stated that it did not consider the factual disputes to be “material to the disposition of [the] motion.” Ratha, 2017 WL 8293174, at *1 n.1. But Rubicon's marketing materials touted that it was in “control [of] every aspect of production,” which is material to a consideration of Rubicon’s participation.

Applying the proper definition of participation, and viewing the facts in the light most favorable to Plaintiffs, a reasonable jury could find that Rubicon participated in a human trafficking venture. Thus, the district court erred in relying on that ground when it denied Plaintiffs' Rule 60(b)(6) motion.

The district court again doubly erred in holding that the record does not contain evidence that Rubicon knew or should have known about Phatthana’s human trafficking at the time of its attempt to benefit. The district court ignored a holding from Ratha I that remained law of the case—that Rubicon had knowledge of Phatthana's alleged human trafficking at least by February 2012. See Ratha I, 35 F.4th at 1177 (stating that 'Rubicon was undisputedly aware of Ratha's whistleblower report' beginning on February 23, 2012). Because Rubicon attempted to sell shrimp to Walmart after that date, it attempted to benefit with the requisite knowledge.

Even if that holding in Ratha I had not been binding, the district court also failed to view the evidence in the light most favorable to Plaintiffs. A reasonable jury could infer that Rubicon should have known about the conditions at the factories as early as October 2011, when Walmart refused to Page 36purchase shrimp because it was concerned about the conditions at the factory. Ratha I, 35 F.4th at 1166.

Because a reasonable jury could have found that Rubicon knew or should have known about Phatthana's alleged human trafficking, the district court erred in relying on that ground when it denied Plaintiffs’ Rule 60(b)(6) motion.

Having found legal error in all three grounds on which the district court relied when denying Plaintiffs' Rule 60(b)(6) motion, we now consider whether the district court’s denial of Plaintiffs’ motion was, itself, error. We hold that it was.

Plaintiffs sought relief “from a final judgment, order, or proceeding” of the district court under Rule 60(b). Fed. R. Civ. P. 60(b). Rule 60(b)(6) allows a court to grant relief for “any other reason that justifies relief.” But relief under the rule is appropriate only when “extraordinary circumstances justify[] reopening the judgment.” Bynoe, 966 F.3d at 979 (citation and internal quotation marks omitted). A clear and authoritative change in law may constitute such extraordinary circumstances, depending on the consideration of certain factors:

For Rule 60(b)(6) motions premised on post-judgment changes in law, we have distilled the extraordinary-circumstances requirement into six factors, considered flexibly and in their totality. We examine (1) the nature of the legal change, including whether the change in law resolved an unsettled legal question; (2) whether the

Page 37movant exercised diligence in pursuing reconsideration of his or her claim; (3) the parties’ reliance interests in the finality of the judgment; (4) the delay between the finality of the judgment and the Rule 60(b)(6) motion; (5) the relationship between the change in law and the challenged judgment; and (6) whether there are concerns of comity that would be disturbed by reopening a case.

Id. at 983 (citing Phelps, 569 F.3d at 1134–40).

The district court considered the six factors and denied Plaintiffs' Rule 60(b)(6) motion. It determined that the second factor (Plaintiffs’ diligence) and the fourth factor (the delay between judgment and motion) weighed in favor of granting Plaintiffs' motion, that the third factor (reliance interests) was neutral, and that the sixth factor (comity) was inapplicable. We see no error in the district court’s weighing of those four factors, nor do the parties ask us to.

The district court determined that the first factor (the nature of the legal change) and the fifth factor (the relationship between the change in law and the challenged judgment) weighed heavily against granting the motion. But the legal errors described above directly affected the court’s assessment of those factors. Thus, the district court abused its discretion by relying on an erroneous view of the law in weighing factors one and five. Moreover, given our legal conclusions above, the first and fifth factors weigh in favor of granting Plaintiffs' motion: the ATRA resolved an ambiguity and had retroactive effect, directly allowing Plaintiff to bring a civil claim for attempt liability. In sum, four factors weigh strongly in favor of granting Plaintiffs’ motion, one factor is neutral, and one is inapplicable.

Page 38Accordingly, Plaintiffs are entitled to relief under Rule 60(b)(6) from the district court’s final order.

REVERSED and REMANDED for further proceedings consistent with this opinion.

CALLAHAN, Circuit Judge, in which M. SMITH, Circuit Judge, joins except as to Part I.A., BADE, Circuit Judge, joins in full, BRESS, Circuit Judge, joins as to Part I only, and JOHNSTONE, Circuit Judge, joins in full, dissenting:

The facts and procedural history of this case are not in question. For that reason, I do not restate them here. My disagreement lies with the majority’s legal reasoning, analysis, and ultimate holding. Therefore, I turn directly to a discussion of the law. This case presents two important issues: (1) the test for determining if a statute is retroactive and (2) the scope of a Federal Rule of Civil Procedure 60(b) appeal. Because the majority errs in addressing both issues, I respectfully dissent.

I agree with the majority that Landgraf v. USI Film Prods. , 511 U.S. 244 (1994), governs our retroactivity analysis and that under Landgraf we will give retroactive effect to a statute (1) if Congress expressly makes the statute retroactive, or (2) if the statute would have retroactive effect and Congress clearly intended that result.¹ Id. at 280. I also Page 39agree that our prior cases that created an exception for clarifying amendments, exempting them from the Landgraf framework, are inconsistent with Landgraf itself, and thus the majority properly overrules them. And finally, I agree that the Abolish Trafficking Reauthorization Act of 2022 ('ATRA') would have retroactive effect and thus the presumption against retroactivity applies. My agreement ends there.

After purging our precedent of past inconsistent decisions, the majority makes the curious choice to fold the reasoning of these 'overruled' cases into its Landgraf analysis, reinfecting the law it claims to cleanse.² Specifically, in its quest to find retroactive intent the majority continues to focus on whether the ATRA was a “clarification,” instead of searching for any real indication of clear congressional intent of retroactivity. In other words,Page 40 the majority is under the mistaken belief that if legislation is clarifying, it automatically demonstrates clear congressional intent of retroactivity. Not so.

The majority cites four factors, which it believes, 'considered together, clearly establish the amendment's retroactivity:” (a) Congress expressly stated that its enactment was a technical and 'clarifying' update; (b) 18 U.S.C. § 1595(a) was ambiguous; (c) Ratha v. Phatthana Seafood Co. ( Ratha I ), 35 F.4th 1159 (9th Cir. 2022), created a circuit split concerning the interpretation of § 1595(a); and (d) Congress acted swiftly, and with immediate effect, following our decision in Ratha I . Maj. Op. at 16. These factors are not found in Landgraf and find no home in the Landgraf progeny. The majority, nonetheless, seems to imply that these factors demonstrate clear congressional intent that the ATRA is retroactive. But a close read of the majority’s analysis dispels this conclusion.

The majority's first factor is that 'Congress expressly stated that its enactment was a technical and 'clarifying update.'' Maj. Op. 16-17. The majority cites Beverly Cmty. Hosp. Ass'n v. Belshe , 132 F.3d 1259 (9th Cir. 1997), to support this factor. Oddly enough, however, Belshe never cites Landgraf and does not purport to apply the Landgraf test. Rather, Belshe is among the cases the majority 'overruled' just paragraphs before as inconsistent with Landgraf . Belshe stands for the proposition that if a law is “clarify[ing]” it “state[s] more clearly what the law already was” and thus “has no retroactive effect that might be called into constitutional question.” Belshe , 132 F.3d at 1265. In other words, Belshe holds that if a law is clarifying, it does Page 41not require a retroactivity analysis because it simply states what the law always was. As we all agree, this holding is inconsistent with Landgraf and the majority’s attempt to fold it into its Landgraf analysis does not cure that inconsistency; it simply makes its reasoning erroneous.

Looking past the majority’s reliance on our now overruled precedent, this factor does not establish clear congressional intent of retroactivity. While it is true that “the title of a statute” is a “‘tool[] available for the resolution of a doubt’ about the meaning of a statute,” that tool is of little use here. Almendarez-Torres v. United States , 523 U.S. 224, 234 (1998) (quoting Trainmen v. Baltimore & Ohio R. Co. , 331 U.S. 519, 528-29 (1947)). As discussed in more detail below, the statute at issue here, 18 U.S.C. § 1595(a), was clear prior to the ATRA—there was nothing about the text of the statute for Congress to “clarify.” The meaning of the statute was not in doubt. So why then did Congress label the legislation as “clarifying?” Under the majority’s view, Congress was using “our own language, directing us to apply the amendment retroactively.' Maj. Op. at 19. But if Congress wanted to 'direct[] us to apply the amendment retroactively,' then Congress would simply state that the statute applies retroactively. Here, it did not. What Congress likely meant is that it recognized that § 1595(a) did not impose liability on those who attempt to benefit from unlawful conduct. Upon realizing that, Congress amended the statute to create such liability, saying nothing of the retroactive effect. Under these circumstances, the word 'clarifying' in the title of the ATRA does not show that Congress was expressing its intent, sub silentio , for the statute to apply retroactively.

Additionally, the majority relies on Rivers v. Roadway Exp., Inc. , 511 U.S. 298 (1994), to bolster its analysis Page 42regarding the “clarifying nature of Congress’s amendment.” Maj. Op. at 17-19. The majority's reliance on Rivers is ironic, because Rivers is a blueprint outlining precisely why the majority is wrong. In Rivers , the Supreme Court determined that § 101 of the Civil Rights Act of 1991 was not retroactive. Id. at 300. The Court held that 'Congress' decision to alter the rule of law established in [ Patterson v. McLean Credit Union , 491 U.S. 164 (1989)]—as petitioners put it, to 'legislatively overrul[e]' -does not, by itself, reveal whether Congress intends the ‘overruling’ statute to apply retroactively.” Id. at 304 (internal citation omitted). Indeed, as the Court noted “the choice to enact a statute that responds to a judicial decision is quite distinct from the choice to make the responding statute retroactive.” Id. at 305.

The petitioners in Rivers argued 'that the structure and legislative history of § 101 indicate that Congress specifically intended to ‘restore’ prior law even as to parties whose rights would otherwise have been determined according to Patterson 's interpretation of § 1981.' Id. They also pointed 'to evidence in the 1991 Act's legislative history indicating legislators’ distress with Patterson ’s construction of § 1981 and their view that [the Court's] decision had narrowed a previously established understanding of that provision.” Id. at 305-06. Accepting these arguments, the Court assumed “that § 101 reflect[ed] congressional disapproval of Patterson 's interpretation of § 1981' and also assumed 'that many or even most legislators believed that Patterson was not only incorrectly decided but also represented a departure from the previously prevailing understanding of the reach of § 1981.' Id. at 30607. But the Supreme Court held that those assumptions only explain 'why Congress might have wanted to legislate Page 43retroactively” not whether Congress intended the legislation to be retroactive. Id. at 307. The Court held that those assumptions were insufficient to demonstrate clear congressional intent of retroactivity. Id. Because there was no clear expression of congressional intent that the statute apply retroactively, the Court declined to apply the statute retroactively. Id. at 313.

Applying Rivers to the instant case makes quick work of the majority’s position. Pursuant to Rivers we may assume for purposes of our retroactivity analysis that the ATRA was a clarification, that Congress disapproved of our decision in Ratha I , and that 'many or even most legislators believed that” Ratha I was incorrectly decided, but nonetheless still find that the ATRA does not apply retroactively. That is because, as Rivers makes clear, those assumptions do not demonstrate a clear expression of congressional intent that the statute should apply retroactively. For this reason, our role is to determine, outside of these assumptions, if there is a clear expression of congressional intent. As outlined in more detail below, no such clear expression exists here.

“Congress . . . has the power to amend a statute that it believes [a court has] misconstrued” and even “make such a change retroactive and thereby undo what it perceives to be the undesirable past consequences of a misinterpretation of its work product.” Rivers , 511 U.S. at 313. But Congress’ “intent to reach conduct preceding the ‘corrective’ amendment must clearly appear.' Id. Because no clear expression of retroactive intent exists in this case, we should apply the reasoning in Rivers and conclude that the ATRA does not apply retroactively.

The majority's second factor is that 'the statute was ambiguous.” Maj. Op. at 16, 20. The majority cites no case supporting the use of this factor in its Landgraf analysis but nonetheless uses this factor as a vehicle to arrive at its desired destination. This is likely because the majority needs the statute to be ambiguous for it to conclude that the ATRA was “clarifying.” But again, the question is not whether the ATRA clarified an ambiguity, the question is whether Congress clearly intended it to be retroactive. Indeed, the rationale of Rivers applies with equal force to this factor. Congress may have thought that Ratha I created an ambiguity and enacted the ATRA in response, but identifying a possible reason for congressional action “does not supply sufficient evidence of clear congressional intent to overcome the presumption against statutory retroactivity.” Rivers , 511 U.S. at 309. There is simply no evidence that Congress clearly intended the ATRA to reach cases that arose before its enactment.

Even accepting ambiguity as a relevant factor in determining legislative intent, a review of the statute makes clear that it is not, in fact, ambiguous. The majority strives to demonstrate that before the ATRA, the statute was ambiguous. But the statute's pre-amendment clarity is revealed by simply reading its text and understanding Chapter 77’s statutory framework.

The statute: At the time Plaintiffs filed suit, the statute at issue, 18 U.S.C. § 1595(a), provided:

An individual who is a victim of a violation of this chapter may bring a civil action against the perpetrator (or whoever

Page 45knowingly benefits, financially or by receiving anything of value from participation in a venture which that person knew or should have known has engaged in an act in violation of this chapter) in an appropriate district court of the United States and may recover damages and reasonable attorneys fees.

18 U.S.C.A. § 1595 (2008). The violations referred to in § 1595(a) are the criminal violations contained in Chapter 77. Thus, a victim of a criminal violation of Chapter 77 could sue (1) a perpetrator of the violation and (2) a knowing beneficiary of the violation. The statute, however, does not impose civil liability on someone who attempts to benefit from that violation. Therefore, a victim could not sue an “attempted beneficiary.”

In general terms, a perpetrator violates Chapter 77 if they (1) engage in peonage or obstruct enforcement against peonage (§ 1581), (2) are involved with a vessel for slave trade (§ 1582), (3) engage in enticement of others into slavery (§ 1583), (4) engage in the sales of others into slavery (§ 1584), (5) engage in the seizure, detention, or transportation of slaves (§ 1585), (6) are involved with service on vessels in slave trade (§ 1586), (7) possess slaves aboard a vessel (§ 1587), (8) transport slaves from the United States (§ 1588), (9) engage in forced labor (§ 1589), (10) engage in trafficking with respect to peonage, slavery, involuntary servitude, or forced labor (§ 1590), (11) engage in sex trafficking of children or sex trafficking by use of force, fraud, or coercion (§ 1591), (12) engage in unlawful conduct regarding documents to further trafficking, peonage, slavery, involuntary servitude, or forced laborPage 46 (§ 1592), or (13) knowingly benefit financially from peonage, slavery, and trafficking in persons (§ 1593A).

Additionally, a perpetrator can violate Chapter 77 if they attempt to (1) engage in peonage or obstruct enforcement against peonage (§ 1581), (2) engage in enticement of others into slavery (§ 1583), (3) engage in the sales of others into slavery (§ 1584), (4) engage in forced labor (§ 1589), (5) engage in sex trafficking with respect to peonage, slavery, involuntary servitude, or forced labor (§ 1590), or (6) engage in sex trafficking of children or sex trafficking by use of force, fraud, or coercion (§ 1591). See 18 U.S.C. § 1594 (2008).

How the statute works: Recall that a victim could sue (1) a perpetrator, and (2) a knowing beneficiary. But a victim could not sue an attempted beneficiary. For example, a victim of peonage could sue (1) the person who held the victim in peonage (the perpetrator), and (2) anyone who knowingly benefited from the victim’s peonage (the knowing beneficiary). But the victim could not sue someone who attempted to benefit from the victim's peonage (the attempted beneficiary). Similarly, a victim of attempted peonage could sue (1) the person who attempted to hold the victim in peonage (the perpetrator), and (2) anyone who knowingly benefited from that attempt (the knowing beneficiary). But the victim could not sue someone who attempted to benefit from the attempted peonage (the attempted beneficiary).

Majority’s confusion: The majority confuses this statutory framework. The majority asserts that there is a parallelism between the criminal violations and civil liability—that for any criminal violation, including attempt, there must be a corresponding civil liability. Maj. Op. at 22Page 4723. This is largely correct and should have led the majority to the unremarkable conclusion that a perpetrator could be held liable for attempted conduct (because attempted conduct could amount to a violation of Chapter 77), but a knowing beneficiary could not. Instead, the majority concluded that this parallelism somehow imposed attempt liability on beneficiaries. It did not.

The majority calls upon § 1589 for support, but § 1589 does not and cannot answer that call. Maj. Op. at 22. Far from supporting the majority’s position, § 1589 proves its error. Section 1589 establishes two theories of criminal liability for forced labor: (a) to knowingly provide or obtain forced labor, and (b) to knowingly benefit from forced labor. 18 U.S.C. § 1589(a), (b). If someone violates either subsection (a) or subsection (b) they are a perpetrator for purposes of § 1595(a) liability. Section 1594 makes it a criminal violation to attempt to violate § 1589. Id. at § 1594. Accordingly, if someone violates § 1594 by either (a) attempting to provide or obtain forced labor, or (b) attempting to knowingly benefit from forced labor, they too are a perpetrator for purposes of § 1595(a) liability. Here, however, Plaintiffs alleged claims against Rubicon Resources, LLC (“Rubicon”) under a knowing beneficiary theory of liability, not a perpetrator theory.

Plaintiffs may have been able to allege a claim against Rubicon as a perpetrator of forced labor under § 1589(b) and § 1594, rather than as a knowing beneficiary. Plaintiffs, however, did not plead such a claim and have never argued that Rubicon was a perpetrator of forced labor under §§ 1589(b) and 1594. Indeed, there are no factual allegations in the complaint that Rubicon attempted to do anything, much less that they attempted to violate § 1589(b). Nonetheless, Plaintiffs' ability to allege such a claim Page 48highlights the majority’s misunderstanding of the statutory framework while also demonstrating the framework’s clarity. Thus, to the extent that ambiguity is relevant to our Landgraf analysis, it weighs against retroactivity because the operation of § 1595(a), while perhaps complex, was not ambiguous.

The majority's third factor is that the statute was the subject of a circuit split. Maj. Op. at 16, 24. The majority asserts that our previous decision in Ratha I created a circuit split with both Ricchio v. McLean , 853 F.3d 553 (1st Cir. 2017), and Roe v. Howard , 917 F.3d 229 (4th Cir. 2019). To be clear, whether a circuit split existed does not tell us anything meaningful about whether Congress intended the ATRA to be retroactive. Regardless, the majority is mistaken—there is no circuit split. The majority’s misunderstanding here stems largely from its confusion, discussed above, about how the statute works.

In Ricchio , a victim sued a series of individuals under a perpetrator theory of liability.³ Ricchio , 853 F.3d at 556-57. The First Circuit reversed the district court’s dismissal of the plaintiff’s claims, finding that the complaint alleged sufficient facts to sustain a theory of perpetrator liability under the applicable statutes. Id. at 558. This case did not establish attempt liability for beneficiaries, as the majority would have you believe. Nor could it. Ricchio has nothing to do with beneficiary liability; it is a case about perpetrator liability. If anything, Ricchio confirms that Plaintiffs here Page 49may have been able to bring a claim against Rubicon under a theory of perpetrator liability-but again, they did not. Therefore, Ricchio does not stand as a direct contradiction to our decision in Ratha I . Indeed, in Ratha I , we explicitly stated that “[n]either Rubicon nor Wales are alleged to have perpetrated any [] violations against Plaintiffs.' Ratha I , 35 F.4th at 1175. Comparing Ricchio to Ratha I is a classic case of comparing apples (a case about perpetrator liability) to oranges (a case about beneficiary liability).

The majority disagrees with this reading of Ricchio . It believes that Ricchio did not allege claims of actual and attempted conduct and mistakenly thinks that Ricchio is not a case about perpetrator liability. The majority’s confusion here is consistent with its confusion about the statutory framework of Chapter 77 as discussed above in the section labeled “Majority’s confusion.” To once again try to dispel this confusion, I reiterate what I explained above, this time using the claims in Ricchio as examples.

Claim One in Ricchio alleges a violation of §§ 1589 and 1595(a).⁴ The violation of § 1589 provides the Chapter 77 violation which gives rise to civil liability under § 1595(a). A person can violate § 1589 if they knowingly provide or obtain forced labor or if they knowingly benefit from forced labor. A person who violates § 1589 is a perpetrator for purposes of § 1595(a). Therefore, Claim One is a claim against an alleged perpetrator who allegedly violated § 1589 by knowingly benefiting from forced labor. Ricchio , 853 F.3d. at 556.

Page 50At the risk of belaboring the point, I turn to some of the remaining claims in Riccio to hammer the explanation home. Claim Two alleges a violation of §§ 1590 and 1595(a). The violation of § 1590 provides the Chapter 77 violation which gives rise to civil liability under § 1595(a). Section 1590 makes it unlawful to, inter alia , harbor any person for labor or services in violation of Chapter 77. A person who violates § 1590 is a perpetrator for purposes of § 1595(a). Therefore, Claim Two is a claim against an alleged perpetrator of § 1590. Id. Claim Three alleges a violation of §§ 1591 and 1595(a). The violation of § 1591 provides the Chapter 77 violation which gives rise to civil liability under § 1595(a). A person violates § 1591, inter alia , if they knowingly benefit, financially or by receiving anything of value, from participation in a venture which has engaged in an act in violation of § 1591(a)(1). A person who violates § 1591 is a perpetrator for purposes of § 1595(a). Therefore, Claim Three is a claim against an alleged perpetrator of § 1591. Id.

Claims Five and Six bring in conspiracy and attempt liability. Specifically, Claim Five alleges violations of § 1594(b) and (c), and § 1595(a). The violations of § 1594(b) and (c) provide the Chapter 77 violations which give rise to civil liability under § 1595(a). A person violates § 1594(b) and (c) if they, inter alia , conspire with another to violate §§ 1589, 1590, and 1591. A person who violates § 1594(b) and (c) is a perpetrator for purposes of § 1595(a). Therefore, Claim Five is a claim against an alleged perpetrator of § 1594(b) and (c). Id. at 556-57. Claim Six alleges a violation of §§ 1594(a) and 1595(a). The violation of § 1594(a) provides the Chapter 77 violation which gives rise to civil liability under § 1595(a). A person violates § 1594(a) if they attempt to violate §§ 1589, 1590, andPage 51 1591.⁵ Therefore, Claim Six is a claim against an alleged perpetrator of § 1594(a). Id. at 557. Thus, Ricchio is a case about perpetrator liability, which includes both actual and attempted conduct.

In Howard , a victim brought a case against her employer, Linda Howard, for abuses suffered while in her employment, including sexual assault. Howard , 917 F.3d at 237-38. In affirming the judgment against the employer, the Fourth Circuit concluded that 'the text of § 1595 shows that it applies coextensively with its predicate offenses . . . .' Id. at 243. This is correct and not in dispute. Perpetrators may be held liable both for their conduct and their attempted conduct because both are violations of Chapter 77. It does not follow that attempt liability applies to knowing beneficiaries. Ratha I did not reject the notion that § 1595 “applies coextensively with its predicate offenses.” Id. It simply recognized that there is no attempt liability for beneficiaries. These cases are thus in harmony.

Ricchio and Howard are not in tension with our previous decision in Ratha I . More importantly though, neither Ricchio nor Howard support the majority's position on retroactivity. These are two cases from two different circuits that properly understood the statutory framework of Chapter 77. Both cases are demonstrative of a clear functioning statutory scheme, not an ambiguous one, and using them to Page 52assert that there was confusion or ambiguity is an affront to their sound reasoning.

The majority's fourth factor is that 'Congress acted swiftly.' Maj. Op. at 16, 28. To support this factor the majority points to three cases, Rivers , 511 U.S. at 304-05, Callejas v. McMahon , 750 F.2d 729, 731 (9th Cir. 1984), and McCoy v. Chase Manhattan Bank, USA, Nat’l Ass’n , 654 F.3d 971, 974 (9th Cir. 2011). The majority’s reliance on Rivers continues to perplex. As discussed above, Rivers only highlights why the majority is wrong and importantly here, Rivers says nothing about the speed of congressional action. Callejas was decided preLandgraf and does not speak to the speed of congressional action, and McCoy was decided well after Landgraf but declined to apply the Landgraf test to its retroactivity analysis. But even putting aside the shortcomings of the majority's authority here, Congress’ purportedly “swift” action tells us little about the retroactive effect of its amendment.⁶ Here, § 1595(a) simply did not provide for a lawsuit against an attempted Page 53beneficiary and Congress acted to change that, creating new liability for attempted beneficiaries. The speed with which Congress created this new form of liability tells us nothing about whether Congress intended it to be retroactive.

Finally, the majority states that it is “worth noting that the ATRA took effect immediately.” Maj. Op. at 30. The majority relies on another preLandgraf decision to support this factor, Fitzgerald v. Century Park, Inc. , 642 F.2d 356, 357 (9th Cir. 1981). In Fitzgerald the court noted that '[i]t is unlikely that Congress would delay the effective date of amendments which are to be applied retroactively.” Id. at 359. Thus, if an amendment is retroactive, then it is unlikely that its effective date would be delayed. This sheds no light on whether an amendment is retroactive, it simply says that if it is, it will likely apply immediately.

Landgraf demands that Congress speak clearly if it wishes for a statute to apply retroactively. Landgraf , 511 U.S. at 268. In this case, our role is to search for that clear intent of retroactivity. I do not fault the majority for recognizing this as our duty. I fault the majority for divining clear congressional intent where it does not exist.

The majority opinion is rooted in a misunderstanding of the statutory framework of Chapter 77 and a pervasive confusion that I have been unable to remedy despite my best efforts. Compounding errors is the majority’s focus on the alleged clarifying nature of the ATRA, causing it to rely on erroneous reasoning from cases it has itself overruled. The cherry on top, perhaps, is that several of the cases the majority relies upon demonstrate why the majority is wrong. See Rivers , 511 U.S. 298; see also Ricchio , 853 F.3d 553.

Page 54These errors undermine the foundation of the majority's analysis, rendering the opinion irreparably flawed.

Far from correcting our precedent to align with Landgraf , the majority flees from it, abandoning Landgraf as quickly as it claimed to embrace it. The majority’s focus on whether the ATRA was “clarifying” an ambiguity blinded it from the real question: whether Congress clearly intended the statute to be retroactive. As discussed below, under a faithful application of Landgraf we should decline to give the ATRA retroactive effect because there is no clear indication that Congress intended it to be retroactive.

Landgraf is based on the simple and long-established presumption that legislation does not apply retroactively. Landgraf , 511 U.S. at 265 (observing that “the presumption against retroactive legislation is deeply rooted in our jurisprudence, and embodies a legal doctrine centuries older than our Republic”). The Supreme Court has instructed that, “[e]lementary considerations of fairness dictate that individuals should have an opportunity to know what the law is and to conform their conduct accordingly; settled expectations should not be lightly disrupted.” Id.

To overcome the presumption against retroactivity, the Supreme Court constructed a two-step framework. At step one, we determine whether Congress expressly provided that the statute be retroactive. Id. at 280. If it did, then our analysis ends, and we apply the statute retroactively. If “however, the statute contains no such express command,” the court moves to step two. Id. At step two, the court must “determine whether the new statute would have retroactive effect,” by asking whether the statute “would impair rights a party possessed when [they] acted, increase a party’s Page 55liability for past conduct, or impose new duties with respect to transactions already completed.” Id. If we find that 'the statute would operate retroactively, [then] our traditional presumption teaches that it does not govern absent clear congressional intent favoring such a result.” Id. In other words, '[i]f a new statute would 'impair rights a party possessed when [they] acted, increase a party’s liability for past conduct, or impose new duties with respect to transactions already completed,' then courts' should only give retroactive effect to the statute when “clear congressional intent favor[s] such a result.” Talaie v. Wells Fargo Bank, NA , 808 F.3d 410, 412 (9th Cir. 2015) (quoting Landgraf , 511 U.S. at 280).

Here, Congress did not expressly state that the ATRA was retroactive. Accordingly, we move to step two. Before the ATRA, § 1595(a) allowed a victim of a criminal violation of Chapter 77 to sue (1) a perpetrator and (2) a knowing beneficiary, but not an attempted beneficiary. The ATRA changed this statutory framework by adding the word “attempt” to § 1595(a). Specifically, it changed § 1595(a) from:

An individual who is a victim of a violation of this chapter may bring a civil action against the perpetrator (or whoever knowingly benefits, financially or by receiving anything of value from participation in a venture which that person knew or should have known has engaged in an act in violation of this chapter) in an appropriate district court of the United States

Page 56and may recover damages and reasonable attorneys fees,

(18 U.S.C. § 1595(a) (2008)), to

An individual who is a victim of a violation of this chapter may bring a civil action against the perpetrator (or whoever knowingly benefits, or attempts or conspires to benefit , financially or by receiving anything of value from participation in a venture which that person knew or should have known has engaged in an act in violation of this chapter) in an appropriate district court of the United States and may recover damages and reasonable attorneys fees.

(18 U.S.C. § 1595(a) (2023) (emphasis added)). This change created an entirely new form of liability. Following the ATRA, § 1595(a) allows a victim of a criminal violation of Chapter 77 to sue (1) a perpetrator, (2) a knowing beneficiary, and (3) an attempted beneficiary. This substantively changed § 1595 and more importantly increased the liability for beneficiaries of unlawful conduct. Because the ATRA increases a party's liability, here an alleged beneficiary of the statutory violation, we give it retroactive effect only if there is clear congressional intent favoring retroactivity. Landgraf , 511 U.S. at 280.

There is no indication from the text of the ATRA or from the legislative history of the ATRA that it was intended to apply retroactively. The majority clings to the word “clarifying” in the title of the amendment—noting Page 57throughout the opinion that the ATRA was a clarification— as if that is somehow a clear indication of congressional intent on retroactivity. As discussed above, it is not. Because we have properly washed away our precedent exempting “clarifications” from the Landgraf analysis, we must look beyond musing about clarifications to find “clear congressional intent.” See id. . Here, there is none.

What Congress did is quite simple, it changed § 1595(a) to impose attempt liability on beneficiaries. It is unclear exactly why Congress did so. Perhaps because Congress decided such liability should exist; perhaps Congress thought it had created such liability but realized that it had not; or perhaps Congress simply found something its members could all agree upon and passed the ATRA in a display of unity. What is clear, however, is that there is no evidence that Congress intended to impose this new form of liability retroactively, much less clear evidence of that intent. For these reasons, I would find that the ATRA does not apply retroactively and affirm the district court.⁷

The majority creates a novel rule for Rule 60(b) appeals. It holds that in an appeal from the denial of a Rule 60(b) motion we can 'review the district court's reasons for granting summary judgment because the court incorporated them into the [Rule 60(b) order] under review . . . .” Maj. Op. at 32-33. Let us pause to take stock of what the majority has done. The majority has reached beyond the boundaries of this appeal, grabbed the district court's summary Page 58judgment order and hoisted it upon the back of the district court’s Rule 60(b) order currently under review. How, one may ask, does the majority justify this? It cites no authority but rather posits a theory of incorporation by reference. Although the idea of such a piggyback appeal may be attractive to some, it runs counter to well-established law, and I cannot endorse it.

In 1978, the Supreme Court held that “an appeal from denial of Rule 60(b) relief does not bring up the underlying judgment for review.” Browder v. Dir., Dep’t of Corr. of Ill. , 434 U.S. 257, 263 n.7 (1978). Understanding this, we have held that although '[w]e review an appeal from the denial of Rule 60(b) relief for an abuse of discretion in denying the motion; the appeal does not bring the entire underlying judgment up for review.” Harman v. Harper , 7 F.3d 1455, 1458 (9th Cir. 1993). Indeed, every circuit in the country has recognized this rule. See Frew v. Young , 992 F.3d 391, 397 n.11 (5th Cir. 2021) ('[An] [a]ppeal from denial of Rule 60(b) relief does not bring up the underlying judgment for review' (quoting Browder , 434 U.S. at 263 n.7)); United States v. Hassebrock , 21 F.4th 494, 497 (7th Cir. 2021) ('Indeed, an appeal from the denial of a Rule 60(b) motion does not allow us to review the underlying decision.”); Crawford v. United States , 611 F. App’x 47, 48 (3d Cir. 2015) (“Our review is limited to the denial of Rule 60(b) relief and does not extend to the District Court's underlying order dismissing Crawford's habeas petition because ‘an appeal from the denial of Rule 60(b) relief does not bring up the underlying judgment for review.’” (quoting Browder , 434 U.S. at 263 n.7)); Perry v. United States , 558 F. App'x 1004, 1007 (Fed. Cir. 2014) ('Mr. Perry's Rule 60(b)(6) Motion cannot reopen this court’s prior jurisdictional decision for review.' (citing Browder , 434Page 59 U.S. 263 n. 7)); Aikens v. Ingram , 652 F.3d 496, 501 (4th Cir. 2011) (“‘[A]n appeal from denial of Rule 60(b) relief does not bring up the underlying judgment for review.'' (quoting Browder , 434 U.S. at 263 n.7)); United States v. Zorrilla-Echevarria , 671 F.3d 1, 9 (1st Cir. 2011) (“[A]ppeal from order denying request for relief from judgment does not resurrect appellants’ expired right to contest the merits of the underlying judgment, nor bring the judgment itself before us for review.” (citation and quotation omitted)); Chambers v. City of Fordyce , 508 F.3d 878, 881 (8th Cir. 2007) ('An appeal from the denial of a Rule 60(b) motion does not raise the underlying judgment for our review but only the question of whether the district court abused its discretion in ruling on the Rule 60(b) motion.' (citation and quotation omitted)); St. Fleur v. City of Fort Lauderdale , 149 F. App’x 849, 851-52 (11th Cir. 2005) (“The scope of an appeal of a ruling on a Rule 60(b) motion is narrow: the appeal addresses only the propriety of the denial or grant of Rule 60(b) relief and does not raise issues in the underlying judgment for review.' (citation modified)); Johnson v. Unknown Dellatifa , 357 F.3d 539, 543 (6th Cir. 2004) ('An appeal from an order denying a Rule 60(b) motion does not bring up for review the underlying judgment disposing of the complaint.'); Smith v. Mallick , 10 F. App'x 1, 1 (D.C. Cir. 2001) ('[A]n appeal from denial of Rule 60(b) relief does not bring up the underlying judgment for review.' (quoting Browder , 434 U.S. at 263 n.7)); S.E.C. v. McNulty , 137 F.3d 732, 741 (2d Cir. 1998) (“The appeal from the denial of a motion to vacate pursuant to Rule 60(b) brings up for review only the validity of that denial, not the merits of the underlying judgment itself.”); United States v. 31.63 Acres of Land , 840 F.2d 760, 761 (10th Cir. 1988) (“When a party appeals from an order denying a Rule 60(b) motion, the appeal ‘does not bring up Page 60the underlying judgment for review.'' (quoting Browder , 434 U.S. at 263 n.7)).

Here, the majority has taken the unprecedented step of departing from this long-standing rule, placing us at odds with every other circuit in the country and with the Supreme Court. Instead of limiting its review to the denial of Plaintiffs’ Rule 60(b) motion, the majority chose to review both the district court’s order denying Plaintiffs’ Rule 60(b) motion and the merits of the underlying judgment. This error cannot be overstated.

Recognizing the gravity of its error, the majority attempts to reframe its analysis, asserting that “the district court denied Rule 60(b) relief based, in part, on reasons set forth more fully in its decision granting summary judgment.” Maj. Op at 33 n.9. While clever, this reframing is a fallacy. Should a case arise where a district court truly relied upon a previous order, granting summary judgment or otherwise, to deny a party Rule 60(b) relief, this court should determine whether that previous order is sufficiently intertwined with the Rule 60(b) denial to make it reviewable. But that is not this case.

Here, the district court referenced its previous summary judgment order only when discussing the first and fifth factors of the Phelps v. Almeida , 569 F.3d 1120, 1134-40 (9th Cir. 2009) test used to analyze Rule 60(b) relief. See Bynoe v. Baca , 966 F.3d 972, 980-86 (9th Cir. 2020) (noting that the first factor is the nature of the intervening law, and the fifth factor is the relationship between the change in law and the challenged judgment). Specifically, the district court referenced its previous summary judgment order only to show that even if the ATRA was retroactive, Plaintiffs would not be entitled to relief because such retroactivity would only Page 61affect one of the three independent bases on which the district court rested its summary judgment order. Ratha v. Phatthana Seafood Co. , No. CV 16-4271-JFW, 2023 WL 2762044, at *4 (C.D. Cal. Mar. 3, 2023). This reference to the summary judgment order occurs once in a single paragraph of the district court’s Rule 60(b) order.⁸

Id. The reference was immediately followed by this analysis:

Although the change to Section 1595(a) might potentially apply to the third ground of the Court's decision granting summary judgment, the Court agrees with Rubicon that the Court’s decision is fully supported by the first and second grounds and the change to Section 1595(a) does not apply to or in any way affect the first and second grounds of the Court’s decision. As a result, the Court concludes that because the change to Section 1595(a) would not alter the Court’s decision in favor of Rubicon, the first and fifth factors weigh heavily against granting relief pursuant to Rule 60(b)(6).

Page 62Critically, the district court never adopted the reasoning of the summary judgment order as a basis for the denial of Rule 60(b) relief.

In reviewing the fifth factor regarding the relationship between the change in law and the challenged judgment, of course a district court will reference the underlying order related to the judgment. Simply referencing the underlying order, however, cannot be enough to pull that underlying judgment up for review on a Rule 60(b) appeal. But the majority's decision here says that it can. Following this decision, it is hard to see how a district court could meaningfully apply the Phelps factors without opening the door for appellate review of both the Rule 60(b) order and the underlying judgment.

We should have treated this appeal like every other appeal from a denial of a Rule 60(b) motion and confined our review to the district court’s order on that motion. Had the majority done so, it would have come to the inevitable conclusion that the district court did not abuse its discretion.

Here, the district court appropriately applied the Phelps factors to find that Plaintiffs are not entitled to relief from judgment. Indeed, the only error the majority identifies in the district court’s order denying Rule 60(b) relief is that the ATRA was retroactive. But we know from our discussion above that the ATRA is not, in fact, retroactive. Therefore, the district court did not err in so finding. Because the district court did not err in finding that the ATRA is not retroactive, it did not abuse its discretion in denying Plaintiffs' Rule 60(b) motion. Our review should have ended there.

But, even if the ATRA was retroactive, we cannot say that the district court abused its discretion in denying RulePage 63 60(b) relief. This is because, as noted above, the district court's order granting Rubicon's motion for summary judgment rested on three independent and sufficient grounds, only one of which related to the change made by the ATRA. In other words, even if the district court agreed that the ATRA applied to Plaintiffs’ claims, the underlying judgment would remain unchanged. Indeed, the district court noted this in its order denying Rule 60(b) relief, stating that even if the ATRA was retroactive “the change to Section 1595(a) would not alter the Court's decision in favor of Rubicon.” This determination supported the district court’s conclusion that the first and fifth Phelps factors weighed against granting Rule 60(b) relief. That is precisely the kind of analysis a district court should engage in when deciding a Rule 60(b) motion.

The majority, however, cannot accept that the district court did not abuse its discretion here, so it searches for error. Finding insufficient error in the district court’s Rule 60(b) order, the majority makes the extraordinary decision to review the underlying judgment, breaking a rule established by the Supreme Court over four decades ago, and departing from all twelve of our sister circuits. It is hard to make an error more striking than this. Ultimately, the district court did not abuse its discretion in denying Plaintiffs’ Rule 60(b) motion and the majority’s holding otherwise is error.

Because I cannot stand idly by while our court makes such errors, I respectfully dissent.