Appeal from the United States District Court for the Western District of Louisiana USDC No. 2:16-CV-146

Before Stewart, Graves, and Oldham, Circuit Judges .

James E. Graves, Jr., Circuit Judge :

After presiding over a bench trial by consent, U.S. Magistrate Judge Kathleen Kay awarded IFG Port Holdings almost $125 million. The losing side, the Lake Charles Harbor and Terminal District (the Port), later discovered that Judge Kay and IFG ’s lead trial counsel had been close family friends for nearly forty years. All Judge Kay had disclosed about the friendship was that IFG’s lead counsel’s daughter was her law clerk, who would be screened from the case. At best, this disclosure was incomplete.

Page 2The Port seeks to vacate the magistrate judge referral. Yet IFG claims that the Port's post-consent constructive knowledge of the friendship can sustain the referral. It cannot.

To waive Article III jurisdiction, a party must knowingly, voluntarily, and intelligently consent. We construe this standard consistently with other waivers of fundamental rights, such as pleading guilty or waiving a jury trial. A party must fully understand what the waiver connotes and its consequence. And we must indulge every reasonable presumption against waiver. Because a constructive knowledge standard would presume a waiver without full understanding, we reject it.

Hence the district court applied the right standard: For the Port's consent to be knowing, it must have actually known the nature and extent of Judge Kay's conflict before judgment. We AFFIRM .

This dispute concerns permits to dredge a ship channel in Lake Charles, Louisiana. IFG Port Holdings, L.L.C. v. Lake Charles Harbor & Terminal Dist. ( First Panel Op. ), 82 F.4th 402, 405–06 (5th Cir. 2023). IFG hoped to use the channel to develop a grain terminal. Id. So it leased the channel from the Port in 2011. Id. at 406. The lease never specified who would secure a permit for the necessary dredging. Id.

By 2015, the terminal was complete, but the dredging was not. Id. Without the dredging, IFG could not fully use the terminal. Id. In IFG ’s view, it was excused from performing under the lease because the Port was responsible for securing the permits. Id. The Port demanded that IFG perform, reasoning that the permits were IFG’s responsibility. Id.

IFG responded by suing in the Western District of Louisiana. Id. District Judge Patricia Minaldi took the case. In February 2017, the parties consented to a bench trial before Judge Kay. After a twenty-day trial, the Port Page 3lost, and Judge Kay awarded IFG nearly $125 million. The posttrial order was unusually harsh, aiming personal attacks at the Port and its lawyers.¹

After the posttrial order, the Port investigated Judge Kay and discovered her mostly undisclosed friendship with IFG’s lead counsel: William Monk. See IFG Port Holdings LLC v. Lake Charles Harbor & Terminal Dist. ( Ord. Vacating Referral ), 742 F. Supp. 3d 548, 562–65 (W.D. La. 2024), reconsideration denied , 2025 WL 1255155 (W.D. La. Feb. 6, 2025). When the Port moved to vacate Judge Kay’s referral, the district court denied the motion. See id.

The Port appealed. First Panel Op. , 82 F.4th at 408. A previous panel held that the Port’s allegations about Judge Kay’s failure to fully disclose her relationship with Monk might be grounds to vacate the referral. Id. at 414–17. But the panel remanded for further factual development. Id. at 417, 442–43. On remand, the district court held a three-day evidentiary hearing, found that the Port’s consent was not knowing, and vacated the referral. Ord. Vacating Referral , 742 F. Supp. 3d at 551-87.

IFG appeals the district court orders vacating Judge Kay's referral. Finding no abuse of discretion, we affirm.

A. We review a motion to vacate a referral for abuse of discretion.

Page 4When the parties consent, a magistrate judge may conduct proceedings and enter judgment in a civil lawsuit. 28 U.S.C. § 636(c)(1). Still, a district court may vacate the referral if the party shows extraordinary circumstances. Id. § 636(c)(4); Fed. R. Civ. P. 73(b)(3). Whether there are extraordinary circumstances depends on the eight Carter factors. See Carter v. Sea Land Servs., Inc. , 816 F.2d 1018, 1021 (5th Cir. 1987).² Only one is relevant here: “whether consent was voluntary and uncoerced.” See First Panel Op. ¸82 F.4th at 415–17.

We review the denial of a motion to vacate a magistrate judge referral for abuse of discretion. Id. at 414. A court abuses its discretion when it bases its decision 'on an erroneous view of the law or on a clearly erroneous assessment of the evidence.” Tollett v. City of Kemah , 285 F.3d 357, 363 (5th Cir. 2002) (citation modified).

We resolve IFG ’s factual challenges first. In the previous appeal, we charged the district court to resolve three essential issues: (1) How close was the Monk-Kay relationship? First Panel Op. , 82 F.4th at 416–17. (2) What did Judge Kay disclose about it? See id. And (3) when did the Port discover the relationship’s nature and extent? See id.

The district court met its charge. In a fifty-nine-page order, it found that: Judge Kay, Monk, and their families shared “a close, multifaceted, and personal friendship spanning nearly forty years.” Ord. Vacating Referral , 742Page 5 F. Supp. 3d at 578. Yet Judge Kay disclosed only that she employed Monk’s daughter as a law clerk. Id. at 579.The Port discovered the friendship years later, after Judge Kay’s post-trial order. Id . IFG challenges these findings, but we find no clear error.

Judge Kay is a Lake Charles native. Id. at 568 . After leaving for law school, she returned for a federal clerkship and soon began dating her future husband, Scott. Id. He introduced her to Monk in 1985, and Judge Kay quickly learned that she and Monk's wife (Aimee) were childhood acquaintances. Id. Over the next few years, the couples' relationship deepened. While Judge Kay clerked, they would sometimes gather with other young couples for social events, such as game nights or meals out. Id. at 569. In 1989, Judge Kay and Scott married and Monk was a groomsman. Id.

The couples stayed close. Id. at 569–72. Judge Kay and Scott joined the Monks for periodic happy hours. Id. at 569. Scott played golf with Monk, and sometimes offered him concert tickets. Id. The couples also got to know each other’s parents. Id. They always attended the same annual Christmas Eve party. Id. When children came, the families grew even closer. Their sons were in the same grade and became close friends Id. at 569–70. Judge Kay and Aimee once drove the boys to camp. Id. at 570. And in 2010, the Monks took Judge Kay’s son on a family vacation. Id.

The Monk-Kay friendship remained strong into the 2010s-even after the Port consented to the referral. Id. at 570–72. The families often had dinners together. Id. at 571-72. Judge Kay and Monk also frequently exchanged texts and Facebook messages. Id. at 570. They wished one another happy birthday, discussed concerts, and shared updates about their children and grandchildren. Id. Some of these messages “indicate[] a familiarity and Page 6lack of formality more at home in a personal friendship than in a professional relationship.” Id. For example, Monk texted Judge Kay in late-2017 to “describe[] a music festival lineup as ‘aimed more for post -summer, pre-fall relaxation than drunken debauchery.'' Id. Judge Kay replied, 'What's wrong with drunken debauchery??” Id.

In the 2010s, Judge Kay and the Monks shared major life events. Id. at 570–72 . In 2014, Judge Kay and Scott threw a party celebrating several family accomplishments, and Monk joined. Id. at 571. In 2015, Judge Kay officiated Monk's daughter Lucie's wedding. Id. Lucie later threw Judge Kay a surprise birthday party. Id. When Judge Kay’s daughter married in 2017, Monk went. Id. At this wedding, Scott’s band played, and Monk introduced it—Judge Kay drafted these remarks. Id.

IFG seeks to minimize this relationship. It relies on Judge Kay and Monk's attempts to distance themselves at the hearing. See id. at 572. But the district court did not find this credible, and we owe this finding 'great[] deference.” Id. ; see Kenai Ironclad Corp. v. CP Marine Servs., LLC , 84 F.4th 600, 605 (5th Cir. 2023) (citation modified). Besides, Judge Kay’s actions over the last forty years belie any claim she and Monk were not close. See Ord. Vacating Referral , 742 F. Supp. 3d at 569-72.

Despite all of this contact, Judge Kay’s disclosure was minimal. At a 2016 status conference, she disclosed only that Monk's other daughter, Margaret, was her law clerk and would be screened from the case. Id. at 553. Judge Kay disclosed nothing more about her friendship with Monk at the conference-or ever. Id. Relying on this incomplete disclosure, the Port consented to the referral in January 2017.

Page 7The Port discovered the extent of the friendship only after judgment. After the ruling, the Port investigated Judge Kay. Ord. Vacating Referral , 742 F. Supp. 3d at 562-65. Through Judge Kay's Facebook page, the Port attorneys discovered photos of Monk at Judge Kay’s wedding, and Judge Kay officiating Lucie Monk’s wedding. Id. at 564. These photos formed the basis of the Port’s initial motion to vacate. Id. at 568.

IFG claims that the Port’s lawyers knew the extent of the Monk-Kay friendship earlier. But there is no compelling evidence that they did. Judge Kay had cordial, professional relationships with the Port’s lawyers, such as Rick Norman and Michael Dees, and the lawyers sporadically interacted with Judge Kay’s Facebook. Id. at 572–78. These contacts are not conclusive that the Port lawyers learned the extent of the Monk-Kay friendship. See id. The district court found that this evidence was simply insufficient to show the Port lawyers’ actual knowledge of the friendship before judgment. In this, we find no clear error.

So, we reject IFG ’s factual challenges. Judge Kay’s disclosure was incomplete, at best, and perhaps misleading. The partial disclosure implied that Monk's daughter's clerkship was the extent of their ties, when the friendship was far closer. Relying on this, the district court found that the Port “lacked crucial information” about the Monk-Kay relationship, so its consent was “not knowing[] and intelligent[].” Ord. Vacating Referral , 742 F. Supp. 3d at 585.

IFG responds that the Port's knowledge of the relationship before judgment forecloses its attempt to vacate the referral. As the previous panel explained, 'consent to a magistrate-judge referral can be inferred from a party’s conduct during litigation.” First Panel Op. , 82 F.4th at 417 (quoting Roell v. Withrow , 538 U.S. 580, 582 (2003)). So 'if the Port learned of the Page 8intimacy of the friendship before . . . judgment and chose to remain silent, then its consent would remain valid.” Id.

We have already rejected IFG ’s argument that the Port, in fact, knew the extent of the Monk-Kay friendship before judgment. But IFG urged the district court, and now us, to adopt a constructive knowledge standard. Ord. Vacating Referral , 742 F. Supp. 3d at 583. Under it, the Port lawyers' constructive knowledge of the Monk-Kay friendship could imply their ongoing consent to the referral. The district court rejected this standard, and so do we.

right.

Consent to magistrate judge jurisdiction waives the right to Article III adjudication, so consent is 'a constitutional imperative.' See First Panel Op. , 82 F.4th at 415. The standard to waive this right matches 'the standard for . . . other waivers of rights.' Carter , 816 F.2d at 1021. Carter drew the consent factor from the guilty-plea context. See id. (citing United States v. Carr , 740 F.2d 339 (5th Cir. 1984)). Like a guilty plea, consent to a referral must be knowing, intelligent, and voluntary. First Panel Op. , 82 F.4th at 416; see Grabowski v. Jackson Cnty. Pub. Defs. Off. , 47 F.3d 1386, 1389 (5th Cir. 1995).

Constructive knowledge does not suffice in the guilty plea context, and it does not here. To plead guilty, a defendant must fully understand 'what the [waiver] connotes and . . . its consequence.' United States v. Hernandez , 234 F.3d 252, 254–55 (5th Cir. 2000). Hence a district court must “inform the defendant accurately of the proper minimum sentence that will result from the plea.” United States v. Carreon-Ibarra , 673 F.3d 358, 364 (5th Cir. 2012) (citation modified). This ensures that the defendant understands the plea's 'direct consequences.' See id. Analogously, to knowingly consent Page 9to a referral, a party must actually know the extent of a magistrate judge’s conflicts. Otherwise, they cannot fully understand the waiver’s consequence.

Carter also drew from the jury-trial waiver context. See Carter , 816 F.2d at 1021 (citing Cox v. C.H. Masland & Sons, Inc. , 607 F.2d 138 (5th Cir. 1979)). Waiver of the right to a civil jury trial “may also be inferred from a party’s conduct.” McDonald v. Steward , 132 F.3d 225, 229 (5th Cir. 1998) (citation modified). Yet “any seeming curtailment of the right . . . should be scrutinized with the utmost care.' Id. 'Courts should indulge every reasonable presumption against waiver.” McAfee v. Martin , 63 F.3d 436, 437 (5th Cir. 1995). So waiver “should not be found in a ‘doubtful situation.’” Id. (quoting Aetna Ins. Co. v. Kennedy , 301 U.S. 389, 393 (1937)). A constructive knowledge standard inverts this principle: it presumes waiver in a doubtful situation.

By illustration, we have declined to infer consent to waive a jury trial when a plaintiff had constructive knowledge that he was participating in a bench trial. See McAfee , 63 F.3d at 437-38. In McAfee , a magistrate judge found that an unrepresented plaintiff had failed to carry his ultimate burden of proof after “an evidentiary hearing” that was effectively a bench trial. Id. The judge “did not use the term trial until well into the proceeding and then without explaining the difference between a trial and an evidentiary hearing.” Id. (citation modified and italics added). Surely, the plaintiff constructively knew the difference between a trial and a hearing, and that he was participating in a bench trial. Yet our court declined to imply a waiver through his conduct in this “doubtful” situation. Id.

This case presents another doubtful situation. IFG claims that certain contacts between Judge Kay and the Port’s lawyers—while not sufficient to show actual knowledge of the Monk-Kay friendship-were sufficient to trigger a duty to investigate the relationship, thereby implying their client’s Page 10consent to the referral. This presumes that the Port's inaction waived its consent, and illustrates precisely the problem with a constructive knowledge standard: It presumes a waiver of a fundamental right. For this reason too, we reject it.

What is more, 'litigants and counsel should be able to rely upon judges to comply with their own Canons of Ethics.” Am. Textile Mfrs. Inst., Inc. v. The Ltd., Inc. , 190 F.3d 729, 742 (6th Cir. 1999) (quoting Porter v. Singletary , 49 F.3d 1483, 1489 (11th Cir. 1995)). It was Judge Kay's duty to inform the Port “about a friendship with a lawyer . . . that [she] believe[d] the parties or their lawyers might reasonably consider relevant to a possible motion for disqualification.” First Panel Op. , 82 F.4th at 418 n.9 (quoting ABA Standing Comm. on Ethics & Pro. Resp., Formal Op. 488, Judges’ Social or Close Personal Relationships with Lawyers or Parties as Grounds for Disqualification or Disclosure (2019)). She failed to do so. Without good reason, we will not adopt a standard that requires a party to “pore through [a] judge’s private affairs,” to prevent a waiver of a constitutional right. Am. Textile Mfs. , 190 F.3d at 742.

IFG attempts to analogize superficially similar contexts that use a constructive knowledge standard. But the analogies fail because the contexts concern forfeitures of a statutory right, not waivers of a constitutional right.

Take judicial recusal, for example. IFG relies on In re Kerns , where a litigant argued that a bankruptcy judge should have recused because 'he served as [a] lawyer in the matter” in “private practice.” Kerns v. First State Bank of Ben Wheeler ( In re Kerns ), 130 F.4th 455, 463 (5th Cir. 2025) (quoting 28 U.S.C. § 455(b)(2)). Kerns had constructive knowledge that the judge served as a lawyer in the matter previously. Id. at 463-64. Because Kerns Page 11should have known about the statutory recusal ground and failed to raise it, he forfeited it. Id.

Of course, recusal has a constitutional dimension: The Due Process Clause guarantees an impartial judge. See, e.g. , Republican Party of Minn. v. White , 536 U.S. 765, 775-76 (2002). But 'not all questions of judicial qualification involve constitutional validity.' Aetna Life Ins. Co. v. Lavoie , 475 U.S. 813, 820 (1986) (citation modified). '[M]atters of kinship, personal bias . . . [and] remoteness of interest [are] generally' left to 'legislative discretion.” Id.

Thus, even if a judge is disqualified under federal law, the Constitution might not mandate recusal.³ See id. at 821. Kerns claimed a statutory-not a constitutional-right to recusal, so his protection was diminished. See In re Kerns , 130 F.4th at 463–64. The Port’s right to consent to a referral is constitutional. See Carter , 816 F.2d at 1021 . Naturally, this right enjoys greater protection.

Juror disqualification and arbitration make worse analogies. Whether trial counsel had a duty to investigate with “reasonable diligence” a juror’s potential bias during voir dire has little relevance to a duty to investigate a magistrate judge’s potential bias after consent. Cf. United States v. Wilson , 116 F.3d 1066, 1086 (5th Cir. 1997), vacated in part on other grounds sub nom. United States v. Brown , 161 F.3d 256 (5th Cir. 1998). Nor are the statutory grounds to vacate an arbitration award analogous to the constitutional right to Page 12Article III adjudication. Cf. Light-Age, Inc. v. Ashcroft-Smith , 922 F.3d 320, 322–23 (5th Cir. 2019).

Besides, Carter provides clear guidance on analogous contexts. They must involve a waiver of another fundamental right. See Carter , 816 F.2d at 1021. It gave examples: guilty pleas, jury-trial waivers, and the right to be represented by counsel of choice. See id. (citation modified). We should be wary of drawing parallels other contexts, and do so only when they concern an analogous fundamental right.

Thus, we reject IFG ’s constructive knowledge standard. For consent to be knowing, a party must actually know the nature and extent of a magistrate judge’s potential conflict. Because the district court applied this standard, we see no error. See First Panel Op. , 742 F. Supp. 3d at 572–79; IFG Port Holdings LLC v. Lake Charles Harbor & Terminal Dist. ( Reconsideration Ord. ), No. 2:16-CV-00146, 2025 WL 1255155, at *12-14 (W.D. La. Feb. 6, 2025).⁴

* * *

Overwhelming evidence shows a close family friendship between Judge Kay and Monk. And vanishingly little direct or circumstantial evidence suggests the Port actually knew about the friendship before judgment. The district court applied the right knowledge standard to these facts: Only the Port’s actual knowledge of the nature and extent of the Monk-Kay friendship before judgment is sufficient to show their knowing consent.

Page 13While vacation of a referral is an extraordinary remedy, this case warrants it. Judge Kay’s actions raise serious concerns about the fairness of her proceedings. Her unusually harsh order, coupled with the windfall verdict, exacerbates this impression. We AFFIRM .

Page 14Andrew S. Oldham , Circuit Judge , dissenting:

The Lake Charles Port lost a trial worth more than $120 million. The Port's lawyers were understandably upset. What's less understandable is what they chose to do about it.

The Port’s lawyers attempted to vacate the judgment based on the fact that opposing counsel’s daughter clerked for the magistrate judge who tried the case. Everyone-including the Port's lawyers-knew this fact before trial. The law clerk was recused from the matter. And the Port consented to the magistrate judge trying the case anyway.

Still, the point of the Hail Mary pass is that it sometimes works. And for the Port, that sometime is today.

I respectfully dissent. The majority's legal rule conflicts with the statutory text and Supreme Court precedent. It also contradicts circuit authority, encourages parties to scrutinize judges' private affairs, and jeopardizes virtually all magistrate and bankruptcy judge orders.

This story has three parts. First, a contractual dispute. Second, a trial. Third, a years-long—and now successful—scheme by the loser of that trial to undo a large, adverse judgment by Facebook-stalking a jurist.

Between Beaumont and Lafayette lies a port formally known as the Lake Charles Harbor & Terminal District (“the Port”). IFG Port Holdings, L.L.C. v. Lake Charles Harbor & Terminal Dist. ( IFG I ) , 82 F.4th 402, 405– 06 (5th Cir. 2023). This case began when a port tenant, IFG Port Holdings LLC (“ IFG ”), and the Port began talks about a new project. In brief, IFG wanted to build an export grain terminal. Ibid . This project would require IFG to build a new facility on a relevant berth and the Port to dredge the Page 15waterway so that it was deep enough for the ships that would load and unload their cargo at the new facility. The Port, for its part, 'agree[d] to take reasonable steps to work with” the U.S. Army Corps of Engineers to deepen the waterway approaching the relevant berth. ROA.34345. And IFG agreed to construct the grain-export facility. ROA.21709–17. IFG held up its end of the bargain. ROA.21723.

But the Port did not get the permits from the U.S. Army Corps of Engineers that were necessary to begin dredging. No dredging occurred by the time that the facility finished construction in 2015, so it was unusable. ROA.21723, 21749, 21756. After much back and forth, IFG sought relief in federal court. It alleged that the Port had breached the relevant contract and violated the Louisiana Unfair Trade Practices Act. ROA.118–50.

IFG retained William Monk to represent it. The Port retained William J. “Joe” Mize and Matt Mize, along with Merrick “Rick” Norman. ROA.28850; ROA.28853–54.

The case was originally placed on the docket of United States District Judge Patricia Minaldi and Magistrate Judge Kathleen Kay. ROA.28850. The parties consented to have Magistrate Judge Kay try the case. ROA.28847, 28852.

The attorneys and Judge Kay were all familiar. Norman's son had done an externship in Judge Kay's chambers. ROA.28883. Norman had chaired the panel that recommended Judge Kay's appointment (and her reappointment). ROA.28854. He “practiced law with her in the same firm for a while.' Ibid. And Monk's daughter Margaret Ellen was clerking for Judge Kay—a fact that both Joe and Matt Mize knew. ROA.45172–73. Judge Kay also informed the parties of that fact and said she would screen Margaret Page 16from the case. ROA.28851. It is thus undisputed that the Port’s lawyers had detailed knowledge about Judge Kay and this potential conflict.

Judge Kay then held a twenty-day bench trial. IFG I , 82 F.4th at 406. During the trial, the Port's counsel were concerned about an adverse outcome. They were so concerned, in fact, that their lawyers exchanged a series of increasingly desperate emails. Norman emailed, “[w]e don’t know the result yet but all signs point to a disaster.” ROA.28858. Joe Mize also got in on the action, writing about Judge Kay that '[t]his woman is insane.' ROA.28859. Norman wrote that “[i]f Kathy screws us, we may want to take the gloves off and ask for a new trial.” ROA.28858. Matt Mize then asked if they could get a new trial, and “if we moved for and we’re granted a motion for new trial could we revoke consent to have Kay try to [ sic ] the case??' ROA.28859.

The Port lost. Judge Kay issued a lengthy opinion finding that the Port was liable. ROA.21702–65. And Judge Kay determined that IFG was entitled to treble damages, resulting in a damage award of $124,531,652. IFG I , 82 F.4th at 407.

The Port's lawyers were upset. Norman compared the ruling to 'losing the Super Bowl by eight touchdowns.' ROA.28862. Matt Mize texted his wife that it was “the worst case scenario.” ROA.28861.

Norman, 'grasping at straws,' emailed Matt Mize about filing a motion to recuse based on Margaret’s work as a clerk. ROA.28863. The idea was to 'tak[e] the case away from Judge Kay before she could award damages.” Ibid. At this point, Norman also reached out to a local attorney, Mike Vernon, for advice about how this scheme would work. ROA.45286. Norman asked whether Margaret Monk’s employment would “help us at the 5th Circuit.” Ibid. Vernon said that “if the port’s lawyers consented to trial Page 17before Kathy knowing that Bill’s daughter worked for her, they are going to have a problem complaining about it down the line.' Ibid. But Vernon advised, “[s]till, I would mention it quite a bit. The feds don’t like that kind of stink.” ROA.45286. The upshot was clear: the alleged conflict was legally irrelevant, but the Port’s lawyers thought it might “help [them] at the 5th Circuit” anyway. Ibid .

Joe and Matt Mize tried the same strategy—employing their wives, Maura and Leslie. Maura searched Judge Kay’s public Facebook page and came across a photo of Judge Kay's wedding in 1989. ROA.28691; ROA.28864–65. Monk was in some of those photos. ROA.28864–65. Leslie found similar photos showing Monk in what she thought was groomsman’s attire, and she also saw a photo of Judge Kay officiating the wedding of Monk’s other daughter. ROA.28866–67.

But this was all small potatoes. Norman himself said in response to the photos, “I’ve read some law review articles. I don’t see grounds for recusal.” ROA.28865. So Norman turned to his wife Polly. Polly was no help, saying that 'there is no close friendship' between Monk and Judge Kay. ROA.28865; ROA.45273. Recognizing this weakness, the dominant theme among the Port’s legal team was that the grounds for appeal would be a “bias issue based on Judge Kay's tone in the ruling'-not any alleged conflict. ROA.28866.

That was not good enough for the Port. So the Port retained yet another lawyer, Michael McKay. McKay picked up where Norman and the Mizes had left off. He informed the “Port Board that he and his team were looking into the relationship between Judge Kay and Monk.” ROA.28869.

Then he also struggled to find anything material. With no firsthand witnesses to the relationship between Monk and Judge Kay, McKay did not have a single person who was willing to sign an affidavit about the putative Page 18misconduct based on that supposed relationship. ROA.28873. Struggling, McKay’s team had Matt Mize sign an affidavit based on “[s]creenshots of the Facebook content located by Maura Mize and Leslie Mize.” Ibid. This material was undeniably skimpy, and the Port knew it. So their motion to vacate hinged on a single question: was it enough “stink” for the “feds” to take issue with the judgment? ROA.45286.

The answer was no, at least initially. Chief Judge Terry Doughty denied the Port’s motion, noting that “some prior social and/or professional interactions” were coin-of-the-realm in “relatively small legal communities” like Lake Charles, Louisiana. ROA.23153. When the Port appealed, this court reversed. Our court explained that the Port's allegations 'remain[ed] untested,” and we directed the district court to determine the extent of the relationship between Judge Kay and Monk. IFG I , 82 F.4th at 412, 417-21.

On remand, the district court set out a new legal standard for what the parties had to show. The court stated that 'only actual knowledge of a relationship suffices for knowing and voluntary consent to a magistrate judge.” ROA.28897. That standard meant that IFG had to show the Port had actually known about the alleged conflict of interest. What’s more, the court also demanded “direct evidence” of actual knowledge of the potential conflict—apparently setting aside any circumstantial evidence or inference in proving knowledge. ROA.29115.

The court also directed discovery. The short summary of that discovery is that lawyers for both sides had attended various social events over the course of several decades, including weddings and parties, with Judge Kay. (The long version is discussed below.) But two events were most relevant: (1) Monk served as a substitute groomsman in Judge Kay’s 1987 wedding, and (2) Judge Kay officiated the wedding of Monk's daughter, which the Normans also attended. ROA.28875–76, 28878.

Page 19IFG The Port

The Monks Here is a summary of the relevant conduct: would attend the The Normans would attend the

wedding of Judge Kay's daughter. wedding of Judge Kay's daughter.

ROA.28879; ROA.33458. ROA.28884; ROA.33458.

Notably, the Monks and Normans Notably, the Monks and Normans

would apparently walk into the would apparently walk into the

event together. ROA.33457–58. event together. ROA.33457–58.

Monk served as a substitute groomsman for Scott McPherson, who married Kay. ROA.28875. Kay and McPherson attended Monk's 1987 birthday party. Ibid.

Judge Kay apparently gossiped with Norman about Monk’s desire to be a federal judge, noting she did not “think [he] would have been thrilled at the prospect of Mr. Monk being the Article III judge.' ROA.33345. Over decades, “[t]exts between Judge Kay and Monk, which there are admittedly only a few of , show that Judge Kay wished Monk happy birthday, asked Monk for someone’s contact information, and reminisced about a case Monk had tried in front of the judge she clerked for.” ROA.28876 (emphasis added).

Judge Kay had interacted with the Port's General Counsel, Michael Dees, at LSU events. ROA.28851; ROA.28886. The two enjoyed a “friendly, professional relationship,” ROA.28887, and would exchange emails, including when Judge Kay asked Dees via Facebook for his daughter-in-law's contact information. Ibid.

Page 20IFG The Port Judge Kay officiated Lucie Monk’s wedding. ROA.28878. Joe Mize, Maura Mize, as well as Norman and his wife received invitations to the wedding. ROA.28889, 33455-56. Norman's wife co-hosted a pre-wedding event and went to the wedding. ROA.33455–56. Polly Norman would state that she “knew from the very beginning' that Judge Kay would officiate the wedding. ROA.34041–42. Judge Kay, her husband, Monk, and Monk's wife would play Trivial Pursuit in a “bigger group of young couples” during the 1980s. ROA.28874–75. Maura Mize liked photos on Facebook, including one showing Judge Kay as an officiant at Lucie Monk’s wedding. ROA.28889.

The district court then asked whether the Port actually knew about the supposed relationship between Judge Kay and Monk. ROA.28890; 28886– 88. The district court looked for 'direct evidence' that the Port had knowledge of any relationship between Monk and Judge Kay, apparently excluding any inferences to that effect. For example, the court discounted that the Port's general counsel, Michael Dees, commented on Judge Kay's photos, including close in time to when she changed her profile photo to an image with Monk. ROA. 20109–10, 29113. What’s more, Dees had even liked a post that had photos showing Monk in Judge Kay’s wedding. ROA.29104. But the district court said that it would “apply the actual knowledge standard to the new evidence, not the forgiving, inference-laden standard that IFG so adamantly prefers.' ROA.29114. The district court would accept neither inferences nor constructive knowledge.

Under this heightened standard, the district court held that the Port did not have actual knowledge of the conflict. True, there was evidence the Page 21Port’s lawyers saw photos of Monk and Judge Kay, including at the wedding in question. ROA.29104. And true, the Port’s lawyers were invited to—and Norman attended—Lucy Monk’s wedding. ROA.33455–56. And true, the Port’s own emails showed the Port thought it was engaged in a meritless lastditch attempt to overturn the judgment. ROA.28865 (“I’ve read some law review articles. I don’t see grounds for recusal.”). Nonetheless, the district court dismissed all of this as circumstantial evidence of knowledge- not direct evidence. So it vacated the prior judgment.

To sum up, even though there were conflicts on both sides, and even though the natural inference is that all the lawyers knew about any possible conflict, and even though there was plenty of circumstantial evidence to support that inference, IFG lost. IFG timely appealed.

As a matter of plain text, this is a simple case. Title 28 provides that a court “may, for good cause shown on its own motion, or under extraordinary circumstances shown by any party, vacate a reference of a civil matter to a magistrate judge under this subsection.' 28 U.S.C. § 636(c)(4); see also Fed. R. Civ. P. 73(b)(3). Thus, to vacate the judgment, the Port had to prove that this case involves “extraordinary circumstances.”

There are two reasons that this case does not involve the requisite extraordinary circumstances. First, this case involves commonplace relationships between bench and bar. Those relationships do not constitute grounds for vacatur. Second, the relevant statute’s reference to “extraordinary circumstances” requires parties to exercise diligence and act in a “faultless” way to get vacatur. See BLOM Bank SAL v. Honickman , 605 U.S. 204, 212 (2025) (quotation omitted). Such a requirement necessarily excludes the “actual knowledge” standard proffered by the Port and adopted by the district court.

Start with the text. Bartenwerfer v. Buckley , 598 U.S. 69, 74 (2023). The phrase “extraordinary” means “[b]eyond or out of the common order or method; not usual, customary, regular, or ordinary.” Webster’s New International Dictionary 903 (2d ed. 1934; 1950) (' Webster’s Second ”).

Nothing about this case is extraordinary or out of the common order. Consider the 'worst' fact: IFG 's lawyer (Monk) served as a substitute groomsman in Judge Kay's wedding thirty years ago. ROA. 28875. It is difficult to understand how attending a long-ago wedding is an extraordinary event. As the district court in this case originally found, such participation would not require recusal or even disclosure to the parties. ROA.23153. And its potential for creating a conflict of interest is particularly weak here: There is evidence in the record that Judge Kay did not even like Monk and gossiped about Monk's desire to become a federal judge with the Port's lawyer. ROA.28877 n.78; ROA.33345–46.

At the very most, the Port has shown that Monk and Judge Kay occasionally attended the same social events, most of which were decades ago, and had a reasonably warm professional relationship that sometimes included social activities or text messages. As I see it, attending a few group meals and social events, exchanging friendly text messages, and the like over the course of a few decades is ordinary—not extraordinary. That notion is supported by the fact that lawyers for both sides had various personal relationships with Judge Kay. If the Monk-Kay relationship was improper, so was the Norman-Kay relationship. And so was the Dees-Kay relationship. Rather than accept the absurd result that the entire Lake Charles Bar somehow violated their oaths to the court, we should acknowledge reality.

Page 23Members of a small legal community all knew each other, recommended each other for jobs, and occasionally crossed paths at social events.

At bottom, ordinary relationships by definition cannot be the “extraordinary circumstances” required to vacate a magistrate order. That simple principle should resolve this case.

Perhaps attempting to obscure this commonsense result, the Port says they needed to have 'actual knowledge' of any alleged conflict between Judge Kay and Monk to consent to trial. Supposedly, the lack of that actual knowledge is an extraordinary circumstance mandating vacatur. But this standard is not consistent with the statutory text, either.

In legal contexts, the term “extraordinary” designates “[t]he degree of care, caution, diligence, or the like, which is exercised by very or extremely prudent and thoughtful persons, exceeding the care exercised by persons of ordinary prudence . . . .” Webster’s Second , supra , at 903. It means “extra; additional” action. Ibid. The Supreme Court has defined “extraordinary” circumstances in much the same way: It has said that where 'the basis for relief in [a] case is extraordinary,' it is '[o]f particular importance” that a party did not show “neglect or lack of due diligence.” Liljeberg v. Health Servs. Acquisition Corp. , 486 U.S. 847, 863 n.11 (1988).

Take for example the “extraordinary circumstances” standard that has long governed Rule 60(b)(6) motions. See Ackermann v. United States , 340 U.S. 193, 199–200 (1950); BLOM Bank , 605 U.S. at 212. In discussing that requirement, the Court has 'underscored the stringency of the ‘extraordinary circumstances’ test.” BLOM Bank , 605 U.S. at 212. It has said that a movant must be “faultless,” ibid. (quotation omitted), and that a litigant who does not discover facts it should have known is not “completely without fault for his or her predicament.” Ibid. (quoting 12 J. Moore et Page 24al., Moore's Federal Practice § 60.48[3][b], p. 60-188 (3d ed. 2024)).

The upshot is clear: Whether under Rule 60(b)(6) or under § 636(c)(4), the party seeking to vacate a magistrate judge award must be faultless and must have exercised more care than even a reasonable party.

An actual knowledge standard is incompatible with both requirements. Under the actual knowledge standard, a party could simply do nothing , sit on significant circumstantial evidence, or ignore attempts to provide direct evidence. They could exercise no diligence or care at all and still win. That no-care-at-all approach cannot fit the textual definition of extraordinary circumstances.

The Port cannot meet this extraordinary-circumstance standard here either. Indeed, the Port’s conduct was either ( 1 ) malicious, as it deliberately cooked up a scheme to discredit a magistrate judge in order to avoid paying treble damages after judgment, ROA.45286 (explaining “the feds don’t like that kind of stink,' so the lawyer should 'mention [Margaret's clerkship with Judge Kay] quite a bit” in the hopes of winning at the Fifth Circuit), or (2) so plainly incompetent that their General Counsel and three trial lawyers (Norman and the Mizes) missed probative evidence that their wives found on Facebook. If the former, the Port is the opposite of faultless. If the latter, the Port’s directors should pursue a malpractice claim. But neither scenario fits the textual definition of “extraordinary circumstances” that justifies relief. So the Port’s “actual knowledge” standard fails as a matter of the statute’s plain text.

If this court were to look beyond the statutory text, there are at least three reasons to reject the Port's 'actual knowledge' standard. First, precedent. Second, incentives. And third, results.

If the plain text and recent Supreme Court precedent are not enough, our own precedent also points to a constructive knowledge standard. That result is unsurprising, since an actual knowledge standard contravenes the sweep of American law.

Under circuit precedent, a party can vacate a magistrate award if its “consent was obtained involuntarily.” Carter v. Sea Land Servs., Inc. , 816 F.2d 1018, 1021 (5th Cir. 1987). And that determination requires parsing an eight-factor balancing test. Ibid. 'Among the things a court may consider' are: ( 1 ) undue delay; ( 2 ) inconvenience to the court and witnesses; ( 3 ) prejudice to the parties; ( 4 ) whether the movant is acting pro se ; ( 5 ) whether consent was voluntary and uncoerced; ( 6 ) whether the motion is made in good faith or is dilatory and contrived; ( 7 ) the possibility of bias or prejudice on the part of the magistrate; and ( 8 ) whether the interests of justice would best be served by holding a party to its consent. Ibid. Carter derived these factors from various areas of law: criminal law, see United States v. Carr , 740 F.2d 339 (5th Cir. 1984), habeas corpus, see Gandy v. Alabama , 569 F.2d 1318 (5th Cir. 1978), tax law, see United States v. Lochamy , 724 F.2d 494 (5th Cir. 1984), secured transactions and credit agreements, United States v. Unum, Inc. , 658 F.2d 300 (5th Cir. Unit A Oct. 1981); Chanofsky v. Chase Manhattan Corp. , 530 F.2d 470 (2d Cir. 1976), and so on. See Carter , 816 F.2d at 1021.

Looking to these same legal areas, an actual knowledge standard makes even less sense. For example, in the criminal law context from which Carter draws, constructive knowledge is sufficient. In United States v. Dobey , for example, the Tenth Circuit held that the defendants' consent to trial before a magistrate was voluntary even though they claimed not to know about the magistrate’s non-Article III status. 751 F.2d 1140, 1142–43 (10th Cir. 1985). Similarly, in the habeas context, a pro se petitioner's decision to Page 26proceed before a magistrate judge was knowing because the petitioner should have known he had a right to an Article III adjudicator. Norris v. Schotten , 146 F.3d 314, 326 (6th Cir. 1998).

Another apt comparator is the judicial recusal context. In the Fifth Circuit, a party seeking a judge's recusal must move for recusal when he “knows or should know” of the facts on which he bases the request. Kerns v. First State Bank of Ben Wheeler (In re Kearns) , 130 F.4th 455, 464 (5th Cir. 2025) (quoting Health Servs. Acquisition Corp. v. Liljeberg , 796 F.2d 796, 802 (5th Cir. 1986)). In Kearns , this circuit said a lack of diligence can prove fatal to recusal motions-which obviously implies a constructive knowledge standard. Ibid. ; see United States v. Whorley , 550 F.3d 326, 339 (4th Cir. 2008).

The juror disqualification context is much the same. Parties cannot bring a juror disqualification challenge if the offending information could have been discovered before the verdict through reasonable diligence. United States v. Wilson , 116 F.3d 1066, 1086-87 (5th Cir. 1997), vacated in part on other grounds sub nom. , United States v. Brown , 161 F.3d 256, 257 n.1 (5th Cir. 1998) (en banc).

Same with arbitration. Our court has recognized that parties may not object to an arbitrator’s supposed conflict of interest if they have “constructive knowledge of the conflict at the time of the arbitration hearing but fails to object.” Light-Age, Inc. v. Ashcroft-Smith , 922 F.3d 320, 322 (5th Cir. 2019) (per curiam). All told, in the judicial recusal, juror disqualification, and arbitration arenas, constructive knowledge is sufficient.

Carter is far from perfect.¹ But if this court is going to apply such tests, they must be applied consistently. And here, a wide range of contextsPage 27 (identical to those in Carter ) ground consent in a constructive knowledge standard.

This focus on constructive knowledge makes sense. After all, constructive knowledge is merely a type of knowledge: “[k]nowledge that one using reasonable care or diligence should have, and therefore that is attributed by law to a given person.” Constructive Knowledge , Black’s Law Dictionary (12th ed. 2024); Intel Corp. Inv. Pol’y Comm. v. Sulyma , 589 U.S. 178, 185 (2020) (similar); Herbst v. Scott , 42 F.3d 902, 906 (5th Cir. 1995) (litigant who “should have known” about his claims had “constructive knowledge” of them).

At bottom, this court’s precedent looks to disparate areas of law to determine whether a party can vacate a magistrate award for lack of consent. And those areas of law employ a constructive knowledge standard. In my view, such authorities demonstrate that the default posture should be constructive—not actual—knowledge.

One reason that courts employ constructive knowledge is that it offers the only sensible set of incentives.

The constructive notice standard places the burden of litigation on the parties ex ante . The parties are supposed to be active participants in litigation, and there should be a penalty for doing nothing. By placing the onus on parties, the parties are discouraged from doing what the Port did here: last- Page 28minute, hail-Mary recusal motions that were either disingenuous attempts to overturn judgments or confessions of professional incompetence so embarrassing that they could be remedied only by wives with Facebook accounts. See Delesdernier v. Porterie , 666 F.2d 116, 121 (5th Cir. 1982).

In contrast, an actual-knowledge rule would misallocate resources and create obvious, perverse incentives. First, a magistrate or bankruptcy judge would have to disclose any and all social relationships dating back decades. Even humdrum connections could void years of hard-fought, expensive, and honest litigation. Second, losing parties would be incentivized to conduct the sort-of “investigation” of jurists that occurred here. Attendance at a single event, decades ago, could result in overturned judicial proceedings. Old weddings? Old friends? Old board game partners? All grounds for the sore loser to complain-and win. And third, the actual knowledge standard entitles losers to disingenuously say they lacked knowledge of any conflict. And they can do so in the face of strong circumstantial and inferential knowledge without sacrificing anything.

Usually, the story of the ostrich burying its head in the sand is a cautionary tale. But it turns up gold under the actual knowledge standard.

Even worse, such incentives could have serious consequences for magistrate and bankruptcy adjudications nationwide. Both magistrate and bankruptcy judges draw their power to adjudicate disputes in part from party consent. So a decision empowering parties to retroactively revoke that consent based on spurious legal “conflicts” undermines the system. See 28 U.S.C. § 636(c); Fed. R. Civ. P. 73(b); 28 U.S.C. § 157(c)(2).

Recall the facts. The Port's lawyers repeatedly searched for a legitimate, on-the-merits way out of this judgment. And after exhausting all Page 29legally meritorious options, they were able to overturn a $112 million judgment based primarily on wedding photos from the 1980s.

As amici explain, such gamesmanship could be the rule in future cases, not the exception. That’s because magistrate judges are “by virtue of their selection process, drawn from the legal communities in which they serve.” Br. of Amicus Curiae Former Magistrate & Dist. Ct. Judges at 17. They are often “selected after years of public service” and, as here, “likely to have connections to many parties and litigants.' Ibid. Such backgrounds are “especially prevalent in small legal communities like Lake Charles,” but “far from exclusive to those settings.” Ibid. So too with bankruptcy courts: after all, there is “tremendous local variation in the practice of bankruptcy.” St. Angelo v. Victoria Farms, Inc. , 38 F.3d 1525, 1536 (9th Cir. 1994) ( Poole, J. , concurring in part and dissenting in part), amended , 46 F.3d 969 (9th Cir. 1995).

At a minimum, the majority’s actual knowledge standard means that magistrate and bankruptcy judges might face an evidentiary hearing like the one Judge Kay suffered in this case. And on the merits, it could lead to crass gamesmanship in which parties strategically delay their concerns until they lose and then use those purported concerns to overturn lengthy, expensive, and fully litigated trials.

Finally, three specific responses to the majority.

First, it is black letter law that 'most matters relating to judicial disqualification [do] not rise to a constitutional level.' Caperton v. A.T. Massey Coal Co. , 556 U.S. 868, 876 (2009) (quoting FTC v. Cement Inst. , 333 U.S. 683, 702 (1948)). And the Port explicitly conceded that its challenge “is not based on any intentional acts of fraud, bias, or undue influence by the Page 30Magistrate Judge.' ROA.22916. If that is true, the Port's constitutional arguments are meritless.

Second, and with deep respect to the majority, even if I were to accept the majority’s constitutional framing of this case, I remain unconvinced by the majority’s analogy to jury trials. That’s because parties can expressly or impliedly waive their rights to a jury. Pizza Hut L.L.C. v. Pandya , 79 F.4th 535, 540 (5th Cir. 2023). There is broad agreement amongst the circuits on this score. See, e.g. , Smith Eng’g Co. v. Pray , 61 F.2d 687, 693 (9th Cir. 1932); Badger Meter, Inc. v. Grinnell Corp. , 13 F.3d 1145, 1159 n.7 (7th Cir. 1994); Waldman v. Stone , 698 F.3d 910, 920 (6th Cir. 2012). So it is odd, to say the least, to ground an actual knowledge standard in a context where implication and constructive waiver suffice. Indeed, even the majority’s cases support constructive wavier so long as it is 'scrutinized with the utmost care.' McDonald v. Steward , 132 F.3d 225, 229 (5th Cir. 1998) (quotation omitted). It is impossible to derive an 'actual knowledge' standard from such permissive language.

Third, how precisely will anyone ever satisfy this brand-new actual knowledge standard? For starters:

• · Can courts infer actual knowledge of a conflict?
• o If such inferences are possible, what is the difference between actual knowledge and implied or constructive knowledge?
• o If such inferences are not possible, just how much direct proof is necessary to show actual knowledge? Must the actual knowledge be evidenced by—
•  A declaration?
•  An affidavit?
•  Advice of counsel?
•  Audio?

• Page 31 Video?
•  Blood?
• · If a General Counsel had actual knowledge of a conflict, is that knowledge imputed to the corporate entity?
• · If a chief executive officer had actual knowledge of a conflict, is that knowledge imputed to the corporate entity?
• o Would it matter if the chief executive was on the board of directors?
• · If an interested party sent letters alerting each board member to the conflict, does that act give rise to actual knowledge?
• o Could the board simply say they did not receive letters? Would it matter if the letters were sent return receipt requested?
• o What if one sent letters to only a minority of board members?
• · What if the shareholders knew about the conflict but the board denied knowledge?

I could go on. But at bottom, there is no better hypothetical than what occurred in this case. Assume, for a moment, that IFG lost at trial. Could it argue that Judge Kay’s order should be vacated because of the Port ’s actions? After all, in the Port’s framing, Judge Kay also “failed to disclose” she was ( 1 ) formerly an attorney at Norman’s firm; ( 2 ) was selected by a committee Norman served on; ( 3 ) was apparently friends with Dees; and ( 4 ) gossiped with Norman about Monk’s chances at becoming a federal judge? Under the majority’s standard, the answer is yes.

And therein lies the rub. It is hard to imagine how any magistrate or bankruptcy judge adjudication could survive the standard set out in today’s opinion.

Page 32* * *

The Port set a trap: muddy the waters with facts and hope the resulting mess justifies some rule that vacates the judgment.

It is a mistake to follow that course. Rather than craft a new standard that risks jeopardizing magistrate and bankruptcy judge adjudication, we should apply the constructive knowledge standard that the statutory text, our precedent, and commonsense demand. With deepest regard for my esteemed colleagues who see the case differently, I respectfully dissent.