75 F.4th 1290

11th Cir.

2023

SKANSKA USA CIVIL SOUTHEAST INC. AND SKANSKA USA, INC., as owners of the Barge KS 5531 praying for exoneration from or limitation of liability v. BAGELHEADS, INC., FLOWERS BY YOKO, GULF BREEZE BAIT & TACKLE, INC. DOG HOUSE DELI II, INC. JLO, INC., et al.

No. 21-13850

United States Court of Appeals For the Eleventh Circuit

August 2, 2023

[PUBLISH]

Appeal from the United States District Court for the Northern District of Florida

D.C. Docket No. 3:20-cv-05980-LC-HTC

SKANSKA USA CIVIL SOUTHEAST INC. AND SKANSKA USA, INC., As owners of the Barge KS 5531 praying for exoneration from or limitation of liability v. BAGELHEADS, INC., FLOWERS BY YOKO, GULF BREEZE BAIT & TACKLE INC, DOG HOUSE DELI II INC, JLO INC, et al.

No. 22-10203

United States Court of Appeals For the Eleventh Circuit

August 2, 2023

Appeal from the United States District Court for the Northern District of Florida

D.C. Docket No. 3:20-cv-05980-LC-HTC

Before BRANCH, GRANT, Circuit Judges, and SCHLESINGER,* District Judge.

GRANT, Circuit Judge:

Hurricane Sally hit Pensacola Bay with a vengeance in September 2020, and 28 barges moored in the Bay were not spared. The barges slammed around the Bay after their moorings snapped, leading to significant damage—including to the Pensacola Bay Bridge, which was closed for months. Skanska, the construction company that owned the barges (and was working on replacing the Bay Bridge) faced hundreds of potential lawsuits. Some were directly related to property damage, but most were economic loss claims from nearby businesses that lost customers during the months-long closure of the bridge.

Skanska filed what are called petitions for limitation of liability, one for each of its 28 barges. These petitions invoked the Limitation Act, a federal law that allows the owner of a maritime vessel to limit its damages in a negligence suit to the combined value of the vessel and its cargo—but only where the owner has neither privity nor knowledge of the negligent acts at issue. After extensive discovery and a bench trial, the district court decided that Skanska could not limit its liability because its own corporate officials were responsible for the negligent acts that led to the barges getting loose in the storm. And once it decided that no limitation of damages could apply, it dismissed the Limitation Act petitions—freeing the claimants to pursue litigation in state court.

Skanska says the district court acted too fast, because the Limitation Act entitles it to more than a decision on limitation of damages (the denial of which it does not contest). It claims that the Limitation Act required the district court to decide whether it was liable to each and every claimant and only then to determine whether it had a right to limit that liability. Here, it says, that would have meant dismissing the economic loss claims because it had no duty as the owner of the barges to prevent that sort of indirect damage.

Skanska‘s approach would turn the Limitation Act on its head, and our precedents have already rejected it. We have been clear that the purpose of the Act is limitation, not exoneration. And the statute‘s text is equally clear—we see no mandate to enforce the two-step process that the company insists is necessary. The Limitation Act allows a federal court to take over all negligence claims to preserve the vessel owner‘s right to limit its liability and then proportionally distribute the available assets to the successful claimants. But only to the extent necessary to protect the right to limitation; it does not create an independent right to have the full merits of each individual claim decided in federal court when no limitation is available.

Skanska also disputes several of the district court‘s other decisions, including multiple evidentiary rulings, the conclusion that it committed negligent acts when it left the barges in the Bay, and the imposition of spoliation sanctions for the destruction of cellphone data. Here too we disagree. We see no reversible error in the district court‘s evidentiary rulings, its findings of fact, or its spoliation sanctions. We therefore affirm the district court.

I.

A.

Pensacola is a city in the westernmost part of the Florida panhandle. It is known for its access to beaches, which attract both locals and tourists from across the country. But Pensacola Bay separates the actual city of Pensacola from the area‘s major beaches (and from several smaller towns such as Gulf Breeze). Pensacola remains connected to its beaches thanks to the Pensacola Bay Bridge—a roughly three-mile bridge across the Bay that has existed in some form since the 1930s.

Around 2010, the bridge needed replacing. Skanska USA Civil Southeast Inc.—which is wholly owned by Skanska USA, Inc.—won a contract to build two new spans and then to destroy the old bridge.¹ Those tasks required construction barges—lots of them. Skanska had a hurricane preparedness plan for the project. If winds of 58 miles per hour or greater were expected within the next 72 hours, the plan called for the construction barges to be moved to Butcherpen Cove on the south side of the Bay, a process that would take at least 30 hours.

The first major warning sign of Hurricane Sally came on Thursday, September 10, 2020, five days before Skanska‘s barges began to break loose. That‘s when the National Hurricane Center issued its notice of a potential storm. Skanska had 55 barges working on the project, and the record does not show that anyone at Skanska was yet aware of the notice. The next day, the National Hurricane Center issued Advisory 1 about what it then called “Tropical Depression 19.” Though it projected that the storm would most likely land at the border between Louisiana and Mississippi, Pensacola Bay fell in the possible 5-day path. Still no sign that Skanska was aware of the storm.

By Saturday morning, that had changed. The National Hurricane Center‘s 72-hour report showed a 16% chance that winds of 58 miles per hour or greater would reach Naval Air

Station Pensacola (a few miles west of the bridge) by Tuesday. Skanska decisionmakers conferenced on Saturday afternoon. By that point, the storm had been upgraded to Tropical Storm Sally and forecasters thought it would be a hurricane by Monday evening. The storm was still projected to make landfall near the Louisiana-Mississippi border—but Pensacola Bay remained on the outskirts of the storm‘s three-day probable path. Skanska management decided to begin preparations to move and secure the barges if necessary. But—in part because of logistical issues with moving the barges over the weekend—the group largely took a “wait and see” approach and agreed to meet again the next morning.

By then, Sunday, the National Hurricane Center had issued a tropical storm warning for the panhandle coast. The 72-hour forecast was slightly better than it had been on Saturday, but it still showed a 9% chance that winds of 58 miles per hour or greater would hit Naval Air Station Pensacola by Wednesday. Skanska began moving its barges, but decided not to take them all the way to Butcherpen Cove. Instead, it moored the barges to various pipe pilings in the Bay—generally within 500 feet of the bridge. Most, but not all, of the barges were tied down by Monday morning, which is also when Tropical Storm Sally was upgraded to Hurricane Sally and the Governor of Florida issued a State of Emergency for the greater Pensacola area. By Monday afternoon, Pensacola Bay was squarely within the hurricane warning zone.

On Tuesday morning, Naval Air Station Pensacola reported winds from 15–26 miles per hour, with gusts reaching almost 45 miles per hour. Skanska‘s barges began to break loose in the Bay, and multiple barges crashed into the bridge. At first, Skanska tried to secure its unmoored barges, but by Tuesday evening it was too dangerous to continue the recovery attempts. On Wednesday, when Hurricane Sally passed through the Bay, it brought winds from 47 to 74 miles per hour, with gusts as high as 92 miles per hour—causing even more barges to slip their moorings. In the end, 28 barges came loose, slamming into the bridge and other property along the Bay‘s edge.² The bridge was closed to traffic for many months because of the damage. This dramatically increased the amount of time it took to get between Pensacola and the towns and beaches on the other side of the Bay.

B.

In short order, Skanska found itself facing state-court lawsuits—many lawsuits—about the damage caused by its barges. Seeking both to limit its liability and to consolidate the multiplying claims in a single forum, Skanska initiated proceedings in federal district court under the Limitation Act. The core of that Act is relatively simple: when a vessel causes “loss, damage, or injury by

collision” without the vessel owner‘s “privity or knowledge,” the owner‘s liability is limited to the value of the vessel and its pending freight. 46 U.S.C. § 30523.

To initiate a limitation proceeding, a vessel owner brings a civil action and deposits the value of the vessel and its freight with the district court. Id. § 30529; Fed. R. Civ. P. Supp. Admiralty & Mar. Claims (Supplemental Rules) F(1), F(2), F(9). This creates a single equitable proceeding about the damage caused by the vessel. All other litigation about the incident—whether in state or federal court—must “cease.” 46 U.S.C. § 30529(c); Supplemental Rule F(3); Beiswenger Enters. Corp. v. Carletta, 86 F.3d 1032, 1036 (11th Cir. 1996). At the same time, those injured by the vessel retain rights under the “saving to suitors” clause. 28 U.S.C. § 1333. That clause, which accompanied the original grant of admiralty jurisdiction to federal district courts, ensures that federal admiralty jurisdiction does not eliminate traditional remedies, including state jury trials. Lewis v. Lewis & Clark Marine, Inc., 531 U.S. 438, 443–46, 453–55 (2001).

Between December 2020 and March 2021, Skanska filed 28 petitions for exoneration or limitation of liability—one for each barge that broke loose during the storm. Following Skanska‘s unopposed motion, all 28 proceedings were consolidated.³ The consolidated proceedings followed standard Limitation Act

procedure: Skanska deposited the value of each of the barges with the district court, initiating a Limitation Act proceeding for each barge, and the court enjoined the filing or prosecution of all related proceedings in other courts. Parties alleging that they had suffered harm from the loose barges then filed claims in the consolidated proceeding, generally with “reservation of rights to proceed in state court.”

Some of these claims came from people and entities whose property was directly damaged by the barges—the United States, for example, alleged that multiple barges caused millions of dollars of damage when they allided with property at Naval Air Station Pensacola. But most of the filings came from businesses claiming not direct physical damage, but economic loss suffered after the closure of the Pensacola Bay Bridge. As an example, the top-line plaintiff here is Bagelheads, Inc., a bagel shop immediately next to the bridge on the Pensacola side. It says it suffered more than $90,000 in damages from the bridge outage because the dramatic increase in travel time from Pensacola to the beach and to the residential towns on the other side of the bridge cut the shop‘s sales by 35%. And Bagelheads was not alone—hundreds of similar claims followed. Skanska moved to dismiss these claims from the limitation proceeding, arguing that it owed no duty to any party whose property did not suffer direct physical damage from the barges.

The district court entered a scheduling order setting a deadline for claimants to join the proceeding and directing all claimants to join in two amended master complaints—one for direct property damage and one for economic damage arising from the bridge outage. The court also announced a bifurcated bench trial. First, the court would consider whether Skanska was entitled to limitation of liability. That involved deciding whether the accident was caused by any negligent acts and, if so, whether Skanska had privity or knowledge of the negligence. Next—but only if that first inquiry showed that Skanska had a right to limit its liability—the parties would conduct damages discovery and the court would schedule any necessary further proceedings. Neither Skanska nor the claimants objected to this scheduling order.

In the end, over 1,000 claimants joined the proceedings—more than 900 of whom alleged economic loss damages from the bridge outage. After the window closed for new filings, Skanska renewed its motion to dismiss the economic loss claims. The court deferred consideration of that issue, in part to better protect the claimants’ right to litigate in the forums of their choice if Skanska were denied limitation.

During discovery, Skanska sought evidence about the claimants’ storm preparations, including extensive discovery from the United States Navy about steps it had taken at the nearby naval air station. A magistrate judge handling pretrial issues held that the Navy‘s preparations were of minimal relevance because it operated seagoing vessels (not barges) and that Skanska‘s discovery of the Navy was not proportional to the needs of the case. See Fed. R. Civ. P. 26(b). Despite that discovery order, Skanska requested depositions from fifteen individuals under the control of the United States Navy, and the magistrate judge granted the Navy a protective order.

After discovery, the district court conducted a five-day bench trial on two issues: Skanska‘s negligence and Skanska‘s right to limit its liability. Yet again, an evidentiary dispute arose. Skanska examined a tugboat captain who had two barges moored to a wharf in Pensacola Bay during Hurricane Sally, with the court repeatedly warning Skanska‘s counsel that his questions about the captain‘s personal experience up to and during the hurricane were mostly irrelevant and directing him to get to the point. After about 14 minutes, the court dismissed the captain, but Skanska was allowed to proffer the rest of his testimony, which showed that the captain would have testified about his experience with storms and about his decision to moor his barges in Pensacola Bay during Hurricane Sally.

The district court ultimately concluded that Skanska had acted negligently. It began with “the Louisiana rule,” which creates a rebuttable presumption that a vessel is negligent when it collides with a stationary object. See In re Skanska USA Civil Se. Inc., 577 F. Supp. 3d 1302, 1313–14 (N.D. Fla. 2021) (citing Bunge Corp. v. Freeport Marine Repair, Inc., 240 F.3d 919, 923 (11th Cir. 2001)); The Louisiana, 70 U.S. (3 Wall.) 164, 173 (1866). The court rejected Skanska‘s assertion that it was “caught off guard” by the storm and found that Butcherpen Cove would have been a safer place to moor the barges. 577 F. Supp. 3d at 1314–18. The “only surprise” to Skanska, the court said, “was that its unreasonable choice to discount—or even ignore—the clear warnings of an approaching tropical storm turned out to have harsh consequences.” Id. at 1322.

The court quickly rejected Skanska‘s “perfunctory” argument that it lacked privity or knowledge of the negligent acts, pointing out that those acts “sprung wholly from executive decision-making that resulted in the failure to take reasonable measures to protect its barges from the impending storm.” Id. at 1324. It then dismissed Skanska‘s petitions for exoneration from or limitation of liability and dissolved the injunction barring prosecution of all related litigation. Id. The court never ruled on Skanska‘s motion to dismiss the economic loss claims.

C.

Along with the adverse verdict, Skanska was sanctioned for spoliating electronic evidence under Rule 37(e); the data from five out of thirteen discovery custodians’ cell phones was destroyed. Even with an active litigation hold and actual litigation, Skanska did not back up the relevant employees’ cell phones. Nor did it suspend its ordinary cell phone data destruction policies.

Two phones were deliberately reset according to Skanska‘s ordinary employee departure procedures when their owners left the company. Another was somehow “disabled” and became inaccessible after the owner left Skanska. Yet another was allegedly lost overboard. And still another had all text messages deleted under disputed circumstances.

The district court ruled that Skanska “acted with the intent to deprive” the claimants of the cell phone data and sanctioned the company under Rule 37(e)(2).⁴ The court emphasized that it saw no cogent explanation, apart from bad faith, for Skanska‘s systematic failure to make any effort to preserve cell phone data until at least seven months after the litigation hold was (technically) in place. So it made two adverse inferences against Skanska and ordered it to pay the claimants’ costs and attorneys’ fees for the sanctions motion.

Skanska appeals both the district court‘s final judgment and its sanctions order, and we consolidated the appeals. Skanska also moved to stay the dissolution of the district court‘s injunction pending appeal to prevent the lawsuits from proceeding against it in state court, but we denied its request.

II.

We begin by addressing Skanska‘s core argument—that the district court violated the Limitation Act by failing to first adjudicate the full merits of the claims before deciding whether its liability could be limited. In particular, we consider Skanska‘s argument that it was entitled to have the economic loss claims dismissed, which would have barred the claimants from bringing

those claims in another forum. We review questions of law about the scope of the Limitation Act de novo. See Dresdner Bank AG v. M/V Olympia Voyager, 446 F.3d 1377, 1381 (11th Cir. 2006). And we review the district court‘s decision to dismiss the entire Limitation Act proceeding for abuse of discretion. See Lewis, 531 U.S. at 454.

A.

The Limitation Act is a difficult statute. It was written in 1851 to address economic realities that are unrecognizable today.⁵ Lewis, 531 U.S. at 446–47. And it was “badly drafted even by the standards of the time.” Id. 447 (quoting 2 T. Schoenbaum, Admiralty and Maritime Law 299 (2d ed. 1994)). Of course, that does not change this Court‘s job—we apply the law as it exists, no matter how poorly it may have aged. But, keenly aware of the inscrutability of this statute, we offer relevant background about the basic history and substance of the Act (and its implementing rules), the petitioner‘s ability to demand exoneration, the

injunction against ongoing litigation, and the Act‘s relationship with the equally mysterious saving to suitors clause.

First, the history. In 1851, Congress thought the American shipping industry was disadvantaged as compared to its foreign competitors. See Lewis, 531 U.S. at 446–47. Unlike their counterparts in England, American courts did not limit the potential liability of vessel owners. See generally Joseph C. Sweeney, Limitation of Shipowner Liability: Its American Roots and Some Problems Particular to Collision, 32 J. Mar. L. & Com. 241, 243–52 (2001). Congress stepped in, enacting the Limitation Act “to encourage ship-building and to induce capitalists to invest money in this branch of industry.” Lewis, 531 U.S. at 446 (quoting Norwich Co. v. Wright, 80 U.S. (13 Wall.) 104, 121 (1872)). To that end, the Act limits a vessel owner‘s liability for any “loss, damage, or injury by collision . . . done, occasioned, or incurred, without the privity or knowledge of the owner.” 46 U.S.C. § 30523(b).⁶ And “consistent with the statutory purpose to protect innocent investors, ‘privity or knowledge’ generally refers to the vessel owner‘s personal participation in, or actual knowledge of, the specific acts of negligence or conditions of unseaworthiness which caused or contributed to the accident.” Suzuki of Orange Park, Inc.

v. Shubert, 86 F.3d 1060, 1064 (11th Cir. 1996).⁷ In other words, the Act limits the liability of vessel owners who were not in some sense responsible for the specific negligent acts or conditions of unseaworthiness that caused the accident. Id.

If a vessel owner is entitled to limitation, then its liability is limited to the post-accident value of the vessel and its pending freight. 46 U.S.C. § 30523(a); Tug Allie–B, Inc. v. United States, 273 F.3d 936, 939 (11th Cir. 2001). That means that the owner‘s total loss—its liability plus the loss from the accident itself—will be no greater than the value of the vessel and its cargo. The Act also recognizes that multiple would-be plaintiffs might need to share in these limited funds, so it provides that—if the funds are insufficient to pay all claims—injured parties will be “paid in proportion to their respective losses.” 46 U.S.C. § 30525(1).

When courts first began to apply the Limitation Act, they encountered a major problem: it had no procedures. In the Supreme Court‘s first case about the Act, the Court observed that it was “reduced to the dilemma of inferring that the legislature has

passed a law which is incapable of execution.” Norwich, 80 U.S. (13 Wall.) at 123. So the Court took matters into its own hands, promulgating rules of admiralty to implement the Act. Lewis, 531 U.S. at 447 (citing Supplementary Rules of Practice in Admiralty, 80 U.S. (13 Wall.) xii–xiv (1872)). A vessel owner seeking limitation of liability would deposit the value of his interest in the vessel and its freight before the court. Id. Then, the district court would order all claimants to appear. Id. If the vessel owner was entitled to limitation, the court would distribute the limited fund among the claimants. Id. at 448.

Congress eventually codified the basic bond posting procedure for the Limitation Act, and the Supreme Court over time has amended and restyled the limitation rules. See 46 U.S.C § 30529; Supplemental Rule F. But the same basic structure remains, and the Act‘s procedures are still primarily implemented by court rule rather than statute.

One of those procedures is the injunction against existing litigation. When a Limitation Act petition is filed, the district court enjoins the prosecution of all pending litigation that overlaps with the subject matter of that petition. 46 U.S.C. § 30529(c); Supplemental Rule F(3). A pause in other litigation about the incident is necessary to protect the limitation fund—after all, the plaintiffs’ combined damages may well exceed the value of the vessel and its cargo. But if one or more plaintiffs received final judgments in other litigation that exceeded this value, then the vessel owner‘s right to limit its liability would be meaningless.

Here, the injunction stopped lawsuits like the one the bagel shop filed in Florida state court—but only while the limitation petitions were pending. Skanska‘s objection is that the district court dissolved the injunction and allowed the lawsuits to proceed.

Finally, we discuss the complicated relationship between the Limitation Act and the saving to suitors clause. See 28 U.S.C. § 1333. That clause falls not in the Limitation Act, but in the general grant of admiralty jurisdiction to the federal district courts. Id. It gives federal district courts “original jurisdiction, exclusive of the courts of the States” of any “civil case of admiralty or maritime jurisdiction.” Id. But it does so while “saving to suitors in all cases all other remedies to which they are otherwise entitled.” Id. Both this grant of federal jurisdiction and the clause‘s corresponding protection of suitors have existed in some form since the Judiciary Act of 1789. Lewis, 531 U.S. at 443–44 (citing Judiciary Act of 1789, ch. 20, § 9, 1 Stat. 73, 77). The Supreme Court has “theorized that the saving to suitors clause was ‘inserted, probably, from abundant caution,‘” so that the grant of admiralty jurisdiction was not wrongly understood to eliminate “‘the concurrent remedy which had before existed‘“—state tort lawsuits. Id. at 444 (quoting N.J. Steam Navigation Co. v. Merchs.’ Bank of Bos., 47 U.S. (6 How.) 344, 390 (1848)). This clause, then, makes it clear that the existence of federal admiralty jurisdiction (which offers no jury right) does not generally keep plaintiffs from seeking relief in other forums—perhaps forums that do offer a jury trial. See id. at 445–46, 454–55; see also, e.g., Suzuki, 86 F.3d at 1063.

Make no mistake—“tension exists between the saving to suitors clause and the Limitation Act.” Lewis, 531 U.S. at 448. The former protects the right to a jury trial. Id. at 445–46, 454–55. The latter creates a proceeding with no jury trial and stays all litigation in any other forum—including where a jury trial would be available. Id. at 448. In light of these somewhat contradictory provisions, the district court is charged with safeguarding a vessel owner‘s right to have the question of limitation adjudicated in federal court while simultaneously ensuring that claimants retain their choice of forum under the saving to suitors clause.

Both this Court and the Supreme Court have repeatedly grappled with this interplay. See, e.g., Offshore of the Palm Beaches, Inc. v. Lynch, 741 F.3d 1251, 1257–59 (11th Cir. 2014); Lewis, 531 U.S. at 448–56; Suzuki, 86 F.3d at 1063–64; Beiswenger, 86 F.3d at 1036–40; Lake Tankers Corp. v. Henn, 354 U.S. 147, 150–54 (1957); Langnes v. Green, 282 U.S. 531, 541–44 (1931). At least a few clear principles have emerged. To start, a vessel owner has an “absolute right to claim the Act‘s liability cap, and to reserve the adjudication of that right in the federal forum.” Beiswenger, 86 F.3d at 1037 (quotation omitted). This guarantee means that a primary duty of the limitation court is to ensure that outside lawsuits do not undermine its exclusive authority to decide limitation. And if that right to limitation cannot be adequately protected with other lawsuits proceeding elsewhere, the limitation court may adjudicate the merits of the entire controversy—which means deciding both liability and limitation. Lewis, 531 U.S. at 454. But if the right to litigate and enforce limitation in federal court is protected, the

saving to suitors clause counsels in favor of giving plaintiffs their choice of forum. Id. at 454–55. In that case, the district court has “discretion to stay or dismiss Limitation Act proceedings to allow a suitor to pursue his claims in state court.” Id. at 454.

B.

With this background, we can address Skanska’s argument that the Limitation Act requires a federal court to decide whether it owed a duty to each and every claimant—or, to put it another way, Skanska’s argument that it had a right to be exonerated in federal court.

Skanska primarily leans on a two-step procedure that we have said is “typical” in a Limitation Act case: “First, the court must determine what acts of negligence or conditions of unseaworthiness caused the accident. Second, the court must determine whether the shipowner had knowledge or privity of those same acts of negligence or conditions of unseaworthiness.” Beiswenger, 86 F.3d at 1036 (quoting Hercules Carriers, Inc. v. Florida, 768 F.2d 1558, 1563–64 (11th Cir. 1985)); see also, e.g., Farrell Lines, Inc. v. Jones, 530 F.2d 7, 10 (5th Cir. 1976); Hartford Accident & Indem. Co. v. S. Pac. Co., 273 U.S. 207, 214–15 (1927); Providence & N.Y. S.S. Co. v. Hill Mfg. Co., 109 U.S. 578, 595 (1883).⁸ Skanska reads these cases as creating a mandatory order of decision for nearly all Limitation Act cases. And, in Skanska’s telling, that mandatory first step includes a decision on every element of negligence—duty, breach, injury, and actual and proximate cause—for every claimant. So, Skanska claims, the district court violated this order of decision by reaching the question of privity and knowledge before fully adjudicating its liability to each claimant.

This Court has already rejected Skanska’s rigid reliance on the “typical” Limitation Act procedure. We have held that if it is “impossible under any set of circumstances for the vessel owner to demonstrate the absence of privity or knowledge,” then “the admiralty court may decide the privity or knowledge issue without first deciding the liability issue.” Suzuki, 86 F.3d at 1064. Although Suzuki articulated a summary judgment standard, the same basic principle logically applies at every stage of the proceedings, including after a trial.⁹ Once it is apparent that the vessel owner cannot establish a lack of privity or knowledge, then limitation is not at issue. And if limitation is not at issue, then “the basis for granting exoneration vanishes.” Id. (quoting Fecht v. Makowski, 406 F.2d 721, 723 (5th Cir. 1969)). So whenever the court finds that the vessel owner cannot establish a lack of privity or knowledge, it is appropriate to dismiss the petition to protect the claimants’ rights under the saving to suitors clause—even if that means forgoing (in part or in entirety) a decision on the vessel owner’s liability.

This is consistent with decades of Limitation Act law about claimants’ saving to suitors clause rights. The Act does “not create a freestanding right to exoneration from liability in circumstances where limitation of liability is not at issue.” Lewis, 531 U.S. at 453. It is about limitation of liability, not immunity from liability or the exclusivity of the federal forum. Lake Tankers, 354 U.S. at 152–53. Courts should thus “give effect to both the Limitation Act and the saving to suitors clause whenever possible” and “damage claimants must be allowed to litigate the vessel owner’s negligence in state court, ‘where it is apparent that limitation cannot be granted.’” Beiswenger, 86 F.3d at 1037; Suzuki, 86 F.3d at 1063 (quoting Fecht, 406 F.2d at 722).¹⁰

Here, Skanska’s “perfunctory” claim that it lacked privity or knowledge straightforwardly failed (indeed, Skanska did not even appeal the district court’s finding that the company had privity or knowledge). After all, the decision not to move the barges to safety was made by Skanska executives, and when a corporation owns a vessel, “the privity and knowledge of corporate managers vested with discretionary authority is attributed to the corporation.” Suzuki, 86 F.3d at 1065 (quotation omitted). Once it was clear that Skanska had privity or knowledge of any negligent acts that caused the barges to damage property, it was equally clear that the company had no right to limit its liability. So there was no continued need for a limitation proceeding. The district court thus did not abuse its discretion by dismissing the consolidated proceedings.¹¹

While we can resolve this case on precedent alone, the text of both the Limitation Act and its rules point in the same direction. The operative provisions of the Limitation Act do not mention exoneration. See 46 U.S.C. §§ 30501–30530. The only source of positive law that even plausibly supports Skanska’s argument is Rule F’s statement that a vessel owner “may demand exoneration from as well as limitation of liability.” Supplemental Rule F(2). But this argument misses the mark. The rule’s history clarifies its narrow scope. The limitation rules’ exoneration language was added to clarify that, unlike under the “very onerous” English rules, an American vessel owner did not need to concede liability in order to seek limitation of liability. Lewis, 531 U.S. at 447–48 (quotation omitted). In light of this history, we see no reason to think that this language does anything more than preserve the right to contest liability while also seeking limitation; it certainly does not “create a freestanding right to exoneration from liability in circumstances where limitation of liability is not at issue.” Id. at 453.

Even if there were evidence that Rule F somehow expanded vessel owners’ rights to a federal forum, we would not be able to read it that way. Judicially promulgated rules cannot “abridge, enlarge or modify any substantive right.” 28 U.S.C. § 2072; see also In re Tidewater Inc., 249 F.3d 342, 347 (5th Cir. 2001) (rejecting an aggressive reading of Rule F on this ground). Any tension between the judicially adopted rules and the congressionally enacted saving to suitors clause must be resolved in the clause’s favor. No matter how Skanska frames its argument, it simply has no right to a federal forum beyond the question of limitation.¹²

As a last resort, Skanska points to dicta saying that the Act and its rules protect vessel owners from “being harassed by litigation in other tribunals” which could lead to the “subversion of the whole object and scheme of the statute.” Providence & N.Y. S.S. Co., 109 U.S. at 594–95; see also, e.g., Hartford, 273 U.S. at 215–16. But the Court has long-since abandoned this dicta. For example, in Lake Tankers, “the Court explicitly rejected the argument that the Limitation Act protects the vessel owner against a multiplicity of suits.” Beiswenger, 86 F.3d at 1039 (citing Lake Tankers, 354 U.S. at 153–54). If limitation is not implicated, the Court noted, there is no reason why a vessel owner “should be treated any more favorably than an airline, bus, or railroad company.” Lake Tankers, 354 U.S. at 153. To be sure, if Congress wishes to protect vessel owners from a multiplicity of suits, it can always expand the Limitation Act and contract the saving to suitors clause. But in the meantime, courts and litigants alike must respect the relationship between these laws.

Neither Skanska’s appeals to procedure nor its gestures to policy sway our conclusion that the district court did not abuse its discretion in dismissing Skanska’s limitation petitions without fully addressing the question of Skanska’s liability. Yes, Skanska had the right to have the question of limitation adjudicated in federal court. And the question of limitation was adjudicated in federal court. But that is as far as it goes. Once the district court decided that Skanska was not entitled to limitation, it correctly dismissed the case so that the claimants could pursue remedies in whichever forums that they chose.¹³

III.

Skanska also challenges two evidentiary rulings: the order limiting its discovery of the Navy’s preparations during Hurricane Sally and the court’s mid-testimony dismissal of Captain Towne at trial.¹⁴ “We review a district court’s evidentiary rulings for abuse of discretion.” Great Lakes Ins. SE v. Wave Cruiser LLC, 36 F.4th 1346, 1353 (11th Cir. 2022). And “even a clearly erroneous evidentiary ruling will be affirmed if harmless.” Id. (quotation omitted). An error is harmless unless “it affects the substantial rights of the parties” such that the reviewing court cannot confidently say that “the judgment was not substantially swayed by the error.” Furcron v. Mail Ctrs. Plus, LLC, 843 F.3d 1295, 1304 (11th Cir. 2016) (quotation omitted).

Starting with the Navy, under Rule 26(b), parties may only obtain discovery that is both “relevant” and “proportional to the needs of the case.” Fed. R. Civ. P. 26(b)(1). The magistrate judge’s discovery order limited discovery relating to the Navy because it was irrelevant and because it was not proportional. On appeal, Skanska only argues that its proposed discovery was relevant—it never explains why the burdens were proportional to the benefits. Skanska has therefore forfeited any challenge to proportionality, so we affirm. See Sapuppo v. Allstate Floridian Ins. Co., 739 F.3d 678, 681 (11th Cir. 2014).¹⁵

As for Captain Towne, the district court heard nearly fifteen minutes of his testimony while repeatedly urging Skanska’s counsel to get to the point. It then considered the captain’s testimony via a one-sided proffer, and it explicitly stated that the full testimony would not have affected its decision. Any exclusion of relevant evidence that the captain might have offered was harmless.

IV.

Moving to the factual basis for the district court’s final judgment, Skanska challenges the finding that it did not exercise reasonable care.¹⁶ For maritime tort cases, “we rely on general principles of negligence law” and have consulted “in particular the Restatement (Second) of Torts” to discern these general principles. Tesoriero v. Carnival Corp., 965 F.3d 1170, 1178 (11th Cir. 2020) (quotations omitted). Whether the defendant’s conduct violated the standard of care is a question of fact. See Restatement (Second) of Torts § 328C cmt. b (Am. L. Inst. 1965). Because the district court sat in admiralty without a jury, we review its findings of fact for clear error. Dresdner Bank, 446 F.3d at 1380. A factual finding is clearly erroneous if “the entirety of the evidence leads the reviewing court to a definite and firm conviction that a mistake has been committed.” Id.

The core of the district court’s analysis was simple: Under the Louisiana rule, Skanska bore the burden of disproving a presumption that it failed to exercise reasonable care. Throughout the weekend, the National Hurricane Center’s weather forecasts showed a 9–16% chance of 58 mile per hour or greater winds hitting Pensacola Bay. According to Skanska’s own hurricane preparedness plan, these wind speeds warranted moving the barges to Butcherpen Cove. Skanska’s inability to get its barges to Butcherpen Cove before high winds arrived was one of two things: a deliberate choice or the consequence of a failure to act more quickly. Either way, the court explained, Skanska left its barges in a less safe location than Butcherpen Cove—and its gamble backfired with disastrous consequences. Given the substantial damage that the barges could do to others if the wind unmoored them, the court found that a reasonable company facing a 9–16% chance of these winds would have moved the barges. Thus, the court held, Skanska acted unreasonably by leaving the barges by the bridge.

This conclusion was not clearly erroneous. Skanska does not attempt to show that the district court’s finding that there was a 9–16% chance of high winds hitting the Bay was incorrect. Nor does it present us with evidence that a reasonable company would have left its barges by the bridge notwithstanding those odds of high winds. Indeed, it fails to engage with the factual findings of the lower court, instead repeating its arguments below about how its preferred weather reports did not show that a hurricane was more likely than not to hit Pensacola Bay until Monday morning (without ever suggesting exactly how likely the storm was to hit the Bay) and emphasizing the difficulty of moving the barges over the weekend. This argument is non-responsive to the district court’s finding that reasonable care required Skanska to take precautions even with a 9–16% chance of high winds in the Bay. Skanska has failed to show reversible error.¹⁷

V.

Finally, Skanska challenges spoliation sanctions it received under Rule 37(e)(2) for the destruction of the data from five of its custodians’ cell phones. We review that decision for abuse of discretion. Tesoriero, 965 F.3d at 1177. In the spoliation context, this Court has often articulated our abuse of discretion review as asking whether the district court “has made a clear error of judgment, or has applied the wrong legal standard.” Mendez v. Wal-Mart Stores E., LP, 67 F.4th 1354, 1361 (11th Cir. 2023) (quotation omitted). But we see no difference between this formulation of the abuse of discretion standard and our articulations of that standard in other contexts. See, e.g., Callahan v. United Network for Organ Sharing, 17 F.4th 1356, 1360 (11th Cir. 2021) (“A district court abuses its discretion when it applies an incorrect legal standard, follows improper procedures in making the determination, makes findings of fact that are clearly erroneous, or commits a clear error of judgment.” (quotations omitted)). Accordingly, as is typical for this Court’s abuse of discretion review, we review the district court’s legal conclusions de novo and its subsidiary factual findings for clear error. Common Cause Ga. v. Georgia, 17 F.4th 102, 107 (11th Cir. 2021). That means that we review the district court’s conclusion that Skanska acted in bad faith—a factual finding—for clear error. Cf. Mar. Mgmt., Inc. v. United States, 242 F.3d 1326, 1331 (11th Cir. 2001).

A.

Spoliation sanctions are often imposed under the broad discretion of the district court, which has inherent power to “manage its own affairs and to achieve the orderly and expeditious disposition of cases.” Flury v. Daimler Chrysler Corp., 427 F.3d 939, 944 (11th Cir. 2005). But sometimes other sources of federal law provide more specific authority.

Rule 37(e) from the Federal Rules of Civil Procedure is one such example. It was adopted in 2006 and modified in 2015 in response to the explosion of electronic discovery. 2015 Committee Notes on Rule 37(e). The Rule seeks to balance the centrality of electronic information to modern litigation with the almost always substantial (and sometimes unnecessary) costs of preserving electronic data. Id.

By its text, Rule 37(e) creates a two-tiered sanctions regime—with lesser sanctions under Rule 37(e)(1) and more severe sanctions under Rule 37(e)(2). Both parts of the rule share two preconditions: (1) “electronically stored information that should have been preserved in the anticipation or conduct of litigation” was “lost because a party failed to take reasonable steps to preserve it” and (2) that information “cannot be restored or replaced through additional discovery.” The requirements diverge after that. Rule 37(e)(1) sanctions are centered on the effect of a violation; they apply only where lost electronic evidence causes “prejudice to another party,” which then justifies sanctions “no greater than necessary to cure the prejudice.” Rule 37(e)(2) sanctions, on the other hand, look more to the cause of the violation. They require a finding that “the party acted with the intent to deprive another party of the information’s use in the litigation.” If so, the court is justified in imposing more severe sanctions: adverse jury instructions, and even dismissal or default judgment. Fed. R. Civ. P. 37(e)(2). What’s more, Rule 37(e)(2) sanctions do not require “any further finding of prejudice.” 2015 Committee Notes on Rule 37(e)(2). “This is because the finding of intent required by the subdivision can support not only an inference that the lost information was unfavorable to the party that intentionally destroyed it, but also an inference that the opposing party was prejudiced by the loss of information that would have favored its position.” Id.

This is the first time we have had reason to thoroughly consider Rule 37(e)(2). In previous cases, we have suggested or assumed, but never held in a published opinion, that the Rule’s “intent to deprive” standard is the same as the “bad faith” standard our Court has used in other spoliation contexts. See ML Healthcare Servs., LLC v. Publix Super Mkts., Inc., 881 F.3d 1293, 1308 (11th Cir. 2018); Ala. Aircraft Indus., Inc. v. Boeing Co., No. 20-11141, 2022 WL 433457, at *15 (11th Cir. Feb. 14, 2022) (unpublished). Here, the district court and both parties have assumed that “intent to deprive” and “bad faith” are the same.

We now explicitly agree: the “intent to deprive another party of the information’s use in the litigation” is the equivalent of bad faith in other spoliation contexts. Fed. R. Civ. P. 37(e)(2). In this Circuit’s spoliation precedents, bad faith “generally means destruction [of evidence] for the purpose of hiding adverse evidence.” Tesoriero, 965 F.3d at 1184 (emphasis added) (quotation omitted). This standard is more than mere negligence and aligns with both the text and advisory committee notes for Rule 37(e)(2). See id. at 1184–86; Rule 37(e)(2); 2015 Committee Notes on Rule 37(e)(2). The phrase “intent to deprive” naturally requires that the spoliator has a “purpose of hiding adverse evidence” from other parties. And the advisory committee notes explain that “intent to deprive” is more than negligence or even gross negligence. 2015 Committee Notes on Rule 37(e)(2). So rather than adopt a new law of spoliation from scratch for Rule 37(e)(2) sanctions, we will borrow our identical and well-trodden standard of “bad faith.” See also EEOC v. Jetstream Ground Servs., Inc., 878 F.3d 960, 965–66 (10th Cir. 2017) (conducting a similar analysis).¹⁸

B.

We now review the district court’s imposition of Rule 37(e)(2) sanctions. It is not entirely clear whether Skanska challenges Rule 37(e)’s first requirement, that information that “should have been preserved in the anticipation or conduct of litigation” was “lost because a party failed to take reasonable steps to preserve it.” Skanska concedes that it was “probably negligent.” But it also seems to suggest in its briefs that its preservation failures could be excused because it may not have reasonably anticipated litigation.

The facts reveal that argument’s folly. To start, within days of Hurricane Sally, Skanska’s in-house counsel had orally informed employees of an evidence retention policy. And about a month later, counsel sent a formal legal hold letter about the Pensacola Bay Bridge incident. Litigation started just a few weeks after that. But even with an active litigation hold—and then active litigation—Skanska did not back up its employees’ cell phones. Nor did it suspend its ordinary cell phone data destruction policies—not even for known electronic data custodians. It was, as we see it, entirely predictable that cell phone data would be needed for litigation, and that some of that data would be lost.

Next, Skanska more directly challenges the applicability of Rule 37(e)’s second condition that the information could not have been “restored or replaced through additional discovery.” It argues that, because discovery of other Skanska employees included some of the text messages and because the claimants were able to depose the five custodians whose data was destroyed, nothing was actually lost.

The district court was not moved; nor are we. While some of the lost text messages were discoverable through other Skanska employees’ text messages, others were not. And it should go without saying that deposing workers well after an event is not a perfect substitute for reviewing their contemporaneous text messages.

Finally, Skanska challenges the district court’s finding that it acted in bad faith when it failed to back up the custodians’ text messages and suspend its destruction policies. The court found a “lack of any cogent explanation” for Skanska’s complete failure to make any effort to preserve the destroyed cell phones. In re Skanska USA Civil Se. Inc., 340 F.R.D. 180, 189 (N.D. Fla. 2021). It focused in particular on how the company “took no action” to educate its custodians and administrators about the litigation hold and “made no effort” to collect its custodians’ cell phone data until at least seven months after the litigation hold was in place. Id. at 188–89. And it highlighted other egregious discovery behavior by Skanska, like its representation to the court that “no documents relating to Skanska’s storm preparations or efforts has [sic] been deleted or destroyed”—a representation made before Skanska had even bothered to check if its custodians’ cell phone data was still available. Id. at 189 (alteration adopted and quotation omitted). In the district court’s view, bad faith was the only thing that explained the company’s actions.

If our review were de novo, this would be a close question. On the one hand, we find Skanska’s utter failure to implement even the most basic data-protection safeguards egregious—so egregious that an inference of bad faith is easy to make. On the other, this is not a case with direct evidence of bad faith; it is also plausible from this record that Skanska was “just” grossly negligent.

But we review the district court’s finding of bad faith for clear error. And an inference of bad faith here was not clear error. Skanska can provide no reasonable explanation for its conduct other than to plead negligence. True enough, much of the evidence was destroyed through “routine” document destruction policies. But a hands-off implementation of an ordinary corporate destruction policy is not a silver bullet. We have already explained that we would be “highly skeptical” of a claim that evidence was unintentionally destroyed “pursuant to a routine policy” after a request that the evidence be preserved. Tesoriero, 965 F.3d at 1186. We will not second guess the district court’s skepticism in those very circumstances.

Our conclusion is further confirmed by the advisory committee notes to Rule 37(e). The reasonableness of evidence preservation efforts depends in part on “the party’s sophistication with regard to litigation in evaluating preservation efforts.” 2015 Committee Notes on Rule 37(e). Skanska is a sophisticated entity—a multinational company tasked with completing a construction contract worth nearly $400 million. Skanska USA even boasts on its website that it is “one of the largest, most financially sound construction and development companies in the U.S.” Even so, Skanska did not bother to take the most fundamental of precautions—starting with backing up the custodians’ cell phones and suspending its policy of wiping those phones. And the company is not being held liable for a failure of imagination—Skanska had an active litigation hold, but took no steps to implement it.

Finally, Skanska argues that—as a per se rule—a finding of bad faith premised on circumstantial evidence requires an “affirmative act” by the spoliating party.¹⁹ But that argument flouts both the text of Rule 37 and our bad-faith caselaw. As we have explained, the rule provides for sanctions when “electronically stored information that should have been preserved in the anticipation or conduct of litigation is lost because a party failed to take reasonable steps to preserve it.” Fed. R. Civ. P. 37(e) (emphasis added). Failure to act can thus—by definition—be a violation of Rule 37.

Our caselaw on spoliation also shows that failures to act can be just as harmful as affirmative acts of destruction. In Flury, we affirmed sanctions after the spoliating party “failed to preserve” a vehicle that the other party wished to inspect, “inexplicably ignored” requests for the location of the vehicle, and “allowed the vehicle to be sold for salvage without notification to defendant of its planned removal.” 427 F.3d at 943, 947, 945. These are failures to act—not affirmative actions. Even so, it “is no surprise that we found bad faith on those facts.” Tesoriero, 965 F.3d at 1185.

So too here. Skanska’s passivity does not change the basic fact that the evidence was destroyed. Given that other circumstances pointed to a reasonable inference of bad faith by Skanska, it is irrelevant whether an act or a failure to act directly caused the spoliation. The court’s finding of bad faith thus was not clear error, and its imposition of Rule 37(e)(2) sanctions was not an abuse of discretion.

* * *

Under the Limitation Act, Skanska had a right to have the issue of limitation litigated by a federal court. That’s exactly what Skanska got. The claimants now have the right to determine where any subsequent litigation about the damage caused by Skanska’s barges during Hurricane Sally will occur. And Skanska has failed to show any other reason why the district court’s dismissal of its petitions should be reversed.

We AFFIRM both the district court’s dismissal of the Limitation Act proceedings and its sanctions order.

Notes

The district court sat in admiralty and made findings of fact after a bench trial. “When reviewing the judgment of a district judge sitting in admiralty with no jury, we may not set aside the court‘s findings of fact unless they are clearly erroneous.” . So—other than those findings of fact that Skanska challenges, which we review for clear error—we take all of the district court‘s findings of fact as true.

The district court mentioned 27 barges that broke loose. But Skanska initiated Limitation Act proceedings for 28 barges, and on appeal, both parties refer to 28 barges as having broken loose. We adopt the parties’ characterization that 28 barges broke loose for clarity, but we do not hold one way or the other whether the district court‘s finding that 27 barges broke loose was erroneous.

Twenty-two of the proceedings were consolidated in an order on February 15. The remaining six proceedings were filed on March 26 and were treated as part of the consolidated proceeding.

The claimants’ spoliation motion was referred to a magistrate judge, who issued the opinion sanctioning Skanska. But the district court denied Skanska‘s objections to the order and “wholly concur[red]” with the magistrate judge, so we refer to the magistrate judge‘s order as coming from the district court for clarity.

Nearly fifty years ago, our predecessor Court referred to the Act as “hopelessly anachronistic“—largely because it was a creative way to limit liability “in an era before the corporation, with its limited shareholder liability, had become the standard form of business organization and before the present range of insurance protection was available.” . It also quoted a leading 1950s admiralty treatise that said the Act “has been due for a general overhaul for the past seventy-five years.” at 441 & n.3 (quoting Grant Gilmore & Charles L. Black, Jr., The Law of Admiralty 677 (1st ed. 1957)). If Gilmore and Black were right, then we are at 141 years and counting.

While this appeal has been pending, Congress restructured and renumbered the codified Limitation Act and limited the extent to which the Act applies to certain small passenger vessels. James M. Inhofe National Defense Authorization Act for Fiscal Year 2023, Pub. L. No. 117–263, § 11503, 136 Stat. 2395, 4130–31 (2022).

Linguistically, the idea of a vessel owner having “privity or knowledge” of specific “negligent acts” is a little awkward; ordinarily, one thinks of privity as being with actors, not of acts. But the meaning of this clause can be clarified by thinking about the archetypical case of limitation: a nineteenth century shipowner whose ship was thousands of miles away, completely outside of his control, when the crew unexpectedly acted negligently. In such cases, the vessel owner might be in a formal sense in privity with the negligent actors, but would not be thought to have had “privity or knowledge” of the crew‘s specific negligent acts.

All published cases of the former Fifth Circuit decided before the close of business on September 30, 1981, are precedent in this Circuit. See (en banc).

The district court declined to apply Suzuki at summary judgment, holding that it was still possible at that point in the litigation for Skanska to show a lack of privity or knowledge. But clearly, after trial, the court determined that Skanska could not meet its burden to show a lack of privity or knowledge.

Skanska claims that Beiswenger identifies the only instances in which district courts may depart from the two-step Limitation Act procedure. But Beiswenger did not attempt to identify every scenario where procedural flexibility is appropriate. To the contrary, we read Beiswenger as blessing procedural creativity by district courts in order to give effect to both the clause and the Limitation Act “whenever possible.” See 86 F.3d at 1037.

The parties dispute, for what it is worth, whether the district court even departed from the two-step procedure. Although the court did not determine every element of negligence for every claimant, it did decide that Skanska’s negligence in not moving the barges to Butcherpen Cove caused the damage to the bridge and other property struck by the barges. The claimants argue that this finding was enough to show what “acts of negligence caused the accident,” satisfying the first step of the two-step procedure, while Skanska argues that nothing short of a full adjudication of every element of negligence for every individual claimant will do. But we think that this debate misses the point; it does not matter whether the district court followed the ordinary two-step procedure. All that matters is that Skanska’s right to have limitation adjudicated in federal court was protected—and it was.

None of this is to say that the two-step approach articulated in many Limitation Act cases is obsolete. After all, it is an intuitive way to approach the problem. “Without negligence there can be no privity or knowledge,” because the vessel owner’s privity or knowledge must be of a particular negligent act. (quoting 3 Benedict on Admiralty, § 41, p. 5-5). So it makes sense that courts often decide negligence (either in part or in full) before deciding privity or knowledge. At the same time, courts should not overread our discussions of the “typical” Limitation Act case, including the occasional use of the word “must” when discussing the steps of that procedure. “The language of an opinion is not always to be parsed as though we were dealing with language of a statute.” (alteration adopted and quotation omitted).

Skanska fleetingly briefs another procedural argument, saying that the district court violated Federal Rule of Civil Procedure 12(i) by failing to rule on its 12(b)(6) motion. If Skanska wished to argue that Rule 12(i) required the district court to address the motion to dismiss before the first phase of the bifurcated trial, then it needed to raise that issue before the first phase of the bifurcated trial. It did not do so, instead raising the issue for the first time in post-judgment motions. We review issues not timely raised before the district court only for plain error, and rarely in civil cases. See . “Under the civil plain error standard, we will consider an issue not raised in the district court if it involves a pure question of law, and if refusal to consider it would result in a miscarriage of justice.” (quotation omitted). Refusing to consider this argument does not result in a miscarriage of justice, and we therefore do not consider it.

Skanska says that the order denying discovery of the Navy led them to believe that the court would not allow discovery of other parties. It now argues that, because this constructively denied additional discovery, the entire discovery process was defective. But it goes without saying that Skanska cannot challenge a hypothetical ruling of the district court. We therefore review only the two specific evidentiary rulings that Skanska appeals.

Because we affirm the limitation of the discovery as non-proportional, we do not address whether the information Skanska sought was relevant.

We do not read the district court as having decided whether Skanska violated any duty to any particular claimant. So Skanska is still free to argue in state court that it did not owe a duty to the economic loss claimants.

Relatedly, Skanska challenges the district court’s factual finding that its hurricane preparedness plan was triggered at all. But we agree with the district court that the core inquiry is “whether the decision to demobilize the barges near the bridge was reasonable,” making the question of whether Skanska’s plan was triggered largely beside the point. . In any event, the court’s interpretation of the plan was not clear error.

We reject, however, the applicability of the so-called “Flury factors” to a Rule 37(e) analysis. See (recognizing that this question was unresolved). Flury was decided before Rule 37(e) was written, and that opinion borrowed from Georgia spoliation law because of a lack of federal Circuit precedent on the precise issue in that case and because the parties and the district court had relied upon Georgia law. See .

In Skanska’s defense, the district court framed its inquiry into “bad faith” by applying a test that asked whether Skanska had committed an “affirmative act.” . But we may affirm the court’s ultimate finding of bad faith on any ground supported by the record. .