TJOFLAT, Circuit Judge

Margaret Diane Mathews, Richard H. Martin, Akerman, LLP, Tampa, FL, Dan L. Bagatell, Perkins Coie, LLP, Phoenix, AZ, Heather Ann Boice, Adam Marchuk, Michael R. Osterhoff, Mark T. Smith, Jade R. Lambert, Patrick M. Collins, Perkins Coie, LLP, Chicago, IL, Eric D. Miller, Perkins Coie, LLP, Seattle, WA, for Plaintiff–Counter, Defendant–Appellee.

Paul D. Clement, H. Christopher Bartolomucci, David Zachary Hudson, Bancroft, PLLC, Washington, DC, James R. Carroll, Christopher A. Lisy, Peter Simshauser, Skadden Arps Slate Meagher & Flom, LLP, Boston, MA, Lance A. Etcheverry, Michael K. Loucks, Nili T. Moghaddam, Skadden Arps Slate Meagher & Flom, LLP, Los Angeles, CA, Christopher M. Sacco, Carlton Fields Jorden Burt, PA, Kevin Napper, Tampa, FL, for Defendant–Counter, Claimant–Appellant.

Appeal from the United States District Court for the Middle District of Florida. D.C. Docket No. 8:11–cv–00775–SCB–TBM.

Before TJOFLAT, COX and SENTELLE, Circuit Judges.

Opinion

TJOFLAT, Circuit Judge:

In this three-year high-stakes litigation—a case that “[t]he [medical] industry is watching,” Ameritox assured the jury—the parties tried various state statutory and common law unfair competition claims predicated upon the alleged violation of two federal statutes that provide no private right of action. The parties, however, did not realize that it was an open question whether or not any of the nine states involved would permit the commandeering of state law to create a right of action where none existed. A district court sitting in diversity and faced with such a situation would have had no choice but to use whatever tools it had available to correctly divine and apply state law. See, e.g., Moreno v. Nationwide Ins. Co., 114 F.3d 168, 169 (11th Cir.1997) (answering a question of first impression under Alabama law only after certifying that question to the Alabama Supreme Court and having that court decline to answer the question). The court below, however, did not sit in diversity—or under any other circumstance that would have required it to answer the state-law questions. See, e.g., Price v. Time, Inc., 416 F.3d 1327, 1335 (11th Cir.2005) (answering a potentially dispositive state-law question before reaching a federal constitutional question). Rather, the District Court took supplemental jurisdiction over the state-law claims, see 28 U.S.C. § 1367(a), and accordingly could have dismissed those claims once their novel nature became apparent, see id. § 1367(c)(1). We hold that the District Court's decision to retain jurisdiction over novel and complex state-law claims hailing from nine different states—claims that the parties either did know or should have known were novel and complex—constituted an abuse of discretion.

I.

This litigation was extraordinarily complex. We recount only those facts relevant to our disposition.A.

Ameritox and Millennium are competitors in the drug-testing industry. Both sell physicians a variety of products and services that enable timely and accurate analysis of a patient's drug use. One such product is the point-of-care testing (“POCT”) cup. Appellant's Br. 2; Appellee's Br. 8. POCT cups perform the same primary function as standard urine specimen cups: enabling the safe and sanitary collection, storage, and transport of urine samples. Appellant's Br. 2–3; Appellee's Br. 8. POCT cups, however, provide an additional advantage. POCT cups contain chemically activated strips that indicate the presence of certain drugs in a patient's urine. Appellant's Br. 3; Appellee's Br. 8. Physicians can therefore use POCT cups to perform in-office (or “point-of-care”) “qualitative testing” to learn about their patient's drug use (or misuse). Appellant's Br. 3; Appellee's Br. 8. Though the POCT cups provide only limited information—they cannot, for example, reveal the exact quantity of a drug in a patient's system—some clinical information is better than no clinical information.

To verify the qualitative testing results and acquire more detailed patient information, physicians send patients' POCT cups to clinical laboratories for “confirmatory testing.” Appellant's Br. 3–4. Ameritox and Millennium compete for confirmatory-testing business. And, despite the need for this additional confirmatory testing, the timely information qualitative testing yields is sufficiently valuable that physicians have widely replaced standard specimen cups with pricier POCT cups. See Appellant's Br. 3; Appellee's Br. 8. It also does not hurt that physicians can bill Medicare and other insurers for conducting qualitative testing, such as for using a POCT cup instead of a specimen cup.

Millennium, a relative upstart in the drug-testing industry, saw in POCT cups an opportunity. In 2009, Millennium pioneered a program under which it would provide free POCT cups to physicians in exchange for a promise to send used cups to Millennium for confirmatory testing. Appellant's Br. 4; see Appellee's Br. 9. Interested physicians entered into a Free Cup Agreement (“FCA”) that memorialized this transaction and additionally barred physicians from billing the federal government for in-office, qualitative testing. Appellant's Br. 4; Appellee's Br. 9–10. In Millennium's view, this prohibition was essential to avoid violating two federal laws: the Stark Law (“Stark”), 42 U.S.C. § 1395nn(a), and the federal Anti–Kickback Statute (“AKS”), 42 U.S.C. § 1320a–7b(b). See Appellant's Br. 4; Appellee's Br. 9. Both statutes forbid entities like Millennium from offering any “remuneration” to a physician in exchange for referrals if payment for the referred services can be made under a federal healthcare program. Stark defines “remuneration” as “any remuneration, directly or indirectly, overtly or covertly, in cash or in kind.”See id. § 1395nn(h)(1)(B). The AKS definition is materially indistinguishable and includes “any remuneration (including any kickback, bribe, or rebate) directly or indirectly, overtly or covertly, in cash or in kind.” See id. § 1320a–7b(b). Millennium reasoned that the provision of free POCT cups in exchange for confirmatory-testing referrals did not constitute prohibited remuneration so long as doctors could not bill for the in-office qualitative testing. Appellant's Br. 4. If doctors could bill for qualitative testing, then the provision of free POCT cups would effectively be the provision of a paycheck; doctors could simply have a patient use the free POCT cup, qualitatively test the sample (by inspecting the cup's embedded chemically-activated strips), bill the government for the qualitative testing, and then ship the sample to Millennium for confirmatory testing.

Although neither Stark nor AKS provide private rights of action, see, e.g., U.S. ex rel. Drakeford v. Tuomey Healthcare Sys., Inc., 675 F.3d 394, 395 (4th Cir.2012) (“[T]he Stark Law does not create its own right of action....”); U.S. ex rel. Barrett v. Columbia/HCA Healthcare Corp., 251 F.Supp.2d 28, 37 (D.D.C.2003) (“There is no private right of action under the Anti–Kickback Act.”), the threat of civil and criminal sanctions compelled Millennium to seek expert counsel on the FCAs' legality. Appellant's Br. 4–5. Millennium thus presented its FCAs to a “nationally-recognized legal expert on clinical laboratory regulatory issues” and a law firm with “one of the nation's leading health care practices” for their blessing. Appellant's Br. 4. And, having received that blessing—that the provision of free POCT cups was legal so long as physicians entered into FCAs—Millennium rolled out its plan. While the majority of Millennium's customers continued to purchase POCT cups (and presumably continued to bill for qualitative testing), approximately ten percent entered into FCAs. Appellant's Br. 5.

B.

1.

This ten percent was too much for Ameritox. While Millennium grew, Ameritox shrank or stalled. Prior to Millennium's FCAs, Ameritox alleged that it “experienced record sales year after year.” Appellee's Br. 11. With the advent of the FCAs, Ameritox began experiencing losses in the POCT segment of its business. Appellee's Br. 11. Ameritox, however, did not experience corresponding losses in its non-POCT business. Appellee's Br. 11. Accordingly, Ameritox determined that the FCAs were responsible for its POCT segment losses.

Believing that the FCAs were illegal—and that its refusal to violate Stark or AKS simply to compete on a level-field with Millennium was costing it current and future customers—Ameritox filed a complaint (the “Florida complaint”) against Millennium in the District Court for the Middle District of Florida (the “District Court”) on April 8, 2011. Ameritox alleged that Millennium had “formed a business plan to increase its market share, revenue, and profits” by providing financial inducements and other kickbacks, in violation of both federal and state law. Ameritox argued that Millennium's FCAs, two other schemes not relevant here, and various “false or misleading statements in commercial advertisements for its services” violated the Lanham Act, 15 U.S.C. § 1125, et seq. ; violated Florida's Deceptive and Unfair Trade Practices Act (“FDUTPA”), Fla. Stat. § 501.201, et seq. ; and constituted unfair competition under Florida common law. The District Court exercised federal-question jurisdiction over the Lanham Act claim, see 28 U.S.C. § 1331, and supplemental jurisdiction over the remaining Florida claims, 28 U.S.C. § 1367(a).

Ameritox's complaint was short-lived. Following Millennium's motion to dismiss, Ameritox filed its First Amended Florida Complaint. And, during an additional year of wrangling between the parties, Ameritox filed a substantially similar complaint against Millennium in the District Court for the Southern District of California (the “California court”), amended that California complaint once, and filed the Second Amended Florida Complaint. Like it had done with the First Amended Florida Complaint, Ameritox then attempted to amend the California complaint for a second time. Its aim was “to clarify the scope of its claims,” thereby ensuring “that such claims are congruent with the claims set forth in the ... Second Amended [Florida] Complaint.” Additionally, Ameritox sought to add “new claims” to the First Amended California Complaint based upon law in New Hampshire, Texas, and Tennessee. The California court, however, did not allow the amendment and instead transferred the case to the Middle District of Florida. That court then consolidated it with the Second Amended Florida Complaint. Likely exasperated with Ameritox's inability to draft a comprehensible complaint, the District Court directed Ameritox “to file one[,] final amended complaint” by April 9, 2012, “which includes all counts and allegations it intends to raise in this matter,” concluding with the admonition that “[t]he Court will not permit further amendments without good cause. ”

Ameritox obliged and, on the deadline, filed its Third Amended Florida Complaint. That document bore only a superficial resemblance to either its Florida or California predecessors. Now, Ameritox sought to recover based on Millennium's violations of the Lanham Act, 15 U.S.C. § 1112, the FDUTPA, Fla. Stat. § 501.201, et seq. , the California Business and Professions Code, Cal. Bus. & Prof.Code § 17200, and the New Hampshire Consumer Protection Act, N.H.Rev.Stat. Ann. § 358–a, et seq. Ameritox also alleged that Millennium had committed the common-law torts of unfair competition and interference with business relationships in Arizona, California, Florida, New Hampshire, Texas, and Tennessee. Allegedly false advertising, the provision of free POCT cups, and other irrelevant schemes provided the foundation for these claims.

Notably, Ameritox's Third Amended Florida Complaint referenced the Stark Law only once and the Anti–Kickback statute twice. The only thing tying either federal statute to any of the state-law claims was a conclusory paragraph, virtually copied-and-pasted into the state-law counts:

As set forth in more detail above, Millennium engages in a pattern and practice of offering and providing inducements, kickbacks and other improper financial inducements to Health Care Providers and health care consumers in exchange for the referral of business, in violation of Federal and [state] law.

See, e.g., Plaintiff Ameritox Ltd.'s Third Amended Complaint for Injunctive Relief, Declaratory Relief and Damages at 35,

ECF No. 92. Ameritox did not allege that the existence of a Stark or AKS violation proved that Millennium had violated state law. But cf. Doe v. Fulton–DeKalb Hosp. Auth., 628 F.3d 1325, 1339 (11th Cir.2010) (noting that, under the doctrine of negligence per se, establishing the violation of a statute or regulation can establish that the defendant breached a duty as a matter of law, leaving the plaintiff to prove only causation and damages to make his claim). Rather, Ameritox appeared to allege that Millennium's conduct directly violated both state law, see, e.g., Plaintiff Ameritox Ltd.'s Third Amended Complaint for Injunctive Relief, Declaratory Relief and Damages at 44, ECF No. 92 (“Millennium implements these practices in New Hampshire.... Through these practices, Millennium engages in unconscionable, deceptive or unfair acts or practices in the conduct of commerce.” (emphasis added)), and federal law.

Shortly thereafter, Millennium answered Ameritox's Third Amended Florida Complaint and asserted various affirmative defenses and counterclaims. Millennium's counterclaims contained counts based upon: (I) the FDUTPA, Fla. Stat. § 501.201, et seq., (II)Cal. Bus. & Prof.Code § 17200, et seq., (III)Cal. Bus. & Prof.Code § 17000, et seq., (IV) the Texas Deceptive Trade Practices–Consumer Protection Act, Tex.Code Ann. Bus. & Com.Code § 17.41, et seq., (V) the New York Consumer Protection from Deceptive Acts and Practices Law, N.Y. Gen. Bus. Law § 349, (VI) common law unfair competition in Florida, Texas, and Washington, and (VII) common law tortious interference with business relationships in Florida, California, Texas, Tennessee, New York, Washington, and Oregon. Ameritox moved to dismiss counts II, III, and V. The Court granted Ameritox's motion to dismiss Count III and partially granted and partially denied the motion with respect to Count V. Once the dust had settled, the parties were ready for discovery, and the District Court was prepared to adjudicate one federal claim and a bevy of what appeared to be state statutory and tort claims.

2.

After an additional year of discovery, Ameritox filed a motion for partial summary judgment. In it, Ameritox contended that Millennium “violated the Stark Law and the anti-kickback statute.” This was a curious claim, given that Ameritox did not assert any claims against Millennium predicated upon those laws. Indeed, it could not have done so had it wanted to, for neither law creates a private right of action. See supra Part I.A. Ameritox explained its strategy: demonstrating that Millennium had violated Stark and AKS would prove, as a matter of law, “part of Claims I–VII of its Third Amended Complaint.” See Fed.R.Civ.P. 56(a) (“A party may move for summary judgment, identifying each claim or defense—or the

part of each claim or defense—on which summary judgment is sought.... The court shall grant summary judgment if ... the movant is entitled to judgment as a matter of law.” (emphasis added)). Which part? The part where “Ameritox must show that Millennium engaged in unfair, deceptive, or unconscionable business practices.”

Ameritox had apparently changed course. One would have assumed from the operative complaint that Ameritox intended to prove its state-law claims with sole reference to Millennium's conduct and state-law precedent. Now, it was clear that Ameritox hoped to prove its state-law claims by establishing violations of independent federal healthcare statutes. Supporting Ameritox's novel theory—that proving a violation of either Stark or AKS was legally sufficient to prove the central element of fourteen different causes of action, three of which involved state statutes, and ten of which involved state common law torts —was Millennium Labs., Inc. v. Universal Oral Fluid Labs., Inc., No. 8:11–cv–1757, at 6–7 (M.D.Fla. Apr. 25, 2012), ECF No. 72, an unpublished opinion from a district court in the Middle District of Florida.

Rather than challenge Ameritox's Stark/AKS state-law incorporation theory, Millennium responded to Ameritox's motion for partial summary judgment by disputing that any Stark or AKS violation had occurred. Like Millennium, the District Court accepted at face value Ameritox's Stark/AKS incorporation theory. Indeed, the court merely flagged Ameritox's theory as it proceeded to rule on the motion:

Ameritox contends that the provision of free POCT cups violates th[e] Anti–Kickback Statute (“AKS”) and/or the Stark Law.... If such conduct is found to violate the AKS or Stark Law, then Ameritox contends that such conduct provides the basis for liability under Ameritox's Lanham Act and unfair competition claims in Counts I through VII.

Indeed, the Court later explained that “Ameritox is using the alleged violation of the Stark Law and AKS as the basis for its Lanham Act and unfair competition claims.”

Regardless, the District Court obliged Ameritox in its efforts to get a ruling on the relevance of Stark and AKS. It concluded that “to the extent that the doctors could not bill for the [qualitative] testing done using a POCT cup for other reasons (such as, because they were using chemical analyzers to test the specimens), then Millennium's provision of free POCT cups” constituted remuneration under Stark. When it came to doctors who did not bill for POCT cup qualitative testing because they had agreed not to do so pursuant to an FCA—the issue at the heart of case—the District Court stated the following:

To the extent that the doctors could bill for the [qualitative] testing done using a POCT cup but decline to do so solely because they agreed to Millennium's requirement not to bill for the [qualitative] testing, the Court finds that a genuine issue of material fact exists regarding whether the provision of free POCT cups in this scenario constitutes remuneration. Instead, in such a situation, it appears that the doctors are giving up the ability to bill for [qualitative] testing, which is giving up the opportunity to net approximately $15 per specimen. Thus, whether the free POCT cups constitute remuneration in this scenario must be determined by the jury.

That is, the District Court proposed giving the factfinder an undisputed factual scenario—a physician accepts a free cup from Millennium in exchange for a promise not to bill for POCT testing and a promise to refer confirmatory testing to Millennium—and allowing it to determine whether, as a matter of law, those facts constituted “remuneration.” With the entry of that order, the case proceeded on.

3.

Beyond its bevy of state-law claims, Ameritox also asserted a Lanham Act claim against Millennium based upon allegedly false representations. Over the course of this litigation, Millennium slowly whittled away at this claim until it vanished entirely. After Millennium successfully moved for summary judgment on the damages portion of Ameritox's claim, the District Court realized that “the viability of [Ameritox's Lanham Act] claim is questionable.” The court therefore informed the parties that if diversity jurisdiction did not exist and the Lanham Act claim was resolved prior to trial, the court “could decline to exercise supplemental jurisdiction over the remaining claims in this case.” See 28 U.S.C. § 1367(c)(3) (permitting district courts to decline to exercise supplemental jurisdiction over state-law claims when “the district court has dismissed all claims over which it has original jurisdiction”).

Thus, the Court directed Ameritox—a limited partnership—to provide a list of all of its general partners to determine whether diversity jurisdiction existed and could provide jurisdiction for the state-law claims beyond 28 U.S.C. § 1367(a). In response, Ameritox conceded that diversity jurisdiction was lacking, but argued that the court should retain jurisdiction over the state-law claims even if the Lanham Act claim were dismissed. For its part, Millennium argued that the court should relinquish supplemental jurisdiction for two reasons. First, it continued to argue that the Lanham Act claim should be dismissed and that the state-law claims should go with it pursuant to 28 U.S.C. § 1367(c)(3). Second, it argued that, without regard to the Lanham Act claim, the District Court should dismiss the state-law claims both because they were novel and complex, see id. § 1367(c)(1), and because they substantially predominated over the Lanham Act claim, see id. § 1367(c)(2).

Shortly after briefing the question of supplemental jurisdiction, Millennium offered to enter into a consent decree enjoining it from representing that “offering chemical analyzers at below-market-value is legal.” Millennium hoped that the consent decree would moot Ameritox's Lanham Act claim, thereby removing the only federal claim from the case. The District Court thereafter asked Ameritox to explain why a consent decree would be an insufficient remedy. In the same order—and without any explanation or analysis—the court proclaimed that it “will not decline to exercise supplemental jurisdiction if the Lanham Act claims are removed from trial.” With that order went the last clear chance to avert the spectacle of trial. Two days before trial, the court entered a consent decree, mooting Ameritox's Lanham Act claim. Thus, the stage was set for a trial based entirely upon hypothetical legal theories.

C.

The trial began on June 2, 2014 and lasted for ten days. The jury considered Ameritox's eleven remaining claims and Millennium's ten counterclaims. Once both sides had rested, and prior to closing argument, the court instructed the jury on Stark and AKS:

Both parties have asserted claims that are based, in part, on the contention that the other party violated the Stark Law and the Anti–Kickback statute. If you determine that either party has violated the Stark Law or the Anti–Kickback Statute, such a violation may constitute unfair competition, an unfair trade practice, and/or tortious interference with business relationships under particular state laws. Accordingly, before the Court instructs you on the specific law underlying the parties' claims, the Court will first instruct you on the Stark Law and the Anti–Kickback Statute....

And, although the court did go into specifics on Stark and AKS—including the incorrect statement that whether Millennium's provision of free POCT cups pursuant to FCAs constituted remuneration under Stark and AKS was somehow a question of fact for the jury —the court

provided no additional information linking Stark or AKS to the state-law claims.

That link, however, did come. In its closing argument following the instructions, Ameritox explained that, to prevail on its unfair competition claims, it would need to demonstrate that (1) “Millennium is a competitor of Ameritox,” (2) “Millennium engaged in deceptive or fraudulent conduct contrary to honest practices in commercial matters,” (3) and that these actions caused damages. Stark and AKS related to the second element: “This is where the remuneration question comes in. If you find that this is a Stark violation, the point-of-care test cups constituted prohibited remuneration, we've satisfied number two.” Regarding the tortious interference claims, Ameritox explained that it had to prove: (1) Ameritox had a preexisting business relationship with a doctor in a particular state, (2) Millennium knew about that relationship, (3) “Millennium has to intentionally and unjustifiably interfere with that relationship,” and (4) Millennium caused Ameritox damages. Again, Stark and AKS came to the rescue: “If we can establish that this is prohibited remuneration, we've satisfied element number three.”

Millennium made no objection and the District Court did not intervene; presumably both believed that Ameritox had accurately stated the law. Millennium responded in closing simply by contesting whether any Stark or AKS violations had occurred. It did not seriously press any of its ten counterclaims, beyond suggesting that Ameritox employees had repeatedly violated company policy. It did, however, state that “Ameritox behaves in ways that are expressly prohibited by Stark,” and that “[t]here won't be any doubt about that.”

Given that the parties' closing arguments revolved almost entirely upon whether the FCAs violated Stark and AKS, it was fitting that the verdict form prefaced both Ameritox and Millennium's state-law claims with “Special Interrogatories to the Jury” concerning Stark and AKS. Those interrogatories included nonsensical questions like “[d]o you find by a preponderance of the evidence” that Millennium's provision of free POCT cups pursuant to FCAs “is remuneration” under Stark or AKS. Faced with such confusing instructions, the jury nonetheless persevered and completed its civic duty. It found that the FCAs violated both Stark and AKS, and that, in combination with Ameritox's separate proof of damages and causation, Ameritox had proven: (1) tortious interference in Florida; (2) unfair competition in Florida; (3) tortious interference in Tennessee; (4) tortious interference in Texas; and (5) unfair competition in Texas. Conversely, the jury found that Ameritox had not violated Stark, AKS, or any other law.

In all, the jury awarded Ameritox $14,755,000. After trial, Millennium renewed the motion for judgment as a matter of law it had made in earlier. It also moved for a new trial and to set aside or reduce the award of punitive damages. The District Court denied Millennium's renewed motion for judgment as a matter of law and granted in part and denied in part its motion for a new trial or reduction in the award of punitive damages; upon reconsideration, the District Court reduced the award of punitive damages.

Millennium appealed, raising four primary arguments. First, it argued that the District Court abused its discretion when it decided to retain jurisdiction over the supplemental state-law claims. Second, it argued that the court misapplied Stark and AKS, an error which necessitated either judgment for Millennium or a new trial. Third, it argued that Ameritox had not sufficiently proven damages or causation for any of its claims. Finally, it argued that the District Court's refusal to limit the exemplary damages awarded under Texas law was error. Because we hold that the District Court's decision to retain supplemental jurisdiction over the state-law claims constituted an abuse of discretion, we do not reach Millennium's final three arguments.

II.

The doctrine of supplemental jurisdiction—sometimes referred to as “pendent jurisdiction”—permits “federal courts to decide certain state-law claims involved in cases raising federal questions” when doing so would promote judicial economy and procedural convenience. Carnegie–Mellon Univ. v. Cohill, 484 U.S. 343, 348–49, 108 S.Ct. 614, 618, 98 L.Ed.2d 720 (1988). Now codified at 28 U.S.C. § 1367, the modern doctrine of supplemental jurisdiction derives from United Mine Workers of America v. Gibbs, 383 U.S. 715, 86 S.Ct. 1130, 16 L.Ed.2d 218 (1966), and Carnegie–Mellon University v. Cohill. See City of Chicago v. Int'l Coll. of Surgeons, 522 U.S. 156, 173, 118 S.Ct. 523, 533, 139 L.Ed.2d 525 (1997) (explaining that “[t]he supplemental jurisdiction statute codifies [the] principles” outlined in Gibbs and Cohill ); Parker v. Scrap Metal Processors, Inc., 468 F.3d 733, 742 (11th Cir.2006) (noting that 28 U.S.C. § 1367 codified Gibbs's holding). Those cases continue to guide federal courts as they determine whether to retain or relinquish jurisdiction over supplemental state-law claims. See City of Chicago, 522 U.S. at 172–73, 118 S.Ct. at 533–34 (noting that § 1367 “reflects the understanding that, when deciding whether to exercise supplemental jurisdiction,” courts should look to the principles outlined in Gibbs and Cohill ); Lucero v. Trosch, 121 F.3d 591, 598 (11th Cir.1997) (noting that District Courts “weigh the competing interests set forth in § 1367(c) and Gibbs in deciding whether it is appropriate to exercise supplemental jurisdiction”). “As articulated by Gibbs, the doctrine of pendent jurisdiction ... is a doctrine of flexibility, designed to allow courts to deal with cases involving pendent claims in the manner that most sensibly accommodates a range of concerns and values.” Cohill, 484 U.S. at 350, 108 S.Ct. at 619.

Gibbs itself concerned an inter-union dispute between the United Mine Workers and the Southern Labor Union. Gibbs, 383 U.S. at 718, 86 S.Ct. at 1135. Gibbs brought suit against United Mine Workers in the District Court for the Eastern District of Tennessee based upon alleged violations of § 303 of the Labor Management Relations Act of 1947 and a Tennessee common-law claim for tortious interference with two contracts; jurisdiction for the common-law claims was based upon pendent jurisdiction. Id. at 720, 86 S.Ct. at 1135–36. Although a jury found for Gibbs on both his federal and state claims, the court thereafter held that Gibbs's Labor Management Relations Act claim—the claim that permitted the exercise of pendent jurisdiction over the state claims—was not cognizable. Id., 86 S.Ct. at 1136. The Sixth Circuit affirmed, and the Supreme Court granted certiorari. Id. at 721, 86 S.Ct. at 1136.

The Court began its analysis by examining the propriety of the District Court's exercise of jurisdiction over the state-law claims. Id., 86 S.Ct. at 1136. The Court noted that a federal court's power to hear pendent state-law claims exists wherever a distinct claim arises under federal law, and “the relationship between that [distinct] claim and the state claim permits the conclusion that the entire action before the court comprises but one constitutional ‘case.’ ” Id. at 725, 86 S.Ct. at 1138. The Court, however, immediately cautioned courts against using this power “in every case in which it is found to exist.” Id. at 726, 86 S.Ct. at 1139. It explained that pendent jurisdiction is “a doctrine of discretion, not of plaintiff's right,” and should be exercised only when doing so would promote “judicial economy, convenience and fairness to litigants.” Id., 86 S.Ct. at 1139.

Indeed, it admonished courts that “[n]eedless decisions of state law should be avoided both as a matter of comity and to promote justice between the parties by procuring for them a surer-footed reading of applicable law.” Id., 86 S.Ct. at 1139. One such needless decision would occur when a court decides issues of state law after the federal claims are dismissed before trial; in that situation, “the state claims should be dismissed as well.” Id., 86 S.Ct. at 1139. Another occurs where it appears that the “state claim constitutes the real body of a case, to which the federal claim is only an appendage.” Id. at 727, 86 S.Ct. at 1140. The Court also explained that “the issue whether pendent jurisdiction has been properly assumed is one which remains open throughout the litigation.” Id., 86 S.Ct. at 1139. Indeed, when a litigant is “well aware of the nature of his proofs and the relative importance of his claims,” and a court later discovers this, the court need not “tolerate a litigant's effort to impose upon it what is in effect only a state law case.” Id. at 727, 86 S.Ct. at 1139–40.

As previously noted, Congress codified Gibbs in 28 U.S.C. § 1367. Parker, 468 F.3d at 742. Like Gibbs and Cohill , § 1367 grants federal courts the power to exercise jurisdiction over claims “that are so related to claims in the action within such original jurisdiction that they form part of the same case or controversy under Article III of the United States Constitution.” 28 U.S.C. § 1367(a) ; Lucero, 121 F.3d at 597. Unlike those cases, however, § 1367 mandates that district courts—at least initially—exercise jurisdiction over those supplemental claims that satisfy the case or controversy requirement. See Parker, 468 F.3d at 743 (“Whenever a federal court has supplemental jurisdiction under section 1367(a), that jurisdiction should be exercised unless section 1367(b) or (c) applies.”). Additionally, the doctrine of supplemental jurisdiction under § 1367 is less flexible than it was under Gibbs or Cohill insofar as it permits dismissal of supplemental claims only under particular circumstances. District courts only possess the authority to dismiss claims brought under § 1367(a) if:

(1) the claim raises a novel or complex issue of State law,

(2) the claim substantially predominates over the claim or claims over which the district court has original jurisdiction,

(3) the district court has dismissed all claims over which it has original jurisdiction, or

(4) in exceptional circumstances, there are other compelling reasons for declining jurisdiction.

28 U.S.C. § 1367(c) ; Parker, 468 F.3d at 743 (“Any one of the section 1367(c) factors is sufficient to give the district court discretion to dismiss a case's supplemental state law claims.”). Once any of these factors is satisfied, the district court possesses the discretion to dismiss supplemental claims and must “weigh ... at every stage of the litigation,” whether to dismiss the supplemental claims. See City of Chicago, 522 U.S. at 173, 118 S.Ct. at 534 (quotation marks omitted). Actually determining whether to dismiss the claims calls for the court to weigh the “host of factors” outlined in Gibbs and Cohill: “judicial economy, convenience, fairness, and comity.” Id., 118 S.Ct. at 534.

Against this backdrop, we evaluate first whether the District Court possessed the discretion to dismiss the supplemental claims, and if it did, whether retaining jurisdiction was an abuse of discretion. See Lucero, 121 F.3d at 598. “A district court abuses its discretion if it applies an incorrect legal standard, follows improper procedures in making the determination, or makes findings of fact that are clearly erroneous.” Klay v. United Healthgroup, Inc., 376 F.3d 1092, 1096 (11th Cir.2004) (quotation marks omitted); see also Estate of Amergi ex. rel. Amergi v. Palestinian Auth., 611 F.3d 1350, 1365 (11th Cir.2010) (“A district court does not abuse its discretion when it has a range of choices and the court's choice does not constitute a clear error of judgment.” (quotation marks omitted)).

A.

Millennium's abuse of discretion argument is a nonstarter unless one of the § 1367(c) factors applies. Millennium principally argues that § 1367(c)(3) was animated following the District Court's resolution of Ameritox's Lanham Act claim. Appellant's Br. 19–20 (“Once Ameritox's lone federal claim was definitely resolved before trial, the court should have dismissed the case.”). That provision allows a district court to dismiss supplemental claims after it “dismiss[es] all claims over which it has original jurisdiction.” 28 U.S.C. § 1367(c)(3). Ameritox counters that the District Court did not “dismiss” the Lanham Act claim. Appellee's Br. 26–27. Rather, that claim was “resolved” in Ameritox's favor following Millennium's decision to enter into a consent decree prohibiting certain objectionable conduct. Appellee's Br. 26. Accordingly, Ameritox argues, § 1367(c)(3) is inapposite.

We need not resolve this question of statutory interpretation because the District Court possessed discretion to decline to exercise jurisdiction over the supplemental claims on another basis. Ameritox's state-law claims raise “a novel or complex issue of State law.” See 28 U.S.C. § 1367(c)(1). It may be true that, “[g]enerally, state tort claims are not considered novel or complex.” Appellee's Br. 33 (quotation marks omitted) (quoting Parker, 468 F.3d at 743 ). This, however, is an extraordinary case predicated upon extraordinary legal theories. When Ameritox filed its partial motion for summary judgment, it effectively amended (for a fourth time) its complaint. There, Ameritox sought partial summary judgment that Millennium “violated the Stark Law and the anti-kickback statute.” This finding would purportedly “show that Millennium engaged in unfair, deceptive, or unconscionable business practices,” thus satisfying the central element of all of Ameritox's state-law claims.

That is, rather than proving the sine qua non element of its claims by comparing Millennium's conduct to conduct a state court had already declared to be tortious (or unlawful, in the case of the statutory claims), Ameritox sought to prove this element by demonstrating that Millennium had violated an independent federal statute. Ameritox necessarily travelled on the theory that the supreme courts of Arizona, California, Florida, New Hampshire, Tennessee, and Texas had incorporated (or obviously would incorporate) Stark and AKS into their respective unfair competition statutes and common-law torts. See LeFrere v. Quezada, 582 F.3d 1260, 1263–64 (11th Cir.2009) (“In addressing issues of state law, we are bound by the decisions of the state supreme court. Only where no state court has decided the point in issue may a federal court make an educated guess as to how that state's supreme court would rule.” (citation omitted) (quotation marks omitted)).

Ameritox did not cite to any state-law precedent to support this point. One explanation for Ameritox's total failure to support its Stark/AKS incorporation theory with binding citations is that no such support exists. No Arizona or New Hampshire cases mention either Stark or AKS. No reported Florida decisions consider whether a Stark or AKS violation suffices to prove any element of Florida's Deceptive and Unfair Trade Practices Act, Florida common law unfair competition, or Florida tortious interference with business relationships. And there are no California or Tennessee cases connecting either Stark or AKS to tortious interference with business relationships.

This brings us to Texas. In Texas, a case does exist connecting AKS to unfair-competition torts. In Reliable Ambulance Service, Inc. v. Mercy Hospital of Laredo, No. 04–02–00188–CV, 2003 WL 21972724 (Tex.App. Aug. 20, 2003), Reliable alleged that Mercy committed the “generic tort of wrongful competition” based upon an alleged AKS violation. 2003 WL 21972724, at *6 n. 1 (quotation marks omitted). After reasoning that “recognizing a common law cause of action by a competitor for damages and injunctive relief for violation of the anti-kickback statute would be inconsistent with the intent of Congress,” the court rejected Reliable's efforts to co-opt state common law into providing a private right of action where Congress saw fit to establish none. 2003 WL 21972724, at *6. The only case anchoring Stark or AKS to state laws explicitly rejects using an AKS violation as the basis for a common law unfair competition tort.

To be fair, Ameritox attempted to provide some support for its theory that proof of a Stark Law or AKS violation was sufficient to prove that Millennium engaged in unconscionable, unfair, or deceptive acts or practices (for the statutory and common law unfair competition claims) or unjustifiable interference with Ameritox's business relationships (for the tortious interference claims). Although it did not cite state-law decisions evidencing incorporation of Stark or AKS, Ameritox pointed to a solitary unpublished case from the Middle District of Florida: Millennium Laboratories, Inc. v. Universal Oral Fluid Laboratories, LLC, No. 8:11–cv–1757 (M.D.Fla. Aug. 16, 2013), ECF No. 136. This case crumples under the staggering burden Ameritox places upon it. In Universal Oral Fluid Laboratories, Millennium brought various unfair competition and tortious-interference claims against Universal Oral Fluid, arguing that “Universal engaged in deceptive and unfair trade practices because [contested business] agreements violate the federal Anti–Kickback Statute and the Stark Law.” Universal, at *6.Millennium raised precisely the same theories that Ameritox raises here. In response, Universal argued that Millennium's claims failed because they relied upon “allegations that [Universal's conduct] violate[d] the Anti–Kickback Statute and the Stark Law, and [Millennium] does not have a private right of action under those Statutes.” Id., at *6. Universal also argued that Millennium did not “allege[ ] any actionable conduct independent of the ... allegations that their conduct violates the Anti–Kickback Statute and the Stark law.” Id., at *6.

The court rejected Universal's arguments, explaining—as it had earlier in response to Universal's motion to dismiss—that

case law provides that conduct violating the Anti-kickback Statute and the Stark Law may provide the basis for liability under recognized common law causes of action and other state statutory laws provided all the elements of the causes of action are met.

Id., at *7.

A district court's unremarkable observation that a single course of conduct might give rise to multiple claims provides the entire support for Ameritox's Stark and AKS incorporation theory. As far as Ameritox and Millennium are concerned, this proposition boils down to the following tautology: to prove an unfair competition claim, you must prove every element of that claim as required by established state law. Stark and AKS are simply irrelevant. This much is evident from the court's next statement that Millennium “is entitled to summary judgment in this case if it can establish that Universal's conduct, although in violation of the Anti–Kickback Statute and the [s]tark law, also violates state statutory or common law.”

Indeed, the court was even clearer on this point when ruling on Universal's earlier motion to dismiss. After surveying other decisions regarding the (in)applicability of Stark and AKS to state unfair competition claims, the court arrived at Synthes (U.S.A.) v. Globus Medical, Inc., No. Civ.A. 04–cv–1235, 2005 WL 2233441 (E.D.Pa. Sept. 14, 2005). See Millennium Labs., Inc. v. Universal Oral Fluid Labs., LLC, No. 8:11–cv–1757, at 12 (M.D.Fla. Apr. 25, 2012), ECF No. 72. The court's discussion went as follows:

[T]he Synthes Court, citing Reliable Ambulance, held that the fact the Anti–Kickback Statute prohibited the plaintiff's conduct does not mean the same conduct could not provide a basis for civil liability under another state or federal statute. The critical question for the court in Synthes was whether the underlying conduct also violated state law. The court thus proceeded to analyze each of the defendant's common law causes of action (tortious interference with prospective contractual relationships, unfair competition, civil conspiracy, and violations of the Lanham Act) under the applicable Pennsylvania State law. This Court agrees with the reasoning of Synthes as foreshadowed by Reliable Ambulance. Plaintiff may bring an action for unfair competition in this case provided all of the elements of common law unfair competition are present under the applicable state laws.

Id. at 12–13. The message is clear: the court would allow Millennium's claims to survive so long as, beneath their Stark and AKS packaging, the claims were valid with sole reference to the relevant state statutes and precedent. Of course, this does not mean that state-court decisions could not or did not accept proof of a Stark or AKS violation as a means of proving deception and interference for unfair competition and tortious interference claims, respectively. It just meant that the court would need proof of such an adoption, and in the absence of proof, the claims could still survive so long as the challenged conduct fell within some extant unfair competition or tortious interference cause of action.

Simply put, whether or not Stark or AKS are relevant to any of the claims in this case is a novel question. It is also a complex question, laden with important policy choices. See generally Reliable Ambulance Serv., 2003 WL 21972724. Yes, Ameritox does claim that “[i]n reality, there were no highly complex issues of state law” and that “[t]hese were garden-variety state tort claims.” Appellee's Br. 33. That may have been true before Ameritox's decision to rely upon independent federal statutes to prove its claims. Regardless, whether or not unfair competition and tortious interference claims are novel and complex in the abstract is irrelevant. The novel and complex question here is whether or not these states should hold that a Stark or AKS violation proves part of an unfair competition or tortious interference claim. Indeed, Ameritox's own appellate brief effectively concedes the complexity point. When arguing that equitable factors supported the District Court's decision to retain supplemental jurisdiction, Ameritox noted that dismissal would waste judicial resources because it would require state courts “to master not only the facts but also predicate federal-law issues under the Stark Law and Anti–Kickback Statute.” Appellee's Br. 31. However, deciding whether or not Stark or AKS violations are relevant to any state cause of action necessarily requires state courts first to “master” those same federal issues, and second, to determine whether or not enforcing those laws via state statutes and common law would advance the state's unique interests and policy goals.

In the end, the reality is likely that none of these states—besides Texas—appears to have even considered whether or not a Stark or AKS violation could prove that Millennium “engaged in unfair, deceptive, or unconscionable business practices.” We say with certainty that each one of Ameritox's state-law claims presented the District Court with novel and complex issues of state law. Accordingly, from the moment Ameritox filed its motion for partial summary judgment, thereby making manifest its actual legal theories, the District Court possessed the discretion to dismiss Ameritox's supplemental state-law claims.

B.

Having determined that the District Court possessed the power to either retain or reject supplemental jurisdiction over Ameritox's state-law claims, we must now determine whether the decision to retain jurisdiction was an abuse of discretion. The Supreme Court's teachings in Gibbs, guide us in this endeavor. “Where § 1367(c) applies, considerations of judicial economy, convenience, fairness, and comity may influence the court's discretion to exercise supplemental jurisdiction.” Baggett v. First Nat'l Bank of Gainesville, 117 F.3d 1342, 1353 (11th Cir.1997) ; see Palmer v. Hosp. Auth., 22 F.3d 1559, 1569 (11th Cir.1994) (holding that when “one or more [§ 1367(c) ] factors is present, the additional Gibbs considerations may, by their presence or absence, influence the court in its decision concerning the exercise” of supplemental jurisdiction). Furthermore, once a district court possesses discretion to dismiss the supplemental claims, it must be continuously mindful regarding whether or not the factors favor dismissal. City of Chicago, 522 U.S. at 173, 118 S.Ct. at 534 (holding that “a federal court should consider and weigh in each case, and at every stage of the litigation” the Gibbs and Cohill values to determine whether it is proper to retain jurisdiction over state claims (quotation marks omitted)). Ameritox argues that each of the Gibbs factors—judicial economy, convenience, fairness, and comity—weigh in favor of retaining jurisdiction.

We first consider judicial economy. Two separate district courts have already been involved in this dispute, see supra Part I.B.1., and the Middle District alone invested over three years' time resolving these issues. That court's docket boasts nearly seven hundred entries, about three hundred of which were entered prior to Ameritox's motion for partial summary judgment. Additionally, the litigation spawned a ten-day jury trial. Dismissal of the state-law claims at this point would render much of the efforts below a ity. Accordingly, there can be no doubt whatsoever that Ameritox's complaint ultimately required and received a tremendous amount of attention from the federal courts. Furthermore, dismissal at this stage would potentially continue to consume resources, as it would also require the parties to re-litigate in multiple state courts. For these reasons, Ameritox argues that factors of judicial economy weigh in favor of retaining jurisdiction. If we were to hold that the District Court abused its discretion, the argument goes, we would effectively waste judicial resources.

This myopic view overlooks the fact that from at least January of 2014—five months prior to trial—the onus for this waste lies squarely with Ameritox. From that point on, Ameritox consumed scarce judicial resources so that it could try nonexistent claims. That the case did go to a jury means that Ameritox additionally wasted each of the jurors' time. Furthermore, there is no telling how far in advance of filing its complaint Ameritox knew it was going to rely entirely upon Stark and AKS to prove all of its state-law claims. Ameritox filed its motion for partial summary judgment on the deadline for dispositive motions as set forth in the Second Amended Case Management and Scheduling Order; presumably it did not develop its legal theories that very day.

Here, one of two things occurred. First, Ameritox's attorneys did not know at the time they filed the Third Amended Florida Complaint that proof of its state-law claims would rely upon Stark or AKS. If that were true, it meant that these attorneys bumbled along for up to a year and a half trying to figure out what claims they had before finally settling upon non-existent legal theories. In other words, Ameritox's attorneys wasted both judicial resources and Ameritox's money.

The second possibility is that, when they filed the Third Amended Florida Complaint, Ameritox's attorneys did know that their case would rely entirely upon Stark and AKS. If Ameritox's attorneys intentionally omitted its Stark and AKS theory from the complaint so as to obscure their theories and prolong litigation, then we need not look beyond Gibbs for a reason to dismiss the supplemental claims. Gibbs unequivocally condemns such opportunistic behavior when discussing the appropriate response to litigants who knowingly conceal the dominance of their state-law claims as a means of ensuring that courts exercise supplemental jurisdiction over those claims:

The issue whether pendent jurisdiction has been properly assumed is one which remains open throughout the litigation. Pretrial procedures or even trial itself may reveal a substantial hegemony of state law claims, or likelihood of jury confusion, which could not have been anticipated at the pleading stage. Although it will of course be appropriate to take account in this circumstance of the already completed course of the litigation, dismissal of the state claim might even then be merited. For example, it may appear that the plaintiff was well aware of the nature of his proofs and the relative importance of his claims; recognition of a federal court's wide latitude to decide ancillary questions of state law does not imply that it must tolerate a litigant's effort to impose upon it what is in effect only a state law case.

383 U.S. at 727, 86 S.Ct. at 1139–40. The Court explicitly condoned dismissal of state-law claims after trial if it appeared that a party concealed or omitted facts that, if known, might have resulted in dismissal of the supplemental claims. The point is that, under either scenario—Ameritox knew it would use Stark and AKS versus Ameritox did not know it would use Stark and AKS—the onus for the waste of judicial resources lies with Ameritox. And, if Ameritox purposefully concealed the novel and complex nature of its claims, that alone might merit dismissal of the supplemental claims; the District Court should not have “tolerate[d] [Ameritox's] effort to impose upon it what is in effect” made-up state-law claims. Id., 86 S.Ct. at 1140.

Despite this, Ameritox would essentially have us hold that once a court has poured sufficient resources into a case, we should hesitate before writing off that investment. If we hold that considerations of judicial economy favor retaining jurisdiction, we would provide litigants with perverse incentives to sandbag their own cases in the hope that courts spend enough resources to make decisions to exercise supplemental jurisdiction effectively unreviewable. See Parker, 972 F.2d at 587 (“[A] court cannot obtain jurisdiction over a case merely by trying it; otherwise, its decision to retain jurisdiction would be, effectively, unreviewable.”). This we cannot tolerate. As our sister circuit put it, “[e]valuating considerations of judicial efficiency and duplication of judicial effort is not just a matter of toting up months or motions or the page counts of judicial orders.” RWJ Mgmt. Co. v. BP Prods. N. Am., Inc., 672 F.3d 476, 481 (7th Cir.2012). Ameritox's argument does not take into account that the parties are responsible for waste. Dismissing the supplemental claims at this point would not waste resources; the waste occurred long ago. Ameritox advanced outlandish legal theories and Millennium went along with it. Either party could have halted this farce long ago. Under these circumstances, we find that considerations of judicial economy do not favor retaining jurisdiction over the state-law claims.

By contrast, considerations of convenience favor retaining jurisdiction. It is true that, as far as the parties are concerned, it would be most convenient to try every claim in a single forum. Accordingly, considerations of convenience weigh in favor of retaining jurisdiction over the supplemental state-law claims.

Considerations of fairness, however, do not support continued exercise of jurisdiction. Ameritox has only itself to blame for the problems inherent in its Stark and AKS theory. No one forced Ameritox to introduce novel and complex legal theories. And, as other circuits have pointed out, every litigant who brings supplemental claims in court knowingly risks the dismissal of those claims. See, e.g., Ani v. Panikkar, 200 F.3d 189, 203 (3d Cir.1999) (noting that it was not unfair to dismiss supplemental claims because “Ani and his lawyers knowingly risked dismissal of [Ani's] pendent claims when they filed suit in federal district court and invoked the Court's discretionary supplemental jurisdiction power”); Pitchell v. Callan, 13 F.3d 545, 549 (2d Cir.1994) (noting that declining to exercise jurisdiction was not an abuse in part because “[w]hen Pitchell brought his state-law claims in federal court, he must have realized that the jurisdiction he invoked was pendent and possibly tentative”).

To the extent our decision benefits Millennium, it does so only by eliminating a judgment patently unsupported by actual law. Furthermore, Millennium has certainly suffered insofar as it massively wasted resources arguing that it did not violate Stark or AKS when it could have moved to dismiss Ameritox's various theories for failure to state a claim. See Fed.R.Civ.P. 12(h)(2) (“Failure to state a claim upon which relief can be granted .... may be raised ... by a motion under Rule 12(c) ; or ... at trial.”). At the end of the day, the attorneys here left their clients with gargantuan legal bills and “a tale/ Told by an idiot, full of sound and fury,/ Signifying nothing.” William Shakespeare, Macbeth, act 5, sc. 5.

Well, perhaps they have something. Both parties are free to use evidence obtained during discovery to pursue their state-law claims in a proper forum. See Ani, 200 F.3d at 203 (noting that discovery obtained prior to dismissal of supplemental claims may be utilized in later suits). This is not a case where retrial will necessarily impose extraordinary burdens on the litigants. See Redondo Constr. Corp. v. Izquierdo, 662 F.3d 42, 49–50 (1st Cir.2011) (holding that exercising supplemental jurisdiction was proper where the parties had conducted extensive discovery in English and the need to translate all documents and testimony into Spanish to retry the case in Puerto Rico's courts would “impose huge burdens,” and the state-law claims were neither novel nor complex). Under these circumstances, fairness concerns do not weigh in favor of retaining jurisdiction.

This brings us to comity. It is a bedrock principle that “[n]eedless decisions of state law should be avoided both as a matter of comity and to promote justice between the parties, by procuring for them a surer-footed reading of applicable law.” Gibbs, 383 U.S at 726, 86 S.Ct. at 1139. “State courts, not federal courts, should be the final arbiters of state law” in our federalist system. Baggett, 117 F.3d at 1353 ; see Hardy v. Birmingham Bd. of Educ., 954 F.2d 1546, 1553 (11th Cir.1992) (noting that “[s]tate courts, not federal courts, are the final expositors of state law” in its decision not to exercise pendent jurisdiction). It is for this reason that our usual procedure when confronted with a case that requires the resolution of a novel or complex question of state law, we certify such questions to a State's highest court. See, e.g., Tyne v. Time Warner Entm't Co., 336 F.3d 1286, 1291 (11th Cir.2003) (“Substantial doubt about a question of a state law upon which a particular case turns should be resolved by certifying the question to the state supreme court.” (quotation marks omitted)). Federal courts are (and should be) loath to wade into uncharted waters of state law, and should only do so when absolutely necessary to the disposition of a case. See, e.g., Price v. Time, Inc., 416 F.3d 1327, 1334 (11th Cir.2005) (expressing disappointment that a state Supreme Court had declined to answer a certified question, thereby leaving “the district court with no choice but to answer the state law question itself”).

Allowing the judgment below to stand would be to ratify the District Court's (perhaps inadvertent) pronouncement that nine separate states—Arizona, California, Florida, New Hampshire, New York, Oregon, Tennessee, Texas, and Washington—accept proof of a Stark or AKS violation as sufficient to prove the sine qua non element of various statutory and common-law unfair competition laws. It would directly contravene Gibbs's bedrock teaching that courts should not unnecessarily decide questions of state law. See Gibbs, 383 U.S. at 726, 86 S.Ct. at 1139. And, while a needless decision of law in one state is bad, creating law in nine states is beyond the pale. Affirming the decision would be an open invitation to litigants everywhere to find a way into the Eleventh Circuit to take advantage of almost certainly erroneous constructions of state law—at least, until the relevant states are forced to action to prevent such mischief. See Arizonans for Official English v. Arizona, 520 U.S. 43, 79, 117 S.Ct. 1055, 1074, 137 L.Ed.2d 170 (1997) (noting that a “federal tribunal risks friction-generating error when it endeavors to construe a novel state Act not yet reviewed by the State's highest court”); see also Mosher v. Speedstar Div. of AMCA Int'l, Inc., 52 F.3d 913, 917 (11th Cir.1995) (“[S]tate law is what the state supreme court says it is, and a state supreme court's pronouncements are binding on every state and federal judge.” (quoting Sultenfuss v. Snow, 35 F.3d 1494, 1504 (11th Cir.1994) (en banc) (Carnes, J., dissenting))). This invitation would ultimately tax the judiciary's resources, requiring our courts to consider claims that state courts may decide are not cognizable. It would also signal disrespect towards our state-court brethren who stand ready and able to tackle novel issues. Allowing the judgment below to stand would reward the parties for what is either misplaced tenacity or blissful ignorance of the law and would needlessly damage the judiciary, both state and federal.Convenience is the only factor weighing in favor of retaining jurisdiction. Concerns regarding judicial economy, fairness, and comity favor dismissal. However, we do not review the District Court's decision de novo. That these considerations would have lead us to relinquish jurisdiction does not mean the District Court necessarily abused its discretion when it decided (without explanation) to retain jurisdiction. Rather, it is the fact that the District Court effectively rewarded either ignorant or duplicitous parties by ratifying new legal theories in nine different states that leads us to the conclusion that the District Court abused its discretion. Allowing the state-law claims to be tried using the Stark/AKS theory was egregious, constituting a clear error of judgment. See Estate of Amergi ex. rel. Amergi, 611 F.3d at 1365 (“A district court does not abuse its discretion when it has a range of choices and the court's choice does not constitute a clear error of judgment.” (quotation marks omitted)).

III.

The doctrine of supplemental jurisdiction is “designed to enable courts to handle cases involving state-law claims in the way that will best accommodate the values of economy, convenience, fairness, and comity” and to allow “the Judicial Branch ... to shape and apply the doctrine in that light.” Carnegie–Mellon Univ. v. Cohill, 484 U.S. 343, 351, 108 S.Ct. 614, 619, 98 L.Ed.2d 720 (1988). The District Court's decision here to retain jurisdiction did not further these goals. Rather, it resulted in the needless creation of new law for nine states and permitted parties that were either ignorant of the law or disingenuous to waste scarce judicial resources. We conclude that the District Court's decision to retain supplemental jurisdiction over the state-law claims was an abuse of discretion.

We also explicitly note that we express no opinion whatsoever as to whether the FCAs violate Stark, AKS, or some other state law. We also express no view as to whether it would be wise to allow Stark and AKS violations to prove, as a matter of law, elements of state unfair competition law. It certainly does not require any great leap of logic to believe that a company that profits by declining to comply with government regulations enjoys an unfair advantage vis-à-vis its competitors who choose to obey the law. We hold only that exercising supplemental jurisdiction under these (perhaps sui generis ) circumstances was an abuse of discretion.

We VACATE the judgment below and REMAND with instructions to dismiss the state-law claims without prejudice so that the parties can litigate their claims in a proper forum.

SO ORDERED.