ZAGEL, District Judge.

The United States District Court for the Southern District of Indiana (the "district court") The Honorable James Zagel, United States District Court for the Northern District of Illinois, Eastern Division, sitting by designation. dismissed the claims of Tommy L. Morris, personal representative of the Estate of Thomas Lynn Morris (the "Estate") against Salvatore Nuzzo ("Nuzzo"). The lawsuit, originally filed in the Trumbull County Common Pleas Court of the State of Ohio, was removed to the United States District Court for the Northern District of Ohio, and finally was transferred to the district court in Indianapolis. Because we find that the district court erred in its determination that Nuzzo was fraudulently joined, we will vacate the judgment of dismissal and remand to the district court with instructions that the case be further remanded to the Trumbull County Common Pleas Court of Ohio.

Background

This matter arises out of an automobile accident and a subsequent insurance dispute. On December 2, 2004, Daemon Sampson ("Sampson") was operating a vehicle involved in a collision in Brown County, Indiana. One of the passengers in Sampson's vehicle, Thomas Lynn Morris, died as a result of injuries sustained in the accident. At the time of the collision, Sampson was insured under a Mid-Century auto insurance policy issued to his mother. Shortly after the accident, the Estate made a claim under the policy for $50,000, the highest allowable amount. Nuzzo, a citizen of Ohio, was the claims adjustor assigned to handle the Estate's claim.

The Estate was unable to settle its claim under the Mid-Century policy. As a result, it filed a wrongful death suit in Indiana state court against Sampson, which Mid-Century defended. The case went to trial and the jury returned a verdict for the Estate and against Sampson of about $1.2 million. Following the verdict, Sampson executed an assignment of his rights against Mid-Century to the Estate in exchange for an agreement that the Estate would not pursue collection of the verdict against Sampson personally.

In or around February 2011, the Estate filed a claim in California state court against Mid-Century, alleging that its bad faith failure to pay out the Estate's insurance claim resulted in the excess jury verdict against Sampson. In May 2011, the California court dismissed the case on forum non conveniens grounds. Although Mid-Century is incorporated and has its principal place of business in California, the California state court found that the suit belonged in Indiana given that the Estate and Sampson were citizens of Indiana, the insurance policy was issued and performed in Indiana, and the underlying death trial was conducted in Indiana.

Rather than file suit in Indiana, the Estate brought claims against Mid-Century and Nuzzo in Ohio state court, alleging tortious bad faith failure to pay an insurance claim and breach of contract. Mid-Century and Nuzzo removed the case to the federal district court for the Northern District of Ohio, and the Estate immediately moved to remand under the "forum defendant rule" based on Nuzzo's Ohio citizenship. See 28 U.S.C. § 1441(b)(2). Mid-Century and Nuzzo opposed remand on the grounds that Nuzzo had been fraudulently joined to the lawsuit and his citizenship could therefore be disregarded in determining the propriety of removal. Mid-Century and Nuzzo also moved to dismiss the claim against Nuzzo or, alternatively, to transfer the case to the Southern District of Indiana pursuant to 28 U.S.C. § 1404. The district court for the Northern District of Ohio granted the motion to transfer the case to the district court in Indiana and denied all other pending motions, including the Estate's motion to remand to the Ohio state court, as moot.

Once in the district court, the Estate argued that the Northern District of Ohio lacked subject matter jurisdiction due to the forum defendant removal defect, which rendered the § 1404(a) transfer invalid. Alternatively, the Estate argued that if the transfer was valid, it could still seek remand. Mid-Century and Nuzzo claimed that the transfer did moot the Estate's remand motion, and renewed its contention that Nuzzo was fraudulently joined and should be dismissed from the case.

The district court found that diversity jurisdiction was properly exercised because the requirements of 28 U.S.C. § 1332(a) were met, and because the alleged forum defendant rule violation was a procedural error that does not affect jurisdiction. So the § 1404(a) transfer was valid, but the Estate's motion to remand was not foreclosed. In seeking remand, the Estate argued that the fraudulent joinder doctrine did not apply to Nuzzo because his presence did not compromise the parties' complete diversity — it simply prevented removal under the forum defendant rule. The Estate also argued that, in any event, Nuzzo was not fraudulently joined because the claims against him stood a "reasonable possibility of success" under Ohio law. See Poulos v. Naas Foods, 959 F.2d 69, 73 (7th Cir.1992).

While acknowledging that "most fraudulent joinder cases involve a defendant who is non-diverse to the Plaintiff and who the removing parties contend was joined to defeat diversity," the district court could find "no principled basis" for refusing to extend the doctrine to a diverse resident defendant joined for purposes of triggering the forum defendant rule. The district court agreed that the Estate's claims against Nuzzo were potentially viable under Ohio law, but determined that Indiana law governed both claims because Indiana had the most significant relationship to the events underlying the claims. See generally Restatement (Second) of Conflict of Laws (1973). Under Indiana law, the district court concluded, the claims against Nuzzo were not cognizable and stood no chance of success. Accordingly, Nuzzo was fraudulently joined. The district court dismissed all claims against Nuzzo and denied the Estate's motion to remand. This appeal followed.

Analysis

On appeal, the Estate contends: (1) that the district court lacked diversity jurisdiction over the case because its removal from Ohio state court was prohibited under the forum defendant rule, (2) that the district court erred in applying the fraudulent joinder doctrine to a diverse resident defendant, and (3) that the district court erred in making a choice of law determination as part of the fraudulent joinder analysis. We review issues involving removal of an action from state to federal court de novo. See Schwartz v. State Farm Mut. Auto. Ins. Co., 174 F.3d 875, 878 (7th Cir.1999).

I

The first question before us is whether the district court exercised proper diversity jurisdiction over this case. A defendant removing a case on diversity grounds must not only demonstrate that the case satisfies the requirements of 28 U.S.C. § 1332(a), but must also clear the "additional hurdle" of 28 U.S.C. § 1441(b)(2), or the "forum defendant rule." Hurley v. Motor Coach Industries, Inc., 222 F.3d 377, 378 (7th Cir.2000).

Section 1441(b)(2) provides:

The forum defendant rule is "designed to preserve the plaintiff's choice of forum, under circumstances where it is arguably less urgent to provide a federal forum to prevent prejudice against an out-of-state party." Hurley, 222 F.3d at 380. In other words, the forum defendant rule disallows federal removal premised on diversity in cases where the primary rationale for diversity jurisdiction — to protect defendants against presumed bias of local courts — is not a concern because at least one defendant is a citizen of the forum state. See Dresser Indus., Inc. v. Underwriters at Lloyd's of London, 106 F.3d 494, 499 (3d Cir.1997) ("If diversity jurisdiction exists because of a fear that the state tribunal would be prejudiced towards the out-of-state plaintiff or defendant, that concern is understandably allayed when the party is joined with a citizen from the forum state."); Lively v. Wild Oats Mkts., Inc., 456 F.3d 933, 940 (9th Cir.2006).

The Estate's argument that the district court lacked diversity jurisdiction over this case because its removal violated the forum defendant rule is wrong. The argument ignores our decision in Hurley v. Motor Coach Industries, Inc., where we joined "the longstanding line of authority that holds that the forum defendant rule is not jurisdictional." 222 F.3d at 379 (listing cases). See also Holmstrom v. Peterson, 492 F.3d 833, 836 (7th Cir.2007). Neither side disputes that there is complete diversity between the parties — the Estate is alleged to be a citizen of Indiana, Nuzzo is alleged to be a citizen of Ohio, and Mid-Century is alleged to be incorporated and have its principle place of business in California — and the amount in controversy exceeds $75,000. Like the plaintiff in Hurley, there is no question that if the Estate's "case had been filed in the first instance in federal court, jurisdiction under § 1332 would have been clear." 222 F.3d at 380. Whether the case was properly removed to federal court, therefore, is a matter of removal procedure, not jurisdiction. Id.

The Estate attempts to distinguish Hurley on the grounds that it filed a timely motion to remand and thus did not waive its right to invoke the forum defendant rule. See 28 U.S.C. § 1447(c). But the fact that the Estate's motion to remand was timely does not change the nonjurisdictional nature of the forum defendant rule — it simply means that the district court had to address the motion on its merits, which it did. Indeed the Estate's acknowledgment that it could have waived its objection is a tacit admission that the forum defendant rule is nonjurisdictional. See Hurley, 222 F.3d at 379 ("Of course, waiver is possible only if the forum defendant rule is nonjurisdictional; true jurisdictional flaws are nonwaivable and can be raised at any time." (internal citation omitted)).

More to the point, subject matter jurisdiction is not the threshold issue in this unusual case. The district court for the Northern District of Ohio did not need to determine its own subject matter jurisdiction prior to transferring the case to the Southern District of Indiana. See In re LimitNone, LLC, 551 F.3d 572, 576 (7th Cir.2008) (a district court is not required to determine its own subject matter jurisdiction prior to transferring a case under 28 U.S.C. § 1404(a)). And once the case arrived in the Southern District of Indiana, the district court was allowed to "assume" subject matter jurisdiction for purposes of conducting the fraudulent joinder analysis. Schur v. L.A. Weight Loss Centers, Inc., 577 F.3d 752, 763 (7th Cir.2009). The real question here is whether the district court erred in applying the fraudulent joinder doctrine to dismiss Nuzzo, a diverse resident defendant.

II

Under the fraudulent joinder doctrine (the "doctrine"), an out-of-state defendant's right of removal premised on diversity cannot be defeated by joinder of a nondiverse defendant against whom the plaintiff's claim has "no chance of success." Poulos v. Naas Foods, Inc., 959 F.2d 69, 73 (7th Cir.1992); see also Walton v. Bayer Corp., 643 F.3d 994 (7th Cir.2011); Schur v. L.A. Weight Loss Centers, Inc., 577 F.3d 752 (7th Cir.2009); Schwartz v. State Farm Mut. Auto. Ins. Co., 174 F.3d 875 (7th Cir.1999); Gottlieb v. Westin Hotel Company, 990 F.2d 323 (7th Cir.1993). The doctrine is designed to "strike a reasonable balance among the policies to permit plaintiffs the tactical prerogatives to select the forum and the defendants they wish to sue, but not to reward abusive pleading by plaintiffs, and to protect the defendants' statutory right to remove." 14B Wright, Miller, Cooper & Steinman, § 3723 pp. 788-93; see also Poulos, 959 F.2d at 73; Schur, 577 F.3d at 763 ("A plaintiff typically may choose its own forum, but it may not join a nondiverse defendant simply to destroy diversity jurisdiction").

To establish fraudulent joinder, a removing defendant "must show that, after resolving all issues of fact and law in favor of the plaintiff, the plaintiff cannot establish a cause of action against the in-state defendant." Poulos, 959 F.2d at 73 (emphasis in original). If the removing defendant can meet this "heavy burden," Poulos, 959 F.2d at 73, the federal district court considering removal may "disregard, for jurisdictional purposes, the citizenship of certain nondiverse defendants, assume jurisdiction over a case, dismiss the nondiverse defendants, and thereby retain jurisdiction." Schur, 577 F.3d at 763. Because the district court may "disregard" the nondiverse defendant, we have described the fraudulent joinder doctrine as an "exception" to the requirement of complete diversity. See Walton, 643 F.3d at 999.

The question before us is whether the district court erred in applying the fraudulent joinder doctrine to Nuzzo, whose presence triggers the forum defendant rule but does not compromise the parties' complete diversity. In other words, we are asked to determine whether the fraudulent joinder doctrine creates an exception to the forum defendant rule. It does not appear that any court of appeals has answered this question. The few district courts that have decided the issue are split. See, e.g., Yellen v. Teledne Continental Motors, Inc., 832 F.Supp.2d 490 (E.D.Pa. Dec. 6, 2011) (fraudulent joinder doctrine applies to a diverse forum defendant); Sargent v. Cassens Corp., No. 06 CV 1042, 2007 WL 1673289 (S.D. Ill. June 7, 2007) (same); but see Yount v. Shashek, 472 F.Supp.2d 1055 (S.D.Ill.2006) (fraudulent joinder doctrine does not apply to a diverse forum defendant); Davenport v. Toyota Motor Sales, No. 09 CV 532, 2009 WL 4923994 (S.D.Ill. Dec. 14, 2009) (same).

The district court's view that the doctrine extends to diverse resident defendants was based largely on language from one of the earliest Supreme Court cases to address fraudulent joinder, Wilson v. Republic Iron & Steel Co., 257 U.S. 92, 42 S.Ct. 35, 66 L.Ed. 144 (1921). The plaintiff in Wilson was an Alabama citizen who filed a negligence suit against his employer, a citizen of New Jersey, and a coemployee, a citizen of Alabama, after sustaining an injury on the job. The employer removed the case to federal district court based on diversity of citizenship, arguing that the co-employee had nothing to do with the plaintiff's injury and was joined solely for purposes of obstructing the employer's right of removal. Id. at 94, 42 S.Ct. at 36. The plaintiff moved to remand on the grounds that the district court lacked diversity jurisdiction. The Alabama district court denied the motion to remand, finding that the co-employee did not belong in the lawsuit. The plaintiff obtained a direct writ of error to the Supreme Court to review the decision to dismiss the co-employee and retain jurisdiction. The Supreme Court upheld the decision, stating:

Wilson, 257 U.S. at 97, 42 S.Ct. at 37 (citation omitted).

Although the co-employee in Wilson was a nondiverse defendant, the district court determined that the above passage applied to Nuzzo. Specifically, the district court interpreted the Supreme Court's use of the term "resident defendant" strictly in terms of the defendant's relationship to the forum state, and not as a reference to his shared citizenship with the plaintiff. Because Nuzzo is a "resident" of Ohio, the district court reasoned, Wilson holds that his fraudulent joinder could not be used to defeat Mid-Century's right of removal, regardless of the fact that his presence did not compromise the parties' complete diversity. In quoting from Wilson, however, the district court did not include the clause "if not a resident of the state in which the case is brought."

The district court also relied on our decision in Poulos v. Naas Foods, Inc., 959 F.2d 69 (7th Cir.1992), the first case in which we addressed the fraudulent joinder doctrine. Like Wilson, Poulos involved an alleged fraudulent joinder of a nondiverse resident defendant, and makes no mention of the forum defendant rule. The district court interpreted the following passage from Poulos as supporting the doctrine's extension to a diverse resident defendant:

Id. at 73. The district court found that this rationale "applies equally to a fraudulently joined forum defendant."

We are unwilling to rely on Wilson and Poulos to extend the doctrine to the forum defendant rule. As mentioned, both cases involved an alleged fraudulent joinder of a nondiverse resident defendant and therefore shed little light upon the question before us. While it is true that the Supreme Court's use of the term "resident defendant" in Wilson could arguably encompass diverse resident defendants, we do not believe that was the Court's intended meaning given its unqualified reference to the forum defendant rule in the same passage. Further, it does not appear that any federal court in the country had considered whether the doctrine could apply to a diverse resident defendant at the time Wilson was decided — we doubt the Court had this rather uncommon scenario on its radar. Whatever ambiguity Wilson might raise, we do not agree that our explanation in Poulos of the doctrine's rationale applies with equal force to diverse resident defendants.

We also do not accept the proposition that the district court could "identify no principled basis for refusing to apply the principles of fraudulent joinder" to a diverse resident defendant. The party seeking removal bears the burden of proving the propriety of removal; doubts regarding removal are resolved in favor of the plaintiff's choice of forum in state court. See, e.g., Schur, 577 F.3d at 758; Boyd v. Phoenix Funding Corp., 366 F.3d 524, 529 (7th Cir.2004). The better question to ask is: what principled basis does exist to extend the fraudulent joinder doctrine to the forum defendant rule? Nuzzo does not offer substantive argument on this critical question — his briefings consist mostly of out-of-context quotations and conclusory statements.1

It seems to us that extending the fraudulent joinder doctrine to diverse resident defendants would constitute a nontrivial expansion of the removal right. To offer a first appellate resolution of a question not often considered even in district courts in a case in which the briefs are, at best, unhelpful, would be unwise, especially when another clear ground exists for a final disposition of this appeal. Accordingly, we will lay out the policy factors we deem to be most relevant in considering whether to extend the doctrine, but ultimately do not decide the issue.

The fraudulent joinder doctrine is designed to strike a "reasonable balance" between competing policy interests. See 14B Wright, Miller, Cooper & Steinman, § 3723 pp. 788-93. At one end of the scale is the plaintiff's right to select the forum and the defendants, as well as the general interest in confining federal jurisdiction to its appropriate limits. Id. At the other end of the scale is the defendant's statutory right of removal, and associated interest in guarding the removal right against abusive pleading practices. Id. To determine whether the fraudulent joinder doctrine ought to extend to diverse resident defendants, it is necessary to consider how these interests balance out in the context of the forum defendant rule.

We begin with the right of removal. When an out-of-state defendant's right of removal is destroyed by the presence of a diverse resident co-defendant there is no reason to "presume bias" on the part of the local courts in favor of an in-state plaintiff because, by definition, there are no in-state plaintiffs. Cf. Poulos, 959 F.2d at 73. The out-of-state defendant, therefore, does not "need access to federal court" in the same way we described in Poulos, because the case can proceed in state court in only one of two ways. Id. First, the resident co-defendant could remain in the case alongside the out-of-state defendant, in which case any local bias would run against the out-of-state plaintiff. Alternatively, if the claims against the local defendant truly are meritless, the state court will, presumably, dismiss that defendant and there will be no local parties on either side of the lawsuit, a scenario in which the local court should be neutral. Absent any threat of local bias to the out-of-state defendant, federal courts arguably have a diminished interest in protecting the removal right against abusive pleading tactics designed to trigger the forum defendant rule.

On the other hand, the actual right of removal is not limited to situations involving a possible risk of local bias. An out-of-state defendant may remove regardless of whether a suit has been brought in the plaintiff's home state so long as there is complete diversity and no resident co-defendants. In other words, we do not give automatic deference to a plaintiff's choice of state forum simply because the plaintiff has filed suit outside of his or her home state. Nor have we limited application of the fraudulent joinder doctrine to cases in which the local bias rationale is implicated. In fact, Poulos itself involved an out-of-state plaintiff who was found to have engaged in fraudulent joinder.

So even though a plaintiff can never secure home-court advantage by joining a diverse resident defendant, limiting the doctrine to nondiverse defendants could lead to some troubling inconsistencies and potential loopholes. Consider the following example: if a plaintiff from State A ("Plaintiff A") sues a defendant from State A ("Defendant A") and a defendant from State C ("Defendant C") and does so in state court in State B ("Court B"), Defendant C can remove the case and argue to the federal district court that Defendant A was fraudulently joined.2 It is difficult to explain why we should not allow Defendant C to do the same if, instead of blocking removal by joining Defendant A, Plaintiff A joins a defendant from State B ("Defendant B") to trigger the forum defendant rule. If it is Defendant C's right of removal that concerns us, and that right is equally frustrated in both scenarios, why should federal courts police against one potential abusive pleading tactic but not the other? Concern about local bias in favor of Plaintiff A cannot explain the discrepancy, because Plaintiff A is out of state in both scenarios. Nor is it satisfactory to invoke the rationale behind the forum defendant rule itself — that we need not protect Defendant B "from the prejudices of its own local courts." See Davenport, 2009 WL 4923994, at *3. That is of course true, but it is Defendant C's right of removal that is at issue. If the claim against Defendant B is truly meritless, Court B will presumably dismiss Defendant B, and then Plaintiff A will face Defendant C in Court B — a situation in which Defendant C would have been entitled to remove but for the wrongful joinder of Defendant B!

This, we believe, is what the district court had in mind when it cited to Wecker v. National Enameling & Stamping Co., 204 U.S. 176, 27 S.Ct. 184, 51 L.Ed. 430 (1907). Like Wilson and Poulos, Wecker involved an alleged fraudulent joinder of a nondiverse resident defendant. In explaining the competing policy interests that underlie the fraudulent joinder doctrine, however, the Supreme Court used language that we find to be of greater relevance to this case:

Id. at 186, 27 S.Ct. at 188. As our example above demonstrates, a plaintiff could potentially use the forum defendant rule as a "device" to defeat removal where an out-of-state defendant would otherwise have that right. Extending the doctrine to preclude this conduct, therefore, seems consistent with Wecker's directive that federal courts vigilantly protect the removal right against abusive pleading practices.

Despite the logical inconsistency, we are reluctant to expand the fraudulent joinder doctrine absent a better understanding of the need to do so. Such a move would be in tension with long-established precedent that the removal statutes are to be strictly construed to preserve the limited jurisdiction of federal courts. See Shamrock Oil & Gas Corp. v. Sheets, 313 U.S. 100, 108-09, 61 S.Ct. 868, 872, 85 L.Ed. 1214 (1941) ("The power reserved to the states under the Constitution to provide for the determination of controversies in their courts" calls for narrow construction of removal statutes); Healy v. Ratta, 292 U.S. 263, 270, 54 S.Ct. 700, 78 L.Ed. 1248 (1934) ("Due regard for the rightful independence of state governments ... requires that [federal courts] scrupulously confine their own jurisdiction to the precise limits which the statute has defined"); Syngenta Crop Protection, Inc. v. Henson, 537 U.S. 28, 32, 123 S.Ct. 366, 369-70, 154 L.Ed.2d 368 (2002) ("[The] statutory procedures for removal are to be strictly construed."). Further, it might well substantially increase the number of removal petitions filed in federal court, which would stall the administration of justice at both the state and federal levels as district courts engage in what can often be complex "act[s] of prediction" regarding the viability of a plaintiff's state law claims. Poulos, 959 F.2d at 74. In short, the costs of expanding the doctrine could far outweigh the benefits of policing against what appears to be an exceptionally rare abusive pleading tactic.3

Ultimately, we think it a very close question whether the fraudulent joinder doctrine ought to extend to diverse resident defendants, and we are reluctant to rule definitively on the issue today absent a more thorough and more able presentation of the relevant balance of interests described above. In any event, we are convinced that Nuzzo was not fraudulently joined.

III

For purposes of reaching the choice of law question we assume, without deciding, that the fraudulent joinder doctrine does apply to diverse resident defendants.4 In finding that Nuzzo was fraudulently joined, the district court rested on its determination that Indiana law governed the claims against Nuzzo, even while acknowledging that the claims were potentially viable under Ohio law. We must decide whether this choice of law determination exceeded the bounds of the fraudulent joinder analysis. The question of whether, or to what extent, a federal district court can make choice of law determinations in conducting a fraudulent joinder analysis appears to be a question of first impression for this court.

The district court agreed with the parties that the Estate's claims against Nuzzo had a reasonable possibility of success under Ohio law but not under Indiana law, and assumed it could reach a decision over which law to apply.5 The Estate argues that this assumption was incorrect because a district court cannot engage in choice of law decisions without first establishing diversity jurisdiction. See Abels v. State Farm Fire & Casualty Company, 770 F.2d 26, 33 n. 10 (3d Cir.1985).6 But the fraudulent joinder analysis allows district courts to "assume" limited jurisdiction over an otherwise nonremovable action to consider the viability of claims against an alleged fraudulently joined defendant. Schur, 577 F.3d at 763. We interpret the Estate to argue that choice of law decisions necessarily exceed the scope of this analysis. Poulos, 959 F.2d at 73.

District courts may not be absolutely precluded from considering choice of law questions that may arise in the fraudulent joinder context. In Poulos, we described the fraudulent joinder analysis as "an act of prediction" to determine whether there is "any reasonable possibility" that a state court would rule against the alleged fraudulently joined defendant. 959 F.2d at 73. In making this determination, the district court must necessarily predict what substantive law the state court would apply. If the parties dispute what law governs, therefore, the district court must engage in some type of choice of law decision. If district courts were powerless to do so, plaintiffs could potentially circumvent the fraudulent joinder doctrine by identifying any jurisdiction in the United States in which its claim against the alleged fraudulently joined defendant stood a reasonable possibility of success, even if the jurisdiction bore absolutely no relation to the case. That would substantially undermine the purpose of the doctrine.

We hold that choice of law decisions can be made as part of the fraudulent joinder analysis where the choice of law decision is dispositive to the outcome, and where the removing defendant bears the same "heavy burden" to make the choice of law showing. Poulos, 959 F.2d at 73. A choice of law decision is dispositive to the fraudulent joinder analysis when the plaintiff and the removing defendant disagree over the substantive law that should govern the claim against the alleged fraudulently joined defendant, and where the district court determines that the claim stands a reasonable possibility of success under the plaintiff's suggested choice of law but not under the removing defendant's. In that case, the removing defendant can demonstrate fraudulent joinder only by showing that, after resolving all issues of fact and law in favor of the plaintiff, there is no reasonable possibility that the state court would apply the plaintiff's suggested choice of law.7 Id.

This should be a difficult showing for the defendant to make. If the federal court considering removal determines that the plaintiff could satisfy even one applicable choice of law factor, it should end the analysis there and remand the case unless that one factor is so attenuated, and so obviously outweighed by the other relevant factors, that there is no reasonable possibility that the state court would rely upon it to apply the plaintiff's suggested choice of law. If the district court determines that the plaintiff could satisfy more than one applicable choice of law factor, the district court should necessarily find against the removing defendant and remand.

So the district court did not err by making a choice of law determination. The error was that it treated the choice of law question as if it was deciding it directly, rather than trying to predict whether there was any reasonable possibility the Ohio state court would decide the question against Nuzzo. Applying the proper standard, we find, at the very least, there is a reasonable possibility the state court would have ruled against Nuzzo and applied Ohio law to the Estate's claim of tortious bad faith.

IV

The district court's choice of law analysis consisted of three steps. First, the district court determined that Ohio choice-of-law rules should apply because the case had been transferred from the Northern District of Ohio under 28 U.S.C. § 1404(a). This was the right conclusion perhaps reached for the wrong reason. In support of its decision to apply Ohio choice-of-law rules, the district court cited Edwardsville Nat'l Bank and Trust Co. v. Marion Laboratories, Inc., 808 F.2d 648, 650 (7th Cir.1987) ("A transfer under § 1404(a) changes venue but not law; the transferee court must apply the transferor's choice-of-law rules"). Unlike the instant case, however, Edwardsville Nat'l Bank involved a diversity suit filed directly in federal district court in Illinois and later transferred to the Southern District of Indiana under § 1404(a). This case is very different because the district court was examining a fraudulent joinder allegation — it was engaging in an "act of prediction" over how the state court would resolve the choice of law dispute. Poulos, 959 F.2d at 73. Thus, to determine what choice-of-law rules to apply, the proper question for the district court to ask was what choice-of-law rules the Ohio state court would apply, not what choice-of-law rules federal law required.

Next, the district court correctly determined that, when confronted with a choice-of-law question regarding a tort claim, Ohio courts apply a presumption "that the law of the place of injury controls unless another jurisdiction has a more significant relationship to the lawsuit." Morgan v. Biro Mfg. Co., Inc., 15 Ohio St.3d 339, 474 N.E.2d 286, 289 (1984). It also rightly noted that, to determine the state with the most significant relationship, Ohio courts consider the general principles set forth in § 145 of the Restatement (Second) of Conflict of Laws (1971) (the "Restatement"), which include: (1) the place of the injury; (2) the place where the conduct causing the injury occurred; (3) the domicile, residence, nationality, place of incorporation, and place of business of the parties; and (4) the place where the relationship between the parties, if any, is located. However, the district court did not consider that Ohio courts also apply the more generalized principles listed in § 6 of the Restatement (which § 145 itself references), namely: (a) the needs of the interstate and international systems; (b) the relevant policies of the forum; (c) the relevant policies of other states and the relative interests of those states in the determination of the particular issue; (d) the protection of justified expectations; (e) the basic policies underlying the particular field of law; (f) certainty, predictability and uniformity of result; and (g) ease in the determination and application of the law to be applied. The Restatement at § 6; Morgan, 474 N.E.2d at 289.

Finally, weighing the § 145 factors against the facts of this case, the district court determined "with no doubt" that Indiana law applied to the tortious bad faith claim:

There are several difficulties with the final step in the district court's choice-of-law analysis. First, it erroneously considered "the injuries involved in the car accident" as somehow tied to the bad faith claim. The excess jury verdict was the only relevant injury the court should have weighed — the alleged bad faith failure to pay the insurance claim did not cause any injuries suffered in the car accident.

Second, the district court's determination that the conduct causing the excess jury verdict was "centered in Indiana" is not supported by the record. Resolving all issues of fact in favor of the Estate, the conduct that caused the injury was Nuzzo's failure to "properly investigate, adjust and settle the claims against the insured." The district court had before it a signed declaration from Nuzzo stating that "[a]ll decisions and determinations I made in the adjustment of the Estate of Morris claim were made in Burg Hill, Ohio." Thus, the conduct causing the injury appears to have emanated from Ohio, where the Estate's claim was processed, not Indiana.

Third, the district court did not consider that the tort of bad faith refusal to pay an insurance claim is arguably designed more to deter wrongful conduct than to compensate for injury. See generally Hoskins v. Aetna Life Ins. Co., 6 Ohio St.3d 272, 452 N.E.2d 1315, 1321 (1983) (punitive damages available under bad faith insurance claim "deters refusals on the part of insurers to pay valid claims where the refusals are both unjustified and in bad faith"). The Restatement (Second) of Conflict of Laws, which the Ohio Supreme Court has adopted in its entirety, see Morgan, 474 N.E.2d at 288-89, states that "[i]f the primary purpose of the tort rule involved is to deter or punish misconduct ... the state where the conduct took place may be the state of dominant interest and thus that of most significant relationship." The Restatement at § 145 cmt. c. In our view, this alone is sufficient to satisfy the "any reasonable possibility" test.

Fourth, the district court did not weigh any of the choice of law principles laid out in the Restatement at § 6. Several of these principles, including the relevant policies of the forum, weigh in favor of applying Ohio law. The district court's own interpretation of Ohio law reflects a policy interest in favor of regulating the conduct of individual insurance agents, while Indiana law does not. See Schwartz, 174 F.3d at 878-79 (bad faith insurance claim against individual employee who handled the claim is not cognizable under Indiana law). Given that Nuzzo was operating out of Ohio, there is a reasonable possibility the Ohio state court would have seized on this difference in state policy and applied Ohio law to the bad faith claim.

Based on the above, we believe that there was more than a reasonable possibility that the Ohio state court would have decided against Nuzzo and applied Ohio law to the Estate's bad faith failure to settle claim. Thus, regardless of what law the Ohio state court would have ultimately applied to the breach of contract claim, Nuzzo was not fraudulently joined and his presence prevented removal under 28 U.S.C. § 1441(b)(2). Because the Estate timely objected to this procedural defect in removal, it has a right to remand.

CONCLUSION

For the reasons set forth above, we will VACATE the judgment of the district court and REMAND with instructions that this case be further remanded to the Trumbull County Common Pleas Court of Ohio.