The threshold question presented by this interlocutory appeal, 28 U.S.C. § 1292(b), is whether Indiana's or Pennsylvania's choice-of-law rules govern a suit against the United States by the estates of passengers killed in the crash of a small private aircraft. Two of the decedents lived in Pennsylvania, another worked in Pennsylvania, the plane was hangared in Pennsylvania, and the corporation that owned it was incorporated in Pennsylvania. The plane crash, which occurred during a landing at a Kentucky airport in inclement weather, resulted from: (1) the Government's publication in Washington, D.C. of a chart incorrectly showing that a long-inactive instrument landing approach at the airport was active; and (2) the negligence of Indiana-based federal air traffic controllers in relying on that chart, in neglecting to monitor the radar during the flight's landing approach, and in failing to respond to the pilot's last-minute radio communications.

In a multistate tort action, the Federal Tort Claims Act ("FTCA") requires a federal court to apply the whole law of the place where the acts of negligence occurred, including its choice-of-law rules. 28 U.S.C. §§ 1346(b), 2674; Richards v. United States, 369 U.S. 1, 82 S. Ct. 585, 7 L. Ed. 2d 492 (1962). In a pretrial ruling, the District Court determined that because acts of negligence were committed in both Indiana and the District of Columbia ("D.C."), those jurisdictions' choice-of-law rules were candidates to govern the case. In order to determine whether it was necessary to choose between the two, the Court first had to determine whether there was a true conflict between their choice-of-law rules. It concluded that there was a true conflict because, unlike the District of Columbia, Indiana law does not recognize the doctrine of depecage, i.e., conducting separate choice-of-law analyses for different issues within a single theory of recovery, such as liability and damages. To determine which jurisdiction's rules applied, the District Court employed the methodology set forth in Gould Electronics, Inc. v. United States, 220 F.3d 169 (3d Cir.2000), where we described five different approaches to choice-of-law conflicts. Concluding that these factors on balance favored Indiana, the District Court utilized Indiana's choice-of-law rules to determine whether to apply Indiana's or Pennsylvania's substantive law.

Turning to the question whether Indiana's choice-of-law rules would apply Pennsylvania or Indiana's substantive law to the case (no party urged the application of D.C. or Kentucky substantive law), the District Court first determined that there was a true conflict between the two jurisdictions' damages regimes, as Pennsylvania's damages law (urged by Plaintiffs) allows for more generous compensation of tort victims than Indiana's (urged by the Government). The Court concluded that the Indiana Supreme Court would choose Indiana substantive law over that of Pennsylvania. Reasoning that the choice-of-law issue would likely be outcome determinative, however, the Court certified the choice-of-law question to us under 28 U.S.C. § 1292(b). The papers are not entirely precise, but it is clear from their tenor that the District Court decided both choice-of-law issues and that they have been presented by this appeal.

We are unsure about the correctness of the District Court's threshold ruling — that a true conflict exists between Indiana and D.C.'s choice-of-law rules. Because the Indiana Supreme Court has not addressed the question whether Indiana courts may apply depecage, we will vacate the District Court's conclusion that the Indiana Supreme Court would not permit them to do so, and certify that question to the Indiana Supreme Court itself pursuant to Rule 64 of the Indiana Rules of Appellate Procedure.1 In the ordinary case, that would be sufficient for now. But if the Indiana Supreme Court holds that Indiana does not recognize depecage, meaning that there is a true conflict between Indiana and D.C. law, we will then be obliged to decide which jurisdiction's choice-of-law rules to apply, and if we choose Indiana, another question looms — whether Indiana would apply its own substantive law or Pennsylvania's to the case.2 This is an extremely difficult question, and we are doubtful of the correctness of the District Court's resolution of the issue. Evaluating the Indiana three-pronged test for determining which state's law it would apply would seem to result in a "tie" between two factors, the third being largely neutral. This question too would therefore benefit from a determination by the Indiana Supreme Court as to how it would apply the factors and break the tie.

While logically we might make the initial certification to the Indiana Supreme Court and await its results (which might or might not render it necessary for us to proceed further), we are reluctant to risk the necessity of a second certification to the Indiana high court, especially in view of the fact that this case was fully ready for a four-week trial when the District Court made its § 1292(b) certification.

Avoiding the spectre of a two-step certification requires that we assume that Indiana may decide that it does not recognize depecage, in which case: (1) there will be a true conflict between Indiana and D.C. choice-of-law principles; and (2) this Court will have to decide which applies. To do this, we must clarify our reasoning in Gould Electronics since that opinion did not establish a general rule. We synthesize the Gould approaches into a single inquiry that chooses the rules of the jurisdiction containing the last significant negligent act or omission relevant to the FTCA. Since we conclude that the air traffic controllers in Indiana committed the last significant act, we will apply Indiana's choice-of-law rules. As noted above, under the Indiana Supreme Court's three-part choice-of-law test set forth in Hubbard Manufacturing Co., Inc. v. Greeson, 515 N.E.2d 1071, 1073 (Ind.1987), one factor is indeterminate, one favors Indiana substantive law, and one favors Pennsylvania substantive law. Since Hubbard gives no indication of how to break such a tie, we will also certify this question to the Indiana Supreme Court. Upon receipt of the answers to both certified questions, the trial court can proceed to trial with certainty as to the law.

These cases arise out of a fatal plane crash in Somerset, Kentucky, in January of 2000. Dennis Schalliol, a Pennsylvania resident, was employed as a salesman by Hart National, a Pennsylvania Corporation with its principal place of its national real estate business in Pennsylvania. B. Kenin Hart, also a Pennsylvania resident, was the sole owner, chairman, and CEO of Hart National. In 1998, Hart National established a subsidiary, Hart Delaware, whose main job it was to operate the aircraft in question. (Although called Hart Delaware, it is based in Pennsylvania.) Hart Delaware had but one employee, John Fare, a New Jersey resident and pilot, who was hired to fly the airplane. The airplane was hangared and operated out of the Northeast Philadelphia Airport for the primary purpose of transporting employees of Hart National from Pennsylvania to related business sites.

In January of 2000, the airplane departed from Northeast Philadelphia Airport bound for Ohio State University Airport, via Cincinnati, with pilot Fare and passengers Schalliol and Hart on board. After picking up a broker, Loy Thompson, in Cincinnati, the plane traveled to Ohio State University Airport in Columbus and remained there overnight. The next morning, pilot Fare received a weather briefing indicating that there were bad conditions at the scheduled destination of Somerset, Kentucky. Such conditions required an instrument flight rules (IFR) landing. While en route to Kentucky, pilot Fare communicated with federal air traffic controllers at the Indianapolis Air Route Traffic Control Center who cleared the flight for the SDF (Simplified Directional Facility) runway 4 approach at the Somerset airport. The SDF system is a group of instruments that guides a plane horizontally onto the runway. Together with an altimeter, it allows a plane to land "blind." As it turns out, the SDF approach had been out of service since 1995, a fact reflected in the Airport Facility Directory ("AFD") published by the Federal Aviation Administration ("FAA"), although pilot Fare was unaware of that condition.

The plane's last radio contact with the air traffic controllers occurred when it was about 30 miles north of the Somerset airport. As it descended, weather conditions made it impossible to see the surrounding terrain. With no further contact from the air traffic controllers, pilot Fare relied on the Instrument Approach Procedure ("IAP"), a document compiled and published in Washington, D.C. that lists navigable IFR approaches to various airports around the country, to guide the landing. If operational, the SDF would have guided the plane safely onto the runway, but instead the plane drifted off course and struck a guy wire from a radio communications tower approximately 3.3 miles from the Somerset airport. The resulting crash killed everyone on board.

The personal representatives of the decedents brought suit in the District Court for the Eastern District of Pennsylvania. Plaintiffs' amended complaint under the Federal Tort Claims Act, 28 U.S.C.A. §§ 1346(b), 2671 et al. ("FTCA"), alleged that the U.S. was negligent in publishing the IAP for the SDF approach to Somerset when in fact that approach had been out of service for five years at the time of the crash.3 It further alleged that the air traffic controllers in Indiana acted negligently by issuing a clearance for an approach that was out of service, failing to monitor radar scopes after clearance was given, failing to conduct proper briefings, and failing to issue a safety alert to pilot Fare.

Plaintiff Mary Schalliol, as personal representative of the estate of Dennis Schalliol, not only brought a claim of negligence against the United States under the FTCA, but also sued the estate of pilot Fare (alleging his negligence) and Fare's employer, Hart Corporation/Delaware Division. The estate of pilot Fare, through separate counsel, has also brought a cross-claim of negligence against the United States. In a separate action that has been consolidated on appeal, four personal representatives of the estate of B. Kenin Hart, his three children and his mother (collectively, the "Simon Plaintiffs"), have brought claims against the United States. The United States filed separate motions for determinations of the applicable choice of law in each of the cases, and each motion sought the application of Indiana substantive law to the claims pled against it under the FTCA. Plaintiff Schalliol, defendants Fare and Hart Delaware, and the Simon Plaintiffs each responded separately seeking the application of Pennsylvania substantive law, which is far more favorable to tort victims.

As explained supra, the District Court found a true conflict between Indiana and D.C.'s choice-of-law rules, so it employed the choice-of-law procedure set forth in Gould Electronics to determine which jurisdiction's rules apply. Gould delineates five common approaches in choosing between competing choice-of-law regimes, although it does not express a preference for any of the five because, under the facts of that case, one of the approaches was inapplicable and the remaining four yielded the same outcome. Applying the approaches to the facts in this case, the Court found the first one, "when the injury can be parsed by the acts or omissions in the different states, ... applying the relevant state's choice of law rules for each act or omission," 220 F.3d at 182, inapplicable because the injury — death in a plane crash — cannot be parsed. It concluded that the second approach, "elect[ing] the choice of law rules of the place of the last act or omission having a causal effect," Id., clearly favored Indiana, the location of the air traffic controllers.

The District Court likewise determined that the third approach, electing the rules of "the place of the last act or omission having the most significant causal effect," id., favored Indiana because a direct nexus exists between the conduct of the controllers and the crash. The fourth approach, "select[ing] the choice of law rules of the state in which `physical acts' could have prevented the injury," id. at 183, was found to be indeterminate because just as the air traffic controllers in Indiana could have given different instructions, the federal employees in D.C. could have published an accurate IAP. Finally, the fifth approach, choosing "based on where the `relevant' act or omission occurred," id., was found to be indeterminate because there were two independent acts of negligence, each of which could have been fatal in its own right. In sum, the District Court concluded that of the five Gould approaches, one was inapplicable, two favored Indiana, and two were indeterminate, so it applied Indiana's choice-of-law rules to choose between Indiana and Pennsylvania substantive law.4

Indiana employs a modified lex loci delicti test, which applies the substantive law of the jurisdiction where the tort was committed. Hubbard, 515 N.E.2d at 1073. Generally, in Indiana wrongful death or injury cases, the place where the death occurred is the site of the tort's commission or of the "last event necessary," see, e.g. Lambert v. Yellowbird, Inc., 496 N.E.2d 406, 409 (Ind.Ct.App.1986), which in this case is Kentucky, the site of the crash and resulting deaths. As the District Court explained, however, the Indiana inquiry does not end here:

In those instances where the place of the tort bears little connection to the legal action, this Court will permit the consideration of other factors such as: 1) the place where the conduct causing the injury occurred; 2) the residence or place of business of the parties; and 3) the place where the relationship is centered. These factors should be evaluated according to their relative importance to the particular issues being litigated.

Hubbard, 515 N.E.2d at 1073 (quoting Restatement (Second) of Conflicts of Laws § 145(2) (1971)).

The District Court concluded that Kentucky "has little connection to the legal action" concerning the FTCA claims against the United States's air traffic controllers and chart publishers, so it turned to the three supplemental Hubbard factors. It found that the first, "the place where the conduct causing the injury occurred," favored Indiana because although negligence also occurred in D.C., "the conduct most directly affecting the plane's landing clearly occurred in Indiana." For example, the Indiana controllers were the only actors with direct contact with and control over the pilot and plane.

As to the second Hubbard factor, "the residence or place of business of the parties," they are various — Indiana (plaintiff Schalliol), Pennsylvania (three of the four Simon plaintiffs and passenger Hart), New York (fourth Simon plaintiff), Delaware (defendant Fare/Cross-plaintiff Fare and defendant Hart Delaware), and, with respect to the defendant United States, all states or no state.5 The District Court concluded that this factor favored no state, so it apparently gave that factor relatively little weight. Finally, the District Court concluded that the third Hubbard factor, "the place where the relationship is centered," could likewise point to any of several locations since none of the parties was ever located in the same state as any adverse party.

Based on these three factors, the District Court concluded that Indiana bears the "edge" as the jurisdiction with the most significant contacts to the negligence claims against the United States. However, the District Court recognized that such a decision would effectively dictate the case's outcome, which is essentially a question of damages — it is hard to envision a successful defense to the Government's liability. Because the choice-of-law issues are so critical, the Court certified them for our immediate review pursuant to 28 U.S.C. § 1292(b), which requires that: (1) the issue involve a controlling question of law; (2) as to which there are substantial grounds for difference of opinion; and that (3) an immediate appeal of the order may materially advance the ultimate termination of the litigation. This being a clear case for § 1292(b) certification, we accepted it and, under Rule 5 of the Federal Rules of Appellate Procedure, requested briefing and heard oral argument.

The District Court had diversity jurisdiction over the Simon and Schalliol suits pursuant to 28 U.S.C. § 1332. It had jurisdiction over Fare's cross-claim against the United States by virtue of the Federal Tort Claims Act, 28 U.S.C. §§ 1346(b) and 2671 et seq. We exercise jurisdiction pursuant to 28 U.S.C. § 1292(b). Because choice-of-law is a purely legal determination, our review is plenary.

The Federal Tort Claims Act waives sovereign immunity against persons suing the federal government for commission of various torts. In a multi-state tort action, it requires a federal court to apply the law of the place where the acts of negligence occurred. 28 U.S.C.A. §§ 1346(b), 2674. The Supreme Court, however, has expressly rejected the argument that the law of the place of injury should apply where negligence occurs in one state and results in injury or death in another. Richards v. U.S., 369 U.S. at 1, 82 S. Ct. 585. Richards dictates that either Indiana or D.C.'s choice-of-law rules must apply, for those are the two places where government negligence allegedly occurred — the air traffic controllers were based in Indiana, and the faulty IAP was published in D.C. Although each of the key questions will be certified, for the benefit of the Indiana Supreme Court and the parties, it is important that we lay out the factual and legal background reasoning that animates our decision.

In choosing between Indiana and D.C.'s choice-of-law regimes, we must first determine whether there exists a "true conflict" between them. Gould, 220 F.3d at 179, 181. The parties disagree as to whether Indiana and D.C.'s rules differ in any material respect. Indiana's choice of law utilizes a modified lex loci delicti test. Hubbard, 515 N.E.2d at 1073. It begins with the presumption that the substantive law of the place where the tort was committed governs the case. Id. However, if the place of the tort bears little connection to the legal action, Indiana follows the "most significant relationship" test detailed in the Restatement (Second) Conflicts of Laws. Id. at 1074. The parties agree that Kentucky, the location of the tort, has little connection to the litigation since the plane merely crashed there. Therefore, in this case Indiana law defaults to the "most significant relationship" test, in which it considers three factors: (1) the place where the conduct causing the injury occurred; (2) the residence and place of incorporation and place of business of the parties; and (3) the place where the parties' relationship is centered.

In comparison, D.C. implements a hybrid of the "governmental interest" and Restatement (Second) methodologies that identifies the governmental policies underlying the applicable laws and determines which state's policy would be most advanced by having its laws applied to the facts of the case. Raflo v. United States, 157 F. Supp. 2d 1, 4 (D.D.C.2001). The stronger state interest is found by examining: (1) the place where the injury occurred; (2) the place where the conduct causing the injury occurred; (3) the domicile, residence, place of incorporation and place of business of the parties; and (4) the place where the relationship is centered. Id. at 5. Although D.C.'s analysis also takes into account government policies, plaintiffs contend that the factors D.C. considers do not differ meaningfully from those Indiana considers, and therefore that there is merely a false conflict between D.C. and Indiana law.

The United States disagrees, asserting that a true conflict exists. It points out that Indiana did not adopt the policy component of § 6 of the Second Restatement. See, e.g., Hubbard, 515 N.E.2d at 1074 (declining to balance competing interests of states by identifying policies underlying their differing laws); In re Bridgestone/Firestone, Inc., Tires Prods. Liab. Litig., 155 F. Supp. 2d 1069, 1083 n. 10 (S.D.Ind.2001) (noting that interests of each state are not determinative or generally even considered by Indiana courts). It notes that academic commentators agree that Indiana eschews such policy considerations. See, e.g., Symeon C. Symeonides, The Judicial Acceptance of the Second Conflicts Restatement: A Mixed Blessing, 56 Md. L. Rev. 1248, 1252, 1268 & n.136, & 1272 n.159 (1997) ("[Indiana's] significant-contacts approach differs from section 6 of the Restatement (Second) in that it calls for a consideration of the factual contacts alone, rather than a set of policies in light of the factual contacts as does the Restatement (Second)." (emphasis added)). Plaintiffs, however, assert that Indiana courts do consider state policies in their analysis. See, e.g., In re the Estate of Bruck, 632 N.E.2d 745, 749 (Ind. Ct.App.1994) (holding on the specific point at issue that, although Indiana law would otherwise apply, "only under Ohio law can the distribution fulfill the essential purpose of the damages, which is to benefit the immediate survivors regardless of dependency").

It is of course possible that even if D.C. considers policy while Indiana does not, the two schemes would present a false conflict because they would ultimately select the same state's substantive law. But, the United States argues, the conflict in choice-of-law regimes goes deeper than mere consideration of policy, for D.C.'s choice-of-law rules recognize the concept of depecage while Indiana's do not. Depecage is "the process whereby different issues in a single case arising out of a single set of facts are decided according to the laws of different states." Broome v. Antlers' Hunting Club, 595 F.2d 921, 923 n. 5 (3d Cir.1979). In the case at bar, depecage would enable a court to apply one state's laws to assess liability and another's to assess damages. It is clear that D.C.'s choice-of-law rules permit depecage, see, e.g., Stutsman v. Kaiser Foundation Health Plan of the Mid-Atlantic States, Inc., 546 A.2d 367, 373 (D.C.App.1988), so a critical question for "true conflict" purposes is whether Indiana's do as well.

Plaintiffs identify several Indiana decisions that allegedly employ depecage, including Allen v. Great American Reserve Ins. Co., 766 N.E.2d 1157, 1162 (Ind.2002) ("In analyzing each of the counts of the plaintiffs' complaint, it is first necessary to determine which state's law applies to that count. The answer may differ for different counts and may differ between defendants as to a single count."); Cap Gemini America, Inc. v. Judd, 597 N.E.2d 1272, 1281 (Ind.Ct.App.1992) (undertaking separate choice-of-law analyses on liability and damages); Estate of Sullivan v. United States, 777 F. Supp. 695 (N.D.Ind.1991) (applying Arizona law to the medical malpractice action and Indiana law to the wrongful death claim). None of these cases, however, establishes that Indiana law embraces depecage. Allen was a routine application of different choice-of-law analyses to different counts, as opposed to different issues within a single count. Cap Gemini conducted a choice-of-law analysis for damages purposes, but there were actually two distinct claims at stake in that case, the second of which involved only a question of damages — the liability and damages issues involved different theories of recovery. Finally, although Estate of Sullivan did employ depecage, the Indiana court in that case was applying Arizona's choice-of-law rules, not Indiana's.

Viewing Indiana's caselaw as a whole, the closest an Indiana court has come to employing depecage is in In re the Estate of Bruck, 632 N.E.2d 745 (Ind.Ct.App. 1994). There the court said in dictum that, if it had found a true conflict, it would have applied a New York decision stating that "the recovery and distribution aspects of wrongful death cases may be separated when choice-of-law questions arise." Id. at 749. But as the District Court correctly noted, Bruck's dictum is not very valuable here, especially since no Indiana case has explicitly employed depecage and the dictum itself is from an intermediate appellate court.

Plaintiffs offer one slightly different rationale in favor of depecage: that many Indiana cases adopt § 145 of the Restatement, which in turn incorporates the general criteria of § 6 of the Restatement. See Hubbard, 515 N.E.2d at 1073-74 (Ind. 1987). Because § 6 endorses depecage, plaintiffs submit that depecage has been incorporated indirectly into Indiana law. However, given the possibility that Indiana has adopted only part of the Restatement, we are unwilling to conclude that Indiana has embraced depecage based only on such roundabout reasoning.

In sum, we find no evidence that Indiana courts employ the concept of depecage, and absent depecage it is clear that there is a true conflict between Indiana and D.C.'s choice-of-law rules. We nevertheless note that the Indiana Supreme Court has not definitively resolved the matter, and we believe that a question of such importance is best left to that Court's authoritative disposition. We will therefore certify the issue to the Indiana Supreme Court pursuant to Rule 64 of the Indiana Rules of Appellate Procedure.

In a multistate tort action, the FTCA requires a federal court to apply the law of the place where the acts of negligence occurred. 28 U.S.C. §§ 1346(b), 2674; Richards, 369 U.S. at 1, 82 S. Ct. 585. We must therefore apply either Indiana's or D.C.'s choice-of-law rules, but we need only choose between them if they conflict. For the reasons outlined above, we will assume arguendo that Indiana does not embrace the concept of depecage, and therefore that a true conflict exists. In Gould, 220 F.3d at 169, we addressed the matter of how to resolve such a conflict. Our opinion surveyed the caselaw and literature on the subject and identified five common approaches: (1) applying different rules to different theories of liability; (2) choosing the place of the last allegedly-wrongful act or omission; (3) determining which asserted act of wrongdoing had the most significant effect on the injury; (4) choosing the state in which the United States' physical actions could have prevented injury; and (5) determining where the "relevant" act or omission occurred. Under Gould's facts, however, the first approach was inapplicable and the final four pointed toward the same state, so we declined to adopt any one approach.

The case at bar is much closer, and the approaches yield conflicting results. The parties agree that the first approach, applying different rules to different theories of liability, is inapplicable because there is only one theory of liability — negligence. The second approach, choosing the place of the last allegedly-wrongful act or omission, seems clearly to favor Indiana. Although it is true that the IAP published in D.C. incorrectly depicted as operational an SDF approach that had been out of service for five years, (Plaintiffs' Stipulation of Facts, A96), the air traffic controllers' negligence postdated the chartmakers' negligence. Plaintiffs' pretrial memorandum alleges that the controllers: (1) failed to review various Area and Hazard binders which would have revealed that the SDF approach was out of service; (2) cleared the flight for landing on an out-of-service approach; (3) failed to monitor the plane on radar during the approach, a safeguard which would have revealed that the plane was off course; (4) failed to issue a low altitude safety advisory; and (4) failed to respond to pilot Fare's radio communication after his first pass by the runway. (Report of Expert J. Gary Parham, Schalliol Appendix at 104A).

The plaintiffs, however, submit that the air traffic controllers' last-in-time negligence is relatively unimportant because their location in Indiana was a mere fortuity — they could have been anywhere, and Indiana as a jurisdictional entity had nothing to do with the plane crash. They therefore argue that this factor should be given relatively little weight, and cite to In re Aircrash Disaster Near Roselawn, Indiana on October 31, 1994, 926 F. Supp. 736, 743 n. 9 (N.D.Ill.1996)("The manner in which federal air traffic controllers choose to partition air space has no bearing on the litigation of the decedent's injuries or the interests of the several states in this litigation.").

The United States disagrees that the flight controllers' location was fortuitous, because the plane's flight plan ensured that it would traverse the controllers' airspace, where they would communicate with the pilot. It also argues that the Aircrash Disaster case is inapposite, for in that case no party had sued the United States alleging air traffic controller negligence, so the location of the controllers themselves was not a factor.

We are satisfied that this "last wrongful act" approach favors Indiana law even if the air traffic controllers' location was to some degree fortuitous. It certainly is not entirely fortuitous, for the federal government has undoubtedly positioned air traffic controllers strategically so as to create manageable regions of airspace. Indeed, the publication of the IAP in Washington, D.C. was at least as fortuitous, for there is no reason why its contents could not have been compiled and distributed in any jurisdiction.

The third Gould approach, determining which asserted act of wrongdoing had the most significant effect on the injury, is quite close. The District Court concluded that this test favored Indiana, where the acts of negligence assertedly had the most "direct nexus" to the crash. The Plaintiffs disagree, arguing that this concept of "direct nexus" flirts dangerously close to the second factor, which is the place of the most recent act. In their view, "most significant cause" is not synonymous with "proximate cause," and they argue that this factor actually favors D.C. because the chartmakers in D.C. committed a (in fact, "the") major sin of commission by producing a defective product. The air traffic controllers, on the other hand, merely relied upon this inaccurate map, making their negligence less significant.

As explained supra, however, it is clear that the air traffic controllers' negligence went beyond passive reliance on a faulty chart; among other things, they failed to review mandatory binders, failed to monitor the flight's approach on radar, and failed to respond to the pilot's radio communications. The Government does not meaningfully challenge these points — although it maintains that the air traffic controllers were not negligent, this position is undermined by its concession at oral argument that: "the air traffic controllers cleared the pilot for an approach; the navigational aid for that approach was out of service, [and] they knew it was out of service but forgot." We therefore conclude that with clear negligence by the air traffic controllers counter-balancing that of the chart makers, this third approach is also indeterminate. Each act of negligence might fairly be termed a "but for" cause of the crash, and reasonable minds could disagree on which cause (if any) was most significant.

The fourth approach, choosing the state in which the United States's physical actions could have prevented injury, seems likewise to be indeterminate. The air traffic controllers could have given correct instructions and been more attentive during the plane's approach, and the chart makers could have produced an accurate chart. No one physical act seems to be the clear cause of the accident.

The fifth and final Gould approach, determining where the "relevant" act or omission occurred, seems to us to add nothing unique to the inquiry. Indeed, it seems indistinguishable from the third approach of identifying the "most significant" act of negligence. We therefore conclude that this approach is also indeterminate.

Of the five Gould approaches, then, one is inapplicable, three are indeterminate, and one favors Indiana. Although Gould might therefore counsel choosing Indiana law insofar as it "outscores" D.C. law, the approaches overlap to such an extent that it would not seem judicious to resolve the question in so simplistic a fashion. At all events, Gould itself does not compel that outcome, and we are unwilling to endorse a system that requires an ad hoc balancing of vague and overlapping approaches each time we choose among choice-of-law regimes. Indeed, it is unclear that one method of picking a choice-of-law regime is superior to another ex ante.

Jurisdictions' interests — their policy concerns — are safeguarded by their substantive laws, so there is much at stake in choosing a substantive regime in a case. Likewise, because jurisdictions' choice-of-law regimes differ and might therefore affect which substantive law applies, each jurisdiction has an interest in applying its choice-of-law rules to a case. There is, however, no compelling reason for a federal court to prefer one method to another for choosing among competing jurisdictions' choice-of-law regimes. For example, it is unclear whether a federal court's adoption of a "last act" approach would yield more desirable final outcomes than a "most significant contact" approach, for once the "last act" jurisdiction is identified, its choice-of-law rules might in turn employ a "last act" approach, a "most significant contacts" approach, or another approach entirely.

We therefore conclude that clarity is the most important virtue in crafting a rule by which we choose a jurisdiction. In that sense, Gould's methodology is far from desirable, for it lists several often-conflicting and overlapping approaches without providing any guidance as to which approach should be given the most weight. Today we refine Gould by adopting a single approach to resolving choice-of-law disputes in FTCA cases: we will apply the choice-of-law regime of the jurisdiction in which the last significant act or omission occurred. This has the salutary effect of avoiding the selection of a jurisdiction based on a completely incidental "last contact," while also avoiding the conjecture that "most significant act" inquiries often entail.

In this case, the "last significant act" approach clearly points to Indiana, the location of the air traffic controllers' negligence. We should therefore apply Indiana's choice-of-law rules to determine whether Indiana's or Pennsylvania's substantive laws apply.

Indiana law first looks to determine whether there is a true conflict between the competing jurisdictions' substantive law. Assuming that depecage does not exist, see supra, we must look at the negligence cause of action as a whole. This case is essentially about damages, for liability does not seem difficult and it is clear that Pennsylvania and Indiana allow very different degrees of recovery. Pennsylvania allows joint-and-several liability and right of contribution, while Indiana does not. See 42 Pa.C.S.A. § 8322, 8324; I.C. § 34-51-2-12. Moreover, although both states maintain survival statutes that allow recovery of survival damages for the decedent's personal injuries and consequent damages, Indiana does not permit recovery for both wrongful death and survival damages. Cahoon v. Cummings, 715 N.E.2d 1 (Ind.App.1999). Lastly, unlike Indiana, Pennsylvania damages include the decedent's conscious pain and suffering from the moment of injury to the time of death. It is clear that a true conflict exists.

In the case of a true conflict, Indiana applies the law of the place of the tort, unless that place "bears little connection to the legal action." Hubbard, 515 N.E.2d at 1073-74. Kentucky is the place of the tort, and the parties agree that it has little relevance; in such instances, Indiana employs a three-part test to determine which state's substantive laws apply, under which it considers: "(1) the place where the conduct causing the injury occurred; (2) the residence or place of business of the parties; and (3) the place where the relationship is centered." Id. Each of these factors must be evaluated "according to their relative importance to the particular issues being litigated." Id.

The first factor, "the place where the conduct causing the injury occurred," clearly favors Indiana because no negligence occurred in Pennsylvania. The second factor, "the residence or place of business of the parties," seems to favor Pennsylvania. Hart Delaware, which owned the plane, was incorporated in Pennsylvania, and plaintiff Fare worked in Pennsylvania. Although plaintiff Schalliol lived in Indiana, three of the four Simon plaintiffs lived in Pennsylvania. While the air traffic controllers presumably lived in Indiana, they are not the defendants — the United States is the defendant, and it is assumed to reside in all states or no state. Compare Clawans v. United States, 75 F. Supp. 2d 368, 374 (D.N.J.1999) (noting that U.S. is not a resident of a particular state), with United States v. Whitcomb, 314 F.2d 415, 417 (4th Cir.1963) (finding that the United States is a resident of every state). On balance, the "residence or place of business of the parties" seems to favor Pennsylvania.

The third factor, "the place where the parties' relationship is centered," is somewhat difficult to conceptualize on these facts. At no time were the parties located in the same state — their only relationships involved reliance on a map and communication over a radio, neither of which seems "centered" in a particular place.

Surveying the three Hubbard factors, the first points to Indiana, the second to Pennsylvania, and the third is indeterminate. Like our decision in Gould, Hubbard gives no guidance as to which factor is most important or how to "break a tie," so any decision by this court on which substantive law Indiana would apply would be little more than a guess. We will therefore certify this issue to the Indiana Supreme Court under Rule 64 of the Indiana Rules of Appellate Procedure.

For the above reasons, we will certify the following questions to the Indiana Supreme Court: (1) whether a true conflict exists between Indiana's and D.C.'s choice-of-law rules; and (2) if there is a true conflict and Indiana's choice-of-law rules therefore control per our "last significant act" test, how to resolve a split among the Hubbard factors in choosing a jurisdiction's substantive law when one factor points toward Indiana, another toward Pennsylvania, and the third is indeterminate, and what substantive law Indiana would choose under these facts.6