Firestone tires on Ford Explorer SUVs experienced an abnormally high failure rate during the late 1990s. In August 2000, while the National Highway Transportation Safety Administration was investigating, Firestone recalled and replaced some of those tires. Ford and Firestone replaced additional tires during 2001. Many suits have been filed as a result of injuries and deaths related to the tire failures. Other suits were filed by persons who own (or owned) Ford Explorers or Firestone tires that have so far performed properly; these persons seek compensation for the risk of failure, which may be reflected in diminished resale value of the vehicles and perhaps in mental stress. The Judicial Panel on Multidistrict Litigation transferred suits filed in, or removed to, federal court to the Southern District of Indiana for consolidated pretrial proceedings under 28 U.S.C. § 1407(a). Once these have been completed, the cases must be returned to the originating districts for decision on the merits. See Lexecon Inc. v. Milberg Weiss Bershad Hynes & Lerach, 523 U.S. 26, 118 S. Ct. 956, 140 L. Ed. 2d 62 (1998). In an effort to prevent retransfer, counsel representing many of the plaintiffs filed a new consolidated suit in Indianapolis and asked the judge to certify it as a nationwide class action, which would make all other suits redundant. The district court obliged and certified two nationwide classes: the first includes everyone who owns, owned, leases, or leased a Ford Explorer of model year 1991 through 2001 anytime before the first recall, and the second includes all owners and lessees from 1990 until today of Firestone ATX, ATX II, Firehawk ATX, ATX 23 Degree, Widetrack Radial Baja, or Wilderness tire models, or any other Firestone tire "substantially similar" to them. In re Bridgestone/Firestone, Inc., Tires Products Liability Litigation, 205 F.R.D. 503 (S.D.Ind.2001); see also 155 F. Supp. 2d 1069 (S.D.Ind.2001). More than 60 million tires and 3 million vehicles fit these definitions.

No class action is proper unless all litigants are governed by the same legal rules. Otherwise the class cannot satisfy the commonality and superiority requirements of Fed.R.Civ.P. 23(a), (b)(3). Yet state laws about theories such as those presented by our plaintiffs differ, and such differences have led us to hold that other warranty, fraud, or products-liability suits may not proceed as nationwide classes. See, e.g., Isaacs v. Sprint Corp., 261 F.3d 679 (7th Cir.2001); Szabo v. Bridgeport Machines, Inc., 249 F.3d 672 (7th Cir. 2001); In re Rhone-Poulenc Rorer Inc., 51 F.3d 1293 (7th Cir.1995). See also In re Mexico Money Transfer Litigation, 267 F.3d 743, 746-47 (7th Cir.2001). The district judge, well aware of this principle, recognized that uniform law would be essential to class certification. Because plaintiffs' claims rest on state law, the choice-of-law rules come from the state in which the federal court sits. See Klaxon v. Stentor Electric Manufacturing Co., 313 U.S. 487, 61 S. Ct. 1020, 85 L. Ed. 1477 (1941). The district judge concluded that Indiana law points to the headquarters of the defendants, because that is where the products are designed and the important decisions about disclosures and sales are made. Ford and Firestone engaged in conduct that was uniform across the nation, which the district court took to imply the appropriateness of uniform law. This ruling means that all claims by the Explorer class will be resolved under Michigan law and all claims by the tire class will be resolved under Tennessee law. According to the district court, other obstacles (such as the fact that the six named tire models represent 67 designs for different sizes and performance criteria, and that half of all 1996 and 1997 model Explorers came with Goodyear tires) are worth overcoming in light of the efficiency of class treatment. Nor did the district court deem it important that Firestone's tires were designed in Ohio, and many were manufactured outside Tennessee, as many of Ford's vehicles are manufactured outside Michigan.

Both Ford and Firestone petitioned for interlocutory review under Fed. R.Civ.P. 23(f). We granted these requests because, as in Rhone-Poulenc and other cases (e.g., West v. Prudential Securities, Inc., 282 F.3d 935 (7th Cir.2002)) the suit is exceedingly unlikely to be tried. Aggregating millions of claims on account of multiple products manufactured and sold across more than ten years makes the case so unwieldy, and the stakes so large, that settlement becomes almost inevitable — and at a price that reflects the risk of a catastrophic judgment as much as, if not more than, the actual merit of the claims. Permitting appellate review before class certification can precipitate such a settlement is a principal function of Rule 23(f). See Blair v. Equifax Check Services, Inc., 181 F.3d 832, 834-35 (7th Cir.1999). Another function is permitting appellate review of important legal issues that otherwise might prove elusive. The district court's conclusion that one state's law would apply to claims by consumers throughout the country — not just those in Indiana, but also those in California, New Jersey, and Mississippi — is a novelty, and, if followed, would be of considerable import to other suits. Our review of this choice-of-law question is plenary, so we start there.

Indiana is a lex loci delicti state: in all but exceptional cases it applies the law of the place where harm occurred. See Hubbard Manufacturing Co. v. Greeson, 515 N.E.2d 1071 (Ind.1987). Those class members who suffered injury or death as a result of defects were harmed in the states where the tires failed. As a practical matter, these class members can be ignored; they are sure to opt out and litigate independently. These classes therefore effectively include only those consumers whose loss (if any) is financial rather than physical: it is the class of persons whose tires did not fail, whose vehicles did not roll over. Many class members face no future threat of failure either, because about 30 million tires were recalled and replaced, while other tires have been used up and discarded. Financial loss (if any, a qualification we will not repeat) was suffered in the places where the vehicles and tires were purchased at excessive prices or resold at depressed prices. Those injuries occurred in all 50 states, the District of Columbia, Puerto Rico, and U.S. territories such as Guam. The lex loci delicti principle points to the places of these injuries, not the defendants' corporate headquarters, as the source of law.

Plaintiffs concede that until 1987 this would have been Indiana's approach. They contend, however, that Hubbard changed everything by holding that when the place of the injury "bears little connection to the legal action" a court may consider other factors, such as the place of the conduct causing the injury and the residence of the parties. It is conceivable, we suppose, that Indiana might think that a financial (or physical) injury to one of its residents, occurring within the state's borders, "bears little connection to the legal action", but the proof of that pudding is in the eating. Has Indiana since 1987 applied the law of a state where a product was designed, or promotional materials drafted, to a suit arising out of an injury in Indiana? As far as we can tell, the answer is no — not even once, and the state has had plenty of opportunities. Yet since 1987 both Indiana and this court have routinely applied Indiana law when injury caused by a defective product occurred in Indiana to Indiana residents. See, e.g., Land v. Yamaha Motor Corp., 272 F.3d 514, 517 (7th Cir.2001) (Indiana law); Morgen v. Ford Motor Co., 762 N.E.2d 137 (Ind.App.2002). Neither Indiana nor any other state has applied a uniform place-of-the-defendant's-headquarters rule to products-liability cases. It is not hard to devise an argument that such a uniform rule would be good on many dimensions, but that argument has not carried the day with state judges, and it is state law rather than a quest for efficiency in litigation (or in product design decisions) that controls.

"Ah, but this is not a products-liability case!" So plaintiffs respond to the conspicuous lack of support from state decisions. And indeed it is not a products-liability suit, since all who suffered physical injury are bound to opt out. No injury, no tort, is an ingredient of every state's law. See, e.g., Fogel v. Zell, 221 F.3d 955, 960-62 (7th Cir.2000); In re Orthopedic Bone Screw Products Liability Litigation, 193 F.3d 781, 789 (3d Cir.1999); Travelers Insurance Co. v. Eljer Manufacturing, Inc., 197 Ill. 2d 278, 258 Ill. Dec. 792, 757 N.E.2d 481 (2001). Plaintiffs describe the injury as financial rather than physical and seek to move the suit out of the tort domain and into that of contract (the vehicle was not the flawless one described and thus is not merchantable, a warranty theory) and consumer fraud (on the theory that selling products with undisclosed attributes, and thus worth less than represented, is fraudulent). It is not clear that this maneuver actually moves the locus from tort to contract. If tort law fully compensates those who are physically injured, then any recoveries by those whose products function properly mean excess compensation.1 As a result, most states would not entertain the sort of theory that plaintiffs press. See, e.g., Briehl v. General Motors Corp., 172 F.3d 623, 628 (8th Cir. 1999) (Mississippi, New York, Pennsylvania, and Texas law); Angus v. Shiley, Inc., 989 F.2d 142, 147-48 (3d Cir.1993) (Pennsylvania law); Willett v. Baxter International, Inc., 929 F.2d 1094, 1099-1100 (5th Cir.1991) (Louisiana law); Carlson v. General Motors Corp., 883 F.2d 287, 298 (4th Cir.1989) (South Carolina law); American Suzuki Motor Corp. v. Superior Court, 37 Cal. App. 4th 1291, 44 Cal. Rptr. 2d 526 (Ct. App.1995); Ford Motor Co. v. Rice, 726 So. 2d 626, 627, 631 (Ala.1998); Yu v. IBM Corp., 314 Ill.App.3d 892, 247 Ill. Dec. 841, 732 N.E.2d 1173 (1st Dist.2000); Capital Holding Corp. v. Bailey, 873 S.W.2d 187, 192 (Ky.1994).

Obviously plaintiffs believe that Michigan and Tennessee are in the favorable minority; we need not decide. If recovery for breach of warranty or consumer fraud is possible, the injury is decidedly where the consumer is located, rather than where the seller maintains its headquarters. A contract for the sale of a car in Indiana is governed by Indiana law unless it contains a choice-of-law clause, and plaintiffs do not want to enforce any choice-of-law clause. Plaintiffs have not cited, and we could not find, any Indiana case applying any law other than Indiana's to warranty or fraud claims arising from consumer products designed (or contract terms written) out of state, unless a choice-of-law clause was involved. State consumer-protection laws vary considerably, and courts must respect these differences rather than apply one state's law to sales in other states with different rules. See BMW of North America, Inc. v. Gore, 517 U.S. 559, 568-73, 116 S. Ct. 1589, 134 L. Ed. 2d 809 (1996). We do not for a second suppose that Indiana would apply Michigan law to an auto sale if Michigan permitted auto companies to conceal defects from customers; nor do we think it likely that Indiana would apply Korean law (no matter what Korean law on the subject may provide) to claims of deceit in the sale of Hyundai automobiles, in Indiana, to residents of Indiana, or French law to the sale of cars equipped with Michelin tires. Indiana has consistently said that sales of products in Indiana must conform to Indiana's consumer-protection laws and its rules of contract law. See, e.g., A.J.'s Automotive Sales, Inc. v. Freet, 725 N.E.2d 955, 963 (Ind.App.2000) (consumer-protection law); Dohm & Nelke v. Wilson Foods Corp., 531 N.E.2d 512, 513-14 (Ind. App.1988) (contract law). It follows that Indiana's choice-of-law rule selects the 50 states and multiple territories where the buyers live, and not the place of the sellers' headquarters, for these suits.

Against all of this plaintiffs set a single decision: KPMG Peat Marwick v. Asher, 689 N.E.2d 1283 (Ind.App.1997). This decision holds that the adequacy of services rendered by an accountant in Missouri to a business whose headquarters were in Missouri is governed by Missouri law, even when a suit is filed by unpaid lenders who live in Indiana. This is a straightforward application of lex loci delicti. The injury, if any, was suffered by the business, which hired and paid the accountant for professional services rendered directly to the client; those who dealt with the audited firm, such as the plaintiffs in KPMG Peat Marwick, suffer a derivative injury. Similarly a malpractice claim against a firm's lawyer is determined by the law of the state where the services are performed, for that state's law supplies the standard of performance and that is where the client normally would suffer injury. Investors may be able to step into a corporation's shoes and assert a derivative claim, and in some states (those that have rejected the Ultramares doctrine, see Ultramares Corp. v. Touche, Niven & Co., 255 N.Y. 170, 174 N.E. 441 (1931) (Cardozo, J.)) investors may have a direct claim too; but because the firm remains the lawyer's or accountant's client one body of law must apply to this single transaction. Sales of a consumer product in 50 states do not lead to derivative claims, and each sale is a separate transaction in the place of the sale. KPMG Peat Marwick accordingly has no bearing on consumers' suits against manufacturers of allegedly defective products.

Because these claims must be adjudicated under the law of so many jurisdictions, a single nationwide class is not manageable. Lest we soon see a Rule 23(f) petition to review the certification of 50 state classes, we add that this litigation is not manageable as a class action even on a statewide basis. About 20% of the Ford Explorers were shipped without Firestone tires. The Firestone tires supplied with the majority of the vehicles were recalled at different times;2 they may well have differed in their propensity to fail, and this would require sub-subclassing among those owners of Ford Explorers with Firestone tires. Some of the vehicles were resold and others have not been; the resales may have reflected different discounts that could require vehicle-specific litigation. Plaintiffs contend that many of the failures occurred because Ford and Firestone advised the owners to underinflate their tires, leading them to overheat. Other factors also affect heating; the failure rate (and hence the discount) may have been higher in Arizona than in Alaska. Of those vehicles that have not yet been resold, some will be resold in the future (by which time the tire replacements may have alleviated or eliminated any discount) and some never will be resold. Owners who wring the last possible mile out of their vehicles receive everything they paid for and have claims that differ from owners who sold their Explorers to the second-hand market during the height of the publicity in 2000. Some owners drove their SUVs off the road over rugged terrain, while others never used the "sport" or "utility" features; these differences also affect resale prices.

Firestone's tires likewise exhibit variability; that's why fewer than half of those included in the tire class were recalled. The tire class includes many buyers who used Firestone tires on vehicles other than Ford Explorers, and who therefore were not advised to underinflate their tires. (Note that this description does not reflect any view of the merits; we are repeating rather than endorsing plaintiffs' contention that Ford counseled "underinflation.") The six trade names listed in the class certification order comprise 67 master tire specifications: "Firehawk ATX" tires, for example, come in multiple diameters, widths, and tread designs; their safety features and failure modes differ accordingly. Plaintiffs say that all 67 specifications had three particular shortcomings that led to excess failures. But whether a particular feature is required for safe operation depends on other attributes of the tires, and as these other attributes varied across the 67 master specifications it would not be possible to make a once-and-for-all decision about whether all 60 million tires were defective, even if the law were uniform. There are other differences too, but the ones we have mentioned preclude any finding "that the questions of law or fact common to the members of the class predominate over any questions affecting only individual members, and that a class action is superior to other available methods for the fair and efficient adjudication of the controversy." Fed.R.Civ.P. 23(b)(3). Regulation by the NHTSA, coupled with tort litigation by persons suffering physical injury, is far superior to a suit by millions of uninjured buyers for dealing with consumer products that are said to be failure-prone.

The district judge did not doubt that differences within the class would lead to difficulties in managing the litigation. But the judge thought it better to cope with these differences than to scatter the suits to the winds and require hundreds of judges to resolve thousands of claims under 50 or more bodies of law. Efficiency is a vital goal in any legal system—but the vision of "efficiency" underlying this class certification is the model of the central planner. Plaintiffs share the premise of the ALI's Complex Litigation Project (1993), which devotes more than 700 pages to an analysis of means to consolidate litigation as quickly as possible, by which the authors mean, before multiple trials break out. The authors take as given the benefits of that step. Yet the benefits are elusive. The central planning model — one case, one court, one set of rules, one settlement price for all involved — suppresses information that is vital to accurate resolution. What is the law of Michigan, or Arkansas, or Guam, as applied to this problem? Judges and lawyers will have to guess, because the central planning model keeps the litigation far away from state courts. (Ford asked us to certify legal questions to the Supreme Court of Michigan, to ensure that genuine state law was applied if Michigan's law were to govern the whole country; the plaintiffs stoutly resisted that proposal.) And if the law were clear, how would the facts (and thus the damages per plaintiff) be ascertained? One suit is an all-or-none affair, with high risk even if the parties supply all the information at their disposal. Getting things right the first time would be an accident. Similarly Gosplan or another central planner may hit on the price of wheat, but that would be serendipity. Markets instead use diversified decisionmaking to supply and evaluate information. Thousands of traders affect prices by their purchases and sales over the course of a crop year. This method looks "inefficient" from the planner's perspective, but it produces more information, more accurate prices, and a vibrant, growing economy. See Thomas Sowell, Knowledge and Decisions (1980). When courts think of efficiency, they should think of market models rather than central-planning models.

Our decision in Rhone-Poulenc Rorer made this point, and it is worth reiterating: only "a decentralized process of multiple trials, involving different juries, and different standards of liability, in different jurisdictions" (51 F.3d at 1299) will yield the information needed for accurate evaluation of mass tort claims. Once a series of decisions or settlements has produced an accurate evaluation of a subset of the claims (say, 1995 Explorers in Arizona equipped with a particular tire specification) the others in that subset can be settled or resolved at an established price. See David Friedman, More Justice for Less Money, 39 J.L. & Econ. 211 (1996).

No matter what one makes of the decentralized approach as an original matter, it is hard to adopt the central-planner model without violence not only to Rule 23 but also to principles of federalism. Differences across states may be costly for courts and litigants alike, but they are a fundamental aspect of our federal republic and must not be overridden in a quest to clear the queue in court. See BMW v. Gore, 517 U.S. at 568-73, 116 S. Ct. 1589; Szabo (reversing a nationwide warranty class certification); Spence v. Glock, G.m.b.H., 227 F.3d 308 (5th Cir.2000) (reversing a nationwide tort class certification); Larry Kramer, Choice of Law in Complex Litigation, 71 N.Y.U. L.Rev. 547, 579 (1996); Linda S. Mullenix, Mass Tort Litigation and the Dilemma of Federalization, 44 DePaul L.Rev. 755, 781 (1995); Robert A. Sedler, The Complex Litigation Project's Proposal for Federally-Mandated Choice of Law in Mass Torts Cases: Another Assault on State Sovereignty, 54 La. L.Rev. 1085 (1994). Tempting as it is to alter doctrine in order to facilitate class treatment, judges must resist so that all parties' legal rights may be respected. Amchem Products, Inc. v. Windsor, 521 U.S. 591, 613, 117 S. Ct. 2231, 138 L. Ed. 2d 689 (1997).

The motion to certify questions of law to the Supreme Court of Michigan is denied as unnecessary in light of this opinion. The district court's order certifying two nationwide classes is REVERSED.