George E. Hemenway v. Peabody Coal Company and Peabody Development Company

EASTERBROOK, Circuit Judge. A mineral lease signed in 1969 requires Peabody Coal Company to pay a royalty based on the coal’s “sales price”. During the 1970s Congress enacted two excise taxes. One, for the benefit of the Black Lung Disability Trust Fund, is 55<t per ton of surface-mined coal, with a cap at 4.4% of the coal’s selling price. 26 U.S.C. § 4121. (The amount of this tax has varied since its enactment in 1978; 55<c is the current rate.) The other, designated a “reclamation fee” by the Surface Mining Control and Reclamation Act of 1977, is 35c per ton of surface-mined coal, capped at 10% of “the value of the coal at the *257 mine”. 30 U.S.C. § 1232(a). In 1983 Peabody began to charge its customers a “mine closing and final reclamation payment” of 25$ per ton. Peabody’s invoices list a per-ton charge for coal, the two taxes, and the mine closing fee. In an invoice that Peabody calls typical, the charge for coal was $28,635 per ton, and the three additional charges were $1.10 per ton. Peabody paid royalties only on the coal charge, which represented 96.3% of the invoice total. In this suit under the diversity jurisdiction, plaintiffs (assignees of the original lessors’ royalty interests) contend that the two taxes and the mine closing fee are part of the “sales price” for royalty-calculation purposes. The district court granted summary judgment to plaintiffs on this theory but rejected their further contention that Peabody defrauded them. To arrive at a judgment the district court also had to select a statute of limitations. Plaintiffs contend that the period is 20 years because this is a suit on a written contract; Peabody contends that it is 6 because an underpaid mineral royalty is delinquent rent in Indiana (whose law governs); the district court chose 8, adopting Peabody’s position but adding 2 years because these plaintiffs were members of the class in an abortive suit against Peabody in Kentucky. The final judgment is about $88,-000, plus $67,000 in prejudgment interest. Before we take up the merits, a few words about jurisdiction are in order. The complaint alleges plaintiffs’ residence rather than their citizenship. This defect is fixed by a proposed amendment tendered after the oral argument on appeal. See 28 U.S.C. § 1653. Post-argument motions explored a second jurisdictional issue: a national banking association, as trustee, is among the plaintiffs. The citizenship of a trust is the citizenship of the trustee. Navarro Savings Association v. Lee, 446 U.S. 458, 100 S.Ct. 1779, 64 L.Ed.2d 425 (1980). Peabody does not dispute that a trust exists as Navarro defines it (not always an easy question, see Indiana Gas Co. v. Home Insurance Co., 141 F.3d 314 (7th Cir.1998)), but there is a potential problem with the identity of the trustee. Some documents in the ease call the trustee “Bank One Trust Company, n.a.” This bank has a branch in Kentucky, where Peabody has its principal place of business. Some district courts have concluded that in light of 28 U.S.C. § 1348 national banking associations are citizens of every state in which they have branches, Ferraiolo Construction, Inc. v. KeyBank, n.a., 978 F.Supp. 23 (D.Me.1997) (collecting authority), and if this is right— and if Bank One Trust Company, n.a., is the trustee—then complete diversity of citizenship is missing and the case must be dismissed. But the amended complaint identifies “Bank One Ohio Trust Company, n.a.” as the trustee. Plaintiffs maintain (a) that Bank One Ohio Trust Company was the trustee when the suit began (the relevant time, see Freeport-McMoRan, Inc. v. K N Energy, Inc., 498 U.S. 426, 111 S.Ct. 858, 112 L.Ed.2d 951 (1991)), and (b) that Bank One Ohio Trust Company does not have a branch outside Ohio. Peabody informs us that it has no reason to doubt that Bank One Ohio Trust Company is indeed the trustee, so the parties have dodged a bullet. Plaintiffs’ motion to amend the complaint is granted. Jurisdiction now is secure—though it would have been much better had the district court handled these issues when the suit began. I The lease defines “sales price” unhelpfully as the “average invoice price of coal mined, removed and sold”. Are excise taxes part of the “average invoice price”? In one sense this is a trivial question. Of course they are. They appear on the invoice as part of the price the customer must pay. But the idea that a mine operator must pay the mineral owner a royalty on taxes, as opposed to coal, is jarring, and Peabody insists that the parties could not have contemplated this result in 1969. (The “mine closing” charge does not appear to represent an excise tax, but the parties make nothing of this. Like the parties, we treat all three surcharges as excise taxes.) A tax lowers the demand for coal, and its burden is distributed according to the elasticity of demand. ■ The district court supposed that demand for coal is perfectly inelastic, and if so buyers bear the whole tax and sellers are unaffected. Richard A. Posner, *258 Economic Analysis of Law 525-29 (5th ed.1998); Charles E. McLure, Jr., Incidence Analysis and the Supreme Court: An Examination of Four Cases from the 1980 Term, 1982 S.Ct. Econ. Rev. 69. If demand is not perfectly inelastic, then the sellers bear some of the loss, which participants on the seller’s side normally would share (or would have so agreed in advance, had they thought about the possibility). But if the excise taxes are included in “sales price,” then if elasticity is less than 1 the mineral owners will be better off and Peabody worse off. (If elasticity exceeds 1, both the mineral owners and Peabody can be worse off, because total revenues will fall.) Consider a simple' example. Suppose coal sells for $10 per ton, the royalty paid to the mineral owners is $1 per ton (10% of sales price), and Peabody makes a profit of $1 per ton. Now Congress adds an excise tax of $1 per ton, so that (at first approximation) the price rises to $11. If the tax is part of the “sales price” then the mineral owners’ royalty rises to $1.10 per ton and Peabody’s profit falls to 90$ per ton. If because of elastic demand Peabody cannot raise the price to $11 (or sells fewer tons at that price than at $10) then • Peabody’s share will be less than 90<t per ton, it will make that return on fewer tons, or both. Economists believe that many mineral and (other) property taxes fall on the land owner, because real property lacks alternative uses. See Richard A. Musgrave & Peggy B. Musgrave, Public Finance in Theory and Practice 412-13, 419-20 (5th ed.1989). This implies that if mineral owners and mine operators could bargain in advance they would be likely to assign the economic effects of future taxes to the mineral owners. (The tax is not on any asset specific to coal extraction, so operators that can bargain with many mineral owners won’t deal with those who seek to shift its incidence to wine operators.) A pact that makes the mineral owner better off, and the mine operator worse off, as a result of a tax makes little economic sense—plaintiffs have not suggested why they would have bargained for this result, or why a mine operator would have agreed to it—and leads to caution in employing a plain-language reading of the text. Especially when plaintiffs agree that a sales tax, nominally imposed on the buyer but collected and remitted by the seller, would not be part of the “sales price” under the contract. The choice between a tax formally on the seller (as these two taxes are, cf. Gurley v. Rhoden, 421 U.S. 200, 95 S.Ct. 1605, 44 L.Ed.2d 110 (1975)), and one formally on the buyer is arbitrary; ■ the two have identical economic and practical effects. Yet according to plaintiffs everything turns on this designation. But it takes more than an economic puzzle to justify giving an unusual spin to contractual language. Contracts allocate risks, and judicial decisions changing those allocations after the fact not only lead to expensive litigation (as each side invests in the pursuit of advantage) but also make the institution of contract less useful ex ante. Parties to contracts may prefer simple-minded textualism to costly disputes later on, even knowing that a plain-language reading sometimes goes wrong, and they may make adjustments—in other parts of the contract, or in the royalty rate—to compensate for this possibility. Or perhaps they just made a mistake and now rue the choice. Learned Hand remarked that “in commercial transactions it does not in the end promote justice to seek strained interpretations in aid of those who do not protect themselves” (James Baird Co. v. Gimbel Bros., Inc., 64 F.2d 344, 346 (2d Cir.1933)), and Indiana follows this precept by insisting that contracts be implemented as written even if the results are unexpected. P.C. Management, Inc. v. Page Two, Inc., 573 N.E.2d 434, 438 (Ind.App.1991). Indiana also is of the view that federal excise taxes are part of a product’s selling price. Sun Oil Co. v. Gross Income Tax Division, 238 Ind. 111, 149 N.E.2d 115 (1958). Although Sun Oil did not involve the interpretation of a contract, the state’s understanding of excise taxes as part of a product’s price is the norm. See Agron v. Illinois Bell Telephone Co., 449 F.2d 906 (7th Cir.1971) (same understanding as a matter of federal law). To justify excluding from “average invoice price” elements of price that appear on the invoice, Peabody needs more than an observation that the contractual outcome surprises the bench. *259 One of Peabody’s two themes emphasizes “of coal” in the formula “average invoice price of coal”. The excise taxes were not part of the price of the coal; they were surcharges. Yet this line of argument proves far too much. It gives Peabody the power to curtail royalties by changing how it designs the invoice. Recall our example, in which coal was $10 per ton. Peabody’s position implies that if it broke out the components — coal $3, labor $3, machinery $3, environmental precautions and waste removal $1 — it would need to pay a royalty only on the component explicitly identified as “coal.” That’s nót a plausible reading of the contract, especially not given that other taxes (income and payroll taxes, plus the sales and excise taxes Peabody pays on the purchase of equipment) and the costs of satisfying the epa’s environmental regulations are indubitably part of the “average invoice price of coal”. Peabody’s other theme is that federal excise taxes on coal were introduced after 1969, which, Peabody insists, necessarily renders the contract ambiguous. How could the parties have resolved the treatment of nonexistent taxes?, Peabody asks. Peabody wants us to deem the treatment of excise taxes an omitted term, on which “the court will admit evidence to show how the parties would have dealt with the contingency had they made specific provision for it.”. Amoco Oil Co. v. Ashcraft, 791 F.2d 519, 521 (7th Cir.1986). This in turn, Peabody insists, means that summary judgment is inappropriate. Yet contracts often handle contingencies. Peabody’s own contracts with the utilities that use its coal to make electricity permit it to add excise taxes to the price of coal (which is why the invoices were structured as they were). Many of these were negotiated before 1977, when the first federal excise tax was enacted, so the contention that no contract signed in 1969 can mean anything with respect-to excise taxes falls flat. The 1969 contract does not address taxes with the same directness as the sales contracts Peabody negotiated, but plaintiffs insist that this is because it did not have to; the “average invoice price” formula does the necessary work. What evidence has Peabody to undermine this conclusion? It does not want to argue industry custom or usage of trade; its fundamental position is that there was in 1969 no relevant custom or usage, and anyway the formula “average invoice price” is an unusual one. (A search of online legal databases turned up no references to this formula.) Peabody does not want to offer parol evidence of the negotiations preceding the contract’s formation; its position that no one thought about the subject is incompatible with this kind of evidence. Nor does Peabody seek to show how other mineral leases from the 1960s handle taxes, when the parties dealt with that possibility expressly. Such evidence could enable a court to address the question posed in Amoco Oil, but it is not on the horizon. (Documents in the record suggest that mineral leases from the 1990s cover excise taxes explicitly; some include the taxes in the royalty base while others do not.) Evidence about the beliefs, wishes, hopes, and fears of its negotiators and managers would be inadmissible, for only objective evidence may be used to elaborate on a contract’s meaning. See AM International, Inc. v. Graphic Management Associates, Inc., 44 F.3d 572 (7th Cir.1995); Bristow v. Drake Street Inc., 41 F.3d 345 (7th Cir.1994). The sort of objective evidence that Peabody does proffer has no bearing on the meaning of a 1969 contract. For example, Peabody wants to argue to a jury that Congress differentiated the excise taxes from the price of coal. Each statute limits the tax to a percentage of the price or value of coal; this supposes, Peabody believes, that the “price” of coal does not include the tax. This evidence would be unhelpful — and not only because federal judges do not hold trials so that jurors may deliberate on the meaning of federal statutes. See DePaepe v. General Motors Corp., 141 F.3d 715, 717 (7th Cir.1998). Congress excluded the tax from “price” to make the formula determinate: a cap of 10% is ambiguous unless we know the denominator — the price of the coal alone, or the price of coal and tax together? Nothing in either 26 U.S.C. § 4121 or 30 U.S.C. § 1232 suggests that the term “sales price” *260 in a private contract must mean the same thing as “price” for the purpose of calculating a maximum tax, let alone that the federal definition of “price” from the 1970s applies retroactively to contracts signed in the 1960s. Many features of the statutes depart from contractual norms. For example, a “ton” in the statutory framework is a ton of dry coal, rather than of coal plus excess moisture (which the contract may permit the seller to deliver). See Amax Coal Co. v. United States, 128 F.3d 613 (7th Cir.1997). Nothing but grief could come from trying to impress the statutory diction on an older contract that uses the same term in a different context. In the end, the parties do not dispute any issue of material fact; they just offer different interpretations of their agreement. Whatever meaning the contract conveys must be found within its four corners plus the uncontested economic context (such as the tax provisos in Peabody’s sales contracts). That the task of drawing out this meaning is difficult does not transfer the job from a judge to a jury. See PSI Energy, Inc. v. Exxon Coal USA, Inc., 17 F.3d 969 (7th Cir.1994) (another complex coal contract dispute under Indiana law). Context implies that the phrase “average invoice price” includes excise taxes. These parties gave a good deal of thought to the potential for change in the coal market during the contract’s lengthy term. Peabody must pay a royalty of 12$ per ton, plus 10% of the amount by which the “sales price” exceeds a “base price.” The base price was pegged at $2.28 per ton in 1969, but it does not stay there. It rises (or falls) as the cost of mine labor rises (or falls). There is a separate exclusion of 50$ per ton for coal shipped on barges at the Yankeetown Dock Corporation. That, however, is the end of the list. Higher wages for workers thus yield a higher “base price” (and prevent a change in royalties even if the labor costs are fully passed on to Peabody’s customers), but higher costs of heavy earth-moving equipment do not change the base price. If Peabody substitutes capital for labor it must pay greater royalties than if it substitutes labor for capital. Peabody pays many taxes that do not affect the “base price,” and which therefore lead to greater royalties for the mineral owners if the taxes affect other firms in the energy business and permit an increase in the price of coal without reducing consumption so much that total income falls. Equally apropos, since 1969 Peabody has been required to follow an increasingly expensive regimen of regulatory constraints designed to make the coal business cleaner and healthier. The Clean Air Act of 1970 and the Clean Water Act of 1970 were but opening salvos. The Surface Mining Control and Reclamation Act of 1977 is only one among many follow-up statutes, cercla, SARA, RCRA, and the rest of the alphabet soup of environmental regulation all postdate this contract. Sometimes the federal government levies taxes to fund environmental cleanup, but more often it requires firms in the business to bear these costs directly. Many of the costs of environmental compliance require capital expenditures (or outlays to third-party contractors, such as environmental engineers) and therefore do not affect the “base price” even though they elevate the “sales price.” It is- safe to say that these costs of environmental compliance and cleanup substantially exceed 35$ per ton, yet all (other than labor outlays) increase the gap between “base price” and “sales price”, and therefore increase plaintiffs’ royalties even as they decrease Peabody’s profits. We observed above that the choice between an excise tax on the seller and a sales tax on the buyer (but collected by the seller) is arbitrary; the political choice between a tax and additional regulations can be equally arbitrary. But this contract unambiguously includes regulatory costs in the royalty base, which implies that excise taxes in lieu of regulations also are in that base. This contract effectively requires Peabody to pay royalties on its gross billings, less the “base price” and the Yankeetown barge allowance. Excise taxes are omitted from both the “base price” and the barge allowance. It would not have required prescience for parties to subtract taxes as well, had they wanted to do so; even in 1969, death and taxes had a certain inevitability. One more consideration supports this conclusion. Although few of Peabody’s con *261 tracts use “average invoice price of coal mined, removed and sold” as a benchmark, many more use the term “gross realization”. Peabody’s brief treats these terms as equivalent, which suggests that the treatment of excise taxes under “gross realization” contracts is instructive. A number of contracts with Indian tribes define “gross realization” as “the gross sales price at the mining site without any deduction therefrom of overhead sales costs or other business expenses.” The Interior Board of Land Appeals concluded that the excise taxes must be included in the “gross sales price” because the buyer pays the tax as part of the price, even if the United States is the ultimate recipient. Peabody Coal Co., 53 I.B.L.A. 261 (1981); Peabody Coal Co. v. Hopi Indian Tribe, 72 I.B.L.A. 337 (1983). A panel of arbitrators reached the same conclusion regarding a commercial coal contract containing that term, and several state and federal judges in states other than Indiana have done likewise. Peabody emphasizes that none of these cases has been tested on appeal, and that one of the gross realization cases ended in its favor when a jury concluded that the excise taxes are not part of “gross sales price”. Willits v. Peabody Coal Co., No. 4:90-CV-34-O(C) (W.D.Ky. Mar. 12, 1998), appeal pending (6th Cir.). We do not put any weight on these decisions, for we must decide the question independently. Salve Regina College v. Russell, 499 U.S. 225, 111 S.Ct. 1217, 113 L.Ed.2d 190 (1991). They show, however, that royalties on excise taxes are not such a startling proposition that courts have bridled at giving the contractual language its most natural reading. Like the district court, we hold that the excise taxes are included in the “sales price” for purposes of the 1969 lease. II Plaintiffs want to add punitive damages to their winnings. Indiana does not permit the use of punitive damages to redress breach of contract, see Miller Brewing Co. v. Best Beers of Bloomington, Inc., 608 N.E.2d 975, 981 (Ind.1993), so plaintiffs characterize Peabody’s conduct as fraud, an independent tort. They face an uphill battle, because Peabody never told a lie. The contract specifies the contents of royalty statements: Peabody must disclose how many tons of coal were mined and removed from the land. It need not state the “sales price” or the “base price” in any given month, or reveal how either was determined. Peabody remained silent about these matters, as was its right under the contract, and as Peabody is not plaintiffs’ fiduciary it is hard to portray silence as a deceptive omission. Plaintiffs essentially say that Peabody committed fraud because it did not alert them to the breach of contract, but if this were fraud then Indiana’s prohibition of punitive damages in contract eases would be undone. The district court rejected this claim, as has the only other court to address a fraud claim under Indiana law arising out of Peabody’s handling of the excise taxes. Hendrickson v. Peabody Coal Co., No. 4:95-CV-224-M (W.D.Ky. Oct. 7, 1997), appeal pending, No. 98-5066 (6th Cir.). Unlike the contract claim, which was decided on summary judgment, the fraud claim was dismissed under Fed. R. Civ. P. 12(b)(6). One ground of this decision—the district court’s observation that plaintiffs did not plead reliance on Peabody’s silence—is mistaken. Complaints need not spell out every element of a legal theory; that’s the big difference between notice and code pleading. See Bennett v. Schmidt, 153 F.3d 516 (7th Cir.1998); Bartholet v. Reishauer A.G. (Zurich), 953 F.2d 1073 (7th Cir.1992). What a claim of fraud must include, however, is “particularity” about the details—“the who, what, when, where, and how: the first paragraph of any newspaper story.” DiLeo v. Ernst & Young, 901 F.2d 624, 627 (7th Cir.1990). See Fed. R. Civ. P. 9(b). Plaintiffs’ complaint does not allege, with or without particularity, that Peabody told a lie, so the district judge was entitled to limit his discussion to the possibility of deceptive omissions. And because plaintiffs related exactly what those omissions were, the judge was entitled to resolve the claim on the merits. Plaintiffs pleaded themselves out of court on the fraud theory by revealing that they located the “fraud” in omissions from the royalty statements—for the statements conformed to the contract the parties had negotiated. Peabody did not include with royalty statements before the excise taxes *262 any information about “sales price,” “base price,” or the calculation of these figures; the identical omission after 1976 did not suddenly become fraudulent concealment. Plaintiffs stress that under Indiana law a non-fiduciary may have a duty to reveal information to its trading partner, if it has superior access to the facts and the other side is entitled to rely on its disclosure. Colonial Penn Insurance Co. v. Guzorek, 690 N.E.2d 664, 675 (Ind.1997); see also Midwest Commerce Banking Co. v. Elkhart City Centre, 4 F.3d 521, 524 (7th Cir.1993) (collecting Indiana cases in which failure to reveal a fact implied a belief in its opposite). Indiana insists that the “entitlement” or “duty” part of this formulation for a constructive fraud claim be taken seriously. Rice v. Strunk, 670 N.E.2d 1280, 1284 (Ind.1996); see also Vaughn v. General Foods Corp., 797 F.2d 1403, 1413-14 (7th Cir.1986) (Indiana law). Parties are free to bargain about disclosure obligations, and these parties did so. Plaintiffs agreed (well, their predecessors agreed) to disclosure of the tons of coal mined (but not price information), plus an opportunity to audit Peabody’s books on demand. The audit clause speaks directly only to tonnage, but by allowing the audit to extend to the “sale” of coal it implies that the mineral owners can review the shipping documents, such as the invoice's. Plaintiffs do not contend that Peabody has ever refused any mineral owner access to invoices and other price information on demand; it is undisputed that when in 1978 Beaver Dam Coal Company, a mineral owner situated similarly to plaintiffs, inquired how Peabody was handling the new excise taxes and asked to see the books, they were promptly handed over. The existence of the taxes was public knowledge, so mineral owners had good reason to ask how Peabody was handling them. Cf. Eckstein v. Balcor Film Investors, 58 F.3d 1162 (7th Cir.1995) (failure to reveal something already in the public record is not securities fraud); Wielgos v. Commonwealth Edison Co., 892 F.2d 509 (7th Cir.1989) (same). Peabody neither lied to nor stonewalled Beaver Dam, and plaintiffs do not offer any reason to suppose that Peabody would have interfered with their exercise of their own audit rights. But plaintiffs, unlike Beaver Dam, did not ask, even after two Indian tribes had litigated their contractual claims to a successful (and public) conclusion in 1981 and 1983. Indiana disfavors punitive damages even when outright lies have been established. “The correct standard ... is whether ... the defendant acted with malice, fraud, gross negligence or oppressiveness which was not the result of a mistake of fact or law, honest error of judgment, overzealousness, mere negligence, or other human failing.” Budget Car Sales v. Stott, 662 N.E.2d 638, 639 (Ind.1996). See also Erie Insurance Co. v. Hickman, 622 N.E.2d 515, 520 (Ind.1993); Bud Wolf Chevrolet, Inc. v. Robertson, 519 N.E.2d 135, 137-38 (Ind.1988). We held in Creative Demos, Inc. v. Wal-Mart Stores, Inc., 142 F.3d 367, 371 (7th Cir.1998), that the phrase “which was not the result of a mistake of fact or law, honest error of judgment, overzealousness, mere negligence, or other human failing” modifies everything in the preceding list, including “fraud.” This standard independently supports the district judge’s conclusion, for the difficulty of the underlying contract question, and the disparate results that have marked the litigation between Peabody and other mineral owners, show that Peabody’s conduct is best characterized as mistaken rather than fraudulent. Only a few words need be said about plaintiffs’ contention that the district court should have compelled Peabody to produce certain documents in discovery before dismissing the complaint. Discovery and Rule 12(b)(6) are unrelated. A complaint may be dismissed under Rule 12(b)(6) only if it would be impossible to prevail, consistent with the pleading. Application of this standard requires the court to assume that the plaintiffs will be able to prove whatever allegations they care to make, and we have assumed for purposes of this opinion that plaintiffs would be able to prove everything they sought to establish by discovery. What they wanted to obtain—Peabody’s internal documents showing that at least some of its executives and lawyers viewed its handling of the excise taxes as problematic under “gross realization” contracts, and papers from the settlement Peabody reached with Beaver Dam—is *263 more in the nature of litigation analysis than proof of fraud. Large corporations generate many options papers, but the choice of one option over another in a world of legal uncertainty is a long way from proof of fraud. Plaintiffs’ fraud claim fails because they were not lied to, because their contract specifies what information Peabody was to disclose (and when), and because there really is legal uncertainty about the meaning of the contract. None of these is undercut in the slightest by the documents plaintiffs wanted to use. Ill Plaintiffs began this suit in 1992, fifteen years after Congress enacted the first of the excise taxes. They want damages for the entire period, but the district court limited the recovery to eight years—six from the statute of limitations the court selected, plus two years of tolling. We’ll come to tolling eventually but start with the question whether the basic period of limitations is six years or twenty. A Two statutes are in the picture. I.C. § 34-11-2-7 (formerly I.C. § 34-1-2-1) provides: The following actions must be commenced within six (6) years after the cause of action accrues: (1) Actions on accounts and contracts not in writing. (2) Actions for use, rents, and profits of real property. (3) Actions for injuries to property other than personal property, damages for detention of personal property and for recovering possession of personal property. (4) Actions for relief against frauds. And I.C. § 34-11-2-11 (formerly I.C. § 34-1—2—2(6)) provides: An action upon contracts in writing other than those for the payment of money, and including all mortgages other than chattel mortgages, deeds of trust, judgments of courts of record, and for the recovery of the possession of real estate, must be commenced within ten (10) years after the cause of action accrues. However, an action upon contracts in writing other than those for the payment of money entered into .before September 1, 1982, not including chattel mortgages, deeds of trust, judgments of courts of record, or for the recovery of the possession of real estate, the action must be commenced within twenty (20) years after the cause of action accrues. The district court concluded that plaintiffs’ action is one “for use, rents, and profits of real property”, § 34-11-2-7(2), and therefore covers only the period of six years before its filing. Plaintiffs deny that coal royalties are “rents” or “profits” of real property, but they would be content with a conclusion that if § 34-11-2-7(2) applies, then so does § 34-11-2-11 (the lease is in writing); and because the lease was signed before September 1, 1982, this statute supplies a 20-year period. When two statutes of limitations apply, plaintiffs contend, Indiana uses the longer. See Northern Indiana Public Service Co. v. Fattore Construction Co., 486 N.E.2d 633, 634 (Ind.App.1985), disapproved on other grounds by Berns Construction Co. v. Miller, 516 N.E.2d 1053 (Ind.1987). Peabody is entirely content with the idea that both statutes apply; it appeals to the maxim that when two statutes apply to the same conduct the more specific governs. See State v. Pulaski Circuit Court, 264 Ind. 37, 338 N.E.2d 634 (1975); Ferguson v. Modern Farm Systems, Inc., 555 N.E.2d 1379, 1383-84 (Ind.App.1990). On this ground the only other excise tax case that has arisen under Indiana law applied the six-year period. Hendrickson, slip op. 5-8. It has the support of Hellyer Communications, Inc. v. WRC Properties, Inc., 888 F.Supp. 94 (S.D.Ind.1995), which selected the 6-year period over the 20-year period when both applied. Both statutes cover underpayment of mineral royalties under a lease. Because the lease is a “contract in writing other than [one] for the payment of money entered into before September 1, 1982,” the 20-year period of § 34-11-2-11 applies directly. Because the contract is a lease, providing for (in its own terms) “royalties or rentals”, § 34-11-2-7(2) fits equally well. Mineral leases *264 often use rents and royalties as synonyms, see Bicknell Minerals, Inc. v. Tilly, 570 N.E.2d 1307, 1313 (Ind.App.1991), and although Bicknell Minerals observes that equation between the two terms is not inevitable, they were equated in this lease. Even if the tonnage payments are not “rents” (though the lease calls them rents), they represent from plaintiffs’ perspective the “profits of real property”. Plaintiffs contend that § 34-11-2-7(2) applies only to oral leases, but the statute is not so restricted, even if at the time of its (first) enactment many leases and sharecropping arrangements were unwritten. The statute has been reenacted and recodified since — most recently last year — and must be understood to govern contemporary transactions affecting the use, rents, and profits of real property, which today are overwhelmingly written. Moreover, § 34-11-2-7(1) covers unwritten agreements directly; if § 34-11-2-7(2) is limited to oral contracts, it is redundant with § 34-11-2-7(1). To accept plaintiffs’ construction would be to consign § 34-11-2-7(2) to the garbage bin. Thus we arrive at the battle of maxims— and what an unedifying spectacle it is when each side lobs volleys with its own legal canons. Take the maxim that the more specific statute prevails. That’s all well and good when one statute is a subset of another. Suppose § 34-11-2-7(2) applied only to written leases. Then every suit within the scope of § 34-11-2-7(2) would be within § 34-11-2-11, and unless Indiana applied § 34-11-2-7(2) to suits seeking rents or profits from real estate, the statute would yield to § 34-11-2-11 in every case. When one statute applies to a subset of another, the canon about general and specific statutes does not really do anything; all the useful work is done by the presumption that every statute serves a function. When the two statutes apply to intersecting sets, however, neither is more specific (or perhaps both are more specific). Section 34-11-2-7(2) applies to every action concerning rents and profits from real property, whether or not the contract is in writing; § 34-11-2-11 applies to every action based on a written contract, whether or not the contract concerns real property. Which is more “specific”? That depends on which end of the telescope you look through. Compare Radzanower v. Touche Ross & Co., 426 U.S. 148, 153-54, 96 S.Ct. 1989, 48 L.Ed.2d 540 (1976), with id. at 164-66, 96 S.Ct. 1989 (Stevens, J., dissenting). As for the supposed canon that the longer statute prevails: that’s not a principle of interpretation but a decision by fiat. It subordinates one law to the other without asking whether that would defeat the law’s function (indeed, leave it any role at all). Although Fattore announces this approach, it does not cite any Indiana case in support. Fattore’s holding was that a 10-year period of limitations for suits against architects displaces the state’s 2-year period of limitations for personal injuries, when an architect is the defendant in a personal-injury case. Bems Construction disapproved that holding and directed Indiana courts to use the two-year period. Although it did not mention the supposed canon that led Fattore to its conclusion, the canon cannot coexist with the holding of Bems Construction. The Supreme Court of Indiana treated the 10-year period as shortening longer periods (such as the 20-year period in § 34-11-2-11) that might otherwise apply to suits against architects, not as lengthening shorter periods. Only that approach gave the architect statute any function. So we are confident that Fattore does not represent the law of Indiana, and that state courts would try to determine which of two competing statutes of limitations more closely fits the situation at hand. Which leaves the fundamental question. Like the district court, we think that § 34-11-2-7(2) must win the conflict between periods of limitations. Otherwise it has become surplusage, and the anti-surplusage canon is in Indiana’s arsenal. Indiana Department of Environmental Management v. Chemical Waste Management, 643 N.E.2d 331, 339 (Ind.1994). Section 34-11-2-7(1) sets a six-year period for actions on oral contracts, so the only scope for independent operation of § 34-11-2-7(2) is written agreements that affect the use, rents, or profits of real property. If the existence of a written agreement were enough to set a' 10-year or 20-year period under § 34-11-2-11, however, then § 34-11-2-7(2) might as well be erased from *265 the statute books. It is impossible to be certain how Indiana would resolve the question; no state case that we could find addresses the issue, and it is unfortunate that the problem recurs in federal litigation while the state courts are silent. Still, our best estimate is that Indiana would use the six-year period. B Indiana offers only a little more guidance on the final issue: whether the period of limitations was tolled during the pendency of a putative class action filed in 1990 on behalf of Peabody’s lessors. After that suit had been pending for three years, it was dismissed for lack of subject-matter jurisdiction. Kelce v. Peabody Coal Co., No. 90-0148-0, 1993 WL 35249 (W.D.Ky. July 29, 1993). A year before the district court dismissed Kelce, plaintiffs filed their own action in Indiana (and have managed, more by luck than by design, to avoid their own jurisdictional pitfall). Relying on American Pipe & Construction Co. v. Utah, 414 U.S. 538, 94 S.Ct. 756, 38 L.Ed.2d 713 (1974), the district court concluded that the time between the filing and the dismissal of Kelce must be added to the period of limitations; as a result, the district judge held that plaintiffs are entitled to damages for the six-year period before Kelce was filed, rather than the six-year period before this suit was filed. Both American Pipe and its sequel Crown, Cork & Seal Co. v. Parker, 462 U.S. 345, 103 S.Ct. 2392, 76 L.Ed.2d 628 (1983), dealt with claims under federal law, for which the period of limitations was also federal; this enabled the Supreme Court to craft tolling rules as a matter of federal law. When state law supplies the period of limitations, it also supplies the tolling rules. Hardin v. Straub, 490 U.S. 536, 109 S.Ct. 1998, 104 L.Ed.2d 582 (1989); Chardon v. Fumero Soto, 462 U.S. 650, 103 S.Ct. 2611, 77 L.Ed.2d 74 (1983); Gonzalez v. Entress, 133 F.3d 551 (7th Cir.1998). Only one Indiana case has addressed the question whether the filing of a class action satisfies the period of limitations for all class members. Arnold v. Dirrim, 398 N.E.2d 426 (Ind.App.1979), holds that it does, relying on American Pipe, but there is a critical difference: in Arnold the court permitted the case to proceed as a class action, while in American Pipe and Kelce the original ease did not go forward as a class action. Arnold does express the view, however, that the period of limitations is tolled between the filing of a class action and the order deciding whether the suit may be maintained as a class action, a principle that covers both grants and denials. What remains to be decided is whether this approach covers a ease in which the class allegation came to nothing, and was never ruled on one way or the other, because the court lacked subject-matter jurisdiction. Indiana has never expressed a view on that question — nor, as far as we can tell, has any federal court. Peabody asks us to limit American Pipe to cases in which, after the order refusing to certify a class, the members of the class intervene in the original case, but neglects to mention that this possibility was considered and rejected by the Supreme Court in Crown, Cork & Seal, 462 U.S. at 349-51, 103 S.Ct. 2392. Arnold suggests that Indiana is likely to follow the Supreme Court’s understanding of American Pipe. The Justices treated the American Pipe doctrine as designed to protect reliance interests and dissuade potential plaintiffs from filing a plethora of suits to protect their interests while the chance remains that one suit will be enough, not necessarily to funnel all litigation to the court in which the original suit began. Perhaps, then, one could say that tolling is limited to plaintiffs who know about and rely on the putative class action — but a knowledge- and-reliance requirement would make a mess of any subsequently filed action by making it necessary for the judge to determine the contents of the parties’ heads, and a rule that American Pipe applies only if the class members refrain from filing their own suits would set a trap for the unwary and turn reasonable precautions (such as filing standby suits just in case class certification is denied) against the class members. In American Pipe the court remarked that tolling ends as soon as the district court declines to certify the case as a class action, and the clock begins ticking with what may be very little time left (in American Pipe, only 11 days, *266 see 414 U.S. at 561, 94 S.Ct. 756). Later cases have confirmed that time begins immediately, rather than after final judgment or decision on appeal. See Armstrong v. Martin Marietta Corp., 138 F.3d 1374 (11th Cir.1998) (en banc) (collecting authority). (Armstrong explains why dicta to the contrary in Jimenez v. Weinberger, 523 F.2d 689, 696 (7th Cir.1975), should not be followed.) Other class members may not learn of this action until too late, and concern about delay in notification may lead them to file protective suits. Precautions of this kind cannot sensibly be used to abbreviate the period of limitations. Well, then, should it matter that Kelce was dismissed for want of subject-matter jurisdiction rather than because class status was inappropriate? Both Armstrong and Basch v. Ground Round, Inc., 139 F.3d 6 (1st Cir.1998), which holds that tolling from multiple unsuccessful class actions can’t be “stacked,” suggest that Crown, Cork & Seal sets the outer limit of the American Pipe doctrine. A concurring opinion by three Justices in Crown, Cork & Seal observed that the tolling doctrine “is a generous one, inviting abuse.” 462 U.S. at 354, 103 S.Ct. 2392 (Powell, J., concurring). But the “abuse” about which these Justices expressed concern was raising new or peripheral claims only tangentially connected with the original class suit. As all nine Justices recognized, a complaint seeking certification as a class action notifies the defendant of the claim and the potential scope of relief. For many purposes it is best to speak of such a complaint as satisfying, rather than tolling, the statute of limitations. Many states, of which Indiana is one, have enacted what are quaintly called journeys account statutes, permitting plaintiffs to refile dismissed suits within some period, provided the original suit was timely. In Indiana the plaintiff has three years to refile, unless the dismissal was attributable to negligence in its prosecution, abatement, or an adverse decision on the merits. I.C. § 34-11-8-1 (formerly I.C. § 34-1-2-8). The refiled action is treated as a “continuation of the original action”. So if plaintiffs themselves had been among the named parties in Kelce, they could have refiled the action in Indiana as a “continuation,” using Kelce’s 1990 filing date for limitations purposes. See also 28 U.S.C. § 1367(d) (same rule for claims sent to state court when a federal judge declines to exercise supplemental jurisdiction). What American Pipe and Arnold establish is that until class certification is denied, the members of the putative class receive the same treatment (for statute of limitations purposes) as do the named plaintiffs who seek to represent the class. Because our plaintiffs would be entitled to the benefit of the 1990 filing date if they had been named plaintiffs in Kelce, the district court’s conclusion follows directly. In states such as Indiana with journeys account statutes, the failure of a class action for jurisdictional reasons must be treated like a failure for some other reason, such as lack of numer-osity or typicality. See Fed. R. Civ. P. 23(a). AFFIRMED