EASTERBROOK, Circuit Judge.

This appeal presents the question whether speculative transactions in foreign currency are "contracts of sale of a commodity for future delivery" regulated by the Commodity Futures Trading Commission. 7 U.S.C. § 2(a)(1)(A). Until recently almost all trading related to foreign currency was outside the CFTC's remit, even if an equivalent contract in wheat or oil would be covered. See Dunn v. CFTC, 519 U.S. 465, 117 S.Ct. 913, 137 L.Ed.2d 93 (1997) (describing the Treasury Amendment to the Commodity Exchange Act). But Congress modified the Treasury Amendment as part of the Commodity Futures Modernization Act of 2000, and today the agency may pursue claims that currency futures have been marketed deceitfully, unless the parties to the contract are "eligible contract participants". 7 U.S.C. § 2(c)(2)(B). "Eligible contract participants" under the Commodity Exchange Act are the equivalent of "accredited investors" in securities markets: wealthy persons who can look out for themselves directly or by hiring experts. 7 U.S.C. § 1a(12); 15 U.S.C. § 77b(a)(15). Defendants, which sold foreign currency to casual speculators rather than "eligible contract participants," are not protected by the Treasury Amendment except to the extent that it permits them to deal over-the-counter, while most other futures products are restricted to registered exchanges (called boards of trade) or "derivatives transaction execution facilities" (specialized markets limited to professionals).

The agency believes that some of the defendants deceived some of their customers about the incentive structure: salesmen said, or implied, that the dealers would make money only if the customers also made money, while in fact the defendants made money from commissions and markups whether the customers gained or lost. This allegation (whose accuracy has not been tested) makes it vital to know whether the contracts are within the CFTC's regulatory authority. The district judge concluded that the transactions are sales in a spot market rather than futures contracts. 2003 WL 22284295, 2003 U.S. Dist. LEXIS 17660 (N.D.Ill. Oct. 3, 2003).

AlaronFX deals in foreign currency. Two corporations doing business as "British Capital Group" or BCG solicited customers' orders for foreign currency. (Michael Zelener, the first-named defendant, is the principal owner and manager of these two firms.) Each customer opened an account with BCG and another with AlaronFX; the documents made it clear that AlaronFX would be the source of all currency bought or sold through BCG in this program, and that AlaronFX would act as a principal. A customer could purchase (go long) or sell (short) any currency; for simplicity we limit our illustrations to long positions. The customer specified the desired quantity, with a minimum order size of $5,000; the contract called for settlement within 48 hours. It is agreed, however, that few of BCG's customers paid in full within that time, and that none took delivery. AlaronFX could have reversed the transactions and charged (or credited) customers with the difference in price across those two days. Instead, however, AlaronFX rolled the transactions forward two days at a time — as the AlaronFX contract permits, and as BCG told the customers would occur. Successive extensions meant that a customer had an open position in foreign currency. If the dollar appreciated relative to that currency, the customer could close the position and reap the profit in one of two ways: take delivery of the currency (AlaronFX promised to make a wire transfer on demand), or sell an equal amount of currency back to AlaronFX. If, however, the dollar fell relative to the other currency, then the client suffered a loss when the position was closed by selling currency back to AlaronFX.

The CFTC believes that three principal features make these arrangements "contracts of sale of a commodity for future delivery": first, the positions were held open indefinitely, so that the customers' gains and losses depended on price movements in the future; second, the customers were amateurs who did not need foreign currency for business endeavors; third, none of the customers took delivery of any currency, so the sales could not be called forward contracts, which are exempt from regulation under 7 U.S.C. § 1a(19). This subsection reads: "The term `future delivery' [in § 2(a)(1)(A)] does not include any sale of any cash commodity for deferred shipment or delivery." Delivery never made cannot be described as "deferred," the Commission submits. The district court agreed with this understanding of the exemption but held that the transactions nonetheless were spot sales rather than "contracts . . . for future delivery." Customers were entitled to immediate delivery. They could have engaged in the same price speculation by taking delivery and holding the foreign currency in bank accounts; the district judge thought that permitting the customer to roll over the delivery obligation (and thus avoid the costs of wire transfers and any other bank fees) did not convert the arrangements to futures contracts.

In this court the parties debate the effect of Nagel v. ADM Investor Services, Inc., 217 F.3d 436 (7th Cir. 2000), and Lachmund v. ADM Investor Services, Inc., 191 F.3d 777 (7th Cir. 1999). These decisions held that hedge-to-arrive contracts in grain markets, which allow farmers to roll their delivery obligations forward indefinitely and thus to speculate on grain prices (while selling their crops on the cash market), are not futures contracts. The rollover feature offered by AlaronFX gives investors a similar option, and thus one would think requires a similar outcome. The CFTC seeks to distinguish these decisions on the ground that farmers at least had a cash commodity, which they nominally sold to the dealer that offered the hedge-to-arrive contract (though they did not necessarily deliver grain to that entity). AlaronFX and BCG acknowledge this difference but say that it is irrelevant; they rely heavily on Chicago Mercantile Exchange v. SEC, 883 F.2d 537, 542 (7th Cir. 1989), where we wrote:

A futures contract, roughly speaking, is a fungible promise to buy or sell a particular commodity at a fixed date in the future. Futures contracts are fungible because they have standard terms and each side's obligations are guaranteed by a clearing house. Contracts are entered into without prepayment, although the markets and clearing house will set margin to protect their own interests. Trading occurs in "the contract", not in the commodity. Most futures contracts may be performed by delivery of the commodity (wheat, silver, oil, etc.). Some (those based on financial instruments such as T-bills or on the value of an index of stocks) do not allow delivery. Unless the parties cancel their obligations by buying or selling offsetting positions, the long must pay the price stated in the contract (e.g., $1.00 per gallon for 1,000 gallons of orange juice) and the short must deliver; usually, however, they settle in cash, with the payment based on changes in the market. If the market price, say, rose to $1.50 per gallon, the short would pay $500 (50¢ per gallon); if the price fell, the long would pay. The extent to which the settlement price of a commodity futures contract tracks changes in the price of the cash commodity depends on the size and balance of the open positions in "the contract" near the settlement date.

These transactions could not be futures contracts under that definition, because the customer buys foreign currency immediately rather than as of a defined future date, and because the deals lack standard terms. AlaronFX buys and sells as a principal; transactions differ in size, price, and settlement date. The contracts are not fungible and thus could not be traded on an exchange. The CFTC replies that because AlaronFX rolls forward the settlement times, the transactions are for future delivery in practice even though not in form; and the agency insists that fixed expiration dates and fungibility are irrelevant. It favors a multi-factor inquiry with heavy weight on whether the customer is financially sophisticated, able to bear risk, and intended to take or make delivery of the commodity. See Statutory Interpretation Concerning Forward Transactions, 55 Fed.Reg. 39188, 39191 (Sept. 25, 1990). See also CFTC v. Co Petro Marketing Group, Inc., 680 F.2d 573, 577 (9th Cir. 1982).

Instead of trying to parse language in earlier decisions that do not wholly fit this situation, we start with the statute itself. Section 2(a)(1)(A) speaks of "contracts of sale of a commodity for future delivery". That language cannot sensibly refer to all contracts in which settlement lies ahead; then it would encompass most executory contracts. The Commission concedes that it has a more restricted scope, that it does not mean anything like "all executory contracts not excluded as forward contracts by § 1a(19)." What if there were no § 1a(19)? Until 1936 that exemption was limited to deferred delivery of crops. (Compare the Grain Futures Act of 1922, 42 Stat. 998 (1922), with the Commodity Exchange Act of 1936, 49 Stat. 1491 (1936).) Then until 1936 a contract to deliver heating oil in the winter would have been a "futures contract," and only a futures commission merchant could have been in the oil business! (Moreover, those contracts could have been secured only on boards of trade, because with rare exceptions, such as foreign currency under the Treasury Amendment, all futures contracts must be traded on exchanges or not at all.) Can it be that until 1936 all commercial contracts for future delivery of newspapers, magazines, coal, ice, oil, gas, milk, bread, electricity, and so on were unlawful futures contracts? Surely the answer is no, which means that "contract for future delivery" must have a technical rather than a lay meaning.

The Commission's candidate for that technical meaning is a multi-factor approach concentrating, as we have remarked, on the parties' goals and sophistication, plus the likelihood that delivery will occur. Yet such an approach ignores the statutory text. Treating absence of "delivery" (actual or intended) as a defining characteristic of a futures contract is implausible. Recall the statutory language: a "contract of sale of a commodity for future delivery." Every commodity futures contract traded on the Chicago Board of Trade calls for delivery. Every trader has the right to hold the contract through expiration and to deliver or receive the cash commodity. Financial futures, by contrast, are cash settled and do not entail "delivery" to any participant. Using "delivery" to differentiate between forward and futures contracts yields indeterminacy, because it treats as the dividing line something the two forms of contract have in common for commodities and that both forms lack for financial futures.

It may help to recall the text of § 1a(19): "The term `future delivery' does not include any sale of any cash commodity for deferred shipment or delivery." This language departs from the definition of a futures contract by emphasizing sale for deferred delivery. A futures contract, by contrast, does not involve a sale of the commodity at all. It involves a sale of the contract. In a futures market, trade is "in the contract." See Chicago Board of Trade v. SEC, 187 F.3d 713, 715 (7th Cir. 1999); Robert W. Kolb, Understanding Futures Markets (5th ed. 1997); Jerry W. Markham, The History of Commodity Futures Trading and its Regulation (1986); Louis Vitale, Interest Rate Swaps under the Commodity Exchange Act, 51 Case W. Res. L.Rev. 539 (2001).

In organized futures markets, people buy and sell contracts, not commodities. Terms are standardized, and each party's obligation runs to an intermediary, the clearing corporation. Clearing houses eliminate counterparty credit risk. Standard terms and an absence of counterparty-specific risk make the contracts fungible, which in turn makes it possible to close a position by buying an offsetting contract. All contracts that expire in a given month are identical; each calls for delivery of the same commodity in the same place at the same time. Forward and spot contracts, by contrast, call for sale of the commodity; no one deals "in the contract"; it is not possible to close a position by buying a traded offset, because promises are not fungible; delivery is idiosyncratic rather than centralized. Co Petro, the case that invented the multi-factor approach, dealt with a fungible contract, see 680 F.2d at 579-81, and trading did occur "in the contract." That should have been enough to resolve the case.

It is essential to know beforehand whether a contract is a futures or a forward. The answer determines who, if anyone, may enter into such a contract, and where trading may occur. Contracts allocate price risk, and they fail in that office if it can't be known until years after the fact whether a given contract was lawful. Nothing is worse than an approach that asks what the parties "intended" or that scrutinizes the percentage of contracts that led to delivery ex post. What sense would it make — either business sense, or statutory-interpretation sense — to say that the same contract is either a future or not depending on whether the person obliged to deliver keeps his promise? That would leave people adrift and make it difficult, if not impossible, for dealers (technically, futures commission merchants) to know their legal duties in advance. But reading "contract of sale of a commodity for future delivery" with an emphasis on "contract," and "sale of any cash commodity for deferred shipment or delivery" with an emphasis on "sale" nicely separates the domains of futures from other transactions.

Any contention that it is appropriate to ignore the contract's form and focus on economic effects — here, that rollover without full payment (AlaronFX allows customers to use margin while positions are open) can give the buyer the economic equivalent of a long position on a futures exchange — produces a sense of déjà vu. We've been here before, but in securities rather than commodities law. A business can be transferred two ways: the corporation may sell all of its assets, then liquidate and distribute to investors the cash received from the buyer; or the investors may sell their securities directly to the buyers. With sufficient care in drafting, these two forms may be made economically equivalent. This equivalence led to arguments that the sale of stock to transfer a whole business should not be regulated by the federal securities laws. Because the sale of assets would be governed by state contract law, it would upset expectations to handle the functionally equivalent transaction under federal law just because stock played a role. Many courts adopted this sale-of-business doctrine, but the Supreme Court rejected it, ruling that form must be respected. See Landreth Timber Co. v. Landreth, 471 U.S. 681, 105 S.Ct. 2297, 85 L.Ed.2d 692 (1985). One reason is that the securities laws are about form, and one can say much the same about the commodities laws. Another powerful reason was the need for certainty. The sale-of-business doctrine led to all sorts of questions. What if there were a significant minority shareholder? What if the new buyer did not plan to run the business as an entrepreneur? The list of questions turned out to be long and the uncertainty considerable — just as the CFTC's list of factors has made it hard to determine when rollovers turn spot or forward deals into futures contracts. By taking form seriously the Supreme Court was able to curtail, if not eliminate, that uncertainty and promote sensible business planning. See also Reves v. Ernst Young, 494 U.S. 56, 110 S.Ct. 945, 108 L.Ed.2d 47 (1990) (simplifying the approach to determining when notes are securities). Since the main battle in the sale-of-business cases was whether fraud litigation would occur in state or federal court, the Justices saw no reason for prolonged litigation about the forum: fraud is illegal in every state. So, too, for the definition of futures contracts. The Commission's principal substantive contention is that BCG deceived its clients. That could form the basis of a mail-fraud or wire-fraud prosecution, a civil or criminal action under RICO, or fraud litigation in state court. Consumers or state attorneys general could invoke consumer-protection laws as well. It is unnecessary to classify the transactions as futures contracts in order to provide remedies for deceit. Why stretch the Commodity Futures Act — with resulting uncertainty, litigation costs, and potentially unhappy consequences for other economic arrangements that may be swept into a regulatory system not designed for them — when other remedies are ready to hand?

Recognition that futures markets are characterized by trading "in the contract" leads to an easy answer for most situations. Customers of foreign exchange at AlaronFX did not purchase identical contracts: each was unique in amount of currency (while normal futures contracts are for fixed quantities, such as 1,000 bushels of wheat or 100 times the price of the Standard Poors 500 Index) and in timing (while normal futures contracts have defined expiration or delivery dates). Thus the trade was "in the commodity" rather than "in the contract." Cf. Marine Bank v. Weaver, 455 U.S. 551, 102 S.Ct. 1220, 71 L.Ed.2d 409 (1982) (a non-fungible contract that could not be traded on an exchange is not a security); Giuffre Organization, Ltd. v. Euromotorsport Racing, Inc., 141 F.3d 1216 (7th Cir. 1998) (a sports franchise linked to a single owner is not a security).

Citing Chevron U.S.A. Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984), the Commission contends that its position that rollovers turn spot sales into futures contracts should be respected without independent judicial inquiry. Yet when deciding what is (or isn't) a "security," courts have not deferred to the SEC; there is no greater reason to defer to the CFTC when defining futures contracts. In Dunn the Supreme Court addressed de novo the question whether an over-the-counter option on foreign currency was excepted under the pre-2000 version of the Treasury Amendment. Perhaps Dunn could be distinguished on the ground that the very name of the Treasury Amendment implied deference to its namesake, the Treasury Department (a possibility the Court mused about, 519 U.S. at 479 n. 14, 117 S.Ct. 913). But the central point is that deference depends on delegation. See United States v. Mead Corp., 533 U.S. 218, 121 S.Ct. 2164, 150 L.Ed.2d 292 (2001). When Congress has told an agency to resolve a problem, then courts must accept the answer. When, however, the problem is to be resolved by the courts in litigation — which is how this comes before us — the agency does not receive deference. Adams Fruit Co. v. Barrett, 494 U.S. 638, 649-50, 110 S.Ct. 1384, 108 L.Ed.2d 585 (1990). Courts must heed the agency's reasoning and give it the benefit of the doubt. But the CFTC has avoided rather than addressed the central issue: is trading "in the contract" a defining characteristic? The agency has assumed a negative answer without explanation. In Nagel, Chicago Mercantile Exchange, and other decisions, this circuit addressed the subject without extending Chevron deference to the Commission; we adhere to that position today. Both Nagel and Lachmund hold that rollovers of grain sales do not turn them into futures; it is hard to see why we should treat rollovers of currency sales differently.

Our decision in Nagel observed that in the great majority of situations even opinions emphasizing "the totality of the circumstances" boil down to whether trading has occurred in fungible contracts. 217 F.3d at 441. Best to take Occam's Razor and slice off needless complexity. We recognized a qualification, however: "there is an exception for the case in which the seller of the contract promises to sell another contract against which the buyer can offset the first contract, as in In re Bybee, 945 F.2d 309, 313 (9th Cir. 1991), and CFTC v. Co Petro Marketing Group, Inc., supra, 680 F.2d at 580. That promise could create a futures contract." Ibid. A promise to create offsets makes a given setup work as if fungible: although the customer can't go into a market to buy an equal and opposite position, the dealer's promise to match the idiosyncratic terms in order to close the position without delivery means that the customer can disregard the absence of a formal exchange.

Because the parties' briefs did not address this possibility, we inquired at oral argument whether the customers' contracts with BCG and AlaronFX entitled customers to close their positions by offset. Counsel for the CFTC answered with a ringing "yes" and counsel for the defendants with an equally confident "no." So they agree that the answer is clear; they just don't agree on what that answer is. Because neither side has made anything of the contracts between customers and BCG, we limit attention to the contracts that AlaronFX required all customers to sign. The Commission relies on paragraphs 5, 8, and 9 of that contract. Paragraph 8 provides that if a customer's account contains "two or more open and opposite Contracts providing . . . for the purchase and sale of the same Foreign Currency . . . on the same Value Date, such Contracts shall automatically be canceled and replaced by an obligation to settle only the net difference". So far, so good; offset is a standard feature of trading in the contract on a futures market. But the question we posed in Nagel is whether the dealer has promised to sell the offsetting position, and thus allow netting on demand. Such an obligation is harder to find in the AlaronFX contract.

Paragraph 9, on which the Commission places its principal reliance, does not contain any promise. What it says instead is that, if the client fails to give timely instructions about the disposition of the positions, then "AlaronFX is authorized, in AlaronFX's sole discretion, to deliver, roll over or offset all or any portion of the Open Position in Customer's Account at Customer's risk." This paragraph does not give the client a right to purchase an offsetting position and thus close a transaction; that option belongs to AlaronFX rather than to the customer. As for ¶ 5: subsection 5.3 says that "AlaronFX will attempt to execute all Orders that it may, in its sole discretion, accept from Customer in accordance with Customer's instructions. . . ." That's not a promise to close any given position by offset. This subsection continues: "AlaronFX or its affiliates may, at a future date, establish a trade matching system. . . . In that event, AlaronFX . . . shall have the right (but not the obligation) . . . to act for its own account, and as a counter party or as a broker to AlaronFX customers, in the making of markets". The Commission does not contend that such a "trade matching system" was ever established. Customers thus had no assurance that they could close their positions by offset. The only promise was that, if AlaronFX did not buy back the currency (and thus create an offset under ¶ 8), it would deliver. This looks more like the business of a wholesaler in commodities such as metals or rare coins than like the system of trading in fungible contracts that characterizes futures exchanges.

These transactions were, in form, spot sales for delivery within 48 hours. Rollover, and the magnification of gain or loss over a longer period, does not turn sales into futures contracts here any more than it did in Nagel and Lachmund. The judgment of the district court therefore is

AFFIRMED.