EASTERBROOK, Circuit Judge.
The Commodity Futures Trading Commission has authority to regulate trading of futures contracts (including futures on securities) and options on futures contracts. The Securities and Exchange Commission has authority to regulate trading of securities and options on securities. If an instrument is both a security and a futures contract, the CFTC is the sole regulator because “the Commission shall have exclusive jurisdiction with respect to ... transactions involving ... contracts of sale (and options on such contracts) for future delivery of a group or index of securities (or any interest therein or based upon the value thereof)”, 7 U.S.C. § 2a(ii). See also 7 U.S.C. § 2 (“the Commission shall have exclusive jurisdiction, except to the extent otherwise provided in section 2a of this title”); Chicago Board of Trade v. SEC, 677 F.2d 1137 (7th Cir.), vacated as moot, 459 U.S. 1026, 103 S.Ct. 434, 74 L.Ed.2d 594 (1982) (GNMA Options). If, hov/ever, the instrument is both a futures contract and an option on a security, then the SEC is the sole regulator because “the [CFTC] shall have no jurisdiction to designate a board of trade as a contract market for any transaction whereby any party to such transaction acquires any put, call, or other option on one or more securities ... including any group or index of such securities, or any interest therein or based on the value thereof.” 7 U.S.C. § 2a(i).
The CFTC regulates futures and options on futures; the SEC regulates securities and options on securities; jurisdiction never overlaps. Problem: The statute does not define either “contracts ... for future delivery” or “option” — although it says that “ ‘future delivery’ ... shall not include any sale of any cash commodity for deferred shipment or delivery”. See Lester G. Telser, Futures and Actual Markets: How They Are Related, 59 J. Business S5 (1986). Each of these terms has a paradigm, but newfangled instruments may have aspects of each of the prototypes. Our case is about such an instrument, the index participation (IP). We must decide whether tetrahedrons belong in square or round holes.
I
Index participations are contracts of indefinite duration based on the value of a basket (index) of securities. The seller of an IP (called the “short” because the writer need not own the securities) promises to pay the buyer the value of the index as measured on a “cash-out day”. Any index, such as the Standard & Poor's 500, can be used. The buyer pays for the IP in cash on the date of sale and may borrow part of the price (use margin) on the same terms the Federal Reserve sets for stock — currently 50%. The exchange designates a conversion ratio between the index and the IP, so that (say) each IP unit entitles the holder on cash-out day to the value of the index times 100. Until cash-out the IP may trade on the exchange just like any other instru[540]*540ment. At the end of each quarter the short must pay the buyer (the “long”) a sum approximating the value of dividends the stocks in the index have paid during the quarter. From the perspective of the long, then, an IP has properties similar to those of a closed-end mutual fund holding a value-weighted portfolio of the securities in the index: the IPs last indefinitely, pay dividends, and may be traded freely; on cash-out day the IP briefly becomes open-end, and the investor can withdraw cash without making a trade in the market.
Things differ from the short’s perspective. Unlike the proprietor of a mutual fund, the short need not own the securities in the index; it will own them (equivalently, a long futures contract based on the same index) only to reduce risk. The short receives the long’s cash but must post margin equal to 150% of the value of the IP, similar to the margin required for a short sale of stock. The short sees the IP as a speculative or hedging instrument scarcely distinguishable from a futures contract that terminates on the cash-out day, plus an option held by the long to roll over the contract to the next cash-out date. Cash-out days for an IP generally are the third Friday of March, June, September, and December, the expiration dates of the principal stock-index futures contracts, making the link even more apparent.
Longs and shorts do not deal directly with each other. After the parties agree on the price, the Options Clearing Corporation (OCC) issues the IP to the long, receiving the cash; at the same time the OCC pays the short and “acquires” the short’s obligation to pay at cash-out time. OCC guarantees the short’s obligations to the long, to secure which it holds the short’s 150% margin. As the quarter progresses the short must pony up cash to cover dividend-equivalent obligations. When a long exercises the cash-out privilege, the OCC chooses a short at random to make the payment. Any link between the original buyer and seller of an IP thus does not extend beyond the formation of the instrument; after that instant, each person’s rights and obligations run to the OCC exclusively. This arrangement also permits either party to close its position by making an offsetting transaction. If the seller of an IP buys an identical contract in the market, the OCC cancels the two on its books.
The Philadelphia Stock Exchange asked the SEC in February 1988 for permission to trade IPs. The American Stock Exchange and the Chicago Board Options Exchange later filed proposals of their own. Each exchange’s IP differs slightly from the others. Philadelphia’s IP, called a “Cash Index Participation”, allows the long to exercise the cash-out privilege on any business day, at a discount of 0.5% from the value of the index. (The long may cash out on a quarterly date without penalty.) The AMEX’s IP, called the “Equity Index Participation”, permits the long to cash out quarterly for money or shares of stock in a ratio matching the index. Holders of 500 or more EIP trading units based on the S & P 500 index (each the equivalent of 100 multiples of that index) may exercise the right to receive securities, and they must pay a “delivery charge” to be established by the AMEX. Writers of EIPs may volunteer to deliver stock; if not enough do, a “physical delivery facilitator” at the AMEX will buy stock in the market, using money provided by the shorts whose positions have been liquidated. The CBOE’s product, the “Value of Index Participation”, has a semi-annual rather than quarterly cash-out date. CBOE’s wrinkle is that the short as well as the long may cash out, by tendering the value of the index on the cash-out date. If shorts seeking to close their positions exceed the number of longs who want cash, the OCC will choose additional long positions at random to pay off.
The three stock exchanges and the OCC asked the SEC to allow them to trade these varieties of IP. Each contended that the SEC has exclusive jurisdiction because IPs are securities and not futures contracts. The AMEX added that in its view an IP is an option on securities, activating the savings clause of § 2a(i). The Chicago Board of Trade and the Chicago Mercantile Exchange, supported by the CFTC, asked the SEC to deny the requests. Each futures market, and the CFTC, argued that IPs are futures and not securities, so that the CFTC’s jurisdiction is exclusive under 7 U.S.C. §§ 2 and 2a(ii). Complicating the picture, the Investment Company Institute argued that if IPs are securities and not futures, the OCC is an “investment compa[541]*541ny”, offering a product combining features of closed-end and open-end mutual funds, and must register under the Investment Company Act of 1940, 15 U.S.C. §§ 80a-l to 80a-64.
On April 11, 1989, the SEC granted the exchanges’ requests. Release No. 34-26709, 54 Fed.Reg. 15280 (1989). At the same time, its Division of Market Regulation, acting with delegated authority, allowed the OCC to change its rules so that it could issue, settle, and clear IPs. Release No. 34-26713, 54 Fed.Reg. 15575 (1989). The SEC concluded that IPs are “stock” within the meaning of § 3(a)(10) of the Securities Exchange Act of 1934, 15 U.S.C. § 78c(a)(10). IPs are negotiable, pay dividends, may appreciate in value, and may be hypothecated; the only attribute of stock missing from IPs is voting rights, which the SEC thought unimportant. 54 Fed. Reg. at 15285-86. If not stock, the SEC concluded, IPs are “certificates of interest or participation in” stock, another of the instruments defined as “securities” in § 3(a)(10). See 54 Fed.Reg. at 15286.1 Next the SEC found that IPs are not “futures”, id. at 15286-89, because they lack two features the SEC thought essential: “futurity” and “bilateral obligation”. “Futurity” means that value is set in the future, while as the SEC observed the buyer of an IP pays a price fixed at the time of sale; “bilateral obligation” means that the contract is executory on both sides until expiration or settlement, while the long on an IP performs at the time of purchase, leaving only the short with executory obligations. The SEC went on to say, id. at 15289-90, that the OCC need not register under the Investment Company Act because there is no “issuer” within the meaning of § 3(a)(1) of that statute, 15 U.S.C. § 80a-3(a)(l). Concluding that IPs may serve as substitutes for “program trading”, provide “an additional layer of liquidity to the market”, and afford “an alternative vehicle for retail customers to invest in ‘the market’ ”, 54 Fed.Reg. at 15290, the SEC allowed the exchanges to proceed with their plans. We denied the futures markets’ request for a stay but accelerated the hearing of the case on the merits. IPs have been trading on the three exchanges since May.
II
Three preliminary matters.
First, the futures markets are not plagued by regulation of their own activities. They complain, instead, that the SEC has passed up an opportunity to throttle three competitors. The futures markets may obtain relief only if there is a case or controversy within the meaning of Article III, and then only if they fall within the zone of interests protected by the Securities Exchange Act and the exclusivity clauses of the Commodity Exchange Act. The last time a dispute of this character was before us, we held that the futures markets have standing, GNMA Options, 677 F.2d at 1140-41 n. 4, but did not mention the “zone” question. Our opinion in Chicago Board of Trade v. SEC, 883 F.2d 525 (7th Cir.1989) (Delta Options), requires decision in the futures markets’ favor. Delta Options concludes that the futures markets may challenge the registration of a clearing agency that facilitates transactions for their competitors, even though the futures markets’ interests may be adverse to those of investors. Today’s case is easier, because the exclusivity clauses of the Commodity Exchange Act are at least arguably there for the benefit of the futures markets. Too, if the SEC is right that IPs are not futures, then the futures markets not only must suffer competition with their existing contracts but also would be forbidden to trade IPs themselves. The CME and the CBOT consequently are proper parties to invoke judicial review.
Second, the futures markets lodged with the court a lengthy report the CFTC submitted to the Senate Committee on Agriculture, Nutrition and Forestry explaining its position that it (and not the SEC) has jurisdiction of IPs. The stock exchanges ask us to strike this report because it is not part of the SEC’s administrative record. They also want us to strike from the appendix affidavits of Todd E. Petzel that were prepared after the SEC’s decision. We deny these motions. The CFTC’s report to the Senate is a public document, which the court would be free to [542]*542consult whether or not anyone had supplied it; lodging simply reduces the workload of the court’s librarian. The Petzel affidavits, by contrast, are case-specific documents that may be presented only if part of the administrative record. Wisconsin Electric Power Co. v. Costle, 715 F.2d 323, 326-27 (7th Cir.1983); cf. Edison Electric Institute v. OSHA, 849 F.2d 611, 623 n. 16 (D.C.Cir.1988). They are. Although prepared after the SEC’s decision, they were submitted to that agency in support of the futures markets’ request for a stay. The stock exchanges asked the SEC to strike them; it took no action on that request. So the affidavits entered the administrative record and may be included in the record here, for what they are worth. Edison Electric; see also Association of Pacific Fisheries v. EPA, 615 F.2d 794, 811-12 (9th Cir.1980) (Kennedy, J.).
Third, the SEC asks us to dismiss one of the petitions for review, No. 89-1538. The futures markets filed a petition the instant the SEC voted at a public meeting on March 14 to approve the trading of IPs. They filed additional petitions after the SEC issued its orders on April 11. Section 25(a)(1) of the ’34 Act, 15 U.S.C. § 78y(a)(l), which supplies this court’s jurisdiction, authorizes review of “a final order of the Commission entered pursuant to this chapter”. The petition shall be filed “within sixty days after the entry of the order”. The Commission’s Rules of Practice define “entry” as the date the order is adopted, reflected by its caption. Rule of Practice 22(k), 17 C.F.R. § 201.22(k). Until the Commission has “entered” an “order”, there is nothing to review. There is a big difference between a vote and an order. Only the order gives legal effect to the SEC’s vote.
The futures markets contend that the vote is reviewable because the stock exchanges could have started to trade IPs immediately thereafter, and they refer to ITT World Communications, Inc. v. FCC, 621 F.2d 1201 (2d Cir.1980); which accepted jurisdiction of a petition filed prior to the written order. The FCC’s decision in ITT went into force before entry of a written document; bound by the decision, ITT, as the party being regulated, needed to file a petition to obtain a stay. The SEC’s vote did not produce an immediately effective order. It did not require the futures markets to take any action. The vote (without the order) had no effect on the futures markets different in kind from a refusal (for the time being) to take action against their competitors. To see this, suppose that the stock exchanges had started trading IPs forthwith, and the SEC had done nothing. Inaction would be neither final nor reviewable, for the reasons given in Delta Options. The vote of March 14 was at most a clue that if the stock exchanges were to move aggressively, the SEC would not act. Petition No. 89-1538 is dismissed for want of a reviewable order.
Ill
A
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. When the contract involves financial instruments, though, the price is fixed by mechanical computation from the instruments on which the contracts are based. See Leist v. [543]*543Simplot, 638 F.2d 283, 286-87 (2d Cir.1980) (Friendly, J.), affirmed on other grounds under the name Merrill Lynch, Pierce, Fenner & Smith, Inc. v. Curran, 456 U.S. 353, 102 S.Ct. 1825, 72 L.Ed.2d 182 (1982); Philip McBride Johnson & Thomas Lee Hazen, 1 Commodities Regulation §§ 1.03-04, 1.10 (2d ed. 1989); Bryan Byrne, Jr., The Stock Index Futures Market (1987).
A security, roughly speaking, is an undivided interest in a common venture the value of which is subject to uncertainty. Usually this means a claim to the assets and profits of an “issuer”. Shares of stock entitle their holders to receive dividends and payments on liquidation (or a change in corporate form), see Landreth Timber Co. v. Landreth, 471 U.S. 681, 105 S.Ct. 2297, 85 L.Ed.2d 692 (1985); bonds and other debt instruments promise interest plus a balloon payment of principal at the end. Unusual interests such as rights in orange groves still may be “securities” if they represent a pro rata share of a variable pool of earnings. SEC v. W.J. Howey Co., 328 U.S. 293, 66 S.Ct. 1100, 90 L.Ed. 1244 (1946). See generally Louis Loss & Joel Seligman, 2 Securities Regulation 926-89 (3d ed. 1988).
Securities usually arise out of capital formation and aggregation (entrusting funds to an entrepreneur), while futures are means of hedging, speculation, and price revelation without transfer of capital. So one could think of the distinction between the jurisdiction of the SEC and that of the CFTC as the difference between regulating capital formation and regulating hedging. Congress conceived the role of the CFTC in that way when it created the agency in 1974 to assume functions that had been performed by the Department of Agriculture but which were no longer thought appropriate for that Department as futures markets expanded beyond commodities into financial instruments. See GNMA Options for a recap of the history. Unfortunately, the distinction between capital formation and hedging falls apart when it comes time to allocate the regulation of options.
A call option is a promise by the writer to deliver the underlying instrument at a price fixed in advance (the “strike price”) if the option is exercised within a set time. The buyer pays a price (the “premium”) in advance for the opportunity; the writer may or may not own the instrument he promises to deliver. Call options are written “out of the money” — that is, the exercise price exceeds the market price at the outset. The writer will make money if by the time the option expires the market price is less than the strike price plus the premium (plus the interest earned on the premium in the interim); the buyer of the option hopes that the market price will rise above the strike price by enough to cover the premium, the time value of money, and the transactions costs of executing the option. Options play valuable roles in price-discovery, and they also allow the parties to adjust the net riskiness of their portfolios. Writers of call options reduce the risk they bear if the market falls while limiting gains if the market rises; buyers hope for large proportional gains if the market rises while accepting the likelihood that the options will turn out to be worthless. Options are side deals among investors, which do not augment an entrepreneur’s coffers (except to the extent greater liquidity and opportunities to adjust risk increase social marginal propensity to invest). Dwight M. Jaffee, The Impact of Financial Futures and Options on Capital Formation, 4 J. Futures Markets 417 (1984). Unlike financial and index futures, options call for delivery of the underlying instrument — be it a share of stock or a futures contract.
The SEC consistently has taken the position that options on securities should be regulated as securities. For some years the CFTC maintained that options on securities should be regulated as futures because options are extrinsic to capital formation and because it is almost always possible to devise an option with the same economic attributes as a futures contract (and the reverse). Matters came to a head in 1980, when both agencies asserted jurisdiction over options on securities based on pools of notes. The Government National Mortgage Association (GNMA) sold pass-through certificates representing proceeds of mortgage notes, and persons started writing options on them to allow hedging against movements in interest rates. The SEC observed that options written on securities are securities under § 3(a)(10) of the ’34 Act; indeed the SEC contended that [544]*544because options are securities it should regulate all options. The CFTC countered that options on financial instruments are futures under § 4c(b) of the CEA, 7 U.S.C. § 6c(b), and added that because its jurisdiction is exclusive, it is the sole lawful regulator. When the SEC allowed stock exchanges to start trading GNMA options, the futures markets sought review in this court and the CFTC howled bloody murder.
While the case was pending, the agencies reached a pact, which the SEC calls the Shad-Johnson Agreement and the CFTC calls the Johnson-Shad Agreement. (John Shad was the SEC’s Chairman at the time, and Phillip Johnson the CFTC’s.) This Accord (as we shall call it to avoid offending either agency) provided that jurisdiction over options follows jurisdiction over the things on which the options are written. So the SEC received jurisdiction of options on securities, while the CFTC got jurisdiction of options on futures contracts. Things were not quite done, though, because we held in GNMA Options that the agencies could not alter their jurisdiction by mutual agreement. 677 F.2d at 1142 n. 8. Starting from the proposition that options on GNMAs are both securities and futures, we held that the CFTC’s jurisdiction is exclusive in light of 7 U.S.C. §§ 2 and 2a.
Congress then enacted the Accord almost verbatim, producing the explicit reference to options in § 3(a)(10) of the ’34 Act, the SEC savings clause in § 2a(i) of the CEA, and a small change in 7 U.S.C. § 6n to implement an understanding about pools. The legislature thought that this Accord would resolve things and restore a regime in which the SEC supervises capital formation and the CFTC hedging. See S.Rep. No. 97-384, 97th Cong., 2d Sess. 21-24 (1982); H.R.Rep. No. 97-565, 97th Cong., 2d Sess., Part I at 38-40 (1982), U.S.Code Cong. & Admin.News 1982, p. 3871; H.R. Rep. No. 97-626, 97th Cong., 2d Sess., Part II at 3 (1982), U.S.Code Cong. & Admin. News 1982, p. 2780; 128 Cong.Rec. 24910 (1982) (Rep. De La Garza); Loss & Selig-man, 2 Securities Regulation at 1064-80; Jerry W. Markham & David J. Gilberg, Stock and Commodity Options — Two Regulatory Approaches and Their Conflicts, 47 Albany L.Rev. 741 (1983).
The legislation implementing the Accord left in place the premise on which GNMA Options was founded: if an instrument is both a security and a futures contract, then the CFTC’s jurisdiction is exclusive. Section 2a(ii) has no other possible meaning. Like many an agreement resolving a spat, the Accord addressed a symptom rather than the problem. Options are only one among many instruments that can have attributes of futures contracts as well as securities. Financial markets work best when they offer every possible combination of risk and return — a condition financial economists call “spanning” — so that investors can construct a portfolio to each need and taste. Exchanges and professional investors therefore continually devise financial products to fill unoccupied niches. See Dennis W. Carlton, Futures Markets: Their Purpose, Their History, Their Growth, Their Successes and Failures, 4 J. Futures Markets 237 (1984); William L. Silber, Innovation, Competition and New Contract Design in Futures Markets, 1 J. Futures Markets 123 (1981). These products are valuable to the extent that they do not match the attributes of instruments already available. New products, offering a new risk-return mixture, are designed to depart from today’s models.
Which means that the dispute of 1980-82 about options will be played out — is being played out — about each new instrument. Today’s case repeats the conflict. Other novel instruments are being handled by regulation. For example, on July 17, 1989, the CFTC adopted rules exempting from its regulation certain hybrid instruments combining equity or debt with payments based on the price of commodities. 17 C.F.R. Part 34, 54 Fed.Reg. 30684 (1989). Only merger of the agencies or functional separation in the statute can avoid continual conflict. Functional separation is hard to achieve (new instruments will appear at any border). The SEC favors merger; it has asked Congress repeatedly for jurisdiction over all products (including stock-index and financial futures) based on securities, which would relegate the CFTC to its original role as superintendent of commodities futures. The CFTC has so far defended its position, in part with the argument that multiple regulatory bodies allow greater competition and experimentation — a new product can reach market if either agency approves the variant within its domain. See Daniel R. Fisehel, Regulatory Conflict and Entry Regulation of New Futures Contracts, 59 J. Business S85 (1986); Ron-[545]*545aid W. Anderson, The Regulation of Futures Contracts Innovations in the United States, 4 J. Futures Markets 297 (1984).
Unless Congress changes the allocation of jurisdiction between the agencies, the question a court must resolve is the same as in GNMA Options: is the instrument a futures contract? If yes, then the CFTC’s jurisdiction is exclusive, unless it is also an option on a security, in which case the SEC’s jurisdiction is exclusive. So long as an instrument is a futures contract (and not an option), whether it is also a “security” is neither here nor there. Still, if IPs really are “stock” they almost certainly are not “futures contracts”, so the inquiries aren’t so distinct as the statutes imply.
B
From the perspective of the long, IPs look like an interest in a portfolio of stock. IPs last indefinitely (except for the chance that a long may be cashed out involuntarily on the CBOE), may be sold like stock or used to secure margin and other loans, change in value with the market, and pay dividends. IPs lack other common attributes of stock: they do not confer voting rights and are not “certificated”; owners of IPs receive dividend-equivalent payments quarterly, not when the firms pay dividends. We need not debate whether these differences come to anything, for they pale beside the larger difficulties in calling IPs “stock”. The greatest is that IPs are not stock in anything. There isn't an issuer — which the SEC emphasized when concluding that the Investment Company Act is inapplicable, 54 Fed.Reg. at 15289-90. Stock is an equity interest in an issuer, the residual claim to the profits of a venture. United Housing Foundation, Inc. v. Forman, 421 U.S. 837, 95 S.Ct. 2051, 44 L.Ed.2d 621 (1975). Landreth rejected the “sale-of-business doctrine” because the owner of 100% of the equity interest in a firm still owns “stock".2 Purchasers of IPs don’t own equity, directly or indirectly; they don’t have a claim to the proceeds and liquidating distribution of a business; there isn’t an underlying pool of assets; there is only a “short” on the other side. The absence of an issuer — IPs don’t carry votes because they don’t have anything to do with equity — tells all. There is no common venture, not even the commonality represented by a mutual fund (which reinvests in real stock and creates the risk that the stakeholder will join Robert Yesco with the kitty).
IPs do not fit comfortably into any of the other pigeonholes of § 3(a)(10). A “certificate of interest or participation in ... any of the foregoing” securities is a security too, but IPs do not represent an “interest or participation” in the stocks in the index; they are based on the value of stock without creating a legal interest in stock. Perhaps the closest match is the language, part of the Accord in 1982, covering a “privilege on any security ... or group or index of securities (including any interest therein or based on the value thereof)”. Then there’s the catch-all: “in general, any instrument commonly known as a ‘security’ ”. IPs convey privileges based on the value of an index, and what is “commonly known as a security” changes as new instruments come into use. So there is a basis for drawing IPs within § 3(a)(10), even though they do not duplicate a recognized category. See also, e.g., SEC v. United Benefit Life Insurance Co., 387 U.S. 202, 87 S.Ct. 1557, 18 L.Ed.2d 673 (1967).
Although the SEC found IPs to be securities by looking at the promises made to the longs, the CFTC found them to be futures by virtue of the promises made by the shorts, a perspective implied by the CEA’s references to “contracts ... for future delivery” — emphasizing the shorts’ obligation. Shorts on IPs make the same pledge as shorts on stock-index futures contracts: to pay the value of an index on a prescribed day (the expiration date for the futures contract, the cash-out date for the [546]*546IP). The short owes this obligation to the clearing house rather than to the long. IPs may be settled by buying or selling an offsetting obligation, after which the clearing house cancels the two on the books, just as with futures contracts. Shorts on IPs must put up more margin than shorts on futures contracts and must make dividend-equivalent payments, but the CFTC did not find these differences any more dispositive than the SEC found the IPs’ lack of voting rights. Shorts also face an obligation of indefinite duration on the Philadelphia and AMEX IPs, but the CFTC and the futures markets treat this as no more than a prepaid rollover privilege.
Despite the congruence of futures and IPs on the short side, the SEC and the stock exchanges say that both “futurity” and “bilateralism” are missing. According to the SEC, IPs lack “futurity” because an IP is the “present obligation to pay current value”. And IPs are not bilateral because the long performs in full by paying up front, although in a futures contract both sides must perform on settlement or expiration.
With respect to bilateralism, the SEC’s point is inescapable. With respect to futurity, the SEC is wrong. IPs are no more a “present obligation to pay current value” than are futures contracts. The holder of either an IP or a stock-index futures contract may go to market and trade it; the price necessarily tracks current value. Neither the long on an IP nor the long on a futures contract can compel the short to pay current value, however.3 Both the futures contract and the IP are settled quarterly (the same dates for both kinds of instrument, except for the CBOE’s omission of two of the four dates). The short’s obligation is to pay the value of the index on that date — which lies in the future to the same extent as the settlement date of any futures contract. Even from the long’s point of view, IP and futures contract ultimately look the same. The long pays up front for the IP, but the long on a futures contract promises up front to make a defined payment on the settlement date; the difference in the timing of the payment does not affect the fact that valuation comes at the defined future date.
So the IP has futurity but not bilateral-ism. It looks like a futures contract to the short — except that it is of indefinite duration, carries a dividend-equivalent obligation, and requires higher margin. It looks like a mutual fund to the long — except that it has no voting rights, does not represent any interest in an underlying pool of stock, and may be settled by executing an offsetting transaction. Fact is, it is no less a future than it is a security, and no more. It just doesn’t fit. Which is the whole point. It isn’t supposed to be just like something else; the IP was designed as a novel instrument so that it could offer attributes previously missing in the market.
The only thing of which we are sure is that an IP is not an option on a security. The AMEX contends that it is a prepaid option, with a premium equal to the full value and an exercise price of zero. The SEC did not accept this contention, writing:
[WJhile IPs contain some characteristics of stock index options (e.g., the issuance and clearance and settlement features of IPs are analogous to those of stock index options), the Commission believes that IPs predominantly have the attributes of a portfolio of common stock.
54 Fed.Reg. at 15286 n. 57. The only “characteristics of stock index options” that either the AMEX or the SEC identified are those introduced by the presence of a clearing house — characteristics that the IP shares with stock-index futures to the last detail. Unless we were to say that all futures are also options (they aren’t), these features do not make IPs options. The very features that the SEC emphasizes to show that IPs are securities — indefinite duration, payment up front in cash, dividend equivalency, and so on — show that IPs cannot be options. Options are written out of the money, limited in time, and establish a careful balance among premium, strike price, and duration; the writer retains dividends. IPs possess none of these distin[547]*547guishing features. As the AMEX defines an “option”, someone who buys an automobile for cash and drives it away really has obtained an option with a high premium, zero strike price, perpetual duration, and 100% probability of exercise. Words are useful only to the extent they distinguish some things from others; symbols that comprise everything mean nothing. IPs are not options.
C
Having concluded that neither the ’34 Act nor the CEA addresses the status of IPs in a straightforward way, the logical thing to do is to defer to the agency’s resolution of the problem. 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). But which agency? Each claims to be exercising its discretion; each claims entitlement to deference on a subject within its domain.
Although cases frequently say that courts should defer to the judgment of the responsible agency whenever the statute is : ambiguous as applied to a subject, e.g., NLRB v. Hearst Publications, Inc., 322 U.S. 111, 64 S.Ct. 851, 88 L.Ed. 1170 (1944), this is something of an oversimplification. Ambiguity does not necessarily dictate whether the court or the agency has the dispositive word. Many’s the statute that drops a half-resolved dispute in the lap of the courts even though one or more agencies exercise jurisdiction. Think of the anti-trust laws, which courts freely construe even though both the Antitrust Division of the Department of Justice and the Federal Trade Commission may lay claim to greater expertise. When a statute has gaps and uncertainties — the status of all rules — the anterior question is: who is charged with filling the gaps? Often statutes delegate comprehensive powers to agencies, and the meaning of the law is that agencies shall solve novel problems as they arise. Solutions may involve complex and unanticipated adjustments. Courts can be more confident that power has been delegated than that any particular exercise is “right”. Deference to the agency’s conclusion follows naturally from such a determination, for what Congress wanted to obtain is the judgment of the agency — Congress delegates precisely because it cannot foresee and resolve all problems. See Chevron, 467 U.S. at 843-45, 104 S.Ct. at 2781-83; Young v. Community Nutrition Institute, 476 U.S. 974, 981-84, 106 S.Ct. 2360, 2364-66, 90 L.Ed.2d 959 (1986); Homemakers North Shore, Inc. v. Bowen, 832 F.2d 408, 411-12 (7th Cir.1987); Henry P. Monaghan, Marbury and the Administrative State, 83 Colum.L.Rev. 1 (1983). When the agency is the addressee of the statutory command, it takes the leading part in giving structure to the statute; when the court is the addressee, it has the principal role.
When two agencies claim to be the addressees, though, this allocation breaks down. Perhaps a court could say that because the agencies disagree, neither is entitled to deference. Yet disagreement doesn’t make the court the recipient of interpretive powers. One or the other agency is still in charge. Courts readily could accept both the SEC’s application of the ’34 Act to IPs and the CFTC’s application of the CEA. Our difficulty is not any logical conundrum in deferring to both agencies when they disagree. It is instead that a dispute about the agencies’ jurisdiction is a zero-sum game because of the exclusivity clauses in the CEA.
Any distinction between action under delegated powers and fixing the scope of delegation will break down at the edges, and some recent cases suggest that a court should not try to draw such a line in the first place — one of them concerning the scope of the CFTC’s powers. CFTC v. Schor, 478 U.S. 833, 844-46, 106 S.Ct. 3245, 3253-54, 92 L.Ed.2d 675 (1986); see also Mississippi Power & Light Co. v. Mississippi ex rel. Moore, — U.S. —, 108 S.Ct. 2428, 2444, 101 L.Ed.2d 322 (1988) (Scalia, J., concurring); CSX Transportation v. United States, 867 F.2d 1439, 1445 (D.C.Cir.1989) (Edwards, J., dissenting) (suggesting that Chevron disallows inquiry into the extent of delegation, although that should be the right question). But even if unambiguous delegation is not a necessary condition of deference, it is an important ingredient in the formula, else it becomes impossible to distinguish statutes such as the Sherman Act from those such as the Clean Air Act, or to conceive where the boundary between court and commission [548]*548falls. Delegation to agencies is not without its costs to the separation of powers; holding agencies within their delegated scope is an important task in maintaining constitutional structure.4
Difficulties in establishing the competence of the agencies and the judicial branch do not influence the outcome of this case, however. We may assume without deciding that even in this jurisdictional dispute, each agency is entitled to leeway in applying its own statute to IPs.
D
If each agency’s interpretation of its own statute is entitled to some deference, then the IP is both a security and a futures contract. It has some attributes of both, and all attributes of neither, as we have laid out in excessive detail. Neither characterization can be called wrong.
The only element of financial futures contracts that is missing is “bilateralism”. Yet bilateralism is not essential to a futures contract. CFTC v. Co Petro Marketing Group, Inc., 680 F.2d 573 (9th Cir. 1982), held that a contract that imposes performance obligations only on the short may be a futures contract. Co Petro sold interests in gasoline that were designed to look like forward contracts, which under the CEA are not futures contracts. Buyers put down deposits to obtain Co Petro’s promise to deliver gasoline on future dates. These contracts could not be traded on any market, but Co Petro promised to pay the investor in cash if the market price should rise (that is, the investor could sell the contract back to Co Petro). The investor risked no more than 95% of his deposit; if the price of gasoline fell, the investor’s position would be closed. Thus buyers of Co Petro’s contracts performed fully on the date they posted the deposit; thereafter only Co Petro had obligations. Despite that, and despite the fact that the contracts were illiquid, the court of appeals concluded that they were futures contracts because their value depended entirely on the price of the commodity at their expiration date, and they were not formed in contemplation of physical delivery.
The SEC brushes off Co Petro and similar cases in district courts as based on the principle that once it smells sulfur (Co Pe-tro may have been a bucket shop), either agency may protect the investor. No such principle may be found in the ’34 Act or the CEA, however. An instrument either is or is not a futures contract. If it is, the CFTC has jurisdiction; if it is not, the CFTC lacks jurisdiction; if the CFTC has jurisdiction, its power is exclusive. The SEC’s position entails the proposition that if Co Petro Marketing Group, Inc., had approached the CFTC after losing in the Ninth Circuit and applied for permission to trade its gasoline contracts as futures, the CFTC would have had to say no, on the ground that the contracts are not “contracts ... for future delivery ” under the CEA. That can’t be right.
Perhaps this point will be clearer if we ask what would have happened if the CBOT and CME had approached the CFTC in 1987 (before the Philadelphia Stock Exchange filed its proposal with the SEC) seeking permission to trade IPs. When we asked the SEC’s Solicitor during oral argument whether the CFTC could have granted such an application, he said yes — largely on the ground that granting the application would have introduced a new product with benefits for investors. When we persisted with the question whether the CFTC could grant the identical application, filed by the CBOT and CME in 1989 (after the SEC’s decision), the Solicitor said no, on the ground that by 1989 the SEC had asserted jurisdiction. Yet this principle of first-come-first-served finds no support in the ’34 Act or the CEA. Either IPs are futures contracts or they aren’t. If they are futures contracts, then the CFTC could have approved their trading in 1987 (as the Solicitor agreed); if IPs were futures in 1987, they are futures today, and the CFTC still may approve their trading. But if the CFTC may approve their trading because they are futures contracts, then the CFTC’s jurisdiction is exclusive.
Doubtless such a decision gives the futures markets the opportunity to block [549]*549competition from an innovative financial product. The SEC’s order, and its brief in this court, casts much of the argument in the form: “The IP is a desirable product; the futures markets have not proposed to trade IPs; if IPs are futures then the CFTC’s jurisdiction is exclusive and IPs won’t exist, which would be regrettable; therefore IPs are not futures.” Everything works until the “therefore”. Whether IPs are futures can’t depend on who first proposed to trade them, or on whether anyone does. We doubt the premise of the SEC's argument as well as its conclusion, for if IPs are useful to investors, then someone will offer them — if not the CBOT and CME, then a market in New York, or Kansas City, or Tokyo. The CBOT and CME will be compelled to follow or they will lose business. There are too many futures exchanges to suppose that a conspiracy could suppress a beneficial financial instrument, cf. Matsushita Electric Industrial Co. v. Zenith Radio Corp., 475 U.S. 574, 590, 106 S.Ct. 1348, 1357, 89 L.Ed.2d 538 (1986). But whether or not a futures market will seek to trade a financial product does not change the nature of that product, and both the ’34 Act and the CEA define coverage by the attributes of the instrument rather than by the identity of those who own or trade it. This is the central message of Landreth, which at the SEC’s urging rejected the “sale of business doctrine” precisely because that doctrine disregarded the characteristics of the financial instrument in order to go straight to the question whether certain persons needed the protection of the law — that is to say, whether coverage was a good idea.
From time to time, the Supreme Court has looked to the purposes of the ’34 Act to define “securities”, usually with a view to enlarging the definition. (The exception is Marine Bank v. Weaver, 455 U.S. 551, 102 S.Ct. 1220, 71 L.Ed.2d 409 (1982).) With the SEC urging it on, the Court has drawn in orange groves covered by joint harvesting contracts, W.J. Howey, leaseholds in land near oil wells, SEC v. C.M. Joiner Leasing Corp., 320 U.S. 344, 64 S.Ct. 120, 88 L.Ed. 88 (1943), and other unusual instruments that have some (but far from all) attributes of conventional securities. Obviously the SEC does not ask us to abandon this approach — for itself. It demands, however, that we apply strictissimi juris to the CEA, to hold that only an instrument with every attribute of a conventional futures contract may be one. Why? If the interpretive approach is proper for the securities acts, it is no less proper for the futures acts. It has been employed under both statutes — not only in Co Petro but also in redefining futures contracts to omit the delivery obligation. Recall the statutory scope of the CEA: contracts “for future delivery”. Commodity futures contracts may be settled by delivery; financial futures contracts are settled exclusively in cash. One might have thought the prospect of “future delivery” the sine qua non of a “futures” contract. Yet no one, not even the SEC, doubts that a contract may be a futures contract even though it provides for cash settlement. If delivery is not essential, then the “traditional” elements of futures contracts are not invariable ingredients of the CFTC’s jurisdiction.
Perhaps the SEC wants us to put a thumb on the scales, enlarging the category “securities” while shrinking the category “futures” because of the exclusivity clauses in the CEA: if both categories expand, then the SEC’s jurisdiction shrinks. We do not conceive it our function, however, to invent counterweights to statutes; judges should be interpreters rather than sappers and miners. As we said in GNMA Options, 677 F.2d at 1161, “[o]ur task should not reflect a value judgment as to which of the competing agencies is best equipped to regulate these [products].”
To the extent instrumental arguments influence the coverage of the laws, they do not necessarily cut for the SEC. The futures markets’ reply brief invites us to imagine a “Wheat Index Participation” (WIP) having the same characteristics as the IP except that it is based on an index of wheat prices rather than of stock prices. The buyer would pay cash for the WIP and be able to trade it freely; on a date identical to the expiry of the wheat futures contracts, the writer could be required to pay cash measured by the value of the wheat index. According to the SEC, such an instrument would not be a futures contract because it would lack both futurity and bilateralism (and we would agree on the latter point). So the CFTC could not allow it to be traded, no matter how valuable participants in the market might find it. On the other hand, the WIP certainly would not be “stock” and probably would not meet the criteria for being a “security” of any kind. So the SEC could not allow it to be traded on stock exchanges (anyway, the WIP would be a duck out of water on the AMEX!). We could escape from such [550]*550silliness by reaching the logical conclusion that a WIP would be a futures contract. Yet if the WIP is a futures contract, it is hard to avoid the conclusion that the IP is one, too.
The petition for review in No. 89-1538 is dismissed for want of jurisdiction. The Investment Company Institute’s petition, No. 89-2012, presents questions that we need not reach in light of our disposition of the futures markets’ claims. On petition Nos. 89-1763 and 89-1786, the SEC’s orders approving the applications of the stock exchanges and the OCC are set aside.