EASTERBROOK, Circuit Judge.
Futures and options are competing financial instruments often used for hedging. The Chicago Board of Trade and the Chicago Mercantile Exchange trade futures on United States Treasury securities and options on futures on these securities. RMJ Options Trading Corporation, Delta Government Options Corporation, and Security Pacific National Trust Company (SPNTCO) trade options on Treasury bills, bonds, and notes. The futures markets contend that the RMJ-Delta-SPNTCO “System” is really an “exchange” that must register under § 6 of the Securities Exchange Act of 1934, 15 U.S.C. § 78f. The Securities and Exchange Commission is allowing the System to operate without registration under § 6: RMJ is a registered broker, Delta is a registered clearing agency, and the System secured a no-action letter from the Commission’s staff. The futures markets ask us to set aside the no-action letter and the order registering Delta as a clearing agency.
I
Section 3(a)(1) of the ’34 Act, 15 U.S.C. § 78c(a)(l), defines an “exchange” as
[527]*527any organization, association, or group of persons ... which constitutes, maintains, or provides a market place or facilities for bringing together purchasers and sellers of securities or for otherwise performing with respect to securities the functions commonly performed by a stock exchange as that term is generally understood....
Section 5 of the Act, 15 U.S.C. § 78e, makes it unlawful for a broker or dealer to transact business on an “exchange” unless it has been registered under § 6 or is “exempted from such registration ... because, in the opinion of the Commission, by reason of the limited volume of transactions effected on such exchange, it is not practicable and not necessary or appropriate in the public interest or for the protection of investors to require such registration.” Section 5(2), 15 U.S.C. § 78e(2). The futures markets maintain that the System is an “exchange” that must register under § 6 because it “provides a market place” for securities and carries out “the functions commonly performed by a stock exchange”. RMJ and its fellow participants deny this characterization, so we start by describing the System and the instruments traded there.
A call option is a promise to sell a security in the future at a price fixed today. The seller agrees to deliver the security for a set price (the “strike price”) during a limited time. The buyer pays a sum (the “premium”) for the privilege. The strike price exceeds the current market price of the security. Sellers are betting that the price will not exceed the strike price during the duration of the option; buyers are betting that it will. A put option works the other way 'round. The seller of the put option promises to pay a fixed price if the security is tendered during the life of the option. Both put and call options allocate the risk of price movements and allow each side of the transaction to hedge its position in the underlying security. Although the mechanics of futures transactions differ from those of options transactions, futures may serve the same economic functions, and both have the important byproduct of improving the liquidity of markets.
Sellers of call options and buyers of put options are called “shorts” because they do not (necessarily) own the securities — although they have to do so if the other side exercises the call option, or may choose to do so if they want to exercise a put option. The party on the other side is called the “long” because it will own the security if it exercises the call or must honor the put. Longs on call options, and shorts on put options, face a risk in addition to the prospect that the price will not change enough to make exercise worthwhile: if the price does move, and the option is exercised, the other party may not deliver (for a call option) or pay (for a put option). Clearing houses stand between the parties, guaranteeing obligations so that each party shucks the risk of the other’s non-performance. Once parties agree on the terms of an option, the clearing house “issues” the option and acquires the delivery or payment obligation on the other side. Each party deals exclusively with the clearing house, which matches transactions and requires margin and guarantees to minimize its own risk.
Clearing houses at least potentially offer another benefit: anonymity. Neither the buyer nor the seller need know who is on the other side of the transaction. Parties may seek anonymity because their identities (coupled with information about the size of their positions) may enable others to infer information that they want to keep confidential. “Blind brokers” specialize in matching buyers and sellers of securities without identifying them to each other. Options, which are executory contracts on one side, cannot be sold in anonymity without a clearing house. The long on a call option needs assurance that the short will deliver, the short on a put option assurance that the long will pay; an anonymous promisor is not a satisfactory trading partner. Clearing houses can be “in the know” while the traders are in the dark, producing anonymous trades with assured performance.
Securities issued by the United States government (“government securities”) are traded over the counter only, by broker-[528]*528dealers acting as principals. Many brokers specialize in these instruments, and some are “blind brokers”. Until RMJ, Delta, and SPNTCO created their System, however, there was little trading in options on government securities, and anonymity was impossible. “The System” itself, as the parties describe it, is a combination of computer hardware and software that matches offers and keeps track of the obligations of the longs and shorts until the options expire or the positions are closed by offsetting transactions. Subscribers can consult terminals linked to the System’s computer to discover bid and asked prices on options. They may call RMJ Options, the System’s sole broker, to make bids (which RMJ will post on the System) or accept offers they have seen posted. RMJ enters the information into the System. Delta, the clearing house, reads the information from the System and formally “issues” the option to the buyer while “purchasing” an offsetting option from the seller. Such a trade is thus anonymous and fully guaranteed, while Delta’s books are in balance at all times. Trades need not be anonymous, however. Any participant in the System may trade directly with another and notify RMJ, which enters the transaction into the System. Delta then acts as administrator and guarantor.
Delta protects its own position against participants’ defaults by ensuring that each subscriber is creditworthy. No one may join the System without satisfying Delta’s standards. Delta also sets margin requirements, as well as trading and position limits, to protect itself against the risk that participants will hold unbalanced portfolios taxing their resources (increasing the likelihood of default). SPNTCO acts as clearing bank for the System, and Delta does not issue or purchase options until SPNTCO has “accepted” the trade, verifying that the long and short have complied with Delta’s rules on margin and position limits. SPNTCO holds the customers’ (and Delta’s) money and securities, paying traders and delivering securities according to the terms of the options.
After rearranging the structure of the System to satisfy federal banking regulators, RMJ, Delta, and SPNTCO needed the SEC’s approval. Firms may register as brokers without much ado. The registration of a “clearing agency” under § 17A(b) of the Act, 15 U.S.C. § 78q-l(b), is more complicated. Those who seek to act as clearing agencies must satisfy nine statutory criteria, many of them calling for an exercise of discretion by the SEC. And the System faced opposition from the futures markets, which did not savor competition and sought to persuade the Commission that the System must register as an exchange. Registration as an exchange would be fatal to the System’s current organization. The System is proprietary: it is organized for profit, with RMJ Options as its sole broker; only persons admitted to the System may trade. Section 6 of the ’34 Act treats “exchanges” as organizations created and run by broker-members. Section 6(b)(3) provides that the rules of the exchange must “assure a fair representation of its members in the selection of its directors ... and provide that one or more directors shall be representative of issuers and investors and not associated with a member of the exchange”, and § 6(b)(5) says that the exchange’s rules can’t tolerate “unfair discrimination between ... brokers”. There is more, but from the System’s perspective this is enough. Section 6 requires registered “exchanges” to conform to the model of the New York Stock Exchange or the Chicago Mercantile Exchange, and the System’s creators wanted to strike off on a new path. So while Delta sought registration as a clearing agency under § 17A, RMJ asked the Commission to say that the System did not need to register under § 6.
The SEC obliged on both fronts. Richard G. Ketchum, Director of the Commission’s Division of Market Regulation, wrote a letter explaining that the Division would not ask the Commission to take action against the System under § 6, at least not so long as the System operated in line with undertakings counsel had made. Although the letter did not offer a definition of “exchange”, Ketchum singled out three attrib[529]*529utes for particular attention (footnote omitted):
Specifically, the staff may determine that the requirements of exchange registration should be imposed upon the System should it find that, at some point in the future:
(1) there develops a continuous, two-sided market in options that: (a) expire at the same monthly intervals; (b) have identical premiums or strike prices; or (c) are written on underlying securities having the same maturity date or yield;
(2) the operator of the System, or any entity operating on its behalf, elects to standardize terms, such as strike price and expiration month, that are currently subject to negotiation between participants; or
(3) System participants assume control of, or begin to regulate their own participation in, the System.
At a public meeting on January 5, 1989, the Commission voted not to object to the Division’s sending this letter. At the same meeting the Commission also granted Delta a three-year registration as a clearing agency under § 17A, see Release No. 34-26450, 54 Fed.Reg. 2010 (1989), and decided to issue for comment a draft rule that would govern the operation of proprietary trading systems. The proposed rule appeared in April, Release No. 34-26708, 54 Fed.Reg. 15429 (1989), and the comment period closed on June 19. We declined to stay the SEC’s order registering Delta as a clearing agency, and the System has been in operation since March.
II
Director Ketchum’s letter to counsel for RMJ states that the Division of Market Regulation will not recommend that the Commission take action, provided that the System continues to meet certain conditions. The futures markets petition for review of this letter. One would think that if the System is an unregistered exchange, the right thing to do is sue RMJ, Delta, and SPNTCO for violations of the statute. Instead of suing their rivals, the futures markets sued the SEC. At least three hurdles stand between the futures markets and our review of the merits.
First, the futures markets want the Commission to prosecute a stranger. This presents substantial questions of standing under Article III of the Constitution. Linda R.S. v. Richard D., 410 U.S. 614, 93 S.Ct. 1146, 35 L.Ed.2d 536 (1973); Leeke v. Timmerman, 454 U.S. 83, 102 S.Ct. 69, 70 L.Ed.2d 65 (1981); Allen v. Wright, 468 U.S. 737, 104 S.Ct. 3315, 82 L.Ed.2d 556 (1984). Linda R.S. held that a wife lacked standing to obtain an order directing the prosecution of her husband for failure to pay child support, because such a prosecution would not necessarily induce payment. To this “redressability” concern Allen added that orders directing agencies to prosecute violations of the law may interfere with the Constitution’s allocation of functions among branches of the government, 468 U.S. at 750-52, 759-61, 104 S.Ct. at 3324-25, 3328-29. See also Antonin Scalia, The Doctrine of Standing as an Essential Element of the Separation of Powers, 17 Suffolk U.L.Rev. 881 (1983). We suppose that the SEC’s prosecution of the System would be no less effective than the futures markets’ civil suit in closing down a scofflaw (though one might have said the same about the relation between public and private litigation in Linda R.S.), so that the injury is “redressable”; concerns for the constitutional allocation of powers still loom.
Second, the futures markets seek review of a decision that by its terms is tentative. The Division won’t recommend prosecution just now, and just so long as conditions are satisfied. Director Ketchum could change his mind tomorrow, or the Commissioners might elect to proceed no matter what the Director recommends. The SEC has not, in other words, issued a “final” decision concerning the status of the System. Under the Administrative Procedure Act, finality is a precondition to judicial review. 5 U.S.C. § 704. See FTC v. Standard Oil Co., 449 U.S. 232, 101 S.Ct. 488, 66 L.Ed.2d 416 (1980). Section 25(a)(1) of the ’34 Act, 15 U.S.C. § 78y(a)(1), the source of our jurisdiction to review the Commission’s orders, authorizes only “[a] [530]*530person aggrieved by a final order of the Commission” to obtain judicial review. “Finality” is a practical concept, and if the SEC had put the futures markets in jeopardy should they neglect to dance to its tune, even a decision of the staff might be “final” action. Compare Ciba-Geigy Corp. v. EPA, 801 F.2d 430 (D.C.Cir.1986), with Biotics Research Corp. v. Heckler, 710 F.2d 1375 (9th Cir.1983). See Gardner v. Toilet Goods Ass’n, 387 U.S. 167, 87 S.Ct. 1526, 18 L.Ed.2d 704 (1967). Nothing of the sort has occurred, however. The Division’s missive is a no-action letter. The futures markets are not under the gun; the System, which might be, does not want review. Although agency action “unlawfully withheld” may be reviewed under the APA, 5 U.S.C. §§ 551(13), 706(1), the Exchange Act’s § 25(a)(1) does not authorize such review in the court of appeals, and at all events the Commission has not “withheld” action concerning the futures markets.
Third, even final agency action is not reviewable if committed to agency discretion by law, 5 U.S.C. § 701(a)(2). Refusal to prosecute is a classic illustration of a decision committed to agency discretion. Heckler v. Chaney, 470 U.S. 821, 831, 105 S.Ct. 1649, 1655, 84 L.Ed.2d 714 (1985). Section 25(a)(1) does not mention the APA’s committed-to-agency-discretion limitation (it was enacted 12 years before the APA), but the futures markets do not question its pertinence, perhaps because it is in large measure a restatement of the separation-of-powers problem.
Any of these three grounds would require us to dismiss the petition for review without reaching the merits. Because a shortfall of standing would depend on the “prudential” branch of that doctrine, we are not compelled to start there. The parties have devoted their attention to the third ground, which we address without deciding whether the futures markets have standing and whether the no-action letter is “final” action.
Although the futures markets maintain that the SEC used the no-action letter to express a decision that the System need not register as an exchange — in other words, that the letter is a substantive rather than a prosecutorial decision — the letter itself says the opposite.
The foregoing is a staff position regarding enforcement action only and should not be understood to express any legal conclusions regarding the applicability of statutory or regulatory provisions of the federal securities laws. Our position is based solely on the representations you have made; any different facts or conditions might require a different response.
Letter from Richard G. Ketchum to Robert A. McTamaney at 12. Some of the Commissioners’ comments during the public meeting suggested a conclusion that the System is not an exchange, but the Commission as a whole did not render a decision on the point. Its official minutes say: “The Commission, in a unanimous decision, determined not to object to the Division of Market Regulation advising RMJ Securities and RMJ Options of its determination to grant the no-action position with respect to RMJ Securities’ operation of the System.” Comments by Commissioners during a meeting are no more the “decision” of the Commission than comments by judges of this court during oral argument are our opinion or judgment.
Decisions not to prosecute are presumptively non-reviewable under Chaney and many cases preceding it. 470 U.S. at 831, 105 S.Ct. at 1655. Consider, for example, the dogma that the decision of the NLRB’s General Counsel not to prosecute may not be reviewed. NLRB v. United Food Workers, 484 U.S. 112, 108 S.Ct. 413, 98 L.Ed.2d 429 (1987); NLRB v. Sears, Roebuck & Co., 421 U.S. 132, 148, 155, 95 S.Ct. 1504, 1515, 1518, 44 L.Ed.2d 29 (1975); Vaca v. Sipes, 386 U.S. 171, 182, 87 S.Ct. 903, 912, 17 L.Ed.2d 842 (1967). This even though private parties may not enforce the NLRA, and even though the General Counsel’s decision is conclusive (thus “final”) in a way the staff’s recommendations to the SEC are not. Investigations and non-prosecutions by the Nuclear Regulatory Commission also lie within the agency’s discretion and are not reviewable. Arnow v. NRC, 868 F.2d 223, 230-34 (7th Cir.1989). Nothing in [531]*531the ’34 Act suggests that the SEC is subject to greater judicial review. Section 21(a)(1), 15 U.S.C. § 78u(a)(l), provides that the “Commission may, in its discretion, make such investigations as it deems necessary to determine whether any person has violated, is violating, or is about to violate any provision of this chapter”, and § 21(d) adds that when the SEC concludes that someone is “engaged or is about to engage in acts or practices constituting a violation ... it may in its discretion bring an action in the proper district court”. Investigation and prosecution under § 21 are discretionary, not mandatory.
Doing nothing may be the most constructive use of the Commission’s resources. Congress gives the SEC a budget, setting a cap on its personnel. With limited numbers of staff-years, the Commission must enforce several complex statutes. To do this intelligently the Commissioners must assign priorities. Prosecuting the System means less time for something else — investigating claims of fraud in issuing new stock or conducting a takeover contest, resolving disputes under the Investment Company Act, and so on. Agencies may find it worthwhile to give short shrift to a particular claim if the aggrieved party can file its own suit (as the futures markets may), for turning the subject over to private litigation frees up time without necessarily diminishing the enforcement of the statute. Yet even when the aggrieved party cannot vindicate its own rights, as with the National Labor Relations Act — indeed, even when the person complaining about failure to prosecute is a defendant whose business is going down the tubes, see FTC v. Universal-Rundle Corp., 387 U.S. 244, 87 S.Ct. 1622, 18 L.Ed.2d 749 (1967); Moog Industries, Inc. v. FTC, 355 U.S. 411, 78 S.Ct. 377, 2 L.Ed.2d 370 (1958)—decisions about the best use of the staff’s time are for the prosecutor’s judgment.
Courts cannot intelligently supervise the Commission’s allocation of its staff’s time, because although judges see clearly the claim the Commission has declined to redress, they do not see at all the tasks the staff may accomplish with the time released. Agencies must compare the value of pursuing one case against the value of pursuing another; declining a particular case hardly means that the SEC’s lawyers and economists will go twiddle their thumbs; case-versus-case is the daily tradeoff. Judges compare the case at hand against a rule of law or an abstract standard of diligence and do not see the opportunity costs of reallocations within the agency. That fundamental difference in the perspectives of the two bodies is why agencies (and other prosecutors) rather than courts must make the decisions on pursuing or dropping claims. Resource allocation is not a task governed by “law”. It is governed by budgets and opportunities. Agencies “take Care that the Laws be faithfully executed” (Art. II, § 3) by doing the best they can with the resources Congress allows them. Judges could make allocative decisions only by taking over the job of planning the agency’s entire agenda, something neither authorized by statute nor part of their constitutional role.
Undoubtedly some aspects of decisions to press one claim and drop another are governed by law and so may be reviewed. If the SEC had said that it was not going to look into the futures markets’ claims because their members had donated too much money to the Democratic Party, a court might have a legitimate role. Chaney said as much, 470 U.S. at 833 n. 4, 105 S.Ct. at 1656 n. 4. See also Cass R. Sunstein, Reviewing Agency Inaction After Heckler v. Chaney, 52 U.Chi.L.Rev. 653, 669-71, 676-82 (1985). Nothing of the sort has happened here. The futures markets do not say that the Commission took an impermissible consideration into account. Instead they maintain that because (in their view) the System is an exchange, the Commission must take action against it. Not so. The petition for review of the no-action letter, No. 89-1449, is dismissed for want of a reviewable order.
Ill
The Commission’s order granting Delta’s application to be a “clearing agency” under § 17 A is final and judicially reviewable, but only at the behest of a “per[532]*532son aggrieved”, § 25(a)(1). The futures markets suffer injury in fact from competition, and a favorable decision would shut down a competitor so that the futures markets could reap gains; thus the minimum requirements of Article III are met. Whether the futures markets are within the “zone of interests” protected by § 6 is a harder matter. Although the “zone” test developed under § 10 of the APA, the futures markets do not deny that this approach should be applied to § 25(a)(1) of the ’34 Act as well. Many courts, including this one, Chicago Board of Trade v. SEC, 677 F.2d 1137, 1140 n. 4 (7th Cir.), vacated as moot, 459 U.S. 1026, 103 S.Ct. 434, 74 L.Ed.2d 594 (1982), have decided suits filed by the futures markets to obtain relief from competition, but no decision to date has addressed the question whether business rivals are within the “zone of interests” of the ’34 Act.
Section 6 protects investors and broker-dealers that want to influence the operation of the exchanges through which they trade. The futures markets fit in neither category. They maintain nonetheless that as rivals for investors’ custom, they are well-situated to enforce § 6 and should be allowed to do so. As an original matter this would have little appeal. Futures markets benefit from exclusivity; options trading reduces the volume of futures trading, and competition doubtless reduces the commissions too. Investors, on the other hand, gain from the competition the futures markets dislike. Futures markets will lay claims under § 6 only if prevailing would break their rivals’ kneecaps — which also could injure investors. Indeed, unless the creation of a new means of trading benefits investors (by introducing new financial instruments or driving down the price of trading), the futures markets won’t file suit. Cf. Block v. Community Nutrition Institute, 467 U.S. 340, 104 S.Ct. 2450, 81 L.Ed.2d 270 (1984). Litigants pursuing private advantage may seek to mislead the court about the scope of the statute they invoke, arguing that it bars more competition than it does. Or they may litigate to raise their rivals’ costs of doing business, not caring whether they prevail. See Premier Electrical Construction Co. v. National Electrical Contractors Ass’n, Inc., 814 F.2d 358, 371-72 (7th Cir.1987); Robert H. Bork, The Antitrust Paradox 347-64 (1978).
Some bodies of law are designed for the protection of competitors, making competitors the logical plaintiffs. Securities law, though, was not designed to protect futures markets from securities exchanges; the futures markets are not litigating to claim the spoils of politics. Nonetheless, Clarke v. Securities Industry Ass’n, 479 U.S. 388, 107 S.Ct. 750, 93 L.Ed.2d 757 (1987), seems dispositive against this approach. The Securities Industry Association sought to squelch competition by banks, and although the SIA’s motives were as antithetical to consumers’ interests as the futures exchanges’ interests potentially are, the Court allowed the SIA to press the litigation. Starting from the presumption that administrative decisions are reviewable, the Court wrote that the plaintiff may be turned away only when “the plaintiff's interests are so marginally related to or inconsistent with the purposes implicit in the statute that it cannot reasonably be assumed that Congress intended to permit the suit.” 479 U.S. at 399,107 S.Ct. at 757. It added: “there need be no indication of congressional purpose to benefit the would-be plaintiff.” Id. at 399-400, 107 S.Ct. at 757. The SIA is no better champion of banks’ customers (or of the Treasury, the insurer of banks afflicted with a surfeit of competition) than the futures markets are of investors in options. See also City of Milwaukee v. Block, 823 F.2d 1158 (7th Cir.1987) (holding that Clarke allows one port to litigate against rules that cause cargo to be shipped through other ports).
Perhaps if the securities laws were designed for the aid of investors exclusively, a court could conclude that Congress meant to forbid litigation by members of the industry. United States v. Naftalin, 441 U.S. 768, 774-77, 99 S.Ct. 2077, 2082-83, 60 L.Ed.2d 624 (1979), allowing a broker to sue a customer who, the broker claimed, committed fraud, scotches any contention that investors’ interests always come ahead [533]*533of industry insiders’. So although a court should bear in mind that the futures markets aim to protect themselves, and that investors are incidental beneficiaries of this litigation if they are beneficiaries at all, we believe that the Board of Trade and the Mercantile Exchange are entitled to a disposition on the merits.
IV
The futures markets won’t get far, however, unless the status of the System matters to the registration of Delta as a clearing agency. Delta is not an “exchange”; it is only one of three participants in a venture that collectively may be an “exchange”. Nothing in § 17A expressly requires the Commission to decide whether a putative clearing agency will participate in an “exchange” as opposed to over-the-counter trading. Clearing agencies may be used in either ease. Acknowledging this, the futures markets observe that before the SEC may register a clearing agency, it must conclude that the applicant “is so organized and has the capacity ... to comply with the provisions of this chapter and the rules and regulations thereunder ... and to carry out the purposes of this section.” Section 17A(b)(3)(A). A clearing agency at the center of an unregistered exchange does not “comply with the provisions of this chapter”, the futures markets insist, and so cannot be registered.
This argument emphasizes “comply”. Counsel for the SEC emphasize “capacity”. Delta has the capacity to comply with the '34 Act in the sense that it has an elaborate record-keeping and cross-check system to protect the clients’ funds and ensure that its staff knows about the rules; it has a supervisory structure that can enforce the rules, and so on. Delta is not tied by any law to one market. Clearing houses such as the Options Clearing Corporation clear and guarantee trades on many exchanges and markets; Delta, too, could work with entities other than the System, including registered exchanges and markets that are not required to register. Delta’s application must stand or fall on what Delta is and proposes to do, the SEC believes, not on whether Delta will transact business with someone else who fails to comply with the law. Failings in the System therefore do not imply that Delta lacks the “capacity” to comply with the Act. To this the futures markets snort: “Be serious!” Delta was formed with a single purpose — to be the clearing house for the System. It is part of the System, the group of persons that collectively (the futures markets say) are an “exchange”. Delta does not propose to clear trades elsewhere. An integral component of the System cannot be said (the exchanges insist) to have the “capacity ... to comply” with the Act if the larger thing of which it is a part thumbs its nose at the registration requirement.
Section 17A does not reveal whether the emphasis should lie on “capacity” or on “comply”. A statute sensibly might have either slant. Section 17A entered the statute as a minor part of a package of amendments in 1975. The parties have not brought to our attention any legislative history bearing on this point, and we could not find anything that did more than repeat the language of the statute. E.g., S.Rep. 94-75, 94th Cong., 1st Sess. 123 (1975), U.S.Code Cong. & Admin.News 1975, pp. 179, 300. No judicial decision interprets the text, and the SEC itself has paid no attention to the difference in readings.
Unless § 17A operates differently from other registration provisions, however, we doubt that the Commission itself (as opposed to its appellate lawyer) would put the emphasis on “capacity” rather than “comply”. If a broker-dealer sought registration trumpeting that it planned routinely to sell stock in interstate commerce without registering pursuant to § 5 of the Securities Act of 1933, the Commission undoubtedly would decline to let the applicant into the industry. Would the Commission register the clearing agency for a bucket shop or Ponzi scheme just because the applicant had a sophisticated computer system and could do a good job if it chose to go straight? Registration as a clearing agency depends on a prediction about compliance with law. Congress assumed that an [534]*534applicant would produce a business plan that, if faithfully executed, would comply. But would it be carried out as written? Thus the role of “capacity”: unless the SEC finds that the applicant is able and likely to comply, it must deny the application. Diligently investigating and certifying the applicant’s “capacity” to do what’s right at the same time as the applicant announces that it plans to do no such thing is bizarre.
The System proclaims its unwillingness to register under § 6, and Delta is a cornerstone of the System. Delta does not claim to have other clients lined up, so that it could put its “capacity ... to comply” to use in an operation that actually complies. We think that an ordinary understanding of the role statutes such as § 17A play in a system of regulation compels the conclusion that the applicant must present the Commission with a business plan that if carried out actually would comply with the requirements of the ’34 Act, and show in addition that the plan is likely to be carried out.
The Commission’s order granting Delta’s application — as opposed to the brief filed by its counsel — implicitly takes this view of § 17A. Although the order does not interpret § 17A, indeed does not recognize the difference between the readings of the statute, the order also does not say anything like: “It doesn’t matter whether the System is an exchange because Delta has the capacity to comply with the Act when it deals with some other marketplace”. On the contrary, the order reflects an assumption that Delta’s registration turns on the status of the System. After mentioning Delta’s recordkeeping system and its “facilities management arrangements”, the Commission continues:
Commentators argue that Delta’s registration would contravene that portion of section 17A(b)(3)(A) requiring a registered clearing agency to be organized and have the capacity to comply with the Act. Specifically, they argue that the System would operate as an unregistered exchange in violation of sections 5 and 6 of the Act, and that Delta would be an aider and abettor to those violations. As noted above, the Division [of Market Regulation] believes that, at this time, the System is not required to register as an exchange. Based upon that determination, the Commission believes Delta’s role in the formation and administration of the System will not impair Delta’s capacity to comply with the Act.
54 Fed.Reg. at 2020 (footnotes omitted, emphasis added). This is the sum total of the SEC’s discussion of the objections the futures markets lodged. We read it as a conclusion (at least a concession) that actual compliance with the registration requirements, rather than “capacity” to comply, is essential to its decision under § 17A. If the Commission had a more confined view of its function it should have said so. We take it as having bitten off what it thought it is required to chew.
V
At this point the futures markets seek to deliver the knockout blow. Having conceded that it matters whether the System is an exchange, the SEC erred in finding the System not to be one. According to the futures markets, something is an “exchange” within the meaning of § 3(a)(1) if (1) for a price based on transactions executed, it (2) brings together buyers and sellers of (3) fungible instruments (or standardized options), (4) clearing the transactions thus executed and (5) disseminating price and other trading information, and also: (6) admits and (7) disciplines members; and establishes trading rules for (8) members and (9) customers. The futures markets extract these elements from various cases and studies. See, e.g., Leist v. Simplot, 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); LTV Federal Credit Union v. UMIC Government Securities, Inc., 523 F.Supp. 819, 834-35 (N.D.Tex.) (Higginbotham, J.), affirmed, 704 F.2d 199 (5th Cir.1983); General Accounting Office, Securities and Futures—How the Markets Developed and How They are Regulated 25, 124 (1986); [535]*535Board of Governors of the Federal Reserve System, Commodity Futures Trading Commission, & Securities and Exchange Commission, A Study of the Effects on the Economy of Trading in Futures and Options III — 28—29, III — 40—41 (1984). Perhaps an “exchange” may have fewer than all of these attributes. The System satisfies all of these, according to the futures markets. The System is a liquid trading market charging transactions fees and clearing transactions; it admits and disciplines those who can trade there, and it has position limits and related standards. Although options do not have to be fungible (that is, have standard strike prices, premiums, and expiration dates), the futures markets say that the System has standardized their terms.
Members of the System rejoin that much the same could be said about the National Association of Securities Dealers Automated Quotation System (NASDAQ) or any other advanced over-the-counter market. As the System sees things, the very proprietary character — the lack of dealer-members — that makes it oppose registration under § 6 also shows that it is not an “exchange” in the first place. True, the dealers on the NASDAQ act as principals, buying and selling for their own accounts rather than matching customers, while RMJ and Delta serve “matchmaking” functions exclusively, but the sponsors of the System counter that they lack a trading floor, specialists, and other common features of an “exchange”. Although it trades standardized options, the System trades unique instruments too. (Anyway, it might add, standardization of the options’ terms might make the options look more like futures contracts without necessarily making the System look more like an exchange.)
The resolution of such a debate has no foreordained outcome. The System is neither fish nor fowl, neither an exchange after the pattern of the Board of Trade and the New York Stock Exchange nor an over-the-counter market after the fashion of the NASDAQ. Developments in automation and communications are bound to produce more of these hard-to-classify entities. Section 3(a)(1) is a product of the ’30s, the System a product of the ’80s. We could not find a single case under § 3(a)(1) discussing which attributes (if any) are necessary, and which are sufficient, for sorting a trading apparatus into the “exchange” bin. Language in § 3(a)(1) referring to “the functions commonly performed by a stock exchange as that term is generally understood” suggests that there is plasticity to the definition. Administrative agencies are entitled — more, exist — to fill in such blanks, to translate the statute from the books to the economic realm. E.g., 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). When Congress establishes the rules, an agency must carry them out. A desire to keep the law “up to date” does not justify departure from its rules. E.g., Board of Governors v. Dimension Financial Corp., 474 U.S. 361, 106 S.Ct. 681, 88 L.Ed.2d 691 (1986); American Bankers Ass’n v. SEC, 804 F.2d 739 (D.C.Cir.1986). Administrative power depends on delegation, and delegation comes from Congress rather than a desire, however keen, to innovate. But § 3(a)(1) does not cast the definition of an “exchange” in the mold of 1934; it invites reinterpretation as the way the term is “generally understood” evolves. If we had before us a decision by the SEC analyzing the subject, weighing the pros and cons of a particular reading of “exchange”, and reaching a reasoned decision, we would be most reluctant to say nay.
Unfortunately, however, the SEC has never made such a decision. Not in this case, not in any other. Although one or another permutation of the proposal to establish the System has been before the SEC for five years, and although other proprietary trading devices have been operating for longer periods, the SEC has never explained whether it views such entities as “exchanges” — and if not, why not. The SEC’s brief filed in this court stoutly maintains that the SEC has not determined whether the System is an exchange and defends the position (which we have now rejected) that it didn’t have to. The futures markets insist that the SEC did de[536]*536cide that the System is not an exchange and erred in so doing, but we have rejected their argument that the Commission made such a decision when issuing the no-action letter, and it seems tolerably clear that the Commission didn’t make it when registering Delta as a clearing agency either. Recall the language of its order:
As noted above, the Division [of Market Regulation] believes that, at this time, the System is not required to register as an exchange. Based upon that determination, the Commission believes Delta’s role in the formation and administration of the System will not impair Delta’s capacity to comply with the Act.
Readers could understand this in a number of ways, but no matter what one makes of the passage it is impossible to find in it a conclusion by the Commission that the System is not an “exchange”. The Commission imputed such a decision to the Division of Market Regulation, but for reasons we have given (reasons that the Commission’s own lawyer advanced in this court), the Division made no such “determination”. Director Ketchum expressly and repeatedly disclaimed deciding whether the System is an “exchange”, and by voting not to oppose issuance of his letter the Commission held even this non-conclusion at arms’ length.
Even if we read the Commission’s two sentences the way the futures markets do, as adopting the Director’s conclusion that the System is not an exchange, there is a fatal flaw. It didn’t give reasons. The order registering Delta as a cleaning agency spans 13 pages of the Federal Register, which has microscopic type. Despite this torrent of words, the Commission does not even cite § 3(a)(1), let alone interpret it. Adopting a “determination” that the Division of Market Regulation did not make, and which itself does not cite or discuss § 3(a)(1), hardly fills the gap. The order registering Delta as a clearing agency is 100% boilerplate. The Commission never came to grips with the only serious objection to Delta’s application, and an agency is not entitled to duck just because the question is hard.
“If an order is valid only as a determination of policy or judgment which the agency alone is authorized to make and which it has not made, a judicial judgment cannot be made to do service for an administrative judgment.” SEC v. Chenery Corp., 318 U.S. 80, 88, 63 S.Ct. 454, 459, 87 L.Ed. 626 (1943). Before concluding that Delta’s proposed operations' could “comply” with the ’34 Act, the Commission had to determine that the System is not an exchange. It didn’t make this decision. Neither our analysis nor that of counsel, id. at 92, 63 S.Ct. at 461, can supply the deficit — though the Commission’s lawyers did not even try, resting on an argument that the SEC needn’t resolve the point. Unless we were to construe § 3(a)(1) without the assistance of either the agency or its lawyers, we could not sustain its decision.
The SEC rationally could decide either way. If the System is not an exchange then one new form of competition will flourish — but at a price in contraction of the SEC’s regulatory jurisdiction that it may deem excessive. (The Commission may think its powers as regulator of the brokers and clearing houses in proprietary systems inadequate; the proposed rule to govern their operation suggests as much, see 54 Fed.Reg. at 15429-30.) As Chenery limits the judicial part to review of decisions taken elsewhere, silence at the SEC means that we must vacate the order registering Delta as a clearing agency and remand so that the SEC can decide in the first instance. See also South Prairie Construction Co. v. Operating Engineers, 425 U.S. 800, 806, 96 S.Ct. 1842, 1845, 48 L.Ed.2d 382 (1986).
Because this decision has the potential to unsettle the expectations of the many investors who have traded on the System, and to require the closing of all positions Delta has taken and guaranteed, we defer for 120 days after the date of our mandate the effectiveness of our judgment vacating the SEC’s order. Cf. Buckley v. Valeo, 424 U.S. 1, 144, 96 S.Ct. 612, 694, 46 L.Ed.2d 659 (1976). Although the remand to the SEC occurs with our mandate, the SEC has 120 days after that to decide [537]*537whether the System is an “exchange” and to re-register Delta or decline to do so. A decision taken within that time will be effective as a new decision. If, however, the SEC has not made up its mind by then, Delta must cease doing business as a clearing agency, just as it was out of the business as an initial applicant.
Petition No. 89-1449 is dismissed for want of a reviewable order. The order registering Delta as a clearing agency is vacated 120 days from the date of our mandate, and the matter is remanded to the Commission for further proceedings consistent with this opinion.