This is an appeal from an order of the district court invalidating under the Glass-Steagall Act a decision of appellant Board of Governors of the Federal Reserve System that permitted appellant Bankers Trust Company, a state-chartered commercial bank and a member of the Federal Reserve System, to place commercial paper issued by third parties. The Act prohibits commercial banks from engaging in investment banking. The Board of Governors determined that Bankers Trust's activities did not cross the line into investment banking, but the district court concluded that they did. After considering the language and history of the Act and the applicable case law, we reverse the judgment of the district court and reinstate the Board's decision.

"Commercial paper" comprises unsecured, large denomination promissory notes written with maturities of less than nine months to supply the current capital needs of corporate issuers. In privately negotiated transactions, issuers typically place commercial paper with large, financially sophisticated institutional investors (such as insurance companies or pension funds).

Bankers Trust acts as an advisor and agent to commercial paper issuers by advising each issuer of the interest rates and maturities that institutional investors are likely to accept, by soliciting prospective purchasers for commercial paper the client decides to issue, and by placing the issue with the purchasers. Bankers Trust does not make any general advertisement or solicitation regarding any issue it is seeking to place, and does not place any issues with individuals or the general public.

Bankers Trust receives a commission for its services based upon a percentage of the issuer's total outstanding commercial paper during a one-year period. To ensure that it acts solely as an agent without an independent financial stake in the success of issues it places, which would clearly involve it in investment banking, Bankers Trust does not purchase or repurchase for its own account, inventory overnight, or take any ownership interest in any commercial paper it places. Nor does Bankers Trust any longer make loans on or collateralize loans with the paper it places (a practice it formerly followed when necessary to remedy any deficiency in placement of an issue).

This appeal is the latest installment in a dispute that began in 1979 when the Securities Industry Association ("SIA"), a trade association of underwriters, brokers, and securities dealers, petitioned the Board of Governors for a ruling that it was unlawful for Bankers Trust and other commercial banks to sell commercial paper issued by unrelated entities. The Board ruled against the SIA, but ultimately the Supreme Court, disagreeing with the Board of Governors, held that commercial paper is included within the category of "notes or other securities" addressed by the Banking Act of 1933, commonly known as the Glass-Steagall Act, and remanded the case for a determination of an unresolved issue: whether Bankers Trust's placement of commercial paper constituted the "underwriting" or "business of issuing, underwriting, selling or distributing" that the Act prohibits. Securities Indus. Ass'n v. Board of Governors of the Fed. Reserve Sys., 468 U.S. 137, 160 n. 12, 104 S. Ct. 2979, 2992 n. 12, 82 L. Ed. 2d 107 (1984) (SIA ).

Upon remand, the Board of Governors found that Bankers Trust's placement of commercial paper constituted the "selling" of a security without recourse and solely upon the order and for the account of customers, a practice permitted by section 16 of the Act, 12 U.S.C. Sec. 24 (Seventh) (1982). Federal Reserve System, Statement Concerning Applicability of the Glass-Steagall Act to the Commercial Paper Activities of Bankers Trust Company (June 4, 1985) ("Board Statement"), Joint Appendix ("J.A.") at 195. The district court reviewed the Board's decision on the petition of the SIA and granted SIA summary judgment, holding that Bankers Trust's activities involved the "underwriting" and "distributing" prohibited by section 21(a)(1) of the Act, 12 U.S.C. Sec. 378(a)(1) (1982). Securities Indus. Ass'n v. Board of Governors of the Fed. Reserve Sys., 627 F. Supp. 695 (D.D.C.1986). This appeal followed.II.

In reviewing the Board's decision, we owe the agency's determination "the greatest deference." Board of Governors of the Fed. Reserve Sys. v. Investment Co. Inst., 450 U.S. 46, 56, 101 S. Ct. 973, 981, 67 L. Ed. 2d 36 (1981) (ICI ); accord Securities Indus. Ass'n v. Board of Governors of the Fed. Reserve Sys., 468 U.S. 207, 217, 104 S. Ct. 3003, 3009, 82 L. Ed. 2d 158 (1984) (Schwab ) (giving Board "substantial deference"); see also Board of Governors of the Fed. Reserve Sys. v. Agnew, 329 U.S. 441, 450, 67 S. Ct. 411, 415, 91 L. Ed. 408 (1947) (Rutledge, J., concurring) ("[The Board's] specialized experience gives [it] an advantage judges cannot possibly have, not only in dealing with the problems raised for [its] discretion by the system's working, but also in ascertaining the meaning Congress had in mind in prescribing the standards by which [the Board] should administer it."). This principle is not contradicted by SIA, 468 U.S. at 143-44 (according only "little deference"), or Investment Co. Inst. v. Camp, 401 U.S. 617, 626-28, 91 S. Ct. 1091, 1097, 28 L. Ed. 2d 367 (1971) (Camp ) (rejecting a deferential approach).

In the latter cases, the agency involved failed to present the Court with anything to which to defer. In Camp, Justice Stewart, writing for the majority, noted that "courts should give great weight to any reasonable construction of a regulatory statute adopted by the agency charged with the enforcement of that statute," 401 U.S. at 626-27, 91 S.Ct. at 1097, but said the "difficulty" was that the Comptroller of the Currency had promulgated the challenged regulation "without opinion or accompanying statement," id. at 627, 91 S.Ct. at 1097. Without the benefit of any "expressly articulated position at the administrative level," the Court refused to defer to the agency's position, reasoning that "[i]t is the administrative official and not appellate counsel who possesses the expertise that can enlighten and rationalize the search for the meaning and intent of Congress." Id. at 628, 91 S.Ct. at 1097-98.

In SIA, the Board had provided an opinion explaining its view of whether commercial paper constituted "securities" for purposes of the Glass-Steagall Act but failed to analyze the legislative purposes behind the Act. Because of this omission, the Court gave "little deference" to the Board's position that its interpretation ran afoul of none of the purposes of the Act. 468 U.S. at 143-44, 104 S.Ct. at 2983-84. The Court at the same time observed generally that because "[t]he Board is the agency responsible for federal regulation of the national banking system, ... its interpretation of a federal banking statute is entitled to substantial deference." Id. at 142, 104 S.Ct. at 2983.

In the present case, as in ICI and Schwab, the Board has comprehensively addressed the language, history, and purposes of the Act that bear on whether commercial banks should be able to place commercial paper. We consequently owe the Board's determination "substantial deference" or "significant weight," and we must look to Chevron U.S.A. Inc. v. NRDC, 467 U.S. 837, 104 S. Ct. 2778, 81 L. Ed. 2d 694 (1984), to guide our application of such principles of review. See Investment Co. Inst. v. Conover, 790 F.2d 925, 932 (D.C.Cir.1986). Since Congress has not clearly addressed the question of whether activities such as those conducted by Bankers Trust fall within the prohibitions of the Act, we must examine whether the agency, in filling the statutory gap left by Congress, has acted reasonably. Chevron, 467 U.S. at 843-45, 104 S.Ct. at 2781-83.

The question in this case involves the interplay of sections 16 and 21 of the Glass-Steagall Act. These provisions implement what the Supreme Court has described as the Act's "general purpose of separating as completely as possible commercial from investment banking," ICI, 450 U.S. at 70, 101 S.Ct. at 989. Section 16, 12 U.S.C. Sec. 24 (Seventh) (1982), draws the line between permissible and impermissible activities for commercial banks, while section 21(a)(1), 12 U.S.C. Sec. 378(a)(1) (1982), draws this line for investment banks. The Supreme Court has found that "Sec. 16 and Sec. 21 seek to draw the same line." SIA, 468 U.S. at 149. The issue before us is whether the Board has reasonably determined that the activities of Bankers Trust do not cross that line between commercial and investment banking.

Section 16 of the Act provides in relevant part that the "business of dealing in securities and stock by [a commercial bank] shall be limited to purchasing and selling such securities and stock without recourse, solely upon the order, and for the account of, customers, and in no case for its own account, and the [bank] shall not underwrite any issue of securities or stock." 12 U.S.C. Sec. 24 (Seventh) (1982). Section 21(a)(1) makes it "unlawful" for

any person, firm, corporation, association, business trust, or other similar organization, engaged in the business of issuing, underwriting, selling, or distributing, at wholesale or retail, or through syndicate participation, stocks, bonds, debentures, notes, or other securities, to engage at the same time to any extent whatever in the business of receiving deposits subject to check or to repayment upon presentation of a passbook, certificate of deposit, or other evidence of debt, or upon request of the depositor.

12 U.S.C. Sec. 378(a)(1) (1982). Because no one disputes that Bankers Trust constitutes a commercial bank within the meaning of these sections, we must determine, first, if the Board has reasonably concluded that the commercial paper placement activities of Bankers Trust fall within the permissive language of section 16. To determine this, we must look at the question of what the statute means by "underwrite," for underwriting not only triggers section 21's prohibitions but also defeats section 16's permissive effect. We must, in contrast, address the meaning of section 21's terms "issuing, ... selling, or distributing" only if section 16 is inapplicable. In other words, section 21 cannot be read to prohibit what section 16 permits. See ICI, 450 U.S. at 63, 101 S.Ct. at 985 (section 21 not intended to bar banking practices permitted by section 16); see also United States v. Menasche, 348 U.S. 528, 538-39, 75 S. Ct. 513, 519-20, 99 L. Ed. 615 (1955) (rejecting an interpretation of a statutory provision that would nullify the effect of another provision). Therefore, if we find that the Board acted reasonably in concluding that section 16 permits Bankers Trust's activities, that is the end of our analysis.

While this proposition seems obvious, SIA nonetheless argues that we must examine both the prohibitions of section 21 and the permissive phrase of section 16 to determine if the Board has erred. SIA seeks to restrict the scope of section 16 by relying on language (added to section 21 in 1935) that expressly refers to section 16 as an exception to section 21's restrictions, stating that "the provisions of this paragraph shall not prohibit national banks or State banks ... from dealing in, underwriting, purchasing, and selling investment securities, or issuing securities, to the extent permitted to national banking associations by the provisions of [section 16 of the Act]." 12 U.S.C. Sec. 378(a)(1) (1982). SIA directs our attention to a passage in the House Report accompanying the 1935 amendments stating that this language was added to "make it clear that [section 21] does not prohibit any financial institution or private banker from engaging in the securities business to the limited extent permitted to national banks under [section 16]." H.R.Rep. No. 742, 74th Cong., 1st Sess. 16 (1935). The Report goes on to say parenthetically that this provision permits commercial banks to deal in or underwrite only certain enumerated government obligations not at issue in this case. Id. SIA has suggested that the overlap between sections 16 and 21 is restricted to this narrow context, and that we must affirm the district court if we find, as SIA argues, that the language of section 21 covers the transaction permitted by the Board in this case.

SIA's argument is wholly unpersuasive for several reasons. First, the House Report said that the cross-reference to section 16 was being added to "make it clear" that commercial banks could underwrite and deal in certain government obligations, and the title of the relevant passage in that Report was "Section 21 of the Banking Act Clarified." H.R.Rep. No. 742, supra, at 16. That this amendment sought merely to clarify the relationship between section 16 and section 21 necessarily implies that before the amendment the two provisions by their own force had to be read together. Congress amended section 21 simply to leave no doubt of the need to read the two sections harmoniously in a matter of particular congressional concern. Moreover, apart from this apparent purpose, the sweeping and comprehensive language employed explicitly provides that any restriction on the activities of a commercial bank that may arise because of the prohibitions of section 21 is relieved insofar as the activity is permitted by section 16. This unambiguous language controls our reading of the statute.

We reject SIA's position for a second and independently decisive reason. If section 21 prohibited what section 16 explicitly permits, section 21 would render section 16's permissive language entirely nugatory--an absurd result. Section 16 allows a commercial bank to sell securities "without recourse, solely upon the order, and for the account of, customers, and in no case for its own account." Section 21, in sharp contrast, flatly prohibits a commercial bank from "selling" securities. If we permitted the restrictions of section 21 to control, the "selling" of securities would be entirely proscribed, despite the explicit permission contained in the statute. If, on the other hand, we read section 16's permissive provisions as an exception to section 21, the restriction on selling in section 21 would retain its force for all activities not permitted by section 16. We must therefore read section 21's operative terms (issuing, underwriting, selling, distributing) to exclude any activity section 16 allows. (We discuss below the prohibition of underwriting by a commercial bank contained in section 16 itself.)

Finally, because the Supreme Court has stated that sections 16 and 21 "seek to draw the same line" between commercial and investment banking, SIA, 468 U.S. at 149, 104 S.Ct. at 2986, those activities of commercial banks that section 16 places on the acceptable commercial banking side of the line cannot be placed by section 21 on the impermissible investment banking side of the line. Thus, if the Board reasonably found that section 16 permits Bankers Trust's activities, our inquiry ends there.

We believe that the Board's determination is reasonable. The Board found that Bankers Trust's activities fell within section 16's requirement that "[t]he business of dealing in securities and stock by the [bank] shall be limited to purchasing and selling such securities and stock without recourse, solely upon the order, and for the account of, customers, and in no case for its own account, and the [bank] shall not underwrite any issue of securities or stock." 12 U.S.C. Sec. 24 (Seventh) (1982). While the district court did not dispute the Board's finding that Bankers Trust's activities "fit neatly within the literal language of section 16's permissive phrase," and relied instead on an analysis of the legislative purposes of the Act to reverse the Board's holding, see Securities Indus. Ass'n v. Board of Governors, 627 F. Supp. at 701-02, SIA vigorously disputes on several grounds the conclusion that the bank's activities come within the permissive language of that section. We take up these arguments in turn.

SIA argues that section 16's permissive language does not apply to the activities of Bankers Trust because the section 16 exception applies only to "the business of dealing in securities and stock," while SIA asserts that the term "dealing" is typically understood to encompass the purchasing and selling of securities only in the secondary trading market. In support of this argument, SIA cites the definition of "dealer" contained in the Securities Act of 1933. In fact, the Securities Act undercuts SIA's position by defining a "dealer" as "any person who engages ... as agent, broker, or principal, in the business of offering, buying, selling, or otherwise dealing or trading in securities issued by another person," without any exclusion of the primary offering market. 15 U.S.C. Sec. 77b(12) (1982). Moreover, the so-called "dealer's exemption" to registration of securities under the Securities Act makes it clear that one may be a dealer under that statute in both secondary and primary markets. This exemption from registration, contained in section 4(3) of the Securities Act, 15 U.S.C. Sec. 77d(3) (1982), applies to "transactions by a dealer (including an underwriter no longer acting as an underwriter in respect of the security involved in [the] transaction)." This language necessarily implies that one may be a dealer--"in the business of offering, buying, selling, or otherwise dealing or trading in securities"--and still act as an underwriter--one who unquestionably may participate in the primary or new issue market. Congress' ordinary understanding of "the business of dealing" clearly was not restricted to secondary trading; SIA's argument on this point is unsupportable.

SIA also contends that the Board erred in concluding that the activities of Bankers Trust are "upon the order ... of ... customers," claiming that Congress imposed this restriction to make it clear that banks could perform the transactions permitted by section 16 only as an accommodation to the existing customers of the bank. In support of this proposition, SIA relies almost exclusively on the Supreme Court's recent opinion in Schwab, in which SIA unsuccessfully challenged a bank holding company's retail brokerage operations under section 20 of the Glass-Steagall Act. See 12 U.S.C. Sec. 377 (1982) (prohibiting bank affiliation with any firm "engaged principally in the issue, flotation, underwriting, public sale, or distribution" of securities). In approving the retail brokerage operation, run by a non-bank affiliate of the bank holding company as an accommodation to the affiliate's customers, the Court relied in part on the fact that section 16 "allows banks to engage directly in the kind of [retail] brokerage activities at issue here, to accommodate [their] customers." 468 U.S. at 221, 104 S.Ct. at 3011. SIA suggests that this statement amounted to an interpretation of section 16 requiring banks to provide securities services under the relevant language only as an accommodation to the bank's preexisting customers. This argument misses the mark. While the Court did state that section 16 permitted retail brokerage as an accommodation to customers of the bank's other services, it specifically left open the question whether such securities brokerage, if more broadly available, would still satisfy that provision. Id. at 219 n. 20, 104 S. Ct. at 3011 n. 20. Thus, the Court did not, as we must, decide whether section 16 allows the placement of securities only as an accommodation to existing customers of other bank services.

While the meaning of "upon the order ... of ... customers" is decidedly ambiguous, we defer, as Chevron requires, to the Board's reasonable conclusion that section 16 should not be given such a narrow reading. The Board below correctly observed that "[n]othing in the literal terms of section 16 requires a preexisting customer relationship." Board Statement at 16, J.A. at 210. According to SIA, however, the legislative history makes it clear that Congress had such a relationship in mind when it included the language "upon the order, and for the account of, customers." In support of its thesis, SIA directs us to a remark in the committee reports that the purpose of section 16 was to permit "[n]ational banks ... to purchase and sell investment securities for their customers to the same extent as heretofore." S.Rep. No. 77, 73d Cong., 1st Sess. 16 (1933); H.R.Rep. No. 150, 73d Cong., 1st Sess. 3 (1933). Because the Supreme Court in Schwab stated that "[b]anks long have arranged the purchase and sale of securities as an accommodation to their customers," and that section 16, read in conjunction with the above-cited legislative history, "expressly endorsed this traditional banking service," 468 U.S. at 215, 104 S.Ct. at 3008, SIA contends that section 16 was intended only to permit such services as had been traditionally performed as an accommodation to preexisting customers of the bank.

SIA again reads too much into Schwab. Congress, in enacting section 16, may well have intended to endorse traditional banking services as they existed before 1933. This does not mean, however, that section 16 permits the transactions covered by its language only to the extent that identical transactions occurred prior to the enactment of Glass-Steagall. Since nothing in the language or legislative history of section 16 even remotely suggests that the Act meant to freeze particular functions in place as of 1933, we decline to read that meaning into the Act.

Moreover, the history of commercial banking shows that, prior to the Glass-Steagall Act, banks offered the securities brokerage services at stake in Schwab both to existing customers and to persons with no preexisting relationship to the banks. See Securities Indus. Ass'n v. Comptroller of the Currency, 577 F. Supp. 252, 255 (D.D.C.1983), aff'd per curiam, 758 F.2d 739, 740 (D.C.Cir.1985) (affirmed "generally for the reasons stated" by district court), cert. denied, --- U.S. ----, 106 S. Ct. 790, 88 L. Ed. 2d 768 (1986); see also Greenfield v. Clarence Sav. Bank, 5 S.W.2d 708, 708-09 (Mo.Ct.App.1928) (transaction in which plaintiff, having no account with the bank, "went there with the sole purpose of purchasing bonds as an investment"); Smith, Stock Market Service Comes High, Am.Bankers A.J. 965 (Apr. 1929) (bank will "buy and sell securities for its customers and the public in general"). Thus, we may not construe "upon the order ... of ... customers" to require a preexisting relationship between the bank and the user of the services permitted under section 16, since Congress intended that language to ratify banking practices that served at least some persons without any relationship to the bank except as "customers" of the services permitted by section 16.

Despite the conclusion that section 16 requires no preexisting customer relationship as a matter of law, we still must inquire whether the Board reasonably concluded that Bankers Trust places commercial paper solely on the order of the issuer. The Board stated that "[a]ccording to Bankers Trust's submission, the issuer, not the bank, decides whether to raise funds by issuing commercial paper and, if so, in what amount." Board Statement at 15-16, J.A. at 209-10. The Board concluded that this meant that "the bank places commercial paper solely on the request and on the order of its customer, the commercial paper issuer." Id. at 18, J.A. at 212. SIA counters that the bank solicits the business of issuers and gives financial advice about the terms and timing of the potential issue of commercial paper. Neither point upsets the Board's conclusion.

SIA offers no support for its claim that Bankers Trust recruits or solicits the business of issuers beyond the assertion that the bank "touts" its placement services in advertisements.1 "Touting" is not enough to render the Board's conclusion unreasonable. Although the bank may generally solicit customers for its placement services by making it known that such services are available, either in the so-called "tombstone ads" or in more general advertisements (for example, "What do you get when you combine an investment bank with a commercial bank? Bankers Trust Company."), any given placement of commercial paper still takes place solely on the order of the customer. We illustrate by analogy. The Supreme Court in Schwab stated that section 16 permitted banks to engage in retail brokerage operations at least as to their own customers. Even if we assume that a bank makes its retail brokerage operations available only to its depositors, it still must find some way to let them know that these services are available. It would strain credulity to assert that the circulation of a brochure or the running of an advertisement to publicize the availability of these services would mean that the brokerage services performed are now barred since no longer performed solely upon the order of the customer.

Nor are we convinced that the rendering of financial advice itself removes Bankers Trust's placements from the category of transactions made solely upon the order of customers. Nothing in section 16 suggests that the bank may not advise issuers who have decided that they may or do want to raise money by issuing commercial paper. If a customer asks the advice of Bankers Trust but makes its own decision about whether and in what amount to issue commercial paper, the transaction is made solely on the order of that customer. Consider, by contrast, a case in which an investment bank decides that the market is favorable to the refinancing of a bond issue or the conversion of debt to equity and initiates discussions with its customer leading up to the eventual transaction. In such a situation, the initiative of the investment banker itself creates the very demand for the particular transaction. This is a far cry from the type of passive advice that Bankers Trust renders after the issuer has decided that it needs to raise capital and must only decide the best way to do it. Indeed, the legislative history provides support for just this distinction, evincing a concern about bankers who found it "necessary ... to seek for customers to become makers of issues of securities when the needs of those customers for long-term money were not very pressing." 75 Cong.Rec. 9911 (1932) (remarks of Sen. Bulkley). We cannot conclude that the Board acted unreasonably in deciding that the danger identified by Senator Bulkley does not characterize the financial advice rendered by Bankers Trust.

SIA also argues that, because section 16 applies both to "purchasing and selling" of securities, the solicitation of buyers of commercial paper by Bankers Trust means that its activities are not "upon the order ... of ... customers." This argument is meritless. Since Bankers Trust acts as sales agent for the issuer, it is clearly engaged in "selling" securities for its customers and necessarily finds and solicits buyers for those securities, buyers who may be customers of other bank services. This does not mean, however, that Bankers Trust is "purchasing" securities for those investors who buy the paper. The buyers decide upon and make their own purchases, while Bankers Trust has an explicit policy against purchasing for any account that it manages, advises, or serves as trustee--the only accounts for which the bank would even have the authority to make such purchases.2

Moreover, no sensible construction of the statute could say that otherwise permissible selling activities cannot involve the solicitation of buyers. The seller's very purpose in engaging a selling agent and paying a commission is to acquire that agent's superior ability to place the product with buyers. If placement of the product with buyers did not require any solicitation of buyers, no rational business would pay another firm to do what it could without cost to itself: passively wait for orders. This construction of "upon the order ... of ... customers," therefore, would lead to the absurd statutory result of allowing a seller-agent relationship to arise only in circumstances that not only would never actually exist but that also would strip the relationship itself of its purpose. The Board's rejection of such a construction appears eminently reasonable.

The final assault on the Board's conclusion that the activities of Bankers Trust fit within the terms of section 16 suggests that these activities amount to "underwriting" and thus divest Bankers Trust of its exemption. Although the Act and its legislative history are barren of any definition of the term "underwriting," the parties and the district court have spent much effort considering whether the term "underwriting" includes agency, as well as principal, transactions, and whether what is commonly called "best efforts" underwriting, in which the selling group assumes none of the risk of its failure fully to distribute the issue, amounts to statutory underwriting for purposes of the exemption.3 These efforts were needless, since we find that the Board reasonably concluded that an "underwriting" defeats the section 16 exemption only if it includes a public offering; private placements therefore do not for this purpose constitute statutory "underwriting." The Board's reliance on the distinction between public offerings and private placements is reasonable because the distinction derives support from congressional intent embodied in contemporaneous securities legislation and reasonably relates to concerns that the Glass-Steagall Act sought to meet.

1. Contemporaneous Securities Legislation --The Glass-Steagall Act nowhere defines "underwriting," and the legislative history contains nothing to clarify the term. When in SIA the Supreme Court reviewed the case we now consider on remand, similar ambiguity surrounded the definition of the terms "security" and "note." The Court in SIA made it very plain that the meaning of a term in other legislation passed roughly at the same time as the Glass-Steagall Act with the shared purpose of restoring confidence in the nation's financial markets provided "considerable evidence" of the "ordinary meaning" Congress attached to the same term in the Glass-Steagall Act itself. SIA, 468 U.S. at 150, 104 S.Ct. at 2987. The statutes to which the Court resorted in discerning congressional understanding of the term "security" were the Securities Act of 1933, 15 U.S.C. Sec. 77a et seq. (1982), the Securities Exchange Act of 1934, 15 U.S.C. Sec. 78a et seq. (1982), and the Public Utility Holding Company Act of 1935, 15 U.S.C. Sec. 79 et seq. (1982). SIA, 468 U.S. at 150, 104 S.Ct. at 2987. In each of these statutes, the sweeping definition of "security" encompasses commercial paper; the Court accordingly found that when Congress meant to exempt commercial paper from the strictures of one of these statutes, it expressly so provided. Id. at 150-51, 104 S.Ct. at 2987. Congress, the Court concluded, understood "that, unless modified, the use of the term 'security' encompasse[d] [commercial paper]." Id. at 151, 104 S.Ct. at 2987.

Only the Securities Act of 1933 defines the term "underwriter" (although the Securities Exchange Act of 1934 provides useful evidence on the meaning of that term as used in the Securities Act). While the evidence of the ordinary congressional cognizance of the term "underwrite" or "underwriter" thus comes from only one piece of similar legislation, that legislation, the Securities Act of 1933, is the closest to Glass-Steagall in time and purpose of the various statutes relied on in SIA. The Securities Act and the Glass-Steagall Act were signed into law within three weeks of each other and both statutes were among the legislative reforms that marked President Roosevelt's first hundred days in office. Thus, while the precise purposes of the Securities Act and the Glass-Steagall Act may differ, both emerged from the same effort to restructure the American financial markets; absent any contrary indication, we must consider Congress' understanding of the financial terms it used in one statute highly relevant to discovering the meaning attached to similar but ambiguous terms in the other. With that rule in mind, we turn to the Securities Act of 1933.

SIA contends that the significance of the Securities Act for this case is that it provides an express exemption from registration for "transactions by an issuer not involving any public offering." 15 U.S.C. Sec. 77d(2) (1982). SIA asserts that this exemption demonstrates Congress' ability knowingly to provide an exemption from statutory requirements for private offerings; Congress' failure so to provide in section 16 of the Glass-Steagall Act means that the Board acted unreasonably when by interpreting the term "underwrite" it effectively read such a private offering into the Act. This point would have considerable force, except that the history of the Securities Act's exemption betrays SIA's argument and, in fact, establishes the converse--that Congress did understand the concept of underwriting to connote involvement in a public offering of securities.

The Securities Act in section 2(11) defines an "underwriter" as "any person who has purchased from an issuer with a view to, or offers or sells for an issuer in connection with, the distribution of any security, or participates or has a direct or indirect participation in any such undertaking, or participates or has a direct participation in the direct or indirect underwriting of any such undertaking." 15 U.S.C. Sec. 77b(11) (1982). An "underwriter" thus cannot exist unless a "distribution" exists.

As originally introduced in the House bill that was to become the Securities Act of 1933, the exemption relied on by SIA applied to "transactions by an issuer not with or through an underwriter." See H.R. 5480, 73d Cong., 1st Sess. Sec. 4(1) (1933). The House Committee added to this language the phrase "and not involving any public offering." H.R.Rep. No. 85, 73d Cong., 1st Sess. 1 (1933). While the deliberate inclusion of both "not with or through an underwriter" and "not involving a public offering" would ordinarily support the conclusion that Congress viewed the coverage of the two phrases as being different, other legislative history shows that, in fact, both phrases had the same coverage. In interpreting the statute contemporaneously with its passage, the Federal Trade Commission, the agency originally charged with administering the securities laws, observed that a statutory "distribution" necessarily involved a "public offering," thus making it clear that one could not be an "underwriter" in the absence of a public offering. See H.R.Conf.Rep. No. 1838, 73d Cong., 2d Sess. 41 (1934). Acknowledging the correctness of the Commission's interpretation, the same Congress that had passed the Securities Act of 1933 eliminated as "superfluous" the language "not with or through an underwriter" when it amended the Securities Act in Title II of the Securities Exchange Act of 1934. Id.; see ch. 404, Sec. 203(a)(1), 48 Stat. 881, 906 (1934); see also 1 L. Loss, Securities Regulation 551 & n. 307 (2d ed. 1961) (distribution "more or less synonymous with" public offering). While by no means conclusive, this history offers support for the reasonableness of the Board's view that Congress understood "underwriting" (and, for that matter, "distribution") of securities to connote a public offering, and that the private offerings of commercial paper effected by Bankers Trust do not come within the Glass-Steagall Act's meaning of "underwriting." At the least, it refutes SIA's contention that the Securities Act undercuts the Board's conclusion in this regard.

One further Securities Act argument made by SIA deserves only brief mention. SIA contends that the Board acted inappropriately in "importing" the public offering/private placement distinction from the Securities Act without requiring adherence to the regulations adopted by the Securities and Exchange Commission in enforcing that distinction. This argument is wholly meritless; the Board properly determined that the SEC regulations simply were not germane to the question at hand. The Board has not, as SIA asserts, sought to "import" a statutory exemption from the Securities Act, but merely has looked to the use of terms in a contemporaneous financial regulatory statute to assist it in discerning what Congress meant when it used similar terms ambiguously in the Glass-Steagall Act. This resort to legislative history does not compel the Board to adopt every subsequent aspect of the Securities Act's enforcement.

But even if this were not so, the relevant SEC rules, collectively known as Regulation D, do not purport to be a definitive interpretation of what constitutes a non-public offering but merely constitute a "safe-harbor" that guarantees non-public offering status to an issue that complies with their terms. Securities offerings not in compliance with Regulation D may nonetheless be exempt from registration as "not involving any public offering." See 17 C.F.R. Sec. 230.501 (Preliminary Note 3) (1985) (issuer's failure to satisfy Regulation D "shall not raise any presumption that the exemption provided by section 4(2) of the [Securities] Act is not available").

The Board's responsibility in this case was to arrive at a reasonable determination of what should constitute a private offering under the Glass-Steagall Act. The Board found Bankers Trust's activities to constitute a private offering because (1) the bank "places commercial paper by separately contacting large financial and non-financial institutions," (2) the bank "does not place commercial paper with any individuals," (3) "the maximum number of offerees and purchasers of commercial paper placed by the bank in any given case is relatively limited," (4) the bank "makes no general solicitation or advertisement to the public" with respect to the placement of particular paper (though it does advertise its activities in business publications to publicize its availability as an agent to issuers), and (5) "the commercial paper placed with the bank's assistance is issued in very large average minimum denominations, which are not a likely investment of the general public." Board Statement at 29-30, J.A. at 223-24. Such considerations properly determine what distinguishes a private from a public offering of securities; we shall shortly see that they also have a strong relationship to one of the principal concerns that animated the Glass-Steagall Act.

2. Legislative Purposes --As the Supreme Court has amply documented, the legislative history of the Glass-Steagall Act shows that, besides "the obvious risk that a bank could lose money by imprudent investment of its funds in speculative securities," Congress sought to address " 'the more subtle hazards that arise when a commercial bank goes beyond the business of acting as fiduciary or managing agent and enters the investment banking business.' " SIA, 468 U.S. at 145, 104 S.Ct. at 2984 (quoting Camp, 401 U.S. at 630, 91 S.Ct. at 1098). The hazards identified by the Court included danger to the impartiality of the bank as a dispenser of financial advice. For example, "Congress concluded that it was unrealistic to expect a banker to give impartial advice about [whether and how best to issue equity or debt securities] if he stands to realize a profit from the underwriting or distribution of securities." SIA, 468 U.S. at 146, 104 S.Ct. at 2984 (citing 75 Cong.Rec. 9912 (1932) (remarks of Sen. Bulkley)). Moreover, the Court pointed to congressional fears that commercial-bank involvement in investment banking might lead to the use of a bank's credit facilities to "shore up a company whose securities the bank sought to distribute" or to facilitate the purchase of securities of the bank's commercial customers. See id. at 146-47, 104 S.Ct. at 2985. Congress, in sum, did not believe that bankers could act as proper fiduciaries if faced with the "pressures" of "involvement in the distribution of securities." Id. at 146, 104 S.Ct. at 2985.

Congress recognized that these pressures largely resulted from the heavy fixed costs incurred by commercial banks in running investment banking operations. In the period immediately preceding the financial collapse that precipitated the enactment of Glass-Steagall, the distribution of an issue of securities took place through an elaborate syndication involving various tiers of purchase, banking, and selling groups managed by an originating banker who handled the negotiations with the issuer. See 1 L. Loss, Securities Regulation 164-66 (2d ed. 1961). The precise details of the distribution process, as it then existed, are not important for our purposes. What is important is that "a large number of the leading originators of securities, particularly the security affiliates of commercial banks," developed "large selling organizations" in this period. Gourrich, Investment Banking Methods Prior to and Since the Securities Act of 1933, 4 Law & Contemp. Probs. 44, 49 (1937). Indeed, the "bank affiliates were particularly active in constructing substantial retail organizations" to distribute the securities to which they had committed themselves as originators or members of a purchase group. Id. at 48 n. 8.

Congress was well aware of these developments. The heavy overhead incurred by banks to carry these large retail operations caused much of the congressional concern about "subtle hazards" that animated the sponsors of the Glass-Steagall Act. As one of the principal sponsors stated:

In order to be efficient a securities department had to be developed; it had to have salesmen; and it had to have correspondent connections with smaller banks throughout the territory tributary to the great bank. Organizations were developed with great enthusiasm and efficiency. The distribution of the great security issues needed for the development of the country was facilitated, and the country developed. But the sales departments were subject to fixed expenses which could not be reduced without the danger of so disrupting the organization as to put the institution at a disadvantage in competition with rival institutions. These expenses would turn the operation very quickly from a profit to a loss if there were not sufficient originations and underwritings to keep the sales departments busy.

It was necessary in some cases to seek for customers to become makers of issues of securities when the needs of those customers for long-term money were not very pressing. Can any banker, imbued with the consciousness that his bond-sales department is, because of lack of securities for sale, losing money and at the same time losing its morale, be a fair and impartial judge as to the necessity and soundness for a new security issue which he knows he can readily distribute through channels which have been expensive to develop but which presently stand ready to absorb the proposed security issue and yield a handsome profit on the transaction?

It is easy to see why the security business was overdeveloped and why the bankers' clients and country bank correspondents were overloaded with a mass of investments many of which have proved most unfortunate.

75 Cong.Rec. 9911 (1932) (remarks of Sen. Bulkley).

The distinction between public and private offerings meshes well with the congressional goal of eliminating the "subtle hazards" of conflicts of interest and abuse of fiduciary relationships in banking. Senator Bulkley's remarks show a concern with the development of a vast selling apparatus necessary to participate in the distribution of "great security issues needed for the development of the country" and mirror the unchallenged evidence in the literature that banks in the early twentieth century were building that type of large selling organization.

In light of the specific congressional focus on the large fixed costs that accompanied retail participation in public distributions, it seems highly plausible that one line Congress might have drawn in adopting the permissive language of section 16 of the Glass-Steagall Act was at the point of public offering, a line which could well explain the prohibition against underwriting. While regular involvement in private offerings of securities undoubtedly produces some fixed costs and some attendant pressures, it seems reasonable to think that Congress might have found these relatively minor expenses acceptable when compared with the much heavier fixed burden of having a far-flung retail network to distribute securities to the public. Although implementation of this distinction through the prohibition of commercial-bank underwriting would not address all the "subtle hazards" with which Congress was concerned (for example, it would do nothing to meet the fear that a bank would sell securities for an issuer to help the issuer repay loans to the bank), the prohibition of underwriting is only one of the limitations that section 16 imposes on banks that desire to deal in securities. We believe that the distinction between public and private offerings as drawn by the Board reasonably interprets the prohibition of underwriting and reasonably relates to the elimination of some of those hazards.

While SIA has not met its burden of refuting the reasonableness of the Board's conclusion that Bankers Trust's activities fit within the literal terms of section 16, SIA mounts a final, sweeping challenge to the reasonableness of the Board's interpretation of the Act. SIA asks the court to analyze the activities approved by the Board to determine whether they pose the "subtle hazards" that the Act seeks to eliminate. The district court relied on the potential for these hazards to conclude that, while the activities of Bankers Trust come within the literal terms of section 16, those terms should be construed narrowly to exclude an otherwise permissible arrangement that frustrates the policies of the Act. In other words, although the language and history of the specific provisions support the reasonableness of the Board's construction of those provisions, the Board might nonetheless be obligated to adopt a different construction if the background policies of the Act as interpreted by the Supreme Court in cases like Camp and SIA conflict with that construction and render it unreasonable. This admittedly seems at odds with the recent statement by the Supreme Court that

[a]pplication of "broad purposes" of legislation at the expense of specific provisions ignores the complexity of the problems Congress is called upon to address and the dynamics of legislative action. Congress may be unanimous in its intent to stamp out some vague social or economic evil; however, because its Members may differ sharply on the means for effectuating that intent, the final language of the legislation may reflect hard fought compromises. Invocation of the "plain purpose" of legislation at the expense of the terms of the statute itself takes no account of the processes of compromise and, in the end, prevents the effectuation of congressional intent.

Board of Governors of the Fed. Reserve Sys. v. Dimension Fin. Corp., --- U.S. ----, 106 S. Ct. 681, 689, 88 L. Ed. 2d 691 (1986). But the Supreme Court's Glass-Steagall Act cases uniformly consider "subtle hazards" and examine background purposes of the Act. Until the Court indicates that it no longer employes this analysis to interpret the Glass-Steagall Act, we too must take such considerations into account. We therefore turn to that analysis.

We believe that the district court erred in concluding that the private placement of commercial paper by Bankers Trust creates the kind of "subtle hazards" that would require the Board to construe section 16 narrowly to exclude that activity. Since the Supreme Court has already undertaken a "subtle hazards" analysis with respect to commercial paper, albeit without the benefit of the Board's analysis of that issue, we know precisely the concerns the Court has identified in this area. What we must decide is whether the practices of Bankers Trust at issue here sufficiently differ from those in the last round of this litigation to justify concluding that the hazards identified are no longer present, or whether the Board has presented new considerations that were not before the Court in SIA and that meet the concerns expressed by the Court.

Initially, it bears noting that the most obvious hazard reached by the Act--the investment of bank funds in speculative securities--is not at issue in this case. Bankers Trust does not purchase the commercial paper of its customers; it does not inventory the paper overnight; and it makes no loans to provide financing to an issuer when an offering of paper falls short of its goal. Nothing in Bankers Trust's services puts its own resources at risk.

This takes us directly to the "subtle hazards" analysis, which catalogues the various conflicts of interest and dangers that may result from a commercial bank's dealing in "particular" securities. The first set of potential conflicts involves the bank in its role as a lender, raising the dual specter of the bank's making loans to the issuer (to ensure the success of its issue) or to the purchasers of commercial paper placed by the bank. SIA, 468 U.S. at 146-47, 156-57, 104 S. Ct. at 2985, 2990. The Board's analysis adequately answers those concerns.

To avoid any danger of making unsound loans to an issuer, Bankers Trust has, since the Supreme Court decided SIA, adopted a policy of providing no back-up credit or guarantees to facilitate the acceptance of commercial paper; any line of credit now granted to an issuer must have "substantially different timing, terms, conditions and maturities from the commercial paper being placed." Board Statement at 40, J.A. at 234. SIA does not dispute the salutary nature of this change, but argues that the Board's reliance on such representations amounts to "regulation" in a statute that Congress meant to operate through "flat prohibitions." This argument is without merit. The Glass-Steagall Act does impose a system of flat "prohibitions" and "prophylactic" measures, see SIA, 468 U.S. at 147-48, 157, 104 S. Ct. at 2985-86, 2990, but this cannot obviate the need to examine particular factual situations to determine on which side of the prohibitory line they fall. Although the Act may seek to prevent even "potential" conflicts, see ICI, 401 U.S. at 637-38, 91 S.Ct. at 1102, this does not foreclose the Board from deciding that the realities of a situation make even the potential for conflict substantially unlikely. Bankers Trust has made representations about the conduct of its loan department that seem to meet the congressional concerns identified by the Supreme Court; it is perfectly appropriate for the Board to credit the bank's new policies. Moreover, we do not believe that the Board's assumption that Bankers Trust will keep adequate records to substantiate its contentions transforms the Board's decision into an instance of "regulation." If a member of the industry were to file charges with the Board, claiming that Bankers Trust was not adhering to its stated policies, the availability of Bankers Trust's records would facilitate the Board's investigation of that charge. It is in no way an impermissible "regulation" to require Bankers Trust to keep adequate records.

The Board has also advanced an argument not considered by the Court in SIA to explain why the arrangement adopted by Bankers Trust will not lead to the lending of money to "shore up" customers of the bank's commercial paper service. The Board points out that the profit from the placement of commercial paper is small, amounting to a commission on the order of one-eighth of one percent of the total amount of the issuer's commercial paper, computed on an annualized basis. The rewards from these commissions are so small compared to the cost of the loans the bank would have to write to make an unsound issuer's paper more attractive to the market that writing such loans would not be worth the risk. Board Statement at 40-41, J.A. at 234-35. A judgment such as this, that the economic realities of the financial marketplace would preclude banks from making loans to shore up troubled issuers, is precisely the kind of exercise of delegated expertise that deserves our full deference.

The Board has also concluded that there is no appreciable risk of the bank's placing commercial paper to enable a debtor of the bank to repay its loans. The Board's opinion reasons that an issuer unable to repay bank loans will probably be unable to raise money in the commercial paper market in any case; Bankers Trust furthermore has adopted a policy of not providing letters of credit or guarantee arrangements to make such paper more attractive. Board Statement at 44, J.A. at 238. Moreover, the antifraud provisions of the securities laws would compel the disclosure of the intended use of the proceeds to satisfy a potentially bad debt owed to the bank, providing an obvious disincentive to such a transaction. Id. Finally, empirical evidence indicates that the proceeds of private placements by banks have not been used to pay off any loans involving a material risk of nonpayment. Id. at 45, J.A. at 239. The Board's findings as to these factors, which the SIA Court apparently did not consider, are reasonable and accordingly receive our deference.

As for the second "subtle hazard," the possibility of the bank's making self-interested loans to finance the purchase of commercial paper it helps issue, the Board provides a persuasive argument, again not before the Supreme Court in SIA, that no such hazard arises here. Turning again to an analysis of financial markets, the Board asserts that it is wholly impractical for a commercial bank to make such loans because the yields on commercial paper are generally lower than the interest rates the loans would have to bear. Board Statement at 41 n. 39, J.A. at 235 n. 39. In the absence of any evidence that this conclusion is wrong, the Board is again entitled to our deference.

Another category of concerns involves the bank's role as a disinterested financial advisor to its customers. First, there is the potential that the bank will give unsound financial advice to the issuer in order to reap the profits from placement of the issuer's commercial paper. See SIA, 468 U.S. at 146, 104 S.Ct. at 2985. The Board found any such risk to be insignificant because the profit from such placements is so low that the bank has no incentive to offer deliberately unsound advice. Board Statement at 46, J.A. at 240. This rationale, not considered by the Supreme Court, seems consistent with the notion that much of the concern with banks' giving self-interested advice was based on the banks' need to meet the fixed costs of far-flung distribution networks. See supra pp. 1065-1066. When the rewards and incentives are lower, the potential benefits from rendering unsound and self-interested advice seem likely to be outweighed by the damage to the bank's reputation and goodwill that would arise from giving bad advice. The Board's conclusion that bad advice will not result from the scheme at issue here is rational.

The role of disinterested financial advisor to depositors presents different concerns. Congress feared that depositors purchasing securities through their bank might lose confidence in their bank if an issuer using the bank's securities services defaulted on their securities. SIA, 468 U.S. at 155-56, 104 S.Ct. at 2989-90. Although Bankers Trust has prevented any conflict of interest concerning any account managed or advised by the bank or its affiliates or for any account in the bank's trust department by adopting a flat rule that it will purchase none of the commercial paper it places for these accounts, Board Statement at 45, J.A. at 239, Bankers Trust does otherwise place commercial paper with its depositors. The Board argues that because the depositors who purchase commercial paper are large, sophisticated business institutions, they would be unlikely to blame their bank for what really amounts to their own error in judgment, while any harm to the bank that did result would not affect its public reputation. Id. at 42-43, J.A. at 236-37. Though this assessment seems entirely realistic, the Supreme Court in SIA clearly rejected these arguments, stating that the Act makes no distinctions on the basis of financial expertise and that the loss of confidence of a few large depositors might, in fact, prove "especially severe." SIA, 468 U.S. at 156, 159, 104 S. Ct. at 2990, 2991. While the Board also argues that an empirical study has indicated no harm to the reputation of commercial banks from their private placements of securities, Board Statement at 42, J.A. at 236, nowhere does the Board's analysis indicate that the study specifically addressed the effect of issuer defaults on depositor/purchaser confidence in commercial banks.

We believe, however, that despite the existence of this one "subtle hazard," we must still affirm the Board. There are several reasons for that conclusion. First, the "subtle hazards" addressed in Camp and returned to in ICI, Schwab, and SIA have never alone caused the Supreme Court to hold that Glass-Steagall permits or prohibits any particular banking practice. Rather, analysis of the hazards in those cases simply reinforced the Court's conclusion that, as a matter of statutory interpretation, Glass-Steagall permitted or prohibited the questioned practice. Moreover, the Court has concluded that "subtle hazards" counsel prohibition of a banking practice only when the practice gave rise to each and every one of the hazards. See SIA, 468 U.S. at 154-59, 104 S.Ct. at 2989-92. Camp, 401 U.S. at 630-34, 636-38, 91 S. Ct. at 1098-1100, 1101-02. Nor must a hazard be "totally obliterated" to permit a banking practice--avoidance of the hazard "to a large extent" suffices. See ICI, 450 U.S. at 67 n. 39, 101 S. Ct. at 987 n. 39. Finally, our conclusion is reinforced by our view that the "subtle hazards" analysis as a whole is a specific instance of the Chevron principle that requires our deference to an agency's reasonable construction of its statute's ambiguities, see Investment Co. Inst. v. Conover, 790 F.2d 925, 931-33, 935-36. (D.C.Cir.1986) (applying Chevron to agency interpretation of Glass-Steagall), since an agency's statutory interpretation that impairs one of the statute's purposes but not others may surely nonetheless be reasonable. (Indeed, the binding force of the Supreme Court's "subtle hazards" analysis in SIA is unclear, since, as we have already noted, supra p. 1056, the Board failed to offer the Court any rationale concerning those hazards to which the Court could defer. See SIA, 468 U.S. at 155, 104 S.Ct. at 2989; see also ICI, 450 U.S. at 68, 101 S.Ct. at 988 (distinguishing Camp on this ground)). We think, in short, that the Board reasonably concluded not only that Bankers Trust's placements of commercial paper meet the literal requirements of section 16, but also that those placements are consistent with the panoply of the Act's purposes.

We therefore reverse the district court's order and reinstate the Board's decision.

It is so ordered.