United States v. Citizens & Southern National Bank

Mr. Justice Brennan,

with whom Mr. Justice Douglas and Mr. Justice White join, dissenting.

I agree that the District Court erred in holding that the correspondent associate programs are immune from Sherman Act scrutiny because they are subject to the “exclusive primary jurisdiction” of the Federal Reserve Board under the Bank Holding Company Act of 1956, as amended. The District Court also erred, however, in holding that the United States did not prove the violations of § 1 of the Sherman Act, and § 7 of the Clayton Act, alleged, and I therefore dissent from the affirmance of its judgment.

The issues under the Clayton and Sherman Acts, while logically independent, are related; both present the question whether a large commercial bank, already possessing *131a substantial share of the Atlanta market, may lawfully acquire other banks, rather than expand internally. Three banks now control more than 75% of the commercial banking business in Atlanta. Today’s decision assures that their dominions will soon be extended as arrangements they have made with independent banks to operate as “de facto branches” are solidified through merger. I cannot agree with today’s decision that the Government is powerless to prevent this result.

I. The Sherman Act

The “5-percent” banks in this litigation entered into a relationship with C&S far exceeding that of “correspondent banking,” the provision of check clearance, investment advice, personnel training, or other specialized services in arm’s-length transactions.1 From the very inception of these relationships, it was contemplated that *132the 5-percent banks would seek, and C&S would provide, advice and guidance with respect to virtually every business decision of significance. C&S provided advisory directors — treated by all parties as actual directors— made available operating manuals covering banking practices in minute detail,2 and maintained a constant flow of bulletins whose contents ranged from admonitions about the antitrust laws to exhortations to “get the rates [on loans] up.” C&S, through its Branch Supervision Department, monitored the performance of the management of the 5-percent banks and was instrumental in having replaced those who did not measure up. These arrangements had the desired effect. The elaborate fabric of “consultations,” of seeking “advice and guidance,” eliminated the opportunity for rivalry among the defendant banks. The District Court found “no presently existing substantial competition between the five-percent banks and C&S National, or inter sese.” 372 F. Supp. 616, 642 (1974).

A

The Court concludes that antitrust scrutiny of the affiliation of three 5-percent banks is foreclosed by the grandfather provision of § 11 (d) of the Bank Holding Company Act, 12 U. S. C. § 1849 (°d). That holding is plainly a distorted expansion of § 11 (d) beyond its language and purpose.

The concept of an amnesty for unchallenged structural arrangements in commercial banking first appeared *133in the 1966 amendments to the Bank Merger Act, 80 Stat. 7. In those amendments, Congress, responding in part to this Court’s decisions in United States v. Philadelphia National Bank, 374 U. S. 321 (1963), and United States v. First National Bank & Trust Co. of Lexington, 376 U. S. 665 (1964), attempted to mesh antitrust considerations with review of proposed bank mergers by the appropriate regulatory agency. The resulting provisions, which mandate Justice Department participation in the regulatory approval process as well as consideration by the regulatory agencies of “competitive factors,” and permit an antitrust suit within 30 days of regulatory approval, appear today in the Federal Deposit Insurance Act, 12 U. S. C. § 1828. See United States v. First City National Bank of Houston, 386 U. S. 361 (1967); United States v. Third National Bank in Nashville, 390 U. S. 171 (1968). The 1966 amendments also included a grandfather provision, 80 Stat. 10, that conferred immunity from antitrust challenge (except under § 2 of the Sherman Act) upon any “merger, consolidation, acquisition of assets, or assumption of liabilities” consummated before June 17, 1963, the date of the decision in Philadelphia National Bank.

A few months after enactment of the Bank Merger Act amendments, the “antitrust” provisions were written almost verbatim into the Bank Holding Company Act. Unlike their Merger Act counterparts, the 1966 amendments to the Bank Holding Company Act were not principally addressed to integrating antitrust standards with the regulatory process, but rather to expanding the Federal Reserve Board’s jurisdiction and regulatory powers. The antitrust provisions of the Holding Company Act amendments received little legislative attention; the brief reference to them in the legislative history indicates that their pur*134pose was to “apply to bank holding company cases the same procedures as are now provided in bank merger cases . 3 Among the provisions so borrowed from the earlier Bank Merger Act amendments was the grandfather provision, § 11 (d).

Because of congressional preoccupation with the regulatory features of the 1966 amendments to the Bank Holding Company Act, interpretation of the antitrust provisions may involve as much an attribution of congressional intent as a discernment of it. This is particularly the case with respect to § 11 (d), which was transplanted from one regulatory statute to another with seemingly scant attention to the differences in the regulatory environment. Objections that grandfathering holding company acquisitions posed policy questions different from the retroactive immunization of mergers were quickly brushed aside,4 and § 11 (d) was swept into law along with the other antitrust provisions. Thus, despite whatever dissimilarity of underlying policy considerations may have been exposed, Congress indicated that it considered the grandfather provisions in both statutes to advance substantially similar purposes. Accordingly, however difficult may be the discernment of the congressional intent expressed in § 11 (d), we must look for assistance to its counterpart in the Bank Merger Act, the only guidepost Congress has left us.

The grandfather provision of the Bank Merger Act amendments most assuredly did not provide sanctuary *135for then-unchallenged price-fixing, market-division, or other cartel activity by banks. Congressional concern was much more narrowly directed. Philadelphia National Bank rejected a literal interpretation of § 7 of the Clayton Act that would have limited its application to stock acquisitions by banks, an interpretation that nevertheless enjoyed some acceptance prior to the decision. Congress was concerned about the difficulty of unscrambling prePhiladelphia National Bank mergers undertaken in reliance upon the literal interpretation of § 7, which the Court ultimately rejected, and accordingly immunized them from suit under that section.5 But a provision barring suit under § 1 of the Sherman Act was also necessary to safeguard the same mergers because of our decision in Lexington Bank, supra. Thus, although the resulting grandfather provision covered both the Clayton and Sherman Acts (except Sherman Act § 2), its purpose was to shield structural arrangements of the sort the Government challenged in Philadelphia National Bank and was continuing to challenge in the District Courts thereafter.6

Against the foregoing background, we confront the language of the counterpart in the Bank Holding Company Act. As enacted in 1966, § 11 (d) shielded an “acquisition, merger, or consolidation of the kind described in § 3 (a) of this Act.” Section 3 (a) provided then, as today, that:

“(a) It shall be unlawful, except with the prior *136approval of the Board, (1) for any action to be taken that causes any company to become a bank holding company; (2) for any action to be taken that causes a bank to become a subsidiary of a bank holding company; (3) for any bank holding company to acquire direct or indirect ownership or control of any voting shares of any bank if, after such acquisition, such company will directly or indirectly own or control more than 5 per centum of the voting shares of such bank; (4) for any bank holding company or subsidiary thereof, other than a bank, to acquire all or substantially all of the assets of .a bank; or (5) for any bank holding company to merge or consolidate with any other bank holding company.” 7

Section 3 (a) is thus the operative provision of the statute permitting the Federal Reserve Board to regulate the events therein described.

By “grandfathering” an “acquisition, merger, or consolidation of the kind described in § 3 (a),” Congress obviously exempted from antitrust challenge only the events for which Board approval would have been required. None of the transactions defined by § 3 (a), however, includes those features of the “correspondent associate” relationship that the Government is challenging under Sherman Act § 1 in this case. Clauses (4) and (5) of § 3 (a) refer, respectively, to an acquisition of assets and a merger of two holding companies. Clause (3) refers to ownership of voting stock by a holding company; the stock ownership by C&S is not, however, the salient feature of the affiliative relationship and indeed is not challenged in this case. Clauses (1) and (2) address the creation of a holding company-subsidiary rela*137tionship. The definitional provisions of § 2 (d) have undergone recent expansion, but in 1966 they designated a bank as a “subsidiary” if a holding company either (1) directly or indirectly owned or controlled 25% or more of its voting stock, or (2) controlled in any manner the election of a majority of its directors. These two conditions would often be satisfied simultaneously, and indeed shortly after enactment of the forerunner of this provision in 1956 it was suggested that the second condition was redundant. See Note, The Bank Holding Company Act of 1956, 9 Stan. L. Rev. 333, 337, and n. 59 (1957). Congress, however, was apparently concerned that stock interests could be so structured that a holding company could elect a majority of directors without satisfying the 25% ownership requirement.8 Whether or not this fear was well-founded, it is clear that satisfaction of either condition required an arrangement whereby the holding company had the power to vote stock.

In establishing its “correspondent associates” C&S did not engage in the transactions described by § 3 (a) in 1966 and therefore sheltered by § 11 (d). Indeed, because of state-law restrictions C&S could not resort to the methods described by § 3 (a) of the Holding Company Act and turned instead to more informal arrangements, including “understandings.” While the functional equivalent of a holding company-subsidiary relationship could perhaps be created through informal affiliation, § 3 (a), at least until quite recently, has been *138triggered by the formality of control of voting stock. To be sure, § 2 (d) has always referred to a subsidiary as one whose stock is “directly or indirectly” owned or controlled or whose election of directors is controlled “in any manner” by the holding company.9 But there has been no suggestion by Congress, nor by the Board, that this language would embrace the less formal arrangements by which the C&S banks operated in complete harmony with C&S. Indeed, the statutory clues suggest the contrary, that Congress was concerned with powers attached to stock, and that “indirect” ownership or control merely referred to their exercise derivatively, through an intermediary.10

*139In the 1970 amendments to the Holding Company-Act, 84 Stat. 1760, Congress expanded the reach of § 3. The Act now defines “control” to include a relationship whereby a company “directly or indirectly exercises a controlling influence over the management and policies of the bank . . . § 2 (a) (2) (C), 12 ü. S. C. § 1841 (a) (2)(C). Congressional preoccupation with stock is still evident since there is a statutory presumption that “any company which directly or indirectly owns, controls, or has power to vote less than 5 per centum of any class of voting securities of a given bank or company does not have control over that bank or company." §2 (a)(3). Nevertheless the Board has by regulation established a rebuttable presumption of control where a company

“enters into any agreement or understanding with a bank . . . such as a management contract, pursuant to which the company or any of its subsidiaries exercises significant influence with respect to the general management or overall operations of the bank____” 12 CFR § 225.2 (b)(3) (1975).

Arguably, the Board’s interpretation would now bring within § 3 the affiliation of the 5-percent banks with C&S. But the Board’s interpretation is based upon recent legislation expanding the reach of the Board’s regulatory authority.11 Since I do not suppose Congress intended in 1966 to immunize transactions of the kind it had not yet brought within § 3, the 1970 amendment is relevant only because it demonstrates the limited char*140acter of the transactions previously embraced by § 3 and “grandfathered” under § 11 (d).

The conclusion that Congress had traditionally not brought informal arrangements within § 3 (a) was reinforced by the provisions of § 4 (a) (2) of the original Act, 70 Stat. 135, which forbade a bank holding company to

“engage in any business other than that of banking or of managing or controlling banks or of furnishing services to or performing services for any bank of which it owns or controls 25 per centum or more of the voting shares.”

This provision was enacted in 1956, and as early as 1960 the Board by regulation interpreted “services” to include many of the functions C&S has performed for the 5-percent banks. Included in the Board’s interpretation are: “(1) [establishment and supervision of loaning policies; (2) direction of the purchase and sale of investment securities; (3) selection and training of officer personnel; (4) establishment and enforcement of operating policies; and (5) general supervision over all policies and practices.” 12 CFR §225.113 (1975). The differentiation of these activities from “control or management” and their inclusion in § 4 of the Act rather than in § 3 vividly exposes the fallacy of today’s holding invoking § 11 (d) to foreclose scrutiny of the “correspondent associate” relationship of three of the 5-percent banks. Since § 11 (d) shielded only the events then described in § 3 (a), the conclusion is compelled that all the 5-percent banks are properly before us on the Sherman Act counts.12 Accordingly, I turn to the merits.

*141B

The District Court found that there were no express agreements among the defendant banks to fix prices or divide markets that would call for application of the per se rule, United States v. Socony-Vacuum Oil Co., 310 U. S. 150 (1940); United States v. Sealy, Inc., 388 U. S. 350 (1967), but it also found that the effect of the association was to eliminate all competition among the banks involved.

The Court finds the restraints embodied in the “correspondent associate” relationship reasonable because of state-law restrictions that blocked, for a time, the avenue of internal expansion by C&S. If the question before us were the lawfulness of these arrangements at their inception, this solution might be satisfactory. The question would be a close one, however, calling for a delicate balancing of the immediate benefits of expanded banking services against the more distant, but nevertheless real, danger of permitting the restraints necessary to circumvent de jure barriers to expansion to continue longer than the conditions that justified them. The inquiry would, of course, have to take into account the possibility that expansion would occur under less restrictive conditions. New entry by an unaffiliated bank13 or entry *142with a more limited form of sponsorship — a period of initial assistance, followed by a withdrawal of the sponsor's influence, at least to a conventional correspondent relationship14 — might have sufficed to provide the expansion cited here as a justification for incidental restraints. The judicial resources consumed by such an inquiry in any particular case would not be insubstantial, and the very difficulty of making such judgments has in many cases led us to prefer per se rules. United States v. Socony-Vacuum Co., supra, at 220-221; Northern Pacific R. Co. v. United States, 356 U. S. 1, 5 (1958); United States v. Sealy, Inc., supra; United States v. Topco Associates, Inc., 405 U. S. 596 (1972). See also United States v. Philadelphia National Bank, 374 U. S., at 362.

The issue in this case, however, is not whether the affiliation of the 5-percent banks was lawful at its inception, but whether it could lawfully continue, for the Government sought only an injunction. By the time the Government brought suit, Georgia law per*143mitted C&S to branch freely in the Atlanta suburbs. Because the rule of reason requires us to assess the lawfulness of a restraint in light of all the circumstances, Chicago Board of Trade v. United States, 246 U. S. 231, 238 (1918), the lawfulness of the practices at their inception, even if assumed, could not be controlling, for changes in market conditions can deprive once-reasonable arrangements of their justification. United States v. Jerrold Electronics, 187 F. Supp. 545, 560-561 (ED Pa. 1960), aff'd, 365 U. S. 567 (1961). See also United States v. E. I. du Pont de Nemours & Co., 353 U. S. 586, 596-598 (1957). The claimed desirability of the challenged arrangements as a response to now-repealed restrictions of Georgia law is therefore relevant only insofar as it may also be claimed that continuation of such arrangements undisturbed by the Sherman Act would be vital to their creation were Georgia to reinstate its restrictions in the future. Put another way, we need concern ourselves with the lawfulness of “de facto branches” as a response to state-law restrictions only if appellees make a convincing showing that no bank would engage in “de facto branching” without a guarantee of perpetual noninterference from the antitrust laws.

Certainly it is open to C&S to argue that no rational banker would sponsor a de facto branch unless assured that the resulting relationships could continue in perpetuity. But this sort of argument has seldom carried the day in this Court, see United States v. Sealy, supra; United States v. Topco Associates, supra, and I do not find it persuasive in this case. A bank hemmed in by state antibranching restrictions will presumably find it profitable to take a small stock interest in an independent bank, to offer assistance and thereby attempt to win consumer loyalty through an expanded use of its own name. C&S presumably found these arrangements *144profitable at their inception. The record does not show whether C&S actually charged the 5-percent banks for such assistance as site selection, economic surveys, equipment procurement, and other promotional services; there is no suggestion, however, that C&S provided these services at an ultimate loss, and presumably gains ultimately accrued to the provider. True, C&S hoped to cement the relationships through merger, but it is not clear that these expectations were essential to the initial undertaking. Indeed, C&S continued to provide assistance to certain banks as to which there was little prospect of ultimate acquisition by C&S. 2 App. 378-379. Our concern, in any event, lies not with protecting the expectations of C&S but with avoiding disincentives to the provision of desirable services. Sponsorship will be profitable to a sponsor bank assuming that there is a demand for the services of the sponsored bank and that the sponsor can recoup in some fashion a return for its assistance. These conditions should be sufficient to induce a profit-seeking bank, chafing under antibranching restrictions, to sponsor a new entrant even if permanent arrangements are forbidden.

This case, therefore, does not present an occasion for consideration whether the restraints incident to “de facto branching” are lawful when undertaken in response to a prohibition of de jure branching, a position the Court says the Government took last Term in United States v. Marine Bancorporation, 418 U. S. 602 (1974). The restraints incident to the affiliation of the 5-percent banks with C&S must be examined in light of conditions prevailing at the time of suit, which include the ability of C&S to branch freely in the Atlanta suburbs.

The arrangements between C&S and the 5-percent banks resemble a “common brand” marketing agreement or a franchising arrangement in which the franchisor *145itself deals directly with consumers as well as providing entrepreneurial skill and other assistance to franchisees. Such combinations may, under certain circumstances, enhance competition. Common-brand marketing may permit a group of small firms to exploit promotional economies and thereby compete with larger enterprises whose business spans several geographic sub-markets. Franchising may facilitate entry by allowing an entering firm to save on promotional expenses and to purchase needed entrepreneurial assistance. Restraints invariably accompany these combinations for the purpose of promoting product uniformity, for some standardization of product is indispensable to the success of the scheme. Because notwithstanding accompanying restraints such combinations may on balance enhance competition, it would be a mistake to regard them as per se or even presumptively unlawful, and lower courts have not done so. See, e. g., United States v. Topco Associates, Inc., 319 F. Supp. 1031, 1038 (ND Ill. 1970), rev’d on other grounds, 405 U. S. 596 (1972); Siegel v. Chicken Delight, Inc., 448 F. 2d 43 (CA9 1971); Susser v. Carvel Corp., 332 F. 2d 505 (CA2 1964). But the Sherman Act limits the scope of cooperation incident to such arrangements. The participants may not fix prices or divide markets. United States v. Topco Associates, Inc., supra; United States v. Sealy, Inc., supra; United States v. Arnold, Schwinn & Co., 388 U. S. 365 (1967). Such combinations, moreover, warrant careful scrutiny when their participants collectively possess a dominant share of a common market, as to which there are substantial barriers to entry, for these conditions enhance the profitability of price collusion among participants and thus may tempt them to standarize price as well as other product attributes.

Despite the acceptability generally of common-brand *146or franchising arrangements, they pose particular difficulty in the commercial banking context. Many features of a commercial bank’s services are set by regulalation, thus inhibiting competition by restricting the number of product features that individual firms are free to vary. With interest rates on loans fixed by law, for example, competition is confined to such “non-price” features as collateral requirements or repayment policies. With competition thus already delimited, few additional restraints incident to a cooperative scheme can be tolerated before competition is extinguished entirely. Moreover, the entry barriers posed by regulation enhance the danger that incidental cooperation will be extended to abolish all rivalry. These considerations suggest that cooperative arrangements in commercial banking should be permitted only where their competitive benefits are clear, and where the combined market shares of the participants dispel the fear that price collusion will accompany them.

The situation here fails to satisfy the test. The combined shares of C&S and the 5-percent banks are substantial under any of the alternative definitions of the geographic market cited by the Court. Ante, at 122-130.15 Furthermore, the cooperative arrangements in*147volve not a group of small firms allied to challenge a larger rival, United States v. Topeo Associates, supra, but instead the dominant firm which thereby extends its hegemony. In a market so concentrated as is commercial banking in Atlanta, the most must be made of opportunity for rivalry among existing firms. Cf. United States v. Philadelphia National Bank, 374 U. S., at 372. The 5-percent banks are now substantial, thriving enterprises,16 inhibited from competing with C&S only by the “correspondent associate” relationship. I would hold that the Government is entitled to an injunction, specifically against the continued use by the 5-percent banks of the C&S name, the continued use of advisory directors furnished by C&S, and continued “consultations” between the management of the 5-percent banks and C&S, including the flow of memoranda for “advice and guidance.”

II. The Clayton Act

The Court concedes that under our prior decisions the Government has established a prima facie case under § 7. Ante, at 120. But the Court affirms the District Court’s determination that the acquisitions add nothing *148of anticompetitive significance to the pre-existing “correspondent associate” relationship. Since I have concluded that the relationship itself violates the Sherman Act, I also disagree with the Court’s affirmance of the District Court on the Clayton Act issue. Since, in my view, appellees can no longer rely upon the affiliation to rebut the Government’s prima facie case, I would remand to the District Court for consideration of the “convenience and needs” defense of 12 U. S. C. § 1828 (c)(5)(B). But I also disagree with the Court’s conclusion that the acquisitions add nothing of significance to the existing arrangements, and I would therefore reverse even if I accepted the Court’s disposition of the Sherman Act counts. I state briefly my reasons for so concluding.

If not acquired, the 5-percent banks have the power to break their ties with C&S, and the likelihood that any would do so may be expected to increase as the demand for their services grows and as their managements acquire additional business experience. However risky these ventures may have been at their inception, the recent performance of the 5-percent banks attests to their present viability.17 Because of the continuing population growth of the Atlanta area, the banks may anticipate an expanding demand for their services. These circumstances might well induce the management of a 5-per-cent bank to assume a more independent posture, at least to shop around among other large Atlanta banks for more conventional “correspondent” services.18

*149Quite apart from what the managements of the 5-per-cent banks might do, it is most improbable that C&S would long be happy with existing arrangements if acquisition were enjoined. The record demonstrates the aggressive, expansionist performance of C&S, having increased its Atlanta offices from three in 1946 to more than 100 by the time of trial. It is quite inconceivable that such a firm would long be content to continue operations through de facto branches in which its interest was limited to 5%. The formation of de jure branches, ultimately in competition with former “correspondent associates,” would be a plausible result.

The foregoing are not “ephemeral possibilities,” Brown Shoe Co. v. United States, 370 U. S. 294, 323 (1962), that antitrust analysis should ignore. Section 7 was intended, as we have repeatedly said, to “arrest anticompetitive tendencies in their fincipiency.’ ” United States v. Philadelphia National Bank, 374 U. S., at 362. In applying the § 7 standards, we are obliged to hold acquisitions unlawful if a reasonable likelihood of a substantial lessening of competition under future conditions is discernible. E. g., United States v. Continental Can Co., 378 U. S. 441, 458 (1964); FTC v. Procter & Gamble Co., 386 U. S. 568, 577 (1967); United States v. Falstaff Brewing Corp., 410 U. S. 526, 539 (1973) (Douglas, J., concurring in part). While inquiry as to future market conditions and performance inevitably involves speculation, fidelity to the *150congressional purpose requires us to resolve reasonable doubts in favor of the preservation of independent entities. This is perforce true where, as here, the market is highly concentrated and the acquiring firm is the dominant one.

My Brother White reminded us in his dissent last Term in United States v. Marine Bancorporation, 418 U. S., at 653:

“In the last analysis, one’s view of this case, and the rules one devises for assessing whether this merger should be barred, turns on the policy of § 7 of the Clayton Act to bar mergers which may contribute to further concentration in the structure of American business. . . . The dangers of concentration are particularly acute in the banking business, since ‘if the costs of banking services and credit are allowed to become excessive by the absence of competitive pressures, virtually all costs, in our credit economy, will be affected. . . .’ ” (Citations omitted.)

Today’s decision permits C&S, the dominant commercial bank in Atlanta, further to entrench its position. Two other rivals, which together with C&S control more than 75% of the banking business in Atlanta, may now be expected to follow suit, acquiring their own “de facto branches.”19 I believe these developments exemplify the “further concentration in the structure of American business” that § 7 was designed to prevent. Accordingly, I would reverse the judgment of the District Court.

Relationships labeled “correspondent banking” may call for careful scrutiny as the sale of specialized services by the corresponding bank shades into “consultation” by the correspondent on every business decision of significance. Correspondent banking, like other intra-industry interaction among firms or their top management, provides an opportunity both for the kind of education and sharing of expertise that ultimately enhances consumer welfare and for “understandings” that inhibit, if not foreclose, the rivalry that antitrust laws seek to promote. As one commentator on commercial banking practices has observed:

“ [C] ommunication, especially when it comes from those at the top of a power hierarchy, tends to facilitate conflict resolution. Perhaps a great deal should not be made of this, but competition is a form of conflict and, in the present context, conflict resolution is a form of restraint on competition.” Phillips, Competition, Confusion, and Commercial Banking, 19 J. of Finance 32, 42 (1964).
Since the relationship of C&S to the 5-percent banks goes well beyond ordinary “correspondent banking,” this case does not present an occasion for further examination of the lawfulness of these more limited interconnections among firms.

The Consumer Credit Operating Bulletin, 7 App. E-1024 (DX-311), is illustrative. It explains what bank records should be established, the methods for arranging a repayment plan, and the procedures to be followed in perfecting a security interest. In addition, the manual sets forth C&S practice with respect to charges for late payments, extensions of repayment deadlines, and the notification of a borrower’s employer about repayment delinquency.

As initially enacted by the House, the amendments contained no antitrust provisions. See generally H. R. Rep. No. 534, 89th Cong., 1st Sess. (1965). These were added later by the Senate Banking and Currency Committee and subsequently adopted by both Houses. See S. Rep. No. 1179, 89th Cong., 2d Sess., 10 (1966).

See letter from Deputy Attorney General Clark to Sen. Robertson, reprinted at 112 Cong. Rec. 12385 (1966), and accompanying remarks by Sen. Robertson, ibid.

See S. Rep. No. 299, 89th Cong., 1st Sess., 1-7 (1965); H. R. Rep. No. 1221, 89th Cong., 2d Sess., 4 (1966); 111 Cong. Rec. 13304-13305 (1965) (remarks of Sen. Robertson); 112 Cong. Rec. 2454 (1966) (remarks of Rep. Celler).

See United States v. Crocker-Anglo National Bank, 223 F. Supp. 849 (ND Cal. 1963); United States v. Manufacturers Hanover Trust Co., 240 F. Supp. 867 (SDNY 1965), cited in Hearings on S. 1698 before a Subcommittee of the Senate Committee on Banking and Currency, 89th Cong., 1st Sess., 446, 463 (1965).

Section 3 (a) had been in force since enactment of the Bank Holding Company Act in 1956. The 1966 amendment added clause (2) to its provisions.

See H. R. Rep. No. 609, 84th Cong., 1st Sess., 12-13 (1955); 101 Cong. Rec. 8028 (1955) (remarks of Rep. Patman). In the form initially adopted by the House, the Act would have defined as a subsidiary a bank over which another company was found by the Federal Reserve Board to “exercise a controlling influence.” The Senate amendment substituted the provision ultimately enacted, the requirement of control of the election of directors. See S. Rep. No. 1095, 84th Cong., 1st Sess., 5 (1955).

The reference to indirect ownership, though contained in § 2 (a) of the 1956 Act (defining holding company), was inadvertently omitted from § 2 (d). See 70 Stat. 134. The 1966 amendments corrected the omission. See S. Rep. No. 1179, 89th Cong., 2d Sess., 8 (1966).

Section 2 (g) of the Act defined indirect control or ownership:

“For the purposes of this Act—
“(1) shares owned or controlled by any subsidiary of a bank holding company shall be deemed to be indirectly owned or controlled by such bank holding company;
“(2) shares held or controlled directly or indirectly by trustees for the benefit of (A) a company, (B) the shareholders or members of a company, or (C) the employees (whether exclusively or not) of a company, shall, be deemed to be controlled by such company; and
“(3) shares transferred after January 1, 1966, by any bank holding company (or by any company which, but for such transfer, would be a bank holding company) directly or indirectly to any transferee that is indebted to the transferor, or has one or more officers, directors, trustees, or beneficiaries in common with or subject to control by the transferor, shall be deemed to be indirectly owned or controlled by the transferor unless the Board, after opportunity for hearing, determines that the transferor is not in fact capable of controlling the transferee.”
This provision was added by the 1966 amendments to adopt inter*139pretations previously made by the Board. S. Rep. No. 1179, supra, at 8.

Congress specifically noted the expansion. See S. Rep. No. 91-1084, p. 6 (1970);. H. R. Rep. No. 91-1747, p. 12 (1970). See also Note, The Bank Holding Company Act Amendments of 1970, 39 Geo. Wash. L. Rev. 1200, 1213-1214 (1971).

My conclusion that the affiliative relationships are not within the terms of § 3 (a), at least prior to the 1970 amendment, is further supported by the scope and outcome of the 1968 investigation of C&S undertaken by the Federal Reserve Board staff. The investigation was convened specifically to inquire into a possible violation *141of § 3. The staff was principally concerned with the pattern of ownership of the stock of the 5-percent banks, especially by C&S officers and employees. Ultimately the staff found this acceptable, so long as C&S did not finance the purchases. There is no indication, however, that the staff concerned itself with communications between C&S and the 5-percent banks with respect to such matters as interest rates, loan repayment policies, or other terms of business.

There is little doubt that pent-up consumer demand for additional banks would sooner or later induce efforts to organize new ones. More questionable, however, is whether regulatory authorities would respond promptly to permit new entry. In general, reg*142ulatory policy has been thought to retard formation of new banking institutions. See Peltzman, Entry in Commercial Banking, 8 J. Law & Econ. 11 (1965).

The record demonstrates that such a chain of events is possible. Citizens & Southern Bank of Stone Mountain, organized in 1957 with C&S assistance, functioned as a correspondent associate from 1959 until 1970. At that time it declined an offer of acquisition by C&S and became independent of the C&S system. Appellees have argued that Stone Mountain represents a unique case because a majority of voting stock remained in the hands of a single family not intimately tied to the C&S system. This contention is not wholly supported by the record, since in his trial testimony Mr. Mills Lane, President of C&S from 1946 to 1970, referred to three other banks having a similar structure of ownership. (2 App. 378-379, referring to Pelham, Fayetteville, Hogansville). The example of Stone Mountain does, in any event, demonstrate that sponsorship can occur under conditions ultimately leading to independence of the sponsored institution.

The District Court made no finding as to the relevant geographic market, accepting the Government’s contentions arguendo in deciding the case. The Court apparently does the same. A report prepared by the Government’s expert witness concluded that while the Atlanta Standard Metropolitan Statistical Area was too large to be considered an integral geographic market, the constituent counties of DeKalb and Fulton were “reasonable geographic areas within which it is appropriate to analyze the competitive effects of the proposed mergers.” 4 App. E-83. This is an approximation, of course, since the same report revealed that a number of DeKalb residents use Fulton County banks, thus suggesting that in certain respects DeKalb and Fulton County banks compete for the same business. Accordingly, it appears that defining the geographic mar*147ket to include both DeKalb and Fulton Counties would be justified under our cases. See United States v. Phillipsburg National Bank & Trust Co., 399 U. S. 350 (1970); United States v. Philadelphia National Bank, 374 U. S. 321 (1963).

Three of the 5-pereent banks — Sandy Springs, Chamblee, and Tucker — had deposits exceeding $15 million as of January 1, 1970. North Fulton, Park National, and South DeKalb were smaller and more recently organized, but all have experienced vigorous growth. The average annual rate of deposit growth for the two years preceding January 1, 1970, was 102% for North Fulton and 50% for Park National, in contrast to a national average rate for all commercial bank deposits during the same period of slightly more than 10%. South DeKalb, organized in late 1969, had more than doubled its deposits from $1.5 to $3 million during the first half of 1970. 5 App. E-422, E-546.

See n. 16, supra.

Officers of both C&S and the 5-percent banks testified that they had not contemplated a severance of relations, but this testimony does not establish what would happen if the acquisitions were enjoined. Had the managements testified that they would not consider severance under any circumstances, such declarations of an intention to eschew a course dictated by economic self-interest would have to be viewed with skepticism. See United States v. Falstaff *149Brewing Corp., 410 U. S. 526, 568-570 (1973) (Marshall, J., concurring in result).

Whether the 5-percent banks would have been formed at all had their principals expected the Clayton Act to bar ultimate acquisition by C&S is a different question. I am not troubled by it for essentially the same reasons that have led me to conclude above that enjoining continuation of correspondent associate relationships would not deter sponsorship of de facto branches under state-law restrictions on de jure branching. See supra, at 143-144.

The record indicates that at the time C&S applied for regulatory approval of the acquisitions, its two largest competitors, First National Bank of Atlanta and Trust Company of Georgia, had sought and in some cases had obtained, approval for similar acquisitions of affiliated banks. 1 App. E-39.