Petitioner Republic of Texas Corporation ("Republic") is a multibank holding company under the Bank Holding Company Act of 1956, as amended, 12 U.S.C.A. §§ 1841-1850 (West 1980). By order dated August 20, 1980, the Board of Governors of the Federal Reserve System ("the Board") denied Republic's application under section 3(a) of the Act, 12 U.S.C. § 1842(a) (Supp. III 1979), for advance approval of Republic's proposed acquisition of The Citizens National Bank of Waco, a commercial bank located in Waco, Texas ("CNB"). Republic of Texas Corporation, 66 Fed.Res.Bull. 787 (1980). Pursuant to section 9 of the Act, 12 U.S.C. § 1848 (1976), Republic petitions this court for review of the Board's order.

Republic first argues that because the Board failed to comply with the so-called "91-day rule" created by section 3(b) of the Act, 12 U.S.C. § 1842(b) (Supp. III 1979), Republic's application must be deemed to have been granted by operation of law. The proper operation of the 91-day rule is a question of first impression in this circuit. After examining the statute and its legislative history, the related regulations promulgated by the Board, and the decisions of the other circuits that have considered the 91-day rule, we conclude below in part III of this opinion that Republic's application cannot be deemed to have been granted by operation of law.

In the alternative, Republic argues that the record does not contain substantial evidence to support the Board's rejection of Republic's application on anticompetitive grounds. This issue is directly controlled by our recent holding in Mercantile Texas Corp. v. Board of Governors of the Federal Reserve System, 638 F.2d 1255 (5th Cir. 1981). Following the teachings of that case, which we discuss below in part IV of this opinion, we vacate the Board's order and remand Republic's application to the Board for reconsideration and more thorough findings.

I. THE STATUTORY AND REGULATORY FRAMEWORK FOR THE PROCESSING OF APPLICATIONS UNDER SECTION 3 OF THE BANK HOLDING COMPANY ACT

Section 3(a) of the Bank Holding Company Act of 1956, as amended, 12 U.S.C. § 1842(a) (Supp. III 1979), makes it unlawful for any bank holding company to acquire control of more than 5% of any class of voting shares of any bank, unless the bank holding company has obtained prior approval of the acquisition from the Board.1 Section 3(b) of the Act, 12 U.S.C. § 1842(b) (Supp. III 1979), requires that immediately upon receiving an application for Board approval pursuant to section 3(a), the Board must notify either the Comptroller of the Currency or the appropriate state authority of the application.2 The statute explicitly directs that the Comptroller or the state authority shall have 30 days in which to submit his views and recommendations on the proposed acquisition. If the Comptroller or state authority opposes the acquisition, the Board must hold a hearing. The statute explicitly sets out an expeditious timetable within which such a hearing must take place.3

Section 3(b) goes on to impose a deadline on the Board's deliberations upon an application:

In the event of the failure of the Board to act on any application for approval under this section within the ninety-one-day period which begins on the date of submission to the Board of the complete record on that application, the application shall be deemed to have been granted.

12 U.S.C. § 1842(b) (Supp. III 1979) (emphasis added). It is the proper application of this provision, hereinafter referred to as section 3(b)'s "91-day rule," that we must decide in the case at bar.

The Board's implementing regulations require that a bank holding company's application for Board approval be filed initially with the Federal Reserve Bank of the district in which the head office of the bank holding company is located, rather than directly with the Board.4 The Reserve Bank investigates and prepares a report on the relevant facts; the report is submitted to the Board along with the Reserve Bank's recommendations on the application.5 In practice, the Reserve Banks usually cooperate with applicants to "clean up" applications before they are sent to the Board.6 Upon receipt of the application and the Reserve Bank's report and recommendations, the Board's staff considers all relevant matter and submits comments to the Board.7 The Board then "takes such action as it deems appropriate in the public interest."8

The Board's regulations recognize, but do not elaborate upon, section 3(b)'s requirement that the Board notify either the Comptroller of the Currency or the appropriate state banking authority "(u)pon receiving" a section 3 application. One regulation provides that "(t)he Board will consider comments on an application from a bank supervisory authority to which notification of receipt of an application has been given, only if such comment is received by the Secretary of the Board within 30 days of the date of the letter giving such notification."9

Section 3 of the Act does not explicitly mandate the provision of notice to, and an opportunity for comment from, the public at large. Nonetheless, in determining whether to approve an application under section 3, one factor that the Board must consider is whether any anticompetitive effects of the proposed acquisition would be "clearly outweighed in the public interest by the probable effect of the transaction in meeting the convenience and needs of the community to be served." 12 U.S.C.A. § 1842(c) (West 1980).10 Accordingly, Congress must have contemplated that the Board might find it appropriate to solicit public comment on a proposed acquisition. Unfortunately, the Board's regulations relating to public notice are, charitably speaking, somewhat less than clear. The regulations apparently envision two distinct types of public notice the first through newspaper publication, and the second through publication in the Federal Register.

The timing and content of the required notice is clearly detailed only as to the former type i. e., the newspaper publication. The Board's regulations require that a bank holding company which seeks to acquire control of a bank must, during the two weeks before filing with the Reserve Bank its application for Board approval, publish notice of its intentions in newspapers of general circulation in both the community in which the bank holding company has its head offices and the community in which the bank to be acquired is located. The notice must contain the name of the bank holding company, the subject matter of its application, the location at which the bank holding company proposes to engage in business, and an invitation to the public to submit written comments relating to the application to the appropriate Reserve Bank no later than 30 days after the date of publication of the first notice.11

The Board's regulations relating to publication in the Federal Register are considerably less clear, and correspondingly more troublesome. The primary regulation relating to Federal Register publication provides simply that "(n) otice of receipt of all section 3 applications is published in the Federal Register."12 That regulation does not specifically require that this notice be published within any particular period of time after receipt of the application, and it does not distinguish between receipt of the application by a Reserve Bank and receipt of the "cleaned up" application by the Board. Neither does it explicitly provide for a public comment period much less a public comment period of any particular length. While another regulation, which purports to detail the Board's policies on considering public comments and requests for hearings, clearly envisions the receipt of public comments following the Federal Register publication, that regulation cannot fairly be said to evidence a Board policy on the length of such comment periods.13

II. FACTS RELATING TO THE PROCESSING OF REPUBLIC'S APPLICATION TO ACQUIRE CNB

As required by the Board regulations discussed above, Republic published notice of its proposed acquisition of CNB in newspapers of general circulation in both Dallas and Waco. The notices appeared in the Dallas Times Herald on January 23 and January 30, 1980, and in the Waco Tribune-Herald on January 25 and February 1, 1980. The notices provided in part that "(p)ersons wishing to comment on this proposal should submit their views in writing within 30 days of (the) date of (this) notice to the Federal Reserve Bank of Dallas," and gave a mailing address for the Reserve Bank.

On February 15, 1980, Republic filed with the Reserve Bank its application to acquire CNB. During the following months, the Reserve Bank requested and received from Republic supplemental information relating to the merits of the application. This information was incorporated into the record on Republic's application. The Reserve Bank received Republic's last relevant submission on April 24, 1980.

On May 1, 1980, the Reserve Bank notified Republic by letter that its application was being forwarded to the Board for consideration. The Reserve Bank advised Republic that its "application ha(d) been reviewed preliminarily and appear(ed) informationally complete," although it warned that "further analysis of the application by this Reserve Bank or Board staff could reveal information deficiencies that require either additional information or return of the application." On that same day, the Reserve Bank forwarded copies of Republic's application to the Comptroller of the Currency, the Texas Commissioner of Banking, and the Antitrust Division of the Department of Justice. The Reserve Bank's cover letter to the Comptroller requested that he send his written views and recommendations to the Board within 30 days.

The Board received the application and accompanying record from the Reserve Bank on May 2. On May 8, the Board forwarded notice of the application to the Federal Register for publication. This notice was duly published on May 16; it directed that any comments or requests for a hearing should be received by the Board no later than June 9.14

By letter dated May 6, the Texas Commissioner of Banking informed the Board that he had no views or recommendations on Republic's application; this letter was received by the Board on May 9. The period for comments from the Comptroller of the Currency ended on June 2, but the Board received no comments or recommendation from the Comptroller. Nor did the Board receive any comments from the public by June 9 in response to the Federal Register notice.

By letter dated August 18 some 109 days after the Board received Republic's application and the accompanying materials from the Reserve Bank Republic notified the Board that it deemed its application to have been approved by operation of law under the 91-day rule. On August 20, however, the Board issued its order denying Republic's application.

A. General Principles Developed by Our Sister Circuits

The 91-day rule had its origins in the Bank Holding Company Act Amendments of 1970, Pub.L.No.91-607, § 102(2), 84 Stat. 1760, 1763, reprinted at (1970) U.S.Code Cong. & Ad.News 2057, 2060. The legislative history indicates that the 91-day rule was intended to correct one principal criticism of the original version of the Act i. e., that "the Federal Reserve in administering the act ha(d) been unduly slow in rendering decisions on applications made."15 Generally speaking, the formal congressional reports on the 1970 amendments provide little indication of how the 91-day rule was intended to operate in practice.16 The language of section 3(b) does, however, speak of the "submission to the Board of the complete record on (an) application," rather than of the submission of a complete application to the Board or to the Reserve Banks, and the hearings on the 1970 amendments indicate that Congress knew of the Board's practice of requiring that an application first be submitted to a Reserve Bank for "cleaning up" before its formal submission to the Board.17 Thus, it appears that Congress did not intend for the 91-day period to begin running until some point after the cleaning-up process and the receipt by the Board of the "complete record" on an application. The crucial question of statutory construction is what Congress meant when it used the phrase "complete record on (an) application."

Although the meaning to be accorded this language is a question of first impression in this circuit, the Second, Fourth, Seventh, and Ninth Circuits18 have addressed in varying degrees either the 91-day rule of section 3(b) or the comparable 91-day rule of section 4(c) of the Act.19 While their decisions do not easily lend themselves to generalizations, there are certain principles as to which there is no disagreement.

First, after considering both the legislative history discussed above and the implication of the "submission to the Board" language of the statute, our sister circuits have uniformly held that, at a minimum, the 91-day period does not begin to run immediately upon the acceptance by the local Federal Reserve Bank of the completed application. E. g., NCNB Corp. v. Board of Governors of the Federal Reserve System, 599 F.2d 609, 612-13 (4th Cir. 1979); Citicorp v. Board of Governors of the Federal Reserve System, 589 F.2d 1182, 1186-87 (2d Cir.), cert. denied, 442 U.S. 929, 99 S. Ct. 2860, 61 L. Ed. 2d 297 (1979). We agree and so hold.

Second, our sister circuits are also in agreement that "staff reports and recommendations of the (Board) are no part of the 'complete record' within the statutory contemplation. Such internal communications are merely a part of the decisional process in determining the result to be reached on the basis of the record which has been submitted (T)he Congressional intent here was to define 'complete record' as meaning the facts upon which the agency would pass, not the internal documents relating to the process of that passing." Tri-State Bancorporation, Inc. v. Board of Governors of the Federal Reserve System, 524 F.2d 562, 566 (7th Cir. 1975). Accord, BankAmerica Corp. v. Board of Governors of the Federal Reserve System, 596 F.2d 1368, 1378 (9th Cir. 1979); Citicorp, 589 F.2d at 1189. Cf. North Lawndale Economic Development Corp. v. Board of Governors of the Federal Reserve System, 553 F.2d 23, 27 (7th Cir. 1977) (oral discussions between applicant and Board's staff concerning facts already in Board's possession was not part of the "complete record"). Again, we agree, and so hold.

Third, our sister circuits agree on a general definition of "complete record" as comprising all final material needed for the Board's decision that is received from the various interested sources outside of the Board, and that the 91-day period should begin to run upon the Board's receipt of all of this material. BankAmerica, 596 F.2d at 1378; Central Wisconsin Bankshares, Inc. v. Board of Governors of the Federal Reserve System, 583 F.2d 294, 296-97 (7th Cir. 1978); North Lawndale, 553 F.2d at 27; First Lincolnwood Corp. v. Board of Governors of the Federal Reserve System, 546 F.2d 718, 721 (7th Cir. 1976), modified on other grounds on rehearing en banc, 560 F.2d 258 (7th Cir. 1977), rev'd on other grounds, 439 U.S. 234, 99 S. Ct. 505, 58 L. Ed. 2d 484 (1978); Tri-State, 524 F.2d at 566. We agree with this general definition, and hold that as a general matter, the 91-day period begins to run when the Board receives from the various interested, external sources the final material relevant to and necessary for its decision on an application.20

Our sister circuits differ, however, as to more subtle applications of these general principles. For example, the Seventh and Ninth Circuits have held that a "record can be 'complete' on a number of successive occasions. It first becomes complete when the staff of the Board receives the application itself and any information the Board has in good faith requested from the applicant and/or other interested parties." BankAmerica, 596 F.2d at 1378. But the entire 91-day period may be "retriggered" if the record is subsequently supplemented, either by an unrequested submission from outside the Board or by a submission in response to a good faith Board request for additional information, so long as the newly submitted material is neither irrelevant nor cumulative. Id. Accord, Central Wisconsin, 583 F.2d at 296. The Second Circuit, by contrast, has expressed doubts about this "retriggering" theory, noting that "(t)he case for insuring that the Board is able to obtain and use (relevant and necessary) information is obviously a strong one, but whether the receipt of such new data should re-start the entire 91-day period is more questionable." Citicorp, 589 F.2d at 1188. The Second Circuit suggested instead that "a request for new information should not restart the 91-day period, but should merely toll the running of that period until the information is supplied by the applicant," but avoided making a firm commitment to this "tolling" theory over the Seventh and Ninth Circuits' "retriggering" theory. Id. at 1189. Given the circumstances of this case, we need not choose between these theories, and we express no views as to which is preferable.21

B. The Comment Period That Follows Federal Register Publication

One possible divergence among our sister circuits is, however, relevant to the case at bar. As discussed above in part I-B of this opinion, the Board has promulgated regulations that require the publication of notice in the Federal Register of each application under sections 3(a) or 4(c)(8) of the Act; the regulations indirectly contemplate, and the Board in practice uniformly drafts each notice so as to provide for, a period following publication of the notice during which the public may submit written comments or requests for a hearing on each application. What role does this comment period play in defining the "complete record" for purposes of the 91-day rules of sections 3(b) and 4(c) of the Act? The Second Circuit is the only court squarely to have addressed this question.

In the Citicorp decision cited above, the Second Circuit declined to find that the petitioner's section 4(c)(8) application had been granted by operation of law pursuant to section 4(c)'s 91-day rule. The court could have based this result solely upon its holding that the 91-day period contemplated by section 4(c) did not include the "cleaning up" process that occurs prior to formal acceptance of the application by the Reserve Bank and its transmission to the Board. Although the Second Circuit characterized this holding as a "tempting" stopping point since it adequately disposed of the case before it, the court continued in the interests of "sound judicial administration" to consider an alternative theory urged by the Board. 589 F.2d at 1187. In what could be characterized either as an alternate holding or as dicta, the court held that "ordinarily the 91-day period does not begin until at least (the) expiration of the period for comments upon an application." Id. (footnote omitted).22

The court focused on section 4(c)'s requirement that the Board give "due notice and an opportunity for hearing" upon section 4(c)(8) applications. Because of this statutory requirement, the court felt that even if no comments or requests for a hearing were received, "the lack of objection to the proposal is an important fact for the Board, which is not confirmed until the close of the comment period." Id. Since comments or the lack thereof are necessarily a part of the "complete record" under the Second Circuit's analysis of section 4(c), the court also focused on the dilemma that the Board might face if the 91-day period were to begin to run prior to the end of the comment period:

In order to assure the full 91 days for careful deliberation, the Board would have to assume that no comments would be received and begin intensive deliberation during the comment period. Then if there were comments that altered the focus of inquiry or changed the information available to the Board, the initial deliberations might become useless. We do not think Congress intended to force the Board either to waste time in possibly premature deliberations or to sacrifice part of its 91 days for decision-making.

Id. The court noted that there might be some conflict between its holding and that of the Seventh Circuit cases. Id. at 1187 n. 9. The Seventh Circuit, however, has never explicitly addressed the argument that the 91-day period cannot begin until the expiration of the comment period; while in North Lawndale the comment period must have extended beyond the date upon which the court held the 91-day period to have commenced, apparently the Board did not argue that point before the Seventh Circuit.23

The Second Circuit recognized that the chief argument against its holding was that "the purpose of the statute could be undercut if the local Reserve Bank unduly delays in forwarding an application to the Board, or if the Board delays in publishing notice, or allows an unreasonably long period for comments." 589 F.2d at 1187. The court accepted the proposition that "the Board's discretion to determine the start of the 91-day time period is limited by the implicit statutory requirement that these activities be accomplished with reasonable promptness," id. at 1188, and accordingly held that "any undue delay at the Reserve Bank should be subtracted from the 91 days allotted to the Board." Id. But the court was satisfied that the Board's unpublished regulations, which required publication of notice within two weeks of receipt of an application and a comment period of 30 days, were "objective guidelines (that) adequately curtail(ed) the potential for Board control or manipulation of the time limit." Id. Significantly, in calculating the running of the 91-day period, the court held the Board to its internal guidelines, subtracting from the comment period an amount of time equal to the Board's delay beyond its internal deadline for securing Federal Register publication. Id. at 1189.

In the case at bar, the Board argues that we should follow the Second Circuit's Citicorp decision; since the close of the comment period in this case was less than 91 days prior to the Board's decision on Republic's application, that application could not, under Citicorp, be deemed to have been granted by operation of law. Republic argues both that Citicorp is distinguishable and that Citicorp is unsound precedent that we should not follow. We disagree with both of Republic's arguments.

Most courts have assumed or stated in dicta that the 91-day rules under section 3(b) and section 4(c) are interchangeable.24 Republic argues, in effect, that they are not. It is true that there are some differences between section 3(b) and section 4(c). The Citicorp court relied in large measure upon section 4(c)'s "due notice and opportunity for hearing" requirement, but there is no such requirement in the language of section 3(b). In fact, our sister circuits have uniformly recognized the different hearing requirements for the two sections.25 And while another part of the 1970 amendments to the Act, 12 U.S.C. § 1850 (1976), affords competitors of both section 3(a) and section 4(c)(8) applicants the right to join as parties in interest in proceedings before the Board, that provision does not require that such parties be notified or that they be given a comment period of any particular length.26 Finally, as noted above in part I-B of this opinion, section 3(b) does not explicitly mandate the provision of notice to, and an opportunity for comment from, the public at large.

Nonetheless, section 3(c) does mandate that one factor that the Board must consider in determining whether to approve an application under section 3 is whether any anticompetitive effects of the proposed acquisition would be "clearly outweighed in the public interest by the probable effect of the transaction in meeting the convenience and needs of the community to be served." 12 U.S.C.A. § 1842(c) (West 1980). Congress could hardly have expected that the Board would weigh the community's convenience and needs without soliciting some sort of input from the public. We agree with the Second Circuit that even if no public comments are received, the lack of objection to the proposal is an important fact for the Board a fact that cannot be determined until after the close of a public comment period.

Accordingly, we hold that for the purposes of section 3(b)'s 91-day rule, the "complete record on (an) application" is not received by the Board until at least the close of the public comment period that follows the publication of notice in the Federal Register.

Republic argues that this holding would defeat the purposes of the 91-day rule by allowing the Board to determine the running of the 91-day period by altering either the date upon which the Federal Register notice is published or the length of the public comment period. This argument does have considerable persuasive force. Although we join the Seventh Circuit in "choos(ing) to believe that this responsible part of our Government will act in a responsible and good faith manner," Tri-State, 524 F.2d at 567, we are concerned over the Board's utter failure to provide clear, public guidelines that explain its policies and practices regarding the publication of the Federal Register notice and the subsequent comment period, and we agree with the Citicorp court that there must be some sort of check on the Board's discretion in this area.

C. The Need for a Check on the Board's Discretion

We do not accuse the Board of bad faith by pointing out the mystery, confusion, and self-contradiction inherent in its past and present practices and policy pronouncements in this area all of which stand in sharp contrast to the clear, detailed newspaper notice and comment requirements that the Board has imposed upon applicants.27 For example, a 1972 opinion letter from the Board announced the following policy:

It should be noted that, at a minimum, the complete record on a given application cannot be ready for submission to the Board prior to the expiration of any comment period. Normally, notice of an application is published in the Federal Register within five to seven days following acceptance of the application by a Reserve Bank. Generally 30 days are allowed for comment from interested persons, from the public, and from Federal and State supervisory authorities. Persons counseling prospective applicants should avoid any suggestion or assurance that the complete record will be before the Board for final action at any time before expiration of the time for submission of comment.28

But the Second Circuit's opinion in Citicorp, while quoting from this letter, 589 F.2d at 1188 n. 11, indicates "that an internal unpublished Board regulation provides for publication of notice within two weeks of receipt of an application and for a comment period of thirty days." Id. at 1188 (emphasis added). It is not clear whether the Citicorp court meant to refer to receipt of the application by the local Reserve Bank, as contemplated by the Board's 1972 opinion letter, or to receipt of the application by the Board after the "cleaning up" process at the Reserve Bank; in either event, however, there is a discrepancy as to when the Federal Register notice must be published. And at oral argument in this case, the Board's counsel informed us that internal Board regulations provide "that the notice has to be executed within one week of the receipt of the application" presumably meaning receipt by the Board.

The Board's actual practices appear to have been equally as inconsistent as its policy statements. For example, the notice in the North Lawndale case was published over one month after the Board had received the cleaned up application from the Reserve Bank; the Seventh Circuit noted in passing that this did not comply with the Board's 1972 opinion letter, but did not attach particular significance to that fact. 553 F.2d at 26 & n. 3. And in Citicorp, the Board had failed by some ten days to comply with the internal guidelines it described at oral argument before the Second Circuit. 589 F.2d at 1189.

Perhaps in recognition of the confusion surrounding the 91-day rules generally, the Board gave these directions to all Reserve Banks in late 1975, shortly after the Seventh Circuit's decision in Tri-State :

Bank holding companies in your District should be cautioned about consummating any acquisition in reliance upon their own determination, or the opinion of their counsel that the 91-day period has expired and that their application has therefore been granted by operation of law, and should be urged to consult with the Reserve Bank or with the Board's Legal Division in the event they have any question about the effect of the Tri-State decision upon a particular application.29

We think that the Board has misapprehended the entire thrust of the 91-day rules. Congress clearly intended that these statutory provisions would set definite, readily ascertainable time limits for the processing of section 3(a) and section 4(c)(8) applications, upon which parties could rely. In all likelihood, Congress envisioned that in implementing the 91-day rules of section 3(b) and section 4(c), the Federal Reserve System would provide all parties interested in a particular application with some sort of formal, definite notification that the "complete record on (an) application" had been submitted to the Board and that the Board was taking the application and accompanying record under advisement for a period not to exceed 91 days. Such a ritualistic closing of the record would provide an unambiguous reference point from which the running of the 91-day period could be computed. Were the Board to conclude during its deliberations that additional information was necessary for it to make an informed decision, it could effectively reopen the record to receive that information; such a reopening of the record would, according to our sister circuits, either toll or retrigger the running of the 91-day period (although we do not here decide which).

The Board, however, has not chosen to implement the 91-day rules in this manner, and because the Board rather than the judiciary is charged with implementing the statute, it is not within our province to command the Board to do so. But because the Board has failed to create this or some similarly concrete system, the 91-day rules are not the simple mechanisms intended by Congress to apprise the Board and the parties of the deadline for Board action on any given application in advance of that deadline. Instead, they operate as mystical, almost arbitrary sources of confusion and litigation, often adding to the bureaucratic delay they were designed to remedy. In effect, one cannot know whether the 91-day period has run in any given case until after the fact, if then.

The cure for this larger problem is not within our means; it must come from Congress or the Board. Yet we are loath to allow the Board the power to frustrate the clear mandate of the statute by placing unlimited discretion in its hands as to when the Federal Register notice is published and how long a public comment period is specified therein. Accordingly, we follow the Citicorp court's path, and will hold the Board to the internal guidelines described by counsel for the Board at oral argument before this court. We hold that if the Board fails to secure publication of the public notice in the Federal Register within one week from its receipt of the completed application from the Reserve Bank, or if the Board sets a public comment period greater than 30 days from the end of that one-week period, such delay shall not prevent the running of the 91-day period. In effect, any such delays are to come out of the Board's deliberation period. We also follow the Citicorp court in holding that any undue delay by the Reserve Bank in forwarding the completed application to the Board after the conclusion of the cleaning up period is to be subtracted from the 91 days allotted to the Board.30

We emphasize that we do not intend to carve into stone the internal guidelines described by the Board at oral argument before this court. Were the Board to issue clear, reasonable, and unambiguous regulations, consistent with section 3(b), that set out in more detail the Board's policies with regard to the publication of the Federal Register notice and the subsequent public comment period regulations upon which outside parties could rely, and by which the Board considered itself bound we would certainly defer to those regulations.

Applying these holdings to the case at bar, we note that Republic's completed application was received by the Board one day after it was formally accepted by the Reserve Bank; Republic participated voluntarily in the cleaning up process, and never demanded that the Reserve Bank formally accept its application on any earlier date. We therefore find no undue delay by the Reserve Bank in forwarding the completed application to the Board. The Board met its internal guidelines by arranging for the publication of the Federal Register notice within one week after receipt of the completed application, and the public comment period specified therein did not extend more than 30 days beyond the end of that one-week period. Accordingly, the Board did not receive "the complete record on (Republic's) application" until the close of the public comment period on June 9, 1980. Its August 20 decision on Republic's application was well within 91 days from the date it received the complete record. Accordingly, Republic's application cannot be deemed to have been granted by operation of law under section 3(b)'s 91-day rule.31

A. The Standard of Review and the Statutory Prerequisites for Rejecting a Section 3(a) Application

The second issue in this case the merits of the Board's order is controlled by our recent holding in Mercantile Texas Corp. v. Board of Governors of the Federal Reserve System, 638 F.2d 1255 (5th Cir. 1981) (vacating and remanding 66 Fed.Res.Bull. 423 (1980)). The Mercantile court began, as we do, by stressing the standard by which we review such orders of the Board:

The Board's findings as to the facts, if supported by substantial evidence, are conclusive. 12 U.S.C. § 1848 Like decisions of other administrative agencies, however, the Board's order may be affirmed only on the basis of the findings explicitly made in its opinions Its decision cannot be affirmed on the basis of "appellate counsel's post hoc rationalizations for the agency action," or because the contents of the record support conclusions that the Board has not stated. When, as here, an agency makes only minimal findings, its decision rests on precarious ground.

638 F.2d at 1260.

Section 3(c) of the Act forbids Board approval of any bank holding company acquisition that would foster a monopoly, or "whose effect may be substantially to lessen competition" unless the Board shall find "that the anticompetitive effects are clearly outweighed in the public interest by the probable effect of the transaction in meeting the convenience and needs of the community to be served." 12 U.S.C.A. § 1842(c) (West 1980). The Mercantile court rejected the Board's argument that the Board may prohibit a proposed merger between bank holding companies if it finds that the anticompetitive effects of the transaction would thwart the "convenience and needs" of the community even if these effects are insufficient to violate section 3(c)'s antitrust standards, which are largely derived from section 7 of the Clayton Act, 15 U.S.C.A. § 18 (West Supp. 1981): "We hold that the Board may not deny approval to a proposed merger of bank holding companies without finding a violation of the antitrust standards explicitly incorporated into (section 3(c) of the Act)." 638 F.2d at 1263.

Although phrased in terms of a merger of bank holding companies, we see no reason why this holding of Mercantile should not apply equally to a bank holding company's acquisition of a bank, as in the case at bar. Accordingly, for the reasons stated by the Mercantile court, 638 F.2d at 1260-63, we hold that the Board may not deny approval of a bank holding company's section 3(a) application to acquire a bank unless the Board finds that the acquisition would constitute a violation of the antitrust standards explicitly incorporated into section 3(c) of the Act.

In the light of the Mercantile decision, the Board has argued that its order in this case contains such a finding. Specifically, the Board argues that its order reflects a finding that Republic's acquisition of CNB would constitute a violation of section 3(c)'s Clayton Act standards by eliminating "probable future competition" (described by the Mercantile court as "actual potential competition") between CNB and Republic: the order states that "(a)cquisition of (CNB) by (Republic) would eliminate the probability that these two organizations would come into direct competition in the future and the Board would view this competition as desirable because of the present (concentrated) structure of the (Waco) market."B. The Findings Necessary as a Predicate to the Application of the Actual Potential Competition Doctrine

The Mercantile court discussed at some length the theory that underlies, and the policy arguments for and against, the actual potential competition doctrine; we need not repeat that discussion here. After weighing these arguments, the court concluded that "the doctrine has logical force and is consonant with the language and policy of the Clayton Act." 638 F.2d at 1265. The court stressed, however, that "(t)he Clayton Act prohibits more than 'certainties'; it also prohibits 'probabilities,' but not 'ephemeral possibilities.' " Id. at 1266. Because the Board had made only "minimal findings" in that case, the court could not tell "whether more thorough findings (could) separate probabilities from 'ephemeral possibilities.' " Id. Accordingly, the court specifically declined to decide "whether the (actual potential competition) doctrine adequately describes a violation of the Clayton Act standard." Id. at 1265.

The court went on to detail the four types of findings that were necessary before it could evaluate the viability of this controversial antitrust theory: (1) whether the relevant market is performing oligopolistically, with "the dominant participants in the target market engaging in interdependent or parallel behavior and with the capacity effectively to determine price and total output of goods or services";32 (2) whether there was a sufficient number of other likely potential entrants into the market so as to vitiate the applicant's importance as a potential competitor; (3) whether there was a reasonable probability that the applicant would, in fact, enter the market independently, based on the anticipated profitability of independent entry, the anticipated profitability of other investment opportunities available to the applicant, and evidence specific to the applicant demonstrating why it might prefer independent entry; and (4) whether independent entry would result in a substantial likelihood of ultimately producing deconcentration of that market or other significant procompetitive effects. Id. at 1266-72. Rather than assuming arguendo that properly supported findings on these issues would state a Clayton Act violation and reversing the Board's order as not being based on substantial evidence, the court remanded "for a more thorough analysis of the relevant economic data" on the last three of these issues, id. at 1270, because the court believed that "further proceedings (would) better insure that the proper disposition is reached," id. at 1272.

We follow the same path. The Board's opinion in the case at bar contains an inadequate finding on the last three of the four issues identified by the Mercantile court as crucial. We discuss all four issues in sequence.

1. The oligopolistic character of the Waco banking market. With respect to the first necessary finding that the relevant market is performing oligopolistically there is no meaningful distinction between the Board opinion and underlying record in this case and those in the Mercantile case. The relevant markets were defined identically, with the Waco standard metropolitan statistical area (SMSA) defined as the geographic market and commercial banking defined as the product market.33 In Mercantile, the four-firm concentration ratio for the Waco banking market, measured by commercial bank deposits,34 was 73.8%; in the case at bar, the four-firm concentration ratio (measured at a slightly different time) was 72.4%. Concentration ratios of this magnitude35 establish here, as in Mercantile, a prima facie case that the Waco bank market is a candidate for the potential competition doctrine, and shift to Republic the burden to show that the concentration ratios (which may be unreliable indicators of actual market behavior) do not accurately depict the economic characteristics of the Waco market. See Mercantile, 638 F.2d at 1267; see also United States v. Marine Bancorporation, Inc., 418 U.S. 602, 631, 94 S. Ct. 2856, 2874, 41 L. Ed. 2d 978 (1974).

The Mercantile court found that Mercantile Texas Corp. had failed in its attempt to establish a "clear trend to deconcentration";36 accordingly, in the light of the prima facie case of oligopolistic behavior made out by the concentration ratios and the failure of the bank holding company adequately to rebut that prima facie case, the court concluded that there was substantial evidence to support the Board's finding that the Waco market was performing oligopolistically. 638 F.2d at 1267.

In the case at bar, Republic argues that evidence of competition from thrift institutions in the Waco SMSA adequately rebuts the prima facie case established by the commercial bank deposit concentration ratios. The Board rejected this argument, concluding that while "thrift institutions hold substantial deposits in the Waco banking market, the Board in this instance is unable to conclude from the evidence in the record that these institutions compete actively with commercial banks over a sufficient range of financial services to mitigate significantly the anticompetitive effects of (Republic's proposed acquisition of CNB)." In effect the Board thereby refused to redefine the relevant product market to include the thrift institutions, which in turn would have altered the concentration ratios sufficiently to cast doubt upon their significance as an indicator of oligopolistic behavior.37

We believe that this refusal was based on substantial evidence. Republic did little more than to point to the amount of the commercial deposits held by such thrift institutions. While the Marine Bancorporation Court recognized that deposit concentration ratios may establish a prima facie case that an entire market is performing oligopolistically, such ratios reflect only one component of the entire field of competition across a wide range of financial services; they do not by themselves establish that the overall competition from the thrift institutions was sufficient to render improper their exclusion from the defined product market.

Accordingly, Republic cannot be said to have rebutted the prima facie showing that the Waco banking market is performing oligopolistically for purposes of the potential competition doctrine. We affirm the Board's finding on this first key issue.38

2. Other potential entrants. The Mercantile court stressed that "(i)f there are numerous other potential competitors waiting in the wings, elimination of the (applicant bank holding company) as one potential entrant would not be significant." 638 F.2d at 1267. Consequently, the court held that as a predicate for any application of the actual potential competition theory, the Board must make explicit findings in which it identifies the other potential entrants into the target market and evaluates "whether the number of other potential entrants and the likelihood of their potential entry vitiate (the applicant's) importance as a potential competitor." Id. at 1268.

In the case at bar, the Board made no such findings. It argues on appeal that "the record reflects agreement as to the limited number of (potential) entrants" an argument that Republic sharply disputes and suggests that in these circumstances, "the need for an explicit evaluation by the Board does not appear to be compelling." This suggestion simply flies in the face of the Mercantile court's emphasis on the standard for reviewing Board decisions: the Board's decision cannot be affirmed merely because the contents of the record support conclusions that the Board has not stated. Id. at 1260. As in Mercantile, the Board has made no reviewable finding on this issue.

3. The reasonable probability of independent entry. The Board did find that "in view of (Republic's) overall size, financial and managerial resources and history of, and plans for, expansion, (Republic) appears to be a likely entrant into the (Waco) market in the reasonably foreseeable future." But the Board failed to satisfy the Mercantile court's requirement, 638 F.2d at 1269, that it offer a persuasive rationale demonstrating that Republic would prefer the opportunity to enter Waco over other opportunities for expansion or investment. Specifically, the Board did not consider the anticipated profitability of independent entry in Waco; neither did it evaluate the profitability of other opportunities for investment or expansion available to Republic elsewhere. Accordingly, the Board has made no reviewable finding on the likelihood of independent entry.

4. Significant procompetitive effects from independent entry. Finally, the Board failed to make a reviewable finding on whether Republic's independent entry into the Waco banking market would result in a substantial likelihood of ultimately producing deconcentration of that market or other significant procompetitive effects. The Board made no explicit finding that independent entry by Republic "would result in 'a substantial likelihood of ultimately producing deconcentration of that market or other significant procompetitive effects.' " Mercantile, 638 F.2d at 1270 (quoting Marine Bancorporation, 418 U.S. at 633, 94 S.Ct. at 2875). As in Mercantile, the preservation of competition between the applicant and the bank to be acquired would be a procompetitive effect, but the Board's findings "fail to explain the significance of that finding." 638 F.2d at 1270. The Board's findings also fail to assess the significance of Republic's potential competition with banking institutions other than CNB that are already in the Waco market.

Similarly, the Board's prediction that Republic is a likely candidate for independent entry in the "reasonably foreseeable future" falls well short of the Mercantile court's requirement that the Board determine "at what point (the applicant) would be likely to enter the market." Id. at 1272. A more definite finding perhaps specifying a range of months or years within which such entry would likely occur is necessary in order for the Board to predict with any degree of accuracy whether the target market will still be performing oligopolistically at the time of entry, which in turn is a necessary finding en route to the ultimate conclusion that independent entry would result in a substantial likelihood of significant procompetitive effects.

The Board's failure to make sufficient findings with respect to three of the four issues that were listed as crucial by the Mercantile court means that, given the standard of review set forth above in part IV-A of this opinion, we cannot affirm the Board's order. However, because the Board's decision in this case was made without benefit of the Mercantile court's analysis, we decline to reverse the Board's order as not being based on substantial evidence; instead, we vacate that order and remand for reconsideration and more thorough findings in the light of that case and of this opinion.39 As did the Mercantile court, we decline to decide in the absence of such findings whether the actual potential competition doctrine adequately describes a violation of the Clayton Act standard incorporated into section 3 of the Bank Holding Company Act.

VACATED and REMANDED.