Mercantile Texas Corporation is a bank holding company operating nine banks throughout Texas. PanNational Group, Inc., a smaller bank holding company, operates five banks in El Paso and Waco, cities not presently served by Mercantile. Mercantile seeks review of a Federal Reserve Board order denying approval to a proposed merger between the two companies. The Board held that elimination of potential competition between the two companies would have a substantially adverse effect on competition in El Paso and Waco without countervailing benefits to those communities.

On review, the Board argues that Section 1842(c) of the Bank Holding Company Act, 12 U.S.C. § 1842(c), grants it broad discretion to reject proposed mergers upon finding that a merger would have an anticompetitive effect detrimental to the convenience and needs of the community even if the merger would not violate the explicit antitrust standards included in Section 1842(c). We conclude that the statute denies the Board the broad discretion it claims. If the Board rejects a proposed merger on anticompetitive grounds, it must find a violation of the Sherman and Clayton Act standards written into the statute.

Anticipating our interpretation, the Board argues that the merger would violate the Clayton Act standard by eliminating potential competition in the two markets. Because the Board failed to make several important findings of fact (and because we can affirm the Board's decision only on the basis of the findings that the Board actually made), we cannot determine whether the merger would violate the potential competition doctrine. Without adequate findings, we are also unwilling to decide whether the elimination of potential competition constitutes a violation of the Clayton Act standard.

Accordingly, we vacate the Board's order and remand Mercantile's petition to the Board for reconsideration and more thorough findings.

Mercantile Texas Corporation is the fifth largest bank holding company in Texas. Its nine banks have aggregate deposits of $2.8 billion, representing 4.2 per cent of the total commercial bank deposits in the state.1 PanNational controls five banks with deposits of $622 million, representing 0.9 per cent of the commercial bank deposits in Texas. The merged corporation would have 5.1 per cent of Texas commercial bank deposits. Mercantile would remain the fifth largest banking organization, but its total deposits would increase by approximately 22 per cent.

PanNational operates four banks in El Paso; the nearest market served by Mercantile is San Antonio, more than 500 miles away. PanNational also owns one bank in Waco; the nearest market served by Mercantile is Dallas, 95 miles away. Because of these distances, the Board found that the proposed merger would not eliminate any significant present competition between the two companies.

The Board found, instead, that future potential competition in Waco and El Paso between the two companies would probably be eliminated by the merger.2 It predicted that, if the proposed merger were rejected, Mercantile would nevertheless enter the two markets, but in a manner resulting in greater competition. The Board stated this conclusion after only a limited discussion of the relevant economic data.

The Board's findings as to the facts, if supported by substantial evidence, are conclusive. 12 U.S.C. § 1848; First State Bank of Clute v. Board of Governors, 553 F.2d 950 (5th Cir. 1977). See 5 U.S.C. § 706(2)(E) (judicial review under the Administrative Procedure Act). Like decisions of other agencies, however, the Board's order may be affirmed only on the basis of the findings explicitly made in its opinions. SEC v. Chenery Corp. (Chenery I), 332 U.S. 194, 67 S. Ct. 1575, 91 L. Ed. 1995 (1947); Real v. Simon, 514 F.2d 738 (5th Cir. 1975). See Gravois Bank v. Board of Governors, 478 F.2d 546 (8th Cir. 1973). Its decision cannot be affirmed on the basis of "appellate counsel's post hoc rationalizations for agency action;" Burlington Truck Lines Inc. v. United States, 371 U.S. 156, 168, 83 S. Ct. 239, 246, 9 L. Ed. 2d 207, 215 (1962); Real v. Simon, 514 F.2d 738 (5th Cir. 1975); 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.

We begin by examining Section 1842(c) to determine what standard governs the Board's discretion in reviewing proposed mergers.

Section 1842(c) of the Bank Holding Company Act, 12 U.S.C. § 1842(c),3 forbids Board approval of any merger that would foster a monopoly or "whose effect ... may be substantially to lessen competition" unless "it finds that the anticompetitive effects are clearly outweighed in the public interest."4 The phrase "substantially to lessen competition" is, of course, borrowed from Section 7 of the Clayton Act, 15 U.S.C. § 18. Because this repetition was purposeful,5 the principles developed under the Clayton Act are applicable to mergers of bank holding companies under Section 1842(c), Mid-Nebraska Bancshares, Inc. v. Board of Governors, 627 F.2d 266 (D.C. Cir. 1980), as well as to mergers of banks under Section 1828(c)(5)(B) of the Bank Merger Act, which uses the same language. See United States v. Marine Bancorporation, 418 U.S. 602, 94 S. Ct. 2856, 41 L. Ed. 2d 978 (1974); United States v. First City National Bank, 386 U.S. 361, 87 S. Ct. 1088, 18 L. Ed. 2d 151 (1967).

If the antitrust consequences of the merger do not require Board disapproval, then the statute commands the Board to "take into consideration ... the convenience and needs of the community." Counsel for the Board argues that this second provision enables it to reject proposed mergers if it finds that the anticompetitive effects of the merger would thwart the "convenience and needs" of the community even if these effects are insufficient to violate the antitrust standards. While an agency's interpretation of its own governing statutes is entitled to great deference, Board of Governors v. First Lincolnwood Corp., 439 U.S. 234, 99 S. Ct. 505, 58 L. Ed. 2d 484 (1978), we are not persuaded that Congress intended the Board to have nearly unlimited discretion to reject proposed mergers on anticompetitive grounds beyond Clayton Act violations.

Two circuits have already rejected the Board's contention. In Washington Mutual Savings Bank v. F.D.I.C., 482 F.2d 459 (9th Cir. 1973), the Ninth Circuit considered this argument in the context of the identical language of the Bank Merger Act provision, 12 U.S.C. § 1828(c)(5). The court held that Congress intended to impose a uniform antitrust standard for consideration of bank mergers, thereby enabling the agencies and courts to develop a "discernible body of law." 482 F.2d at 464. The community's "convenience and needs," which the Board must also consider, was found to refer exclusively to banking (as distinguished from competitive) factors. Id. at 465. In County National Bancorporation v. Board of Governors, No. 79-1783 (8th Cir. Sept. 3, 1980), rehearing granted, the Eighth Circuit followed Washington Mutual and held that the Clayton Act standard was the exclusive test for evaluating anticompetitive effects under the Bank Holding Company Act provision, 12 U.S.C. § 1842(c).

The legislative history of the two provisions establishes that Congress did not intend for the Board to apply a more stringent standard than ordained by Section 7 of the Clayton Act. Above all, Congress sought to establish "uniform standards" for consideration of anticompetitive effects of bank mergers and acquisitions by bank holding companies. S.Rep.No. 1179, 89th Cong., 1st Sess. 10, reprinted in (1966) U.S. Code Cong. & Ad. News 2385, 2394 (1966 amendment to the Bank Holding Company Act). See H.R.Rep.No. 1221, 89th Cong., 1st Sess. 1, reprinted in (1966) U.S. Code Cong. & Ad. News 1860 (1966 amendment to Bank Merger Act); Austin, The Evolution of Commercial Bank Merger Antitrust Law, 36 Bus.Law. 297, 315 (1981).6 Allowing the Board to employ a more stringent standard tailored to the "convenience and needs" of each community would destroy that uniformity.

The House Banking and Currency Committee designed the language of the statute as a compromise between those seeking the most vigorous possible application of the antitrust laws to banks and those seeking to exempt banks from their operation.7 The two sides compromised by applying the established principles of the antitrust laws to bank mergers, but allowing those mergers whose anticompetitive effects are outweighed by "the convenience and needs of the community." By applying standards more stringent than the antitrust laws, the Board would go beyond the terms of the compromise to gain a victory that Congressional antitrust proponents may not have sought and certainly did not obtain.

The Board argues that Congress established standards that would be uniform only when the Board is required to reject a proposed merger, but did not limit the Board's discretion to apply a standard stricter than the Clayton Act if the "convenience and needs" of the community warrant. This construction, however, gives "convenience and needs" an additional meaning unknown to the bill's proponents.

The statute twice uses the same term, "convenience and needs" only a few lines apart. The term is first used to state circumstances that may outweigh even Clayton Act anticompetitive effects. It is next used to state a criterion to be applied "in every case." There is some indication in the congressional debate that the two phrases are intended to reflect the same values. See 112 Cong.Rec. 2453 (1966) (remarks of Congressman Gonzalez). The Board's interpretation would require the clause to be given a different effect the second time it is used.

In the first usage, the phrase "convenience and needs" is explicitly contrasted with Clayton Act anticompetitive factors and refers to considerations other than competitive impact. These are referred to as "banking factors." They include such considerations as the general availability of banking services, their adequacy for the economic needs of the community, the geographic location of banking facilities and all of the other indicia by which it can be determined that, notwithstanding anticompetitive effect, there is a real social and economic necessity for additional banking service. The Board's interpretation of the second usage would treat it as embracing undefined anticompetitive factors together with the banking factors that measure convenience and needs. Under the Board's interpretation, these factors must necessarily create a standard more stringent than the Clayton Act exacts. Any merger violating the Clayton Act standard would ipso facto fail the test set under Section 1842(c)(2).

The House Report accompanying the 1966 amendments to the Bank Merger Act states that the phrase "convenience and needs" was substituted for "archaic" phraseology describing "the so-called banking factors." H.R.Rep.No. 1221, 89th Cong., 1st Sess. 4, reprinted in (1966) U.S.Code Cong. & Ad.News 1860, 1863.8 Throughout the debate on the Bank Merger Act amendments, the bill's proponents consistently contrasted "convenience and needs" with competitive factors. Congressman Brock, the leading Republican supporter, explained:

... this bill would make the competitive factor, as defined by the antitrust laws, the primary factor to be used both by the bank supervisory agencies and the courts in determining whether to approve a merger.... In place of the six banking factors (contained in the original version of the Bank Merger Act), this bill substitutes "the convenience and needs of the community to be served."

The way in which this factor of convenience and needs of the community to be served is juxtaposed against the antitrust competitive standard is important....

112 Cong.Rec. 2444 (1966).9 It is possible that Congressman Brock was referring exclusively to the use of "convenience and needs" in subparagraph (2). However, no supporter of the bill gave any indication that he regarded the phrase as referring to more than banking factors in either context.

To buttress its interpretation, the Board relies mainly on two statements made during the House debate on the Bank Merger Act amendments. Two opponents of the bill declared that the need for a competitive market in banking was a primary "need" of the community. See 112 Cong.Rec. 2457 (1966) (remarks of Congressman Todd); Id. at 2459 (remarks of Congressman Weltner).

While statements by a bill's opponents are relevant in determining congressional intent, Arizona v. California, 373 U.S. 546, 83 S. Ct. 1468, 10 L. Ed. 2d 542 (1963), they are entitled to little weight. Ernst & Ernst v. Hochfelder, 425 U.S. 185, 203 n.24, 96 S. Ct. 1375, 1386 n.24, 47 L. Ed. 2d 668, 682 n.24 (1976); Holtzman v. Schlesinger, 414 U.S. 1304, 94 S. Ct. 1, 38 L. Ed. 2d 18 (1973) (per Marshall, Circuit Justice). Given greater weight, however, they would not compel the Board's interpretation. While the meaning of the statements is not entirely clear, they appear to be merely rhetorical complaints by two congressmen desiring strict antitrust enforcement in the banking industry, not attempts to construe language to which both congressmen were opposed.

The Board also points to a statement in the House Report accompanying the 1966 Bank Merger Act amendments that the Board may consider "in any particular case such other factors as (it) might deem relevant." H.R.Rep.No. 1221, 89th Cong., 1st Sess. 4, reprinted in (1966) U.S.Code Cong. & Ad.News 1863. This statement, however, is made in the context of a discussion of banking factors; it indicates only that the language substituted for the "archaic ... phraseology" of the previous version of the statute (see id.) does not exhaust the possible banking and financial factors that the Board may consider.10 No mention is made of additional anticompetitive factors.

Finding that the Board's interpretation of Section 1842(c) contradicts the antitrust policy embodied in the statute, 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 the statute.

In rejecting Mercantile's application, the Board concluded that the merger would violate the Clayton Act standard by eliminating "actual potential competition" between Mercantile and PanNational.11 The Board relied on a controversial doctrine whose validity remains a current subject of debate.

In a concentrated market, a small number of firms can dominate the market through their ability to manipulate the supply of a given product or service. Collusion is difficult to prove, but easy to accomplish either by agreement or through consciously parallel action. In banking, for example, a concentrated market enables dominant banks to avoid cutting interest rates, offering new services or extending credit to all but the safest credit risks, secure in the knowledge that their customers cannot go to a competitor and get better service.12

Seeing the possibility of significant profits, companies outside of the market will desire to enter it. One way to enter easily is to acquire a successful company already in the market. Thus, the new entrant does not add to the number of competitors and yet enjoys the opportunity to enjoy the oligopolistic profits that result from the noncompetitive environment.

The potential competition doctrine postulates two benefits if outside firms are prevented from acquiring dominant firms within a concentrated market. "Perceived potential competition" describes the effect exerted on a concentrated market by large firms outside of the market who are "perceived" as likely entrants. The threatened entry of a powerful new competitor may intimidate already dominant firms into maintaining nearly competitive prices to avoid enticing the outside "midget killer" into the market by the prospect of oligopoly profits. The Supreme Court accepted this branch of the theory in United States v. Falstaff Brewing Corp., 410 U.S. 526, 93 S. Ct. 1096, 35 L. Ed. 2d 475 (1973), remanding for a determination whether Falstaff Brewing's threatened entry into the New England beer market deterred local brewers from raising prices. The Board now maintains that Mercantile was also a perceived potential competitor. Because the Board made no finding to that effect in its decision, we cannot consider this post hoc rationalization on appeal. Burlington Truck Lines, Inc. v. United States, 371 U.S. 156, 83 S. Ct. 239, 9 L. Ed. 2d 207 (1962); SEC v. Chenery Corp. (Chenery I ), 332 U.S. 194, 67 S. Ct. 1575, 91 L. Ed. 1995 (1947).

While perceived potential competition describes a present procompetitive effect, "actual potential competition"13 describes the future procompetitive benefit that occurs when an effective new competitor "actually" enters the market on its own. (Hereafter, "potential competition" and "potential competitor" refer exclusively to the actual potential competition doctrine.) The potential competition doctrine holds that, if a strong outsider is prevented from acquiring a dominant insider, the outsider will still enter the market independently, either de novo or through the purchase of a small (i. e. not dominant) company to establish a "toehold" in the market. If the outsider is forced to enter independently, it is more likely to become an aggressive price-cutting competitor as it attempts to increase its share of the market. The addition of a new effective competitor to the band of dominant firms will also make collusion or conscious parallelism more difficult. Had the outsider been allowed to merge with a successful insider, it would have assumed the role of the insider and continued the anticompetitive practices. Because the outsider's full competitive force will never be felt, the merger is said to substantially lessen competition for purposes of the Clayton Act standard.14

The Supreme Court has twice refused to decide whether the elimination of "actual potential competition" violates the Clayton Act. See United States v. Marine Bancorporation, 418 U.S. 602, 94 S. Ct. 2856, 41 L. Ed. 2d 978 (1974) (decided under the Bank Merger Act provision); United States v. Falstaff Brewing Corp., 410 U.S. 526, 93 S. Ct. 1096, 35 L. Ed. 2d 475 (1973). This Circuit has not yet passed on the issue.15 As the Second Circuit noted in BOC International Ltd. v. FTC, 557 F.2d 24 (2d Cir. 1977), we are on the frontiers of antitrust law in attempting to determine the validity and applicability of the theory to specific mergers.

We believe that the doctrine has logical force and is consonant with the language and policy of the Clayton Act. The Act bars any merger that would "substantially lessen" competition. Some commentators have argued that it prohibits only those mergers substantially lessening present competition. See Kaplan, Potential Competition and Section 7 of the Clayton Act, 25 Antitrust Bull. 297, 314-317 (1980); Rahl, Applicability of the Clayton Act to Potential Competition, 12 ABA Section of Antitrust L. 128, 142-43 (1958); Comment, Toehold Acquisitions and the Potential Competition Doctrine, 40 U.Chi.L.Rev. 156, 180-82 & n.105 (1972). The statute, however, does not command us to determine whether only present competition has been substantially lessened. In evaluating a merger under Section 7, the courts have looked at its effect on future competition. In United States v. Penn-Olin Chemical Co., 378 U.S. 158, 84 S. Ct. 1710, 12 L. Ed. 2d 775 (1964), the Court invalidated a joint venture between Pennsalt Chemicals and Olin-Mathieson whereby "Penn-Olin Chemical Co." would enter the sodium chlorate market in the southeastern United States. While the Court relied on several grounds, it noted that the joint venture (in effect, a merger) "may well foreclose any prospect of competition between Olin and Pennsalt in the relevant sodium chlorate market." 378 U.S. at 173, 84 S.Ct. at 1718, 12 L. Ed. 2d at 786. Such competition could, of course, take place only in the future. In FTC v. Procter & Gamble Co., 386 U.S. 568, 87 S. Ct. 1224, 18 L. Ed. 2d 303 (1967), we are instructed to look at a merger's impact on competition "present and future." 386 U.S. at 577, 87 S.Ct. at 1229, 18 L. Ed. 2d at 309. The Clayton Act standard was designed as a "prophylactic measure, intended 'primarily to arrest apprehended consequences of intercorporate relationships before those relationships could work their evil.' " Brunswick Corp. v. Pueblo Bowl-O-Mat, Inc., 429 U.S. 477, 485, 97 S. Ct. 690, 695, 50 L. Ed. 2d 701, 710 (1977), quoting United States v. E.I. Du Pont de Nemours & Co., 353 U.S. 586, 597, 77 S. Ct. 872, 879, 1 L. Ed. 2d 1057, 1069 (1957). See Brodley, Potential Competition Mergers: A Structural Synthesis, 87 Yale L.J. 1, 45-52 (1977).

In the absence of necessary findings by the Board, however, we will not decide whether the doctrine adequately describes a violation of the Clayton Act standard. Professor Posner has argued that it is impossible to develop workable criteria for determining the applicability of the theory to specific mergers. See R. Posner, Antitrust Law: An Economic Perspective 113-25 (1976). We do not yet accept Professor Posner's despair. The Clayton Act prohibits more than "certainties;" it also prohibits "probabilities," but not "ephemeral possibilities." Marine Bancorporation, 418 U.S. at 622-23, 94 S.Ct. at 2870, 41 L. Ed. 2d at 996-97; Brown Shoe Co. v. United States, 370 U.S. 294, 323, 82 S. Ct. 1502, 1522-23, 8 L. Ed. 2d 510, 534 (1962). Because the Board made only minimal findings, we cannot tell whether more thorough findings can separate probabilities from "ephemeral possibilities."

We join the Second and Fourth Circuits in attempting to set adequate standards of proof in preference to rejecting an admittedly problematic doctrine. See United States v. Siemens Corp., 621 F.2d 499 (2d Cir. 1980); BOC International Ltd. v. FTC, 557 F.2d 24 (2d Cir. 1977); FTC v. Atlantic Richfield Co., 549 F.2d 289 (4th Cir. 1977).16

The findings we require necessarily involve economic prediction,17 but this is not unusual. Many of the findings we call on juries to make are equally speculative; for example, the future loss that an injured person will suffer. The Board's expertise entitles its findings to an additional degree of confidence. We cannot determine whether such findings are inherently too speculative until they have been discussed by the parties and employed by the Board in actual fact finding.

To establish a violation of the doctrine, the Board must make explicit findings as to the degree of concentration in the market, the existence of other potential competitors, the probability that the potential competitor will enter the market independently and the probability that the potential competitor's entry will have a significant procompetitive impact in either El Paso or Waco.18 The Board made an adequate finding only as to the degree of concentration in El Paso and Waco.

"The potential competition doctrine has meaning only as applied to concentrated markets." Marine Bancorporation, 418 U.S. at 630, 94 S.Ct. at 2874, 41 L. Ed. 2d at 1000. If the market is competitive, collusion or conscious parallelism cannot affect the availability of credit and other banking services. An agency refusing to sanction a merger must find that the market is highly concentrated. Siemens, 621 F.2d at 505.

The Board found that the El Paso and Waco markets are highly concentrated.19 The four largest banking organizations in El Paso control 86.1 per cent of the market and in Waco control 73.8 per cent of the market for commercial bank deposits. These high concentration ratios establish a prima facie case of a concentrated market. Marine Bancorporation, 418 U.S. at 631, 94 S.Ct. at 2874, 41 L. Ed. 2d at 1001.20

Mercantile countered only with evidence that the markets were deconcentrating but not with evidence that the markets were now deconcentrated to a significant degree. Mercantile failed to establish a "clear trend to deconcentration;" a small decline does not constitute a clear trend. United States v. Black and Decker Mfg. Co., 430 F. Supp. 729, 751 (D.Md.1976). Even with a stronger trend towards deconcentration, years may pass before the influence of the dominant firms is substantially reduced. The addition of more competition could still have the requisite beneficial effect. We conclude that there was substantial evidence to support the Board's conclusion that the markets are concentrated.

If there are numerous potential competitors waiting in the wings, elimination of Mercantile as one potential entrant would not be significant. FTC v. Atlantic Richfield Co., 549 F.2d 289 (4th Cir. 1977); United States v. First National State Bancorporation, 499 F. Supp. 793 (D.N.J.1980); 5 P. Areeda & D. Turner, supra at P 1123; R. Posner, supra at 113-25; Brodley, supra, 87 Yale L.J. at 75-77. The Board made no findings concerning the effect, if any, of Mercantile's independent entry on the willingness and ability of other potential competitors to compete in the two markets.

Economic theory suggests that, where oligopoly profits are available, a multitude of firms will eagerly seek to enter the market. See R. Posner, supra at 124-25. Barriers to entry, however, will reduce the number of potential entrants. In regulating the banking industry, Congress and the states have created a number of such barriers. For example, federal anti-branching legislation prevents banks and bank holding companies domiciled outside Texas from entering El Paso and Waco.21 A potential competitor must also hurdle more universal barriers, for it must have the necessary capital and managerial resources.

Despite these obstacles, there are likely to be other potential competitors in Texas. Mercantile is the fifth largest Texas bank holding company. While two of the four larger Texas bank holding companies have already entered the El Paso market, no outside bank holding company (besides PanNational) has yet entered Waco. A significant number of large Texas bank holding companies remain as possible entrants for both markets.

On remand, the Board should attempt to identify any other potential entrants into El Paso and Waco. In particular, the Board should consider the other large Texas bank holding companies (or those large Texas banks capable of becoming bank holding companies). The Board should then evaluate whether the number of other potential entrants and the likelihood of their potential entry vitiate Mercantile's importance as a potential competitor. We leave to the Board's expertise the task of developing standards for determining the number and importance of the other potential competitors.

While "approval of the (merger) proposal would do nothing to reduce the concentration of banking resources" in El Paso and Waco (Board Op. at 4), forbidding the merger accomplishes nothing unless Mercantile will, in fact, enter the two markets independently.22 Accordingly, the Board had to determine the likelihood that Mercantile will do so. Siemens, 621 F.2d at 506-07.

The Board did make two necessary preliminary findings. It concluded, on the basis of substantial evidence, that no significant barriers to entry prevent an outsider from entering El Paso and Waco. (The Board noted the previous de novo entry of other banks between 1970 and 1979.) The Board also concluded that Mercantile possesses the financial and managerial resources for independent entry, a conclusion supported by the record.

Having concluded that Mercantile could enter El Paso and Waco, the Board failed to make sufficient findings showing that Mercantile would enter independently. The Board merely stated, in conclusory fashion, that there was a "probability" that Mercantile would enter the markets and that the El Paso market was attractive (although it found the Waco market "less attractive"). These findings were insufficient.

Two circuits have divided over how strong the likelihood must be that the outsider will enter the market independently. In Atlantic Richfield, the Fourth Circuit implicitly endorsed Professor Turner's argument that, if the potential competition doctrine is the only ground offered for interdicting a merger, entry of the outside firm must appear to be certain. 549 F.2d at 294, citing Turner, supra note 15, 78 Harv.L.Rev. at 1384-1386.23 The Second Circuit, in contrast, has insisted on only a "reasonable probability" of procompetitive entry. Siemens, 621 F.2d at 506-07; BOC International, 557 F.2d at 28 & n.7. Neither Falstaff nor Marine Bancorporation illuminate the issue.

We agree with the Second Circuit that certainty is too strict a standard in light of the Congressional command that we prevent anticompetitive "probabilities" as well as certainties. BOC International, 557 F.2d at 28-29 & n.6-7. See Marine Bancorporation; Brown Shoe. "Reasonable probability" alone, however, does not adequately describe a finding that will prevent federal agencies and district courts from mistaking possibilities for sufficient probabilities. A probability signifies that an event has a better than fifty percent chance of occurring. A "reasonable" probability presumably represents an even greater likelihood of the event's occurrence. Unfortunately, the threshold between a probability and a "reasonable" probability is difficult to discern, much less articulate. Any other description of probability (e.g. a "strong" probability or a "serious" probability) is equally likely to founder in ambiguity. We adopt the "reasonable" probability standard, but attempt to mitigate its ambiguity by specifying subsidiary findings that must be compared before the ultimate finding of a reasonable probability is made.24

In essence, the Board must offer a persuasive rationale demonstrating that Mercantile would prefer the opportunity to enter El Paso or Waco over other opportunities for expansion or investment. The Board should consider three important factors in determining whether a persuasive rationale exists.

First, and most important, is the anticipated profitability of independent entry, for a firm's assumptions about profitability, either in the short or long term, will determine the direction in which it expands its operations. On remand, the Board should identify the most relevant indicia of profitability in the banking industry, a task of which it is uniquely capable. We would anticipate, however, that the Board will look to the profits enjoyed by banks currently in the two markets as well as to any special opportunities that Mercantile may itself enjoy either by virtue of being an outsider or because of particular competitive advantages it enjoys. In estimating profitability, the Board must take into account the costs of entry, whether incurred in construction, marketing, compliance with federal and state regulations or in any other expense.

The Board should then examine the profitability (as discounted by any costs of entry) of other opportunities for investment and expansion available to Mercantile. A bank holding company faces more limited opportunities than the economist's universe of other activities. The other large Texas markets that Mercantile has not yet penetrated seem prime candidates for consideration, but Mercantile may present evidence evaluating the profitability of any other opportunities that the firm considers attractive.

If the expected profits from independently entering El Paso or Waco are markedly higher than those expected from other opportunities available to Mercantile,25 then it would be fair to regard Mercantile as a reasonably probable entrant. This is not irrebuttable; Mercantile is free to present any evidence, objective or subjective, showing that it would not fulfill the usual expectation in this instance.

If independent entry into El Paso or Waco offers profits similar to those anticipated from the likely alternatives, then there must be evidence specific to Mercantile demonstrating why it might prefer independent entry. Such evidence might take a variety of forms. Mercantile's past behavior or current marketing strategy may indicate an intent to enter every major market in the state. Internal office memoranda could furnish a sufficient basis for inferring intent. El Paso might serve any need for a presence close to the Mexican border. We do not indicate that any of the above evidence exists, but only describe examples of relevant evidence. Mercantile may counter with any evidence of its own.

In requiring these findings, we do not preempt the Board's discretion to determine for itself the probative weight to be given various items of evidence, objective and subjective. "Objective" evidence is essentially circumstantial evidence. It consists largely of observable economic data, such as evidence of a market's profitability or of the previous behavior of a firm. "Subjective evidence" refers to statements by officers of the outsider firm relevant to the likelihood of the firm's independently entering a market. See 5 P. Areeda & D. Turner, supra at PP 1121b, 1121c (1980).

Mercantile argues that the Board should be compelled to give weight to the statement in Mercantile's application that the firm would not enter El Paso or Waco independently. Justice Marshall noted in his concurring opinion in Falstaff that such statements are probative, but may frequently be self-serving and, therefore, entitled to little weight. 410 U.S. at 565, 93 S.Ct. at 1117, 35 L. Ed. 2d at 500.26 The Board evidently found other evidence more probative of Mercantile's intent, a conclusion well within its discretion in considering proposed bank holding company mergers. We have no grounds for preempting its judgment.

Underlying Mercantile's argument is a protest against the Board's exclusive use of objective economic evidence in concluding that Mercantile was a probable independent entrant. Not only is objective evidence undeniably probative, but subjective evidence is not required to establish a violation of the Clayton Act standard. Falstaff, 410 U.S. at 548, 93 S.Ct. at 408, 35 L. Ed. 2d at 490 (Marshall, J., concurring). On remand, the Board may rely exclusively on objective evidence if that evidence is sufficient to support the findings we require.

4. Significant Procompetitive Effects of Independent Entry

The Board must also find that independent entry would result in "a substantial likelihood of ultimately producing deconcentration of that market or other significant procompetitive effects." Marine Bancorporation, 418 U.S. at 633, 94 S.Ct. at 2875, 41 L. Ed. 2d at 1002. The Board opinion makes no finding that an independent entry by Mercantile would deconcentrate either of the two markets. Instead, it found only that the merger "would do nothing to reduce the concentration of banking resources" and that "denial of the proposal would preserve the probability that (Mercantile) and PanNational will be confronting each other in these concentrated markets in the future."

That the merger would do nothing to reduce present concentration is irrelevant. Congress has condemned only those mergers that produce substantial anti-competitive effects. Brunswick Corp. v. Pueblo Bowl-O-Mat, Inc., 429 U.S. 477, 97 S. Ct. 690, 50 L. Ed. 2d 701 (1977). Congress has not required that firms expand by the most procompetitive means; a competitively neutral acquisition does not violate the Clayton Act standard. Comment, Toehold Acquisitions and the Potential Competition Doctrine, 40 U.Chi.L.Rev. 156 (1972).

The preservation of potential competition between Mercantile and PanNational is a procompetitive effect, but the Board's findings fail to explain the significance of that competition. The Board's opinion also fails to assess the significance of Mercantile's potential competition with other banking institutions already active in the two markets. Under Chenery I, we cannot make that assessment for the Board.

Because the Board evidently thought that only minimal findings were necessary, we do not reverse its decision for failure to show a significant procompetitive impact. Instead, we remand for a more thorough analysis of the relevant economic data.

On remand, the Board may again interdict the proposed merger only if Mercantile's independent entry would have the significant procompetitive effect required by Marine Bancorporation. A small addition to existing competition will not suffice if its absence would not substantially lessen future competition. We recognize that the entry of a large firm into a concentrated market necessarily "shak(es) things up." BOC International, 557 F.2d at 27, citing Turner, supra note 14, 78 Harv.L.Rev. at 1383. We doubt that this effect is sufficient in itself, as it may have no lasting impact.27 We also recognize that, to be encouraged, Mercantile's independent entry need not single-handedly deconcentrate the El Paso or Waco market. Significance is the test.

In evaluating procompetitive effects, the Board is not limited to considering gains for economic efficiency. Another gain from increased competition is the further dispersal of discretionary economic authority. Brodley, supra, 87 Yale L.J. at 33-40. Dominant firms daily make decisions affecting not only their shareholders' profits (the presumed primary motivation), but also the lives of their employees, customers and other members of the community. In any economy predicated on the availability of capital, decisions by banking institutions have an especially far-reaching impact on the community served. These decisions often have social as well as economic implications. The congressional policy underlying the Sherman and Clayton Acts favors a wide dispersal of economic power. United States v. Von's Grocery Co., 384 U.S. 270, 274-77, 86 S. Ct. 1478, 1480-82, 16 L. Ed. 2d 555, 558-560 (1966); United States v. Alcoa Aluminum Co., 148 F.2d 416 (2d Cir. 1945) (L. Hand, J.). See Lewis v. BT Investment Managers, Inc., 447 U.S. 27, 46, 100 S. Ct. 2009, 2020, 64 L. Ed. 2d 702 (1980); United States v. Philadelphia National Bank, 374 U.S. 321, 83 S. Ct. 1715, 10 L. Ed. 2d 915 (1963). Mercantile's substantial assets may enable it to facilitate the dispersal of discretionary economic authority even before it obtains a substantial share of the market. See Brodley, supra, 87 Yale L.J. at 39. We do not predict that Mercantile's assets will have this effect, nor do we suggest that this effect is sufficient to support a finding of a probable significant effect. We merely offer a line of analysis for the Board's consideration.

In fact, Mercantile's independent entry is unlikely to have a significant impact unless Mercantile eventually accumulates a market share sufficient to challenge the dominance of the established firms. See Marine Bancorporation, 418 U.S. at 632-39, 94 S.Ct. at 2875-78, 41 L. Ed. 2d at 1001-1006. Mercantile contends that Texas branching restrictions comparable to those faced by Marine Bancorporation in Spokane, Washington, will prevent Mercantile from accumulating a meaningful share of the market. Rather than make our own finding, we instruct the Board to deal explicitly with this and any other cogent arguments presented by Mercantile when the Board makes its additional findings on remand.

The Board should also consider any evidence showing that Mercantile might become as aggressive a competitor and as virulent an enemy of the status quo if it enters the two markets by merging with PanNational. As was recently noted in United States v. First National State Bancorporation, 499 F. Supp. 793 (D.N.J.1980), the certain procompetitive effects of an aggressive competitor's entrance through merger may outweigh the more speculative effects of independent entry. See Carter, Actual Potential Competition under Section 7 of the Clayton Act, 66 Va.L.Rev. 1485, 1509 (1980).

The Board made no finding as to when Mercantile would actually enter either of the two markets. Mercantile urges that we adopt the Second Circuit's rule that independent entry must be shown likely to occur in the near future. See Siemens, 621 F.2d at 505; BOC International, 557 F.2d at 29. We agree that, as petitioner argued in BOC International, the "degree of uncertainty in any economic prediction becomes unacceptably high as it is projected further and further into the future." 557 F.2d at 29. See 5 P. Areeda & D. Turner, supra at P 1121. While concurring with the Second Circuit that a time limit is desirable to limit speculation, we believe that the time limit should be more directly related to the predication being made.

The amount of time needed for successful entry is relevant only if entry will be so delayed that the structure of the market will have changed by the time Mercantile actually enters. If the market has deconcentrated, Mercantile's pro-competitive efforts will no longer be required by the Clayton Act standard. At some point in the future, the degree of concentration in the market becomes so inherently unpredictable that the entire predictive enterprise should be abandoned out of fairness to the outsider firm. Accordingly, the Board should determine at what point Mercantile would be likely to enter the market. If Mercantile can enter either market within two or three years, we assume (absent contrary evidence) that each market will remain sufficiently concentrated so that Mercantile's procompetitive efforts are legally required. If Mercantile will not enter until much later, then the Board should assess the volatility of the two markets to determine the fairness of predicting whether they will still be concentrated at the time when Mercantile is likely to enter. A market's structure may or may not be predictable for five years; after that point, only the most stagnant market will support the extrapolation of previous data. As a heavily regulated industry, banking would not seem a likely candidate for the seer; regulatory change can quickly alter the structure of the market.28

We have attempted to define comprehensible standards in a difficult area of antitrust law. We hope that our suggested analysis will provide a framework for determining the validity and applicability of the potential competition doctrine. We look forward to any refinements that the parties may develop in the Board's further proceedings below. While we regret delaying the final disposition of Mercantile's application, we believe that further proceedings will better insure that the proper disposition is reached.

To summarize: we hold that the Board may invalidate a proposed bank holding company merger on anticompetitive grounds only if the merger would violate the explicit antitrust standards contained in the 1966 amendments to the Bank Holding Company Act. We remand the application to the Board for the following findings:

1. Whether the existence of other actual potential competitors indicates that the elimination of Mercantile from their ranks will have a significant anticompetitive effect.

2. Whether there is a reasonable probability that Mercantile will enter the El Paso and Waco markets independently if the present application (and merger with any other dominant firm in the two markets) is rejected. This finding should be made only after considering what factors would dispose Mercantile to prefer these opportunities for investment to other likely opportunities.

3. Whether Mercantile's entry de novo or by toehold acquisition will result in ultimate deconcentration of the two markets or in other significant procompetitive effects.

REMANDED.