Susan M. Bechtel v. Federal Communications Commission, Anchor Broadcasting Limited Partnership Galaxy Communications, Inc., Intervenors

Opinion for the Court filed by Circuit Judge STEPHEN F. WILLIAMS. STEPHEN F. WILLIAMS, Circuit Judge: In choosing among mutually exclusive applicants seeking to build and operate a new broadcasting station, the Federal Communications Commission prefers applicants who promise that the station’s owners will participate in its management. An applicant who lost because of this policy now attacks it as arbitrary and capricious. The Commission’s preference for the “integration” of ownership and management originated more than 45 years ago in a rather mild form. “Other significant factors being equal,” declared one decision, “the Commission is inclined to prefer an applicant who intends to manage and operate the proposed station personally rather than to entrust its operation to employees.” Homer Rodeheaver, 12 F.C.C. 301, 307 (1947). In these early days, however, the Commission’s focus was not on integration per se, but on whether an applicant would fulfill his promises to the Commission and be responsive to the broadcasting needs of his community. The Commission recognized that integration was not necessarily the most reliable indicator of these things, and it put little weight on integration when it had other reasons to believe that an applicant would be responsible and responsive. See, e.g., Pilgrim Broadcasting Co., 14 F.C.C. 1308, 1349 (1950). The powerful integration preference that now prevails dates back to a policy statement issued in 1965. See Policy Statement on Comparative Broadcast Hearings, 1 F.C.C.2d 393, 394 (1965) [“1965 Policy Statement ”]; Anchor Broadcasting Limited Partnership, 7 F.C.C.Rec. 4566, 4569 n. 6 (1992). The statement declared it “important per se” for station owners to participate in day-to-day station management. 1965 Policy Statement, 1 F.C.C.2d at 396; see also Central Florida Enterprises, Inc. v. F.C.C., 598 F.2d 37, 56 (D.C.Cir.1978) (holding that Commission policy forbids “functional” inquiry into efficacy of substitutes for integration); Committee for Community Access v. F.C.C., 737 F.2d 74, 82 (D.C.Cir.1984) (same). Three decades later, this case squarely calls into question the validity of continued use of the integration preference. I. Procedural History This litigation arises out of several mutually exclusive applications to construct and operate a commercial FM radio station in Sel-byville, Delaware. In 1989 an administrative law judge awarded the necessary permit to Anchor Broadcasting Limited Partnership and rejected the competing applications of .Susan M. Bechtel, Galaxy Communications, Inc., and another company. Anchor Broadcasting Limited Partnership, 4 F.C.C.Rec. 5687 (ALJ 1989), modified 5 F.C.C.Rec. 2432 (Rev.Bd.1990). Bechtel’s application received little attention because, alone among the four applicants, she did not propose to integrate ownership and management of the new station, and neither the ALJ nor the FCC’s Review Board had authority to discard or modify the Commission’s integration policy. Even when the case reached the Commission itself, her arguments were ignored; the Commission said that attacks on the integration preference “would more appropriately be considered in a rule making proceeding”. Anchor Broadcasting Limited Partnership, 6 F.C.C.Rec. 721, 724 n. 4 (1991). That theory did not fare well on review in this court. We noted that an agency relying on a previously adopted policy statement rather than a rule must be ready to justify the policy “just as if the policy statement had never been issued”, Bechtel v. F.C.C., 957 F.2d 873, 881 (D.C.Cir.1992) [“Bechtel I”] (quoting Pacific Gas & Elec. Co. v. FPC, 506 F.2d 33, 38-39 (D.C.Cir.1974)), and explained that the Commission’s broad discretion to choose between rulemaking and adjudication did not justify its applying an undefended policy in adjudications simply on the basis of a hypothetical future rulemaking. Id. We therefore remanded the case to the Commission for it to address Bechtel’s challenge. Id. at 881-82. The Commission then launched a rulemak-ing proceeding that sought comment on mod *878 ification or elimination of the integration test. Notice of Proposed Rule Making, 7 F.C.C.Rec. 2664, 2665-66 ¶¶ 14-15 (1992). On remand in Bechtel’s own case, however, it neglected Bechtel /’s mandate to “demonstrate why its focus on integration is still in the public interest, if indeed the Commission concludes that it is”, and to respond to Bechtel’s claim that “her proposal ... would serve the public interest better than her competitors’ integrated proposals.” Bechtel I, 957 F.2d at 880. Instead, the Commission narrowed the issue to consideration of whether post-1965 regulatory changes had removed the basis for the policy, and put the burden on Bechtel to show that they had clearly done so. Anchor Broadcasting Limited Partnership, 7 F.C.C.Rec. 4566, 4567 ¶ 12 (1992) [“First Remand Order ”]. Concluding that she had not met that burden, it reaffirmed the prior order. In a later case involving attacks on the integration policy- — attacks that the Commission had brushed off without a word of justification — another panel of this court noted the gap between the Bechtel I remand and the Commission’s performance. See Flagstaff Broadcasting Found. v. F.C.C., 979 F.2d 1566, 1571 (D.C.Cir.1992). The Commission responded by modifying its first remand order and producing the decision now before us. Anchor Broadcasting Limited Partnership, 8 F.C.C.Rec. 1674, 1675 ¶ 12 (1993) [“Second Remand Order ”]. Bechtel has appealed from this modified decision, and Galaxy Communications (another disappointed applicant) has intervened in Bechtel’s appeal pursuant to 28 U.S.C. § 2348. 1 We agree with Bechtel that continued application of the integration preference is arbitrary and capricious, and therefore unlawful. See 5 U.S.C. § 706 (1988). Accordingly, we reverse the Commission’s decision and remand this case to the Commission. On remand, the Commission should conduct a proceeding in which it considers Bechtel’s application (and any other application properly before it) under standards free of the integration preference. II. The Status of Policy Statements Policy statements are exempt from the Administrative Procedure Act’s notiee- and-comment requirements, see 5 U.S.C. § 553(b), and hence may take effect without the rigors — and presumed advantages — of that process. The price to the agency is that the policy “is subject to complete attack before it is finally applied in future cases”. Pacific Gas, 506 F.2d at 39; Panhandle Producers & Royalty Owners Ass’n v. Economic Regulatory Admin., 822 F.2d 1105, 1111 (D.C.Cir.1987). Sooner or later the agency must meet its obligation to respond to criticisms. American Mining Congress v. Mine Safety & Health Admin., 995 F.2d 1106, 1111 (D.C.Cir.1993). This does not mean that policy statements have no effect. Although the agency must respond to challenges and be ready to consider “the underlying validity of the policy itself’, Pacific Gas, 506 F.2d at 39; see also Panhandle, 822 F.2d at 1110; Guardian Fed. Sav. & Loan Ass’n v. FSLIC, 589 F.2d 658, 666 (D.C.Cir.1978), it need not repeat itself incessantly. When a party attacks a policy on grounds that the agency already has dispatched in prior proceedings, the agency can simply refer to those proceedings if their reasoning remains applicable and adequately refutes the challenge. But the agency must always stand ready “to hear new argument” and “to reexamine the basic propositions” undergirding the policy. McLouth Steel Products Corp. v. Thomas, 838 F.2d 1317, 1321 (D.C.Cir.1988); see also FCC v. WNCN Listeners Guild, 450 U.S. 582, 603, 101 S.Ct. 1266, 1278, 67 L.Ed.2d 521 (1981). III. The Commission’s Defense of its Integration Policy As summarized in the Second Remand Order, the Commission attributes three basic advantages to integration, which can be described under the headings “incentives”, “interest”, and “information”. First, owner-managers — because of their direct financial and legal stake in-the station’s performance — have better incentives than mere *879 employee-managers, and hence stations run by integrated owners are more likely to respond to community needs and to comply with Commission rules. Second, integrated owners are more likely than absentee owners to have an active interest in the operation of their stations, and an interested owner tends to improve performance. Third, on-site owners are better positioned than absentee owners to gather relevant information about whether the station is fully satisfying community needs and complying with Commission rules. Second Remand Order, 8 F.C.C.Rec. at 1676 ¶ 15; see also 1965 Policy Statement, 1 F.C.C.2d at 395-96. Finally, recognizing that the integration credit is by no means a perfect predictor of whether a licensee will serve “the public interest, convenience, and necessity” (the' statutory standard for assessing applications, see 47 U.S.C. § 309(a)), the Commission also urges that the integration preference has the advantage of objectivity. A. Some Common Problems We will address these arguments one by one, but first we raise some points that apply equally to all the three claimed substantive advantages. 1. Lack of Permanence Whatever the benefits of integration, they would last only if the Commission insisted on licensees maintaining the owner-manager relation or if successful licensees tended to adopt the integrated structure of their own free will. Neither appears to be the case. Perhaps in recognition of integration’s artificiality, the Commission has done little to ensure its continuation once the promise of . integration has carried an applicant to victory. On the first anniversary of the commencement of program tests, people who have won their station in a comparative hearing must report any deviations from their integration proposals. 47 CFR § 73.1620(g). But as long as they did not misrepresent their intentions in their applications, abandonment of those proposals apparently carries no consequences. See Proposals to Reform the Comparative Hearing Process, 6 F.C.C.Rec. 157 ¶22 (1990); Hulse, Ham, Metzger & Wookey, 7 F.C.C.Rec. 5090, 5095 n. 12 (Rev.Bd.1992) (concurrence). After the first anniversary, moreover, no reports are required. Similarly, while successful applicants in comparative hearings generally cannot transfer or assign their stations during the first year of operations, see 47 CFR § 73.3597(a), thereafter a licensee who had won his station through his integration proposal could “turn around and sell it ... without regard to the buyer’s ‘integration’ or lack thereof’. Bechtel I, 957 F.2d at 880. 2 The Commission, while admitting that it has never actually addressed the issue, suggests that “an applicant proposing integration and having, at the time of the proposal, a present intention to sell the station after one year would not be entitled to integration credit.” First Remand Order, 7 F.C.C.Rec. at 4569 n. 10. But denying integration credit to people who manifest a present intention to sell out quickly — or revoking the licenses of the handful who could after the fact be proved to have misrepresented their intentions — is not the same thing as guaranteeing permanent integration. Indeed, Bechtel has repeatedly challenged the Commission to identify a single instance in which an applicant who won his station on the basis of his integration proposal continued to operate the station as promised for an appreciable period of time. Though such examples surely must exist, the Commission has failed to provide one. According to Bechtel, in fact, the Commission never has made any effort to determine “the actual length of time its ‘integrated’ license winners have owned and operated their broadcast stations”. Brief of Petitioner Susan M. Bechtel at 33. The Commission has launched a rulemak-ing proceeding to consider boosting the one- *880 year holding period for stations won in comparative hearings to three years after the start of operations, and perhaps making parallel changes in the reporting requirement. 8 F.C.C.Rec. 5475 (1993). This proposal would mark a partial return to the policy that prevailed from 1962 until 1982, under which the Commission discouraged efforts to transfer or assign any station held for less than three years. See Procedures on Transfer and Assignment Applications, 32 F.C.C. 689, 691 (1962). The Commission abandoned this policy after concluding that the benefits of a three-year holding period did not outweigh the disadvantages of impeding the flow of resources to their most valued use; in broadcasting as in other businesses, the Commission observed, “important services can be performed by people who trade in broadcast properties, rehabilitate ailing stations with new capital and ideas or relieve unwilling licensees of the responsibility of running a station they no longer want.” Amendment of the Commission’s Rules Regarding Applications for Voluntary Assignments or Transfers of Control, 47 Fed.Reg. 55924 ¶ 28 (1982). And in 1989, in declining to open a proceeding to restore its former “anti-traf-fieking” policy, the Commission said, “The buyer who is willing and able to pay the market price for a given facility would be more likely to deliver the service audiences want than the owner unable or unwilling to continue station operation.” Amendment of Section 73.3597, 4 F.C.C.Rec. 1710, 1710 ¶ 4 (1989). But even if the Commission reaches (and adequately defends) a new conclusion in the pending rulemaking, a three-year holding period would still give it no reason to think that integration proposals will “be adhered to on a permanent basis”, as the Commission contemplated in 1965. See 1965 Policy Statement, 1 F.C.C.2d at 395 n. 6. 3 Bechtel, who proposed to build a station that would serve 21% more people than the facility proposed by the applicant that won the Selbyville permit, argues that “the ephemeral period of initial ownership of a broadcast station ... is vastly outweighed as a public interest factor by the lasting impact of a technical facility which provides greater coverage.... ” Reply Brief of Petitioner Susan M. Bechtel at 1. Since the Commission does not know how long the typical successful applicant adheres to his integrated proposal, it can offer no real response to this argument. 2. Lack of Evidence The Commission’s uncertainty about the practical effects of its integration policy is not limited to the question of how long integration persists. Despite its twenty-eight years of experience with the policy, the Commission has accumulated no evidence to indicate that it achieves even one of the benefits that the Commission attributes to it. As a result, the Commission ultimately rests its defense of the integration criterion on the deference that we owe to its “predictive judgments”. See, e.g., Brief for Respondent at 19, 21-22, 27-28; Second Remand Order, 8 F.C.C.Rec. at 1675-76 ¶ 14. But as Bechtel protests, the relevant predictions have now had almost three decades to succeed or fail. There comes a time when reliance on unverified predictions begins to look a bit threadbare. “The Commission’s necessarily wide latitude to make policy based upon predictive judgments deriving from its general expertise implies a correlative duty to evaluate its policies over time to ascertain whether they work — that is, whether they actually produce the benefits the Commission originally predicted they would.” Bechtel I, 957 F.2d at 881 (citations omitted). What is more, the predictions at the root of the integration policy seem rather implausible. As Bechtel’s counsel observed at oral argument, the fact that corporate America generally does not insist upon the integration of ownership and management easts doubt on the Commission’s rosy speculations about the benefits of integration. 4 Without adopting *881 the Panglossian view that all economic arrangements that exist must necessarily be efficient, one should still be skeptical when regulatory agencies promote organizational forms that private enterprise would not otherwise adopt. At least such skepticism is appropriate when the agencies are trying to accomplish something that is essential to the survival and prosperity of firms in an ordinary market — such as ensuring that a business identifies and fills available market niches, is responsive to its customers, and complies with laws whose violation can get its owners into serious trouble and jeopardize the value of their investment. Finally, it is worth noting that the “predictive judgments” at the root of the integration policy concern an area that the Commission has sometimes considered beyond its expertise. In scrutinizing integration proposals asserted by rival claimants to be illegitimate, purely formal, or otherwise inadequate, the Commission has disclaimed any “particular expertise in finance or business management” and accordingly expressed itself “reluctant to second-guess an applicant’s business judgment — so long as it is, in fact, a good faith business decision”. Victory Media, Inc., 3 F.C.C.Rec. 2073, 2075 ¶ 19 (1988). And later: “[Tjhere is difficulty in having the F.C.C. attempt to define what constitutes ‘good’ management. The Commission is reluctant to impose on applicants any one view of what constitutes a well managed broadcast venture.” Omaha TV 15, Inc., 4 F.C.C.Rec. 730, 732-33 ¶20 (1988). Of course the Commission may believe that the goals that Congress has directed it to pursue in the license allocation process — “the public interest, convenience, and necessity”, 47 U.S.C. § 309(a) — -are in the end completely unmeasurable. Especially given the strictures of the First Amendment, that may be so. But see Red Lion Broadcasting Co. v. F.C.C., 395 U.S. 367, 89 S.Ct. 1794, 23 L.Ed.2d 371 (1969). But the Commission seems not to have taken that position; it seems to believe that it is applying testable hypotheses. On that assumption, its failure over a 28-year period to generate a shred of supporting evidence is rather telling. 3. Exclusion of Other Factors Even if integration’s claimed advantages were more plausible than we find them, they would not necessarily justify the extraordinary weight that the Commission assigns to integration. The Commission has identified “two primary objectives” for its comparative process: generating “a maximum diffusion of control of the media of mass communications” and securing “the best practicable service to the public”. 1965 Policy Statement, 1 F.C.C.2d at 394. In the typical case, the integration criterion is “the most important element of best practicable service”. Hassayampa Broadcasting, 92 F.C.C.2d 472, 475 ¶ 7 (Rev.Bd.1982); cf. Northern Sun Corp., 100 F.C.C.2d 889, 891 ¶ 3 (Rev.Bd.1985) (declaring that “no more need be said” about applicant once it was denied integration credit). In other words, the Commission generally deems an applicant’s integration proposal *882 more important than his past broadcast record, his proposed program service, or the efficiency of his proposed use of frequency. 5 Within the framework of the integration criterion, the Commission does take certain “qualitative” factors into account: an applicant’s integration credit can be enhanced if the proposed owner-managers live in the station’s service area, have participated in civic affairs, have broadcast experience, or belong to a minority group. But the “quantitative” portion of the integration credit tends to swamp the qualitative. The Commission calculates the quantitative portion with a numerical precision that masks the fuzziness of the underlying facts (discussed below). It applies a formula conceived by analogy to the Hirschman-Herfin-dahl Index of antitrust law, under which a proposed owner-manager’s integration score is 10000 x (ownership share) x (hours per week/40)^. An applicant’s overall quantitative score is the sum of the scores of each of its proposed owner-managers. Thus, applicants who get credit for full-time integration of all their owners receive the maximum score of 10,000. Because the formula squares the fraction representing the owner’s proposed working portion of a 40-hour week, part-time management is discounted at a more-than-linear rate; a 100% owner who pi’oposed to devote 30 hours a week to the station would get only 5625 points — (10000) x (%)¿. See Omaha TV 15, Inc., 4 F.C.C.Rec. 730, 734 ¶ 30 (1988). Qualitative factors cannot overcome a “clear” quantitative advantage — which the Commission defines as a difference of at least 1250. Miracle Strip Communications, Inc., 4 F.C.C.Rec. 5064, 5066 ¶ 18 (1989). An applicant whose proposed owner-manager knows nothing about either broadcasting or the community but promises to work a 40-hour week, for example, will handily win an integration preference over one whose proposed owner-manager is a veteran broadcaster who has spent his whole life in the station’s community but proposes to work a 36-hour week at the station (scoring only 8100). See also Cannon Communications Corp., 101 F.C.C.2d 169, 181 ¶ 14 (Rev.Bd.1985) (acknowledging that quantitative predominance “provides a ‘leg-up’ for the sole integrated proprietor over a multi-party applicant entity not proposing integration by all of its members, notwithstanding that the latter may bring more in terms of local residence and civic involvement as well as cultural diversity”). All this occurs under a policy whose stated goal is to pick owners who are aware of and responsive to their communities’ special needs. Of course, comparative hearings turn on other issues when the applicants have similar integration proposals, and the integration preference itself can range from slight to substantial. But as the Review Board reports, “Quite frequently, these days, the quantitative difference in the amount of ownership ‘integration’ credit awarded is all that dispositively separates the winning applicant from the also-rans.” Religious Broadcasting Network, 3 F.C.C.Rec. 4085, 4087-88 ¶7 (Rev.Bd.1988). An applicant that secures a “clear” quantitative advantage in integration (the 1250-point edge) will normally win the station, as long as it meets the Commission’s threshold criteria and does not own other media interests. Given all the factors that affect a station’s performance, the Commission faces a difficult task in justifying this remarkable system. With these points in mind, we address the purported advantages of the integration criterion. B. Incentives 1. Financial Incentives The Commission asserts that stations perform better when managed by those with the “most direct financial interest” in the venture. See Second Remand Order, 8 F.C.C.Rec. at 1676 ¶ 15. The Commission has not defined exactly what it means by “financial interest”. However the term is *883 defined, though, the integration policy does not serve this goal. For instance, in calculating integration credit, the Commission does not take the ownership interests of limited partners into account if all the limited partners are sufficiently insulated from influence over the partnership’s affairs. Anax Broadcasting, Inc., 87 F.C.C.2d 483 (1981). When an applicant is organized as a corporation rather than a partnership, the same is true for the interests of nonvoting shareholders. Accordingly, an applicant can get full integration credit even though the general partner or voting shareholder has only a small percentage of the total equity in the firm, the rest of which is held by people who will have nothing to do with the station. For example, a firm has received full integration credit where the sole general partner held only 10% of the firm’s equity, for which he had paid $10. Independent Masters, Ltd., 104 F.C.C.2d 178, 187-92 (Rev.Bd.1986); see also Religious Broadcasting Network, 3 F.C.C.Rec. at 4103 ¶ 61 (awarding dispositive integration credit where station’s proposed Director of Public Affairs held 14.3% of corporate applicant’s voting stock but only 0.84% of its total equity). See generally WHW Enterprises, Inc., 89 F.C.C.2d 799, 816-17 ¶ 26 (Rev.Bd.1982) (noting that voting share, not equity ownership or capital contribution, is focus of integration criterion). Because the Commission’s method of measuring “ownership” focuses on voting power rather than profit share, the Commission does not insist that the proposed “owner-managers” stand to gain much if the station is especially profitable. 6 Other Commission views further obliterate the purported link to financial incentive. Full integration credit is also available to nonprofit corporations, with the corporate directors being treated as “owners” despite their lack of any equity stake in the venture. Reginald A. Fessenden Educational Fund, Inc., 100 F.C.C.2d 440 (Rev.Bd.1985); Roanoke Christian Broadcasting, Inc., 92 F.C.C.2d 1477 (Rev.Bd.1983). And, confronted with a broadcasting subsidiary of a national mutual insurance company owned by 2.5 million policyholders, the Commission has treated the subsidiary’s officers as its “owners” for purposes of the integration preference. Farragut Television Corp., 8 F.C.C.2d 279, 282 ¶¶ 8-9 (1967). In short, whatever the benefits of ensuring that day-to-day management decisions are made by people whose money is directly on the line, the integration policy does not achieve them. 2. Legal Accountability The Commission also considers it “inherently desirable” that day-to-day management decisions be made by people with “legal responsibility” for the station. 1965 Policy Statement, 1 F.C.C.2d at 395. But station employees can be held legally responsible for their acts even if they don’t own the station. Thus, it is a crime for anyone to “utter[ ] any obscene, indecent, or profane language by means of radio communication”, 18 U.S.C. § 1464, or for a station employee to accept money for broadcasting anything without first disclosing the payment to the station, 47 U.S.C. § 508. More broadly, willful and knowing commission of any act prohibited by the chapter of the Communications Act dealing with "radio communication can result in imprisonment and a fine of up to $10,000, id. § 501, and “[a]ny person who willfully and knowingly violates any rule, regulation, restriction, or condition made or imposed by the Commission under authority of this chapter” faces an extra fine of up to $500 “for each and every day during which such offense occurs”, id. § 502. It may be true, as the Commission suggests, that station owners have the most legal accountability for the station. See Second Remand Order, 8 F.C.C.Rec. at 1676. But to a large extent they have this accountability whether or not they work at the station. Stations are not insulated from the *884 threat of license revocation, 47 U.S.C. § 312, or nonrenewal, id. § 307, merely because they are owned by absentee investors. Nor does absentee ownership protect stations from the forfeiture provisions of 47 U.S.C. § 503(b). Since absentee owners thus have strong incentives to ensure that their station complies with the relevant statutes and rules, the incremental contribution of the integration preference on this score appears trivial. C. Interest The Commission also asserts that integrated owners are more likely than absentee owners to have an active interest in the operation of their stations. See Second Remand Order, 8 F.C.C.Rec. at 1676. The Commission does not explain why it believes that an interested owner tends to improve station performance, but perhaps this belief rests on the proposition that owners have more clout than mere employees, and so they can do more good than employees who must clear decisions with an apathetic or hard-to-reach owner. At the outset, one might question the Commission’s decision to stress an owner’s interest in station operations over other factors, such as his skill or his experience. In most cases, broadcast experience merits at most a “qualitative” enhancement of an applicant’s integration score. 7 Yet even assuming that owner-manager integration is a perfect yardstick for the manager’s interest, it is hard to see why a relatively modest differential (12.-5%, or 1250 out of 1Ó000) in “interest” should overwhelm a substantial difference in experience. Although the Commission has argued that broadcast experience should be “of minor significance” because it can come with time, 1965 Policy Statement, 1 F.C.C.2d at 396, it is hard to imagine that anyone seriously interested in “picking winners” would so heavily downgrade the contestants’ track records. In any event, the integration criterion simply measures one form in which owners may express their interest. In general, integration credit is available only for people who hold day-to-day jobs at the station, not for those who make management decisions from afar while running other business activities too. Yet executives routinely supervise a variety of firm activities, a few hours to each, without being the least bit apathetic about the performance of any. Likewise, no integration credit is awarded for proposals to spend less than 20 hours a week at a station, see, e.g., Midwest Broadcasting Co., 70 F.C.C.2d 1489, 1494-95 ¶9 (Rev.Bd.1979); Woods Communications Group, 7 F.C.C.Rec. 78, 79 ¶ 4 (1992), even though an owner who spends each morning at his station is far from passive or inaccessible. Further, the very existence of the integration criterion weakens any correlation between a proposed owner’s integration and his real interest in the station. It may well be that before the Commission began dispensing permits on the basis of integration proposals, only owners interested in broadcast operations would propose to work at the station. But now that applicants have every incentive to declare an intention to manage (and need honor that declaration for only a year at most), the empirical inference is severely weakened. D. Information According to the Commission, on-site owners have better sources of information than *885 absentee owners. By virtue of their presence at the station, the Commission asserts, integrated owners “necessarily have been in a better position than absentee owners” to learn that the station is violating Commission rales or that people have asked the station to address particular community needs. For example, integrated owners are more likely than absentee owners to follow station correspondence or to hear comments by station visitors. Second Remand Order, 8 F.C.C.Rec. at 1676 ¶ 15. Galaxy labels “sheer myth” the notion that “people throng like pilgrims to present their views in person at the nearest broadcast studio”. Reply Brief of Petitioner Galaxy Communications, Inc. at 4. The Commission cites no evidence that station visitors are a major source of information for broadcasters, and the idea seems implausible. Correspondence may be a more likely source of information about community needs than station visitors. But the Commission evidently does not take its own argument very seriously, for it insists that owner-managers spend at least 20 hours a week at the station in order to receive any integration credit. It is hard to see why picking up the week’s influx should take more than a few minutes, and, even assuming an avalanche of mail and only semi-competent staff work, why mastering its content should take more than a few hours. Familiarity with a community seems much more likely than station visitors or correspondence to make one aware of community needs. But even long-time local residence generates at most a “qualitative” enhancement of an applicant’s integration credit. E. Objectivity Aside from the substantive advantages that the Commission attributes to integration, the Commission claims a procedural advantage. As a “structural” factor, it says, the integration criterion can be applied more consistently and objectively than other possible ways of assessing an applicant’s likely responsiveness to community needs. Second Remand Order, 8 F.C.C.Rec. at 1676. Any “objectivity” added by the integration criterion is unfortunately illusory. The Commission’s scores for quantitative integration merely lend the policy a veneer of precision; every step towards the magic number is packed with subjective judgments, some generic, some ad hoe. At a generic level, the Commission’s weighting system, by which it squares the fraction representing the owner’s proposed working portion of a 40-hour week, is simply a fancy way for the Commission to express its view that the values of integration fall off sharply when owner-managers work less than full-time. See Omaha TV 15, Inc., 4 F.C.C.Rec. 730, 734 (1988). It has not a shred of data supporting the basic conjecture, let alone that the relationship is best captured by squaring the fraction rather than cubing it. Similarly, the Commission’s formula seems to assume that the importance of ownership share varies on a linear basis, disregarding issues of control that might seem critical. Thus, if one applicant proposes the full-time integration of its 55% owner and the other applicant proposes the full-time integration of its 45% owner, neither applicant enjoys a “clear” advantage even though only the former owner-manager is assured of controlling station policy. But a 13% difference in the ownership shares of two competing full-time owner-managers translates into a decisive 1300 spread even where both have control (e.g., 75% and 88%), and thus will nullify any qualitative differences. Applying the formula requires more subjective decisions. In calculating the hours to be worked by each applicant-owner, for example, the Commission usually takes the prospective owner’s word at face value, even though it knows that the promise is likely made in large part to please the Commission, that the Commission will do little or nothing to enforce the promise, and that the promise therefore may quite possibly be bogus. For the most part, only if a rival makes out an affirmative case that the promise is unreliable will the Commission bother to consider the problem — and, necessarily, it does so then subject to the usual shortcomings of *886 anyone seeking to penetrate the mysteries of the human heart. Further, many cases turn on whether a proposed job should be classified as part of station “management”. Compare, e.g., Merrimack Valley Broadcasting, Inc., 92 F.C.C.2d 506, 511 ¶ 8 (Rev.Bd.1982) (finding Minority Affairs Director, Women’s Affairs Director, and Community Affairs Director all managerial positions) with Religious Broadcasting Network, 3 F.C.C.Rec. 4085, 4100 ¶ 49 (Rev.Bd.1988) (finding Director of Public and Community Affairs a non-managerial position because occupant would not set any station policies). Not surprisingly, Merrimack makes no mention of policy-making, and Religious Broadcasting makes no mention of Merrimack. So far as measuring ownership is concerned, the Commission’s policy provides rich incentives for the adoption of firm structures that we characterized in Bechtel I (quoting Bechtel) as “strange and unnatural”. 957 F.2d at 880. After all, if a station can be acquired for legal fees and minor engineering services, and can be sold a year later for several million dollars, one would expect to see a good deal of ingenuity. In Bechtel I we recited some of the startling arrangements manifested just in this case itself: In our case, for example, best friends and co-owners of a station swear not to consult with each other; family members with valuable broadcast knowledge and experience agree not to assist the tyro station manager in the family; people with steady jobs and families in one city pledge to leave them and move permanently to another; and wealthy retirees promise to move to and work in small summer towns in Delaware with which they have no former connection. Id. In the Notice of Proposed Rule Making triggered by Bechtel I, the Commission itself acknowledged that comparative hearings “often appear to become bogged down in litigating subjective or trivial distinctions.” 7 F.C.C.Rec. 2664, 2665 ¶9 (1992). It also observed, accurately enough we think, that “[examination of potentially unreliable proposals can be a time-consuming and uncertain process”. See id. at 15. Even in its brief here, while at one moment advancing objectivity as a major advantage of the integration preference, the Commission answers Bechtel’s argument about the impermanence of integration by suggesting that integration credit is denied when an applicant has a “present intention” to sell his station quickly, Brief for Respondent at 19, and touts its ability to detect when integration proposals are “unreliable”, id. at 9, or “irreconcilable with the exercise of sound business judgment”, id. at 17 (quoting Royce Int’l Broadcasting, 5 F.C.C.Rec. 7063, 7065 n. 10 (1990)). Reading the record in this case, and in a range of other cases, we are bound to say that only the Commission’s lamentations about undue subjectivity have the ring of truth. We aré quite aware that the Commission’s task is a difficult one. As it is charged with handing out extremely valuable resources, the number of parties lined up to win them inevitably will greatly exceed the supply. 8 Yet common sense, not to mention the First Amendment, counsel against the Commission’s trying to decide what America should see and hear over the airwaves. Further, the ability to pick persons and firms who will be “successful” at delivering any kind of services is a rare one, however success might be defined; that is why it commands generous rewards in the market. The Commission has often implicitly recognized the difficulty, noting the advantages of allowing stations to shift by voluntary transactions into what are presumptively more capable hands (and if not, to shift again). See supra at 880. *887 Moreover, any effort to infer some sort of recipe for success by studying existing owners would be inadequate: if success could be captured in a formula, the skill of identifying future successes would not be so scarce and well rewarded. Any sort of recipe that could be discerned would necessarily abstract criteria from a complex web of facts; applicants would immediately start to adopt the specified ingredients solely to satisfy the Commission, and would feign them, so that their earlier predictive value would decline. Changing exterior circumstances would also undermine the recipe. And any effort to rely on intuitive judgments about applicants, assuming Commissioners had sound intuition, would provide rich opportunities for graft and corruption in a public agency dispensing valuable resources. All these difficulties flow from the statutory scheme itself. All that said, the integration preference is peculiarly without foundation. While the Commission makes it a central focus of allocation, the Commission takes no interest whatever in the matter when it comes to transfers or even in the continuing conduct of the original licensee. The Commission appears to have no evidence that the preferred structure even survives among the winners, much less that it does so among especially outstanding broadcasters. Because of applicants’ incentive to create a facade of integration, and the difficulty of identifying sound business practices, even the preference’s touted objectivity proves an illusion. Though we owe substantial deference to the Commission’s expertise, we are forbidden to suspend our disbelief totally. We find the integration policy arbitrary and capricious. IV. Remedy At times, the Commission has suggested that the pendency of a rulemaking designed to reconsider the integration criterion permits the Commission to continue applying the criterion in the meantime. See, e.g., Second Remand Order, 8 F.C.C.Rec. at 1676 ¶ 16; First Remand Order, 7 F.C.C.Rec. at 4567 ¶ 12. If a policy is arbitrary and capricious, however, the mere fact that the Commission is reconsidering that policy does not authorize the Commission to continue making arbitrary and capricious decisions. As Bechtel was denied a license on the basis of an arbitrary and capricious policy, she is entitled to a proceeding in which the Commission considers her application (and any other application properly before it) under standards free of that policy. As for the scope of the proceedings on remand, Bechtel asks us to direct the Commission to compare her application with that of Anchor Broadcasting, the applicant to which the Commission awarded the Selbyville permit initially. Galaxy wants to be included too. While Bechtel and Anchor obviously should be involved in the upcoming comparative process, we do not decide whether Galaxy (or any other parties) should also be considered. The only Selbyville applicant who has consistently challenged the integration policy is Bechtel. Which applicants the Commission should consider on remand (other than Bechtel) is not properly before us, but is a question for the Commission to decide in the first instance. The Commission’s decision is reversed and the matter is remanded for further proceedings consistent with this opinion. So ordered. Footnotes [1] Galaxy also has appealed in its own right in one of the cases consolidated here, but it now follows Bechtel's lead by focusing entirely on the Commission's integration policy. [2] Transfers are subject to Commission approval, but the Commission’s inquiry is limited to whether the proposed transfer will serve "the public interest, convenience, and necessity”; the Communications Act bars the Commission from holding a comparative hearing to assess whether transfer to another person might better serve the public interest. See 47 U.S.C. § 310(d). Since integration is not a threshold qualification, stations can readily be transferred to nonintegrated owners. [3] The proposition that ownership integration can be maintained only by persistent Commission monitoring and staunch frustration of station transferability suggests that it is highly antithetical to optimal patterns of station ownership. See infra at 880-881, 885-886. [4] The Commission attempts to play on the same theme, asserting that McDonald's requires its *881 franchisees to participate in the day-to-day operations of their restaurants. Brief for Respondent at 23 n. 9. But the available empirical data about franchising support Bechtel. The most recent comprehensive study of fast-food franchise contracts is now more than two decades old, and may pre-date the McDonald’s requirement. At least as of 1971, though, only 12% of franchise contracts in the fast-food industry required the franchisee to be a full-time manager. See Gillian K. Hadfield, "Problematic Relations: Franchising and the Law of Incomplete Contracts”, 42 Stan.L.Rev. 927, 943 (1990). Even within the McDonald's chain, the average franchisee owns two or three outlets. Kathleen De-veny et al., "McWorld?”, Business Week, Oct. 13, 1986, p. 78. In addition, in 1991 more than 15% of the restaurants were company-owned rather than franchised, and hence were not operated by owner-managers. See Lois Therrien, "McRi-sky”, Business Week, Oct. 21, 1991, p. 114. And despite some recent movement toward flexibility, McDonald's still micromanages the conduct of its franchisees in excruciating detail, even specifying the order in which condiments should be placed on hamburgers. Id. (citing 600-page operations manual distributed to franchisees). We do not understand this to be the Commission’s model for its relation to licensees. In any event, to focus solely on franchises (let alone on the franchises of one company) is to miss the diversity of organizational structures in the manufacturing sector and in the rest of the retail sector. The prevalence of vertical integration in corporate America belies any notion that the Commission is merely enforcing well-accepted principles of business organization. [5] The integration preference becomes secondary in certain unusual circumstances. An applicant who proposes superior coverage in areas that are currently underserved, for instance, can prevail despite an opponent's substantial integration preference. See, e.g., FBC, Inc., 95 F.C.C.2d 256, 261-62 ¶¶ 13-14 (Rev.Bd.1983). [6] Likewise, the Commission does not insist that the owner-managers stand to lose much if the station fails. The limited liability of corporate shareholders is no obstacle to integration credit for corporate applicants, and the general partner in a partnership applicant may be a corporation. See, c.g., Louisiana Super Communications Limited Partnership, 102 F.C.C.2d 1293, 1297-1301 (Rev.Bd.1985). [7] Indeed, even the maximum qualitative credit for experience is small. Northern Sun Corp., 100 F.C.C.2d 889, 892 (Rev.Bd.1985); see also New Continental Broadcasting Co., 96 F.C.C.2d 544, 547 (Rev.Bd.1983) (applicant loses even though its proposed integrated owner had 31 years of broadcast experience and the other applicant’s principals had no broadcast experience at all). Under extremely limited circumstances, past broadcast record can be a separate comparative criterion rather than a minor enhancement of the integration criterion. If a proposed owner-manager of the new station has been an owner-manager (rather than a mere employee) of a station in the past, and if that station's performance was extraordinary rather than merely competent, then the Commission will assess whether the reasons that may have accounted for that extraordinary performance would also apply at the new station. See 1965 Policy Statement, 1 F.C.C.2d at 398. Of course, if the owner-manager has retained his ownership interest in the old station, he may suffer a crippling demerit under the Commission's policies promoting diversification of control of the media. See id. at 394. [8] The Commission has expressed dismay at the notion that people might enjoy lavish pecuniary returns by selling a station without operating it for very long. Indeed, it has described the one-year holding period of 47 CFR § 73.3597 as a rule that "prohibits licensees from profiting on the sale of ... a license for a station that has been on the air less than a year." Rebecca Radio of Marco, 4 F.C.C.Rec. 830, 832 ¶ 20 (1989). It is unclear why these windfalls are more distressing when highlighted by prompt sale than in their ordinary occurrence — every time the Commission issues a license.