Full opinion text
OPINION OF THE COURT AMBRO, Circuit Judge TABLE OF CONTENTS I. Background... 37 II. Jurisdiction and Standard of Review. . .39 A. Delayed agency action... 39 B. Final agency action... 40 III. The Eligible Entity Definition... 40 A. The FCC’s objectives.. .40 B. Prometheus I and its aftermath. . .42 C. Prometheus II.. .43 D. Aftermath of Prometheus II.. .45 E. Discussion.. .48 IV. Failure to Complete Quadrennial Review. . .50 A. The costs of delay.. .51 B. Vacatur is not appropriate... 52 C. Deregulatory Petitioners have waived other forms of relief.. .53 V. The Joint Sales Agreement Rule.. .54 A. Background.. .54 B. Discussion... 57 VI. Conclusion... 60 Twelve years have passed since we first took up challenges to the broadcast ownership rules and diversity initiatives of the Federal Communications Commission (“FCC” or “Commission”). In some respects the Commission has made progress in the intervening years. In key areas, however, it has fallen short. These shortcomings are at the center of this dispute— the third (and likely not the last) round in a protracted battle over the future of the nation’s broadcast industry. Specifically, the parties present challenges to the Commission’s “eligible entity” definition, its Quadrennial Review process, and its rule on television joint sales agreements. Although courts owe deference to agencies, we also recognize that, “[a]t some point, we must lean forward from the bench to let an agency know, in no uncertain terms, that enough is enough.” Public Citizen Health Research Group v. Chao, 314 F.3d 143, 158 (3d Cir. 2002) (emphasis and internal quotation marks omitted). For the Commission’s stalled efforts to promote diversity in the broadcast industry, that time has come. We conclude that the FCC has unreasonably delayed action on its definition of an “eligible entity” — a term it has attempted to use as a lynchpin for initiatives to promote minority and female broadcast ownership — and we remand with an order for it to act promptly. Equally troubling is that nearly a decade has passed since the Commission last completed a review of its broadcast ownership rules. These rules lay the groundwork for how the broadcast industry operates and have major implications for television, radio, and newspaper organizations. Although federal law commands the Commission to conduct a review of its rules every four years, the 2006 cycle is the last one it has finished; the 2010 and 2014 reviews remain open. Several broadcast owners have petitioned us to wipe all the rules off the books in response to this delay — creating, in effect, complete deregulation in the industry. This is the administrative law equivalent of burning down the house to roast the pig, and we decline to order it. However, we note that this remedy, while extreme, might be justified in the future if the Commission does not act quickly to carry out its legislative mandate. Whereas the first two issues before us involve agency delay, the third is a challenge to agency action. The Commission regulates the number of television stations a company can own. In 2014, it determined that parties were evading its ownership limits through the influence exerted by advertising contracts known as joint sales agreements. As a result, it created a rule designed to address this perceived problem. However, we conclude that the Commission improperly enacted the rule; hence we vacate it and remand the matter to the Commission. I. Background This is the third volume in a long-running saga that has its roots in the Telecommunications Act of 1996, Pub. L. No. 104-104, 110 Stat. 56 (1996). Section 202(h) of that law imposes on the Commission what was initially a biennial — and is now a quadrennial — obligation to examine its broadcast ownership rules. Broadly speaking, the purpose of these rules is to limit consolidation in the industry by capping the amount of common control that is permissible. Section 202(h) provides that the Commission “shall” review its rules on broadcast ownership every four years, “shall determine whether any of such rules are necessary in the public interest as the result of competition,” and “shall repeal or modify any regulation it determines to be no longer in the public interest.” Id. § 202(h), 110 Stat. at 111-12, as amended by Pub. L. No. 108-199, § 629, 118 Stat. 3, 99-100 (2004) (making review quadrennial as opposed to biennial). Our first foray into § 202(h) came in 2004 when we ruled on challenges to the Commission’s 2002 review of its ownership rules. See Prometheus Radio Project v. FCC, 373 F.3d 372 (3d Cir. 2004) (“Prometheus I”). Though § 202(h) is limited to a review for whether ownership rules remain necessary in light of competition in the broadcast industry, Prometheus I also involved a challenge to the Commission’s efforts to promote minority and female broadcast ownership. Following our decision in Prometheus I, the Commission set about fine-tuning its minority and female ownership initiatives at the same time that it conducted its 2006 Quadrennial Review. In December 2007, it adopted two orders: the first completing the 2006 review and the second implementing diversity-related efforts. (Though adopted in late 2007, they were released in early 2008.) The former has become known as the “2008 Order.” See 2006 Quadrennial Regulatory Review — Review of the Commission’s Broadcast Ownership Rules and Other Rules Adopted Pursuant to Section 202 of the Telecommunications Act of 1996, Report and Order and Order on Reconsideration, 23 F.C.C.R. 2010 (Dec. 18, 2007). Meanwhile, the latter is called the “Diversity Order.” See Promoting Diversification of Ownership in the Broadcasting Services, 2006 Quadrennial Regulatory Review — Review of the Commission’s Broadcast Ownership Rules and Other Rules Adopted Pursuant to Section 202 of the Telecommunications Act of 1996, Report and Order and Third Further Notice of Proposed Rulemaking, 23 F.C.C.R. 5922 (Dec. 18, 2007). In 2011, we decided challenges to both the 2008 Order and the. Diversity Order. See Prometheus Radio Project v. FCC, 652 F.3d 431 (3d Cir. 2011) (“Prometheus II”). In Prometheus I and Prometheus II we reviewed challenges to completed § 202(h) review cycles. But, in the aftermath of Prometheus II, the process broke down. The 2010 Quadrennial Review, which was underway at the time of Prometheus II, is still not complete. It was rolled into the 2014 review cycle, which remains ongoing. With its Quadrennial Review process backed up and our remand instructions from Prometheus II unfulfilled, the Commission issued a combined order and notice of proposed rulemaking in 2014. See 2011 Quadrennial Regulatory Review— Review of the Commission’s Broadcast Ownership Rules and Other Rules Adopted Pursuant to Section 202 of the Telecommunications Act of 1996, Further Notice of Proposed Rulemaking and Report and Order, 29 F.C.C.R. 4371 (Mar. 31, 2014) (“20U FNPRM & Order”). That brings us to the current dispute. In May 2014, Prometheus filed in our Court a petition for review of the 20H FNPRM & Order. Shortly afterward, Howard Stirk Holdings, LLC; the National Association of Broadcasters; and Nexstar Broadcasting, Inc. filed separate petitions for review in the D.C. Circuit Court of Appeals. The FCC filed a notice of multi-circuit petitions pursuant to 28 U.S.C. § 2112(a)(3), and the Judicial Panel on Multidistrict Litigation randomly selected the D.C. Circuit as the venue for the petitions. That Court consolidated them, and the petitioners and various intervenors fully briefed the case. (Consistent with our past usage, we refer . to petitioners and intervenors raising anti-deregulatory challenges as “Citizen Petitioners” and to groups seeking deregulation as “Deregulatory Petitioners.”) Citizen Petitioners argued that the D.C. Circuit should transfer the case to us because the petitions related, at least in part, to Prometheus II. Our decision there ended in a remand, and we retained jurisdiction over the remanded issues. Prometheus II, 652 F.3d at 472. Deregulatory Petitioners opposed transfer, maintaining that most of the issues in front of the D.C. Circuit were independent of Prometheus II. By order dated November 24, 2015, that Court transferred the consolidated petitions to us, and we heard argument on April 19, 2016. II. Jurisdiction and Standard of Review These petitions present challenges to both agency delay (the eligible entity definition and the failure to complete the Quadrennial Review) and agency action (the joint sales agreement rule). These two categories have different jurisdictional hooks and standards of review. A. Delayed agency action Because circuit courts have exclusive jurisdiction over the FCC’s final rule-making action, see infra Section B, we also have exclusive jurisdiction to review claims of agency inaction. See Oil, Chemical & Atomic Workers Union v. Occupational Safety & Health Admin., 145 F.3d 120, 123 (3d Cir. 1998) (citing Telecommunications Research & Action Ctr. v. FCC, 750 F.2d 70, 75 (D.C. Cir. 1984)). This serves a commonsense purpose: “Because the statutory obligation of a Court of Appeals to review on the merits may be defeated by an agency that fails to resolve disputes, a Circuit Court may resolve claims of unreasonable delay in order to protect its future jurisdiction.” Telecommunications Research & Action Ctr., 750 F.2d at 76. Our standard of review comes from Administrative Procedure Act (“APA”) § 706(1), which allows us to “compel agency action unlawfully withheld or unreasonably delayed.” 5 U.S.C. § 706(1). Under this provision, our “polestar is reasonableness.” Public Citizen Health Research Group, 314 F.3d at 151. We must “balance the importance of the subject matter being regulated with the regulating agency’s need to discharge all of its statutory responsibilities under a reasonable timetable.” Oil, Chemical & Atomic Workers Union, 145 F.3d at 123. We have held that, [w]ith this balance in mind, unreasonable delay should be measured by the following factors: First, the court should ascertain the length of time that has elapsed since the agency came under a duty to act. Second, the reasonableness of the delay should be judged in the context of the statute authorizing the agency’s action. Third, the court should assess the consequences of the agency’s delay. Fourth, the court should consider any plea of administrative error, administrative inconvenience, practical difficulty in carrying out a legislative mandate, or need to prioritize in the face of limited resources. Id. (internal quotation marks omitted). We refer to these here as the Oil, Chemical & Atomic Workers Union factors. B. Final agency action Circuit courts have exclusive jurisdiction over final FCC rulemaking under 47 U.S.C. § 402(a) and 28 U.S.C. § 2342(1). Per § 706(2) of the APA, we can set aside agency action that is arbitrary or capricious. 5 U.S.C. § 706(2). “The scope of review under the ‘arbitrary and capricious’ standard is narrow and a court is not to substitute its judgment for that of the agency.” Motor Vehicle Mfrs. Ass’n of U.S., Inc. v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43, 103 S.Ct. 2856, 77 L.Ed.2d 443 (1983). Despite this deference, we require the agency to “examine the relevant data and articulate a satisfactory explanation for its actionf,] including a rational connection between the facts found and the choice made.” Id. (internal quotation marks omitted). When the FCC conducts a Quadrennial Review under § 202(h), that provision also affects our standard of review, as it requires that “no matter what the Commission decides to do to any particular rule — retain, repeal, or modify (whether to make more or less stringent) — it must do so in the public interest and support its decision with a reasoned analysis.” Prometheus I, 373 F.3d at 395. We have also concluded that, when § 202(h) refers to rules being “necessary,” that term means “useful,” “convenient,” or “helpful.” Id. at 394. III. The Eligible Entity Definition The FCC, in line with its statutory obligation to promote minority and female broadcast ownership, has attempted to give preferences to certain “eligible entities.” Citizen Petitioners support the Commission’s objectives but argue that its definition of eligible entities has not been helpful to minority and female ownership. They contend that the Commission, in failing to adopt a better definition, has violated our instructions from Prometheus I and Prometheus II. They ask us to grant them relief under APA § 706(1) or through a writ of mandamus. For its part, the Commission says that it has tried to comply with our previous two remands but that it lacks the administrative record necessary to implement a different definition. However, this conclusion is tentative, and the Commission has yet to determine in a final order that it cannot do what Citizen Petitioners ask. A. The FCC’s objectives The Commission has a statutory obligation to promote minority and female broadcast ownership. For instance, Congress has provided, in the context of applications for licenses or construction permits, that a “significant preference shall be granted to any applicant controlled by a member or members of a minority group.” 47 U.S.C. § 309(i). And for licenses and permits that the Commission awards using competitive bidding, one of its objectives must be promoting opportunities for “businesses owned by members of minority groups and women.” Id. § 309(j). As early as 1978, the FCC recognized that the inadequate representation of minorities in the broadcast industry was “detrimental not only to the minority audience but to all of the viewing and listening public.” Statement of Policy on Minority Ownership of Broadcasting Facilities, 68 F.C.C.2d 979, 980-81 (May 25, 1978). Twelve years later, in Metro Broadcasting, Inc. v. FCC, 497 U.S. 547, 110 S.Ct. 2997, 111 L.Ed.2d 445 (1990), the Supreme Court upheld under intermediate scrutiny two types of racial preferences designed to remedy this problem. Id. at 566, 110 S.Ct. 2997. Not long afterward, however, the Supreme Court reversed course and held that “all racial classifications, imposed by whatever federal, state, or local governmental actor, must be analyzed by a reviewing court under strict scrutiny.” Adarand Constructors, Inc. v. Peña, 515 U.S. 200, 227, 115 S.Ct. 2097, 132 L.Ed.2d 158 (1995). Adarand did not affect gender preferences, which received intermediate scrutiny prior to that decision and have continued in its aftermath to be subject to that standard. See, e.g., Mississippi Univ. for Women v. Hogan, 458 U.S. 718, 724, 102 S.Ct. 3331, 73 L.Ed.2d 1090 (1982); Tuan Anh Nguyen v. INS, 533 U.S. 53, 60, 121 S.Ct. 2053, 150 L.Ed.2d 115 (2001). Following Adarand, one of the Commission’s major initiatives for fulfilling its statutory obligation has been to give “eligible entities” certain preferences. This dates back to the 2002 § 202(h) review cycle. The Commission had established new caps on radio station ownership but agreed to a “grandfathering” provision that would allow companies that already owned stations in excess of the limits to continue to do so, with the caveat that they were not allowed to transfer the grandfathered combinations. The FCC concluded during the 2002 cycle that, notwithstanding the prohibition, licensees could transfer grandfathered combinations to eligible entities. See 2014 FNPRM & Order ¶ 264 (summarizing background). At the time, the Commission said that an eligible entity was one that qualifies as a small business under the revenue-based definition used by the Small Business Administration (“SBA”). Id. Later, the Commission expanded its use of the eligible entity definition by applying it in other contexts. For instance, its rules have aimed to give these entities financing preferences, provide them with construction extensions, and generally encourage them to own broadcast operations. See Prometheus II, 652 F.3d at 442 n. 9. The Commission, at least until recently, has defended its revenue-based definition as a legally supportable means of promoting minority and female ownership because the criteria are race and gender neutral. However, the main criticism has been that there is a lack of evidence showing that small businesses are more likely to be owned by minorities or females. See Prometheus II, 652 F.3d at 470. Several parties have therefore asked the FCC to use a definition that gives preferences to “socially and economically disadvantaged businesses” (“SDBs”). B. Prometheus I and its aftermath In Prometheus I we remanded portions of the Quadrennial Review proceedings to the FCC, but we rejected as premature a challenge to the revenue-based eligible entity definition. We credited the FCC’s contention that, “because of pending legislation, the definition of SDBs is currently too uncertain to be the basis of its regulation.” Prometheus I, 373 F.3d at 428 n. 70. However, we noted our anticipation that “by the next quadrennial review the Commission will have the benefit of a stable definition of SDBs, as well as several years of implementation experience, to help it reevaluate whether an SDB-based [definition] will better promote the Commission’s diversity objectives.” Id. Our anticipation turned out to be overly optimistic, and to date the Commission still has not employed any alternative definitions for eligible entities. The next Quadrennial Review after Prometheus I was the 2006 cycle. In connection with this review, the Commission released a notice of proposed rulemaking in the summer of 2006. This notice, however, sidestepped questions over how to define SDBs. Commissioner Copps argued at the time that the notice’s failure to address minority and female ownership violated our instructions from Prometheus I. See Prometheus II, 652 F.3d at 466 & n. 33. The next year, the FCC issued another notice of proposed rulemaking that acknowledged the failings of the previous notice. It took note of two arguments that supporters of an SDB-based definition raised against the first notice: 1) that “the concept of SDBs is central to most ... minority ownership initiatives” and 2) that “without a definition for SDBs, the Commission cannot effectively” comply with Prometheus I. Id. at 467 (internal quotation marks omitted). Notwithstanding these contentions, the second notice “did not provide descriptions of any existing proposals for SDB definitions ... or discuss the Commission’s analysis of existing briefing on those proposals’ constitutionality or efficacy.” Id. Ultimately, the second notice did not do much with respect to the eligible entity definition. It “merely called for general comment on [the] proposal that the Commission define SDBs for purposes of analyzing policy initiatives in support of media ownership diversity.” Id. (internal quotation marks omitted). At the same time it overhauled a number of its rules in the 2008 Order based on the 2006 Quadrennial Review, the Commission issued the Diversity Order. That order, however, did not consider proposed SDB-based definitions that had been submitted to the Commission. Id. at 468. A large part of the problem was inadequate data. An independent review concluded that “all the researchers (and the peer reviewers) agree that the FCC’s databases on minority and female ownership are inaccurate and incomplete and their use for policy analysis would be fraught with risk.” Id. (internal quotation marks omitted). Two Commissioners dissented in part from the order. Commissioner Copps specifically emphasized that the Commission already should have started “getting an accurate count of minority and female ownership.” Id. at 469 (internal quotation marks omitted). Ultimately, the order kept the revenue-based eligible entity definition. In connection with the Diversity Order, the FCC also released another notice of proposed rulemaking that sought comment on using the SBA’s definition of an SDB (as opposed to the SBA’s revenue-based definition of a small business). Its SDB definition uses both social and economic criteria. For the former category, “[s]ocially disadvantaged individuals are those who have been subjected to racial or ethnic prejudice or cultural bias within American society because of their identities as members of groups and without regard to their individual qualities. The social disadvantage must stem from circumstances beyond their control.” 13 C.F.R. § 124.103. Members of certain groups, including African Americans, Hispanic Americans, and Native Americans, get a rebuttable presumption that they qualify. Id. Meanwhile, the economic definition encompasses “socially disadvantaged individuals [as defined above] whose ability to compete in the free enterprise system has been impaired due to diminished capital and credit opportunities as compared to others in the same or similar line of business who are not socially disadvantaged.” Id. § 124.104. Two years later, in 2010, the Commission’s own group of experts, the Advisory Committee on Diversity for Communications in the Digital Age, proposed another possible eligible entity definition. It would give a preference “for individuals who have faced substantial disadvantages and overcome those disadvantages.” See Joint Appendix 984. This is known as the overcoming disadvantages preference (“ODP”). As examples of potential disadvantages, the Committee mentioned: 1) physical disabilities or psychological disorders; 2) physical or emotional trauma due to military service; 3) unequal access to higher education; 4) unequal access to credit; 5) unequal treatment in hiring or promotions; 6) exclusion without cause from business, professional, or social associations; 7) retaliatory or discriminatory behavior by an employer or school; and 8) social patterns or pressures. Id. at 987. The Committee concluded that the proposal would not trigger heightened scrutiny because “the program would not award preferences on the basis of an applicant’s ethnicity or gender.” Id. at 993. As a result, it predicted that courts would apply rational basis review. Id. C. Prometheus II Although the Commission sought comments on using the SBA’s social and economic criteria and the Advisory Committee on Diversity for Communications in the Digital Age separately proposed a definition based on having overcome disadvantages, the revenue-based definition was still the one in place (and the only one the FCC had formally considered) by the time we decided Prometheus II. In that decision we found the revenue-based definition to be arbitrary and capricious and remanded the matter to the Commission. Prometheus II, 652 F.3d at 469-71. This put a freeze on all ongoing initiatives that relied on an eligible entity definition. We concluded that, “[f]irst and foremost, the Diversity Order does not explain how the eligible entity definition adopted would increase broadcast ownership by minorities and women.” Id. at 470. We also said that it is hard to understand how measures using this definition would achieve the stated goal. For example, by the Commission’s own calculations, minorities comprise 8.5% of commercial radio station owners that qualify as small businesses [under the revenue-based definition], but 7.78% of the commercial radio industry as a whole — a difference of less than 1%. Thus, these measures cannot be expected to have much effect on minority ownership. Id. (internal citation omitted). We noted also that “the Commission referenced no data on television ownership by minorities or women and no data regarding commercial radio ownership by women. This is because, as the Commission has since conceded, it has no accurate data to cite.” Id. (emphases in original). At the time of Prometheus II, an FCC initiative to fix the data problem was underway. Specifically, in the spring of 2009 the Commission issued another notice of proposed rulemaking along with a report and order that instituted certain changes. The principal efforts related to Form 323, which the FCC uses to track race, ethnic origin, and gender data for broadcast licensees. Prior to 2009, full-power commercial AM, FM, and television broadcast stations typically had to file Form 323 biennially, but many other types of entities were exempt. The 2009 initiative ended the exemption for sole proprietorships, partnerships comprised of natural persons, and low-power television stations. See Promoting Diversification of Ownership in the Broadcasting Services, 2006 Quadrennial Regulatory Review — Review of the Commission’s Broadcast Ownership Rules and Other Rules Adopted Pursuant to Section 202 of the Telecommunications Act of 1996, Report and Order and Fourth Further Notice of Proposed Rulemaking, 24 F.C.C.R. 5896, ¶ 3 (Apr. 8, 2009). The FCC also directed that the format for filing Form 323 be changed so that a searchable database could be created. Id. ¶ 20. The Commission expressed optimism that these changes, along with others outlined in the document, would fix the data gap. Id. ¶ 12 (“We believe that the changes we are adopting today adequately address [the] criticisms and will allow us to undertake studies that reliably analyze minority and female ownership.”). Our reaction to these efforts in Prometheus II was that, while “this is certainly a welcome and long overdue step, it does not remedy the existing data gap in the Diversity Order.” Prometheus II, 652 F.3d at 471. After finding the eligible entity definition to be arbitrary and capricious in Prometheus II, we noted our expectations for how the FCC should proceed on remand. First, we “anticipate[d]” that the Form 323 changes would “lay necessary groundwork for the Commission’s actions.” Id. Next, we directed the Commission to consider proposed eligible entity definitions before completing the 2010 Quadrennial Review. Id. (We conclude ... that the FCC did not provide a sufficiently reasoned basis for deferring consideration of the proposed SDB definitions and remand for it to do so before it completes its 2010 Quadrennial Review.”); see also id. at 472 (“[W]e reemphasize that the actions required on remand should be completed within the course of the Commission’s 2010 Quadrennial Review of its media ownership rules.”). Consideration of the proposals was particularly urgent because, “[d]espite our prior remand requiring the Commission to consider the effect of its rules on minority and female ownership, and anticipating a workable SDB definition well before this rulemaking was completed, the Commission has in large part punted yet again on this important issue.” Id. at 471. We cautioned that the FCC could not merely fall back on Adarand to justify further delays: “Stating that the task is difficult in light of Adarand does not constitute ‘considering’ proposals using an SDB definition.” Id. at 471 n. 42. We also addressed the data gap, concluding that the “FCC’s own failure to collect or analyze data, and lay other necessary groundwork, may help to explain, but does not excuse, its failure to consider the proposals presented over many years. If the Commission requires more and better data to complete the necessary Adarand studies, it must get the data and conduct up-to-date studies.... ” Id. Despite these concerns, we were nonetheless “encouraged that the FCC has taken steps” to plug the data gap, and we anticipated that it would “act with diligence to synthesize and release existing data such that studies will be available for public review in time for the completion of the 2010 Quadrennial Review.” Id. D. Aftermath of Prometheus II In December 2011 (five months after Prometheus II), as part of its report on the still ongoing 2010 Quadrennial Review, the Commission once again punted, citing data concerns. It wrote: We recognize that the data currently in the record of this proceeding are not complete and are likely insufficient either to address the concerns raised in Prometheus II or to support race- or gender-based actions by the Commission. Although we would prefer to be able to propose specific actions in response to the Third Circuit’s remand of the measures relying on the eligible entity definition ..., we believe that making legally sound proposals would not be possible based on the record before us at this time. 2010 Quadrennial Regulatory Review— Review of the Commission’s Broadcast Ownership Rules and Other Rules Adopted Pursuant to Section 2Ó2 of the Telecommunications Act of 1996, Notice of Proposed Rulemaking, 26 F.C.C.R. 17489, ¶ 158 (Dec. 22, 2011). It then committed to undertaking the following initiatives in anticipation of the 2014 Quadrennial Review: 1) Continue to improve our data collection so that we and the public may more easily identify the diverse range of broadcast owners, including women and minorities, in all services we license; 2) Commission appropriately-tailored research and analysis on diversity of ownership; and 3) Conduct workshops on the opportunities and challenges facing diverse populations in broadcast ownership. In addition, we ask interested parties to supplement the record and provide any and all data available that can complete a picture of the current state of ownership diversity, including minority and female ownership in the broadcast industry[,] and to justify any prospective actions the Commission may take on remand. Id. Also in anticipation of the 2014 review, the Commission sought comments about whether it should keep the revenue-based eligible entity definition. Specifically, it asked whether there was any “additional evidence available that would show a stronger connection” between using that definition and promoting minority and female ownership. Id. H.160. It also asked whether “other policy objectives aside from the promotion of minority and female station ownership,” such as encouraging ownership by small businesses, might support using a revenue-based definition. Id. ¶ 161. With respect to alternative definitions of eligible entities, the FCC once again sought comment on whether to use the SBA’s social and economic preferences. Id. ¶ 163. It also asked about any other definitions that commenters might suggest. Id. The next year, the Commission released a document known as the 2012 323 Report, which provides analysis of the first two rounds of reporting under the revised biennial Form 323 requirements. It concluded that the new approach was working: “These data represent the first two snapshots of broadcast ownership in a series of planned biennial data collections that, taken together, should provide a reliable basis for analyzing ownership trends in the industry, including ownership by racial and ethnic minorities and women.” 2012 323 Report, 27 F.C.C.R. 13814, ¶2 (Nov. 14, 2012). The results confirmed the underrep-resentation of minorities and women. For instance, as of the 2011 Form 323 reporting period, racial minorities had majority ownership of just 2.2 percent of full-power commercial television stations. Id. ¶ 7. For women, that number was 6.8 percent. Id. ¶ 5. Finally, in the 201b. FNPRM & Order, the FCC addressed our Prometheus II remand, saying: “As directed by the court, we consider the socially and economically disadvantaged business definition as a possible basis for favorable regulatory treatment, as well as other possible definitions that would expressly recognize the race and ethnicity of applicants.” 2014 FNPRM & Order ¶ 7. However, it tentatively rejected them. And although it “concede[d] that we do not have an evidentiary record demonstrating that this standard specifically increases minority and female broadcast ownership,” the Commission tentatively recommended reinstating the revenue-based standard that we had rejected in Prometheus II as arbitrary and capricious. Id. ¶ 267. It reasoned that, “even in the absence of such evidence, we believe that reinstatement of the revenue-based standard would serve the public interest by promoting small-business participation in the broadcast industry.” Id. In evaluating SDB-based — as opposed to revenue-based — proposals, the Commission never considered whether they would increase minority and female ownership. Rather, it rejected them on the ground that they would not, on the current record, survive constitutional scrutiny. First, the Commission considered whether it could establish a compelling governmental interest, as required to pass strict scrutiny. It concluded that its “interest in promoting a diversity of viewpoints could be deemed sufficiently compelling to survive strict scrutiny analysis.” Id. ¶ 286. However, it determined that there was not enough record evidence to meet strict scrutiny’s narrow tailoring requirement. Id. ¶ 282. In support of that conclusion, the Commission walked through the studies in the administrative record. It determined that two “directly address the impact of minority ownership on viewpoint diversity.” Id. ¶ 292. The first, Media Ownership Study 8A, found that “the relationship between minority ownership and viewpoint diversity is not statistically distinguishable from zero.” Id. The second, Media Ownership Study 8B, found “almost no statistically significant relationship between [minority] ownership and ... viewpoint diversity.” Id. ¶ 293. The exception in 8B was a positive correlation between minority ownership and coverage of minority politicians. Id. The Commission also looked at studies that found a connection between minority ownership and diversity of programming or format. For instance, Media Ownership Study 7 showed that the “presence of minority-owned stations increases the amount of minority-targeted programming and that the availability of minority-targeted formats attracts more minorities to listening.” Id. ¶ 294. Another study determined that “72.5 percent ... of minority-owned radio stations broadcast minority-oriented formats, including Spanish, Urban, Urban News, Asian, Ethnic and Religious formats geared to minority audiences.” Id. ¶ 296. However, the FCC concluded that these studies were not helpful because, though they tie minority ownership to diversity, they do not relate to viewpoint diversity. Id. ¶ 294 (“We do not believe that evidence regarding program or other forms of diversity is as relevant as evidence regarding viewpoint diversity for the purpose of establishing narrow tailoring to a compelling interest.”). The Commission based this determination on its reading of Turner Broadcasting System, Inc. v. FCC, 512 U.S. 622, 114 S.Ct. 2445, 129 L.Ed.2d 497 (1994), which it said stands for the principle that viewpoint diversity in broadcasting is an interest “of the highest order.” 2011 FNPRM & Order ¶ 294 & n. 888 (internal quotation marks omitted). Based on this review, the FCC concluded that the studies “begin to answer” some of the questions necessary for a strict scrutiny analysis but “do not demonstrate the nearly complete or tightly bound nexus between diversity of viewpoint and minority ownership that would be required to justify a race-based eligible entity definition.” Id. ¶ 298 (internal quotation marks omitted). The Commission also tentatively determined that the record did not “reveal a feasible means of carrying out the type of individualized consideration the Supreme Court has held is required to pass constitutional muster under strict scrutiny.” Id. ¶ 299. It noted that, although the Supreme Court has approved the use of racial preferences in higher education, the “manner in which the Commission allocates broadcast licenses is different in many important respects from university admissions, and we believe that implementing a program for awarding or affording preferences related to broadcast licenses based on the ‘individualized reviev/ required in other contexts would pose a number of administrative and practical challenges for the Commission.” Id. The Commission also gave limited attention to the proposed ODP (overcoming disadvantages preference) standard. At the outset, it said that “it is not entirely clear whether the ... standard would be subject to heightened constitutional scrutiny.” Id. ¶ 300. This is so even if the standard would not mention race - or gender and would instead focus on disadvantages that anyone could face. The Commission reasoned that a standard that does not “facially include race-conscious criteria, yet is constructed for the purpose of promoting minority ownership, might be subject to heightened scrutiny.” Id. ¶300 n. 915. It also noted: Among other issues, no commenter provided input on (1) what social or economic disadvantages should' be cognizable under an ODP standard, (2) how the Commission could validate claims of eligibility for ODP status, (3) whether applicants should bear the burden of proving specifically that they would contribute to diversity as a result of having overcome certain disadvantages, (4) how the Commission could measure the overcoming of a disadvantage if an applicant is a widely held corporation rather than an entity with a single majority shareholder or a small number of control persons, and (5) how the Commission could evaluate the effectiveness of the use of an ODP standard. Id. ¶ 300. Next, the Commission considered the possibility of preferences that focused on gender alone, as these would trigger intermediate rather than strict scrutiny. It concluded: While we acknowledge that the data show that women-owned stations are not represented in proportion to the presence of women in the overall population, we do not believe that the evidence available at this time reveals that the content provided via women-owned broadcast stations substantially contributes to viewpoint diversity in a manner different from other stations or otherwise varies significantly from that provided by other stations. Id. ¶301. Finally, the FCC rejected the notion that its approach violated our remand in Prometheus II. The crux of its argument is that we said it needed to consider adopting an SDB-based definition in connection with the 2010 Quadrennial Review, which technically has not ended yet because it was rolled into the 2014 cycle (which, in turn, is not yet complete). As it explained: The Commission intends to follow the Third Circuit’s direction that we consider adopting an SDB definition before completion of this proceeding and evaluate the feasibility of adopting a race-conscious eligibility standard based on an extensive analysis of the available evidence. We do not believe that the Third Circuit intended to prejudge' the outcome of our analysis of the evidence or the feasibility of implementing a race-conscious standard that would be consistent both with applicable legal standards and the Commission’s practices and procedures. Id. ¶ 283. E. Discussion With 12 years having passed since Prometheus I, we conclude that the Commission has had more than enough time to reach a decision on the eligible entity definition. We put it on notice of our concerns five years ago in Prometheus II. 652 F.3d at 471. We directed it to take action in the course of the 2010 Quadrennial Review, id. and then we returned to that topic again to “re-emphasize” our directive, id. at 472. However, the Commission has not complied. In light of this history, the Oil, Chemical & Atomic Workers Union factors for evaluating requests under APA § 706(1) all counsel in favor of finding an unreasonable delay. First, we must consider “the length of time that has elapsed since the agency came under a duty to act.” Oil, Chemical & Atomic Workers Union, 145 F.3d at 123. Here the duty to act arose as early as 2004 as a result of our Prometheus I decision and became even more explicit in Prometheus II. Second, we must judge “the reasonableness of the delay ... in the context of the statute authorizing the agency’s action.” Id. Federal law imposes on the Commission an obligation to promote ownership by minorities and women. See 47 U.S.C. §§ 309(i), (j). As such, we have described promoting minority and female ownership as an “important aspect of the overall media ownership regulatory framework.” Prometheus II, 652 F.3d at 472. Third, we need to “assess the consequences of the agency’s delay.” Oil, Chemical & Atomic Workers Union, 145 F.3d at 123. In our case, the consequence is that several initiatives pegged to a workable eligible entity definition cannot take effect. Finally, we have to look at “any plea of administrative error, administrative inconvenience, practical difficulty in carrying out a legislative mandate, or need to prioritize in the face of limited resources.” Id. (internal quotation marks omitted). We already have determined that difficulty in collecting data does not justify the delay here. See Prometheus II, 652 F.3d at 471 n. 42. In another context, we have ordered action under APA § 706(1) when an agency had delayed for nine years. See Public Citizen Health Research Group, 314 F.3d at 153 (‘We find extreme OSHA’s nine-year (and counting) delay since announcing its intention to begin the rulemaking process, even relative to delays other courts have condemned in comparable cases.”). Here more than a decade has passed since Prometheus I. Although “we are sympathetic to [the] claim that a thorough ... analysis is both highly important and quite difficult, we cannot allow an imperfect analysis to justify indefinite delay.” Id. at 156. The FCC presents two arguments for why we should not order relief. Both fail. The first is that it is not yet in violation of Prometheus II because we instructed it to address the eligible entity definition during the 2010 Quadrennial Review, which is still ongoing. This contention improperly attempts to use one delay (the Quadrennial Review) to excuse another (the eligible entity definition). By this logic, the Commission could delay another decade or more without running afoul of our remand. Simply put, it cannot evade our remand merely by keeping the 2010 review open indefinitely. The second reason is that the FCC’s Chairman has committed to circulating to the other Commissioners by June 30 of this year an order taking final action on the eligible entity definition. Counsel for the Commission said at oral argument that it was reasonable to expect that the order would be finalized and adopted by the end of year. See Oral Arg. Tr. at 46. The D.C. Circuit encountered a similar request from the FCC for forbearance in In re Core Communications, Inc., 531 F.3d 849 (D.C. Cir. 2008). There, as the Court was considering whether to mandate action after the Commission had violated a remand order, counsel said that ordering relief was not necessary because “the Commission [was] on the brink of concluding” its obligations. Id. at 858. The Court rejected the request to stay its hand, observing that “[w]e have heard this refrain before.” Id. So have we. The Commission justified delay in Prometheus I on the basis of pending legislation. In Prometheus II it did the same based on data concerns but said that better data was forthcoming due to the Form 323 revisions. Moreover, although we appreciate the Chairman’s commitment to prompt action, it is not reason enough to refrain from issuing an order, particularly .because his desire to move quickly does not bind his fellow Commissioners. As the D.C. Circuit aptly explained, the Chairman’s doing “everything he can” may not suffice, as he may not be able to enforce his will on his fellow Commissioners. In any event, the representation is not enforceable unless backed up by issuance of [an order]. At some point, promises' are no longer enough, and we must end the game of administrative keep-away. Id. at 859 (internal quotation marks omitted). Thus, although we expect that the Commission will meet the proposed deadline, in light of the previous delays we find it necessary to back up that expectation with an order. We therefore remand and order the Commission, pursuant to APA § 706(1), to act promptly to bring the eligible entity definition to a close. It must make a final determination as to whether to adopt a new definition. If it needs more data to do so, it must get it. We do not intend to prejudge the outcome of this analysis; we only order that it must be completed. Once the agency issues a final order either adopting an SDB- or ODP-based definition (or something similar) or concluding that it cannot do so, any aggrieved parties will be able to seek judicial review. Though we readily conclude that the time for prompt action has arrived, the harder issue is determining what “prompt” should mean. Recognizing that the parties are in a better position than us to understand the complexities of what is required, we proposed at oral argument the possibility of following our approach in Public Citizen Health Research Group, where we turned to mediation after ordering APA § 706(1) relief. The parties agreed to this approach, and as a result we order mediation between Citizen Petitioners and the Commission for purposes of fixing a timetable for final agency action. However, we include here the same caveat that we did in Public Citizen Health Research Group: If the parties are not able to agree within 60 days on an appropriate timeline, we will promulgate a schedule that we deem appropriate. See 314 F.3d at 159. IV. Failure to Complete Quadrennial Review Section 202(h) of the Telecommunications Act of 1996 uses unmistakably mandatory language in describing the Commission’s obligations. It provides that the Commission “shall” review its rules on broadcast ownership every four years, “shall determine whether any of such rules are necessary in the public interest as the result of competition,” and “shall repeal or modify any regulation it determines to be no longer in the public interest.” This repeated use of “shall” creates “an obligation impervious to ... discretion.” Lexecon Inc. v. Milberg Weiss Bershad Hynes & Lerach, 523 U.S. 26, 35, 118 S.Ct. 956, 140 L.Ed.2d 62 (1998). Indeed, the very purpose of § 202(h) — to function as an “ongoing mechanism to ensure that the Commission’s regulatory framework would keep pace with the competitive changes in the marketplace,” Prometheus I, 373 F.3d at 391 (internal quotation marks omitted)— reinforces the need for timeliness. Despite this command from Congress, the last review that the Commission completed was, as noted, the 2006 cycle. The order ending that review was adopted in December 2007 and released in February 2008, and it led to Prometheus II. Since 2008, more than eight years — enough time for two review cycles — have passed without any final action. In 2014, the Commission said it was “cognizant” of its “statutory obligation to review the broadcast ownership rules every four years,” but it concluded, based on deficiencies in the record, that it was unable to wrap up the 2010 review. 2011 FNPRM & Order ¶ 1. As a result, it rolled that review into the 2014 cycle, which is still ongoing. Asked repeatedly at oral argument to explain the reason for the delay, the" Commission was unable to provide a cogent response. And we note that at least one Commissioner shares our confusion. See. Statement of Commissioner Ajit Pai, Dissenting, 29 F.C.C.R. 4371, 4587 (Mar. 31, 2014) (“Pai Dissent”) (“Our decision today — or, to be more accurate, our lack of decision — is a thumb in the eye of Congress and an evasion of our legal obligations. What makes it even worse is that [the 2011 FNPRM & Order] contains no meaningful explanation for why we are not resolving the quadrennial review.”). A. The costs of delay The Commission’s delay keeps five broadcast ownership rules in limbo: the local television ownership rule, the local radio ownership rule, the newspaper/broadcast cross-ownership (“NBCO”) rule, the radio/television cross-ownership rule, and the dual network rule. We provided an overview of these rules in our 2011 decision. See Prometheus II, 652 F.3d at 438-42. Because the challenge here is not to the content of the rules, but rather relates to the consequence of the Commission’s, delay, we need not repeat that summary here. Though a full exploration of the rules is not necessary, a brief discussion of one of them, the NB'CO rule, provides a telling example of why the delay is so problematic. The rule, which dates back to 1975, prohibits common ownership of a daily newspaper and a television or radio station in the same market. See 47 C.F.R. § 73.3555(d). As part of its 2002 review, the FCC determined that a complete ban was no longer in the public interest but that some restrictions were. See Prometheus I, 373 F.3d at 400-01. We agreed with this premise in Prometheus I. Id. at 398-400. However, we concluded that the particular proposal that the FCC adopted to replace the complete ban was arbitrary and capricious. Id. at 402. The result was to leave the 1975 ban in place. The Commission tried again during the 2006 cycle, but we determined in Prometheus II that its new attempt did not comply with the APA’s notice-and-comment requirements. Prometheus II, 652 F.3d at 445. Once again, our conclusion resulted in the complete ban staying on the books. In the 201f FNPRM & Order, the Commission struggled to find a way to make the rule work, and it discussed the possibility of leaving radio/newspaper combinations unregulated and limiting its restrictions to television/newspaper cross-ownership. See 2011 FNPRM & Order ¶¶ 113-20. As a result, the 1975 ban remains in effect to this day even though the FCC determined more than a decade ago that it is no longer in the public interest. This has come at significant expense to parties that would be able, under some of the less restrictive options being considered by the Commission, to engage in profitable combinations. The delay also hampers judicial review because there is no final agency action to challenge. Cf. Core Communications, 531 F.3d at 856 (“Moreover, until the FCC states its explanation for its interim rules in a final order, Core cannot mount a challenge to those rules. In this way, the FCC insulates [itself] ... from further review.”). B. Vacatur is not appropriate Deregulatory Petitioners argue that the result of this delay should be to vacate all five broadcast ownership rules that are subject to § 202(h). This would wipe the slate clean. And, unless the Commission were immediately to issue new rules, it would lead to a degree of deregulation that is unprecedented in the modern broadcast industry. Despite asking for such sweeping relief, Deregulatory Petitioners do not cite a single instance when a court has ordered mass vacatur in similar circumstances. There is a reason for this: Vacatur typically is inappropriate where it is “conceivable” that the Commission can, if given the opportunity, create a supportable rule. See Allied-Signal, Inc. v. U.S. Nuclear Regulatory Comm’n, 988 F.2d 146, 151 (D.C. Cir. 1993). Here, despite the delay, we have no reason to suspect that the Commission cannot justify at least some restrictions on broadcast ownership. We must also consider the “disruptive consequences of vacating.” Id. In this case, vacatur risks creating a temporary free-for-all in the broadcast industry, with companies racing to engage in previously prohibited combinations before the FCC comes forward with new regulations. This would invite chaos, and it presumably would lead to drawn-out litigation over whether combinations entered into during this vacuum could continue to exist even if the Commission later determined to outlaw them once again. Fox Television Stations, Inc. v. FCC, 280 F.3d 1027 (D.C. Cir.), opinion modified on reh’g, 293 F.3d 537 (D.C. Cir. 2002), which Deregulatory Petitioners cite as support for vacatur, is not contrary to our conclusion. There the D.C. Circuit said that if “the reviewing court lacked the power to require the Commission to vacate a rule it had improperly retained [during a Quadrennial Review] and could require the Commission only to reconsider its decision, then the presumption in § 202(h) would lose much of its bite.” Id. at 1048. This is a reference to the notion that “Section 202(h) carries with it a presumption in favor of repealing or modifying the ownership rules.” Id. At the outset, we note that there is little left of this presumption in light of subsequent clarifications from both the D.C. Circuit and our Court. See Cellco P’ship v. FCC, 357 F.3d 88, 97-99 (D.C. Cir. 2004); Prometheus I, 373 F.3d at 394-95. Quibbling aside, we agree that vacatur is sometimes appropriate in the § 202(h) context. It is most likely appropriate in the setting the D.C. Circuit described in Fox — when the Commission has reached a final (but unsupportable) decision to retain, rather than repeal or modify, a rule. Here the Commission has not reached any final determinations. Deregulatory Petitioners are not complaining of improper agency action-, rather, their problem is with agency delay. When an agency gives its reasoning for retaining a rule and a court concludes that it is insufficient and incapable of being cured, vacatur makes sense. With no final justifications to review, however, we cannot be convinced that it would be futile to allow the Commission to try to come to a supportable conclusion. That is not to say that vacatur is never an appropriate response when the allegation is agency delay rather than arbitrary and capricious agency action. For instance, in Core Communications the D.C. Circuit gave the Commission six months to justify certain rules or have them wiped from the books. Core Communications, 531 F.3d at 861. There, however, the Court had twice remanded the issue to the Commission, which had delayed action for six years after the second remand. Id. at 850. That situation is analogous to the Commission’s delays on the eligible entity definition. Here, by contrast, the delay did not begin until after Prometheus II. Though its delay is troubling, the Commission was not on notice that its tardiness might result in vacatur. C. Deregulatory Petitioners have waived other forms of relief There is yet another reason why vacatur is inappropriate: Deregulatory Petitioners could have sought relief in the form of an order under APA § 706(1) or a writ of mandamus. Under either of these provisions, we could have ordered the Commission to resolve promptly its 2010 and 2014 review cycles. This route, rather than va-catur, typically is the appropriate remedy to agency delay. See Cellco P’ship, 357 F.3d at 101. However, unlike Citizen Petitioners, who specifically sought APA § 706(1) and mandamus relief, Deregula-tory Petitioners abandoned those requests. Of the petitions for review, the only one to take issue with the delayed Quadrennial Review process was filed by the National Association of Broadcasters (“NAB”). It charges the Commission with “unlawfully withholding]” action, but it requests vaca-tur as the remedy and never gives any indication (apart from a generic plea for any relief we deem appropriate) that it seeks an order mandating action. See NAB Petition for Review at 3, 5. Similarly, the sole remedy that Deregulatory Petitioners request in their opening brief is vacatur. They cite § 706(1) in their reply brief, but only in a footnote in that document do they even mention remedies other than complete vacatur. Specifically, they suggest that, if “the Court does not vacate the rules outright, it should at minimum vacate with a stay to urge quick action or remand with direction to act immediately.” Deregulatory Petitioners’ Reply Br. at 11 n.2 (internal citations omitted) (emphasis in original). By focusing their energy on vacatur, Deregulatory Petitioners went for a home run. Instead, they have struck out. Now they find themselves with alternative requests that are not singly, but rather doubly, waived: first because they were raised for the first time in a reply brief, and second because they were relegated to a footnote. See United States v. Pelullo, 399 F.3d 197, 222 (3d Cir. 2005) (“It is well settled that an appellant’s failure to identify or argue an issue in his opening brief constitutes waiver of that issue on appeal.”); John Wyeth &Bro. Ltd. v. CIGNA Int’l Corp., 119 F.3d 1070, 1076 n. 6 (3d Cir. 1997) (“[A]rguments raised in passing (such as, in a footnote), but not squarely argued, are considered waived.”). As a result, we decline to order APA § 706(1) or mandamus relief. Our conclusion that Deregulatory Petitioners have litigated themselves out of relief, however, should not be confused with a green light for further agency delay. At oral argument, counsel for the Commission made the same assurance here as he did in the context of the eligible entity definition: that the Chairman intends to circulate an order to the other Commissioners by June 30 of this year, with the expectation being that it would be finalized and adopted by the end of the year. See Oral Arg. Tr. at 46. The order, we are told, would bring the 2010 and 2014 Quadrennial Reviews to a close. Id. at 41, 46. We fully anticipate that the Commission will meet this deadline. However, if it fails to complete the reviews by then, it does so at its own risk. New litigation would likely result, and the outcome may well be different. V. The Joint Sales Agreement Rule The Commission places restrictions on common ownership of broadcast television stations. For instance, the local television ownership rule allows “one entity to own two television stations [with overlapping contours] in a market (a television duopoly) as long as at least one of the stations was not ranked among the market’s four largest stations and at least eight independently owned and operated stations (called ‘eight voices’) would remain post-merger.” Prometheus II, 652 F.3d at 439. Imagine a market with ten overlapping stations: Stations A-J (with Stations A-D being the four largest). Clearly, the local television ownership rule prevents one company from owning Stations D, E, and F because there is a limit of two. But what if a company owns Stations D and E and has a contract that gives it significant influence over station F? To answer this question, the Commission has “attribution” rules. If Station F is attributed to the owner of Stations D and E, it means that the latter is considered to own the former for purposes of the Commission’s ownership caps. The purpose of attribution rules is to prevent companies from circumventing ownership limits by doing through clever contracting what they are not permitted to do through outright purchases of stations. In 2014, the Commission created a new attribution rule for television joint sales agreements (“JSAs”). A JSA is a contract that allows one station (the brokering station) to sell advertising (but not programming) on a second station (the brokered station). Specifically, the FCC determined that same-market JSAs involving the sale of more than 15 percent of the brokered station’s weekly advertising will be attributed (ie., counted toward ownership caps). See 2011 FNPRM & Order ¶ 340. Returning to our example, if the owner of Stations D and E has a JSA over the 15-percent threshold with Station F, that arrangement would violate the Commission’s ownership rules. Importantly, the new rule does not outlaw JSAs over the 15-percent limit; it just means that they must be counted toward ownership caps. For instance, a company that only owns one television station can have an attributable JSA with a second station without running afoul of the ownership limits (as long as the eight-voices and the four-largest-stations rules are not violated). Deregulatory Petitioners challenge the television JSA rule and argue that its enactment violated § 202(h). A. Background The 2014 television JSA rule traces its roots to the 1990s, when the FCC first attributed local marketing agreements (“LMAs”). They are similar to JSAs, except that they involve the sale of programming as well as advertising on the brokered station. In 1992, the Commission attributed same-market radio