Spirit Airlines, Inc. v. DOT

United States Court of Appeals FOR THE DISTRICT OF COLUMBIA CIRCUIT Argued September 16, 2020 Decided May 21, 2021 No. 19-1248 SPIRIT AIRLINES, INC., PETITIONER v. UNITED STATES DEPARTMENT OF TRANSPORTATION AND FEDERAL AVIATION ADMINISTRATION, RESPONDENTS On Petition for Review of an Order of the Federal Aviation Administration Aimee W. Brown argued the cause for petitioner. On the briefs were Joanne W. Young and David M. Kirstein. Kannon K. Shanmugam entered an appearance. Scott P. Lewis and Thomas R. Devine were on the brief for amicus curiae Airports Council International - North America in support of petitioner. Benjamin M. Shultz, Attorney, U.S. Department of Justice, argued the cause for respondents. With him on the brief were Michael S. Raab, Attorney, Steven G. Bradbury, General Counsel, U.S. Department of Transportation, Paul M. Geier, Assistant General Counsel for Litigation and Enforcement, Joy 2 K. Park, Senior Trial Attorney, and Arjun Garg, Chief Counsel, Federal Aviation Administration. Before: HENDERSON and WALKER, Circuit Judges, and GINSBURG, Senior Circuit Judge. Opinion for the Court filed by Senior Circuit Judge GINSBURG. GINSBURG, Senior Circuit Judge: Spirit Airlines, a low- fare passenger carrier, challenges the Federal Aviation Administration’s decision not to reallocate peak-period flight authorizations previously held by Southwest Airlines at Newark International Airport. Spirit argues this decision was arbitrary and capricious because the FAA improperly failed to: (1) consider the effect on competition; (2) consider less burdensome alternatives; and (3) support its decision with substantial evidence. The FAA argues its decision is unreviewable because it is not final agency action and, in the alternative, contests each of Spirit’s objections. We conclude the FAA’s decision was final because it prevented Spirit from operating as many peak-period flights as it would otherwise have done in the Summer 2020 scheduling season. We also conclude the FAA’s decision was arbitrary and capricious because the agency disregarded warnings about the effect of its decision on competition at Newark. We therefore grant Spirit’s petition for review and vacate the FAA’s decision to retire the peak-period flight authorizations previously held by Southwest. I. Background Since 1968, the FAA has exercised varying degrees of control over the scheduling of flights to and from Newark 3 International Airport. See High Density Traffic Airports, 33 Fed. Reg. 17,896 (Dec. 3, 1968). For some years the FAA maintained a formal reservation system known as “slot control” that required each airline to request in advance a “slot” for each takeoff or landing it proposed to schedule. See Republic Airline Inc. v. Dep’t of Transp., 669 F.3d 296, 297-98 & n.2 (D.C. Cir. 2012). The FAA relaxed this requirement in 2016. See Change of Newark Liberty International Airport (EWR) Designation, 81 Fed. Reg. 19,861, 19,862 (Apr. 6, 2016). Under its current policy, the FAA announces hourly and half-hourly caps on takeoffs and landings for a given scheduling season. See Notice of Submission Deadline for Schedule Information for Newark Liberty International Airport for the Summer 2020 Scheduling Season, 84 Fed. Reg. 52,580, 52,581 (Oct. 2, 2019). Each airline then tells the FAA what flights it wants to operate during the upcoming season. Id. The FAA may either approve an airline’s plan or request that it make changes in order to reduce congestion. Id. An airline is not legally barred from operating flights not on its FAA-approved schedule. See 81 Fed. Reg. at 19,862. The FAA has warned, however, that doing so may exacerbate congestion and bring about a return to slot control. 84 Fed. Reg. at 52,582 (noting “if voluntary schedule adjustments are not achievable, consideration may be given to whether [slot control] is necessary...”). Should that happen, the FAA has said it would allocate slots based upon a grandfathering policy: “historic precedence would not be granted,” however, “for any operation conducted without FAA approval” under the current, more relaxed framework. Id. As a result, only flights that currently operate with the FAA’s blessing would be allowed to continue under slot control. 4 Competition – more specifically, the lack of competition among airlines – has long been a problem at Newark. In 2010, when the airport was still under slot control, United and Continental Airlines sought to merge. To prevent harm to competition, the Department of Justice (DoJ) conditioned the merger on United’s transferring 36 slots to Southwest Airlines, a low-fare carrier that was not then operating at Newark. See DoJ Press Release, “United Airlines and Continental Airlines Transfer Assets to Southwest Airlines in Response to Department of Justice’s Antitrust Concerns,” (Aug. 27, 2010), https://www.justice.gov/opa/pr/united-airlines-and- continental-airlines-transfer-assets-southwest-airlines- response. Over the next five years, the DoJ resisted United’s multiple attempts to acquire more slots at Newark. For example, United tried to acquire more slots once in 2014 and twice in 2015 even though it was not using all the slots it already had. Verified Compl. at ¶¶ 3-4, 7-8 21-24, United States v. United Continental Holdings, Inc., No. 2:15-cv-07992 (D.N.J. Nov. 10, 2015). In 2015 the DoJ sued United for attempted monopolization in violation of the Sherman Antitrust Act. Id. at ¶¶ 21, 48-49. United ultimately abandoned each effort. Id. at ¶ 21; Stipulation of Dismissal, United Continental Holdings, No. 2:15-cv-07992 (D.N.J. Apr. 6, 2016). United remained the dominant carrier at Newark nonetheless. In July 2019 Southwest announced it would pull out of Newark in November of that year. Of Southwest’s 36 slots, approximately 16 were in the highly desirable “peak hours,” which run from 7:00 a.m. to 8:59 a.m., and from 1:30 p.m. to 9:59 p.m. Those are the periods in greatest demand. See 84 Fed. Reg. at 52,581. Spirit Airlines immediately asked for 5 them. 1 In meetings with officials from the U.S. Department of Transportation (DoT) and the FAA, Spirit said it would “continue the low-fare service that had been established by the Department of Justice in 2010 and prevent the detrimental effects on competition” that would ensue if Southwest’s peak hour authorizations were simply retired. Others weighed in too. In an August 2019 letter to the FAA, Makan Delrahim, Assistant Attorney General in charge of the Antitrust Division of the DoJ, observed that United then held “approximately 66% of [the] authorizations at Newark.” He also noted over half of all flights at Newark were United flights on “monopoly routes,” meaning no other airline flew the same route. Huntley Lawrence, Director of the Aviation Department of the Port Authority of New York and New Jersey, which operates Newark Airport, shared similar concerns in his own August 2019 letter. He pointed out that United accounted “for 72 percent of [Newark’s] peak hour operations” and, he observed, “the true price of [United’s] dominance ... is borne by consumers in the form of higher ticket prices, or the ‘Newark Premium.’” Both the DoJ Antitrust Division and the Port Authority cautioned the FAA against retiring Southwest’s slots. The Antitrust Division explicitly forewarned that “some stakeholders, particularly United, may urge the DoT and FAA to retire the capacity, ostensibly to alleviate congestion at the airport.” It urged DoT and the FAA to preserve competition by 1 Airlines such as United that already operated during peak hours with the FAA’s blessing did not need to worry about running into the hourly caps. See 84 Fed. Reg. at 52,581. As the agency explained, it would continue to “accept flights above the limits if the approved flights were operated by the same carrier on a regular basis in the previous corresponding season.” Id. 6 reallocating Southwest’s peak period slots and to address the problem of congestion by other means, such as “scheduling reduction meetings” with all carriers operating at Newark, pursuant to 49 U.S.C. § 41722. The Port Authority expressed similar concerns: Allowing [United] to increase its share of peak hour operations at Newark would cement its monopolistic position. The threat to consumer choice and healthy competition can only be mitigated by maintaining meaningful low-fare service options during peak (i.e. marketable) times. Accordingly, Southwest’s authorized operations should be allocated as a package to a new entrant, limited incumbent, or low-cost carrier. And, like the Antitrust Division, the Port Authority urged the FAA to convene a scheduling meeting to address congestion. It further charged that United and the FAA had “largely caused” the congestion problems because the agency had allowed United to operate an increasing number of flights during already busy hours. In October 2019, the FAA issued a Notice announcing the deadline for airlines to submit their proposed Summer 2020 Newark schedules for review. See 84 Fed. Reg. at 52,580-82 (October Notice). At the same time, the FAA announced it would retire Southwest’s entire block of peak period slots: The FAA plans to assess the impacts of the peak period Southwest reductions and other schedule changes at [Newark] on performance, as well as the impacts on competition in close coordination with the Office of the Secretary of Transportation, in the upcoming Winter 2019/2020 and Summer 2020 scheduling seasons. The FAA intends to publish additional information on the 7 outcome of this assessment in future notices related to these airports [sic]. However, the FAA will not during that assessment period be replacing or “backfilling” the peak morning and afternoon/evening operations that Southwest conducted during Winter 2018/2019 and Summer 2019, to the extent the new operations would exceed the current scheduling limits. Id. at 55,582. Spirit petitioned this court to review and vacate the FAA’s decision not to reallocate Southwest’s peak slots. It claims the decision was arbitrary and capricious because the FAA: (1) failed to consider the effect of its decision on competition; (2) did not explain why it could not use a less burdensome tool, such as a schedule reduction meeting, to address congestion; and (3) lacked substantial evidence for its decision. The FAA argues its decision retiring Southwest’s peak slots is not final and hence not reviewable and contests each of Spirits contentions. II. Reviewability Spirit relies upon 49 U.S.C. § 46110(a), which authorizes this court to review an “order” issued by the FAA. City of Dania Beach v. FAA, 485 F.3d 1181, 1187 (D.C. Cir. 2007) (explaining that, as under the Administrative Procedure Act, “a reviewable order under 49 U.S.C. § 46110(a) must possess the quintessential feature of agency decisionmaking suitable for judicial review: finality” (cleaned up)). Because “section 46110 does not impose any explicit finality requirement,” we have “incorporated generally applicable finality principles into the analysis of what counts as an ‘order’ under [that provision].” Flytenow, Inc. v. FAA, 808 F.3d 882, 888-89 (D.C. Cir. 2015). “To be deemed ‘final,’ an order must mark 8 the ‘consummation’ of the agency’s decisionmaking process, and must determine ‘rights or obligations’ or give rise to ‘legal consequences.’” City of Dania Beach, 485 F.3d at 1187 (quoting Bennett v. Spear, 520 U.S. 154, 177–78 (1997)); see also Nat’l Envtl. Dev. Assoc.’s Clean Air Project v. EPA, 752 F.3d 999, 1006 (D.C. Cir. 2014) (“An agency action may be final even if the agency’s position is subject to change in the future.” (internal quotation marks omitted)). The FAA argues its October 2019 Notice “neither determines obligations nor carries legal consequences” because airlines may legally operate flights not included on their preapproved schedules. As a result, any consequence the Notice may have is not “legal” but merely “practical.” 2 See Joshi v. NTSB, 791 F.3d 8, 11 (D.C. Cir. 2015) (explaining an agency’s request for voluntary compliance had practical consequences, but no binding legal effect and therefore did not constitute final agency action). Admittedly, it is a blurry line that separates legal consequences from practical consequences. We can see, however, that an agency’s action need not flatly prohibit a party from acting in order to affect its legal rights; it is enough that the agency action presently and directly limits or 2 The FAA characterizes its finality argument as jurisdictional. See Resp. Br. at 12 (“This Court lacks jurisdiction because the [Notice] is not a final order.”). As we have explained before, however, finality is a prudential doctrine, not a limit on our jurisdiction. Flytenow, Inc. v. FAA, 808 F.3d 882, 889 (D.C. Cir. 2015) (“Because the finality requirement under section 46110(a) is judicially imported from the APA, it is no more jurisdictional than the APA’s own finality requirement. Our precedent confirms that finality under the Federal Aviation Act is a matter of judicial creation, allowing us to avoid premature intervention in the administrative process.”) (internal quotation marks omitted). 9 defeats a party’s ability to enter into an advantageous business arrangement. Two of our cases illustrate this principle. Safe Extensions, Inc. v. FAA involved an “advisory circular” specifying how the bases that secure runway lights to the runway must be tested in order to get on the FAA’s list of approved products. 509 F.3d 593, 595-96 (D.C. Cir. 2007). An airport that receives federal funds (as do all major airports) may not use a product that is not on that list. Id. The FAA issued several circulars that together had the effect of exempting one type of light base from testing and intensified the requirements for another. Id. at 596-97. A manufacturer of the latter type challenged the FAA’s actions as arbitrary and capricious. Id. at 597. The agency argued its circular was not a final order under section 46110 because it “neither imposed a legal obligation upon any person nor created any legal rights.” Id. at 598 (cleaned up). We noted, however, that the circular “effectively prohibits airports from buying light bases that fail the new ... test, and it bars manufacturers like Safe Extensions from selling their products to airports. These are clear legal consequences of enormous significance.” Id. Safe Extensions remained free to manufacture its light bases and sell them to the few airports that might be willing and able to buy them. Indeed, in principle nothing prohibited Safe Extensions from trying to persuade other airports to forgo federal funding in order to buy its light bases, but obviously that would be a fool’s errand. This court applied the same principle more recently in SecurityPoint Holdings, Inc. v. TSA, 769 F.3d 1184 (D.C. Cir. 2014). That case involved a program of the Transportation Security Administration that allowed private venders to place advertisements on checkpoint equipment (such as plastic bins) they provided free of charge. To participate in the program, an airport would sign a Memorandum of Understanding (MoU) 10 with the TSA requiring “participating airports to indemnify TSA from all liability for intellectual property claims related to the checkpoint equipment.” Id. at 1186. One of the contractors, SecurityPoint, asked the TSA to reconsider, arguing no airport would be willing to enter into the MoU. Id. at 1187. When the agency refused, SecurityPoint petitioned this court for review under section 46110. Id. We held the TSA’s refusal was final because it “gave rise to legal consequences by confirming that participating airports will be subject to TSA’s new mandatory MOU language and thereby affected SecurityPoint’s ability to contract with those airports.” Id. at 1187 (cleaned up). The indemnification clause may not have stopped all airports from contracting with SecurityPoint, but it limited the company’s ability to enter into business relationships with many airports. So, too, here: Spirit is legally free to operate unapproved flights or, improbably, to try to persuade other airlines to swap their peak slots for Spirit’s off-peak slots. The FAA’s action, however, effectively forecloses Spirit from operating as many peak-period flights as it would otherwise do. In this way, the FAA’s action hinders Spirit’s ability to pursue business opportunities as surely as would an express prohibition. To emphasize the purportedly voluntary nature of the overall scheduling regime at Newark, 3 the FAA also ignores 3 We question just how “voluntary” this regime is. When it ended slot control at Newark in 2016, the FAA acknowledged “some carriers might operate at times without approval from the airport’s schedule facilitator.” 81 Fed. Reg. at 19,862. Since then, however, the FAA has conveyed its expectation – backed by the threat of a possible return to slot control – that airlines will cooperate with its scheduling efforts. See 84 Fed. Reg. at 52,581-82. A request for help backed by a threat hardly seems a call for voluntary action; at best, the airlines appear to have been “voluntold.” U.S. Army, 11 the value to an airline of having the agency’s approval. The FAA acknowledges operating flights without its blessing could cause it to reimpose slot controls at Newark,4 but it argues this possibility is speculative and therefore of no legal consequence. In the same vein, the FAA maintains that none of the scheduling decisions it makes under the scheme now in place at Newark is or ever could be final. 5 By declaring that “Soldier-Speak: A Brief Guide to Modern Military Jargon” (Mar. 9, 2015), https://www.army.mil/article/144045/soldier_speak_a_brief_guide_ to_modern_military_jargon (noting “a voluntold assignment is technically voluntary,” but “is understood to be mandatory”). In Chamber of Commerce v. Department of Labor, we rejected an agency’s attempt to portray as voluntary a program that allowed certain workplaces to eliminate the risk of costly inspections by implementing a comprehensive safety and health program. 174 F.3d 206, 209 (1999). There, as here, “the voluntary form of the rule is but a veil for the threat it obscures.” Id. at 210. 4 The FAA has said it sets flight limits based upon congestion. See 84 Fed. Reg. at 52,581; see also 81 Fed. Reg. at 19,862. It has also said flights at Newark are already above capacity and that it has observed problematic delays. See 84 Fed. Reg. at 52,581-82. And it has warned continued congestion could result in a return to slot control. Id. at 52,582. So airlines that choose to operate unapproved flights do so knowing the FAA already considers them congestion- enhancing. In other words, were Spirit to try to backfill Southwest’s authorizations without the FAA’s blessing, it would risk hastening a return to slot control. 5 See Oral Arg. at 20:14-20:57 (“[W]e’re not here saying that if at some point in the future the FAA goes to [slot control] and Spirit wants to get a spot and ... doesn’t get that slot in the [slot controlled] regime – we’re not saying that Spirit can’t at that point seek judicial review. A [slot controlled] regime would have legal effect. What we are saying is that in [the current] regime where all the FAA is doing is this voluntary facilitation process ... because those decisions 12 only approved flights would be grandfathered should it reimpose slot control, however, the FAA effectively created two classes of flights of profoundly different value. The first class comprises flights an airline operates with the FAA’s approval; they would be given precedence if congestion worsens and slot controls return. The second class consists of flights operated without the FAA’s approval; they would assuredly be barred under slot control. Thus, the FAA’s decision to retire Southwest’s peak slots rather than allocating them to Spirit denied Spirit (and perhaps other airlines) both the chance to use those slots in the present and the value they would have should the FAA reimpose slot control in the future. That is surely final agency action. The FAA nonetheless argues the October Notice is not the consummation of the FAA’s decisionmaking process because it did not actually approve or disapprove any specific proposed flights. The agency did predict it would not approve new peak period flights but, as it points out, those predictions could prove wrong. The FAA also notes there could be some swaps among airlines, the upshot being that Spirit might end up with approval to operate some peak flights after all. The FAA misses the forest for the trees. Spirit is not challenging the FAA’s decision regarding any individual scheduling authorization; rather, it is challenging the FAA’s decision, in the same October Notice, to retire Southwest’s entire block of peak-period slots. Even if Spirit were to swap its way into all Southwest’s peak authorizations, the Notice would still limit its ability to take full advantage of them for the FAA’s decision denies them the protected status approved don’t have legal effect and Spirit is still free to fly right now then it’s not yet final agency action.”). 13 flights get. Because these legal consequences are the direct result of the FAA’s decision, our precedents, as we have seen, require that we deem the October Notice final agency action. III. Merits Having concluded the FAA’s order was final, we must determine whether the FAA’s decision to retire Southwest’s peak authorizations was arbitrary and capricious, as Spirit claims. “Under this standard, we may reverse only if the agency’s decision is not supported by substantial evidence, or the agency has made a clear error in judgment.” J.A. Jones Mgmt. Servs. v. FAA, 225 F.3d 761, 764 (D.C. Cir. 2000) (internal quotation marks omitted). Although our review is inherently deferential, it is not satisfied by an agency decision that ignores an important aspect of the problem before it or relies upon a threadbare explanation. See Am. Wild Horse Pres. Campaign v. Perdue, 873 F.3d 914, 923 (D.C. Cir. 2017). “An agency is required to consider responsible alternatives to its chosen policy and to give a reasoned explanation for its rejection of such alternatives.” Am. Radio Relay League, Inc. v. FCC, 524 F.3d 227, 242 (D.C. Cir. 2008) (internal quotation marks omitted). This principle goes to the heart of reasoned decisionmaking; it is not limited to rulemaking. See, e.g., Yakima Valley Cablevision, Inc. v. FCC, 794 F.2d 737, 746 n.36 (D.C. Cir. 1986) (collecting cases and noting “[t]he failure of an agency to consider obvious alternatives has led uniformly to reversal”). Ignoring an important aspect of the problem is precisely what the FAA has done. The DoJ, the Port Authority, and Spirit all complained to the FAA that retiring Southwest’s peak authorizations was a drastic measure to address congestion and would do substantial harm to competition and hence passengers at Newark. If cutting flights was necessary, the DoJ urged the 14 DoT to do so through a schedule reduction meeting pursuant to 49 U.S.C. § 41722, rather than by retiring Southwest’s slots. Despite all this, the FAA gave no indication it even considered convening a schedule reduction meeting. Indeed, we find very little to suggest it considered competition at all beyond a single sentence in the October Notice, quoted above, saying it “plans to assess” how its decision to retire Southwest’s slots affects competition at Newark. 84 Fed. Reg. at 52,582. That falls well short of what is needed to demonstrate the agency grappled with an important aspect of the problem before it or considered another reasonable path forward. See Chamber of Commerce v. SEC, 412 F.3d 133, 145 (D.C. Cir. 2005) (“Where a party raises facially reasonable alternatives, the agency must either consider those alternatives or give some reason for declining to do so” (cleaned up and quoting Laclede Gas Co. v. FERC, 873 F.2d 1494, 1498 (D.C. Cir. 1989)); Allied Local & Reg’l Mfrs. Caucus v. EPA, 215 F.3d 61, 80 (D.C. Cir. 2000) (“To be regarded as rational, an agency must ... consider significant alternatives to the course it ultimately chooses”). For this reason, we must vacate the FAA’s decision not to reallocate Southwest’s peak slots as announced in the October Notice and remand this matter to the agency to deal with the issue of competition. We could stop here, but to inform the FAA’s consideration on remand and avoid another round of review, we will address Spirit’s contention that the FAA’s decision was not supported by substantial evidence. The substantial-evidence standard requires such “evidence as a reasonable mind might accept as adequate to support a conclusion.” Biestek v. Berryhill, 139 S. Ct. 1148, 1154 (2019) (quoting Consolidated Edison Co. v. NLRB, 305 U.S. 197, 229 (1938)). The FAA fails to clear this low bar. 15 The FAA argues the record shows it tried to address a hopelessly complex problem and was beset by irreconcilable proposals from stakeholders. It points to United’s submissions, which argued congestion at Newark was so bad that the agency should immediately reimpose slot control. “Faced with ... competing proposals, and accompanying uncertainty surrounding what effect Southwest’s departure would have on [Newark],” the FAA contends it “adopted a middle-of-the road approach.” It also contends an agency faced with uncertainty acts reasonably when it pauses to study an issue further. Of course, that can sometimes be the only rational thing to do. Cf. Commonwealth of Pa. v. Lynn, 501 F.2d 848, 855-56 (D.C. Cir. 1974). In this case the FAA’s ‘pause to study’ explanation runs into three problems. First, the agency assumes embracing a “middle-of-the-road approach” and studying an issue further is self-evidently reasonable. But the FAA points to nothing in the record from which we can conclude it rationally analyzed the various issues before deciding to retire Southwest’s peak- period slots and then study the effect on congestion. See Fla. Power & Light Co. v. FERC, 85 F.3d 684, 689 (D.C. Cir. 1996) (noting “parties are entitled to the agency’s analysis of its proposal, not post hoc salvage operations of counsel”). Second, contrary to the FAA’s suggestion, the record does not show it had to tackle a particularly complex problem. The agency produced a fairly simple model based upon historical data – which Spirit does not challenge – to predict how congestion would change if Southwest’s peak authorizations were retired. The model predicted varying reductions in delay over the course of a day but they were uniformly quite modest, to say the least. For example, the FAA anticipated no delay reduction whatsoever from retiring the flight Southwest operated during the 7:00 a.m. hour or the two flights it operated 16 during the 1:00 p.m. hour. Retiring the seven flights Southwest operated during the 2:00 p.m. or 5:00 p.m. hours was anticipated to reduce the average delay per operation (i.e., per landing and takeoff) by about 20 seconds. Retiring the three flights Southwest operated during the 7:00 p.m. hour was anticipated to reduce delays by about one minute per operation. The greatest reduction would occur if the agency retired the two flights Southwest operated between 8:00 p.m. and 9:59 p.m. In that case, delays would decrease by about four minutes per operation, from approximately 28 minutes to approximately 24 minutes per flight. In total, the FAA’s model suggested retiring all of Southwest’s authorizations during peak hours – as it did – would reduce delays on average by a little over one minute per operation. Meanwhile, the agency also ignored information about the competitive situation at Newark. The Port Authority, in particular, had painted a dire picture. United already accounted for 72 percent of the peak period operations at Newark, and it estimated retiring Southwest’s authorizations would increase that to 75 percent. The Port Authority also observed airfares generally fall by nearly 45 percent when a second airline begins flying what had been a monopoly route. And the Port Authority suggested United’s own scheduling requests – on which the FAA signed off – were “the root cause of ... delay” at Newark. The record provides precious little insight into whether or how the FAA approached the competition problem. The agency has not pointed us to a single page in the record where it analyzed the competition issues highlighted by the DoJ, the Port Authority, and Spirit. Nor did it say in the Notice or anywhere else why it prefers miniscule reductions in delay more than competition that could lower fares for passengers. Hence, we cannot say a reasonable mind would find the record 17 as a whole supports the FAA’s decision, bearing in mind that “[t]he substantiality of evidence must take into account whatever in the record fairly detracts from its weight.” Universal Camera Corp. v. NLRB, 340 U.S. 474, 488 (1951). Here that means we must consider the Port Authority’s contention – which the FAA did not address let alone contest – that other, less drastic measures for reducing delays were preferable to retiring Southwest’s peak-period approvals. It also means we must take with a grain of salt the “self-serving views of the regulated entities,” such as those offered by United upon which the FAA seems to have relied. NetCoalition v. SEC, 615 F.3d 525, 541 (D.C. Cir. 2010), superseded by statute as stated in NetCoalition v. SEC, 715 F.3d 342, 344 (D.C. Cir. 2013). Third, the FAA has pointed to nothing in the record to show why a delay reduction meeting would be impractical or less appropriate than retiring all Southwest’s peak-period authorizations. The agency tells us it “reasonably concluded that convening such a meeting (or engaging in any other process that would involve the extreme measure of forcibly reducing existing flights) would be premature” because it “was still gathering data on competition and delays in the post- Southwest era.” But that explanation makes no sense, nor is it supported by the record. It makes no sense because its decision to retire Southwest’s authorizations did, in fact, forcibly reduce the number of existing flights. And it is unsupported by the record because the FAA’s own predictive model and the information provided by the Port Authority suggested retiring Southwest’s peak-period approvals would do hardly anything to reduce delays. If the FAA again decides to retire Southwest’s peak-period slots, it should be prepared to provide a reasoned explanation for preferring to cut travel time an average of one minute rather 18 than to cut the price of flying by as much as 45 percent on routes that would gain a second carrier. IV. Concluding Remarks We close by touching briefly upon Spirit’s statutory arguments. Spirit claims various statutory provisions obligated the FAA to consider competition when it determined the fate of Southwest’s peak-period authorizations. See 49 U.S.C. §§ 40101(a), 40103(b), and 47101(a)(9). The FAA counters that none of those provisions applies to the decision Spirit is challenging. It argues § 40101(a) imposes an obligation only on the Secretary of Transportation and only when carrying out duties not at issue in this case. It argues § 40103(b) permits, but does not require, the agency to consider competition. And it argues § 47101(a)(9) is “best read as an advisory directive only.” Given our discussion thus far we need not decide whether the FAA was bound by statute to consider competition. But see Am. Airlines v. Civil Aeronautics Bd., 192 F.2d 417, 420 (D.C. Cir. 1951) (“Whatever belittling significance may be attached to the fact that [provisions detailing factors for the agency to consider] were under a title ‘Declaration of Policy’, they are in the statute, are peremptory, and are as much an enactment by the Congress as is any other section of the statute”). Under the APA, it is enough that interested parties raised the lack of competition to the FAA, which essentially ignored the issue. For the foregoing reasons, the petition for review is, Granted.