American Great Lakes Ports Association v. Karl Schultz ( 2020 )


Menu:
  •  United States Court of Appeals
    FOR THE DISTRICT OF COLUMBIA CIRCUIT
    Argued January 22, 2020                 Decided June 16, 2020
    No. 18-5145
    AMERICAN GREAT LAKES PORTS ASSOCIATION, ET AL.,
    APPELLANTS
    v.
    KARL L. SCHULTZ, IN HIS OFFICIAL CAPACITY AS
    COMMANDANT, UNITED STATES COAST GUARD, ET AL.,
    APPELLEES
    Consolidated with 18-5167
    Appeals from the United States District Court
    for the District of Columbia
    (No. 1:16-cv-01019)
    C. Jonathan Benner argued the cause for appellants. With
    him on the briefs was Michael E. Deutsch. Kayla Grant entered
    an appearance.
    Jane M. Lyons, Assistant U.S. Attorney, argued the cause
    for federal appellees. With her on the brief were Jessie K. Liu,
    U.S. Attorney, and R. Craig Lawrence, Assistant U.S.
    Attorney. Jeremy S. Simon, Assistant U.S. Attorney, entered an
    appearance.
    2
    John Longstreth argued the cause for intervenor-
    appellees. With him on the brief was Mark Ruge.
    Before: SRINIVASAN, Chief Judge, and ROGERS and RAO,
    Circuit Judges.
    Opinion for the Court filed by Circuit Judge RAO.
    RAO, Circuit Judge: Ships engaged in foreign trade on the
    Great Lakes must use pilots registered pursuant to the Great
    Lakes Pilotage Act of 1960. The Coast Guard administers this
    licensing monopoly and sets rates for the American pilots,
    which has resulted in ongoing disputes between the Pilots and
    the Great Lakes commercial shipping and port interests
    (“Shippers”). This case requires us to resolve an
    Administrative Procedure Act challenge by the Shippers to the
    pilot rates for the 2016 commercial shipping season (“2016
    Rule”). The Shippers claim the 2016 Rule set an artificially
    inflated pilot rate that caused significant harm to the industry.
    The district court upheld parts of the 2016 Rule setting higher
    compensation targets for the Pilots, but held several parts of the
    Rule to be unsupported by the administrative record and
    remanded to the Coast Guard without vacating the Rule. We
    affirm the district court’s decision in full. Although remand
    without vacatur is the exception rather than the rule, in these
    circumstances, the district court acted within its discretion,
    given the disruption likely to occur from reallocating rates paid
    several years ago.
    I.
    The Great Lakes Pilotage Act requires foreign vessels and
    American vessels participating in foreign trade to hire an
    American or Canadian maritime pilot to assist in navigating the
    3
    difficult waters of the Great Lakes. See 46 U.S.C. §§ 9301–
    9308. The Act authorizes the Coast Guard to certify pilots,
    establish conditions of service, and set the rates that pilots must
    charge for their services. See 46 U.S.C. § 9303. Pursuant to this
    statutory authority, the Coast Guard has certified three pilotage
    associations to be the exclusive American providers of Great
    Lakes pilotage services in their assigned regions. See 81 Fed.
    Reg. 11,908, 11,910 (Mar. 7, 2016). When setting rates, the
    Coast Guard must consider “the public interest and the costs of
    providing the services.” 46 U.S.C. § 9303(f). The Coast Guard
    must “establish new pilotage rates by March 1 of each year,”
    id.,1 which the agency does through notice and comment
    rulemaking.
    After requests from the Pilots and Shippers, the Coast
    Guard proposed a new methodology to calculate Great Lakes
    pilot rates for the 2016 shipping season. The Coast Guard’s
    proposed rule was based largely upon the recommendations of
    the Great Lakes Pilotage Advisory Committee (GLPAC), an
    entity created by Congress in 1983 for the purpose of assisting
    the Coast Guard in formulating rates. See 80 Fed. Reg. 54,484,
    54,486 (Sept. 10, 2015); 46 U.S.C. § 9307(d)(2) (“The
    Secretary shall consider the information, advice, and
    1
    46 U.S.C. § 9303(f) reads in full:
    The Secretary shall prescribe by regulation rates and
    charges for pilotage services, giving consideration
    to the public interest and the costs of providing the
    services. The Secretary shall establish new pilotage
    rates by March 1 of each year. The Secretary shall
    establish base pilotage rates by a full ratemaking at
    least once every 5 years and shall conduct annual
    reviews of such base pilotage rates, and make
    adjustments to such base rates, in each intervening
    year.
    4
    recommendations of the Committee in formulating policy
    regarding matters affecting Great Lakes pilotage.”). The
    agency identified two reasons for changing the methodology.
    First, both the Pilots and the Shippers identified
    methodological issues that distorted the ratemaking
    calculation. The Pilots argued that the methodology resulted in
    artificially low rates that made it difficult to attract and retain
    pilots (harming “the public interest”) and the Shippers argued
    that the rates were artificially inflated (ignoring “the costs of
    providing the services”). See, e.g., 80 Fed. Reg. at 54,486.
    Second, the Coast Guard previously relied on union
    compensation data for similarly situated merchant marine
    masters and mates to help determine target pilot compensation;
    however, such data was no longer available from the union. See
    id. at 54,484.
    After the public comment period, the Coast Guard
    finalized the 2016 Rule largely along the lines initially
    proposed and consistent with the GLPAC recommendations.
    81 Fed. Reg. at 11,908. In adopting a new methodology, the
    Coast Guard found that the prior ratesetting undercompensated
    pilots, which resulted in pilot shortages and threats to vessel
    safety. The agency concluded rates must be increased to ensure
    a well qualified pool of pilots.
    Id. at 11,910.
    The new
    methodology was designed “to generate sufficient revenue for
    the pilots to provide the service [the public] require[s].” See
    id. at 11,909.
    To accomplish this, the Coast Guard, as relevant to
    this appeal, switched to the Peak Staffing Model, which pegged
    the number of necessary pilots to peak traffic periods in order
    to ensure the availability of rested pilots at all times. See 80
    Fed. Reg. at 54,489; 81 Fed. Reg. at 11,908–909. The agency
    also employed Canadian pilot compensation as a benchmark
    for compensation, plus a ten percent cost of living upward
    adjustment to incentivize American pilots to remain in the
    Great Lakes region. 81 Fed. Reg. at 11,914–915. Finally, the
    5
    Coast Guard estimated the rule would cost the shipping
    industry an additional $1.87 million annually, as well as a one-
    time $1.65 million expense to cover training.
    Id. at 11,937–38.
    The Shippers, represented by the American Great Lakes
    Ports Association, filed a lawsuit challenging the 2016 Rule
    under the Administrative Procedure Act in the United States
    District Court for the District of Columbia. The Shippers
    disputed the overall justification for the new methodology,
    questioning the agency’s conclusion that there was a
    compensation-driven pilot shortage in the Great Lakes region
    that could be remedied by increasing pilot rates. They also
    challenged the Coast Guard’s failure to consider “weighting
    factors”2 in the methodology; the setting of American pilot
    rates using Canadian pilot compensation with a ten percent
    upward adjustment; and the use of the new Peak Staffing
    Model to calculate rates.
    The district court rejected the Shippers’ overarching
    challenge to the Coast Guard’s new methodology. See Am.
    Great Lakes Ports Ass’n v. Zukunft, 
    296 F. Supp. 3d 27
    , 39–41
    (D.D.C. 2017). As an initial matter, the court held the Coast
    Guard’s decision to increase rates was not arbitrary and
    capricious because it rested on sufficient record evidence “even
    absent the empirical evidence demanded by Plaintiffs.”
    Id. at 39–41.
    The court also affirmed the Coast Guard’s use of the
    Peak Staffing Model to determine the number of rested pilots
    2
    Weighting factors “are multipliers that are used by pilotage
    associations to calculate the actual pilotage fees that the associations
    will charge for any given voyage. In essence, the larger the vessel,
    the higher the weighting factor, and the more the pilotage
    associations can charge.” Am. Great Lakes Ports Ass’n v. Zukunft,
    
    296 F. Supp. 3d 27
    , 35 n.5 (D.D.C. 2017). The Shippers contended
    that the Coast Guard’s failure to account for these factors artificially
    inflated the pilot rates.
    6
    needed throughout the season because one of the agency’s
    rationales—safety—was “amply supported” by the record.
    Id. at 42–43.
    The court, however, held two aspects of the Coast
    Guard’s new methodology to be unsupported by the record.
    First, the court held there was no reasoned basis for setting
    American pilot compensation by reference to a ten percent
    increase over the base Canadian compensation rate because the
    figure came from unidentified comments during a GLPAC
    meeting.
    Id. at 46–48.
    Second, the court held that the Coast
    Guard acted arbitrarily by failing to account for increased pilot
    revenue from vessel weighting factors, resulting in a potential
    overcharge to the Shippers.
    Id. at 51–52.
    Noting the difficulty
    of crafting a remedy, the district court instructed the parties to
    file supplemental briefing on the appropriate remedy.
    Id. at 56.
    Following additional briefing, the district court
    determined, in a separate published opinion, that remand
    without vacatur was the appropriate remedy. See Am. Great
    Lakes Ports Ass’n v. Zukunft, 
    301 F. Supp. 3d 99
    , 104–05
    (D.D.C. 2018). The Shippers urged the court to vacate the 2016
    Rule and order various forms of prospective and retroactive
    relief to compensate for the unlawfully inflated rates.
    Id. at 102.
    The district court noted the Shippers’ requests demonstrated
    they “fundamentally misunderstand th[e] Court’s prior ruling”
    because the court had not held that the rates were too high, but
    instead that the rates were not justified by the administrative
    record.
    Id. The court
    found the errors in the 2016 Rule were
    substantial because the Coast Guard provided no support for
    pegging compensation to Canadian pilots or for its failure to
    include weighting factors (which were included in the 2017
    rate review).
    Id. at 103.
    Although the court determined the
    Coast Guard’s errors to be significant, it remanded without
    vacatur because the “disruptive consequences” of vacatur
    outweighed the seriousness of the errors.
    Id. at 103–05.
    The
    court noted it was unclear “whether and to what extent the
    7
    pilotage associations might be required to issue refunds” in
    response to vacatur but that “to the extent that they would be
    required to do so, the disruptive consequences are clear.”
    Id. at 104.
    The court thus remanded to the Coast Guard to “evaluate
    and justify an appropriate adjustment to benchmark
    compensation for its ratemaking methodology going forward.”
    Id. at 105.
    The Shippers appeal, arguing the district court erred in
    affirming parts of the 2016 Rule and further abused its
    discretion in declining to vacate the Rule despite finding
    significant parts of it unsupported by the record.
    II.
    As an initial matter, this court must assure itself that it has
    jurisdiction over the Shippers’ appeal of the district court’s
    remand order. See Steel Co. v. Citizens for a Better Env’t, 
    523 U.S. 83
    , 94–95 (1998). The courts of appeals have jurisdiction
    over “all final decisions of the district courts of the United
    States,” 28 U.S.C. § 1291, and “[a] remand order usually is not
    a final decision,” NAACP v. United States Sugar Corp., 
    84 F.3d 1432
    , 1436 (D.C. Cir. 1996). We have noted, however, that
    “[t]he requirement of finality is to be given a practical rather
    than a technical construction.” Limnia, Inc. v. Dep’t of Energy,
    
    857 F.3d 379
    , 385 (D.C. Cir. 2017) (internal quotation marks
    omitted). A remand order “that terminate[s] an action fall[s]
    within the core of Section 1291’s requirement of finality.”
    Id. (internal quotation
    marks omitted).
    Here, the district court’s remand order effectively
    terminates the Shippers’ action. This appeal presents the only
    opportunity for the Shippers to challenge the remand order.
    Although “a remand for significant further proceedings”
    generally constitutes a nonfinal order, Pueblo of Sandia v.
    Babbitt, 
    231 F.3d 878
    , 881 (D.C. Cir. 2000), the remand order
    8
    under review does not instruct the Coast Guard to reopen the
    2016 rate review and conduct further proceedings. See Am.
    Great Lakes Ports 
    Ass’n, 301 F. Supp. 3d at 105
    . The order
    thus finally disposed of the Shippers’ APA challenges to the
    2016 Rule. See In re Long-Distance Tel. Serv. Fed. Excise Tax
    Refund Litig., 
    751 F.3d 629
    , 633 (D.C. Cir. 2014) (“[T]reating
    the district court’s remand order as unappealable would
    effectively preclude … plaintiffs from ever challenging the
    district court’s decisions.” (internal quotation marks omitted)).
    Moreover, it is relevant, although not dispositive, see 
    Limnia, 857 F.3d at 386
    , that the district court characterized its remand
    order as “a final appealable Order.” J.A. 119. In sum, we have
    jurisdiction over the Shippers’ appeal under 28 U.S.C. § 1291.
    III.
    Turning to the merits, the Shippers challenge the Coast
    Guard’s determination that a compensation-driven pilot
    retention crisis necessitated increased rates and that the Peak
    Staffing Model should be used to calculate rates. We review
    such challenges “de novo, evaluating the administrative record
    directly and invalidating the [agency’s] actions only if, based
    on that record, they are arbitrary, capricious, an abuse of
    discretion, or otherwise not in accordance with law.” Stand Up
    for California! v. Dep’t of Interior, 
    879 F.3d 1177
    , 1181 (D.C.
    Cir. 2018) (internal quotation marks omitted).
    A.
    The Shippers first challenge the Coast Guard’s finding that
    a compensation-driven pilot shortage was developing on the
    Great Lakes, endangering both safety and efficient commerce.
    We agree with the Coast Guard that the record contains ample
    evidence of a pilot shortage on the Great Lakes and that the
    shortage was caused by low compensation.
    9
    As to the shortage, the Coast Guard explained that the
    number of pilots has decreased significantly, with thirty-one
    pilots leaving the system. The rate of attrition has exceeded the
    rate of replacement, resulting in a net decrease of twenty-two
    percent, from forty-four to thirty-six pilots in less than a
    decade. See 81 Fed. Reg. at 11,919. The 2016 Rule includes
    abundant evidence that this decline has brought pilot numbers
    below acceptable levels, impairing the safety of Great Lakes
    navigation. See, e.g.,
    id. at 11,918–921.
    The agency also cited significant evidence demonstrating
    that low compensation drove this pilot shortage. For instance,
    comments from affected individuals and entities supported the
    Coast Guard’s conclusion that compensation was a key factor
    in the retention and recruitment of pilots. Former pilots and
    trainees stated they “resigned because of low pay and long
    hours.” 81 Fed. Reg. at 11,919. Hiring agents noted that “low
    compensation and long hours … kept many highly qualified
    mariners from signing on as pilots.”
    Id. Finally, one
    pilot
    commented that “10 pilots in his association took early
    retirement to escape the low compensation and long hours.”
    Id. The administrative
    record thus indicated that the Great Lakes
    have a “retention and attraction problem” because its pilots are
    “the lowest paid pilots in America” and “have the highest
    workload in America.” J.A. 468 (statement from Pilots’
    representative at GLPAC meeting).
    The Shippers primarily challenge the agency’s reliance on
    public comments, rather than empirical evidence. Appellant Br.
    37–40. The APA, however, “imposes no general obligation on
    agencies to produce empirical evidence.” Stilwell v. Office of
    Thrift Supervision, 
    569 F.3d 514
    , 519 (D.C. Cir. 2009). The
    Coast Guard’s heavy reliance on comments was legitimate in
    this proceeding. The Great Lakes pilotage labor market is
    small—approximately thirty-six pilots—and the agency
    10
    received comments from a large segment of current and former
    pilots, who consistently cited compensation as a leading reason
    for the pilot shortage. See 81 Fed. Reg. at 11,917–920. “A
    degree of agency reliance on [comments from affected parties]
    is not only permissible but often unavoidable.” Nat’l Ass’n of
    Regulatory Util. Comm’rs v. FCC, 
    737 F.2d 1095
    , 1124 (D.C.
    Cir. 1984). The Shippers provide no contrary evidence,
    empirical or otherwise, to suggest these firsthand accounts are
    not representative of pilots at large. “[A]n agency need not—
    indeed cannot—base its every action upon empirical data;
    depending upon the nature of the problem.” Chamber of
    Commerce of U.S. v. SEC, 
    412 F.3d 133
    , 142 (D.C. Cir. 2005).
    It was entirely reasonable for the agency to rely on reasons
    provided by the hiring agents, pilots, and former pilots to
    determine that an increase in compensation would aid retention
    and recruitment of Great Lakes pilots.
    Moreover, the Coast Guard’s conclusion that
    compensation is a major cause of the shortage was supported
    by the recommendations of the Great Lakes Pilotage Advisory
    Committee (GLPAC). See 80 Fed. Reg. at 54,486; 81 Fed. Reg.
    at 11,911 (citing GLPAC recommendation). Congress directed
    that the Coast Guard “shall consider” GLPAC’s
    recommendations, 46 U.S.C. § 9307(d)(2), and here GLPAC
    unanimously recommended many aspects of the methodology
    the agency adopted in the 2016 Rule. See 80 Fed. Reg. at
    54,486. In addition, the agency relied upon an expert evaluation
    in the Coast Guard-commissioned Bridge Hour Study, which
    identified growing concern “regarding the available candidate
    pool to replace pilots who will soon be retiring” and cited
    inadequate pay “as a leading issue” alongside “stability of
    pay,” “quality of life,” and other issues. J.A. 495; 81 Fed. Reg.
    at 11,908.
    11
    The administrative record thus included a wealth of
    evidence supporting the agency’s conclusion that if it “fail[ed]
    to implement this methodology and new rates, … the pilot
    associations will not be able to recruit experienced mariners
    and retain their registered pilots.” 81 Fed. Reg. at 11,925; cf.
    FCC v. Nat’l Citizens Comm. for Broad., 
    436 U.S. 775
    , 814
    (1978) (“[C]omplete factual support in the record for the
    [agency’s] judgment or prediction is not possible or required; a
    forecast of the direction in which future public interest lies
    necessarily involves deductions based on the expert knowledge
    of the agency.” (internal quotation marks omitted)). In sum, the
    Coast Guard reasonably concluded that “increased pilot rates
    are the best and quickest way to attract and retain more
    qualified pilots.” 81 Fed. Reg. at 11,921.
    B.
    The Shippers also challenge the Peak Staffing Model as
    arbitrary and capricious. Part of the Coast Guard’s new
    ratesetting methodology, the Model seeks to ensure the pilot
    rate is sufficient to cover the cost of employing “the number of
    pilots needed to meet each shipping season’s peak pilotage
    demand periods without interruption to service.” 80 Fed. Reg.
    at 54,489. In the 2016 Rule, the Coast Guard predicted the
    Model would reduce delays caused by insufficient staffing and
    also increase safe pilotage practices. See 81 Fed. Reg. at
    11,922. The district court found the delay rationale
    “inconclusive” but could not “say that the evidence necessarily
    runs counter to the Coast Guard’s conclusion.” Am. Great
    Lakes Ports 
    Ass’n, 296 F. Supp. 3d at 42
    . The court, however,
    found the record “amply supported the Coast Guard’s
    conclusion that greater staffing levels were needed to improve
    safe pilotage on the Great Lakes.”
    Id. at 43.
    The Shippers
    12
    maintain that the Coast Guard’s explanation for its reliance on
    the Model is inadequate.
    We find that the agency adequately supported the Peak
    Staffing Model on the grounds that it furthered safe pilotage.
    When setting rates, the Coast Guard must consider “the public
    interest,” 46 U.S.C. § 9303(f), and this court has interpreted the
    public interest to include safe pilotage. See Menkes v. DHS, 
    637 F.3d 319
    , 334 (D.C. Cir. 2011); accord
    id. at 351
    (Brown, J.,
    dissenting in part) (noting that the text and structure of the Act
    “promotes maritime safety”). The Coast Guard concluded the
    Peak Staffing Model is necessary to ensure there are enough
    pilots on hand throughout the season to cover surge periods.
    See 81 Fed. Reg. at 11,922. Without pegging staffing needs to
    the peak periods, which may occur anytime during the season,
    situations will inevitably arise in which pilots will have to forgo
    the necessary rest periods recommended by the National
    Transportation Safety Board (“NTSB”). See
    id. (“Setting pilot
    numbers high enough to accommodate all these peak periods is
    essential … to provide the recuperative monthly rest periods
    recommended by the NTSB in the interests of safety.”). Citing
    multiple reports, commenters, and the NTSB recommendation,
    the agency’s decision rests on an ample record supporting the
    conclusion that the shortage of rested pilots endangered safety.
    See, e.g., 81 Fed. Reg. at 11,918–922. Although the record does
    not support the reducing delay rationale,3 the public safety
    3
    Like the district court, we find the record inconclusive regarding a
    correlation between delays and staffing levels. The Coast Guard’s
    support for the delay rationale consists of a series of charts
    comparing pilot staffing levels and delay hours. 81 Fed. Reg. at
    11,920–921. The charts compare only two variables, pilot staffing
    and delay hours, and fail to demonstrate a correlation between the
    two. See
    id. The Coast
    Guard acknowledged that “[o]ther factors
    contribute to delays,” but provided no analysis of the various factors
    that cause delays.
    Id. at 11,921.
                                     13
    rationale provides an independent and sufficient reason for the
    Peak Staffing Model.
    In lieu of the Peak Staffing Model, which required
    maintaining a greater number of pilots at all times, the Shippers
    assert that part time contractors could supplement full time
    Great Lakes pilots during peak periods. Yet the Coast Guard
    specifically determined this alternative would be unsafe
    because contract pilots would not have knowledge of local
    waters and weather conditions in the Great Lakes and could not
    gain the necessary experience on short notice. See 81 Fed. Reg.
    at 11,922. Moreover, the agency explained that other pilotage
    systems are not proper comparators because “those systems
    cover smaller areas in which those pilots more easily can
    maintain the necessary knowledge without impacting safety.”
    Id. With ample
    record evidence, we decline to second guess the
    Coast Guard’s expertise over maritime safety and uphold the
    use of the Peak Staffing Model.
    IV.
    The Shippers prevailed below in their challenges to the
    failure to consider weighting factors and to the use of an
    upwardly adjusted Canadian benchmark for target pilot
    compensation. See Am. Great Lakes Ports Ass’n, 
    301 F. Supp. 3d
    at 101. Neither the Coast Guard nor the Pilots attempt to
    defend these aspects of the 2016 Rule on appeal.4 The Shippers,
    4
    In subsequent rate adjustments, the Coast Guard addressed both
    issues by including weighting factors in the calculation and by
    pegging target compensation to union data (available from earlier
    years) adjusted for inflation and Great Lakes conditions, rather than
    the Canadian pilots’ compensation. See 84 Fed. Reg. 20,551, 20,553
    (May 10, 2019) (“In 2017, we added … new steps that accurately
    account for the … weighting factors …. In 2018, we revised the
    methodology by which we develop the compensation benchmark.”).
    14
    however, argue that the district court’s decision to remand
    without vacatur was an abuse of discretion because it
    effectively precludes the Shippers from obtaining relief from
    the agency’s arbitrary and capricious action.
    Although “vacatur is the normal remedy” under the APA,
    our precedents permit a court to remand without vacating the
    agency’s action in limited circumstances. Allina Health Servs.
    v. Sebelius, 
    746 F.3d 1102
    , 1110 (D.C. Cir. 2014). To
    determine whether to remand without vacatur, this court
    considers first, “the ‘seriousness of the [action’s]
    deficiencies,’” and, second, the “likely ‘disruptive
    consequences’ of vacatur.”
    Id. (quoting Allied-Signal,
    Inc. v.
    Nuclear Regulatory Comm’n, 
    988 F.2d 146
    , 150–51 (D.C. Cir.
    1993)). The Coast Guard and Pilots do not dispute the
    seriousness of the errors on appeal, and so we see no need to
    disturb the district court’s finding. See Am. Great Lakes Ports
    Ass’n, 
    301 F. Supp. 3d
    at 103 (“[T]he Court cannot conclude
    that the defects in the Coast Guard’s 2016 Rule were not
    serious.”). We focus on the parties’ main point of contention,
    namely the district court’s assessment of the disruptive
    consequences that would flow from vacating the 2016 Rule.
    The Shippers contend that the district court could “easily
    remedy” the error by “ordering a review and repricing of the
    invoices from the Pilotage Associations to the shipping
    companies under the rates established under the prior 2015
    Rule.” Appellant Br. 49. Yet the district court identified
    This controversy remains live, however, because the agency has not
    attempted to compensate the Shippers for potential overcharges
    caused by the errors in the 2016 Rule. Cf. Cape Cod Hosp. v.
    Sebelius, 
    630 F.3d 203
    , 210 (D.C. Cir. 2011) (“[S]ince the 2008 rule
    in no way compensated for any underpayments that might have been
    made in 2007, a live controversy remains regarding the hospitals’
    objection to the 2007 rule.”).
    15
    numerous disruptive consequences that would follow from
    vacating the 2016 Rule. Our review comes nearly four years
    after rates have been paid in reliance on the 2016 Rule. As the
    district court found, vacatur would mean that “every payment
    that was made in the 2016 season was erroneous,” Am. Great
    Lakes Ports Ass’n, 
    301 F. Supp. 3d
    at 104, and may involve the
    Coast Guard and the Shippers attempting to recoup and
    redistribute funds that changed hands years ago in numerous
    separate transactions. Cf. Milk Train v. Veneman, 
    310 F.3d 747
    ,
    756 (D.C. Cir. 2002) (noting disruptive consequences from
    attempting to retrieve funds erroneously disbursed and likely
    invested by the recipients because “those moneys may not be
    recoverable three years later”). Further, the precise amount of
    each refund would be unclear given the lack of an operative
    2016 rate. See Am. Great Lakes Ports Ass’n, 
    301 F. Supp. 3d
    at 104.
    Under our precedents, a quintessential disruptive
    consequence arises when an agency cannot easily unravel a
    past transaction in order to impose a new outcome. We have
    rejected approaches similar to the Shippers’ proposed
    reinvoicing as “an invitation to chaos.” Sugar Cane Growers v.
    Veneman, 
    289 F.3d 89
    , 97 (D.C. Cir. 2002). More generally,
    although remand without vacatur remains an exceptional
    remedy, we have held that it is appropriate when vacatur would
    disrupt settled transactions. See Milk 
    Train, 310 F.3d at 756
    ;
    Sugar Cane 
    Growers, 289 F.3d at 97
    (“The egg has been
    scrambled and there is no apparent way to restore the status quo
    ante.”). Because the district court acted within our precedents
    and its remedial discretion, we affirm its order remanding the
    2016 Rule to the agency without vacatur. Cf. Stand Up for
    
    California!, 879 F.3d at 1190
    (holding that “the district court
    acted well within its discretion in finding vacatur unnecessary
    to address any harm the defect had caused”).
    16
    We recognize that a remand without vacatur does not
    provide the Shippers with the complete relief they sought, and
    agencies often delay or decline to take action in these
    circumstances. See In re Core Commc’ns, Inc., 
    531 F.3d 849
    ,
    862 (D.C. Cir. 2008) (Griffith, J., concurring) (“[E]xperience
    suggests that [remand without vacatur] sometimes invites
    agency indifference.”). The limits of a judicial remedy here
    stem in part from this particular statutory scheme, which
    requires the Coast Guard to review and adjust pilot rates
    annually. By the time an appeal winds its way to us, the egg
    will often be scrambled, leaving us with few options other than
    telling the Coast Guard, in effect, to “try better next time.”5 In
    its next annual rate review, the agency should consider if it has
    the statutory authority to remedy the harms from the 2016 Rule
    and if doing so would comport with its mandate to consider
    “the public interest and the costs of providing the services.” 46
    U.S.C. § 9303(f).6
    5
    Government ratesetting in this regulatory program frequently
    leaves either the Pilots or Shippers dissatisfied. As one district court
    recently observed, “each year, it seems, either the shipping
    companies or the associations that supply the pilots sue the Coast
    Guard to challenge aspects of the rulemaking. The shippers
    perennially complain that the rates are too high, while the pilots gripe
    that they are too low.” Am. Great Lake Ports Ass’n v. Coast Guard,
    18-cv-2650, 
    2020 WL 1157028
    , at *1 (D.D.C. Mar. 10, 2020); see
    also St. Lawrence Seaway Pilots Ass’n v. Coast Guard, 
    357 F. Supp. 3d
    30 (D.D.C. 2019); St. Lawrence Seaway Pilots Ass’n, Inc. v. Coast
    Guard, 
    85 F. Supp. 3d 197
    (D.D.C. 2015).
    6
    At oral argument, the government suggested the agency could
    account for potential overcharges to the Shippers in a future rate
    adjustment. See Oral Arg. 27:40–29:35. We take no position on
    whether the Great Lakes Pilotage Act provides the Coast Guard with
    the authority to take such an action.
    17
    ***
    For the foregoing reasons, we affirm the district court’s
    decision in full.
    So ordered.