United States Court of Appeals FOR THE DISTRICT OF COLUMBIA CIRCUIT
Argued January 10, 2022 Decided August 16, 2022
No. 21-1088
BLOOMBERG L.P., PETITIONER
v.
SECURITIES AND EXCHANGE COMMISSION, RESPONDENT
FINANCIAL INDUSTRY REGULATORY AUTHORITY, INTERVENOR
On Petition for Review of an Order of the Securities and Exchange Commission
Keith Bradley argued the cause for petitioner. With him on the briefs was ScheLeese Goudy.
Thomas A. Burns was on the brief for amicus curiae Healthy Markets Association in support of petitioner.
Theodore J. Weiman, Senior Litigation Counsel, Securities and Exchange Commission, argued the cause for respondent. With him on the brief were Michael A. Conley, Solicitor, and Tracey A. Hardin, Assistant General Counsel. 2 Nowell D. Bamberger was on the brief for intervenor Financial Industry Regulatory Authority, Inc. in support of respondent. Giovanni P. Prezioso entered an appearance.
Paul S. Mishkin and David B. Toscano were on the brief for amici curiae Intercontinental Exchange, Inc., et al. in support of respondent.
Before: WILKINS, KATSAS and JACKSON, Circuit Judges.
Opinion for the Court filed by Circuit Judge WILKINS.
WILKINS, Circuit Judge: Petitioner Bloomberg L.P. (“Bloomberg”) seeks review of the Securities and Exchange Commission’s (the “Commission” or “SEC”) decision to approve new reporting requirements proposed by the Financial Industry Regulatory Authority, Inc. (“FINRA”), Intervenor- for-Respondent, affecting underwriter members in the corporate bond market. FINRA represented to the SEC that market inefficiencies in the corporate bond market reduce market participation, decrease liquidity, and increase transaction and opportunity costs. To address these problems, FINRA proposed to consolidate and provide market-wide access to “core” reference data for new issues of corporate bonds through a subscription-based service. The Commission ultimately concluded that FINRA’s proposal would impose a limited burden on competition and enable market participants to obtain broad, uniform access to corporate bond reference data before the first transaction in a new-issue bond. Accordingly, the Commission approved FINRA’s proposal.
Circuit Judge, now Justice, Jackson was a member of the panel at the time the case was argued but did not participate in this opinion. 3 Importantly, though, during the rulemaking process, various commenters raised concerns about FINRA’s proposed data service. In relevant part, Bloomberg commented that FINRA did not provide any information about how much it will cost to build and maintain the database, and to what extent FINRA will pass those costs along to market participants.
For the reasons explained below, we find that pursuant to the Administrative Procedure Act, 5 U.S.C. § 706(2)(A), the Commission’s approval of FINRA’s proposal was arbitrary and capricious because the Commission neglected to give a reasoned explanation in response to Bloomberg’s significant concerns about the costs that FINRA, as well as market participants, will incur in connection to the creation and maintenance of the data service. Accordingly, we grant Bloomberg’s petition for review on the grounds that the Commission’s failure to respond to significant public comments about the costs associated with FINRA’s proposal was arbitrary and capricious. We deny Bloomberg’s petition for review with respect to its remaining arguments. We remand without vacatur for the Commission to respond appropriately.
I.
FINRA is a private association of securities broker-dealers that regulates the conduct of broker-dealers—people or firms in the business of buying or selling securities on behalf of customers, their own accounts or both—and formulates and enforces standards for trading and brokerage. Because of its status as a self-regulatory organization, Section 19(b)(2)(C)(i) of the Securities and Exchange Act of 1934 (the “Exchange Act” or “Act”), 15 U.S.C. § 78s(b)(2)(C)(i), imposes on FINRA “a duty to promulgate and enforce rules governing the conduct of [its] members, under the oversight of the SEC.” See NetCoalition v. SEC, 715 F.3d 342, 344–45 (D.C. Cir. 2013) 4 (“NetCoalition II”) (internal quotation marks and citations omitted). The Act provides that FINRA must submit any proposed rules or rule changes to the SEC, which “shall approve” proposals if it finds them “consistent with the requirements of” the Exchange Act and applicable SEC rules and regulations. 15 U.S.C. § 78s(b)(2)(C)(i).
Bloomberg is a global business and financial information company. Among other things, it is a private data vendor that dominates the market for private data services. See Pet’r’s Opening Br. at 8. Bloomberg obtains newly issued bond reference data from underwriters—usually before other data vendors and market participants—and then sells that data to market participants.
Reference data is very important. When corporations issue bonds, the underwriters who buy and sell those bonds to market participants compile certain reference data—details and terms, such as the name of the issuer and identification numbers—and provide them to data vendors like Bloomberg. Without reference data, trading platforms cannot list a bond for trading.
Timing matters, too. The hours and days that immediately follow a bond offering are generally a highly active trading period. Accordingly, market participants, namely investors, seek out timely reference data about new corporate bonds coming to market from vendors like Bloomberg. According to FINRA, because each reference data provider collects and distributes new issue reference data “from different sources and at different speeds[,]” the reference data is not necessarily “consistent, timely and accurate across reference data providers.” J.A. 1–2. 5 II.
In 2017, the SEC established the Fixed Income Market Structure Advisory Committee (“FIMSAC” or “Advisory Committee”) to offer advice and recommendations to the Commission on the structure of the fixed income market. (Corporate bonds are considered fixed income securities.) FIMSAC is composed of individuals representing a range of perspectives on the fixed income markets, including corporate bond investors, broker-dealers, underwriters, academics, and data vendors. FIMSAC’s Technology and Electronic Trading Subcommittee (the “Subcommittee”), which represents a cross-section of fixed income market participants, determined that there are gaps in the corporate bond reference data market, both in terms of when data vendors make data available and the data’s accuracy. The Subcommittee attributed these disparities to (1) the fact that private data vendors—like Bloomberg—are not obligated to provide impartial access to new issue reference data; (2) vendors lacking equal access to information from underwriters; and (3) the resulting confusion, which increases transaction costs and impedes competition in the corporate bond market.
The Subcommittee met several times over the course of seven months to fill the gaps in the corporate bond reference data market. It ultimately recommended that FINRA, as an impartial entity, establish a consolidated new issue reference data service that makes core reference data available to all subscribers in a timely and commercially reasonable manner.
The Subcommittee’s recommendation paralleled another proposal the SEC had previously approved. In 2008, the SEC accepted the Municipal Securities Rulemaking Board’s proposal to establish a centralized, self-regulatory organization-mandated data service that would provide 6 participants in the municipal bond market with timely access to “core” reference data and streamline the trade of municipal bonds. The petition before us concerns a proposal for a similar service in the corporate bond market.
FIMSAC unanimously endorsed the Subcommittee’s recommendation. In October 2018, FIMSAC recommended that FINRA establish a corporate bond new issue reference data service—similar to the municipal bond market service that the SEC approved in 2008—to mitigate the disparities in the market for corporate new issue reference data.
A.
To that end, FINRA performed an Economic Impact Assessment: an analysis of the corporate bond market. FINRA coordinated independent outreach to eleven participants in the corporate bond market: four data vendors, three underwriters, two trading platforms, and two clearing firms. Insights from these entities led FINRA to discover the same problems FIMSAC did: data vendors receive reference data through different channels at different times, which leads to untimely, inconsistent, and inaccurate distribution of corporate bond new issue reference data to market participants. Thereafter, FINRA set out to develop a solution that would address gaps in the availability of accurate, complete, and timely access to corporate bond new issue reference data.
FINRA developed a proposal to establish a centralized new issue reference data service for corporate bonds. First, FINRA would require its member underwriters to report to FINRA—before trading on a new issue begins—32 unique data elements that are integral to the valuation, trade, and settlement of corporate bonds. Collectively, these elements constitute “core” reference data. Next, FINRA would create a database 7 to consolidate the data and provide market-wide access through a fee subscription.
B.
On March 27, 2019, pursuant to Section 19(b)(1) of the Exchange Act, 15 U.S.C. § 78s(b)(1), FINRA proposed to the SEC the establishment of a consolidated fee-based data service that would provide market-wide access to new issue reference data.
FINRA Rule 6760 “requires certain data elements—those sufficient to identify the security accurately—to be reported before the execution of the first transaction, and all remaining data elements to be reported within 15 minutes of the Time of Execution of the first transaction.” J.A. 2 n.4. Rule 6760(b) currently requires underwriters to provide certain new issue reference data to FINRA. FINRA’s proposal would amend Rule 6760(b) to require underwriters to report additional data elements to FINRA. In relevant part, the amended version of Rule 6760(b) would require underwriters to report the following core information:
(A) The International Securities Identification Number (ISIN); (B) the currency; (C) the issue date; (D) the first settle date; (E) the interest accrual date; (F) the day count description; (G) the coupon frequency; (H) the first coupon payment date; (I) a Regulation S indicator; (J) the security type; (K) the bond type; (L) the first coupon period type; (M) a convertible indicator; (N) a call indicator; (O) the first call date; (P) a put indicator; (Q) the first put date; (R) the minimum increment; (S) the minimum piece/denomination; (T) the issuance amount; 8 (U) the first call price; (V) the first put price; (W) the coupon type; (X) rating (TRACE Grade); (Y) a perpetual maturity indicator; (Z) a Payment-In-Kind (PIK) indicator; (AA) first conversion date; (BB) first conversion ratio; (CC) spread; (DD) reference rate; (EE) floor; and (FF) underlying entity ticker.
J.A. 281. FINRA represented to the SEC that it would submit a separate filing to establish fees for the service and would implement its proposed service if those fees are adopted.1 Id.
C.
In April 2019, the Commission published notice of FINRA’s proposed rule change in the Federal Register. See Notice of Filing of a Proposed Rule Change to Establish a Corporate Bond New Issue Reference Data Service, 84 Fed. Reg. 13,977 (Apr. 8, 2019). On July 1, 2019, the Commission instituted proceedings under Section 19(b)(2)(B) of the Exchange Act, 15 U.S.C. § 78s(b)(2)(B), to determine whether to approve the rule change.
1 Initially, in its March 2019 proposal, FINRA proposed that it would make the new-issue data available to subscribers for a fee of $250 per month (for internal use) or $6,000 per month (to retransmit the data to other users). On October 3, 2019, FINRA filed Partial Amendment No. 1, which was withdrawn because of a non- substantive administrative error. That same day, FINRA filed Partial Amendment No. 2, in which it removed the fee component and tweaked some of the core data fields required from underwriters. FINRA withdrew the proposed subscription fees because commenters urged FINRA to provide more information to justify the fees. Upon withdrawal, FINRA stated that it would further evaluate the appropriate fee structure. 9 The Commission received 30 comments in response to FINRA’s proposal, including five from Bloomberg, one from FINRA, and one from FIMSAC. Supporters of FINRA’s proposal, including Intercontinental Exchange, Inc. (“ICE”) Bonds, ICE Data Services, and Charles River Development, all of whom are market participants, commented “that currently there is no uniform, universally available mechanism for providing market participants with consistent and timely access to reference data about corporate bonds on the day a newly issued corporate bond commences trading,” J.A. 193, and explained that timely access to reference data is crucial for the valuation, trade, and settlement of corporate bonds. Supporters of FINRA’s proposal also pointed to another issue with the current market for reference points: reference data is not available to all market participants before the trading of a new issue, which puts certain market participants at a competitive disadvantage. Ultimately, supporters of FINRA’s proposal contended that a FINRA-operated centralized data reporting requirement for new corporate bond issues would increase the efficiency of the corporate bond market and reduce trading and research costs.
Opponents of FINRA’s proposal, including Bloomberg, the Heritage Foundation, and the Healthy Markets Association, submitted comments arguing that FINRA failed to establish the existence of a market structure problem that requires regulatory intervention, as mandated under Section 15A(b)(6) of the Exchange Act, 15 U.S.C. § 78o–3(b)(6). Assuming that a problem exists, Bloomberg suggested that it was “questionable whether a single [self-regulatory organization] would provide more accurate, complete and timely service than competing private sector providers” and FINRA provided no evidence that its proposal would reduce broken trades and errors. J.A. 193 n.28 (internal quotation marks and citation omitted). Further, Bloomberg submitted that the impact of any errors in a 10 centralized system would be magnified. Id. The U.S. Chamber of Commerce’s Center for Capital Markets Competitiveness commented that FINRA’s proposal would increase regulatory and liability burdens for underwriters without any clear benefit. Bloomberg echoed this concern, contending that the rule’s imposition of additional burdens on underwriters would disproportionately impact smaller underwriters.
In addition, Bloomberg commented that FINRA’s proposal was “antithetical to the most foundational principles of administrative law and cost-benefit analysis” because FINRA failed to quantify the direct and indirect costs of its proposed service (or explain why certain costs could not be quantified). J.A. 162. Bloomberg cautioned that self- regulatory organizations and agencies should not be “permitted to ignore regulatory burdens in this manner.” Id. Otherwise, “agencies could propose laudable programs heedless of their pricetags, seek their provisional approval without respect to cost, and then—once established—propose a fee that was by now necessary to sustain an already approved program.” Id. “[A]t a minimum,” Bloomberg commented, “the Commission must . . . know what [the] costs [of the proposed rule] are.” See id.
On December 4, 2019, the SEC’s Division of Trading and Markets (the “Division”), which is authorized by the SEC and federal regulations to approve proposed rule changes, accepted FINRA’s proposal and concluded that it was consistent with Sections 15A(b)(6) and 15A(b)(9) of the Act. See 15 U.S.C. §§ 78o–3(b)(6), (9).
On December 18, 2019, Bloomberg petitioned the Commission for review of the Division’s December 2019 approval order. The Commission conducted a de novo review 11 of whether FINRA’s proposal is consistent with the Exchange Act and applicable rules and regulations.
III.
Section 15A(b) of the Exchange Act, 15 U.S.C. § 78o– 3(b), lays out numerous requirements the SEC must consider when reviewing a rule change. Section 15A(b)(5) requires that “[t]he rules of the association provide for the equitable allocation of reasonable dues, fees, and other charges among members and issuers and other persons using any facility or system which the association operates or controls.” Id. § 78o– 3(b)(5). Section 15A(b)(6) provides that FINRA’s rules must be “designed” “to promote just and equitable principles of trade”; “to foster cooperation and coordination with persons engaged in regulating, clearing, settling, processing information with respect to, and facilitating transactions in securities”; “to remove impediments to and perfect the mechanism of a free and open market and a national market system”; and “in general, to protect investors and the public interest.” Id. § 78o–3(b)(6). Finally, Section 15A(b)(9) states that FINRA’s rules may not “impose any burden on competition not necessary or appropriate.” Id. § 78o–3(b)(9).
As a general matter, when the SEC is engaged in the review of a rule of a self-regulatory organization like FINRA, Section 3(f) of the Exchange Act provides that the SEC must consider “whether the action will promote efficiency, competition, and capital formation.” Id. § 78c(f). As mentioned, the SEC “shall approve” a rule proposal if it is “consistent with the requirements of” the Exchange Act and applicable SEC rules and regulations. Id. § 78s(b)(2)(C)(i). It is ultimately FINRA’s burden to demonstrate that its proposed rule change is consistent with the Act and applicable rules and regulations. 17 C.F.R. § 201.700(b)(3). 12
On January 15, 2021, the Commission upheld the Division’s approval order, finding FINRA’s proposed rule change consistent with the Exchange Act, namely Section 3(f), 15 U.S.C. § 78c(f), Section 15A(b)(6), id. § 78o–3(b)(6), and Section 15A(b)(9), id. § 78o–3(b)(9).
In its order, the Commission declared that “there is an inefficiency in the collection and availability of reference data for newly issued corporate bonds.” J.A. 281 (footnote omitted). “[T]his inefficiency results in an information asymmetry in the market for newly issued corporate bond reference data that can disadvantage many market participants.” Id. Further, “[t]his information asymmetry inhibits these market participants from transacting in the secondary market for newly issued bonds. . . .” Id. The Commission concluded that “FINRA’s proposal is reasonably designed to address this information asymmetry to the benefit of the marketplace” by “mak[ing] certain reference data available to market participants in a timely, accessible, and impartial manner.” Id.
The Commission also addressed several issues raised by commenters, which included, among other things, (1) “whether information asymmetry [actually] exists in the current marketplace”; (2) “the proposal’s burden on underwriters”; (3) “the proposal’s effect on competition among reference data vendors”; and (4) “the lack of information regarding fees” for FINRA’s data service. Id. at 282.
First, the Commission reasoned that the record—namely comments provided by reference data providers at the October 2018 FIMSAC meeting and FINRA’s outreach to market 13 participants through its Economic Impact Assessment— establishes the existence of a reference data access problem. FIMSAC’s meeting transcript provides that Bloomberg “typically has the timeliest access to newly issued bond reference data on the first day a bond trades, as it enjoys the voluntary cooperation of underwriters,” id. at 286 n.91, while other data vendors face challenges with collecting and distributing reference data. The meeting transcript also reflects that Bloomberg’s dominance is attributable to underwriters’ unwillingness to distribute reference data to all market participants because widespread distribution increases the risk of inaccuracies. Accordingly, data vendors “must expend substantial time and effort gathering information from multiple sources . . . ultimately resulting in an unnecessary market inefficiency.” Id. at 286–87. This hinders timely market-wide participation in corporate bond trading.
Next, the Commission assuaged concerns about the burden of additional reporting requirements on underwriters. The Commission observed that the impact of FINRA’s proposal on how underwriters currently distribute data or how data vendors conduct business is uncertain, but underwriters “already have . . . data reporting processes in place and have incurred the costs of establishing those processes.” Id. at 298. Accordingly, underwriters would have the choice to continue providing new issue reference data to data vendors as they see fit.
The Commission then turned to the proposal’s general burden on competition. The Commission conceded that “the impact on competition is uncertain,” but nevertheless concluded that “any burden on competition would both be limited and justified by the evidence in the record demonstrating an information asymmetry that can disadvantage many market participants due to the lack of 14 timely access to basic information that is important for the identification, valuation and settlement of newly issued corporate bonds.” Id. at 282. Further, FINRA’s proposal would not require market participants to purchase data from FINRA. The Commission noted that even Bloomberg acknowledged “that market participants currently demand more reference data fields than FINRA is proposing to collect.” Id. at 298 n.249.
Finally, the Commission addressed concerns about fees associated with FINRA’s proposal as well as the overall cost of the proposed service. The Commission first noted that FINRA cannot charge fees for its data service without the Commission’s authorization, and the Commission would not approve FINRA’s fee proposal absent a showing that it complies with the Exchange Act and the SEC’s rules regarding proposed fee changes. Moreover, proposed fee changes are subject to public notice and comment. Even if FINRA’s fee proposal becomes effective upon filing with the Commission, the Commission has authority to suspend the fee and order proceedings to assess whether the fee proposal is consistent with the Exchange Act.
Further, the Commission disagreed with commenters— including Bloomberg—who suggested that the agency could not adequately assess the proposal’s economic effects without knowing the fees FINRA would charge and the costs FINRA would incur in connection to building the service. In response to such comments, the Commission reasoned that FINRA’s later “fee filing . . . will merit a consideration of FINRA’s cost to build the New Issue Reference Data Service.” J.A. 301. “[T]he costs of the system, which will be better known once the system is built, will be necessary to assess whether FINRA has proposed a fee for that service that is consistent with the Act, including Section 15A(b)(5).” Id. All in all, the 15 Commission reasoned that FINRA’s proposal is justified because it will resolve a widespread corporate bond reference data problem to the benefit of the market and market participants.
On March 15, 2021, Bloomberg filed a timely petition for this Court to review the Commission’s January 2021 order. Bloomberg contends that the Commission’s decision should be vacated as arbitrary and capricious and unsupported by substantial evidence pursuant to the Administrative Procedure Act, 5 U.S.C. § 706(2)(A), because (1) there is insufficient evidence of an actual reference data access problem in the corporate bond market; (2) FINRA’s data service will impede competition in the corporate bond market by imposing undue burdens on underwriters and ousting incumbent data vendors, in violation of Sections 3(f) and 15A(b)(9) of the Exchange Act, 15 U.S.C. §§ 78c(f), 78o–3(b)(9); (3) the Commission failed to conduct a cost-benefit analysis of FINRA’s proposal, in violation of its obligation under Section 3(f) of the Exchange Act, id. § 78c(f); and (4) the Commission improperly allowed FINRA to evade review of the fees associated with its data service, thereby ignoring its obligation under Section 15A(b)(5) of the Act, id. § 78o–3(b)(5).
For the reasons explained below, we find that the Commission’s approval of FINRA’s proposal was arbitrary and capricious under the Administrative Procedure Act, 5 U.S.C. § 706(2)(A), because the Commission failed to respond to significant and relevant concerns Bloomberg raised in its comments objecting to FINRA’s proposal. Specifically, the agency did not provide a reasoned response to Bloomberg’s comments that FINRA failed to quantify the direct and indirect costs of its proposed data service (or explain why certain costs 16 could not be quantified), and failed to explain how the costs incurred for building the service will be paid if the Commission disapproves FINRA’s proposed fee structure in subsequent proceedings. As such, the Commission’s decision to approve FINRA’s proposal was not the product of reasoned decisionmaking.
IV.
This Court has jurisdiction to review the Commission’s January 2021 order pursuant to Section 25(a) of the Exchange Act, 15 U.S.C. § 78y(a). We review the order under the Administrative Procedure Act, which requires us to hold unlawful agency action that is “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law” or that is “unsupported by substantial evidence.” 5 U.S.C. §§ 706(2)(A), (E); Susquehanna Int’l Grp., LLP v. SEC, 866 F.3d 442, 445 (D.C. Cir. 2017) (internal quotation marks omitted). “To satisfy the ‘arbitrary and capricious’ standard, ‘the agency must examine the relevant data and articulate a satisfactory explanation for its action including a rational connection between the facts found and the choice made.’” Susquehanna, 866 F.3d at 445 (quoting Motor Vehicle Mfrs. Ass’n v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43 (1983)).
At the outset, Bloomberg contends that certain documents in the Joint Appendix—namely the transcript and other materials from an October 2018 FIMSAC meeting—are not a part of the official administrative record in this case because the SEC did not list these materials as a separate exhibit in the certified index of the record. Pet’r’s Opening Br. at 29–30; Pet’r’s Reply Br. at 5–6. 17 Bloomberg is right that the SEC neglected to include certain materials in connection to the October 2018 FIMSAC meeting in the administrative record. But the SEC’s apparent error does not justify vacatur of the Commission’s January 2021 order. To be sure, federal courts base their review of agency action on the record before the agency. See 5 U.S.C. § 706. But, in this instance, the SEC’s “omission” of certain FIMSAC documents “is of little consequence, and certainly does not warrant vacatur” because the “parties did include the [omitted materials] in the Joint Appendix, making [them] available for our perusal” and referred to them throughout their briefs. MD Pharm., Inc. v. Drug. Enf. Admin., 133 F.3d 8, 15 (D.C. Cir. 1998) (“The Rules allow parties to supply materials that were improperly omitted from the record.”); see also FED. R. APP. P. 16(b). In short, we will not vacate the order because of the SEC’s omission or disregard the omitted record material in our resolution of this petition.
We turn next to Bloomberg’s substantive challenges to the Commission’s January 2021 order.
First, Bloomberg contends that FINRA—and by extension, the SEC—relied on “anonymous sources, corroborated by complaints from competing data vendors that they struggle to get data from underwriters.” Pet’r’s Opening Br. at 15–16. Bloomberg characterizes FINRA’s outreach effort as biased, resulting in “anecdotes [that] cannot . . . constitute substantial evidence” of reference data access issues. See id. at 16. Rather than discovering evidence that traders or investors lack access to reference data, Bloomberg says FINRA’s “process was simply a mechanism for Bloomberg’s rivals to complain about their challenges.” Id. at 30 (emphasis in original). Further, Bloomberg invokes 18 Susquehanna to bolster its claim that the SEC “uncritically accepted” FINRA’s evidence. Id. (citing Susquehanna, 866 F.3d at 448).
In Susquehanna, a clearing agency proposed to make changes to the fees that exchanges pay to the agency. 866 F.3d at 443–44. The SEC accepted the clearing agency’s capital plan proposal. The SEC conceded in Susquehanna that it relied on an assessment performed by the clearing agency’s consultant; it did not undertake its own analysis. Id. at 446–47. The Court criticized the SEC’s “unquestioning reliance” on the clearing agency’s defense of its own actions and the SEC’s failure to “critically review[] [the clearing agency’s] analysis or perform[] its own [analysis].” Id. at 447. The Court concluded that the SEC failed to ascertain whether the dividend level proposed by the clearing agency was reasonable. Id. Instead, “the SEC took [the clearing agency’s] word for it.” Id. For that reason, the Court determined that the SEC’s approval of the clearing agency’s proposal to change exchange fees was arbitrary and capricious. Id. at 449.
The case before us is nothing like Susquehanna. First, it was FIMSAC—an SEC-backed advisory committee—that independently discovered the existence of a reference data access problem in the market for new issue corporate bond reference data. FIMSAC is comprised of 23 diverse members, including investors, broker-dealers, underwriters, the SEC’s former chief economist, and data vendors; these participants unanimously agreed that participants in the corporate bond market lack timely, uniform, and accurate access to new issue reference data. Moreover, FIMSAC’s recommendation to create a solution preceded FINRA’s undertaking of an Economic Impact Assessment and market participant outreach. The January 2021 order reflects that the SEC found FINRA’s analysis consistent with other evidence, including materials 19 from the October 2018 FIMSAC meeting. Additionally, contrary to Bloomberg’s assertions, the participants in the FIMSAC proceedings are not anonymous; many are identified on the SEC’s website in FIMSAC’s public meeting transcripts and subcommittee meeting minutes.
All in all, by suggesting that the Commission’s decision rests on speculation that a reference data access problem exists, Bloomberg overlooks substantial evidence submitted by FINRA—not merely anecdotes from supposedly biased market participants—that there are information asymmetries and inefficiencies in the market for new issue corporate bond reference data. For all these reasons, Bloomberg’s argument— that the SEC merely took FINRA’s word for the existence of a reference data access problem—does not hold water.
The parties also dispute whether FINRA’s proposal offends Section 15A(b)(9) of the Exchange Act, which provides that FINRA’s rules should “not impose any burden on competition not necessary or appropriate in furtherance of the purposes of” the Exchange Act, 15 U.S.C. § 78o–3(b)(9), and Section 3(f), which requires the SEC to “consider . . . whether the action will promote efficiency, competition, and capital formation” in its review of FINRA’s proposed rule. Id. § 78c(f).
Bloomberg contends that FINRA’s data service constitutes “government-compelled data collection” that is inconsistent with Section 15A(b)(9), 15 U.S.C. § 78o–3(b)(9), because it will burden competition by “nationaliz[ing] corporate-bond reference data, replacing multiple private enterprises with a single regulatory utility.” Pet’r’s Opening Br. at 16, 31. In Bloomberg’s view, the Commission downplayed the risk of 20 creating a “quasi-governmental monopoly [that] will replace competitive private providers,” which rendered the Commission’s January 2021 order arbitrary and capricious. See id. at 22. Further, Bloomberg says FINRA’s proposed amendment to Rule 6760(b) will require underwriters to provide data to the FINRA-mandated database, thereby disincentivizing underwriters from giving data to private vendors. See id. at 21–22. Finally, Bloomberg suggests that implementation of FINRA’s proposed rule change could render data service vendors unviable. Id. at 34.
In its response brief, the SEC contends that the Commission’s approval of FINRA’s proposal is consistent with Section 15A(b)(9), 15 U.S.C. § 78o–3(b)(9), and Section 3(f), id. § 78c(f), because any limited burden on competition is outweighed “by the benefits of alleviating the information asymmetry that disadvantages many market participants.” Resp.’s Br. at 34. Moreover, the SEC reiterates that FINRA’s proposal will promote market participation and suggests that the Exchange Act “does not limit its focus on competition to one narrow sector of the market; it focuses on the impact of competition in the market as a whole.” Id. Furthermore, the SEC does not share Bloomberg’s concern that FINRA’s data service—which will only require underwriters to report core reference data—will oust private data vendors like Bloomberg. Id. at 37. Finally, the SEC points out that FIMSAC’s recommendation to establish a new issue reference data service was based on the success of a similarly centralized, self- regulatory organization-mandated service in the municipal bond market, where competition continues to exist. See id. at 17. For these reasons, the SEC contends that FINRA’s data service will foster competition, not burden it.
The SEC’s argument is compelling, and it accurately encapsulates the Commission’s requirements under the 21 Exchange Act. The Exchange Act obligates the Commission to consider the effects of proposed rules and regulations on the market as a whole, not just narrow sections of the market. For example, in Business Roundtable v. Securities & Exchange Commission, this Court emphasized the Commission’s “unique obligation to consider the effect of a new rule upon ‘efficiency, competition, and capital formation’” as well as its economic consequences. 647 F.3d 1144, 1148 (D.C. Cir. 2011) (quoting 15 U.S.C. § 78c(f)). Failure to fulfill this obligation “makes promulgation of the rule arbitrary and capricious and not in accordance with law.” Id.; see also Chamber of Com. v. SEC, 412 F.3d 133, 143 (D.C. Cir. 2005) (“the Commission [has a] statutory obligation to determine as best as it can the economic implications of the rule it has proposed”); New York Stock Exch. LLC v. SEC, 962 F.3d 541, 559 (D.C. Cir. 2020).
All in all, Bloomberg’s argument that FINRA’s data service will impose an unnecessary burden on competition lacks merit. First, Bloomberg mischaracterizes FINRA’s proposed data service as a competitive endeavor that will displace incumbent data vendors, and not a supplementary service that will foster competition and improve efficiency and timeliness in the reference data market. Also, Bloomberg overlooks the existence of underwriters’ existing data reporting processes, which make continued data collection “less burdensome than if new processes had to be established.” Resp.’s Br. at 40.
Moreover, Bloomberg’s own comments during the agency proceedings undermine its arguments here. As the Commission mentioned in its January 2021 order, Bloomberg acknowledged that “market participants currently demand more reference data fields than FINRA is proposing to collect.” J.A. 298 n.249. In other words, FINRA’s proposal will facilitate additional competition by imposing data reporting 22 requirements on underwriters, which will in turn level the playing field for private data providers who face a reference data access problem. Underwriters would still be free to provide a broader scope of reference data elements to data vendors, and in turn, data vendors can meet the market’s demand for more comprehensive data.
To be sure, the Commission conceded in its January 2021 order that “the impact [of FINRA’s proposal] on competition is uncertain” and dependent on market participants’ response. J.A. 282. Nevertheless, given the narrow scope of FINRA’s proposed amendment to Rule 6760(b) and significant evidence of a potentially positive impact on competition in the corporate bond market, including input from diverse market participants with expertise in the corporate bond market, we hold that the Commission concluded appropriately that “the limited set of data proposed to be reported and disseminated to allow for the identification, valuation and settlement of new issue corporate bonds is unlikely to supplant the demand for a more comprehensive reference database.” J.A. 298–99. At its core, the Administrative Procedure Act’s substantial evidence standard “requires more than a scintilla, but can be satisfied by something less than a preponderance of the evidence.” Epsilon Elecs., Inc. v. U.S. Dep’t of Treasury, 857 F.3d 913, 925 (D.C. Cir. 2017) (internal quotation marks and citations omitted). “If that threshold is met, we must uphold the agency’s judgment regarding the relevant facts. . . .” Id.
In sum, on the record before us, we hold that the Commission had substantial evidence to support its determination that FINRA’s proposal is consistent with Sections 3(f) and 15A(b)(9) of the Exchange Act, 15 U.S.C. §§ 78c(f), 78o–3(b)(9), because it will impose only a limited, justifiable burden on competition. 23
D.
We next address Bloomberg’s final two arguments—that the Commission failed to weigh the proposed rule’s costs and benefits, in violation of Section 3(f) of the Exchange Act, 15 U.S.C. § 78c(f), and arbitrarily allowed FINRA to evade review of the fee component of its data service, in violation of Section 15A(b)(5) of the Exchange Act, which requires the Commission to assess whether FINRA’s proposed rule “provide[s] for the equitable allocation of reasonable . . . fees.” 15 U.S.C. § 78o–3(b)(5).
First, we reject Bloomberg’s argument that the Commission arbitrarily allowed FINRA to evade review of the fee component of its data service, in violation of Section 15A(b)(5) of the Exchange Act. See id. As mentioned, FINRA’s proposed fee is subject to public notice and comment and the Commission retains the authority to suspend FINRA’s fee-based service and order proceedings to assess whether the fee proposal is consistent with the Exchange Act. In its order, the Commission acknowledged its obligation to ensure that FINRA’s pending fee proposal is consistent with the Exchange Act. See J.A. 301. In sum, the SEC will consider fees in a future rulemaking, and the agency reasonably concluded that it has control over those fees and will not approve them unless they are reasonable and equitable such that they are consistent with the Exchange Act.
We turn next to Bloomberg’s argument that the Commission’s decision to approve FINRA’s proposal was arbitrary and capricious because the Commission failed to consider the costs FINRA will incur in building the system and the extent to which FINRA will pass along those costs to market participants, especially if FINRA’s data service is 24 unreasonably expensive. See Pet’r’s Opening Br. at 42. In Bloomberg’s view, Section 3(f) of the Exchange Act, 15 U.S.C. § 78c(f), obligated the Commission to evaluate the costs and benefits of FINRA’s proposed data service, but the Commission abdicated this responsibility. See Pet’r’s Opening Br. at 35 (citing Chamber of Com., 412 F.3d at 143, and Bus. Roundtable, 647 F.3d at 1149).
Although Bloomberg has raised the issue of whether Section 3(f) of the Exchange Act, 15 U.S.C. § 78c(f), imposed on the Commission a requirement to include costs incurred by FINRA in a cost-benefit analysis of FINRA’s proposed rule, we need not reach that question because the Commission’s failure to respond adequately to Bloomberg’s comments rendered its decision arbitrary and capricious under the Administrative Procedure Act. Bloomberg raised relevant concerns about the direct and indirect costs of FINRA’s proposal, but the Commission brushed them aside. As such, the Commission’s decision was not the product of reasoned decisionmaking.
“Under the APA, whenever agencies promulgate a rule that intends to create new law, rights, or duties . . . they must engage in a process known as notice-and-comment rulemaking.” Oakbrook Land Holdings, LLC v. Comm’r of Internal Revenue, 28 F.4th 700, 710 (6th Cir. 2022) (cleaned up); see also 5 U.S.C. § 553(b). Pursuant to this process, “an agency must respond to comments ‘that can be thought to challenge a fundamental premise’ underlying the proposed agency decision.” Carlson v. Postal Regul. Comm’n, 938 F.3d 337, 344 (D.C. Cir. 2019) (quoting MCI WorldCom, Inc. v. FCC, 209 F.3d 760, 765 (D.C. Cir. 2000)); see also City of Waukesha v. EPA, 320 F.3d 228, 257 (D.C. Cir. 2003) (“The agency need not address every comment, but it must respond in a reasoned manner to those that raise significant problems.”) 25 (internal quotation marks and citation omitted). Indeed, “[t]he requirement that agency action not be arbitrary or capricious includes a requirement that the agency adequately explain its result and respond to relevant and significant public comments.” Pub. Citizen, Inc. v. FAA, 988 F.2d 186, 197 (D.C. Cir. 1993) (internal quotation marks and citations omitted). In sum, “[a]n agency’s response to public comments . . . must be sufficient to enable the courts ‘to see what major issues of policy were ventilated … and why the agency reacted to them as it did.’” Carlson, 938 F.3d at 344 (quoting Del. Dep’t of Nat. Res. & Env’tl Control v. EPA, 785 F.3d 1,17 (D.C. Cir. 2015)).
As mentioned, during notice-and-comment rulemaking, Bloomberg “raised concerns regarding FINRA’s costs to build and operate the new reference data service.” J.A. 301. In response, the Commission reasoned that if FINRA “hypothetically build[s] a New Issue Reference Data Service at a high cost that would be unreasonable to pass on to end- users,” then “FINRA would not be able to make a showing that any such fees proposed to be assessed on the basis of its cost to build the service are reasonable, as required by Section 15A(b)(5) of the Act.” Id.; 15 U.S.C. § 78o–3(b)(5). “In such a case . . . the Commission would suspend and disapprove the proposal.” J.A. 301.
The Commission’s analysis overlooks a key problem: if FINRA’s data service ends up being unreasonably expensive, then the agency cannot protect market participants from footing the bill for it at the fees stage. To be sure, the Commission is right that it could suspend and disapprove FINRA’s proposal at the fees stage, see id., but at that point, FINRA will have already incurred the financial burden of building the service. That cost—which could be millions, or even tens of millions, of dollars—must be paid by someone, whether the subscribers 26 of the service or the broker-dealers who make up FINRA. In short, the Commission approved FINRA’s proposal without responding to comments that urged it to assess not only the financial impact of the service on FINRA, but also the entities that fund FINRA. That is not reasoned decisionmaking.
All in all, if public comments raise relevant and significant concerns about the costs associated with a proposed rule, then the agency should provide a reasoned response to those comments. See, e.g., Ne. Md. Waste Disposal Auth. v. EPA, 358 F.3d 936, 950 (D.C. Cir. 2004) (per curiam) (holding that the EPA responded adequately to members of the municipal waste combustor industry who submitted comments complaining about the high costs associated with regulating pollutants produced by municipal waste combustor units). In this case, the Commission’s failure to respond to relevant and significant comments about the direct and indirect costs of FINRA’s proposed data service was sufficient to render its decision arbitrary and capricious.
Furthermore, we find that the appropriate remedy for the agency’s error is to remand the Commission’s order approving the proposal without vacating the order. “The decision to vacate depends on two factors: the likelihood that ‘deficiencies’ in an order can be redressed on remand, even if the agency reaches the same result, and the ‘disruptive consequences’ of vacatur.” Black Oak Energy, LLC v. FERC, 725 F.3d 230, 244 (D.C. Cir. 2013) (quoting Allied-Signal v. Nuclear Regul. Comm’n, 988 F.2d 146, 150–51 (D.C. Cir. 1993)). In this case, both factors weigh against vacatur. See Vecinos para el Bienestar de la Comunidad Costera v. FERC, 6 F.4th 1321, 1331–32 (D.C. Cir. 2021).
First, we find that on remand, “the Commission can redress its failure of explanation” by analyzing the costs 27 FINRA will incur in building and maintaining its data service and how the costs of building the data service will be remunerated if the fee proposal is ultimately disapproved by the Commission. See id. at 1332. Second, we find that vacatur of the order would “needlessly disrupt” the Commission and FINRA’s efforts to address market inefficiencies resulting from untimely, inconsistent, and inaccurate collection and dissemination of new issue reference data in the corporate bond market. See id.
In sum, we find that the Commission’s approval of FINRA’s proposed reference data service was arbitrary and capricious in one respect: the Commission failed to respond adequately to Bloomberg’s concerns about the cost of building and maintaining the program and the extent to which those costs—which could conceivably amount to millions, or tens of millions, of dollars—will be borne by market participants. As such, the Commission violated the Administrative Procedure Act and failed to engage in reasoned decisionmaking. In this regard, Bloomberg’s petition for review is granted.
Bloomberg’s remaining arguments lack merit. Therefore, Bloomberg’s petition for review is otherwise denied. For the foregoing reasons, we remand to the Commission without vacatur for further proceedings consistent with this opinion.
So ordered.