United States Court of Appeals FOR THE DISTRICT OF COLUMBIA CIRCUIT
Argued May 15, 2025 Decided October 14, 2025
No. 24-1350
CBOE GLOBAL MARKETS, INC., ET AL., PETITIONERS
v.
SECURITIES AND EXCHANGE COMMISSION, RESPONDENT
On Petition for Review of a Final Rule of the Securities and Exchange Commission
Amir C. Tayrani argued the cause for petitioners. With him on the briefs were Paul E. Greenwalt III, Michael K. Molzberger, Alex Gesch, and Cameron J.E. Pritchett.
Kevin Carroll and Daniel J. Feith were on the brief for amicus curiae Securities Industry and Financial Markets Association (SIFMA) in support of petitioners.
Daniel E. Matro, Senior Appellate Counsel, U.S. Securities and Exchange Commission, argued the cause for respondent. With him on the brief were Tracey A. Hardin, 2 Solicitor, and Brooke J. Wagner, Appellate Counsel. Dominick V. Freda, Assistant General Counsel, entered an appearance.
Stephen W. Hall was on the brief for amicus curiae Better Markets, Inc. in support of respondent.
William M. Jay was on the brief for amicus curiae Law Professor J.W. Verret in support of respondent.
Keith Bradley, Kayla Marie Mendez, and Kathryn M. Brown were on the brief for amici curiae We the Investors and Urvin Finance, Inc. in support of respondent.
Thomas A. Burns was on the brief for amici curiae Healthy Markets Association, et al. in support of respondent.
Before: SRINIVASAN, Chief Judge, MILLETT and GARCIA, Circuit Judges. Opinion for the Court filed by Circuit Judge GARCIA. GARCIA, Circuit Judge: In 2005, the Securities and Exchange Commission capped the fees that national securities exchanges could charge investors for accessing their services. In 2024, the SEC lowered that cap. Several exchanges petitioned for review, arguing that the SEC exceeded its statutory authority and acted arbitrarily or capriciously. We disagree and deny the petition for review. I The Securities Exchange Act of 1934 established the SEC and charged it with overseeing and regulating the securities industry. See Pub. L. No. 73–291, § 4, 48 Stat. 881, 885. In 1975, Congress amended the Act to expand the SEC’s regulatory authority. See Pub. L. No. 94-29, 89 Stat. 97. Among other things, the amendments tasked the SEC with 3 establishing a “national market system for securities.” 15 U.S.C. § 78k-1(a)(2). Congress envisioned that new data- processing and communications technologies would connect securities markets nationwide, in ways that would boost efficiency and competition. Id. § 78k-1(a)(1). It gave the Commission “intentionally broad and clear power,” plus “discretion,” to shape the development of that system. Bradford Nat’l Clearing Corp. v. SEC, 590 F.2d 1085, 1095 (D.C. Cir. 1978) (cleaned up). A few basics are in order about how the system works today. Exchanges provide markets where investors can buy securities from other investors. See All. for Fair Bd. Recruitment v. SEC, 125 F.4th 159, 169 (5th Cir. 2024). Investors submit orders to buy and sell stocks. Exchanges, in turn, rank, display, and match those orders. Some investors participate by posting bids to buy (or offers to sell) a security. 89 Fed. Reg. 81620, 81623 (Oct. 8, 2024). In doing so, these investors create opportunities for trading—they “make” liquidity in the market. Id. Other investors then “take” liquidity by accepting these standing bids to buy (or offers to sell). Id. Exchanges coordinate and regulate this activity. Most exchanges charge liquidity takers a fee for executing transactions. The exchanges then use most of the fee to pay a rebate to liquidity makers, leaving the remainder (the “net capture”) for the exchange to keep as revenue. Id. at 81645.1 In 2005, the SEC sought to “modernize and strengthen the national market system” for securities. 70 Fed. Reg. 37496, 37496 (June 29, 2005). To that end, the Commission adopted a series of initiatives called Regulation NMS (short for National Market System). Three aspects of Regulation NMS
1 Some exchanges follow an inverted “taker-maker” model, under which liquidity makers pay a fee while liquidity takers receive a rebate. 89 Fed. Reg. 81620, 81645 n.368 (Oct. 8, 2024). 4 are relevant here. One measure prevented exchanges from executing orders at prices worse than those displayed at another exchange (a phenomenon the Commission referred to as “trade-throughs”). Id. at 37501; see 17 C.F.R. § 242.611. Another measure required exchanges to display securities prices in minimum pricing increments (or “tick sizes,” in industry-speak). That requirement prevents investors from outbidding each other by posting trivially higher prices (for instance, by beating out a $2 bid to buy with a $2.000001 bid). 70 Fed. Reg. at 37503; see 17 C.F.R. § 242.612. Most important for our purposes, a third rule limited the fees that exchanges could charge investors for executing orders. The rule capped fees at 30 mils ($0.003) per share for stocks priced at or above $1. And it capped fees at 0.3% of the quotation price for stocks priced below $1. Id. at 37502–03 & n.36; see 17 C.F.R. § 242.610(c) (2005). These caps, the SEC explained, would mitigate manipulative fee practices and reduce the gap between a security’s “published” price and its “true” price. 70 Fed. Reg. at 37545.2 In the ensuing years, modern pricing and fee practices “attracted considerable attention and generated significant debate.” 84 Fed. Reg. 5202, 5202–03 (Feb. 20, 2019). Some commentators insisted that predominant fee arrangements had harmfully distorted national securities markets. 83 Fed. Reg. 13008, 13010–11 (Mar. 26, 2018). Others maintained that those fee models had “positive effects” on the national market
2 Technically, Regulation NMS applies to all “trading centers,” which include not only exchanges but also entities such as “alternative trading system[s].” 17 C.F.R. § 242.600(b)(106). In the final rule at issue here, however, the SEC explains that “exchanges are the only trading centers that have quotations that are subject to the access fee caps” that petitioners challenge. 89 Fed. Reg. at 81623 n.40. Like the parties, we therefore refer to exchanges for simplicity. 5 system. Id. at 13011. And still others urged the SEC to gather more data to assess those market effects. Id. In 2019, the SEC responded by adopting a pilot program seeking to “gather data” on the effect that regulatory fee caps and rebates have on “market quality and execution quality.” 84 Fed. Reg. at 5203. This initiative, the SEC hoped, would help it decide “whether any changes in the current regulatory framework [were] appropriate.” Id. The pilot would sort stocks into three groups: (1) a control group subject to the prevailing 30-mil cap, (2) a test group subject to a lower, 10- mil cap, and (3) another test group subject to the prevailing 30- mil cap and a ban on rebates paid by exchanges. Id. The program would run for two years, covering stocks with average daily trading volumes of at least 30,000 shares that were priced at or above $2 per share and that did not close below $1 per share. Id. After several exchanges petitioned for review, in 2020, our court set aside the pilot program. The SEC, we held, had no authority to promulgate a “one-off” regulation “merely to secure information that might indicate . . . whether there is a problem worthy of regulation.” N.Y. Stock Exch. LLC v. SEC, 962 F.3d 541, 546 (D.C. Cir. 2020). In 2022, the SEC undertook another rulemaking effort. This time, instead of a pilot study, the Commission proposed concrete changes to Regulation NMS. 87 Fed. Reg. 80266, 80266 (Dec. 29, 2022) Those changes included reductions to both the minimum tick sizes and the access-fee caps. After considering comments, in 2024, the SEC unanimously adopted the proposed rules (with certain modifications). The SEC reduced the minimum tick sizes, anticipating that smaller ticks would improve price competition. 89 Fed. Reg. at 81632. The Commission also lowered the cap on access fees—from 30 mils to 10 mils for 6 stocks priced at or above $1, and from 0.3% of the price quotation to 0.1% for stocks priced below $1. The Commission explained that this reduction went hand in hand with its decision to reduce minimum tick sizes: “[A]n access fee that is too high when compared to the tick size can create pricing distortions.” Id. at 81624. The SEC also concluded that lower fees would address various market problems “that have developed under the access fee caps.” Id. at 81651. Several exchanges petitioned our court to review the access-fee amendment. See 15 U.S.C. § 78y(b). The SEC then agreed to stay the amendment pending review. In re Mot. for Stay of Effect of Amends. to Rules 610 & 612 of Regul. NMS, Exchange Act Release No. 101899, 2024 WL 5094955, at *1– 2 (Dec. 12, 2024). II Petitioners seek to set aside the access-fee amendment as beyond the SEC’s statutory authority and, alternatively, as arbitrary and capricious. We disagree on both fronts. A We start with petitioners’ claim that the SEC exceeded its legal authority in setting an across-the-board access-fee cap. To assess whether the Commission “acted within its statutory authority,” we must exercise our own “independent judgment,” using every interpretive tool at our disposal “to determine the best reading of the statute.” Loper Bright Enters. v. Raimondo, 603 U.S. 369, 373, 400, 412 (2024). “When the best reading of a statute is that it delegates discretionary authority to an agency,” our role is to ensure the agency has acted within “the boundaries of the delegated authority.” Id. at 395 (cleaned up). With that approach in mind, we conclude that the SEC acted within its statutory authority. 7 To support its new rule, the SEC relies on several statutory provisions. Section 11A of the Act directs the Commission to “facilitate the establishment of a national market system for securities,” “in accordance with the findings and to carry out the objectives set forth” in the statute. 15 U.S.C. § 78k-1(a)(2). Those objectives include the “economically efficient execution of securities transactions” and “fair competition” among market actors. Id. § 78k-1(a)(1)(C). Section 11A also authorizes the SEC to prescribe “rules and regulations” that “assure” the “fairness and usefulness of the form and content” of “information with respect to quotations for and transactions in” securities. Id. § 78k-1(c)(1)(B). Section 23 of the Act then gives the SEC the “power to make such rules and regulations as may be necessary or appropriate to implement the provisions” of the statute. Id. § 78w(a)(1). Those grants of power confer “broad, discretionary” authority to regulate the national market system. Nasdaq Stock Mkt. LLC v. SEC, 38 F.4th 1126, 1131 (D.C. Cir. 2022) (quoting S. Rep. 94-75, at 7 (1975)). Through the 1975 amendments, Congress “explicitly endeavored to leave the Commission substantial flexibility of choice in boldly and effectively accomplishing the herculean task of rapidly restructuring an entire industry.” Bradford Nat’l, 590 F.2d at 1104 (cleaned up). In keeping with that mandate, the amendments gave the SEC broad license to “flesh out the contours of the NMS” and “ensure that exchanges have rules designed to promote the NMS policy that SEC develops.” All. for Fair Bd. Recruitment, 125 F.4th at 177. This broad regulatory authority includes the power to set a universal access-fee cap. The SEC determined that an access- fee cap would promote the “fairness and usefulness” of price- quotation “information.” Fee caps, the Commission reasoned, help ensure that a security’s displayed price does not stray too far from its true price. 70 Fed. Reg. at 37545. The Commission 8 refers to this concept as “price transparency” and explained that it promotes the “usefulness” of a security’s posted price. See 89 Fed. Reg. at 81646–49; see also id. at 81644. The SEC also determined that an access-fee cap was a “necessary” and “appropriate” means of maintaining an “efficient” and “compet[itive]” national market system. 89 Fed. Reg. at 81644–45, 49. The Commission’s “trade-through protections,” for instance, aim to promote efficiency and fair competition by funneling investors to the exchanges displaying the best prices. 70 Fed. Reg. at 37584. A fee cap prevents those exchanges from imposing “exorbitant fees” on investors who are “required to access” their posted prices. 89 Fed. Reg. at 81656; 70 Fed. Reg. at 37503. Looking just at Sections 11A and 23, then, imposing an access-fee cap seems to fall well within the SEC’s regulatory authority. Petitioners resist this conclusion in two main ways. They argue that (1) the SEC cannot cap access fees at all and (2) to the extent it can, the Commission must do so on an exchange- by-exchange basis. Neither argument persuades. Start with petitioners’ claim that the Act does not authorize the SEC to cap access fees at all. Taking Section 11A on its own terms, petitioners argue that the provision permits only the regulation of quotation “information,” 15 U.S.C. § 78k-1(c)(1)(B), not the regulation of fees. But Section 11A is not so limited: The provision empowers the SEC to “assure” the “fairness and usefulness” of quotation “information.” Id. And as just described, the SEC has discretion to determine that capping access fees is a “necessary” or “appropriate” means of ensuring the “fairness and usefulness” of quotation “information.” Petitioners next insist that statutory context forecloses this reading of the SEC’s regulatory authority. Congress, petitioners emphasize, included the term “fees” in certain parts 9 of the statute but omitted it from the provisions that the SEC has invoked. For example, Section 6(b) directs the SEC to ensure that each exchange’s rules “provide for the equitable allocation of reasonable dues, fees, and other charges.” Id. § 78f(b)(4). And Section 6(e) permits the SEC to approve or abrogate exchange rules “impos[ing] a schedule or fix[ing] rates of commissions, allowances, discounts, or other fees to be charged by its members for effecting transactions.” Id. § 78f(e). Yet Section 11A does not mention “fees” at all. Petitioners argue that the SEC may regulate fees only in the specific ways set out in Section 6, invoking the rule that “‘[w]hen Congress includes particular language in one section of a statute but omits it in another section of the same Act,’ we generally take the choice to be deliberate.” Bartenwerfer v. Buckley, 598 U.S. 69, 78 (2023) (quoting Badgerow v. Walters, 596 U.S. 1, 11 (2022)). That rule, however, is “not absolute”—and here, we are hesitant “to read much into” the Act’s selective use of the term “fees.” Id. For one thing, the omission of “fees” is not the “sole difference” between the provisions the SEC invoked and the provisions petitioners cite. City of Columbus v. Ours Garage & Wrecker Serv., Inc., 536 U.S. 424, 435–36 (2002). The provisions “use different words,” speak to different contexts, and “are not otherwise parallel.” Nat’l Postal Pol’y Council v. Postal Regul. Comm’n, 17 F.4th 1184, 1191 (D.C. Cir. 2021). Neither Section 6(b) nor Section 6(e), for instance, refers to the SEC’s authority to “assure . . . the fairness and usefulness” of “information with respect to quotations for and transactions in” securities. 15 U.S.C. § 78k-1(c)(1)(B). Those differences weaken the “presumption” that the presence of “fees” in one part of the statute and its absence in another “reveals Congress’ design.” Ours Garage, 536 U.S. at 436. That presumption grows even weaker in statutes where Congress presumably “left to reasonable agency discretion 10 questions that it has not directly resolved,” especially through “a broad grant of authority contained within the same statutory scheme.” Adirondack Med. Ctr. v. Sebelius, 740 F.3d 692, 697 (D.C. Cir. 2014) (quoting Cheney R.R. Co. v. ICC, 902 F.2d 66, 68–69 (D.C. Cir. 1990)); see also Creekstone Farms Premium Beef, L.L.C. v. Dep’t of Agric., 539 F.3d 492, 500 (D.C. Cir. 2008). Congress did just that here. The Act’s provisions, as we have explained, give the SEC broad discretion to regulate the national market system. And Section 23 in particular confers especially broad regulatory authority, empowering the SEC “to make such rules and regulations as may be necessary or appropriate to implement” the statute’s provisions. 15 U.S.C. § 78w(a)(1). Petitioners also briefly suggest that the SEC cannot claim statutory powers “absent clear congressional authorization,” citing Loper Bright Enterprises v. Raimondo, 603 U.S. 369 (2024). That is not what Loper Bright said. It instructed courts to “exercise their independent judgment in deciding whether an agency has acted within its statutory authority,” mindful that Congress can and sometimes does “confer discretionary authority on agencies.” Id. at 404, 412. That imperative does not require agencies to point to “clear congressional authorization” for all exercises of authority. Cf. West Virginia v. EPA, 597 U.S. 697, 723 (2022) (requiring “clear congressional authorization” for “extraordinary” assertions of regulatory authority). Next, petitioners argue that even if the SEC can cap exchange fees, it cannot do so on an industry-wide basis. Sections 19(b) and (c) of the Act, petitioners argue, set forth special procedures for reviewing and overriding exchange rules on an individualized basis. See 15 U.S.C. § 78s(b), (c). That “specific” authority to regulate rules (including fees) on an exchange-by-exchange basis, as petitioners see it, displaces any “general” authority to regulate fees across the board. 11 Neither provision, however, limits the SEC’s authority in the way petitioners urge. Section 19(b) allows the SEC to review “any proposed rule change” submitted by individual self-regulatory organizations (or SROs, a category that includes exchanges). Id. § 78s(b). It also describes the special procedures that the SEC must follow when conducting such review. See id. That case-by-case assessment of SRO-specific rule changes co-exists comfortably with the SEC’s authority to set generally applicable default rules or regulatory boundaries. There is, in other words, no conflict for us to resolve between a provision that empowers the SEC to (reactively) review proposals by a specific exchange and one that empowers the SEC to (proactively) set regulatory policy for all exchanges. Cf. Am. Hosp. Ass’n v. NLRB, 499 U.S. 606, 612 (1991) (“[E]ven if a statutory scheme requires individualized determinations, the decisionmaker has the authority to rely on rulemaking to resolve certain issues of general applicability unless Congress clearly expresses an intent to withhold that authority.”). Section 19(c) does not help petitioners either. True, Section 19(c) specifically provides that the SEC may “abrogate, add to, and delete from” the rules of an SRO. 15 U.S.C. § 78s(c). True also, that provision outlines a special notice-and-comment process by which the Commission must amend SRO rules. See id. Yet at the same time that it describes one way the SEC may regulate the industry, Section 19(c) explicitly states that it should not be interpreted to affect other regulatory authorities. That provision contains a savings clause clarifying that “[n]othing in this subsection shall be construed to impair or limit the Commission’s power to make . . . rules and regulations pursuant to any other authority under this chapter.” Id. § 78s(c)(4)(B). Lest there be any doubt, the clause adds that Section 19(c) does not “modify or alter the procedures the Commission may follow” in making rules under these other sources of authority. Id. 12 Petitioners’ exchange-by-exchange theory of regulation also proves too much. It would compel the SEC to accomplish through multiple, separate rulemakings what it could do in one, forcing “the agency continually to relitigate issues that may be established fairly and efficiently in a single rulemaking proceeding.” Heckler v. Campbell, 461 U.S. 458, 467 (1983). And it would require the SEC to do so for a wide range of regulatory matters. Section 19 governs SEC supervision over any SRO rules, not just rules pertaining to fees. See 15 U.S.C. § 78s. So, if Section 19 compelled the SEC to proceed on a case-by-case basis anytime it seeks to regulate fees, then Section 19 would also compel the SEC to proceed on a case- by-case basis anytime it wished to implement any policy that could conflict with an exchange’s own rules. Given the broad text of Sections 11A and 23, we doubt that Congress intended to so severely constrain the SEC’s rulemaking powers. B Next, petitioners argue that the access-fee amendment is arbitrary and capricious. We assess that claim by asking whether the SEC “reasonably considered the relevant issues and reasonably explained the decision.” FCC v. Prometheus Radio Project, 592 U.S. 414, 423 (2021). Our review under this standard is generally “deferential.” Id. It is “particularly deferential” in a case like this, which “implicate[s] competing policy choices, technical expertise, and predictive market judgments.” AD HOC Telecomms. Users Comm. v. FCC, 572 F.3d 903, 908 (D.C. Cir. 2009). Petitioners claim that the SEC acted arbitrarily and capriciously in two main ways. First, they argue that the SEC changed its position on the market effects of fees without adequate explanation. In petitioners’ view, the Commission was unwilling to assess those effects in the 2019 pilot, but it was willing to do so in the new rule yet failed to explain this 13 purported change. Second, petitioners argue that the SEC failed to reasonably explain its decision to lower the access-fee cap to 10 mils for stocks priced at or above $1 per share. We reject both challenges. 1 The SEC did not change its position between the pilot and this rule without adequate explanation. Agencies may change their positions “as long as they provide a reasoned explanation for the change.” Encino Motorcars, LLC v. Navarro, 579 U.S. 211, 221 (2016). In explaining the change, the agency need not show “to a court’s satisfaction that the reasons for [its new position] are better than the reasons for the old one.” FCC v. Fox Television Stations, Inc., 556 U.S. 502, 515 (2009). It “suffices” that “there are good reasons” for the new position and that “the agency believes” the new position to be “better.” Id. At the same time, the agency must still consider all “relevant data.” Id. at 513 (quoting Motor Vehicles Mfrs. Ass’n of United States, Inc. v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43 (1983)). It cannot “disregard contrary or inconvenient factual determinations that it made in the past, any more than it can ignore inconvenient facts when it writes on a blank slate.” Mozilla Corp. v. FCC, 940 F.3d 1, 55 (D.C. Cir. 2019) (cleaned up). Nor can an agency ignore “serious reliance interests” that its prior position has engendered. Encino, 579 U.S. at 222 (cleaned up). Petitioners argue that the SEC violated these principles. In the 2019 pilot, the SEC expressed uncertainty about the market effects of access fees. The SEC acknowledged the “fundamental disagreement” among commentators over the issue. 84 Fed. Reg. at 5238. But it concluded that it lacked the “data” necessary to “evaluate” their “competing claims.” Id. 14 As a result, the SEC explained, it could not “predict whether investors will face higher or lower transaction costs” because of the pilot. Id. at 5277. Nor was it “clear,” the Commission added, how the program would affect trading volume on exchanges. Id. at 5283. Then, in the rulemaking under review here, the SEC seemed to take a different view, reaching definitive conclusions about the new rule’s market effects. The Commission predicted, for instance, that reducing access fees would not increase transaction costs at all. See, e.g., 89 Fed. Reg at 81656–57, 81736–37, 81767. And the SEC claimed that “liquidity providers would not be deterred from quoting on exchange.” Id. at 81657. Petitioners contend that, under the change-in-position doctrine, the Commission needed, but failed, to adequately explain this shift in its thinking. We disagree. At the outset, we have serious doubts that the change-in-position doctrine applies at all here. For one thing, the SEC’s progression from the pilot to the new rule reflects more continuity than inconsistency. In the pilot, the SEC claimed that it needed to gather more information before it could properly assess the market effects of access fees. But even after the pilot program fell through, the SEC proceeded to collect relevant information. Between 2019 and 2024, commentators continued to study the market effects of fee-and-rebate pricing models, and the SEC solicited their views as part of its new rulemaking effort. See, e.g., id. at 81737; see also Tr. of Oral Arg. at 20–21, 28–29. The Commission then incorporated this new data into its analysis of the new rule, using it as part of its definitive assessment of the rule’s effects on various market dynamics. See, e.g., 89 Fed. Reg. at 81698–99, 81737. We are skeptical that the change-in-position doctrine speaks to “change” of this sort—when an agency approaches a policy problem by expressing interest in gathering more 15 information, solicits and obtains new information, and only then takes a definitive view on the matter. That sequence of events seems to describe a single, natural course of rational agency decision-making. It does not reflect any “inconsistency” in an agency’s thinking. FDA v. Wages & White Lion Invs., L.L.C., 604 U.S. 542, 581 (2025) (cleaned up). Nor does it raise any of the traditional concerns implicated by the change-in-position doctrine—such as the risk that the agency has “misle[d] regulated entities,” id. at 567, or the possibility that the agency has ignored “serious reliance interests” in its decision-making process, Fox, 556 U.S. at 515. Another fundamental difference between the situations the SEC confronted in 2019 and 2024 also gives us pause about applying the change-in-position doctrine. In 2019, the SEC felt no compulsion to change the access-fee caps—that was why the Commission felt free to proceed with a pilot program in the first place. But circumstances changed in 2024. As part of its new rulemaking effort, recall, the SEC also reduced the minimum tick sizes for certain stocks. This policy decision, all agree, required a corresponding reduction in the access-fee cap. See 89 Fed. Reg. at 81651. Reducing the initial tick sizes without also lowering the initial fee caps would create serious pricing distortions. Id. at 81651, 81733. So once the SEC decided to reduce the minimum tick size, the Commission could no longer afford to take a wait-and-study approach to the fee caps; it needed to adjust those caps based on the information it possessed at the time. The SEC was (repeatedly) explicit on this point: The fee caps were lowered to “accommodate the change in tick sizes.” Id. at 81621; see also, e.g., id. at 81624, 81732. And as part of that effort, the SEC took another hard look at the market effects of fees—this time with the added benefit of new data—and made its decision. That context, too, explains the SEC’s initial tentativeness in 2019 and its choice to take a more definitive stance in 2024. The SEC did not “change” its position on the same issue—whether and how 16 access-fee caps should be adjusted—so much as it resolved a different issue—how access-fee caps should be adjusted given a change in the tick sizes. We doubt that the change-in-position doctrine required the SEC to formally acknowledge these obvious differences in context and further explain why it took action in the final rule. But even setting those doubts aside and assuming that the SEC did “change” its position in the relevant sense, the SEC did not violate the change-in-position doctrine. The Commission offered “good reasons” for its current views of fees’ market effects. Fox, 556 U.S. at 515. And through it all, the SEC did not “ignore” any contrary factual premises that may have motivated its initial uncertainty. Id. To illustrate the point, consider the three specific market effects that, in petitioners’ view, exemplify the SEC’s shifting views: (1) the effect of access fees on transaction costs, (2) the effect of fees on order flow, and (3) the effect of fees on broker incentives. Transaction costs. In the new rule, the SEC considered whether lowering access fees would increase the overall cost of accessing liquidity. The Commission offered “good reasons” to think that lowering fees would not increase the transaction costs that investors face. Id. at 515. It acknowledged that lowering fees would raise the sticker price of liquidity for some securities. As the SEC explained: Exchanges use the fees they charge to fund the rebates they pay out to liquidity makers. So lowering fees would likely reduce rebates. 89 Fed. Reg. at 81736. To compensate for those rebate reductions, liquidity makers would in turn raise the price of liquidity—by, for example, raising the posted price at which they would offer to sell a security. At the same time, however, liquidity takers would also pay less in the form of lower fees. Id. at 81687, 81736–37, 81767. Those effects—higher market 17 prices but lower fees—would roughly offset one another: Rebates would fall by however much fees fell. And market prices would rise by however much rebates fell. Id. at 81738. As part of its analysis, the SEC grappled with the reasoning that motivated its initial uncertainty in 2019. In the pilot, petitioners note, the SEC tentatively acknowledged that reducing fees might increase transaction costs (for at least some stocks) by discouraging liquidity makers from posting liquidity, prompting “a reduction in quoted depth.” 84 Fed. Reg. at 5277. The SEC explicitly addressed that possibility in the new rule. It predicted that lower fees would not discourage liquidity makers from posting liquidity on exchanges in the first place—liquidity makers could maintain their current revenues simply by raising the sticker price of liquidity. See 89 Fed. Reg. at 81685, 81737, 81765. And in the aggregate, the SEC concluded, the total quantity of liquidity supplied would not change. Id. at 81686–87. To support its position, the SEC relied on basic supply-and-demand principles. Id. And to reinforce the point, the SEC pointed to recent empirical data that was “directionally consistent” with its economic model. Id. at 81737. Petitioners suggest that the SEC needed to, at a minimum, expressly “acknowledge” that it indicated uncertainty on this subject in the 2019 pilot. It is true that “the requirement that an agency provide reasoned explanation for its action would ordinarily demand that it display awareness that it is changing position.” Fox, 556 U.S. at 515. But this “ordinarily” applicable requirement does not apply robotically in every circumstance. An agency that changes position need show only that “there are good reasons” for its new position and that “the agency believes” the new position to be better. Id. An explicit and “conscious change of course” is one way of signaling that belief. Id. But it is not the only way. An explanation that considers and rejects an alternative position, even without 18 expressly acknowledging that the position was once the agency’s own, can likewise express the agency’s belief that its current position is “better.” We see no reason to require a formal display of awareness in the unique circumstances of this case. The SEC has already acknowledged the disagreement among commenters over transaction costs. See 89 Fed. Reg. at 81685 n.1010. It has already reasonably explained why it favors one side of the debate over the other. And it has already reasonably explained why it disagrees with any contrary premises that it might have embraced in the pilot. In doing so, the SEC has examined all the “relevant data” and articulated a “satisfactory explanation” for its action. Fox, 556 U.S. at 513 (quoting State Farm, 463 U.S. at 43). The APA does not impose some “magic words” acknowledgment requirement where, as here, the agency grappled directly and comprehensively with the contrary point of view in a reasoned way and there is doubt about whether the agency even broke from a past position in the ordinary sense. In these unusual circumstances, the agency’s analysis is reasonable and sufficient even without a more explicit acknowledgment that the agency once held that contrary view. Cf. Physicians for Soc. Resp. v. Wheeler, 956 F.3d 634, 646 (D.C. Cir. 2020) (emphasizing that “the rule that an agency must display awareness that it is changing position is simply a species of the more general requirement—present in all APA cases—that an agency provide a reasoned explanation for its action” (cleaned up)).3
3 Of course, an agency must consider any “serious reliance interests” its prior policy may have engendered. Encino, 579 U.S. at 222 (cleaned up). And it would presumably be difficult (if not impossible) for an agency to consider such reliance interests without also acknowledging that it previously adopted a different policy. Here, however, petitioners have not argued that the SEC’s initial uncertainty engendered any reliance interests (which is unsurprising, 19 Reinforcing the point, our decisions remanding agency action because of “unexplained inconsistencies” across proceedings typically involve situations where the agency said nothing at all about contrary points it raised before. See, e.g., id. at 647; U.S. Sugar Corp. v. EPA, 830 F.3d 579, 651 (D.C. Cir. 2016) (per curiam); Dist. Hosp. Partners, L.P. v. Burwell, 786 F.3d 46, 59 (D.C. Cir. 2015). This case is different. The SEC left no mystery as to how it might respond to any “contrary or inconvenient factual determinations that it made in the past.” Mozilla, 940 F.3d at 55. Order flow. The same goes for the SEC’s views on the effect of access fees on order flow—that is, the amount of trading activity that occurs on exchanges relative to non- exchange venues. The SEC provided good reasons to conclude that lower fees would not drive trading activity off exchanges, drawing on the same analysis outlined above. The Commission also grappled with the reasons that had prompted its initial uncertainty in 2019. In the pilot, petitioners note, the SEC stated that the program’s effects on venue competition were “unclear.” 84 Fed. Reg. at 5282. Those effects, the SEC concluded, would depend on whether the pilot discourages liquidity makers from providing liquidity on exchanges. Id. But, as we have already discussed, the SEC explicitly addressed that issue in the new rule, thoroughly explaining why it did not think the new rule would decrease the quantity of liquidity provided on exchanges. See, e.g., 89 Fed. Reg. at 81738.
as the SEC statements petitioners focus on were embedded in a rule that simply proposed an information-gathering pilot program). In these unique circumstances, the Commission was able to offer a coherent explanation of its current views without explicitly acknowledging its prior uncertainty. 20 Petitioners argue that the SEC failed to offer even good reasons for its current view. They claim that the Commission “minimized empirical evidence” showing that lower access fees would harm exchanges competitively. But petitioners have failed to identify which specific pieces of evidence the SEC discounted. Their brief also refers to “fundamental differences” and “regulatory disparities between exchanges and off-exchange venues.” Petitioners’ Brief 39–40. The SEC addressed those concerns head-on, explaining why the amendment would not alter that competitive landscape. The net costs of trading on an exchange, the SEC concluded, would generally remain unchanged. See 89 Fed. Reg. at 81657, 81738. So market participants would continue to face “the same” “underlying economic tradeoff” in choosing where to transact. Id. at 81738. Petitioners also insist that the SEC acted unreasonably by resting its order-flow predictions on “conjecture” and “theoretical supply-and-demand analysis” instead of “concrete data.” Petitioners’ Brief 39. But “it is perfectly legitimate for the Commission to base its findings on basic economic theory,” including by relying on “generic factual predictions, as long as the agency explains and applies the relevant economic principles in a reasonable manner.” Xcel Energy Servs. Inc. v. FERC, 41 F.4th 548, 560–61 (D.C. Cir. 2022) (cleaned up). The SEC did so here. Broker incentives. That leaves the SEC’s views on the effects of access fees (and rebates) on broker incentives— specifically, the possibility that brokers will route orders to chase higher rebates or avoid higher fees, even if those routing decisions come at the expense of their clients. A broker might, for instance, choose to route an order to a venue offering market-high rebates, even though a different venue provided higher trade-execution rates and lower transaction costs. 21 The SEC did not change its position at all on this issue: In the pilot, the SEC acknowledged the “theoretical” possibility that transaction-based fee models could create “potential conflicts of interest” between brokers and their clients. 84 Fed. Reg. at 5247 (emphasis added). It likewise acknowledged that lowering fees could mitigate these potential conflicts. Id. at 5277–78. Still, the SEC concluded that theory could take it only so far. The Commission could not “determine from existing empirical evidence the impact, if any, of exchange transaction fee models on order routing decisions by broker- dealers or on market and execution quality.” Id. at 5244; see also, e.g., id. at 5253, 5277–78. In this rulemaking, the SEC reiterated those same points. The Commission (again) acknowledged that existing fee-and-rebate arrangements might create “potential conflict[s] of interest between broker-dealers and their customers with respect to broker-dealer order routing.” 89 Fed. Reg. at 81624 (emphasis added); see also, e.g., id. at 81646, 81682, 81741–42, 81761. And it (again) noted that lower fees could “alleviate” these “potential conflicts of interest.” Id. at 81647 (emphasis added). Nevertheless, the SEC (again) concluded that “there is uncertainty regarding to what degree those potential conflicts of interest are being acted upon.” Id. at 81742; see also, e.g., id. at 81682 n.986, 81761, 81768. We see no shift in the SEC’s thinking on this issue. Petitioners also challenge two other aspects of the SEC’s discussion of broker incentives. First, petitioners assert that the SEC acted unreasonably by failing to identify “any real-world examples” of broker-dealers acting on potential conflicts of interest. But, again, an agency “need not—indeed cannot— base its every action upon empirical data.” Nasdaq, 38 F.4th at 1142; see also Am. Great Lakes Ports Ass’n v. Schultz, 962 F.3d 510, 516 (D.C. Cir. 2020) (concluding that the APA “imposes no general obligation on agencies to produce empirical evidence”). The SEC was “entitled to conduct a 22 general analysis based on informed conjecture.” Nasdaq, 38 F.4th at 1142 (cleaned up). It “highlighted a plausible conflict of interest” based on “its experience” and “industry comments.” Id. Second, petitioners argue that the SEC acted arbitrarily by failing to consider other regulatory measures that could mitigate potential conflicts of interest. This rule, however, addresses different aspects of the access-fee cap issue: accommodating the change in minimum tick size, price transparency, and so on. The policy measures that petitioners highlight do not plausibly tackle these issues. The SEC did not need to consider a host of policy alternatives that target a different and unproven potential problem. See Nat’l Shooting Sports Found., Inc. v. Jones, 716 F.3d 200, 215 (D.C. Cir. 2013) (requiring agencies to “consider only ‘significant and viable’ and ‘obvious’” policy alternatives (collecting cases)). 2 Petitioners next challenge the SEC’s decision to set the new access-fee cap at 10 mils for stocks priced at or above $1 per share. We disagree with their arguments on this point too. The SEC settled on a 10-mil cap after “consider[ing] and balanc[ing] many factors.” 89 Fed. Reg. at 81660. Three key considerations framed the Commission’s analysis: It sought to (1) “mitigate the market distortions” that had developed under the preexisting caps, (2) “accommodate the new minimum pricing increments” set by the tick-size amendment, and (3) “preserv[e] the viability of the agency market business model.” Id. Those aims set the following constraints: The cap would need to stay below 25 mils (50% of the new minimum pricing increment) to “preserve price coherence.” Id. at 81768. And the cap would need to stay above 2 mils (the current net capture rate) to avoid “creat[ing] strain on exchange business models.” Id. at 81767. The lower the cap, moreover, the more the 23 amendment would reduce market distortions. See, e.g., id. at 81759 n.1730. Weighing these factors, the SEC concluded that setting the cap at 10 mils “strikes an appropriate balance.” Id. at 81659. That figure, the Commission also noted, fell within the range of rates charged by alternative trading systems (or ATSs) for trade execution services. Id. at 81659 & n.612; see also id. at 81661, 81698, 81735. The SEC “considered the relevant issues” and “reasonably explained” its decision-making. Prometheus, 592 U.S. at 423. We have no basis to second-guess the Commission’s choice, especially given agencies’ “wide discretion” in making “line- drawing” decisions. Nat’l Shooting Sports, 716 F.3d at 214. Petitioners raise a host of challenges to the SEC’s analysis. Preliminarily, petitioners insist that we owe no deference to the SEC’s line-drawing determinations after Loper Bright. Loper Bright, however, recognized that Congress can and does “confer discretionary authority on agencies,” 603 U.S. at 404, including the authority to draw lines when lines must be drawn. Indeed, the Supreme Court recently reiterated that courts should defer to an agency’s reasonably explained line-drawing choices that “involve[] primarily issues of fact”—matters that an agency “is better equipped to assess . . . than a court is.” Seven Cnty. Infrastructure Coal. v. Eagle County, 145 S. Ct. 1497, 1512 (2025); see also Loper Bright, 603 U.S. at 392 (“[The APA] does mandate that judicial review of agency policymaking and factfinding be deferential.”). The question whether to cap access fees at 10 mils or 15 or some other number certainly fits that description. Petitioners also insist that the SEC was required to identify empirical evidence to support its line-drawing conclusions about the 10-mil cap. No case or principle requires such evidence. Agency leeway in this context flows from the reality that Congress vested agencies—not courts—with the authority 24 to make discretionary policy judgments. See id. at 1512–13. That discretion does not turn on the agency’s ability to procure specific types of data in a given case. Cf. Am. Great Lakes Ports Ass’n, 962 F.3d at 516. Next, petitioners attack one of the many factors the SEC considered in reaching its conclusion, namely the SEC’s observation that a 10-mil cap “would be consistent with the access fees charged by ATSs.” 89 Fed. Reg. at 81661. Petitioners claim that exchanges need to be able to charge higher fees (and promise higher rebates) than ATSs do, to offset inherent competitive disparities between exchanges and ATSs. But, as noted above, the SEC explained why it believed that exchanges did not need to rely on rebates to attract trading activity. See, e.g., id. at 81656. And the SEC had no obligation to set a fee cap guaranteed to mitigate ATSs’ competitive advantages, especially given the agency’s countervailing policy objectives. See id. at 81662 (explaining that the level of ATS rates was “not the only factor” in the SEC’s “balancing” analysis). We may not substitute petitioners’ preferred “policy judgment for that of the agency.” Prometheus, 592 U.S. at 423; see also AD HOC Telecomms., 572 F.3d at 908 (emphasizing our “particularly deferential” review in cases that “implicate competing policy choices”). Petitioners also argue that the 10-mil figure does not accurately reflect the fees that ATSs actually charge. The SEC reasonably explained otherwise, relying on the submissions of several commenters. 89 Fed. Reg. at 81661 & n.651. Petitioners counter that one commenter failed to cite his sources and that another commenter misinterpreted an underlying source. But petitioners omit the many other sources that commenters relied on, including disclosures from several ATSs, feedback from various exchange members, and reports from a leading brokerage firm, all of which support the SEC’s conclusion that the 10-mil figure accurately reflects ATSs’ 25 fees. See J.A. 563; see also 89 Fed. Reg. at 81698. Petitioners insist that the SEC could have acted on even better data by directly obtaining this information from ATSs. The APA, however, “imposes no general obligation on agencies to conduct or commission their own empirical or statistical studies.” Prometheus, 592 U.S. at 427. Petitioners also doubt that a 10-mil cap will permit exchanges to maintain the same “net capture rates,” arguing that current net capture rates are both hard to assess and likely to change once exchanges adjust to the lower caps. The SEC reasonably addressed both points. The SEC acknowledged the uncertainty in measuring net capture rates. Still, the Commission concluded, based on several comments and “staff conversations with industry members,” that the current figure was “likely close to 2 mils on most exchanges.” 89 Fed. Reg. at 81696 & n.1103. It was not arbitrary for the SEC to “make tough choices about which of the competing estimates is most plausible” or to at least “hazard a guess as to which is correct.” Bus. Roundtable v. SEC, 647 F.3d 1144, 1150 (D.C. Cir. 2011). The SEC likewise acknowledged “uncertainty” over whether this capture rate would “persist” post-amendment. 89 Fed. Reg. at 81733. The Commission reasonably concluded that it would. A 2-mil capture rate, it explained, would “remain possible under the adopted amendments.” Id.; see also id. at 81656, 81658. A 10-mil fee gives exchanges enough room to pay out a rebate and keep 2 mils for themselves. That assumption, moreover, comports with the SEC’s general view that lowering access fees would not drastically alter the competitive landscape. Petitioners offer no basis to second- guess the SEC’s predictive judgment. See Sorenson Commc’ns, LLC v. FCC, 897 F.3d 214, 230 (D.C. Cir. 2018). 26 Next, petitioners dispute the SEC’s occasional assertions that the fee-cap amendment would preserve the ability of exchanges to “innovate.” The SEC reasonably explained this point too. The SEC anticipated that lowering the fee cap for stocks priced at or above $1 would have no effect on transaction revenue; exchanges could continue to retain a 2-mil net capture rate going forward. 89 Fed. Reg. at 81661, 81767. At the same time, the Commission acknowledged that lowering the fee cap for stocks priced below $1 would reduce transaction revenue. For those stocks, exchanges “typically” do not pay out rebates, so the entire fee comprises the exchange’s “net capture.” Id. at 81734. Further, the SEC acknowledged that this drop in revenue—totaling around $55 million per year, id.—“could impact exchange investment in new technologies,” id. at 81740. Contrary to petitioners’ assertion, those statements are not inconsistent; the SEC identified one dynamic that may reduce exchange revenue, but predicted that securing the same net capture rate for stocks priced at or above $1 would allow for continued investment and innovation. Further, the SEC noted that other parts of the rule could “offset[]” any negative revenue effects by, for instance, boosting trading volume on exchanges. Id. Finally, petitioners insist that the SEC did not meaningfully explore alternative fee arrangements— specifically, a multi-tiered fee model that would apply a 15-mil cap to half-penny-tick stocks priced at or above $1. But the Commission considered that very proposal (and others), raising concerns that a higher fee would increase transaction costs and worsen price transparency. Id. at 81766–68. We have no reason to second-guess that exercise of judgment. III The petition for review is denied. So ordered.